Talk:BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

From Lojban
Jump to navigation Jump to search

Posted by brablonau on Sat 07 of Jan., 2006 17:30 GMT posts: 80 Does anybody have a body of example uses which were actually used for creating the gismu?

Lacking the ability to read the minds of those who created the gismu list (several years ago, at that), it would be nice to have access to some historical documentation which may elucidate what went on in the process.

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

Re: Re: BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

arj Posted by arj on Sat 07 of Jan., 2006 18:34 GMT posts: 953 > Does anybody have a body of example uses which were actually used for creating the gismu? > > Lacking the ability to read the minds of those who created the gismu list (several years ago, at that), it would be nice to have access to some historical documentation which may elucidate what went on in the process. >

Ask the mailing list.

--arj

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

Re: BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Wed 09 of Aug., 2006 00:18 GMT posts: 14214 I have an actual proposal. I expect this to be mighty controversial, but _someone's_ got to do it.

I propose that the BPFK's Offiicial Word be as follows:

All paranthetical or square bracketed remarks in gismu definitions are descriptive rather than proscriptive. These are divided into two basic types:

1. Square bracketed comments, which are normally intended to narrow the scope of definition of a word as a whole, as in:

x1 is a physical page/leaf of book/document/bound mass of pages x2

These give a sensible default, and warn you that if you are talking about something that doesn't match what is in square brackets you may confuse your listener somewhat. Having said that, the BPFK specifically asserts that a document that exists only in a word processor is a papri (or possibly a se papri, depending on how the document is organized).

2. Parenthized comments, which are normally intended to narrow the scope of individual places in the gismu, as in:

x1 (event) modifies/alters/changes/transforms/converts x2 into x3

x1 (physical object) is a list/catalog/register of sequence/set x2 in order x3 in medium x4

Here, again, these give a sensible default, and warn you that confusion may result if you violate these suggestions. For things like "(physical object)", no more force exists than for square bracketed comments discussed above. For parenthetical comments that specifically call for a particular abstraction, not using that abstraction is considered to be implicit sumti raising (that is, should be treated by a non-hostile listener as though you had put tu'a in front of the sumti you actually used), but the BPFK specifically states that this is OK, as long as you're willing to have your listener come back with a question about implicit abstraction. The BPFK, therefore, specifically asserts that "mi" can "galfi" and a shopping list jotted on a virtual post-it on your desktop is a "liste".

To make this work, we really should go through and crop as many of the abstraction places as we think we can live with out. IMO, they should only be present in places where they are really needed, like ckaji, unless someone can come up with a coherent system.

-Robin

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

Re: BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Eppcott Posted by Eppcott on Wed 09 of Aug., 2006 01:51 GMT posts: 4740 That's great, Robin! I'm in support of it. -epkat Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 09 of Aug., 2006 14:08 GMT My only problem with this is the auto-sumti-raising. Kinda destroys usage like "ko senva lo melbi" vs "ko senva tu'a lo melbi".

-- Theodore Reed (treed/bancus) www.surreality.us


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Wed 09 of Aug., 2006 18:08 GMT posts: 14214 On Wed, Aug 09, 2006 at 07:23:30AM -0700, Theodore Reed wrote: > My only problem with this is the auto-sumti-raising. Kinda > destroys usage like "ko senva lo melbi" vs "ko senva tu'a lo > melbi".

senva itself seems rather confused on this point:

x1 dreams about/that x2 (fact/idea/event/state);

IIRC, your intended distinction is "Dream beautiful dreams" vs. "Dream about beautiful things", yeah? I can certainly think of other ways to make that distinction, but I don't think that really addresses your point.

It seems to me that either we have to allow auto sumti raising, or we need to clean up an awful lot of definitions that have abstractions required for no readily apparent or useful reason. But the problem of sumti without abstractions in them that semantically map to abstractions is one I had not considered.

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 09 of Aug., 2006 18:23 GMT On 8/9/06, Robin Lee Powell wrote: > On Wed, Aug 09, 2006 at 07:23:30AM -0700, Theodore Reed wrote: > > My only problem with this is the auto-sumti-raising. Kinda > > destroys usage like "ko senva lo melbi" vs "ko senva tu'a lo > > melbi". > > senva itself seems rather confused on this point: > > x1 dreams about/that x2 (fact/idea/event/state);

I don't see much confusion. Explain?

> IIRC, your intended distinction is "Dream beautiful dreams" vs. > "Dream about beautiful things", yeah? I can certainly think of > other ways to make that distinction, but I don't think that really > addresses your point.

That is the intended distinction, yes. The idea that an abstraction can be "a beautiful thing" and that there for lo melbi can go into a place normally required to be an abstraction.

> It seems to me that either we have to allow auto sumti raising, or > we need to clean up an awful lot of definitions that have > abstractions required for no readily apparent or useful reason. But > the problem of sumti without abstractions in them that semantically > map to abstractions is one I had not considered.

Honestly, I think just rewriting definitions is the long-term simplest route. The difficulty is getting the work done.

-- Theodore Reed (treed/bancus) www.surreality.us


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Wed 09 of Aug., 2006 18:28 GMT posts: 14214 On Wed, Aug 09, 2006 at 11:34:48AM -0700, Theodore Reed wrote: > On 8/9/06, Robin Lee Powell wrote: > >On Wed, Aug 09, 2006 at 07:23:30AM -0700, Theodore Reed wrote: > >> My only problem with this is the auto-sumti-raising. Kinda > >> destroys usage like "ko senva lo melbi" vs "ko senva tu'a lo > >> melbi". > > > >senva itself seems rather confused on this point: > > > >x1 dreams about/that x2 (fact/idea/event/state); > > I don't see much confusion. Explain?

There are 4 different abstractions mentioned; this is pretty unusual. Possibly even unique.

> >IIRC, your intended distinction is "Dream beautiful dreams" vs. > >"Dream about beautiful things", yeah? I can certainly think of > >other ways to make that distinction, but I don't think that > >really addresses your point. > > That is the intended distinction, yes. The idea that an > abstraction can be "a beautiful thing" and that there for lo melbi > can go into a place normally required to be an abstraction. > > >It seems to me that either we have to allow auto sumti raising, > >or we need to clean up an awful lot of definitions that have > >abstractions required for no readily apparent or useful reason. > >But the problem of sumti without abstractions in them that > >semantically map to abstractions is one I had not considered. > > Honestly, I think just rewriting definitions is the long-term > simplest route. The difficulty is getting the work done.

I still think we should say something about non-abstract sumti in abstract place, especially when they can't semantically map to abstractions (like "le gerku"). At least something to the effect of "Your listener may take this as implicit sumti raising, or may just get confused.".

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Wed 09 of Aug., 2006 19:03 GMT posts: 2388 At a guess (based on past cases), the problematic abstractions are the ones put in to solve the opaque context problems. In that case, the number of them should be increased rather than diminished, since some unmarked cases of opaque contexts remain (mainly where English subject raising has been given a separate place in Lojban — very malglico indeed). For the rest, Lojban speakers just need to overcome their sloppy English (etc.) prejudices and move over to what is required by Logic (what the language is supposed to show up, after all) and see that what we desire (etc.) are events not the things in them. And so on. And then be consistent about it. Or, of course, copy English and just scrap the pitiful remains of a claim to be related to Logic.



> On Wed, Aug 09, 2006 at 11:34:48AM -0700, Theodore Reed wrote: > > On 8/9/06, Robin Lee Powell wrote: > > >On Wed, Aug 09, 2006 at 07:23:30AM -0700, Theodore Reed wrote: > > >> My only problem with this is the auto-sumti-raising. Kinda > > >> destroys usage like "ko senva lo melbi" vs "ko senva tu'a lo > > >> melbi". > > > > > >senva itself seems rather confused on this point: > > > > > >x1 dreams about/that x2 (fact/idea/event/state); > > > > I don't see much confusion. Explain? > > There are 4 different abstractions mentioned; this is pretty > unusual. Possibly even unique. > > > >IIRC, your intended distinction is "Dream beautiful dreams" vs. > > >"Dream about beautiful things", yeah? I can certainly think of > > >other ways to make that distinction, but I don't think that > > >really addresses your point. > > > > That is the intended distinction, yes. The idea that an > > abstraction can be "a beautiful thing" and that there for lo melbi > > can go into a place normally required to be an abstraction. > > > > >It seems to me that either we have to allow auto sumti raising, > > >or we need to clean up an awful lot of definitions that have > > >abstractions required for no readily apparent or useful reason. > > >But the problem of sumti without abstractions in them that > > >semantically map to abstractions is one I had not considered. > > > > Honestly, I think just rewriting definitions is the long-term > > simplest route. The difficulty is getting the work done. > > I still think we should say something about non-abstract sumti in > abstract place, especially when they can't semantically map to > abstractions (like "le gerku"). At least something to the effect of > "Your listener may take this as implicit sumti raising, or may just > get confused.". > > -Robin > > > >


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 09 of Aug., 2006 19:08 GMT On 8/9/06, John E Clifford wrote: > At a guess (based on past cases), the problematic abstractions are the ones put in to solve the > opaque context problems. In that case, the number of them should be increased rather than > diminished, since some unmarked cases of opaque contexts remain (mainly where English subject > raising has been given a separate place in Lojban — very malglico indeed). For the rest, Lojban > speakers just need to overcome their sloppy English (etc.) prejudices and move over to what is > required by Logic (what the language is supposed to show up, after all) and see that what we > desire (etc.) are events not the things in them. And so on. And then be consistent about it. > Or, of course, copy English and just scrap the pitiful remains of a claim to be related to Logic.

As much as I tend to disagree with John (when I can understand him, which is rare), I have to agree on this point.

-- Theodore Reed (treed/bancus) www.surreality.us


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Wed 09 of Aug., 2006 19:58 GMT posts: 14214 On Wed, Aug 09, 2006 at 12:23:28PM -0700, Theodore Reed wrote: > On 8/9/06, John E Clifford wrote: > >At a guess (based on past cases), the problematic abstractions > >are the ones put in to solve the opaque context problems. In > >that case, the number of them should be increased rather than > >diminished, since some unmarked cases of opaque contexts remain > >(mainly where English subject raising has been given a separate > >place in Lojban — very malglico indeed). For the rest, Lojban > >speakers just need to overcome their sloppy English (etc.) > >prejudices and move over to what is required by Logic (what the > >language is supposed to show up, after all) and see that what we > >desire (etc.) are events not the things in them. And so on. And > >then be consistent about it. Or, of course, copy English and just > >scrap the pitiful remains of a claim to be related to Logic. > > As much as I tend to disagree with John (when I can understand > him, which is rare), I have to agree on this point.

I didn't actually understand him, because I don't know what an "opaque context problem" is.

If you guys are proposing increasing the number of abstraction places, though, someone please explain to me what is illogical about "mi galfi le karce". Of course there's an event involved, but there's an event involved in "mi cusku", too, and binxo and cenba and citka. I don't get why galfi should be special here. If, on the other hand, you are proposing that all current places for an agent get turned into event place, you'll need to explain that, too.

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 09 of Aug., 2006 20:18 GMT On 8/9/06, Robin Lee Powell wrote: > On Wed, Aug 09, 2006 at 12:23:28PM -0700, Theodore Reed wrote: > > As much as I tend to disagree with John (when I can understand > > him, which is rare), I have to agree on this point. > > I didn't actually understand him, because I don't know what an > "opaque context problem" is. > > If you guys are proposing increasing the number of abstraction > places, though, someone please explain to me what is illogical about > "mi galfi le karce". Of course there's an event involved, but > there's an event involved in "mi cusku", too, and binxo and cenba > and citka. I don't get why galfi should be special here. If, on > the other hand, you are proposing that all current places for an > agent get turned into event place, you'll need to explain that, too.

I suppose I should have clarified my position. I agree that we should not modify the logic inherent in Lojban simply to make it easier for English (or any other language) speakers. I'm honestly still annoyed by the "le mi broda" construct.

As far as adding more specificity, that's another argument altogether. I think the specificity of each word should probably be debated on its own merits. For instance, I tend to agree that the agent in a galfi-ing is probably more important than the process by which the galfi takes place, but djica or troci really should require an abstraction.

-- Theodore Reed (treed/bancus) www.surreality.us


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 09 of Aug., 2006 20:57 GMT On 8/9/06, Robin Lee Powell wrote: > > I didn't actually understand him, because I don't know what an > "opaque context problem" is.

http://en.wikipedia.org/wiki/Opaque_context "An opaque context is a linguistic context in which it is not possible to substitute co-referential terms while guaranteeing the preservation of truth values."

The idea that abstractions in Lojban "solve" the opaque context "problem" is at the very least debatable. It is true that in general you cannot substitute co-referential terms inside abstractions, and so using an abstraction is one way of creating an opaque context. But how does one jump from there to making it obligatory to use abstractions in places that sometimes might need opacity?

Another related question is whether for example "Mary dreamt of unicorns" involves any opacity at all. In my opinion it does not. It simply means "Mary dreamt of those mythical creatures that look like horses with a horn on their forehead". No problem with substitution of co-referential terms there. It only causes problems for those who confuse universe of discourse, where "unicorns" does have a referent, with the real world, where we know that unicorns don't exist. It also causes problems to those who think generic terms don't have referents, or that they somehow must refer all the way down to specimens, so that they want to go from "Mary dreamt of unicorns" to "there is some unicorn (among all unicorns) such that Mary dreamt of that unicorn". But that's only a problem if you think {lo pavyseljirna} is a quantified term {su'o lo pavyseljirna}, not if you think it's a simple referring term.

And events in Lojban are not special sumti. So if you insist that {la meris pu senva lo pavyseljirna} is {la meris pu senva su'o lo pavyseljirna}, then {la meris pu senva lo nu lo pavyseljirna cu blabi} is {la meris pu senva su'o lo nu lo pavyseljirna cu blabi}, i.e. there is at least one event among all the events of unicorns being white, such that Mary dreamt of that event, which is in no way less problematic that the equivalent thing with plain unicorns.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Wed 09 of Aug., 2006 21:18 GMT posts: 2388 A brief review of "opaque context problems." A principle of logic is that like things are treated alike in their logics. So, in particular, any argument (sumti) should be open to generalization: e.g. {lo broda cu brode} to {da brode} and even {da poi broda cu brode}. Likewise, if two names refer to the same ting, we ought to be able exchange one name for the other, wihtout changing the meaning — and certainly the truth — of the claim. In English, there are situations where this will not work: all kinds of modal situations (which take us to other worlds where what exists here or what are identical here may not), but also just the object (and even the subject) places of various predicates: "believe in," "talk about," "want," "look for," "is a picture of" and a seemingly endless list of others. Logic has had to deal with these situations, since they seem to violate the principle above. There are two possibilities (there may be more but these are what logic typically does) 1) mark certain places as being different from others, so that arguments in them are not subject to the usual rules or 2)insist that the arguments that go there logically are not what they seem to be in English but some abstraction related to these. Lojban — when it remembers to do anything about this at all — has chosen the second approach (more or less), requiring that (except for cases where the principle causes no problem — and there usually are some) the places take abstractions as arguments, rather than the concreta as in English. The point is that the abstractions always exist (a given in Lojban)so the generalizations go through and the identity of abstractions is not dependent upon the identity of things mentioned in them, typically, so the failure of Leibniz's Law is expalined (or rather does not arise). So, if I say that I want to have a dog, then there is something, indeed, the event of my having a dog, that I want. If I just say that I want a dog, in one sense, my wish is granted since there are dogs, but in another sense, there need not be a dog that I want (there might not be any dogs at all or no one of them is such that I can say "That is the one I want."). The problem cases that remain are mainly raisings as when we get from "He believes that Bill is wise" to "He believes of Bill that he is wise," where Bill may not exist or may have an alias that he thinks is stupid but "Bill" is now outside the intensional flag "that." Lojban, alas, follows English in some of these cases. There are also a few cases which both English and Lojban just don't notice, like {prami2}.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Wed 09 of Aug., 2006 21:33 GMT posts: 2388 Yes, xorxes claims to have a different solution to the problem of opaque contexts — or rather a theory in which this problem does not arise. He has, however,refused for the better part of the last decade to say what that theory and how it works and has rejected a number of suggestions that would in fact accomplish at least some of the things he claims to be able to do. In the absdence of an actual spelled out theory, I think we are stuck with working with what we — and Logic generally — have.


> On 8/9/06, Robin Lee Powell wrote: > > > > I didn't actually understand him, because I don't know what an > > "opaque context problem" is. > > http://en.wikipedia.org/wiki/Opaque_context > "An opaque context is a linguistic context in which it is not possible > to substitute co-referential terms while guaranteeing the preservation > of truth values."

Ther is also the problem that particular generalization does not work into these contexts.

> The idea that abstractions in Lojban "solve" the opaque context > "problem" is at the very least debatable. It is true that in general > you cannot substitute co-referential terms inside abstractions, and > so using an abstraction is one way of creating an opaque context. > But how does one jump from there to making it obligatory to use > abstractions in places that sometimes might need opacity?

One doesn't, of course, and one of the problems with the comments is exactly that they are taken as absolutes. Obviously, if you know that what you want exists and which thing it is you want, then there is nothing gained — and something lost — by using an abstraction in place of a simple sumti. But this is — within standard semantics — a trick, since the deep structurte still contains the abstraction but the interpretation rules now add that the referent is trannsparent in this case. The point is that the transparent cases are logically the unusual ones and so get special treatment (which here curiously is to be treated like normal places). If you don't have the assurances, then you need to use the abstraction.

> Another related question is whether for example "Mary dreamt of > unicorns" involves any opacity at all. In my opinion it does not. > It simply means "Mary dreamt of those mythical creatures that > look like horses with a horn on their forehead". No problem with > substitution of co-referential terms there. It only causes problems > for those who confuse universe of discourse, where "unicorns" does > have a referent, with the real world, where we know that unicorns > don't exist. It also causes problems to those who think generic terms > don't have referents, or that they somehow must refer all the way > down to specimens, so that they want to go from "Mary dreamt of > unicorns" to "there is some unicorn (among all unicorns) such that > Mary dreamt of that unicorn". But that's only a problem if you think > {lo pavyseljirna} is a quantified term {su'o lo pavyseljirna}, not if you > think it's a simple referring term.

> And events in Lojban are not special sumti. So if you insist that > {la meris pu senva lo pavyseljirna} is {la meris pu senva su'o lo pavyseljirna}, > then {la meris pu senva lo nu lo pavyseljirna cu blabi} is {la meris pu senva > su'o lo nu lo pavyseljirna cu blabi}, i.e. there is at least one event among > all the events of unicorns being white, such that Mary dreamt of that event, > which is in no way less problematic that the equivalent thing with plain > unicorns. > > mu'o mi'e xorxes > > > >


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Thu 10 of Aug., 2006 00:13 GMT posts: 14214 On Wed, Aug 09, 2006 at 01:30:41PM -0700, Theodore Reed wrote: > On 8/9/06, Robin Lee Powell wrote: > >On Wed, Aug 09, 2006 at 12:23:28PM -0700, Theodore Reed wrote: > >> As much as I tend to disagree with John (when I can understand > >> him, which is rare), I have to agree on this point. > > > >I didn't actually understand him, because I don't know what an > >"opaque context problem" is. > > > >If you guys are proposing increasing the number of abstraction > >places, though, someone please explain to me what is illogical > >about "mi galfi le karce". Of course there's an event involved, > >but there's an event involved in "mi cusku", too, and binxo and > >cenba and citka. I don't get why galfi should be special here. > >If, on the other hand, you are proposing that all current places > >for an agent get turned into event place, you'll need to explain > >that, too. > > I suppose I should have clarified my position. I agree that we > should not modify the logic inherent in Lojban simply to make it > easier for English (or any other language) speakers. I'm honestly > still annoyed by the "le mi broda" construct. > > As far as adding more specificity, that's another argument > altogether. I think the specificity of each word should probably > be debated on its own merits. For instance, I tend to agree that > the agent in a galfi-ing is probably more important than the > process by which the galfi takes place, but djica or troci really > should require an abstraction.

I agree an all points, and am willing to go with "we need to do a bunch of work to clean things up" rather than the implicit sumti-raising solution if that is the BPFK's majority stance.

xorxes? jcowan? Arnt? others?

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 12:54 GMT On 8/9/06, Robin Lee Powell wrote: > On Wed, Aug 09, 2006 at 01:30:41PM -0700, Theodore Reed wrote: > > > > As far as adding more specificity, that's another argument > > altogether. I think the specificity of each word should probably > > be debated on its own merits. For instance, I tend to agree that > > the agent in a galfi-ing is probably more important than the > > process by which the galfi takes place, but djica or troci really > > should require an abstraction. > > I agree an all points, and am willing to go with "we need to do a > bunch of work to clean things up" rather than the implicit > sumti-raising solution if that is the BPFK's majority stance. > > xorxes? jcowan? Arnt? others?

I agree with "we need to do a bunch of work to clean things up". In particular I would like to see an exhaustive list of gismu places that take du'u-abstractions only, an exhaustive list of gismu places that take ka/ni-abstractions only, an exhaustive list of gismu places that take li-abstractions only, an exhaustive list of gismu places that take nu-abstractions only, and an exhaustive list of places that take abstractions of some other kind besides those mentioned, if there is any. These lists are actually shorter than one might imagine, so it's not such a daunting task. That would probably be more fruitful than discussing one gismu at a time, because it would let us see more easily what the reasonable generalizations are, and where the crazy idiosyncrasies lie.

mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

Re: BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Eppcott Posted by Eppcott on Thu 10 of Aug., 2006 14:06 GMT posts: 4740 > I agree with "we need to do a bunch of work to clean things up". > In particular I would like to see an exhaustive list of gismu places > that take du'u-abstractions only, an exhaustive list of gismu places > that take ka/ni-abstractions only, an exhaustive list of gismu places > that take li-abstractions only, an exhaustive list of gismu places > that take nu-abstractions only, and an exhaustive list of places that > take abstractions of some other kind besides those mentioned, if > there is any. These lists are actually shorter than one might imagine, > so it's not such a daunting task. That would probably be more fruitful > than discussing one gismu at a time, because it would let us see more > easily > what the reasonable generalizations are, and where the crazy > idiosyncrasies lie. > > mi'e xorxes

I will do it. -epkat

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 10 of Aug., 2006 14:13 GMT posts: 2388


> On 8/9/06, Robin Lee Powell wrote: > > > > I didn't actually understand him, because I don't know what an > > "opaque context problem" is. > > http://en.wikipedia.org/wiki/Opaque_context > "An opaque context is a linguistic context in which it is not possible > to substitute co-referential terms while guaranteeing the preservation > of truth values." > > The idea that abstractions in Lojban "solve" the opaque context > "problem" is at the very least debatable. It is true that in general > you cannot substitute co-referential terms inside abstractions, and > so using an abstraction is one way of creating an opaque context. > But how does one jump from there to making it obligatory to use > abstractions in places that sometimes might need opacity? > > Another related question is whether for example "Mary dreamt of > unicorns" involves any opacity at all. In my opinion it does not. > It simply means "Mary dreamt of those mythical creatures that > look like horses with a horn on their forehead". No problem with > substitution of co-referential terms there. It only causes problems > for those who confuse universe of discourse, where "unicorns" does > have a referent, with the real world, where we know that unicorns > don't exist. It also causes problems to those who think generic terms > don't have referents, or that they somehow must refer all the way > down to specimens, so that they want to go from "Mary dreamt of > unicorns" to "there is some unicorn (among all unicorns) such that > Mary dreamt of that unicorn". But that's only a problem if you think > {lo pavyseljirna} is a quantified term {su'o lo pavyseljirna}, not if you > think it's a simple referring term.

The move from dreaming of unicorns to dreaming of mythical creatures…forehead is not a case of substitution under identity, for, while the two expressions are in fact coreferential that is not what allows the replacement. The replacement is legitimated rather by the fact that they are synonymous (let’s suppose, so that the replacement is legitimate), that they have the same sense not merely the same reference. Suppose that instead of “the mythical creature… forehead” we replaced “unicorn” with “the critter that was the subject of my daughter’s favorite movie when she was six.” Since that movie was The Last Unicorn, the expressions are coreferential, have the same referent: unicorns, but the substitution is no longer valid: Mary doesn’t know what my daughter’s favorite movie was in 1982 – indeed probably doesn’t know I have a daughter or even that I exist – so she didn’t dream about that (she may not know that unicorns are mythical, for example, and so even the first substitution may not be legitimate).

As for the part about universe of discourse, while xorxes for some reason keeps denying it, the default universe of discourse just is the real world – or a part of it – as understood by the speaker (and, hopefully, the other interlocutors). So, lacking any sign (explicit or implicit) to the contrary, the claim about Mary is made against the standard of the real world. To be sure, in Mary’s dream world there are, apparently, unicorns, but that does not change what there are in the basic universe of discourse here, i.e., the real world. The problem remains: if the subject of Mary’s dream is just a normal sumti then it can be validly generalized upon (a basic rule of logic), but the quantification is now in the real world, the basic domain, and there there are no unicorns. So the generalization is in fact invalid, going from a true premise (suppose) to a false conclusion (the fact that there are no unicorns is not essential here; the case can be made as well if there are the sorts of things involved, since there is even then no guarantee that the thing mMary dreamed about is real). So that subject must to keep the logic — be something other than a normal sumti: either the place must be marked as special (but that is problematic, since some cases in that place are just normal sumti) or the subject must be buried so that is no longer primary. Lojban generally (though not thoroughly) takes the latter tack. I suppose – and this seems to be xorxes’ plan though that is not perfectly clear – that one could also create a sort of entity and have the problem reference be to these sorts of entities that would somehow justify all the moves but would not require subordinating constructions. Of course, they would require somehow indicating that these entities are involved and that would seem to be as much a deviation from ordinary English (the driving force here, it seems) as requiring abstractions. In addition, at least some of the abstractions are clearly actually involved – as the case of wanting a dog shows (you don’t just want a dog, you want to have a dog, i.e, and event not a thing).

Notice, the move involved here does not depend upon taking {lo pavyseljirna} as the same as {su’o lo pavyseljirna}, it merely requires that {lo pavyseljirna} is a sumti and that it means something like what it appears to mean. I suppose that xorxes believes it means something unlike what it appears to mean (the non-specific analog of {le pavyseljirna}) and, instead refers to one of those strange entities discussed in the last prargraph. But, even if this is the case, it does not solve the problem of opaque contexts, for these affect {le pavyseljirna} and {la meris} and indeed {da} as much as indefinite descriptions. Of course, even in the case of {lo pavyseljirna}, the xorxces’ interpretation (insofar as I understand it) creates the problem that indefinite descriptions are now ambiguous: sometimes referring to these special entities, sometimes to what one would expect, the things named. That is, in some circumstances, {lo broda} just refers to brodas (usually ordinary extensional objects), in others it refers to this entity (which has to be intensional as much as abstractions are). And, of course, with extensional objects the move from {lo broda}to {su’o lo broda} is always legitmate.

> And events in Lojban are not special sumti. So if you insist that > {la meris pu senva lo pavyseljirna} is {la meris pu senva su'o lo pavyseljirna}, > then {la meris pu senva lo nu lo pavyseljirna cu blabi} is {la meris pu senva > su'o lo nu lo pavyseljirna cu blabi}, i.e. there is at least one event among > all the events of unicorns being white, such that Mary dreamt of that event, > which is in no way less problematic that the equivalent thing with plain > unicorns.

Actually it is less problematic, since all events exist and, thus, the one that Mary dreamed of does. Admittedly, the claim that all events exist may seem objectionable, but these are events in the abstract, so the universe does not get cluttered by this and, as here, it solves a number of otherwise somewhat intractable problems. Xorxes argument here is quite correct, but no one insists on the premise (that {lo broda} is just {su’o lo broda}) and, even if they did, this has no bearing on the actual problem under discussion (the inference from {la meris senva lo pavyseljirna} to {da poi pavyseljirna go’u la meris senva da}).


> > I agree with "we need to do a bunch of work to clean things up". > In particular I would like to see an exhaustive list of gismu places > that take du'u-abstractions only, an exhaustive list of gismu places > that take ka/ni-abstractions only, an exhaustive list of gismu places > that take li-abstractions only, an exhaustive list of gismu places > that take nu-abstractions only, and an exhaustive list of places that > take abstractions of some other kind besides those mentioned, if > there is any. These lists are actually shorter than one might imagine, > so it's not such a daunting task. That would probably be more fruitful > than discussing one gismu at a time, because it would let us see more > easily what the reasonable generalizations are, and where the crazy > idiosyncrasies lie. > Amen, though I would qualify that “always take” is probably too strong for any place (“typically take” is about right) and add that a list should be made of those places which really need to take such abstractions often enough to be worth mentioning (see cases cited earlier for examples).



Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Thu 10 of Aug., 2006 16:28 GMT posts: 14214 On Thu, Aug 10, 2006 at 07:06:23AM -0700, Eppcott wrote: > Re: BPFK gismu Section: Parenthetical Remarks in Brivla Definitions > > I agree with "we need to do a bunch of work to clean things up". > > In particular I would like to see an exhaustive list of gismu > > places that take du'u-abstractions only, an exhaustive list of > > gismu places that take ka/ni-abstractions only, an exhaustive > > list of gismu places that take li-abstractions only, an > > exhaustive list of gismu places that take nu-abstractions only, > > and an exhaustive list of places that take abstractions of some > > other kind besides those mentioned, if there is any. These lists > > are actually shorter than one might imagine, so it's not such a > > daunting task. That would probably be more fruitful than > > discussing one gismu at a time, because it would let us see more > > easily what the reasonable generalizations are, and where the > > crazy idiosyncrasies lie. > > > > mi'e xorxes > > I will do it.

Don't; you don't have the skill I do at getting computers to produce lists like thing, which means what I can do in minutes will likely take you hours.

As usual, no offense intended; just trying to work to our strengths (as a group). I'll make the list.

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 16:53 GMT On 8/10/06, John E Clifford wrote: > > Suppose that instead of “the mythical creature… forehead” we > replaced “unicorn” with “the critter that was the subject of my daughter’s favorite movie > when she was six.” Since that movie was The Last Unicorn, the expressions are coreferential, > have the same referent: unicorns, but the substitution is no longer valid: Mary doesn’t know > what my daughter’s favorite movie was in 1982 â€" indeed probably doesn’t know I have a > daughter or even that I exist â€" so she didn’t dream about that (she may not know that unicorns > are mythical, for example, and so even the first substitution may not be legitimate).

I can tell your daughter: "Last night Mary dreamt of the critter that was the subject of your favorite movie when you were six" and I'd be telling her the same as if I told her "last night Mary dreamt of unicorns". The fact that Mary would not describe her dream that way is not really relevant here.

> Of course, even in the case of {lo pavyseljirna}, > the xorxces’ interpretation (insofar as I understand it) creates the problem that indefinite > descriptions are now ambiguous: sometimes referring to these special entities, sometimes to what > one would expect, the things named.

They always refer to the things named, not just in some cases. The problem is in the assumption that the only things that can be named are things existing in the real world. That's obviously not the case.

> That is, in some circumstances, {lo broda} just refers to > brodas

Always, not just in some circumstances.

>(usually ordinary extensional objects),

Often, I don't know if usually, but often yes.

> > And events in Lojban are not special sumti. So if you insist that > > {la meris pu senva lo pavyseljirna} is {la meris pu senva su'o lo pavyseljirna}, > > then {la meris pu senva lo nu lo pavyseljirna cu blabi} is {la meris pu senva > > su'o lo nu lo pavyseljirna cu blabi}, i.e. there is at least one event among > > all the events of unicorns being white, such that Mary dreamt of that event, > > which is in no way less problematic that the equivalent thing with plain > > unicorns. > > Actually it is less problematic, since all events exist and, thus, the one that Mary dreamed of > does. Admittedly, the claim that all events exist may seem objectionable, but these are events in > the abstract, so the universe does not get cluttered by this and, as here, it solves a number of > otherwise somewhat intractable problems.

Do you make any distinction between events that take place in the real world and events that don't? What exactly is the force of "exists" in "all events exist"? Why is it any less problematic having these two types of events (assuming you accept that there are these two types) than having other creatures some of which exist and some of which don't exist in the real world?

mu'o mi'e xorxes

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Thu 10 of Aug., 2006 16:58 GMT posts: 14214 On Thu, Aug 10, 2006 at 10:10:12AM -0300, Jorge Llamb?as wrote: > > I agree with "we need to do a bunch of work to clean things up". > In particular I would like to see an exhaustive list of gismu > places that take du'u-abstractions only, an exhaustive list of > gismu places that take ka/ni-abstractions only, an exhaustive list > of gismu places that take li-abstractions only,

There is exactly one of those, interestingly enough:

namcu nac na'u number x1 (li) is a number/quantifier/digit/value/figure (noun); refers to the value and not the symbol

(cf. lerfu, mekso, klani)

> an exhaustive list of gismu places that take nu-abstractions only, > and an exhaustive list of places that take abstractions of some > other kind besides those mentioned, if there is any.

http://teddyb.org/~rlpowell/media/regular/abstractions/

"rest.txt" is all the things with parens in the base definitions not in the other lists; it could use some by-hand weeding, but I'm not sure on what basis.

> These lists are actually shorter than one might imagine,

Actually, I found them rather longer than I imagined.

> so it's not such a daunting task. That would probably be more > fruitful than discussing one gismu at a time, because it would let > us see more easily what the reasonable generalizations are, and > where the crazy idiosyncrasies lie.

Indeed.

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 18:13 GMT On 8/10/06, Robin Lee Powell wrote: > On Thu, Aug 10, 2006 at 10:10:12AM -0300, Jorge Llamb?as wrote: > > > > I agree with "we need to do a bunch of work to clean things up". > > In particular I would like to see an exhaustive list of gismu > > places that take du'u-abstractions only, an exhaustive list of > > gismu places that take ka/ni-abstractions only, an exhaustive list > > of gismu places that take li-abstractions only, > > There is exactly one of those, interestingly enough: > > namcu nac na'u number > x1 (li) is a number/quantifier/digit/value/figure > (noun); refers to the value and not the symbol

I wasn't very clear with my "I would like to see". What I had in mind is that I would like to see the BPFK defining these lists rather than to see what's already marked as such in the gi'uste. For example, in the case of li-abstractions, which are probably the easiest to deal with: all the x2 places of unit gismu (mitre, grake, snidu, etc) can only be filled with li-abstractions, even if the gi'uste doesn't mark this in any explicit way. Also the x1 of {detri} and {tcika}, the x3 of {jbena}, and a couple of oddballs like the x2 of {verba}. But your lists are clearly the right place to start.

ki'e mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 18:28 GMT On 8/10/06, Jorge Llambías wrote: > > I wasn't very clear with my "I would like to see". What I had in mind is > that I would like to see the BPFK defining these lists rather than to see > what's already marked as such in the gi'uste. For example, in the case > of li-abstractions, which are probably the easiest to deal with: all the x2 > places of unit gismu (mitre, grake, snidu, etc) can only be filled with > li-abstractions, even if the gi'uste doesn't mark this in any explicit way. > Also the x1 of {detri} and {tcika}, the x3 of {jbena}, and a couple of > oddballs like the x2 of {verba}. But your lists are clearly the right place > to start.

I'd like to reiterate my misgivings about *requiring* certain abstractions. I can see requiring that something be a "number". But, for instance, couldn't I name a number Bob? (Don't ask why, but figure it as that.) Or, hell say I did some calculations and assigned the result to ty. If the structure *requires* the use of li, that also means that I couldn't use ty or any other variable. (Or la bab.) I think it's much safer to specify what something in a place is, rather than the exact type of abstraction needed. I should absolutely be able to answer "You want what?" with "Something grand.", where grand could be an event, without requiring an abstraction. (Although, in a way, I suppose that is an abstraction, but it wouldn't be permitted with restrictive sumti-typing.)

-- Theodore Reed (treed/bancus) www.surreality.us

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

adamgarrigus Posted by adamgarrigus on Thu 10 of Aug., 2006 18:38 GMT posts: 92 On 8/10/06, Robin Lee Powell wrote: > > On Thu, Aug 10, 2006 at 10:10:12AM -0300, Jorge Llamb?as wrote: > > > > I agree with "we need to do a bunch of work to clean things up". > > In particular I would like to see an exhaustive list of gismu > > places that take du'u-abstractions only, an exhaustive list of > > gismu places that take ka/ni-abstractions only, an exhaustive list > > of gismu places that take li-abstractions only, > > There is exactly one of those, interestingly enough: > > namcu nac na'u number > x1 (li) is a number/quantifier/digit/value/figure > (noun); refers to the value and not the symbol > > (cf. lerfu, mekso, klani) > > > an exhaustive list of gismu places that take nu-abstractions only, > > and an exhaustive list of places that take abstractions of some > > other kind besides those mentioned, if there is any. > > http://teddyb.org/~rlpowell/media/regular/abstractions/ > > "rest.txt" is all the things with parens in the base definitions not > in the other lists; it could use some by-hand weeding, but I'm not > sure on what basis. > > > These lists are actually shorter than one might imagine, > > Actually, I found them rather longer than I imagined. > > > so it's not such a daunting task. That would probably be more > > fruitful than discussing one gismu at a time, because it would let > > us see more easily what the reasonable generalizations are, and > > where the crazy idiosyncrasies lie. > > Indeed.


Some small points:

mikce and katna are in the ka list, mistakenly, I think. gunka is in the ni list, also mistakenly ? I didn't bother to get exhaustive, 'cause maybe the algorithm needs tweaking. mu'o mi'e komfo,amonan


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Thu 10 of Aug., 2006 18:38 GMT posts: 14214 On Thu, Aug 10, 2006 at 02:51:50PM -0400, komfo,amonan wrote: > On 8/10/06, Robin Lee Powell wrote: > > > >On Thu, Aug 10, 2006 at 10:10:12AM -0300, Jorge Llamb?as wrote: > >> > >> I agree with "we need to do a bunch of work to clean things up". > >> In particular I would like to see an exhaustive list of gismu > >> places that take du'u-abstractions only, an exhaustive list of > >> gismu places that take ka/ni-abstractions only, an exhaustive list > >> of gismu places that take li-abstractions only, > > > >There is exactly one of those, interestingly enough: > > > >namcu nac na'u number > > x1 (li) is a number/quantifier/digit/value/figure > > (noun); refers to the value and not the symbol > > > > (cf. lerfu, mekso, klani) > > > >> an exhaustive list of gismu places that take nu-abstractions only, > >> and an exhaustive list of places that take abstractions of some > >> other kind besides those mentioned, if there is any. > > > >http://teddyb.org/~rlpowell/media/regular/abstractions/ > > > >"rest.txt" is all the things with parens in the base definitions not > >in the other lists; it could use some by-hand weeding, but I'm not > >sure on what basis. > > > >> These lists are actually shorter than one might imagine, > > > >Actually, I found them rather longer than I imagined. > > > >> so it's not such a daunting task. That would probably be more > >> fruitful than discussing one gismu at a time, because it would let > >> us see more easily what the reasonable generalizations are, and > >> where the crazy idiosyncrasies lie. > > > >Indeed. > > > Some small points: > * mikce and katna are in the ka list, mistakenly, I think. > * gunka is in the ni list, also mistakenly ?

Fixed.

> I didn't bother to get exhaustive, 'cause maybe the algorithm > needs tweaking.

It's unlikely to be re-run; it was a pretty nasty regex as it was.

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 18:39 GMT On 8/10/06, Theodore Reed wrote:

> I'd like to reiterate my misgivings about *requiring* certain > abstractions. I can see requiring that something be a "number". But, > for instance, couldn't I name a number Bob?

Of course. In that case, "Bob" will be the name of a li-abstraction, there is nothing wrong with that. You can also have li-abstractions as values of variables, and of course {lo namcu} refers to a li-abstraction too. The point is that anything that fills a place that requires a li-abstraction has to refer to a li-abstraction. This is not about the form that the sumti there can have but about the type of value that the sumti there can refer to.

>I should absolutely be able > to answer "You want what?" with "Something grand.", where grand could > be an event, without requiring an abstraction.

(I don't think the x2 of {djica} has to be an event abstraction, but that's a different issue.) Certainly if x2 of djica always required a nu-abstraction that would not mean you could not put {lo fasnu} in it.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

Re: BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Eppcott Posted by Eppcott on Thu 10 of Aug., 2006 18:42 GMT posts: 4740 > On Thu, Aug 10, 2006 at 07:06:23AM -0700, Eppcott wrote: > > > > I will do it. > > Don't; you don't have the skill I do at getting computers to produce > lists like thing, which means what I can do in minutes will likely > take you hours. > > As usual, no offense intended; just trying to work to our strengths > (as a group). I'll make the list. > > -Robin > Point taken. Thanks. -epkat

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 10 of Aug., 2006 18:43 GMT posts: 2388


> On 8/10/06, John E Clifford wrote: > > > > Suppose that instead of “the mythical creature… forehead” we > > replaced “unicorn” with “the critter that was the subject of my > daughter’s favorite movie > > when she was six.” Since that movie was The Last Unicorn, the expressions are > coreferential, > > have the same referent: unicorns, but the substitution is no longer valid: Mary doesn’t > know > > what my daughter’s favorite movie was in 1982 â€" indeed probably doesn’t > know I have a > > daughter or even that I exist â€" so she didn’t dream about that (she may not know > that unicorns > > are mythical, for example, and so even the first substitution may not be legitimate). > > I can tell your daughter: "Last night Mary dreamt of the critter that was the > subject of your favorite movie when you were six" and I'd be telling > her the same > as if I told her "last night Mary dreamt of unicorns". The fact that Mary would > not describe her dream that way is not really relevant here.

We have an ambiguity here in English (I hope not in Lojban, but presently apparently so)between an teernal account of a dream and its content and an internal one. In the first, what Mary would say does not matter; in the second it does. I was, of course, talking about the second. The first is monstrously uninteresting because it does not get to the dream itself; it is some one else's description and so extensional in a boring sort of way. In that sense then whar you say may be true, but who cares?

> > Of course, even in the case of {lo pavyseljirna}, > > the xorxces’ interpretation (insofar as I understand it) creates the problem that > indefinite > > descriptions are now ambiguous: sometimes referring to these special entities, sometimes to > what > > one would expect, the things named. > > They always refer to the things named, not just in some cases. The problem is > in the assumption that the only things that can be named are things existing > in the real world. That's obviously not the case.

No, the only things that there are are things in the domain of discourse, some of which exist and some do not. The problem arises — if not elesewhere (but it does) — especially when we talk about things of which there are none (in the domain). "There are no unicorns but I want one." is perfectly intelligible and may even be true, but you can'tdodge the problem by saying that the unicorns referred to are in the domain but don't exist: we just said that they are not in the domain at all.

I seem to recall that you (or someone)tried to dodge this by insisting that anytime unicorns were mentiooned the domain thereby expanded to include some. But this has very strange consequences, for example that "There are no unicorns" can never be true if uttered, although it is often true (so long as we keep quiet about it).

Actually, that there are things — even existing things — that do not exist in the real world is not obvious and, indeed, there are many arguements to deny. However, for now let us assume that it is the case. The "problems" with opaque contexts still arise.

> > That is, in some circumstances, {lo broda} just refers to > > brodas > > Always, not just in some circumstances.

But then you have not solved the problem, for if there are no brodas, there are no brodas that you want.

> > >(usually ordinary extensional objects), > > Often, I don't know if usually, but often yes. > > > > And events in Lojban are not special sumti. So if you insist that > > > {la meris pu senva lo pavyseljirna} is {la meris pu senva su'o lo pavyseljirna}, > > > then {la meris pu senva lo nu lo pavyseljirna cu blabi} is {la meris pu senva > > > su'o lo nu lo pavyseljirna cu blabi}, i.e. there is at least one event among > > > all the events of unicorns being white, such that Mary dreamt of that event, > > > which is in no way less problematic that the equivalent thing with plain > > > unicorns. > > > > Actually it is less problematic, since all events exist and, thus, the one that Mary dreamed > of > > does. Admittedly, the claim that all events exist may seem objectionable, but these are > events in > > the abstract, so the universe does not get cluttered by this and, as here, it solves a number > of > > otherwise somewhat intractable problems. > > Do you make any distinction between events that take place in the real > world and events that don't?

I suppose that is the difference between events that actually occur and those that don't and that is a distinction I would want to make, of course. If you want to equate occurring with existing, the I will simply say that all events are in every domain and work from that.


> What exactly is the force of "exists" in "all > events exist"? Why is it any less problematic having these two types of > events (assuming you accept that there are these two types) than having > other creatures some of which exist and some of which don't exist in the > real world?

Probably not much, but then I don't have any problem with having creatures some of which do and some don't exist in the real world. The problems arise with things that are not in the domain altogether.



Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

rlpowell Posted by rlpowell on Thu 10 of Aug., 2006 18:43 GMT posts: 14214 On Thu, Aug 10, 2006 at 03:29:01PM -0300, Jorge Llamb?as wrote: > On 8/10/06, Robin Lee Powell wrote: > >On Thu, Aug 10, 2006 at 10:10:12AM -0300, Jorge Llamb?as wrote: > >> > >> I agree with "we need to do a bunch of work to clean things > >> up". In particular I would like to see an exhaustive list of > >> gismu places that take du'u-abstractions only, an exhaustive > >> list of gismu places that take ka/ni-abstractions only, an > >> exhaustive list of gismu places that take li-abstractions only, > > > >There is exactly one of those, interestingly enough: > > > > namcu nac na'u number > > x1 (li) is a number/quantifier/digit/value/figure > > (noun); refers to the value and not the symbol > > I wasn't very clear with my "I would like to see". What I had in > mind is that I would like to see the BPFK defining these lists > rather than to see what's already marked as such in the gi'uste.

Yeah, that's a lot different than what you asked for. I have no idea how to generate such a list, by hand or otherwise.

> For example, in the case of li-abstractions, which are probably > the easiest to deal with: all the x2 places of unit gismu (mitre, > grake, snidu, etc) can only be filled with li-abstractions, even > if the gi'uste doesn't mark this in any explicit way. Also the x1 > of {detri} and {tcika}, the x3 of {jbena}, and a couple of > oddballs like the x2 of {verba}.

Yeah, that means going through every single gismu by hand.

Perhaps while we are at it a less ambigious way of marking such places could be devised, like (x3=nu).

> But your lists are clearly the right place to start.

Thanks.

-Robin


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 19:03 GMT On 8/10/06, Jorge Llambías wrote: > On 8/10/06, Theodore Reed wrote: > > > I'd like to reiterate my misgivings about *requiring* certain > > abstractions. I can see requiring that something be a "number". But, > > for instance, couldn't I name a number Bob? > > Of course. In that case, "Bob" will be the name of a li-abstraction, > there is nothing wrong with that. You can also have li-abstractions > as values of variables, and of course {lo namcu} refers to a li-abstraction > too. The point is that anything that fills a place that requires a > li-abstraction > has to refer to a li-abstraction. This is not about the form that the sumti > there can have but about the type of value that the sumti there can refer to.

Given that, I think that putting (nu) in definitions is somewhat misleading. Either a preface needs to be added clarifying this, or we need to use words like "event" that doesn't so strongly indicate the requirement of an exact phrasing. Preferably both.

-- Theodore Reed (treed/bancus) www.surreality.us

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 19:16 GMT On 8/10/06, Theodore Reed wrote: > On 8/10/06, Jorge Llambas wrote: > > This is not about the form that the sumti > > there can have but about the type of value that the sumti there can refer to. > > Given that, I think that putting (nu) in definitions is somewhat > misleading. Either a preface needs to be added clarifying this, or we > need to use words like "event" that doesn't so strongly indicate the > requirement of an exact phrasing. Preferably both.

True. That was probably done in part because of the strict length limitations that the gismu definitions were forced to have. The gi'uste is also not very consistent since sometimes it uses (ka) and sometimes (property), etc.

mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 20:13 GMT On 8/10/06, John E Clifford wrote: > > We have an ambiguity here in English (I hope not in Lojban, but presently apparently so)between an > teernal account of a dream and its content and an internal one. In the first, what Mary would say > does not matter; in the second it does.

Mary can tell me: {mi pu senva lo pavyseljirna} I can report that to your daugther as {la meris pu senva lo danlu poi se skina lo se nelrai be do}

{lo pavyseljirna} and {lo danlu poi se skina lo se nelrai be do} refer in both cases to the same thing, the subject of Mary's dream, unicorns.

Another dream report might go like this:

Mary tells me: {mi pu senva lo du'u lo pavyseljirna cu blabi} I _cannot_ report that to your daugther as {la meris pu senva lo du'u lo danlu poi se skina lo se nelrai be do cu blabi} because the terms are not co-referential in the subordinate bridi.


> > The problem is > > in the assumption that the only things that can be named are things existing > > in the real world. That's obviously not the case. > > No, the only things that there are are things in the domain of discourse, some of which exist and > some do not. The problem arises — if not elesewhere (but it does) — especially when we talk > about things of which there are none (in the domain).

It is impossible to talk about things not in the domain of discourse. That's a contradiction in terms.

> "There are no unicorns but I want one." is > perfectly intelligible and may even be true, but you can'tdodge the problem by saying that the > unicorns referred to are in the domain but don't exist: we just said that they are not in the > domain at all.

In that context, the stilted expression "there are no unicorns" has to be interpreted as "unicorns don't exist (in the real world)", and of course there have to be unicorns in the universe of discourse for me to say something about unicorns.

> I seem to recall that you (or someone)tried to dodge this by insisting that anytime unicorns were > mentiooned the domain thereby expanded to include some.

By definition of domain of discourse, yes. The domain of discourse must include anything that is talked about.

> But this has very strange consequences, > for example that "There are no unicorns" can never be true if uttered, although it is often true > (so long as we keep quiet about it).

"There are no unicorns (in the real world)" is a reasonable thing to say.

> Actually, that there are things — even existing things — that do not exist in the real world is > not obvious and, indeed, there are many arguements to deny. However, for now let us assume that > it is the case. The "problems" with opaque contexts still arise.

There are other issues with opaque contexts that still arise, but the issues concerning existence in the real world all seem like non-issues to me.

> > > That is, in some circumstances, {lo broda} just refers to > > > brodas > > > > Always, not just in some circumstances. > > But then you have not solved the problem, for if there are no brodas, there are no brodas that you > want.

If there are no brodas in the world, I might still want brodas, yes, and I can still talk about brodas, and refer to brodas.


> > Do you make any distinction between events that take place in the real > > world and events that don't? > > I suppose that is the difference between events that actually occur and those that don't and that > is a distinction I would want to make, of course. If you want to equate occurring with existing, > the I will simply say that all events are in every domain and work from that.

But that's not reasonable. Sometimes the domain might be contextually restricted to real numbers for example, so I can say "for every x, there is some y such that x*y = 1". If every domain must include events, then that would be nonsense.

> > What exactly is the force of "exists" in "all > > events exist"? Why is it any less problematic having these two types of > > events (assuming you accept that there are these two types) than having > > other creatures some of which exist and some of which don't exist in the > > real world? > > Probably not much, but then I don't have any problem with having creatures some of which do and > some don't exist in the real world. The problems arise with things that are not in the domain > altogether.

In any given context, everything referred to is in the domain. (Trivially, everything referred to is among the things being talked about.)

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 10 of Aug., 2006 20:38 GMT posts: 2388


> On 8/10/06, John E Clifford wrote: > > > > We have an ambiguity here in English (I hope not in Lojban, but presently apparently > so)between an > > teernal account of a dream and its content and an internal one. In the first, what Mary would > say > > does not matter; in the second it does. > > Mary can tell me: {mi pu senva lo pavyseljirna} > I can report that to your daugther as {la meris pu senva lo danlu > poi se skina lo se nelrai be do} > > {lo pavyseljirna} and {lo danlu poi se skina lo se nelrai be do} refer in both > cases to the same thing, the subject of Mary's dream, unicorns.

Quite true but not relevant to the issue at hand. There is a sense of "Mary dreamed about unicorns" that depends upon unicorns playing a major role in the dream. Now in this sense, it may or may not be the case that some things that are extensionally equivalent to unicorns play a role. And the fact that it they may not is enough to make the case. To take a clearer example (I hope): in the appropriate world, if Lois Lane dreams about Superman it is incorrect to say she dreams (in the relevant sense) about Clark Kent, even though Clark Kent is Superman (and vice versa) — the thing she dreamed about does not have the characteristics of CK (shy, needs glasses, ...) and so he plays no role in her dream, even though, ouside of the dream, the person who does play a role is also CK. So, being the critter in my daughter's favorite film plays no role in Mary's dream even though what does play a role is on the outside that critter.

> Another dream report might go like this: > > Mary tells me: {mi pu senva lo du'u lo pavyseljirna cu blabi} > I _cannot_ report that to your daugther as {la meris pu senva lo du'u > lo danlu poi se skina lo se nelrai be do cu blabi} because > the terms are not co-referential in the subordinate bridi. > > > > > The problem is > > > in the assumption that the only things that can be named are things existing > > > in the real world. That's obviously not the case. > > > > No, the only things that there are are things in the domain of discourse, some of which exist > and > > some do not. The problem arises — if not elsewhere (but it does) — especially when we talk > > about things of which there are none (in the domain). > > It is impossible to talk about things not in the domain of discourse. > That's a contradiction in terms.

See the case below and the problems with this view. It is obviously not a contradiction in terms unless you take the extreme position that everything mentioned has to be in the domain of discourse.


> > > "There are no unicorns but I want one." is > > perfectly intelligible and may even be true, but you can'tdodge the problem by saying that the > > unicorns referred to are in the domain but don't exist: we just said that they are not in the > > domain at all. > > In that context, the stilted expression "there are no unicorns" has to be > interpreted as "unicorns don't exist (in the real world)", and of course there > have to be unicorns in the universe of discourse for me to say something > about unicorns.


Not at all; it is deliberately taken as the standrad way of saying thatnowhere in the range of variables is there anything which has the property of being a unicorn, where the variable range exactly over the universe of discourse.


> > I seem to recall that you (or someone)tried to dodge this by insisting that anytime unicorns > were > > mentiooned the domain thereby expanded to include some. > > By definition of domain of discourse, yes. The domain of discourse must include > anything that is talked about.

That is not the standard definition of a universe of discourse. If you want to use the expression that way, fine — but give us warning. And accept the consequences.

> > But this has very strange consequences, > > for example that "There are no unicorns" can never be true if uttered, although it is often > true > > (so long as we keep quiet about it). > > "There are no unicorns (in the real world)" is a reasonable thing to say.

But that is not what "There are no unicorns" says.

> > Actually, that there are things — even existing things — that do not exist in the real > world is > > not obvious and, indeed, there are many arguements to deny. However, for now let us assume > that > > it is the case. The "problems" with opaque contexts still arise. > > There are other issues with opaque contexts that still arise, but the issues > concerning existence in the real world all seem like non-issues to me.

But there is no issue about existence which is confined to the real world.

> > > > That is, in some circumstances, {lo broda} just refers to > > > > brodas > > > > > > Always, not just in some circumstances. > > > > But then you have not solved the problem, for if there are no brodas, there are no brodas that > you > > want. > > If there are no brodas in the world, I might still want brodas, yes, > and I can still > talk about brodas, and refer to brodas.

Yes, but those are all now in a different universe of discourse, not the one that applies in non-subordinate positions (maybe you identitfy the real world with the top level universe of discourse?)

> > > > Do you make any distinction between events that take place in the real > > > world and events that don't? > > > > I suppose that is the difference between events that actually occur and those that don't and > that > > is a distinction I would want to make, of course. If you want to equate occurring with > existing, > > the I will simply say that all events are in every domain and work from that. > > But that's not reasonable. Sometimes the domain might be contextually > restricted to real numbers for example, so I can say "for every x, there is > some y such that x*y = 1". If every domain must include events, then that > would be nonsense.

Alright then in every domain that is nto specifically restricted (impliicitly or explicitly). Actually, events don't cause that much trouble in in the domain of numbers — it just means that there are some strange number, but there are anyhow. \ > > > What exactly is the force of "exists" in "all > > > events exist"? Why is it any less problematic having these two types of > > > events (assuming you accept that there are these two types) than having > > > other creatures some of which exist and some of which don't exist in the > > > real world? > > > > Probably not much, but then I don't have any problem with having creatures some of which do > and > > some don't exist in the real world. The problems arise with things that are not in the domain > > altogether. > > In any given context, everything referred to is in the domain. (Trivially, > everything referred to is among the things being talked about.)

I suppose that at some level of universe something like this is true, but put the way you seem to be using it it gives a very static view of the universes of discourse — or, at best, an expanding one. The universe of discourse is dynamic and expands and contracts as the discourse proceeds (I don't actually hold this view, but I am trying to get some sort of accomodation with your strange view). The problem about "There are no unicorns" remains even here however, since there are no shifts of level involved. Of course, you could argue that this is not a referring expression and so falls outside the stricture, but the problem remains for referring expression "a unicorn."


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 10 of Aug., 2006 21:28 GMT On 8/10/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > The domain of discourse must include > > anything that is talked about. > > That is not the standard definition of a universe of discourse. If you want to use the expression > that way, fine — but give us warning. And accept the consequences.

Could you give the standard definition, so that we can communicate more effectively?

A quick search gives me:

universe of discourse n. Logic. A class containing all the entities referred to in a discourse or an argument. Also called universe.

everything stated or assumed in a given discussion Synonym: universe

domain of discourse The domain of discourse, sometimes called the universe of discourse, is an analytic tool used in deductive logic, especially predicate logic. It indicates the relevant set of entities that are being dealt with by quantifiers.


None of these definitions seem to allow for talk about things that are outside the universe of discourse.


> > If there are no brodas in the world, I might still want brodas, yes, > > and I can still talk about brodas, and refer to brodas. > > Yes, but those are all now in a different universe of discourse, not the one that applies in > non-subordinate positions (maybe you identitfy the real world with the top level universe of > discourse?)

I certainly don't, you sometimes seem to. The real world is not a set or a class, like the universe of discourse is. I can certainly talk about unicorns while being in a world without unicorns. My talking about unicorns automatically introduces unicorns into the universe of discourse. Obviously it does not create unicorns in the real world, but that's not relevant. If I tell you that I dreamt of unicorns, I am bringing unicorns into the universe of discourse, I am not claiming that unicorns exist in the real world. There is now something in the universe of discourse which is both unicorn and dreamt about by me. That's all that's needed for {mi senva lo pavyseljirna} to be true. There is no requirement that unicorns exist in the real world.


> > In any given context, everything referred to is in the domain. (Trivially, > > everything referred to is among the things being talked about.) > > I suppose that at some level of universe something like this is true, but put the way you seem to > be using it it gives a very static view of the universes of discourse — or, at best, an expanding > one. The universe of discourse is dynamic and expands and contracts as the discourse proceeds (I > don't actually hold this view, but I am trying to get some sort of accomodation with your strange > view). The problem about "There are no unicorns" remains even here however, since there are no > shifts of level involved.

I don't know how to interpret "there are no unicorns" without an implicit "in the real world". It is an incomplete sentence unless interpreted as "no unicorns exist". It cannot mean that there are none in the universe of discourse because by talking about unicorns you are bringing them into the universe of discourse.

> Of course, you could argue that this is not a referring expression and > so falls outside the stricture, but the problem remains for referring > expression "a unicorn."

Not sure what you mean.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 11 of Aug., 2006 15:18 GMT posts: 2388

> On 8/10/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > The domain of discourse must include > > > anything that is talked about. > > > > That is not the standard definition of a universe of discourse. If you want to use the > expression > > that way, fine — but give us warning. And accept the consequences. > > Could you give the standard definition, so that we can communicate more > effectively? > > A quick search gives me: > > universe of discourse > n. Logic. > A class containing all the entities referred to in a discourse or an > argument. Also called universe. > > everything stated or assumed in a given discussion > Synonym: universe > > domain of discourse > The domain of discourse, sometimes called the universe of discourse, > is an analytic tool used in deductive logic, especially predicate > logic. It indicates the relevant set of entities that are being dealt > with by quantifiers.

Is this wikipedia? Only the last sentence seems to me to be on the mark (which is pretty good if it is wikipedia). > ----- > > None of these definitions seem to allow for talk about things that are > outside the universe of discourse. > But they allow one to talk at one time about things that are outside the universe of discourse at another time (or place, more accurately)and here is where the problem comes. In these terms (not really since these definitions appear to be about static once-for-all-this-section universes) the problem is that, while talking about Mary's dream, there are unicorns in the universe of discourse, but once we step away from that there are not (assuming there are none in the initial universe — the "real world" or some other place wherein the conversation starts). The illegitimate generalization goes from the one broader universe to the narrower and hence from true to false in this case.

Notice again that the problem is not confined to things that are not; it holds in other cases as well. I need a pair of pliers, say. It may be that the pair I need is not in the initial domain (I need needle-nosed and there are only the other kind). Or, of course, it may just be that although there are no such problems but the inference still does not go through because there is no pair of pliers that I need: whichever I point to any other one will do as well, so I didn't need that one) > > > > If there are no brodas in the world, I might still want brodas, yes, > > > and I can still talk about brodas, and refer to brodas. > > > > Yes, but those are all now in a different universe of discourse, not the one that applies in > > non-subordinate positions (maybe you identitfy the real world with the top level universe of > > discourse?) > > I certainly don't, you sometimes seem to. The real world is not a set > or a class, like the universe of discourse is/

Picky, picky. OK, the domain composed of just those things that exist in the real world (or some subset of them). This is the default initial domain of discourse; moving from it takes a sign or some other sort of conventional indication.

I can certainly talk > about unicorns > while being in a world without unicorns.

I think I understand this, which verbally goes against a point you were insisting upon a moment ago. You mean that you can be in one world but be talking using another universe than that defined by the wor;d you are actually in. Of course you can; we do it all the time. This is not a special insight or an attack on anything I have said ot implied. So what is your point here.

> My talking about unicorns > automatically introduces unicorns into the universe of discourse. Obviously > it does not create unicorns in the real world, but that's not relevant. > If I tell you that I dreamt of unicorns, I am bringing unicorns into the > universe of discourse, I am not claiming that unicorns exist in the real > world.

OK, you have changed the universe of discourse by expanding it. But this is ofyrn a very temporary expansion/ The unicorn are added for talking about your dream but disappear when you get off that (well, they can be held in various ways, but it takes an effort)-- and they were not there before you started talking about the dream (suppose). The illicit inference involves taking something from the expanded domain and claiming it is in the restricted. (It actually turns out to work better if rather than expanding and contracting domains you pass from one world (domain with all the interpretations in place) to another, since that allows an explanation of the idnetity failures in the same system).


> There is now something in the universe of discourse which is both > unicorn and dreamt about by me. That's all that's needed for {mi senva > lo pavyseljirna} to be true. There is no requirement that unicorns exist > in the real world.

And no one ever said there was — in fact it is the fact that there aren't any that creates the problem.

> > > > In any given context, everything referred to is in the domain. (Trivially, > > > everything referred to is among the things being talked about.) > > > > I suppose that at some level of universe something like this is true, but put the way you seem > to > > be using it it gives a very static view of the universes of discourse — or, at best, an > expanding > > one. The universe of discourse is dynamic and expands and contracts as the discourse proceeds > (I > > don't actually hold this view, but I am trying to get some sort of accomodation with your > strange > > view). The problem about "There are no unicorns" remains even here however, since there are > no > > shifts of level involved. > > I don't know how to interpret "there are no unicorns" without an implicit > "in the real world".

"Suppose there are no unicorns..." clearly not about the real world because flagged as about some other world. It is about the current domain, whatever it is — not just the "real world" domain.

> It is an incomplete sentence unless interpreted as > "no unicorns exist". It cannot mean that there are none in the universe of > discourse because by talking about unicorns you are bringing them into > the universe of discourse.

But possibly only temporarily and possibly only offering the change — which offer the other side may reject, as "there are no unicorns" does.

> > Of course, you could argue that this is not a referring expression and > > so falls outside the stricture, but the problem remains for referring > > expression "a unicorn." > > Not sure what you mean.

"There are no unicorns" does not contain a referring expression (as we have at some point agreed to use that phrase). The same problem can be generated, however, with "Nothing is a unicorn."

Let me bring all this together in one place. The universe of discourse is dynamic. At the beginning of a conversation there is already an established domain (the one of the "real world" as the interlocutors conceive it as default). As the conversation proceeds, the domain may change, grow or shrink, at the instigation of the speaker at the moment. There are two kinds of changes, cooperative and unassailable. Cooperative changes require the acceptance of the others in the conversation to take effect. If I say "I just hit a unicorn" when unicorns are not in the domain at the beginning of the utterance, I propose extending that domain to include some unicorns. My hearer may say something like "Wow! How bad was it hurt?" accepting the change. Or he may say "Don't be silly! There aren't any unicorns," rejecting it. In the first case the domain is changed, in the second not. Whether to accept or reject a proposed change is largely a matter of conversational etiquette: refusing to restrict "bear" from "all bears ever anywhere" to "the bears in this zoo now" is usually churlish; refusing to countenance adding unicorns is probably acceptable, even noble. I would like to say that cooperative changes are permanent (until explicitly reversed) but in fact some of them seem to just fade away and I don't have a good grip on how that works (but right hand ends are always a problem). Unassailable changes take place regardless of the hearers' attitude. They are, however, temporary. They involve, as it were, detours on the conversational path, moves into other realms (wishdom, dreamdom, beliefdom and so on), and endure only until the return to the mainstream. Thus: main domainI want possible shift a unicorn yes, a shift even though shift back there aren't any unicorns. In every language (I suppose) there are certain situations (like saying "I want" above) that lay open the possibility of a shift and there are certain expressions that usually show that a shift has taken place (like "that..."). There are, it turns out, in English (and probably most other languages) exceptions: shifts that are not in any way marked (after "I saw" for example in reproting sense data or dubious perceptions) and possible shifts not taken ("I want that bear there" pointing to a bear there). In English, also, (and most other languages)many of the markings are merely lexical, not explicitly set apart(the second place of "want" is not different from the second place of "hit" in any obvious way). And the supposed markers sometimes are used where no shift is actually taken (or where it is the zero-shift, if you want to solve the matter technologically). In a logical language, it ought to be otherwise. That is, the unassailable shifts should be marked and non-shifts should be unmarked (as much as possible in this case, for there are other consideration: {djuno2} pretty much has to be a proposition, even if that proposition involves no change in the domain — as is the case with {djuno}). Hopefully this will prteey much settle the question of the scope of the shift (but see the issues with {da'i} and bound variables). Lojban has chosen to do this by setting unassailable shifts in abstraction sumti. This is a rather clever choice since 1) abstraction regularly do involve shifts, 2) in some cases it is clear that abstractions are involved (propositions, events in wishes and the like) and 3) abstractions are always available to be used (If you don't like this axiom of Lojban, you might coonsider that all abstractions can be reduced to propositional functions, which seem less objectionable than events, say, and that availability need not mean actual presence so much as guaranteed (unassailable but permanent) production at need.) Consistent use of these devices will make Lojban much more nearly a logical language (there are some problems as noted to be solved). As far as the word lists go, what needs to be marked are those cases where there is a string potential for a shift and how that shift might be dealt with. If no shift is intended, then the device need not be used, unless otherwise required. Further, we have to allow that the shift can be used in other places under special circumstances: {mi viska tu'a lo pavyseljirna} for "I swear I saw a unicorn — or at least it sure looked like one to me then."


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 11 of Aug., 2006 19:48 GMT On 8/11/06, John E Clifford wrote: ... > In these terms (not really since these definitions appear to be about > static once-for-all-this-section universes) the problem is that, while > talking about Mary's dream, there are unicorns in the universe of > discourse, but once we step away from that there are not (assuming there > are none in the initial universe — the "real world" or some other place > wherein the conversation starts).

When I say "Mary dreamt of unicorns", the reason there are unicorns in the universe of discourse is not that there are unicorns in Mary's dream. The reason there are unicorns in the universe of discourse is that I mentioned unicorns. If I say "Mary has never dreamt of unicorns", or "Mary dreamt that unicorns didn't exist" there are also unicorns in the universe of discourse for the very same reason. The existence of unicorns in Mary's dream world or in the speaker's real world is irrelevant to whether or not the universe of discourse has unicorns as one of its members.

... > Let me bring all this together in one place. The universe of discourse > is dynamic. At the beginning of a conversation there is already an > established domain (the one of the "real world" as > the interlocutors conceive it as default).

Any conversation takes place in a world, yes. From there to saying that any conversation is by default about only the things that exist in that world there is a big jump, one that I do not make. Lots of conversations are from the start about things that don't exist in the world where the conversation takes place, and probably no conversation ever is about _all_ the things that exist in that world.

> As the conversation proceeds, the domain may change, > grow or shrink, at the instigation of the speaker at the moment. There are > two kinds of changes, cooperative and unassailable. > Cooperative changes require the acceptance of the others in the conversation > to take effect. If I say "I just hit a unicorn" when unicorns are not in the > domain at the beginning of the utterance, I propose extending that domain > to include some unicorns. My hearer may say something like "Wow! > How bad was it hurt?" accepting the change.

I cannot imagine a context in the real world in which such a conversation would be meaningful, but perhaps I'm just being unimaginative.

> Or he may say "Don't be silly! There aren't any > unicorns," rejecting it. In the first case the domain is changed, in the > second not.

When he says "there aren't any unicorns" he has to know what unicorns are, he is referring to them, and he is claiming of them that there aren't any (in the real world, which is what would be required for you to hit one). He has certainly accepted your introduction of unicorns into the domain of discourse, as he is indeed now talking about unicorns himself. He is rejecting your claim about unicorns, not your introducing unicorns into the domain of discourse. In order to not accept the introduction into the universe of discourse he would have to say something like "I don't follow you, what does 'unicorn' mean here?". In that case he would not be making any claims about unicorns.

> Whether to accept or reject a proposed change is largely a matter > of conversational etiquette: refusing to restrict "bear" from "all bears ever > anywhere" to "the bears in this zoo now" is usually churlish; > refusing to countenance adding unicorns is probably acceptable, even > noble.

Should I ask now what do you mean by 'unicorns'? That doesn't seem noble. I know what you mean by that word, why should I refuse to countenance your adding them to the things you say something about?


... > Lojban > has chosen to do this by setting unassailable shifts in abstraction sumti. > This is a rather clever choice since 1) abstraction regularly do involve shifts,

Shifts in what? Shifts in universe of discourse, or shifts in the world in which a given proposition is to be evaluated? These are two different things. Abstractions regularly involve a shift in the world where a proposition is evaluated, not a shift in universe of discourse.

> 2) in some cases it is clear that > abstractions are involved (propositions, events in wishes and the like)

It is clear that a propositional abstraction is involved in "I want that ...". It is not at all clear that a propositional abstraction must be involved in "I want a ...". The claim that there is a deep structure there where a propositional abstraction is involved is just one proposed way of dealing with some issues. It is not some obvious and unquestionable fact.

> and 3) abstractions are always available to be used (If you don't like this > axiom of Lojban, you might coonsider that all abstractions can be reduced > to propositional functions, which seem less objectionable than events, > say, and that availability need not mean actual presence so much as > guaranteed (unassailable but permanent) production at need.)

But I have absolutely no problem with the availability of abstractions. My only comment here is that non-abstractions are always as available as abstractions to be entered into the universe of discourse. There is no relevant distinction to be made here between events and animals. Both can be real or imaginary, and both are always available to be introduced into the universe of discourse.

> Consistent use of these devices will make Lojban much more nearly > a logical language (there are some problems as noted to be solved).

I agree there are problems yet to be solved. I strongly disagree that cluttering Lojban with {tu'a}'s will make it more logical.

> As far as the word lists go, what needs to be marked are those cases > where there is a string potential for a shift and how that > shift might be dealt with. If no shift is intended, then the device need > not be used, unless otherwise required. Further, we have to allow that > the shift can be used in other places under special circumstances: > {mi viska tu'a lo pavyseljirna} for "I swear I saw a unicorn — or at least > it sure looked like one to me then."

I see no logic in that interpretation of {tu'a}.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 11 of Aug., 2006 21:18 GMT posts: 2388

> On 8/11/06, John E Clifford wrote: > ... > > In these terms (not really since these definitions appear to be about > > static once-for-all-this-section universes) the problem is that, while > > talking about Mary's dream, there are unicorns in the universe of > > discourse, but once we step away from that there are not (assuming there > > are none in the initial universe — the "real world" or some other place > > wherein the conversation starts). > > When I say "Mary dreamt of unicorns", the reason there are unicorns > in the universe of discourse is not that there are unicorns in Mary's dream. > The reason there are unicorns in the universe of discourse is that I mentioned > unicorns. If I say "Mary has never dreamt of unicorns", or "Mary dreamt that > unicorns didn't exist" there are also unicorns in the universe of discourse for > the very same reason. The existence of unicorns in Mary's dream world or > in the speaker's real world is irrelevant to whether or not the universe of > discourse has unicorns as one of its members.

I find this somewhat perverse and apparently confusing predication with reference. The "unicorn"s in your hypothetic sentences are not references but merely predications attached to boudj variables (or some such thing). Using the predicate "unicorn" in no way commits one to there being unicorns in the domain of discourse (the solution to the original version of the "there are no unicorns" paradox — which I am sorry you missed). As for the first point, I did not say that there were unicorns in the domain because they were in her dream (although presumably they were); domains are connected with language and arise and change only in (or at least around) conversation. But the expansion of a domain at a point in a conversation occurs only when a reference is made, not a predication. The predicate "unicorn" is already covered in the domain both extensionally and intensionally (we could fiddle with the intension a bit if we want to insist that only intensions the conversants know are in the domain). The shift them is in whether the extension is null or not and whether it has a certain member or not. But that only arises when a refernce is made to that or to some member of that extension. Notice as aways that this applies to an domain, not merely the one of the real world.

> ... > > Let me bring all this together in one place. The universe of discourse > > is dynamic. At the beginning of a conversation there is already an > > established domain (the one of the "real world" as > > the interlocutors conceive it as default). > > Any conversation takes place in a world, yes. From there to saying that > any conversation is by default about only the things that exist in that world > there is a big jump, one that I do not make. Lots of conversations are from > the start about things that don't exist in the world where the conversation > takes place, and probably no conversation ever is about _all_ the things that > exist in that world.

Yes, not every conversation (perhaps not many) begins in the default domain. That does not mean that it is not the default; it means that there are a large number of ways to set different inital domains and that these means are often employed. As for "all" the things in the world — or any domain for that matter — if they are never mentioned or involved, they do not harm; if they would be involved in a negative way they are probably removed by some device (as witrness that they sometimes are not and we get trouble). But I am quite content to say that the default domain is drawn from the "real world" (although this is not quite enough for a real "default"). I suppose that another approach would be to start with an essentially empty domain and add things as the need arose (initially only the speaker and hearer and perhaps a few other items to make conversation possible).

> > As the conversation proceeds, the domain may change, > > grow or shrink, at the instigation of the speaker at the moment. There are > > two kinds of changes, cooperative and unassailable. > > Cooperative changes require the acceptance of the others in the conversation > > to take effect. If I say "I just hit a unicorn" when unicorns are not in the > > domain at the beginning of the utterance, I propose extending that domain > > to include some unicorns. My hearer may say something like "Wow! > > How bad was it hurt?" accepting the change. > > I cannot imagine a context in the real world in which such a conversation > would be meaningful, but perhaps I'm just being unimaginative.

Who said this was in the domain of the real world? To be sure, there is no domain specified for this snippet, but the most that can be said in that case is that it draws upon the then current domain ("the domain at the beginning of the utterance"), whatever it may be (except it apparently does not contain unicorns). I think you are reading too muc into "default" ("obligatory" maybe?).

> > Or he may say "Don't be silly! There aren't any > > unicorns," rejecting it. In the first case the domain is changed, in the > > second not. > > When he says "there aren't any unicorns" he has to know what unicorns > are, he is referring to them, and he is claiming of them that there > aren't any (in > the real world, which is what would be required for you to hit one).

He does have to know what "unicorn" means, but he is not referring to unicorns and he is refusing to extend the domain at this point, very likely because he believes that there aren't any unicorns to hit.

> He has > certainly accepted your introduction of unicorns into the domain of discourse, > as he is indeed now talking about unicorns himself.

No he has not, he has in fact rejected it. To be sure, he is also calling the claim false — a very economical utterance. Notice, he does not say that unicorns do not exist, but rather uses the form that denies that anything which might be the value of a variable is in the unicorn cam, i.e., makes a claim about the domain (or rather refuses a change in it).

> He is rejecting your claim > about unicorns, not your introducing unicorns into the domain of discourse. In > order to not accept the introduction into the universe of discourse he would > have to say something like "I don't follow you, what does 'unicorn' > mean here?".

He has to understand what "unicorn" means as much to reject the change as to accept it. "What does "unicorn" mean here?" is a request for information needed to decide whether to accept or reject.

> In that case he would not be making any claims about unicorns.

True, but he is not — in your sense — making any claim about unicorns in the other either. At msost he making a claim about the predicate "unicorn", that its extension in the domain ie empty.

> > Whether to accept or reject a proposed change is largely a matter > > of conversational etiquette: refusing to restrict "bear" from "all bears ever > > anywhere" to "the bears in this zoo now" is usually churlish; > > refusing to countenance adding unicorns is probably acceptable, even > > noble. > > Should I ask now what do you mean by 'unicorns'? That doesn't seem noble. > I know what you mean by that word, why should I refuse to countenance > your adding them to the things you say something about?

I didn't say it mas noble, I said it might be, if it saved the discussion in any of a variety of time-wasting opaths, for example. > > ... > > Lojban > > has chosen to do this by setting unassailable shifts in abstraction sumti. > > This is a rather clever choice since 1) abstraction regularly do involve shifts, > > Shifts in what? Shifts in universe of discourse, or shifts in the world in which > a given proposition is to be evaluated? These are two different things.

Not really. If the domain is to be sensitive to predicates then it needs to contain the interpretations of thos predicates, otherwise, how can we add say unicorns to the domain — adding another x doesn't fill the bill. To work, the new individual (and of course, it need not be new as an individual, it merely needs a new location, moving into the unicorn extension) has to be a unicorn. Maybe I ought to be more careful and reserve the word "domain" for just the things and "universe" for the whole package (or "world" if you prefer). Then what I said about domains is about worlds instead. For the present purpose, the differnce is miniscule and doesn't affect matters that I can see.

> Abstractions regularly involve a shift in the world where a proposition is > evaluated, not a shift in universe of discourse.

A shift in a world is, almsot by definition a shift in domain, because, even if the same individuals occur in both worlds, their properties are different and hence they come in as referents in different ways.

> > 2) in some cases it is clear that > > abstractions are involved (propositions, events in wishes and the like) > > It is clear that a propositional abstraction is involved in "I want that ...". > It is not at all clear that a propositional abstraction must be involved in > "I want a ...". The claim that there is a deep structure there where a > propositional abstraction is involved is just one proposed way of dealing > with some issues. It is not some obvious and unquestionable fact.

As they say, "If you want a dog, you are in luck, because there are dogs." If your want is not satisfied — as it usually would not be — then you must want something other than a dog, presumably to have a dog. And that seems to be an event. I haven't seen anybody offer a good dodge arounf this, so I will be interested in what you propose. I didn't say anything about deep structure and the claim (which comes later on) that all abstractions can be defined in terms of propositional function is presented only as one possible way to alleviate someone's discomfort about saying that all events are in every domain.

> > and 3) abstractions are always available to be used (If you don't like this > > axiom of Lojban, you might consider that all abstractions can be reduced > > to propositional functions, which seem less objectionable than events, > > say, and that availability need not mean actual presence so much as > > guaranteed (unassailable but permanent) production at need.) > > But I have absolutely no problem with the availability of abstractions. My > only comment here is that non-abstractions are always as available as > abstractions to be entered into the universe of discourse. There is no > relevant distinction to be made here between events and animals. Both > can be real or imaginary, and both are always available to be introduced > into the universe of discourse.

The point is that abstractions are always already in the world or can be called up without changing the world, so the problems of not having what one wants (etc.) or of not having the right one disappear. This isn't the only reason for doing this, but it is a handy side effect. The case of ordinary things is quite different: they are added (in the interesting cases) only in unassailable and temporary world shifts and so are not around once we leave that detour world. If they are added cooperatively then they cannot be defined on the basis of the world shift and expect to be accepted. Indeed, it turns out that defining additions in terms of the world shift creates a world shift of the appropriate sort — one of the ways that the scope of shifts can be extended. If I talk about the dog that I want, then I am into a shifted world and whatever comes up then will not apply to the world before this locution (or its kin) entered the conversation.

> > Consistent use of these devices will make Lojban much more nearly > > a logical language (there are some problems as noted to be solved). > > I agree there are problems yet to be solved. I strongly disagree that > cluttering Lojban with {tu'a}'s will make it more logical.

Well we diagree there and you owe me a way of solving the problem of opaque contexts without this device (I woulds be happy with a shorter one but I would not be happy for these problem cases to be unmarked).

> > As far as the word lists go, what needs to be marked are those cases > > where there is a strong potential for a shift and how that > > shift might be dealt with. If no shift is intended, then the device need > > not be used, unless otherwise required. Further, we have to allow that > > the shift can be used in other places under special circumstances: > > {mi viska tu'a lo pavyseljirna} for "I swear I saw a unicorn — or at least > > it sure looked like one to me then." > > I see no logic in that interpretation of {tu'a}.

The logic is just that it allows me both to tell you what I experienced and to recognize that it may not be what really happened. This is at least occasionally a useful sort of thing to be able to say and having a convenient way of saying it — one that fits a pattern and does the job — is handy.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Sat 12 of Aug., 2006 16:12 GMT On 8/11/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > The "unicorn"s > in your hypothetic sentences are not references but merely predications > attached to boudj > variables (or some such thing).

This is something we will probably never agree on. For me, "lo pavyseljirna" is a referring expression, and for you it is a quantified term. Since we disagree about something so basic, we probably won't advance much along this direction.

> > He has > > certainly accepted your introduction of unicorns into the domain of > discourse, > > as he is indeed now talking about unicorns himself. > > No he has not, he has in fact rejected it. To be sure, he is also calling > the claim false — a > very economical utterance. Notice, he does not say that unicorns do not > exist, but rather uses > the form that denies that anything which might be the value of a variable is > in the unicorn cam, > i.e., makes a claim about the domain (or rather refuses a change in it).

What if instead of saying "but there are no unicorns" he had said "but unicorns don't exist"? Would you agree that in that case he is accepting unicorns in the universe of discourse and referring to them?

> > > 2) in some cases it is clear that > > > abstractions are involved (propositions, events in wishes and the like) > > > > It is clear that a propositional abstraction is involved in "I want that > ...". > > It is not at all clear that a propositional abstraction must be involved > in > > "I want a ...". The claim that there is a deep structure there where a > > propositional abstraction is involved is just one proposed way of dealing > > with some issues. It is not some obvious and unquestionable fact. > > As they say, "If you want a dog, you are in luck, because there are dogs." > If your want is not > satisfied — as it usually would not be — then you must want something > other than a dog, > presumably to have a dog.

By that reasoning, if I say "mi djica lo nu mi ponse lo gerku" I'm in luck again, because there are nu mi ponse lo gerku. So if my want is not satisfied I must want something other than lo nu mi ponse lo gerku.

> And that seems to be an event. I haven't seen > anybody offer a good > dodge arounf this, so I will be interested in what you propose.

"lo gerku" and "lo nu mi ponse lo gerku" may both refer generically to dogs and events of my having dogs, and my wanting them is not satisfied just by their existence.

> The point is that abstractions are always already in the world or can be > called up without > changing the world, so the problems of not having what one wants (etc.) or > of not having the right > one disappear.

Does that mean one always has what one wants, if what one wants is an event?

> > > Further, we have to allow that > > > the shift can be used in other places under special circumstances: > > > {mi viska tu'a lo pavyseljirna} for "I swear I saw a unicorn — or at > least > > > it sure looked like one to me then." > > > > I see no logic in that interpretation of {tu'a}. > > The logic is just that it allows me both to tell you what I experienced and > to recognize that it > may not be what really happened.

So events that don't take place in the real world can still be visible?

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Sat 12 of Aug., 2006 18:25 GMT posts: 2388

> On 8/11/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > The "unicorn"s > > in your hypothetic sentences are not references but merely predications > > attached to boudj > > variables (or some such thing). > > This is something we will probably never agree on. For me, "lo pavyseljirna" > is a referring expression, and for you it is a quantified term. Since we > disagree about something so basic, we probably won't advance much > along this direction.

We actually agree perfectly on this; the ap[oint is that your examples contained nothing that would likely be presented as {lo pavyseljirna}.

> > > He has > > > certainly accepted your introduction of unicorns into the domain of > > discourse, > > > as he is indeed now talking about unicorns himself. > > > > No he has not, he has in fact rejected it. To be sure, he is also calling > > the claim false — a > > very economical utterance. Notice, he does not say that unicorns do not > > exist, but rather uses > > the form that denies that anything which might be the value of a variable is > > in the unicorn cam, > > i.e., makes a claim about the domain (or rather refuses a change in it). > > What if instead of saying "but there are no unicorns" he had said > "but unicorns don't exist"? Would you agree that in that case he is > accepting unicorns in the universe of discourse and referring to them?

That is iffy. I take "unicorns do not exist" as a narrower statement than "there are no unicorns", indeed as a statement about whatever is passing for the real world at the moment (by definition, I think - that is, the domain of the real world is just the extesnion of "exist"). I don't think that I would want to draw a distinction on the basis of form either: between say {no pavyseljirna cu zasti} and {lo pavyseljirna na zasti}. But in neither case would I say that in saying this he was committed to having unicorns in his universe of discourse. But he is not prevented from this by saying that they don't exist. I think the main difference between us is that you think that referring expressions always refer and I think that, while that is what they are designed to do, they don't always live up to it. I hold this for even primary occurrences; so I certainly hold it for secondary ones like these.

> > > > 2) in some cases it is clear that > > > > abstractions are involved (propositions, events in wishes and the like) > > > > > > It is clear that a propositional abstraction is involved in "I want that > > ...". > > > It is not at all clear that a propositional abstraction must be involved > > in > > > "I want a ...". The claim that there is a deep structure there where a > > > propositional abstraction is involved is just one proposed way of dealing > > > with some issues. It is not some obvious and unquestionable fact. > > > > As they say, "If you want a dog, you are in luck, because there are dogs." > > If your want is not > > satisfied — as it usually would not be — then you must want something > > other than a dog, > > presumably to have a dog. > > By that reasoning, if I say "mi djica lo nu mi ponse lo gerku" I'm in luck > again, because there are nu mi ponse lo gerku. So if my want is not > satisfied I must want something other than lo nu mi ponse lo gerku.

Alright, your luck with dogs is because dogs EXIST, but lo nu mi ponse lo gerku does not.

> > And that seems to be an event. I haven't seen > > anybody offer a good > > dodge arounf this, so I will be interested in what you propose. > > "lo gerku" and "lo nu mi ponse lo gerku" may both refer generically > to dogs and events of my having dogs, and my wanting them is not > satisfied just by their existence.

Well, what more than the existence of lo nu mi ponse lo gerku would youwant; you would then have your dog after all, so the condition for satisfying the want is met. Did we not work out that for an event to exist was just to have it obtain? If not, I propose that for a working rule.

> > The point is that abstractions are always already in the world or can be > > called up without > > changing the world, so the problems of not having what one wants (etc.) or > > of not having the right > > one disappear. > > Does that mean one always has what one wants, if what one wants is > an event?


You seem to be confusing what you have tried before to keep separate, the world (= universe of discourse) and "the real world" in which things are actual, etc. Things can be in the world but not in the real world as you well know (and are constantly insisting).

> > > > Further, we have to allow that > > > > the shift can be used in other places under special circumstances: > > > > {mi viska tu'a lo pavyseljirna} for "I swear I saw a unicorn — or at > > least > > > > it sure looked like one to me then." > > > > > > I see no logic in that interpretation of {tu'a}. > > > > The logic is just that it allows me both to tell you what I experienced and > > to recognize that it > > may not be what really happened. > > So events that don't take place in the real world can still be visible?

In at least one sense of "see," yes. We tend to call those illusions or delusion or mirages or ...


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Mon 14 of Aug., 2006 13:03 GMT On 8/12/06, John E Clifford wrote: > I don't think that I would want to draw a distinction on the basis of > form either: between say {no pavyseljirna cu zasti} and {lo pavyseljirna > na zasti}. But in neither case would I say that in > saying this he was committed to having unicorns in his universe of > discourse.

Whereas for me, both sentences are meaningless without unicorns being a member of the universe of discourse.

> > "lo gerku" and "lo nu mi ponse lo gerku" may both refer generically > > to dogs and events of my having dogs, and my wanting them is not > > satisfied just by their existence. > > Well, what more than the existence of lo nu mi ponse lo gerku would youwant; you would then have > your dog after all, so the condition for satisfying the want is met.

Yes, sometimes the existence of X in the world is enough to satisfy my want of X, but sometimes it is not. It depends on what kind of thing X is.

> Did we not work out that for > an event to exist was just to have it obtain?

I thought you said all events existed, whether they take place in the real world or not.

> If not, I propose that for a working rule.

I don't have a problem with equating existence of an event in the world with its taking place in the world. I do have a problem with generalizing from "existence of an event X satisfies my want of X" to "existence of any object X satisfies my want of X".

> > > The point is that abstractions are always already in the world or can be > > > called up without > > > changing the world, so the problems of not having what one wants (etc.) or > > > of not having the right > > > one disappear. > > > > Does that mean one always has what one wants, if what one wants is > > an event? > > You seem to be confusing what you have tried before to keep separate, the world (= universe of > discourse) and "the real world" in which things are actual, etc.

I didn't equate the world with the universe of discourse. You do that (sometimes).

>Things can be in the world but > not in the real world as you well know (and are constantly insisting).

All I said is that things can be members of the universe of discourse without existing in the world.

> > > > > Further, we have to allow that > > > > > the shift can be used in other places under special circumstances: > > > > > {mi viska tu'a lo pavyseljirna} for "I swear I saw a unicorn — or at > > > least > > > > > it sure looked like one to me then." > > > > > > > > I see no logic in that interpretation of {tu'a}. > > > > > > The logic is just that it allows me both to tell you what I experienced and > > > to recognize that it > > > may not be what really happened. > > > > So events that don't take place in the real world can still be visible? > > In at least one sense of "see," yes. We tend to call those illusions or delusion or mirages or ...

But a mirage can be object-type as much as event-type. There is no relevant distinction between them. If I can delusionally see unicorns running, I can equally well delusionally see unicorns. In the sense that {mi viska tu'a lo pavyseljirna} can be true, {mi viska lo pavyseljirna} can also be true. {tu'a} cannot serve as a mark of delusion.

In summary, for {X wants Y} to be true (or false), we need X and Y to be members of the universe of discourse. There is no requirement that Y exists in the world where X exists. There is no requirement for Y to be of event type. If Y is of event type, its existence in the world where X exists will probably be enough to satisfy X's want. If Y is of object type, its existence in the world where X exists will usually not be enough to satisfy X's want. In the case of objects, to satisfy X's want Y not only has to exist in X's world but also be somehow available to X.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Mon 14 of Aug., 2006 16:03 GMT posts: 2388

> On 8/12/06, John E Clifford wrote: > > I don't think that I would want to draw a distinction on the basis of > > form either: between say {no pavyseljirna cu zasti} and {lo pavyseljirna > > na zasti}. But in neither case would I say that in > > saying this he was committed to having unicorns in his universe of > > discourse. > > Whereas for me, both sentences are meaningless without unicorns > being a member of the universe of discourse.


For me, on the other hand, they are not only meaningful but true (if they are not in the domain, they don't exist, a fortiori). By the same reasoning, {lo pavyseljirna cu zasti} is false if there are no unicorns even in the domain.

For the rest here, I see that I have been exceedingly careless (not to say confused) and I apologize for the confusion this caused in turn. I can't seem to keep the technical terms straight between "the real world" (or the world of the highest level in a conversation) and the domain of sicourse (at any given point) which may be more inclusive. Add in the worlds that arise at variouas points in the discourse and there is plenty of rooom for muddle — all of which i seem to have found. So, I'll try to say it completely correctly here.


> > > "lo gerku" and "lo nu mi ponse lo gerku" may both refer generically > > > to dogs and events of my having dogs, and my wanting them is not > > > satisfied just by their existence. > > > > Well, what more than the existence of lo nu mi ponse lo gerku would you want; you would then > have your dog after all, so the condition for satisfying the want is met. > > Yes, sometimes the existence of X in the world is enough to satisfy my want > of X, but sometimes it is not. It depends on what kind of thing X is.

But surely having a dog (the event of your having a dog existing, i.e., occurring,r in the real world) satisfies your want in the real world for a dog. And the mere existence of a dog in the real world does not — only having a dog satisfies what is commonly said to be a want for a dog. That is, "I want a dog" is merely shorthand for "I want to have a dog" and that is reflected in Lojban quite properly (some of the other cases are less clear, admittedly).

> > Did we not work out that for > > an event to exist was just to have it obtain? > > I thought you said all events existed, whether they take place in the real > world or not.

Well, I did say that existing is just what being in the real world meant. What I meant for events is that all are in any domain of discourse. That is, they can be the values of variables. Only those that occur exist in the real world.

> > If not, I propose that for a working rule. > > I don't have a problem with equating existence of an event in the world > with its taking place in the world. I do have a problem with generalizing > from "existence of an event X satisfies my want of X" to "existence of > any object X satisfies my want of X".

I do too and have never suggested it. In fact, my point has been exactly that the existence of dogs does not satisfy my want of a dog and so my want has to be for something more than a dog, namely what does seem to hold, having a dog.

> > > > The point is that abstractions are always already in the world or can be > > > > called up without > > > > changing the world, so the problems of not having what one wants (etc.) or > > > > of not having the right > > > > one disappear. > > > > > > Does that mean one always has what one wants, if what one wants is > > > an event? > > > > You seem to be confusing what you have tried before to keep separate, the world (= universe of > > discourse) and "the real world" in which things are actual, etc. > > I didn't equate the world with the universe of discourse. You do > that (sometimes).

Well, I don't actually confuse the things, but I do mix up the terminology a lot.

> >Things can be in the world but > > not in the real world as you well know (and are constantly insisting). > > All I said is that things can be members of the universe of discourse > without existing in the world.

Precisely.

> > > > > > Further, we have to allow that > > > > > > the shift can be used in other places under special circumstances: > > > > > > {mi viska tu'a lo pavyseljirna} for "I swear I saw a unicorn — or at > > > > least > > > > > > it sure looked like one to me then." > > > > > > > > > > I see no logic in that interpretation of {tu'a}. > > > > > > > > The logic is just that it allows me both to tell you what I experienced and > > > > to recognize that it > > > > may not be what really happened. > > > > > > So events that don't take place in the real world can still be visible? > > > > In at least one sense of "see," yes. We tend to call those illusions or delusion or mirages or > ... > > But a mirage can be object-type as much as event-type. There is no > relevant distinction between them. If I can delusionally see unicorns running, > I can equally well delusionally see unicorns. In the sense that {mi viska > tu'a lo pavyseljirna} can be true, {mi viska lo pavyseljirna} can also be true. > {tu'a} cannot serve as a mark of delusion. > {tu'a} is vague as to what abstraction it is — the appropriate one seems to be the best guess. So, it need not be {nu}. In fact there once was an abstractor exactly for (potential) illusion cases but I think it is gone now, giving the content of one's experience without commitment to veridicality. {li'i} is what is left of it, I think, but it seems to have a different meaning now (insofar as I can figure out what it means). In any case, I would take "I see a unicorn" in a context where the real world of the moment does not contain unicorns as I gave it above — just to be on the safe side — even if the universe of discourse did contain unicorns.

> In summary, for {X wants Y} to be true (or false), we need X and Y to be > members of the universe of discourse. There is no requirement that Y exists > in the world where X exists. There is no requirement for Y to be of event type. > If Y is of event type, its existence in the world where X exists will probably > be enough to satisfy X's want. If Y is of object type, its existence > in the world > where X exists will usually not be enough to satisfy X's want. In the case of > objects, to satisfy X's want Y not only has to exist in X's world but also > be somehow available to X.

In counter summary, for "X wants Y" to be true, there must be some event involving Y such that,1) if tu'a Y were to occur, X's want would be satisfied and 2) if X's want is satisfied then tu'a Y has occurred. Being someohow available to X IS an event, the one wanted.



Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Mon 14 of Aug., 2006 20:28 GMT On 8/14/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > On 8/12/06, John E Clifford wrote: > > > I don't think that I would want to draw a distinction on the basis of > > > form either: between say {no pavyseljirna cu zasti} and {lo pavyseljirna > > > na zasti}. But in neither case would I say that in > > > saying this he was committed to having unicorns in his universe of > > > discourse. > > > > Whereas for me, both sentences are meaningless without unicorns > > being a member of the universe of discourse. > > For me, on the other hand, they are not only meaningful but true (if they are not in the domain, > they don't exist, a fortiori). By the same reasoning, {lo pavyseljirna cu zasti} is false if there > are no unicorns even in the domain.

(1) X does (not) exist.

(2) The referent of "X" is a member of the domain of discourse of utterance (1).

(1) is a sentence in some object language that we are analyzing. (2) is a sentence in the metalanguage we are using to describe the object language that contains (1). Comparing their truths is somewhat weird, given that they belong to different language systems. In order for (1) to be meaningful in the object language (whether true or false), (2) must be true in the metalanguage.

Sentence (2) is not something that the speaker of (1) has to commit to, simply because the domain of discourse of utterance (1) is not something that needs to be a member of the domain of discourse of utterance (1). A speaker only talks about the things that are members of the universe of discourse (by definition of universe of discourse), but the speaker does not talk about the universe of discourse itself. Only the metalinguistic analysis introduces this universe of discourse.

> > > > "lo gerku" and "lo nu mi ponse lo gerku" may both refer generically > > > > to dogs and events of my having dogs, and my wanting them is not > > > > satisfied just by their existence. > > > > > > Well, what more than the existence of lo nu mi ponse lo gerku would you want; you would then > > have your dog after all, so the condition for satisfying the want is met. > > > > Yes, sometimes the existence of X in the world is enough to satisfy my want > > of X, but sometimes it is not. It depends on what kind of thing X is. > > But surely having a dog (the event of your having a dog existing, i.e., occurring,r in the real > world) satisfies your want in the real world for a dog. And the mere existence of a dog in the > real world does not — only having a dog satisfies what is commonly said to be a want for a dog. > That is, "I want a dog" is merely shorthand for "I want to have a dog" and that is reflected in > Lojban quite properly (some of the other cases are less clear, admittedly).

The fact that "I want a dog" and "I want to have a dog" are more or less synonymous is not reason enough to make the first a shorthand for the second. While we are of course free to decree that the first form ({mi djica lo gerku}) is incorrect in Lojban, there is nothing in logic or in Logic that requires that we do that. And my point in all of this is that doing so does not solve the "problem" it is set up to solve anyway.


> > I do have a problem with generalizing > > from "existence of an event X satisfies my want of X" to "existence of > > any object X satisfies my want of X". > > I do too and have never suggested it.

That's what you seem to be doing here:

> In fact, my point has been exactly that the existence of > dogs does not satisfy my want of a dog and so my want has to be for > something more than a dog, > namely what does seem to hold, having a dog.

I don't follow the step:

"the (mere) existence of dogs does not satisfy my want of a dog" => my want has to be for something more than a dog.

The only way I can see that following is if you assume that the mere existence of X must be enough to satisfy the want of X.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Mon 14 of Aug., 2006 23:13 GMT posts: 2388

> On 8/14/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > On 8/12/06, John E Clifford wrote: > > > > I don't think that I would want to draw a distinction on the basis of > > > > form either: between say {no pavyseljirna cu zasti} and {lo pavyseljirna > > > > na zasti}. But in neither case would I say that in > > > > saying this he was committed to having unicorns in his universe of > > > > discourse. > > > > > > Whereas for me, both sentences are meaningless without unicorns > > > being a member of the universe of discourse. > > > > For me, on the other hand, they are not only meaningful but true (if they are not in the > domain, > > they don't exist, a fortiori). By the same reasoning, {lo pavyseljirna cu zasti} is false if > there > > are no unicorns even in the domain. > > (1) X does (not) exist. > > (2) The referent of "X" is a member of the domain of discourse of utterance (1). > > (1) is a sentence in some object language that we are analyzing. (2) is > a sentence in the metalanguage we are using to describe the object > language that contains (1). Comparing their truths is somewhat weird, > given that they belong to different language systems. In order for (1) to > be meaningful in the object language (whether true or false), (2) must be > true in the metalanguage.

Well, we just disagree about that. My system works OK (and is spared a lot of meaningless sentences). Yours works OK, too (but with either meaningless sentences or a bloated ontology). And yours is subject to the paradox that I can't say that there are no whatevers, even if there aren't. "No one can get across that chasm""Superman can""Well, we're not talking about Superman here" (vs. "Not even him; the bottom is lined with kryptonite." or "But he doesn't exist."). So there are differences, though I am not sure what difference those differences make.

> Sentence (2) is not something that the speaker of (1) has to commit to, > simply because the domain of discourse of utterance (1) is not something > that needs to be a member of the domain of discourse of utterance (1). A > speaker only talks about the things that are members of the universe of > discourse (by definition of universe of discourse), but the speaker does not > talk about the universe of discourse itself. Only the metalinguistic analysis > introduces this universe of discourse.

All perfectly true, but not relevant so far: no one has suggested that the speaker of 1 also utters (or even thinks) 2. "There are no Xs" is a perfectly good object language sentence which will be true (moving up to the metalanguage now) just in case the extension of "X," however you want to define that, in the u8niverse of discourse is null. You've set it up so that it can't ever be true, if uttered. And that seem odd to me.

> > > > > "lo gerku" and "lo nu mi ponse lo gerku" may both refer generically > > > > > to dogs and events of my having dogs, and my wanting them is not > > > > > satisfied just by their existence. > > > > > > > > Well, what more than the existence of lo nu mi ponse lo gerku would you want; you would > then > > > have your dog after all, so the condition for satisfying the want is met. > > > > > > Yes, sometimes the existence of X in the world is enough to satisfy my want > > > of X, but sometimes it is not. It depends on what kind of thing X is. > > > > But surely having a dog (the event of your having a dog existing, i.e., occurring, in the > real > > world) satisfies your want in the real world for a dog. And the mere existence of a dog in > the > > real world does not — only having a dog satisfies what is commonly said to be a want for a > dog. > > That is, "I want a dog" is merely shorthand for "I want to have a dog" and that is reflected > in > > Lojban quite properly (some of the other cases are less clear, admittedly). > > The fact that "I want a dog" and "I want to have a dog" are more or less > synonymous is not reason enough to make the first a shorthand for the > second. While we are of course free to decree that the first form > ({mi djica lo gerku}) is incorrect in Lojban, there is nothing in logic > or in Logic that requires that we do that. And my point in all of this is > that doing so does not solve the "problem" it is set up to solve anyway.

My point is that I can figure out how my want is satisfied by what I want existing in the event case, but not in the bare dog case (every attempt comes down to another event). That seems a point in favor of the event scenario, though perhaps not decisive. It also turns out that, if I want events, then I can do generalization validly, which I cannot do with bare things. That is decisive (for a logical language) for something like events, though not necessarily events, of course (but I don't know of another candidate). What part of the problem is left unsolved: generalization works, relieving want works, Leibniz's Law works (events are intensional, so, even though the referent of "unicorn" is the same as the referent of "centaur," seeing a unicorn is not the same event as seeing a centaur)?

> > > > I do have a problem with generalizing > > > from "existence of an event X satisfies my want of X" to "existence of > > > any object X satisfies my want of X". > > > > I do too and have never suggested it. > > That's what you seem to be doing here: > > > In fact, my point has been exactly that the existence of > > dogs does not satisfy my want of a dog and so my want has to be for > > something more than a dog, > > namely what does seem to hold, having a dog.

But the point of that (it is a reductio) is just that the existence of an X does NOT satisfy my want of an X. Since it does not, what I want (what would satisfy my want) has to be something else.

> I don't follow the step: > > "the (mere) existence of dogs does not satisfy my want of a dog" > => my want has to be for something more than a dog. > > The only way I can see that following is if you assume that the mere > existence of X must be enough to satisfy the want of X.

Why would I assume what I am working to deny? At most, you could object to the "more" and say I have only justified "other than." But I assume that if I want a dog, a dog is going to enter into the saisfaction of that want somehow. The "somehow" introduces the "more," since it cannot (I think) be less than a dog involved. And, of course, I think that what I really want is what would satisfy that want, regardless of what I say.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Tue 15 of Aug., 2006 15:18 GMT On 8/14/06, John E Clifford wrote: > > "There are no Xs" is a perfectly good object language sentence which > will be true (moving up to the metalanguage now) just in case the > extension of "X," however you want to define that, in the u8niverse of > discourse is null. You've set it up so that it can't > ever be true, if uttered. And that seem odd to me.

{no da broda} is true whenever nothing in the universe of discourse is a broda.

{lo broda na zasti} is true whenever the brodas (which must be members of the universe of discourse for us to refer to them with {lo broda}) don't exist.

In either case, you get true when you want true. How have I set things up so that "there are no Xs" can't ever be true?

What I would find extremely odd is that in order for {lo broda na zasti} to be true, you would need {lo broda} to have no referent. In that case, I wouldn't be able to tell what it is that is being claimed not to exist.


> My point is that I can figure out how my want is satisfied by what I want > existing in the event case, but not in the bare dog case (every attempt > comes down to another event).

How is figuring out that "I want X" is satisfied when X exists different from figuring out that "I want X" is satisfied when X is available to me? Isn't the condition of X existing as much an event as the condition of X being available to me?

> That seems a > point in favor of the event scenario, though perhaps not decisive. It also turns out that, if I > want events, then I can do generalization validly, which I cannot do with bare things.

Whatever you can do with one you can do with the other. If what worries you is that from "I want object X" you cannot conclude that there is at least one among all the objects X such that that is the one that I want, the same applies to "I want event X", you cannot conclude from it that there is at least one among all the events X such that that is the the one I want. So you must allow generic reference to the events X, in which case there is no problem, but the same applies to generic reference to objects X.


> > I don't follow the step: > > > > "the (mere) existence of dogs does not satisfy my want of a dog" > > => my want has to be for something more than a dog. > > > > The only way I can see that following is if you assume that the mere > > existence of X must be enough to satisfy the want of X. > > Why would I assume what I am working to deny? At most, you could object to the "more" and say I > have only justified "other than." But I assume that if I want a dog, a dog is going to enter into > the saisfaction of that want somehow. The "somehow" introduces the "more," since it cannot (I > think) be less than a dog involved. And, of course, I think that what I really want is what would > satisfy that want, regardless of what I say.

So, if you really want X and X exists, your want is thereby satisfied?

That argument seems to me to rely on the assumption that when you _really_ want something, your want will be satisfied by the something existing.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Tue 15 of Aug., 2006 16:42 GMT posts: 2388

> On 8/14/06, John E Clifford wrote: > > > > "There are no Xs" is a perfectly good object language sentence which > > will be true (moving up to the metalanguage now) just in case the > > extension of "X," however you want to define that, in the universe of > > discourse is null. You've set it up so that it can't > > ever be true, if uttered. And that seem odd to me. > > {no da broda} is true whenever nothing in the universe of discourse is a broda.

I'm glad to hear it. You seem to have said otherwise earlier, namely that since brodas were mentioned, there had to be some.

> {lo broda na zasti} is true whenever the brodas (which must be members > of the universe of discourse for us to refer to them with {lo broda}) > don't exist. > > In either case, you get true when you want true. How have I set things > up so that "there are no Xs" can't ever be true?

Well, you have now clearly (as opposed to less so)avoided it. But what about (what I take to be equivalent) {no da me lo broda}?

> What I would find extremely odd is that in order for {lo broda na zasti} to > be true, you would need {lo broda} to have no referent. In that case, > I wouldn't be able to tell what it is that is being claimed not to exist.

To tell what something is is about the sense of the expression, not about its referents. Senses are another range of things altogether (rather like properties)and are always available. I know what {broda} means and thus know what (sort of thing) I am denying to be in the domain (I have to know this to affirm they are in the domain as well). > > > My point is that I can figure out how my want is satisfied by what I want > > existing in the event case, but not in the bare dog case (every attempt > > comes down to another event). > > How is figuring out that "I want X" is satisfied when X exists different from > figuring out that "I want X" is satisfied when X is available to me? Isn't > the condition of X existing as much an event as the condition of X being > available to me?

Well, the difference is that X merely existing (without any specified association to me) does not generally satisfy my want, where as X being available to me does. They are both events, if you will, but one works and the other doesn't to satisfy me.

> > That seems a > > point in favor of the event scenario, though perhaps not decisive. It also turns out that, if > I > > want events, then I can do generalization validly, which I cannot do with bare things. > > Whatever you can do with one you can do with the other. If what worries you > is that from "I want object X" you cannot conclude that there is at least one > among all the objects X such that that is the one that I want, the same applies > to "I want event X", you cannot conclude from it that there is at > least one among > all the events X such that that is the the one I want. So you must allow > generic reference to the events X, in which case there is no problem, but the > same applies to generic reference to objects X.


On the contrary, since I want an event and all events are in the domain, the one I want is in the domain. If I want a dog, unless all possible dogs are in the domain (which is not guaranteed, as the events are), the one I want may not be there, and that mere "may" is enough to declare the inference invalid/ And, of course, I may want something of which none are in the domain(well, you won't allow that, but that particular prohibition doesn't help here). What is "generic reference?" > > > > I don't follow the step: > > > > > > "the (mere) existence of dogs does not satisfy my want of a dog" > > > => my want has to be for something more than a dog. > > > > > > The only way I can see that following is if you assume that the mere > > > existence of X must be enough to satisfy the want of X. > > > > Why would I assume what I am working to deny? At most, you could object to the "more" and say > I > > have only justified "other than." But I assume that if I want a dog, a dog is going to enter > into > > the saisfaction of that want somehow. The "somehow" introduces the "more," since it cannot (I > > think) be less than a dog involved. And, of course, I think that what I really want is what > would > > satisfy that want, regardless of what I say. > > So, if you really want X and X exists, your want is thereby satisfied?

I suppose so, but I can't think of a case where all I want is X (some object, not an event) rather than some event involving it (and usually me). Even if I did come up with such a rare case, I would probably say "I want X to exist."

> That argument seems to me to rely on the assumption that when you _really_ > want something, your want will be satisfied by the something existing.

As noted several times, the argument is exactly the opposite. If all I want is X, then X's existing would satisfy the want, but it doesn't, so I don't really want just X. I gather your point is that the initial conditional is not quite right: "if I really want X, then some more complex state of affairs involving X occurring satisfies my want" That is, of course, what I come to, adding only that, since what satisfies my want is not the thing but the state of affairs involving that thing and since satisfaction of a want is getting what one wants, then it must be that it is really the state of affairs that one wants. That is, your claim seems to concede mine.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Tue 15 of Aug., 2006 19:36 GMT On 8/15/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > {no da broda} is true whenever nothing in the universe of discourse is a broda. > > I'm glad to hear it. You seem to have said otherwise earlier, namely that since brodas were > mentioned, there had to be some.

No, I'm pretty sure what I said was that if you use a referring term (such as {lo broda}) then its referent must be a member of the universe of discourse.

> But what about (what I take to be > equivalent) {no da me lo broda}?

That's either false, or you have to assume some implicit contextual restriction, such as {no da poi nenri le kumfa/poi zasti/poi ... cu me lo broda}.

> > What I would find extremely odd is that in order for {lo broda na zasti} to > > be true, you would need {lo broda} to have no referent. In that case, > > I wouldn't be able to tell what it is that is being claimed not to exist. > > To tell what something is is about the sense of the expression, not about its referents. Senses > are another range of things altogether (rather like properties)and are always available. I know > what {broda} means and thus know what (sort of thing) I am denying to be in the domain (I have to > know this to affirm they are in the domain as well).

But {zasti} is not about membership in a metalinguistic domain. It's a predicate that must be predicated of something.


> > > My point is that I can figure out how my want is satisfied by what I want > > > existing in the event case, but not in the bare dog case (every attempt > > > comes down to another event). > > > > How is figuring out that "I want X" is satisfied when X exists different from > > figuring out that "I want X" is satisfied when X is available to me? Isn't > > the condition of X existing as much an event as the condition of X being > > available to me? > > Well, the difference is that X merely existing (without any specified association to me) does not > generally satisfy my want, where as X being available to me does. They are both events, if you > will, but one works and the other doesn't to satisfy me.

That doesn't address what I said. Isn't the condition of (event) X existing as much an event as the condition of (object) X being available to me? In other words, why is it more of a problem having to consider the condition "object X is available to me" as a satisfaction condition for "I want X" than having to consider the condition "event X exists (=occurs)" as a satisfaction condition for "I want X". They seem to be exactly parallel.


> > Whatever you can do with one you can do with the other. If what worries you > > is that from "I want object X" you cannot conclude that there is at least one > > among all the objects X such that that is the one that I want, the same applies > > to "I want event X", you cannot conclude from it that there is at > > least one among > > all the events X such that that is the the one I want. So you must allow > > generic reference to the events X, in which case there is no problem, but the > > same applies to generic reference to objects X. > > On the contrary, since I want an event and all events are in the domain, the one I want is in the > domain.

But there could be many events of "I have a dog" in the domain. When I want to have a dog, it is not the case that some particular one of them is such that I want that event. I want any of them. Just as with "I want a dog".

> If I want a dog, unless all possible dogs are in the domain (which is not guaranteed, as > the events are), the one I want may not be there, and that mere "may" is enough to declare the > inference invalid/

Whereas when *I* want a dog, the single referent of "a dog" is in the domain, just as when I want to have a dog the single referent of "(my) having a dog" is in the domain. This is irrespective of how many dogs exist in the world, how many events of my having a dog exist in the world, how many events of my having a dog might or might not exist in the world, etc.

> And, of course, I may want something of which none are in the domain(well, you > won't allow that, but that particular prohibition doesn't help here).

I have no prohibitions against introducing anything in the domain of discourse. I thought it was you who imposed those kind of prohibitions.

> What is "generic reference?"

For example {lo nu mi ponse lo gerku} in {mi djica lo nu mi ponse lo gerku} refers generically to events of my having a dog. It would not (usually) refer to any one particular event of my having a dog.

> > So, if you really want X and X exists, your want is thereby satisfied? > > I suppose so, but I can't think of a case where all I want is X (some object, not an event) rather > than some event involving it (and usually me).

I can't think of a case where seeing an object does not involve seeing the object in some situation or other either. That doesn't mean that I can't say {mi viska lo gerku} instead of the (also allowed) {mi viska tu'a lo gerku}.

> Even if I did come up with such a rare case, I > would probably say "I want X to exist."

That's your choice, and it's always available. All I'm saying is that there is no logical reason to restrict the x2 of {djica} to events. No more than there is to restrict the x2 of {viska} to events. It could be done, but for no gain.

> > That argument seems to me to rely on the assumption that when you _really_ > > want something, your want will be satisfied by the something existing. > > As noted several times, the argument is exactly the opposite. If all I want is X, then X's > existing would satisfy the want, but it doesn't, so I don't really want just X. I gather your > point is that the initial conditional is not quite right: "if I really want X, then some more > complex state of affairs involving X occurring satisfies my want" That is, of course, what I come > to, adding only that, since what satisfies my want is not the thing but the state of affairs > involving that thing and since satisfaction of a want is getting what one wants, then it must be > that it is really the state of affairs that one wants. That is, your claim seems to concede mine.

Which claim of mine concedes which claim of yours?

My position: If I want object X, and I get X, then my want is satisfied. If I want object X, and X exists, but I don't get X, then my want is not satisfied.

The step that makes no sense to me is "since what satisfies my want is not the thing but the state of affairs involving that thing and since satisfaction of a want is getting what one wants, then it must be that it is really the state of affairs that one wants". Is that meant as a syllogism? The premises seem true, but the conclusion does not seem to follow.

True premise: ~Satisfies(W(x),x) & Satisfies(W(x),G(x)) True premise: Satisfaction(W(x)) = G(x) ?Conclusion: W(G(x)) & ~W(x)

Could you explain how the (false for me) conclusion follows from the (true for both of us) premises?

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Tue 15 of Aug., 2006 20:44 GMT posts: 2388

> On 8/15/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > {no da broda} is true whenever nothing in the universe of discourse is a broda. > > > > I'm glad to hear it. You seem to have said otherwise earlier, namely that since brodas were > > mentioned, there had to be some. > > No, I'm pretty sure what I said was that if you use a referring term (such > as {lo broda}) then its referent must be a member of the universe of > discourse. > > > But what about (what I take to be > > equivalent) {no da me lo broda}? > > That's either false, or you have to assume some implicit contextual > restriction, such as {no da poi nenri le kumfa/poi zasti/poi ... > cu me lo broda}.

So the paradox is still there, just not in its simplest form.

> > > What I would find extremely odd is that in order for {lo broda na zasti} to > > > be true, you would need {lo broda} to have no referent. In that case, > > > I wouldn't be able to tell what it is that is being claimed not to exist. > > > > To tell what something is is about the sense of the expression, not about its referents. > Senses > > are another range of things altogether (rather like properties)and are always available. I > know > > what {broda} means and thus know what (sort of thing) I am denying to be in the domain (I have > to > > know this to affirm they are in the domain as well). > > But {zasti} is not about membership in a metalinguistic domain. It's a predicate > that must be predicated of something.

Not in the strong sense you seem to mean. There are many predicates which apply to nothing in the domain (unless of course, every possible — and impossible — thing is in the domain, which you don't seem to want. "unicorn" is a perfectly good predicate at all times and in all domains; sometimes its referent is unicorns, sometimes it has no referent. This difference does not affect its sense nor make it more or less a predicate. I suppose that {zasti} is a fairly rare kind of predicate in that it always has a referent (well, maybe one case not, but that one is very odd in other ways as well).

> > > > > My point is that I can figure out how my want is satisfied by what I want > > > > existing in the event case, but not in the bare dog case (every attempt > > > > comes down to another event). > > > > > > How is figuring out that "I want X" is satisfied when X exists different from > > > figuring out that "I want X" is satisfied when X is available to me? Isn't > > > the condition of X existing as much an event as the condition of X being > > > available to me? > > > > Well, the difference is that X merely existing (without any specified association to me) does > not > > generally satisfy my want, where as X being available to me does. They are both events, if you > > will, but one works and the other doesn't to satisfy me. > > That doesn't address what I said. Isn't the condition of (event) X existing as > much an event as the condition of (object) X being available to me? In other > words, why is it more of a problem having to consider the condition > "object X is available to me" as a satisfaction condition for "I want X" > than having to consider the condition "event X exists (=occurs)" as > a satisfaction condition for "I want X". They seem to be exactly parallel.

I agree that I have trouble figuring out what your point is here. This seems to add little light, but maybe some. Lets see. X being available to me, the existence of the event of X being available to me, does satisfy the want I loosely express as "I want X." The existence of X does not satisfy the want I express by saying "I want X." Do we agree so far? But now what. It is claimed, I gather, that there is a problem with the object X being available to me that there is not with event X existing as satisfactions for my want. But the first X is not an event (it's a dog in the original example) and the event that works to satisfy my want is exactly that dog being available to me. That is, the two things you are really talking aobut — as far as I can understand — are in fact the same thing: that the event of a dog being available to me exists.So, obviously there is no problem with one that there is not with the other and coversely. Nor I can I find anything in what I have said that suggests there is. Sorry if I am being obtuse here, but I just don't see what you are driving at. The problem I have been pointing to is the inadequacy of X existing (or the event of X existing existing) to satisfy a want and the adequacy of X being available to me (or the vent of X being available to me existing). the crucial difference is the "available to me" part.

> > > > Whatever you can do with one you can do with the other. If what worries you > > > is that from "I want object X" you cannot conclude that there is at least one > > > among all the objects X such that that is the one that I want, the same applies > > > to "I want event X", you cannot conclude from it that there is at > > > least one among > > > all the events X such that that is the the one I want. So you must allow > > > generic reference to the events X, in which case there is no problem, but the > > > same applies to generic reference to objects X. > > > > On the contrary, since I want an event and all events are in the domain, the one I want is in > the > > domain. > > But there could be many events of "I have a dog" in the domain. When I want > to have a dog, it is not the case that some particular one of them is such > that I want that event. I want any of them. Just as with "I want a dog".

Well, one of the weird things about events is that they come in trees. If, indeed, there a bunc of events any one of which would satisfy my want, there is also an event that occurs exactly when any one of these events occurs — the meta-event to just those events. This is probably why we want them all in. The same is not necessarily (nor at all, I would say, though I know you disagree) true of dogs. With dogs, you have to be able to say "I want that one" (you can say it as often as you like, that is not the issue) and there is no guarantee — unless you get all possible dogs in the domain — that there is even one you would say that of. That is, there is no contradiction or personal inconsistency in someone saying he wants a dog and that no dog in the world satisfies him. So, from "I want a dog" we cannot infer that there is a dog I want.

> > If I want a dog, unless all possible dogs are in the domain (which is not guaranteed, as > > the events are), the one I want may not be there, and that mere "may" is enough to declare the > > inference invalid. > > Whereas when *I* want a dog, the single referent of "a dog" is in the domain, > just as when I want to have a dog the single referent of "(my) having a dog" > is in the domain. This is irrespective of how many dogs exist in the world, > how many events of my having a dog exist in the world, how many events > of my having a dog might or might not exist in the world, etc.

The trouble is that "a dog" needn't have a single referent in these contexts — not in the basic domain anyhow, since I can want a dog even if there are no dogs in the universe of discourse (and saying "a dog" doesn't introduce one there, only in the wish world, which is different at least). Presumably among the events of my having a dog are all the particular events of my having this dog and that dog and so on, where what goes on in the event world is independent of what goes on in the domain of discourse. Esistence in the world need not enter into it at all except subjunctively: what would satisfy me were it to occur.

> > And, of course, I may want something of which none are in the domain(well, you > > won't allow that, but that particular prohibition doesn't help here). > > I have no prohibitions against introducing anything in the domain of discourse. > I thought it was you who imposed those kind of prohibitions.

The point is, if I say "I want an X" you would claim that there then is an X in the domain of discourse and I just don't see that: it certainly isn't necessary and it does not seem to me to be even plausible.

> > What is "generic reference?" > > For example {lo nu mi ponse lo gerku} in {mi djica lo nu mi ponse lo gerku} > refers generically to events of my having a dog. It would not (usually) refer > to any one particular event of my having a dog.

Well, I suppose that the referents of {lo nu mi ponse lo gerku} are events — several of them, even though there is one that could do for all. The phrase just refers to them in the ordinary way that indefinite descriptions refer to several (well, at least one) things. This is not a special kind of referentce, as "generic reference" seems to imply, but just ordinary reference to things of the approriate sort. {lo nu mi ponse lo gerku} refers to events of my having a dog in the same way as {lo tadni} refers to students. The trick here is just that, while we can be sure there are the right events, we can't be sure there are dogs, let alone the right ones.

> > > So, if you really want X and X exists, your want is thereby satisfied? > > > > I suppose so, but I can't think of a case where all I want is X (some object, not an event) > rather > > than some event involving it (and usually me). > > I can't think of a case where seeing an object does not involve seeing > the object in some situation or other either. That doesn't mean that I can't > say {mi viska lo gerku} instead of the (also allowed) {mi viska tu'a lo gerku}. > > > Even if I did come up with such a rare case, I > > would probably say "I want X to exist." > > That's your choice, and it's always available. All I'm saying is that there is > no logical reason to restrict the x2 of {djica} to events. No more than there > is to restrict the x2 of {viska} to events. It could be done, but for no gain.


Well, there is one reason for the restriction in the case of {djica} and not of {viska}: the fact that generalization goes through with the restriction and not without. For a logical language, that is enough.

> > > That argument seems to me to rely on the assumption that when you _really_ > > > want something, your want will be satisfied by the something existing. > > > > As noted several times, the argument is exactly the opposite. If all I want is X, then X's > > existing would satisfy the want, but it doesn't, so I don't really want just X. I gather your > > point is that the initial conditional is not quite right: "if I really want X, then some more > > complex state of affairs involving X occurring satisfies my want" That is, of course, what I > come > > to, adding only that, since what satisfies my want is not the thing but the state of affairs > > involving that thing and since satisfaction of a want is getting what one wants, then it must > be > > that it is really the state of affairs that one wants. That is, your claim seems to concede > mine. > > Which claim of mine concedes which claim of yours? > > My position: > If I want object X, and I get X, then my want is satisfied. > If I want object X, and X exists, but I don't get X, then my want is > not satisfied. > > The step that makes no sense to me is "since what satisfies my want is > not the thing but the state of affairs involving that thing and since > satisfaction > of a want is getting what one wants, then it must be that it is really > the state > of affairs that one wants". Is that meant as a syllogism? The premises seem > true, but the conclusion does not seem to follow. > > True premise: ~Satisfies(W(x),x) & Satisfies(W(x),G(x)) > True premise: Satisfaction(W(x)) = G(x) > ?Conclusion: W(G(x)) & ~W(x)

Well, I am not sure just what the symbolism means here; since they don't seem to correspond in any natural way to the words earli9er (we are up several levels here). It appears that the first says "x does not satisfy wanting x and getting x satisfies wanting x" But that puts getting x on a par witrh wanting x, whereas it is on a par with x itself, as an object of desire. Or rather, the notion of getting comes in at two different levels (so call one — on a par with wanting --m getting and the other having.) So what you want to say is getting x doesn; 't satisfy wanting x but getting having x does satisfy wanting x. Then what you really want is what satisfies the want. So, when you say you want x, what you really want is having x.


> Could you explain how the (false for me) conclusion follows from > the (true for both of us) premises? No false steps, just bad notation so that ti doesn't say what is wanted (I think adequate notation would just confuse this issue more).


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Tue 15 of Aug., 2006 21:53 GMT On 8/15/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > On 8/15/06, John E Clifford wrote: > > > But what about (what I take to be > > > equivalent) {no da me lo broda}? > > > > That's either false, or you have to assume some implicit contextual > > restriction, such as {no da poi nenri le kumfa/poi zasti/poi ... > > cu me lo broda}. > > So the paradox is still there, just not in its simplest form.

I don't see the paradox there.

> I agree that I have trouble figuring out what your point is here.

My point is that {mi djica lo gerku} has none of the problems you attribute to it.

... > Well, one of the weird things about events is that they come in trees. If, indeed, there a bunc > of events any one of which would satisfy my want, there is also an event that occurs exactly when > any one of these events occurs — the meta-event to just those events. This is probably why we > want them all in. The same is not necessarily (nor at all, I would say, though I know you > disagree) true of dogs.

Why does not the same thing work for dogs? What is it about events that makes them susceptible to this kind of treatment, and which dogs lack?

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Wed 16 of Aug., 2006 01:28 GMT posts: 2388

> On 8/15/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > On 8/15/06, John E Clifford wrote: > > > > But what about (what I take to be > > > > equivalent) {no da me lo broda}? > > > > > > That's either false, or you have to assume some implicit contextual > > > restriction, such as {no da poi nenri le kumfa/poi zasti/poi ... > > > cu me lo broda}. > > > > So the paradox is still there, just not in its simplest form. > > I don't see the paradox there.

That you can't say {no da me lo broda} truthfully, even though it might be true so long as unsaid.

> > I agree that I have trouble figuring out what your point is here. > > My point is that {mi djica lo gerku} has none of the problems you attribute > to it.

What does it lack? Generalization does not apply nor does interchange of equivalents. And, of course, it turns out that getting what you want is not satisfying.

> ... > > Well, one of the weird things about events is that they come in trees. If, indeed, there a > bunc > > of events any one of which would satisfy my want, there is also an event that occurs exactly > when > > any one of these events occurs — the meta-event to just those events. This is probably why > we > > want them all in. The same is not necessarily (nor at all, I would say, though I know you > > disagree) true of dogs. > > Why does not the same thing work for dogs? What is it about events that > makes them susceptible to this kind of treatment, and which dogs lack?

I'm not sure which you are talking about: there is no guarantee that the dog you want is among the ones in the domain as there is for events, ther is no summary dog as there is for events. dogs are comntingent beings, events are not — you can have a universe without dogs but not without events (by definition, admittedly, but we are talking about the language in which it is a given).


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 16 of Aug., 2006 13:53 GMT On 8/15/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > On 8/15/06, John E Clifford wrote: > > > --- Jorge Llambas wrote: > > > > On 8/15/06, John E Clifford wrote: > > > > > But what about (what I take to be > > > > > equivalent) {no da me lo broda}? > > > > > > > > That's either false, or you have to assume some implicit contextual > > > > restriction, such as {no da poi nenri le kumfa/poi zasti/poi ... > > > > cu me lo broda}. > > > > > > So the paradox is still there, just not in its simplest form. > > > > I don't see the paradox there. > > That you can't say {no da me lo broda} truthfully, even though it might > be true so long as unsaid.

But you can say {no da me lo broda} truthfully. All you need is some implicit contextual restriction.

There's nothing paradoxical in the fact that using a referring term introduces its referent as a member of the domain of discourse.


> > > I agree that I have trouble figuring out what your point is here. > > > > My point is that {mi djica lo gerku} has none of the problems you attribute > > to it. > > What does it lack? Generalization does not apply nor does interchange of equivalents. And, of > course, it turns out that getting what you want is not satisfying.

Let's take them one by one.

(1) Existential generalization applies to {mi djica lo broda} as much as to {mi djica lo nu ponse lo broda}.

>From {mi djica lo gerku} it follows that {mi djica su'o da}. >From {mi djica lo nu mi ponse lo gerku} it follows that {mi djica su'o da}.

All this means is that {lo gerku}/{lo nu mi ponse lo gerku} must have in each case a referent that is a member of the universe of discourse. In neither case does this mean that said referent must be something that satisfies the {zasti} predicate. It is possible to want things (be it objects or events) that satisfy {na zasti}. There is no difference in generalization that I can see between objects and events.

(2) Interchange of equivalents. Mary told me yesterday that she wanted a dog. There is a dog now in front of us. {lo gerku} and {lo danlu poi klesi tu} are equivalents in this context. {lo nu ponse lo gerku} and {lo nu ponse lo danlu poi klesi tu} are equivalents in this context. Mary could not have used the words {lo danlu poi klesi tu} to tell me what she wanted because there were no dogs around when she told me what she wanted. (She might not have used the words {lo gerku} either.)

{la meris djica lo gerku} and {la meris djica lo danlu poi klesi tu} are equivalent. {la meris djica lo nu ponse lo gerku} and {la meris djica lo nu ponse lo danlu poi klesi tu} are equivalent.

There is no difference between objects and events here.

(3) "Getting what you want is not satisfying"

That makes no sense. Similarly with the statement you had made before: > As they say, "If you want a dog, you are in luck, because there are dogs."

It is perfectly possible that there is X, that you want X, that you don't get X, and therefore your want is not satisfied. There is nothing contradictory or inconsistent in any of that.

> > ... > > > Well, one of the weird things about events is that they come in trees. If, indeed, there a > > bunc > > > of events any one of which would satisfy my want, there is also an event that occurs exactly > > when > > > any one of these events occurs — the meta-event to just those events. This is probably why > > we > > > want them all in. The same is not necessarily (nor at all, I would say, though I know you > > > disagree) true of dogs. > > > > Why does not the same thing work for dogs? What is it about events that > > makes them susceptible to this kind of treatment, and which dogs lack? > > I'm not sure which you are talking about: there is no guarantee that the dog you want is among the > ones in the domain as there is for events, ther is no summary dog as there is for events.

The dog you want is a member of the domain of discourse as soon as you mention it, just as for the event you want. That both are in the domain of discourse as soon as you mention them does not mean in either case that your want will be satisfied. The satisfaction condition is not related to membership in the domain of discourse in either case.

> dogs > are comntingent beings, events are not — you can have a universe without dogs but not without > events (by definition, admittedly, but we are talking about the language in which it is a given).

I don't see how dogs are more or less contingent than events. An event of having a dog does not exist (it does not occur) in a world where dogs don't exist. Both dogs and events of having dogs are automatically guaranteed to be members of the universe of discourse for any discourse in which a want of dogs or a want of having dogs is expressed. That's all that's required. No difference between dogs and events here.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Wed 16 of Aug., 2006 21:20 GMT posts: 2388

> On 8/15/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > On 8/15/06, John E Clifford wrote: > > > > --- Jorge Llambías wrote: > > > > > On 8/15/06, John E Clifford wrote: > > > > > > But what about (what I take to be > > > > > > equivalent) {no da me lo broda}? > > > > > > > > > > That's either false, or you have to assume some implicit contextual > > > > > restriction, such as {no da poi nenri le kumfa/poi zasti/poi ... > > > > > cu me lo broda}. > > > > > > > > So the paradox is still there, just not in its simplest form. > > > > > > I don't see the paradox there. > > > > That you can't say {no da me lo broda} truthfully, even though it might > > be true so long as unsaid. > > But you can say {no da me lo broda} truthfully. All you need is some > implicit contextual restriction.

But you can't do it literally. And this is odd, since it is equivalent (I think you have said) to {noda broda}, which can be said truthfully, when it is true. You can (and you) put in implicit restrictions all over the place to make the most strange things happen — or, as in this case, to prevent strange things from happening — but the pure cases remain as problems.

> There's nothing paradoxical in the fact that using a referring term introduces > its referent as a member of the domain of discourse.

Well, in this case there seems to be, since it makes a previously true sentence true false as soon as it is uttered.

> > > > I agree that I have trouble figuring out what your point is here. > > > > > > My point is that {mi djica lo gerku} has none of the problems you attribute > > > to it. > > > > What does it lack? Generalization does not apply nor does interchange of equivalents. And, > of > > course, it turns out that getting what you want is not satisfying. > > Let's take them one by one. > > (1) Existential generalization applies to {mi djica lo broda} as much > as to {mi djica lo nu ponse lo broda}.

Assume that {mi djica lo broda} is true. It does not follow that {da poi broda go'u mi djica da} is true (there may not be any brodas or the one I want may not be one of the ones there are). So it doesn't kn this case. But from {mi djica lo nu mi ponse lo broda} it does follow {da poi nu mi ponse lo broda go'u mo djica da} (neither of the previous objections arises).

> >From {mi djica lo gerku} it follows that {mi djica su'o da}. > >From {mi djica lo nu mi ponse lo gerku} it follows that {mi djica su'o da}. > > All this means is that {lo gerku}/{lo nu mi ponse lo gerku} must have > in each case a referent that is a member of the universe of discourse. > In neither case does this mean that said referent must be something > that satisfies the {zasti} predicate. It is possible to want things > (be it objects or events) that satisfy {na zasti}. There is no difference > in generalization that I can see between objects and events.

{zasti} does not enter into this at all.

> (2) Interchange of equivalents. > Mary told me yesterday that she wanted a dog. There is a dog now in > front of us. {lo gerku} and {lo danlu poi klesi tu} are equivalents in this > context. {lo nu ponse lo gerku} and {lo nu ponse lo danlu poi klesi tu} > are equivalents in this context. Mary could not have used the words > {lo danlu poi klesi tu} to tell me what she wanted because there were > no dogs around when she told me what she wanted. (She might not > have used the words {lo gerku} either.)

Actually they aree not, even if your original claim, that {lo gerku} and {lo danlu poi klesi tu}, which I see no reasion to believe — especially since Mary's claim was yesterday and the dog is here today. There two (at least) different contexts involved here and jumbling together makes a fairly plain situation murky. But more to the point, wanting a dog and wanting a creature in front of here are different wants: they are satisfied by different things, a dog in one case a creature in front (any creature, note) in the other. To be sure, if the creature in fromnt happens to be a dog, then both will be satisfied (if being in front is enough for a dog to satisfy the want). But that is one very particular case, in at least one other case (actually in all other cases) the eexchange does not work, so interchange of equivalence is not valid/

> {la meris djica lo gerku} and {la meris djica lo danlu poi klesi tu} > are equivalent. > {la meris djica lo nu ponse lo gerku} and {la meris djica lo nu ponse lo danlu > poi klesi tu} are equivalent. > > There is no difference between objects and events here.

Well, since the first are not equivalent, the second aren't either, so I agree that they are the same. > (3) "Getting what you want is not satisfying" > > That makes no sense. Similarly with the statement you had made before: > > As they say, "If you want a dog, you are in luck, because there are dogs." > > It is perfectly possible that there is X, that you want X, that you don't get X, > and therefore your want is not satisfied. There is nothing contradictory or > inconsistent in any of that.

This is a new one. First please correct — as I have done — the first claim to read "If you want a dog, you are in luck because dogs exist." The joker here is that you smuggle in an event in the phrase "get a dog" that is "come to have a dog," whereas you are only entitled to the dog, not the having a dog (which you have explicitly denied is involved). So, this argument again makes my point: a dog existing is not enough to satisfy a want, there has to exist an event involving a dog

(and me, in this case) before satisfaction can occur. So, what I really want — what will sayisfy my want — is the event of having a dog.

> > > ... > > > > Well, one of the weird things about events is that they come in trees. If, indeed, there > a > > > bunc > > > > of events any one of which would satisfy my want, there is also an event that occurs > exactly > > > when > > > > any one of these events occurs — the meta-event to just those events. This is probably > why > > > we > > > > want them all in. The same is not necessarily (nor at all, I would say, though I know you > > > > disagree) true of dogs. > > > > > > Why does not the same thing work for dogs? What is it about events that > > > makes them susceptible to this kind of treatment, and which dogs lack? > > > > I'm not sure which you are talking about: there is no guarantee that the dog you want is among > the > > ones in the domain as there is for events, therr is no summary dog as there is for events. > > The dog you want is a member of the domain of discourse as soon as you > mention it, just as for the event you want. That both are in the domain of > discourse as soon as you mention them does not mean in either case that > your want will be satisfied. The satisfaction condition is not related to > membership in the domain of discourse in either case.

Ah, a fundamental disagreement again. The dog you mention in subordinate place is in a subordinate universe, not in the main one. It disappears with the subordinate expression -- unless moved up in some way (there are many such but not are used here). For the rest, I don't get the point: the event is in the primary universe of discourse, the dog — insofar as it was introduced in the secondary place — is not. The point about satisfaction is not that it will occur (in the primary) but that it defines what it is you want — the one that would satisfy did it exist. It is for identification only.

> > dogs > > are comntingent beings, events are not — you can have a universe without dogs but not without > > events (by definition, admittedly, but we are talking about the language in which it is a > given). > > I don't see how dogs are more or less contingent than events. An event of having > a dog does not exist (it does not occur) in a world where dogs don't exist. Both > dogs and events of having dogs are automatically guaranteed to be members of > the universe of discourse for any discourse in which a want of dogs or a want > of having dogs is expressed. That's all that's required. No difference between > dogs and events here.

Conyingrny in the sense that there can be a perfectly functional universe without any dogs in it, but not one without events — they are built into the universe from the get-go. The want for a dog guarantees (let's suppose — I'm not sure this is the way to go) that a dog is in the domain of the want-world, the situation generated by a want. That does not mean the dog is in the domain of the world in which the want occurs. On the other hand, the event of having a dog is in the universse of discourse in which the wish occurs. The one will go away with the wish or be raised -- and when it goes away, generalization fails. I think that you want to have a single universe that incorporates the primary universe and all the subordinate ones into one. This leaves several phenomena unexplained and probably generate paradoxes. It certainly puts some odd restrictions on universes (no predicate evaluated can be evaluated null, apparently, for example). Even if it can get through all the other problems, it still gives the wrong results on generalization and Leibnitz' law.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 17 of Aug., 2006 14:17 GMT On 8/16/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > There's nothing paradoxical in the fact that using a referring term introduces > > its referent as a member of the domain of discourse. > > Well, in this case there seems to be, since it makes a previously true > sentence true false as soon as it is uttered.

Since the utterance changes the domain of discourse, there is no paradox. There is only a paradox if you don't accept that an utterance may have an effect on the domain of discourse.


> > (1) Existential generalization applies to {mi djica lo broda} as much > > as to {mi djica lo nu ponse lo broda}. > > Assume that {mi djica lo broda} is true. It does not follow that {da poi broda > go'u mi djica da} is true (there may not be any brodas or the one I want may > not be one of the ones there are).

(go'u -> zo'u)

Existential generalization from {mi djica lo broda} gives you {su'o da zo'u mi djica da}.

What you give does follow too, although pragmatically it is a bad move.

It may be the case that no brodas exist in the world, but it may not be the case that brodas are not a member of the domain of discourse. The use of the referring term {lo broda} ensures that brodas are a member of the domain of discourse. That is all that is required for existential generalization.

Conflating existence in the world with membership in the domain of discourse leads to all sorts of problems, I agree.

> > (2) Interchange of equivalents. > > Mary told me yesterday that she wanted a dog. There is a dog now in > > front of us. {lo gerku} and {lo danlu poi klesi tu} are equivalents in this > > context. {lo nu ponse lo gerku} and {lo nu ponse lo danlu poi klesi tu} > > are equivalents in this context. Mary could not have used the words > > {lo danlu poi klesi tu} to tell me what she wanted because there were > > no dogs around when she told me what she wanted. (She might not > > have used the words {lo gerku} either.) > > Actually they aree not, even if your original claim, that {lo gerku} and > {lo danlu poi klesi tu}, which I see no reasion to believe — especially > since Mary's claim was yesterday and the dog is here today.

"Dogs" and "animals of that kind" (when pointing to a dog) are not co-referential?


> First please correct — as I have done — the first claim > to read "If you want a dog, you are in luck because dogs exist."

Why should dogs existing be enough to satisfy my want of a dog? Do you think that existence of X is always a sufficient condition for satisfaction of a want of X? If you don't think that (and I know you don't think that) then why say that the fact that dogs exist means I'm in luck?

> The joker > here is that you smuggle in an event in the > phrase "get a dog" that is "come to have a dog," whereas you are only > entitled to the dog, not the having a dog (which you have explicitly denied > is involved).

I have not denied the having of a dog is involved at all. I have explicitly agreed that the satisfaction condition for wanting X involves getting X. The satisfaction condition of an event of wanting is not the same thing as the object of the want. You seem to be conflating the two.

> So, this argument again makes my > point: a dog existing is not enough to satisfy a want, there has to exist an > event involving a dog (and me, in this case) before satisfaction can occur.

We agree so far. But then you jump to:

> So, what I really want — what will sayisfy my want — is the event of having > a dog.

But events are always there, even if they don't occur. So by your own reasoning what you really really want is the event of having a dog to occur.

But wait, the event of an event occurring is also always there. What you really really really want is the event of the event of having a dog occurring to occur.

No, that can't be it either, because there is always that too. What you really really really really want is...

You can see how this line of argument never ends. The mistake is in the very first step. From "X existing is not enough to satisfy a want" it does not follow that "what I really want then must be not X but getting X".

If I want X, X is what I really really want, and satisfaction of that want involves my getting X. No contradiction.

That I also thereby want getting X, and also thereby want the event of getting X ocurring, and also thereby want the event of the event of getting X occurring occurring, and so on, does not change the fact that what I really want is X.

If I see X, then it must follow that I see some event involving X, and it must also follow that I see that some event involving X occurs, and it must also follow that I see that some event of some event involving X ocurring occurs, and ... That does not change the fact that I really see X.

> > I don't see how dogs are more or less contingent than events. > > Conyingrny in the sense that there can be a perfectly functional universe > without any dogs in it, but not one without events — they are built into > the universe from the get-go.

What does "built into the universe" mean? Are we talking about domains or about worlds here? Any event, just like any dog, can be a member of the universe of discourse, all it takes is a mention of them.

No event, just like no dog, necessarily exists in a given world. Worlds can be perfectly functional with and without dogs existing in them, and with and without events of having dogs taking place in them.

No difference between dogs and events.

> The want for a dog guarantees (let's suppose — I'm not sure this is the > way to go) that a dog is in the domain of the want-world, the situation > generated by a want. That does not mean the dog is in the domain > of the world in which the want occurs.

What are "the domain of the want-world" and "the domain of the world in which the want occurs"? Are they related to the domain of discourse?

> On the other hand, the event of having a dog is in the > universse of discourse in which the wish occurs.

If the utterance is {mi djica lo nu mi ponse lo gerku}, then events of having dogs are in the universe of discourse of the utterance, yes.

If the utterance is {mi djica lo gerku}, then dogs are in the universe of discourse of the utterance.

In neither case is it necessary that the referred member of the universe of discourse exists/occurs in the world in which the speaker makes the utterance.

> The one will go away with the wish or be raised > — and when it goes away, generalization fails. I think that you want to have a single universe > that incorporates the primary universe and all the subordinate ones into one.

What do you mean by "universe" here? The domain of discourse that includes all things/events mentioned as its members, or a world where these things/events may or may not exist/occur?

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 17 of Aug., 2006 17:27 GMT posts: 2388

> On 8/16/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > There's nothing paradoxical in the fact that using a referring term introduces > > > its referent as a member of the domain of discourse. > > > > Well, in this case there seems to be, since it makes a previously true > > sentence true false as soon as it is uttered. > > Since the utterance changes the domain of discourse, there is no paradox. > There is only a paradox if you don't accept that an utterance may have an > effect on the domain of discourse. Which is exactly my point. It does not change the fundamental domain of discourse but shifts one into a temporary one whihc endures omnly so long as the conversation stays in the wish-world (or whatever hypothetical realm is involved). To be sure, that domain can be folded into the fundamental domain, but that takes a dekliberate step and is subject to the acceptance of the other conversants (we've been over this before).

> > > > (1) Existential generalization applies to {mi djica lo broda} as much > > > as to {mi djica lo nu ponse lo broda}. > > > > Assume that {mi djica lo broda} is true. It does not follow that {da poi broda > > go'u mi djica da} is true (there may not be any brodas or the one I want may > > not be one of the ones there are). > > (go'u -> zo'u)

Thanks; my dictionary is down somehow and memory always is.

> Existential generalization from {mi djica lo broda} gives you {su'o da zo'u > mi djica da}.

That, too, and of course not poblematic. The more specific one remains.

> What you give does follow too, although pragmatically it is a bad move. > > It may be the case that no brodas exist in the world, but it may not be the > case that brodas are not a member of the domain of discourse. The use > of the referring term {lo broda} ensures that brodas are a member of the > domain of discourse. That is all that is required for existential > generalization.

You keep saying that but give no reason why the mere occurrence of a referring expression should guarantee that it refers (other than a misunderstanding of what "referring expression" means). That the referring expression is also in a (unmarked) secondary place makes the change in the fundamental domain even more difficult to explain.

> Conflating existence in the world with membership in the domain of > discourse leads to all sorts of problems, I agree.


But that is not what we are talking about; this is a matter of permanently changing the fundamental domain of discourse.

> > > (2) Interchange of equivalents. > > > Mary told me yesterday that she wanted a dog. There is a dog now in > > > front of us. {lo gerku} and {lo danlu poi klesi tu} are equivalents in this > > > context. {lo nu ponse lo gerku} and {lo nu ponse lo danlu poi klesi tu} > > > are equivalents in this context. Mary could not have used the words > > > {lo danlu poi klesi tu} to tell me what she wanted because there were > > > no dogs around when she told me what she wanted. (She might not > > > have used the words {lo gerku} either.) > > > > Actually they aree not, even if your original claim, that {lo gerku} and > > {lo danlu poi klesi tu}, which I see no reasion to believe — especially > > since Mary's claim was yesterday and the dog is here today. > > "Dogs" and "animals of that kind" (when pointing to a dog) are not > co-referential?

They are in that case, of course, but that was not the case at hand (the pointing was not at the same time or circumstance as the wanting involved). Notice though that even in that case the exchange does not go through as witness what satisfies the want (the fact that this is token reflexive complicates matters even more, since either way of entering it into the whole creates a peculiar outcome). Notice that a dream about a dog has a dog in it, but not necessarily you; a dream about the animal you are pointing at has you and some animal or other in it.

> > > First please correct — as I have done — the first claim > > to read "If you want a dog, you are in luck because dogs exist." > > Why should dogs existing be enough to satisfy my want of a dog?

It isn't; that's my point.

> Do you think that existence of X is always a sufficient condition for > satisfaction of a want of X?

It will be only in specific cases, i.e., if what I want is an event.

> If you don't think that (and I know you don't > think that) then why say that the fact that dogs exist means I'm in luck?


Because that is all you wanted: a dog and, lo, dogs exist. There was nothing in your want about any rellation that the dog might have to you or anything else, so the bare dog ought to be satisfactory. Notice, the only way you can make a dog satisfactory is to add in another fact, moving from the dog to an event involving a dog. If that is what you mean, then in a logical language you should say so.

> > The joker > > here is that you smuggle in an event in the > > phrase "get a dog" that is "come to have a dog," whereas you are only > > entitled to the dog, not the having a dog (which you have explicitly denied > > is involved). > > I have not denied the having of a dog is involved at all. I have > explicitly agreed > that the satisfaction condition for wanting X involves getting X.

Where getting X means having X or some such. But there is nothing about that in what was said; where does it come from.

>The > satisfaction > condition of an event of wanting is not the same thing as the object of the > want. You seem to be conflating the two.

What more is it and where is it in the whole scheme of things. What you want, you say, is a dog. So, presumably a dog would satisfy your want. What is missing here? My suggestion is that what is missing that "the thing you want" is embedded in an event and so far everything you have said has backed this up, since you always talk about some event, not just the object, in describing what is going on. Of course, it helps that having wanteds be events solves some other problems as well.

> > So, this argument again makes my > > point: a dog existing is not enough to satisfy a want, there has to exist an > > event involving a dog (and me, in this case) before satisfaction can occur. > > We agree so far. But then you jump to: > > > So, what I really want — what will satisfy my want — is the event of having > > a dog. > > But events are always there, even if they don't occur. So by your own > reasoning what you really really want is the event of having a dog to occur.


Sure, satisfaction requires that wehat you want exists. I take it you want this to be another event over and above the event in question. That seems OK with me, though a tad convoluted. If you want to make an infinite regress out of it, it will apply as well to the simple dog case.

> But wait, the event of an event occurring is also always there. What you > really really really want is the event of the event of having a dog occurring > to occur. > > No, that can't be it either, because there is always that too. What you > really really really really want is...


Of course, what you want is an event of which a dog is a constituent. That stops off at the first step. And then satisfaction is explained in terms of the existence of what you want. So I suppose I need to modify my point slightly to say that what you really want is that whose existence releieves the want. Thanks.

> You can see how this line of argument never ends. The mistake is in the > very first step. From "X existing is not enough to satisfy a want" it does > not follow that "what I really want then must be not X but getting X".

But of course the same line works for the simple dog case: having a dog is neot enough, since that always is there (introduced by mentioning it if not otherwise)so it must be that having a dog exists and then the rest follows in order. Best, as laozi says, stop off at the one.

> If I want X, X is what I really really want, and satisfaction of that want > involves my getting X. No contradiction.

Except that now there is this event which was not anywhere mentioned popping up. It is not that every want is satisfied by getting something so this is not an inherent part of the logic of "want;" rather it is something brought in ad hoc top make this case work. I say if you are going to bring it in, do it explicitly.

> That I also thereby want getting X, and also thereby want the event of getting > X ocurring, and also thereby want the event of the event of getting X occurring > occurring, and so on, does not change the fact that what I really want is X. > > If I see X, then it must follow that I see some event involving X, and it > must also follow that I see that some event involving X occurs, and it > must also follow that I see that some event of some event involving > X ocurring occurs, and ... That does not change the fact that I really > see X.

But seeing is a different sort of thing; it is not generally opaque, to cite the most immediately relevant difference.

> > > I don't see how dogs are more or less contingent than events. > > > > Contingency in the sense that there can be a perfectly functional universe > > without any dogs in it, but not one without events — they are built into > > the universe from the get-go. > > What does "built into the universe" mean? Are we talking about domains > or about worlds here? Any event, just like any dog, can be a member of > the universe of discourse, all it takes is a mention of them.

Every universe of discourse contains all events or (if you want) can generate any one wyou want without any extra moves; to add dogs requires not merely mentioning them but also getting concurrence from the conversants — and requires mentioning them in primary space.

> No event, just like no dog, necessarily exists in a given world. Worlds can > be perfectly functional with and without dogs existing in them, and with and > without events of having dogs taking place in them.

Exists yes, bes not.

> No difference between dogs and events.

That difference.

> > The want for a dog guarantees (let's suppose — I'm not sure this is the > > way to go) that a dog is in the domain of the want-world, the situation > > generated by a want. That does not mean the dog is in the domain > > of the world in which the want occurs. > > What are "the domain of the want-world" and "the domain of the world in > which the want occurs"? Are they related to the domain of discourse?

These make the point that the universe of discourse is variable and subdivided. The want-world is a subdivision of a universe of discourse that deals with what someone wants. It is brought in as the want is expressed and dropped when that topic is dropped. The world in which the want occurs is the universe of discourse in effect when the want is expressed; it continues after the want has ceased to be a topic.

> > On the other hand, the event of having a dog is in the > > universse of discourse in which the wish occurs. > > If the utterance is {mi djica lo nu mi ponse lo gerku}, then events of > having dogs are in the universe of discourse of the utterance, ye > > If the utterance is {mi djica lo gerku}, then dogs are in the universe of > discourse of the utterance. > > In neither case is it necessary that the referred member of the universe of > discourse exists/occurs in the world in which the speaker makes the > utterance.

I am not sure what you are saying here. We are talking about universes of discourse, not real worlds (however that might be defined in these cases). The point is that making a wish may change the universe of discourse (on your plan) so that the wished for thing enters a universe of discoure. The point is that it is not (necessaarily) in the universe in effect before the wish is expressed. Events, however, are always in the prior universe and so moves with them that circumvent the wish will go through, but not moves that involve something only entering with the expressed wish.

> > The one will go away with the wish or be raised > > — and when it goes away, generalization fails. I think that you want to have a single > universe > > that incorporates the primary universe and all the subordinate ones into one. > > What do you mean by "universe" here? The domain of discourse that > includes all things/events mentioned as its members, or a world where > these things/events may or may not exist/occur?

The former, properly assembled.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 17 of Aug., 2006 19:32 GMT On 8/17/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > On 8/16/06, John E Clifford wrote: > > > --- Jorge Llambas wrote: > > > > > > > There's nothing paradoxical in the fact that using a referring term introduces > > > > its referent as a member of the domain of discourse. > > > > > > Well, in this case there seems to be, since it makes a previously true > > > sentence true false as soon as it is uttered. > > > > Since the utterance changes the domain of discourse, there is no paradox. > > There is only a paradox if you don't accept that an utterance may have an > > effect on the domain of discourse. > Which is exactly my point. It does not change the fundamental domain > of discourse but shifts one into a

It's hard to maintain a meaningful discussion when you make use of a "fundamental domain of discourse" in your metalanguage and I don't.

> You keep saying that but give no reason why the mere occurrence of > a referring expression should guarantee that it refers (other than a > misunderstanding of what "referring expression" means).

It is guaranteed to refer by construction. You figure what the universe of discourse is by examination of the referring expressions used in the discourse. The universe of discourse is not a given prior to the utterance.

> That the referring expression is also in a (unmarked) secondary place > makes the change in the fundamental domain even more difficult to explain.

I have neither secondary places nor a fundamental domain in my analysis.

> > Conflating existence in the world with membership in the domain of > > discourse leads to all sorts of problems, I agree. > > But that is not what we are talking about; this is a matter of permanently > changing the fundamental domain of discourse.

In what sense "permanently"? The domain of discourse is a metalinguistic construct that is only useful in the analysis of a particular discourse. There is nothing fundamental or permanent about it, as far as I can tell.

> > > > (2) Interchange of equivalents. > > > > Mary told me yesterday that she wanted a dog. There is a dog now in > > > > front of us. {lo gerku} and {lo danlu poi klesi tu} are equivalents in this > > > > context. {lo nu ponse lo gerku} and {lo nu ponse lo danlu poi klesi tu} > > > > are equivalents in this context. Mary could not have used the words > > > > {lo danlu poi klesi tu} to tell me what she wanted because there were > > > > no dogs around when she told me what she wanted. (She might not > > > > have used the words {lo gerku} either.) > > > > > > Actually they aree not, even if your original claim, that {lo gerku} and > > > {lo danlu poi klesi tu}, which I see no reasion to believe — especially > > > since Mary's claim was yesterday and the dog is here today. > > > > "Dogs" and "animals of that kind" (when pointing to a dog) are not > > co-referential? > > They are in that case, of course, but that was not the case at hand > (the pointing was not at the same time or circumstance as the wanting > involved).

The two utterances being compared were {la meris djica lo gerku} and {la meris djica lo danlu poi klesi tu}. The important thing here is that both utterances express the same fact, so substitution of co-referential terms (at least in this case) works. It is irrelevant when the speaker learned about this fact.

> Notice though that even in that case the > exchange does not go through as witness what satisfies the want

You would say the two utterances express different facts then?

> (the fact that this is token reflexive complicates matters even more, > since either way of entering it into the whole creates a > peculiar outcome).

I don't understand what you mean there.

> Notice that a dream about a dog has a dog in it, > but not necessarily you; a dream about the animal you are pointing > at has you and some animal or other in it.

Why would a dream about an animal I'm pointing at have to include me? If I ask "have you ever dreamt about that animal?" (pointing at some animal) am I asking if you have ever dreamt about me?


> > > First please correct — as I have done — the first claim > > > to read "If you want a dog, you are in luck because dogs exist." > > > > Why should dogs existing be enough to satisfy my want of a dog? > > It isn't; that's my point.

Then you agree with me that the quoted claim is nonsensical. Why bring it up as if it showed something then?

> > Do you think that existence of X is always a sufficient condition for > > satisfaction of a want of X? > > It will be only in specific cases, i.e., if what I want is an event.

Right.

> > If you don't think that (and I know you don't > > think that) then why say that the fact that dogs exist means I'm in luck? > > Because that is all you wanted: a dog and, lo, dogs exist.

But how is the fact that dogs exist relevant here? What has that got to do with satisfaction of my want? I said I wanted a dog, I didn't say I wanted a dog to exist. Why would you say that the fact that dogs exist means I'm in luck?

> There was nothing in your want about > any rellation that the dog might have to you or anything else, so the bare > dog ought to be satisfactory.

The satisfaction of "X wants object Y" requires Y to be available to X, it is not enough for Y to exist for the want to be satisfied.

> Notice, the only way you can make a dog satisfactory is to add in > another fact, moving from the dog to an event involving a dog. If that > is what you mean, then in a logical language you should say so.

You are not offering any kind of logical argument here.

> > I have not denied the having of a dog is involved at all. I have > > explicitly agreed > > that the satisfaction condition for wanting X involves getting X. > > Where getting X means having X or some such.

Yes.

> But there is nothing about that in what was said; > where does it come from.

It's part of what wanting something means.

> >The > > satisfaction > > condition of an event of wanting is not the same thing as the object of the > > want. You seem to be conflating the two. > > What more is it and where is it in the whole scheme of things. > What you want, you say, is a dog.

Exactly.

> So, presumably a dog would satisfy your want.

Having a dog would satisfy my wanting a dog, yes. A dog existing would not be enough to satisfy my wanting a dog.

> What is missing here?

Nothing, as far as I can tell.

> Of course, it helps that having wanteds be events solves some other > problems as well.

None so far.


> > If I want X, X is what I really really want, and satisfaction of that want > > involves my getting X. No contradiction. > > Except that now there is this event which was not anywhere mentioned > popping up.

Popping up where? In the explanation of what the condition that will satisfy a want is? What's wrong with an event popping up in such an explanation? Events pop up in all sorts of explanations of utterances that don't make any reference to them. There is no logical requirement that says that events may not pop up in explanations of utterances that make no reference to them.

> It is not that > every want is satisfied by getting something so this is not an inherent > part of the logic of > "want;" rather it is something brought in ad hoc top make this case work. > I say if you are going to bring it in, do it explicitly.

That the want of an object is satisfied by getting the object is of course part of the inherent logic of "want". Why do you say it is not? Similarly, that the want of an event is satisfied by the event occurring is part of the inherent logic of "want".

> > > > I don't see how dogs are more or less contingent than events. > > > > > > Contingency in the sense that there can be a perfectly functional universe > > > without any dogs in it, but not one without events — they are built into > > > the universe from the get-go. > > > > What does "built into the universe" mean? Are we talking about domains > > or about worlds here? Any event, just like any dog, can be a member of > > the universe of discourse, all it takes is a mention of them. > > Every universe of discourse contains all events or (if you want) can generate any one wyou want > without any extra moves; to add dogs requires not merely mentioning them but also getting > concurrence from the conversants — and requires mentioning them in primary space.

It takes the same effort to get the conversants to concur with the incorporation of dogs into the domain of discourse as it takes them to concur with the incorporation of dogs barking, as far as I can tell.

"Do you like dogs?" "Do you like the barking of dogs?" "Have you ever seen dogs?" "Have you ever heard dogs barking?" "Mary yesterday dreamt of dogs." "Mary yesterday dreamt of dogs barking."

No difference in introducing the referents of {lo gerku} or of {lo nu lo gerku cu bacru} into the domain of discourse.


> The point is that making a wish may change > the universe of discourse (on your plan) so that the wished for thing > enters a universe of discoure.

It's not the making of a wish that changes the domain. Referring to something wished for, either by the speaker or by anyone else, may change the domain of discourse. Making a wish does not in itself touch the universe of discourse.

> The point is that it is not (necessaarily) in the universe in effect before > the wish is expressed. Events, however, are always in the prior universe > and so moves with them that circumvent the wish will go through, but not > moves that involve something only entering with the expressed wish.

Are you saying that for example the event of twenty-seven blue unicorns flying around the moon is a member of every domain of discourse even if nobody has ever mentioned it, but that unicorns are not a member of any domain of discourse until someone brings them up? That's not a useful metalinguistic construct, as far as I can tell.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 18 of Aug., 2006 01:46 GMT posts: 2388

> On 8/17/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > On 8/16/06, John E Clifford wrote: > > > > --- Jorge Llambías wrote: > > > > > > > > > There's nothing paradoxical in the fact that using a referring term introduces > > > > > its referent as a member of the domain of discourse. > > > > > > > > Well, in this case there seems to be, since it makes a previously true > > > > sentence false as soon as it is uttered. > > > > > > Since the utterance changes the domain of discourse, there is no paradox. > > > There is only a paradox if you don't accept that an utterance may have an > > > effect on the domain of discourse. > > Which is exactly my point. It does not change the fundamental domain > > of discourse but shifts one into a > > It's hard to maintain a meaningful discussion when you make use of > a "fundamental domain of discourse" in your metalanguage and I don't.

It is hard to maintain a discussion if you do not adhere to fundamental principlr in the area we are talking about. When a conversation begins it does so with a given universe of discourse. As the conversation goes along.. Oh, Hwll, we;ve been through this half a dozen time before. What about it don't you understand or — if you do understand it — what reason do you have for not accepting it and going off on some harebrained scheme of your own? One that leaves major areas unexplained and gives obviously wrong results?

> > You keep saying that but give no reason why the mere occurrence of > > a referring expression should guarantee that it refers (other than a > > misunderstanding of what "referring expression" means). > > It is guaranteed to refer by construction. You figure what the universe of > discourse is by examination of the referring expressions used in the > discourse. The universe of discourse is not a given prior to the utterance.

But this is at best a way to find out what are the things referred to by Logical Proper Names (whatever they are) and says nothing about things referred to by descriptions, whose reference depends upon what there is, not the other way around. To be sure, we can enrich the universe by admitting (in cooperation with the other conversants) new things, but that is more complicated than just mentioning them. You can admit things by just mentioning them in secondary environements, but then you admit them only temporarily and in that environement. They carry no weight outside.

> > That the referring expression is also in a (unmarked) secondary place > > makes the change in the fundamental domain even more difficult to explain. > > I have neither secondary places nor a fundamental domain in my analysis.

Which is a fairly good reason why your analysis does such a poor job of explaining linguistic phenomena.

> > > Conflating existence in the world with membership in the domain of > > > discourse leads to all sorts of problems, I agree. > > > > But that is not what we are talking about; this is a matter of permanently > > changing the fundamental domain of discourse. > > In what sense "permanently"? The domain of discourse is a metalinguistic > construct that is only useful in the analysis of a particular discourse. There > is nothing fundamental or permanent about it, as far as I can tell.

Permanent for the course of a conversation as opposed to temporary for the course of a secondary context. And, of course, not really even permanent then since it is possible to get rid of things from a conversation as well as to add them. But that too takes cooperation.

> > > > > (2) Interchange of equivalents. > > > > > Mary told me yesterday that she wanted a dog. There is a dog now in > > > > > front of us. {lo gerku} and {lo danlu poi klesi tu} are equivalents in this > > > > > context. {lo nu ponse lo gerku} and {lo nu ponse lo danlu poi klesi tu} > > > > > are equivalents in this context. Mary could not have used the words > > > > > {lo danlu poi klesi tu} to tell me what she wanted because there were > > > > > no dogs around when she told me what she wanted. (She might not > > > > > have used the words {lo gerku} either.) > > > > > > > > Actually they are not, even if your original claim, that {lo gerku} and > > > > {lo danlu poi klesi tu}, which I see no reasion to believe — especially > > > > since Mary's claim was yesterday and the dog is here today. > > > > > > "Dogs" and "animals of that kind" (when pointing to a dog) are not > > > co-referential? > > > > They are in that case, of course, but that was not the case at hand > > (the pointing was not at the same time or circumstance as the wanting > > involved). > > The two utterances being compared were {la meris djica lo gerku} and > {la meris djica lo danlu poi klesi tu}. The important thing here is that > both utterances express the same fact, so substitution of co-referential > terms (at least in this case) works. It is irrelevant when the speaker learned > about this fact.

But they don't express the same fact, since they are about different wants: one is about wanting a dog and the other is about wanting an animal (any animal) of the sort pointed at. It may happen, because a dog is pointed at, that both wants are simultaneously satisfied. But one could be satisfied while the other was not: the animal there might be a cat and the dog not pointed at.

> > Notice though that even in that case the > > exchange does not go through as witness what satisfies the want > > You would say the two utterances express different facts then?

Yes, because they attribute different wants so Mary.

> > (the fact that this is token reflexive complicates matters even more, > > since either way of entering it into the whole creates a > > peculiar outcome). > > I don't understand what you mean there.

Defining something on the basis of what happens on a particular occasion, especially when that occasion is determined by the utterance of the token that contains the defining builds up tangles. So, we have (for clarity at least) to extract a bit: the thing I pointed at at dd/mm/yyyy/hh/mm/ss was a dog and then we contrast "Mary wants a dog" with "Mary wants the animal of the kind I pointed tto at dd/mm/yyyy/hh/mm/ss" but Mary probably didn't even know (or believe) that you pointed at some animal (or were going to, most likely) at that time, nor what sort of animal it would be, so she almost certainly did not want it. Bake to the old story: in the appropriate universe, Lois Lane wants Superman (probably "wants to be intimate with") but she certainly does not want (to be intimate with) Clark Kent, even though ... (in wish worlds, the referents are principle world senses).

> > Notice that a dream about a dog has a dog in it, > > but not necessarily you; a dream about the animal you are pointing > > at has you and some animal or other in it. > > Why would a dream about an animal I'm pointing at have to include > me? If I ask "have you ever dreamt about that animal?" (pointing at > some animal) am I asking if you have ever dreamt about me?

But notice that you have changed the context — and indeed the description of the dream — from dreaming about that class of animals right there right now (with all the problems of token reflexivity) to one about a kind of animal defined by an act separate from that dream. > > > > > First please correct — as I have done — the first claim > > > > to read "If you want a dog, you are in luck because dogs exist." > > > > > > Why should dogs existing be enough to satisfy my want of a dog? > > > > It isn't; that's my point. > > Then you agree with me that the quoted claim is nonsensical. Why > bring it up as if it showed something then?

I forget what the quoted claim was. "If you want a dog, then you are in luck, because dogs exist"?It is not nonsense; it, combined with the factr that the existence of dogs does not assuage your want shows that you want something other than/more than dogs. A look at the actual satisfaction shows, as you have pointed out several times, that what cures your want is to have (get) a dog. When you object to this, do you mean that dogs do assuage your desire or that having a dog does not?

> > > Do you think that existence of X is always a sufficient condition for > > > satisfaction of a want of X? > > > > It will be only in specific cases, i.e., if what I want is an event. > > Right. > > > > If you don't think that (and I know you don't > > > think that) then why say that the fact that dogs exist means I'm in luck? > > > > Because that is all you wanted: a dog and, lo, dogs exist. > > But how is the fact that dogs exist relevant here? What has that got > to do with satisfaction of my want? I said I wanted a dog, I didn't > say I wanted a dog to exist. Why would you say that the fact that dogs > exist means I'm in luck?

Well, what exactly does wanting a dog mean. Nothing is said about any event involving a dog and satisfaction is going to involve an event. The only event that goes unstated (or even implied) is existence, so ...

> > There was nothing in your want about > > any relation that the dog might have to you or anything else, so the bare > > dog ought to be satisfactory. > > The satisfaction of "X wants object Y" requires Y to be available to X, > it is not enough for Y to exist for the want to be satisfied.

Precisely. And if that is so, why not say so?

> > Notice, the only way you can make a dog satisfactory is to add in > > another fact, moving from the dog to an event involving a dog. If that > > is what you mean, then in a logical language you should say so. > > You are not offering any kind of logical argument here.

You mean, nothing compels you to say what you mean when talking about your wants. I agree and haven't claimed that something does. The compulsion (weak in your case) comes from the desire to have a logical language in even a fairly narrow sense of that term (solving semantic problems in the way that Logic does, for example) and that seems to involve saying what you mean. The fact that doing it that way also solves some other problems is a plus, of course.

> > > I have not denied the having of a dog is involved at all. I have > > > explicitly agreed > > > that the satisfaction condition for wanting X involves getting X. > > > > Where getting X means having X or some such. > > Yes. > > > But there is nothing about that in what was said; > > where does it come from. > > It's part of what wanting something means.

? You mean that having what you want is always part of what what wanting something means, that "I have.." is built into wanting. But that doesn't seem to be true in general or is so only in an very attenuated way: "I want my great grandchildren to be happy." or even "I want great grandchildren" seems to have nothing to do with my having anything (I being seventy and and my daughter still without children). I think that what is true is that in a lot of cases where we say, in English, etc., that we want X, we mean that we want to have X. In other cases we mean other things, equally clear from context. That is to say, we often — in English — abbreviate what we mean to say. The point is that Lojban shouldn't do that — to the same extent at least ({tu'a} is still pretty abbreviated). As i say, the fact that it forms part of a uniform way to deal with a logical problem is icing on the cake (and, hopefully — but not really — an added incentive to do it that way).

> > >The > > > satisfaction > > > condition of an event of wanting is not the same thing as the object of the > > > want. You seem to be conflating the two. > > > > What more is it and where is it in the whole scheme of things. > > What you want, you say, is a dog. > > Exactly. > > > So, presumably a dog would satisfy your want. > > Having a dog would satisfy my wanting a dog, yes. > A dog existing would not be enough to satisfy my wanting a dog.

Agreed. So it is an event that satisfies your want. So you are saying you really want something that does not satisfy that want. This seems odd to me.

> > What is missing here? > > Nothing, as far as I can tell. > > > Of course, it helps that having wanteds be events solves some other > > problems as well. > > None so far. >

I am sorry to keep harping on the problem of opaque contexts, but they do not go away because you come up with com cockamamie scheme in which they appear to vanish (being driven underground and ignored).

> > > If I want X, X is what I really really want, and satisfaction of that want > > > involves my getting X. No contradiction. > > > > Except that now there is this event which was not anywhere mentioned > > popping up. > > Popping up where? In the explanation of what the condition that will > satisfy a want is? What's wrong with an event popping up in such an > explanation? Events pop up in all sorts of explanations of utterances > that don't make any reference to them. There is no logical requirement > that says that events may not pop up in explanations of utterances that > make no reference to them.

For instance?

> > It is not that > > every want is satisfied by getting something so this is not an inherent > > part of the logic of > > "want;" rather it is something brought in ad hoc top make this case work. > > I say if you are going to bring it in, do it explicitly. > > That the want of an object is satisfied by getting the object is of course > part of the inherent logic of "want". Why do you say it is not? > Similarly, that the want of an event is satisfied by the event occurring is > part of the inherent logic of "want".

Why are these two kinds of wants so different? The stateds involved seem to be the same both phenomenologically and logically. And as noted, I can want things that I could never get and which I yet thing can be satisfied.

> > > > > I don't see how dogs are more or less contingent than events. > > > > > > > > Contingency in the sense that there can be a perfectly functional universe > > > > without any dogs in it, but not one without events — they are built into > > > > the universe from the get-go. > > > > > > What does "built into the universe" mean? Are we talking about domains > > > or about worlds here? Any event, just like any dog, can be a member of > > > the universe of discourse, all it takes is a mention of them. > > > > Every universe of discourse contains all events or (if you want) can generate any one wyou > want > > without any extra moves; to add dogs requires not merely mentioning them but also getting > > concurrence from the conversants — and requires mentioning them in primary space. > > It takes the same effort to get the conversants to concur with the > incorporation of dogs into the domain of discourse as it takes them to > concur with the incorporation of dogs barking, as far as I can tell.

Nice use of a vagueness. It takes no effort to get them to admit that there is an event of dogs barking, it takes quite a bit to get them to admit that that event is or has at some time under discussion occurred. That is, getting it into the universe of discourse is free, getting it into the real world takes ome doing. You seem to have confused the two.

> "Do you like dogs?" "Do you like the barking of dogs?" > "Have you ever seen dogs?" "Have you ever heard dogs barking?" > "Mary yesterday dreamt of dogs." "Mary yesterday dreamt of dogs barking."

The first two pairs introduce (or at least offer for introduction) dogs and the barking of dogs. For the first in each case, the gambit can be refused, in each second case it cannot. Although the person may refuse to answer the question, he can't say "There are no events of dogs barking" and mean to make a remark about the universe ogf discourse rather than about the "real world."

> No difference in introducing the referents of {lo gerku} or of {lo nu lo gerku > cu bacru} into the domain of discourse.

Yes, because you refuse to cooperate with your fellow conversants. That is simply rude. > > > The point is that making a wish may change > > the universe of discourse (on your plan) so that the wished for thing > > enters a universe of discoure. > > It's not the making of a wish that changes the domain. Referring to > something wished for, either by the speaker or by anyone else, may change > the domain of discourse. Making a wish does not in itself touch the universe > of discourse.

True. The point is that stating the wish doesn't either — at lest not in the way you seem to claim.

> > The point is that it is not (necessarily) in the universe in effect before > > the wish is expressed. Events, however, are always in the prior universe > > and so moves with them that circumvent the wish will go through, but not > > moves that involve something only entering with the expressed wish. > > Are you saying that for example the event of twenty-seven blue unicorns > flying around the moon is a member of every domain of discourse even > if nobody has ever mentioned it, but that unicorns are not a member of any > domain of discourse until someone brings them up? That's not a useful > metalinguistic construct, as far as I can tell.

Actually, the last part seems to be a part of what you are saying. Yes, the event with all the unicorns is there for the asking. But unicorns can be part of the initial universe of discourse even if nobody ever mentions them. It is more likely that two conversants have slightly different universes in mind and much of a conversation may involve fusing the two into one. If the unicorns are never mentioned primarily, that aspect of the common universe will not get sorted out and the two may end with different understandings of what the universe was — even though they agree on all that was explicitly said.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 18 of Aug., 2006 16:52 GMT On 8/17/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > On 8/17/06, John E Clifford wrote: > > > You keep saying that but give no reason why the mere occurrence of > > > a referring expression should guarantee that it refers (other than a > > > misunderstanding of what "referring expression" means). > > > > It is guaranteed to refer by construction. You figure what the universe of > > discourse is by examination of the referring expressions used in the > > discourse. The universe of discourse is not a given prior to the utterance. > > But this is at best a way to find out what are the things referred to by > Logical Proper Names (whatever they are) and says nothing about things > referred to by descriptions, whose reference depends upon what there is, > not the other way around.

Here you seem to be saying that "what there is" is a given, independent of what an utterance is about. That seems wrong to me, because it seems clear that different utterances call for different "what there is"s.

> To be sure, we can enrich the universe by > admitting (in cooperation with the other conversants) new things, but that > is more complicated than just mentioning them.

OK, it might be more complicated than just mentioning them, it does take some negotiation with the other conversants.

But you claim that it is easier to incorporate events than objects into the domain of discourse (or sometimes you say that they don't need to be incorporated because they are there from the get go). Yet there does not seem to be anything special about events that would make them any easier to admit into the domain of discourse. There are events that exist (occur) in the world just as there are objects that exist in the world, there are events that don't exist (occur) in the world just as there are objects that don't exist in the world. There does not seem to be any reason for events to be easier than objects to incorporate into the domain of discourse

> You can admit things by > just mentioning them in secondary environements, but then you admit them > only temporarily and in that environement. They carry no weight outside.

Here you introduce "environments". Are these subsets of the domain of discourse? Or are they events, such as an event that can be dreamed about? If the latter, then what you may be saying is that some member of the universe of discourse is restricted to exist in some world which is not the world where the speaker exists. That seems perfectly reasonable. But if by "environment" you mean a subset of the universe of discourse, then I don't understand what it is for. That some referent exists in a certain restricted world and does not exist in some other world is quite acceptable. That applies both to objects and to events. But the referent, be it object or event, before we can decide whether it exists or not in this world or that, must be a member of the domain of discourse.


> Bake to the old story: in the appropriate universe, Lois Lane wants > Superman (probably "wants to be intimate with") but she certainly > does not want (to be intimate with) Clark Kent, even though ... (in > wish worlds, the referents are principle world senses).

Are you saying that {la superman} and {la klark kent} have the same referent for us, whereas {lo nu la lois lein cu gletu la superman} and {lo nu la lois lein cu gletu la klark kent} have different referents for us?

If so, then {la lois lein cu djica la superman} says the same thing as {la lois lein cu djica la klark kent} for us, whereas {la lois lein cu djica lo nu la lois lein cu gletu la superman} and {la lois lein cu djica lo nu la lois lein cu gletu la klark kent} say different things for us. So where is the problem?

Lois would not say {mi djica la klark kent} instead of {mi djica la superman}, because for her {la klark kent} and {la superman} are not co-referential.

But none of this is different for objects than for events. If Lois does not know, for example, that {la klark kent cu djica lo nu la lois cu gletu la superman}, then she will not know that {lo nu la lois cu gletu la superman} and {lo se djica be la klark kent} have the same event as referent. So she will not say {mi djica lo se djica be la klark kent}. But we will know and can say that {la lois djica lo se djica be la klark kent}.

Descriptions of events can be co-referential without someone knowing that they are co-referential as much as object descriptions. Objects and events don't present any difference here.

> > > > > First please correct — as I have done — the first claim > > > > > to read "If you want a dog, you are in luck because dogs exist." > > > > > > > > Why should dogs existing be enough to satisfy my want of a dog? > > > > > > It isn't; that's my point. > > > > Then you agree with me that the quoted claim is nonsensical. Why > > bring it up as if it showed something then? > > I forget what the quoted claim was. "If you want a dog, then you are in luck, because dogs > exist"?

Yes.

> It is not nonsense; it, combined with the factr that the existence of dogs > does not assuage your want shows that you want something other > than/more than dogs.

But why should we take the claim "If you want a dog, then you are in luck, because dogs exist" seriously, be it alone or in conjunction with something else? The claim sounds like nonsense.

> A look at the actual > satisfaction shows, as you have pointed out several times, that what cures > your want is to have (get) a dog.

Exactly, so the mere existence of dog is not reason enough to say that someone who wants a dog is in luck.

> When you object to this, do you mean that dogs do assuage your desire > or that having a dog does not?

I mean that the mere existence of dogs does not assuage my desire.

> Well, what exactly does wanting a dog mean. Nothing is said about any > event involving a dog and satisfaction is going to involve an event. The only > event that goes unstated (or even implied) is existence, so ...

Aha! "The only event that goes unstated (or even implied) is existence".

Where does that come from?


> I think that what is true is that in a lot of cases where we > say, in English, etc., that we want X, we mean that we want to have X. > In other cases we mean other things, equally clear from context. That > is to say, we often — in English — abbreviate what we mean to say. The > point is that Lojban shouldn't do that — to the same extent at least > ({tu'a} is still pretty abbreviated).

That's not a logical consequence of your assumption.

Let's grant, for the sake of argument, that in English "X wants object Y" is always an abbreviation for "X wants some event involving Y". Let's now define in Lojban the predicate {djicrxeventa} that means "X wants event Y". There is nothing in Logic that prevents having another predicate in a logical language that means "X wants object Y" (let's call it {djicrxobjekta} and which is used to describe situations where the referent of X and the referent of Y are in a relationship such that if the condition "X gets Y" were to come about, X would be satisfied. And there is nothing in Logic that prevents having a predicate {djica} such that "X djica Y" is true whenever Y is an event and "X djicrxeventa Y" is true or whenever Y is an object and "X djicrxobjekta Y" is true.


> > > > If I want X, X is what I really really want, and satisfaction of that want > > > > involves my getting X. No contradiction. > > > > > > Except that now there is this event which was not anywhere mentioned > > > popping up. > > > > Popping up where? In the explanation of what the condition that will > > satisfy a want is? What's wrong with an event popping up in such an > > explanation? Events pop up in all sorts of explanations of utterances > > that don't make any reference to them. There is no logical requirement > > that says that events may not pop up in explanations of utterances that > > make no reference to them. > > For instance?

Well, let's say "I eat an apple". If I have to explain what that means, I might mention events like taking a bite, chewing and swallowing, that are not mentioned in "I eat an apple".


> And as noted, I can want things that I could never get and > which I yet thing can be satisfied.

I can want things that I could never get, just as I could want events that could never happen, yes. No difference between objects and events here.


> > It takes the same effort to get the conversants to concur with the > > incorporation of dogs into the domain of discourse as it takes them to > > concur with the incorporation of dogs barking, as far as I can tell. > > Nice use of a vagueness. It takes no effort to get them to admit that > there is an event of dogs barking, it takes quite a bit to get them to admit > that that event is or has at some time under discussion occurred. That > is, getting it into the universe of discourse is free, getting it into > the real world takes ome doing. You seem to have confused the two.

No, you seem to confuse the two. Getting them into the universe of discourse is as free for dogs as for events. It takes work to get them to admit that they exist/occur in the real world. No difference between objects and events.

> > "Do you like dogs?" "Do you like the barking of dogs?" > > "Have you ever seen dogs?" "Have you ever heard dogs barking?" > > "Mary yesterday dreamt of dogs." "Mary yesterday dreamt of dogs barking." > > The first two pairs introduce (or at least offer for introduction) dogs and > the barking of dogs.

Right, no difference so far.

> For the first in each case, the gambit can be refused, in each second > case it cannot. Although the > person may refuse to answer the question, he can't say "There are no > events of dogs barking" and > mean to make a remark about the universe ogf discourse rather than > about the "real world."

He can't say "There are no dogs" and mean to make a remark about the universe of discourse rather than about the "real world" either. No difference between objects and events.


> > > Events, however, are always in the prior universe > > > and so moves with them that circumvent the wish will go through, but not > > > moves that involve something only entering with the expressed wish. > > > > Are you saying that for example the event of twenty-seven blue unicorns > > flying around the moon is a member of every domain of discourse even > > if nobody has ever mentioned it, but that unicorns are not a member of any > > domain of discourse until someone brings them up? That's not a useful > > metalinguistic construct, as far as I can tell. > > Actually, the last part seems to be a part of what you are saying. Yes, the > event with all the unicorns is there for the asking. But unicorns can be part > of the initial universe of discourse even if nobody ever mentions them.

The event of unicorns flying can also be part of the initial universe of discourse even if nobody mentions them, with suitable context, and the unicorns are always there for the asking too, at least as much as the event of unicorns flying around the moon is. So no difference so far between objects and events.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by JohnCowan on Fri 18 of Aug., 2006 17:42 GMT posts: 149 Jorge Llamb?as scripsit:

> Are you saying that {la superman} and {la klark kent} have the same > referent for us, whereas {lo nu la lois lein cu gletu la superman} and > {lo nu la lois lein cu gletu la klark kent} have different referents > for us?

Indeed, they have the same and different referents, respectively, for everyone. Superman *is* Clark Kent and when identity holds, it holds in all possible worlds (modulo issues about existence). But nu-contexts are opaque and identical terms may not be substituted in them.

> Lois would not say {mi djica la klark kent} instead of {mi djica la > superman}, because for her {la klark kent} and {la superman} are not > co-referential.

Either that, or else (as I believe) they are co-referential even for her, and she simply has appropriate false beliefs. In other words, she will not point to Superman and say "That's Clark Kent", because she believes that statement to be false even though we know it's true.

> But none of this is different for objects than for events.

+1

-- MEET US AT POINT ORANGE AT MIDNIGHT BRING YOUR DUCK OR PREPARE TO FACE WUGGUMS John Cowan cowan@ccil.org http://www.ccil.org/~cowan


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 18 of Aug., 2006 20:22 GMT posts: 2388 Since a lot of this is irreelevant to the topic under discussion, I will try to be briefere.


> On 8/17/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > On 8/17/06, John E Clifford wrote: > > > > You keep saying that but give no reason why the mere occurrence of > > > > a referring expression should guarantee that it refers (other than a > > > > misunderstanding of what "referring expression" means). > > > > > > It is guaranteed to refer by construction. You figure what the universe of > > > discourse is by examination of the referring expressions used in the > > > discourse. The universe of discourse is not a given prior to the utterance. > > > > But this is at best a way to find out what are the things referred to by > > Logical Proper Names (whatever they are) and says nothing about things > > referred to by descriptions, whose reference depends upon what there is, > > not the other way around. > > Here you seem to be saying that "what there is" is a given, independent > of what an utterance is about. That seems wrong to me, because it seems > clear that different utterances call for different "what there is"s.

In a given universe of discourse, what there is is a given. Desc4ription in that universe refer first and foremost to things in that universe. If they fail, then they are attempts to introduce new thig into that universe — and that take some cooperation.

> > To be sure, we can enrich the universe by > > admitting (in cooperation with the other conversants) new things, but that > > is more complicated than just mentioning them. > > OK, it might be more complicated than just mentioning them, it does take > some negotiation with the other conversants. > > But you claim that it is easier to incorporate events than objects into > the domain of discourse (or sometimes you say that they don't need > to be incorporated because they are there from the get go). Yet there > does not seem to be anything special about events that would make > them any easier to admit into the domain of discourse. There are events > that exist (occur) in the world just as there are objects that exist in > the world, there are events that don't exist (occur) in the world just as > there are objects that don't exist in the world. There does not seem to > be any reason for events to be easier than objects to incorporate into > the domain of discourse

It is a principle of Lojban — which you may not like, of course — that events (all events) are always available as assignables for variables, one of the key featurees of things in the universe of discourse. This availability applies to all abstractions. Dogs are not under the same principle.

> > You can admit things by > > just mentioning them in secondary environements, but then you admit them > > only temporarily and in that environement. They carry no weight outside. > > Here you introduce "environments". Are these subsets of the domain of > discourse? Or are they events, such as an event that can be dreamed about? > If the latter, then what you may be saying is that some member of the universe > of discourse is restricted to exist in some world which is not the world where > the speaker exists.

This sounds about right if by "world" you mean "universe of discourse" and if by "exist" you mean "be in the u/d".

>That seems perfectly reasonable. But if by "environment" > you mean a subset of the universe of discourse, then I don't understand what > it is for. That some referent exists in a certain restricted world and does not > exist in some other world is quite acceptable. That applies both to objects > and to events. But the referent, be it object or event, before we can decide > whether it exists or not in this world or that, must be a member of the domain > of discourse.

Well, you just blew away the distinction. I'll just repeat that things required in a subjunctive universe of discourse need not also be in the no=subjunctive one to which it is subordinate. The do their work in the contrary-to-fact world but do not (by the fact of therir being in that world) affect what is in the ordinary universe to which the subjunctive is subordinate,

Secondary environments are lingusitic items like subjunctive clauses, "that" clauses and so on (including indeed negative clauses). They are not about domains and the like as such, but they are places where mentioning something may make that thing temporarily a part of the universe of discourse.

> > Back to the old story: in the appropriate universe, Lois Lane wants > > Superman (probably "wants to be intimate with") but she certainly > > does not want (to be intimate with) Clark Kent, even though ... (in > > wish worlds, the referents are principle world senses). > > Are you saying that {la superman} and {la klark kent} have the same > referent for us, whereas {lo nu la lois lein cu gletu la superman} and > {lo nu la lois lein cu gletu la klark kent} have different referents for us?

For us, the two events have the same referent. What are not equivalent are {la lois lein djica lo nu gletu la superman} {la lois lein djica lo nu gletu la klark kent}. Lois Lane knows that the two wants are different (they are her wants after all) but, since she doesn't know that the identity holds, she is not puzzled by this. Did she learn the identity, she might well be for a moment. We, knowing both, are not puzzled since we know that wants refer into counter-factual worlds where the identity between the referent of the two names — which holds in the non-counterfactual world — need not hold.


> If so, then {la lois lein cu djica la superman} says the same thing as > {la lois lein cu djica la klark kent} for us,

No, it has the same problem as the event formulation — it refers into a conterfactual realm, were factual eqyuivalences need not hold.

> whereas {la lois lein cu djica > lo nu la lois lein cu gletu la superman} and {la lois lein cu djica lo nu > la lois lein cu gletu la klark kent} say different things for us. So where > is the problem?

> Lois would not say {mi djica la klark kent} instead of {mi djica la superman}, > because for her {la klark kent} and {la superman} are not co-referential.

Very true, but her knowledge comes in only insofar as this difference might briefly puzzle her if she knew it.

> But none of this is different for objects than for events. If Lois > does not know, > for example, that {la klark kent cu djica lo nu la lois cu gletu la superman}, > then she will not know that {lo nu la lois cu gletu la superman} and > {lo se djica be la klark kent} have the same event as referent.

But they don't.

> So she will not > say {mi djica lo se djica be la klark kent}. But we will know and can say that > {la lois djica lo se djica be la klark kent}.


No we can't — truthfully — since she doesn't want that at all.

> Descriptions of events can be co-referential without someone knowing that > they are co-referential as much as object descriptions. Objects and events > don't present any difference here.

I agree; for neither of them does Leibniz' Law hold. However, you do sem to think that some things which are not identical are (implicitly because of Leibniz' Law).

> > > > > > First please correct — as I have done — the first claim > > > > > > to read "If you want a dog, you are in luck because dogs exist." > > > > > > > > > > Why should dogs existing be enough to satisfy my want of a dog? > > > > > > > > It isn't; that's my point. > > > > > > Then you agree with me that the quoted claim is nonsensical. Why > > > bring it up as if it showed something then? > > > > I forget what the quoted claim was. "If you want a dog, then you are in luck, because dogs > > exist"? > > Yes. > > > It is not nonsense; it, combined with the fact that the existence of dogs > > does not assuage your want shows that you want something other > > than/more than dogs. > > But why should we take the claim "If you want a dog, then you are in > luck, because dogs exist" seriously, be it alone or in conjunction with > something else? The claim sounds like nonsense.

Well, what else doe "I want a dog" mean? There is nothing mentioned but the dog, so presumably that is all that is required. It is, to say it for the umpteenth time, a reductio to make the point that something more than even the existence of a dog is mean when you want a dog.

> > A look at the actual > > satisfaction shows, as you have pointed out several times, that what cures > > your want is to have (get) a dog. > > Exactly, so the mere existence of dog is not reason enough to say that > someone who wants a dog is in luck.

The point!

> > When you object to this, do you mean that dogs do assuage your desire > > or that having a dog does not? > > I mean that the mere existence of dogs does not assuage my desire.

Good, we agree so far.

> > Well, what exactly does wanting a dog mean. Nothing is said about any > > event involving a dog and satisfaction is going to involve an event. The only > > event that goes unstated (or even implied) is existence, so ... > > Aha! "The only event that goes unstated (or even implied) is existence". > > Where does that come from?

Elsewhere it is claimed that, if you use a bear thing word where an event is required you are taking the thing as an event (as you can do), the event of that thing existing — maybe even the event of the entire existence of that thing. I know of no other suggestion that has been offered for this sort of expression.

> > > I think that what is true is that in a lot of cases where we > > say, in English, etc., that we want X, we mean that we want to have X. > > In other cases we mean other things, equally clear from context. That > > is to say, we often — in English — abbreviate what we mean to say. The > > point is that Lojban shouldn't do that — to the same extent at least > > ({tu'a} is still pretty abbreviated). > > That's not a logical consequence of your assumption.

Which assumption? But anyhow this story (I take it that "that" refers to the whole paragraph) is not offered as a consequence of anything but as a tentative explanation for a certain linguistic phenomenon: that,when what is said to be wanted is an event, satisfaction is that event occurring, and when what is said to be wanted is a thing, satisfact is the event of having that thing occurring.


> Let's grant, for the sake of argument, that in English "X wants object Y" > is always an abbreviation for "X wants some event involving Y". Let's now > define in Lojban the predicate {djicrxeventa} that means "X wants event Y". > There is nothing in Logic that prevents having another predicate in a > logical language that means "X wants object Y" (let's call it {djicrxobjekta} > and which is used to describe situations where the referent of X and the > referent of Y are in a relationship such that if the condition "X gets Y" were > to come about, X would be satisfied. And there is nothing in Logic that > prevents having a predicate {djica} such that "X djica Y" is true > whenever Y is an event and "X djicrxeventa Y" is true or whenever Y is > an object and "X djicrxobjekta Y" is true. >

Hell, nothing in logic prevents our putting in any predicate whatsoever and defining it any way at all. The issue is what does {djica} actually mean and Lojban seems to have come down on the side of {djicrxeventa}, hence the parenthetical notation.

> > > > > If I want X, X is what I really really want, and satisfaction of that want > > > > > involves my getting X. No contradiction. > > > > > > > > Except that now there is this event which was not anywhere mentioned > > > > popping up. > > > > > > Popping up where? In the explanation of what the condition that will > > > satisfy a want is? What's wrong with an event popping up in such an > > > explanation? Events pop up in all sorts of explanations of utterances > > > that don't make any reference to them. There is no logical requirement > > > that says that events may not pop up in explanations of utterances that > > > make no reference to them. > > > > For instance? > > Well, let's say "I eat an apple". If I have to explain what that means, > I might mention events like taking a bite, chewing and swallowing, that > are not mentioned in "I eat an apple". > > > > And as noted, I can want things that I could never get and > > which I yet think can be satisfied. > > I can want things that I could never get, just as I could want events that > could never happen, yes. No difference between objects and events here.

Not my point, which is that, in the case, of things, it is not always "have" or "get" that is involved. I want a greatgrandchild. I cannot — the world being what it is — have one (I am not even sure what this means but it is certainly a play on an ambiguity or at least a vaguness of "have.") Now, in this case, the mere existence of a child of my child's child would satisfy my want — no "have" reasonably involved. > > > > > It takes the same effort to get the conversants to concur with the > > > incorporation of dogs into the domain of discourse as it takes them to > > > concur with the incorporation of dogs barking, as far as I can tell. > > > > Nice use of a vagueness. It takes no effort to get them to admit that > > there is an event of dogs barking, it takes quite a bit to get them to admit > > that that event is or has at some time under discussion occurred. That > > is, getting it into the universe of discourse is free, getting it into > > the real world takes some doing. You seem to have confused the two. > > No, you seem to confuse the two. Getting them into the universe of > discourse is as free for dogs as for events. It takes work to get them > to admit that they exist/occur in the real world. No difference between > objects and events.

But the events are already there in the universe of discourse. I know you don;t like this axiom, but as long as we are arguing about Lojban, you are stuck with it. Suck it in, and make do.

> > > "Do you like dogs?" "Do you like the barking of dogs?" > > > "Have you ever seen dogs?" "Have you ever heard dogs barking?" > > > "Mary yesterday dreamt of dogs." "Mary yesterday dreamt of dogs barking." > > > > The first two pairs introduce (or at least offer for introduction) dogs and > > the barking of dogs. > > Right, no difference so far. > > > For the first in each case, the gambit can be refused, in each second > > case it cannot. Although the > > person may refuse to answer the question, he can't say "There are no > > events of dogs barking" and > > mean to make a remark about the universe of discourse rather than > > about the "real world." > > He can't say "There are no dogs" and mean to make a remark about the > universe of discourse rather than about the "real world" either. No difference > between objects and events.

But of course he can. There are universes of discourse without dogs in them and this may be one. To be sure, the other conversant may have offered to expand it, in which case, this is a refusal of the offer.

> > > > > Events, however, are always in the prior universe > > > > and so moves with them that circumvent the wish will go through, but not > > > > moves that involve something only entering with the expressed wish. > > > > > > Are you saying that for example the event of twenty-seven blue unicorns > > > flying around the moon is a member of every domain of discourse even > > > if nobody has ever mentioned it, but that unicorns are not a member of any > > > domain of discourse until someone brings them up? That's not a useful > > > metalinguistic construct, as far as I can tell. > > > > Actually, the last part seems to be a part of what you are saying. Yes, the > > event with all the unicorns is there for the asking. But unicorns can be part > > of the initial universe of discourse even if nobody ever mentions them. > > The event of unicorns flying can also be part of the initial universe of > discourse even if nobody mentions them,

And is.

> with suitable context, and the > unicorns are always there for the asking too, at least as much as the > event of unicorns flying around the moon is. So no difference so far > between objects and events.

As I said, suck it in and live with it. Or set out to change it, of course. There are some reasons to be unhappy with it, even though its advantages seem to override tham (and if we moved to the propositional form, even those disadvantages would disappear).


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 18 of Aug., 2006 20:32 GMT posts: 2388

> Jorge Llamb?as scripsit: > > > Are you saying that {la superman} and {la klark kent} have the same > > referent for us, whereas {lo nu la lois lein cu gletu la superman} and > > {lo nu la lois lein cu gletu la klark kent} have different referents > > for us? > > Indeed, they have the same and different referents, respectively, for > everyone. Superman *is* Clark Kent and when identity holds, it holds in > all possible worlds (modulo issues about existence).

Not so: some identities are contingent, when the problems about opaque contexts. With necessary identities (synonymy) Leibniz' Law works even in opaque contexts.

> But nu-contexts > are opaque and identical terms may not be substituted in them. > > > Lois would not say {mi djica la klark kent} instead of {mi djica la > > superman}, because for her {la klark kent} and {la superman} are not > > co-referential. > > Either that, or else (as I believe) they are co-referential even for her, > and she simply has appropriate false beliefs. In other words, she will > not point to Superman and say "That's Clark Kent", because she believes > that statement to be false even though we know it's true. > > > But none of this is different for objects than for events. >

I just noticed this remark in context. Of course, it is the case that if events are mentioned in opaque contexts, LL fails for them. But the events here are not mentioned in opaque contexts; they are the opaque contexts. LL holds for them, it is just that the identity of say Supernman and Clark Kent is not enough to make two events involving the two in the same way to be identical.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by JohnCowan on Fri 18 of Aug., 2006 20:47 GMT posts: 149 John E Clifford scripsit:

> Not so: some identities are contingent, when the problems about opaque > contexts.

I think Kripke's N&N disposes of these. It might have been the case that the thing called "Hesperus" turned out to be different from the thing called "Phosphorus". It could not be the case that what *we* call "Hesperus" (viz. Venus) was something other than what *we* call "Phosphorus". IOW, the former discrepancy is about a possible state of the *language* rather than a possible state of the *world*, like an alternate Middle Ages wherein "Tully" was the name applied to some other philosopher than Tully (i.e. Cicero).

To put it another way, what's contingent is not the claim that Hesperus is Phosphorus, but that "Hesperus" and "Phosphorus" corefer.

> I just noticed this remark in context. Of course, it is the case > that if events are mentioned in opaque contexts, LL fails for them. > But the events here are not mentioned in opaque contexts; they are the > opaque contexts. LL holds for them, it is just that the identity of > say Supernman and Clark Kent is not enough to make two events involving > the two in the same way to be identical.

+1

-- You let them out again, Old Man Willow! John Cowan What you be a-thinking of? You should not be waking! cowan@ccil.org Eat earth! Dig deep! Drink water! Go to sleep! Bombadil is talking. http://ccil.org/~cowan


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 18 of Aug., 2006 21:12 GMT posts: 2388

> John E Clifford scripsit: > > > Not so: some identities are contingent, when the problems about opaque > > contexts. > > I think Kripke's N&N disposes of these. It might have been the case > that the thing called "Hesperus" turned out to be different from the > thing called "Phosphorus". It could not be the case that what *we* call > "Hesperus" (viz. Venus) was something other than what *we* call "Phosphorus". > IOW, the former discrepancy is about a possible state of the *language* > rather than a possible state of the *world*, like an alternate Middle > Ages wherein "Tully" was the name applied to some other philosopher than > Tully (i.e. Cicero). > > To put it another way, what's contingent is not the claim that > Hesperus is Phosphorus, but that "Hesperus" and "Phosphorus" corefer.


Well, I never was a fan of rigid designators across the boeard (but they have some uses in counterfactuals). Still, the contingency of the coreference is enough for my project.

> > I just noticed this remark in context. Of course, it is the case > > that if events are mentioned in opaque contexts, LL fails for them. > > But the events here are not mentioned in opaque contexts; they are the > > opaque contexts. LL holds for them, it is just that the identity of > > say Supernman and Clark Kent is not enough to make two events involving > > the two in the same way to be identical. > > +1 > > — > You let them out again, Old Man Willow! John Cowan > What you be a-thinking of? You should not be waking! cowan@ccil.org > Eat earth! Dig deep! Drink water! Go to sleep! > Bombadil is talking. http://ccil.org/~cowan > > > >


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 18 of Aug., 2006 21:17 GMT On 8/18/06, John E Clifford wrote:

> It is a principle of Lojban — which you may not like, of course — that > events (all events) are always available as assignables for variables, > one of the key featurees of things in the universe of discourse. This > availability applies to all abstractions. Dogs are not under the same > principle.

If you want to call this a principle of Lojban, I have no objection with your doing so. It is not a principle of Lojban for me, but that's neither here nor there.

If you want to say that Logic requires this principle, I object, for there is nothing in Logic that would seem to require this principle, and what's more separating events and objects in such a way seems totally unfounded.

> Hell, nothing in logic prevents our putting in any predicate whatsoever > and defining it any way at all. The issue is what does {djica} actually > mean and Lojban seems to have come down on the side > of {djicrxeventa}, hence the parenthetical notation.

If all you are claiming is that {djica} has been defined that way, then I already knew that. There can be no controversy about that claim, all we need to do to verify it is check the gi'uste. I thought we were discussing the logical necessity of doing it that way.

> But the events are already there in the universe of discourse. I know you > don;t like this axiom, but as long as we are arguing about Lojban, you are > stuck with it. Suck it in, and make do.

I thought you were saying that this was a necessity for a logical language.

> As I said, suck it in and live with it. Or set out to change it, of course. > There are some reasons to be unhappy with it, even though its advantages > seem to override tham

I haven't noticed any of the advantages yet.

> (and if we moved to the propositional form, even > those disadvantages would disappear).

Some of my arguments would indeed not work if this was about propositions rather than events. (A whole other sort of complications would arise in that case.) But the separation of objects and events that Lojban attempts ("attempts" because it does not even carry it through consistently) is, in my view, just pointless.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 18 of Aug., 2006 23:07 GMT posts: 2388

> On 8/18/06, John E Clifford wrote: > > > It is a principle of Lojban — which you may not like, of course — that > > events (all events) are always available as assignables for variables, > > one of the key featurees of things in the universe of discourse. This > > availability applies to all abstractions. Dogs are not under the same > > principle. > > If you want to call this a principle of Lojban, I have no objection with > your doing so. It is not a principle of Lojban for me, but that's neither > here nor there. > > If you want to say that Logic requires this principle, I object, for there > is nothing in Logic that would seem to require this principle, and > what's more separating events and objects in such a way seems totally > unfounded.

I very carefully have never claimed it as a principle of Logic. In logical systems where it might be appropriate, I might try to claim it beloged but since they are built on such different foundations, I am not sure it would go. I could, I think, make an argument for the propositional form of this principle, but that doesn't seem to the point here. We are talking about Lojban and so we have this principle. For now, at least.

> > Hell, nothing in logic prevents our putting in any predicate whatsoever > > and defining it any way at all. The issue is what does {djica} actually > > mean and Lojban seems to have come down on the side > > of {djicrxeventa}, hence the parenthetical notation. > > If all you are claiming is that {djica} has been defined that way, then I > already knew that. There can be no controversy about that claim, > all we need to do to verify it is check the gi'uste. I thought we were > discussing the logical necessity of doing it that way.


Ther is no necessity (nor was one claimed); it is just a convenience, solving as it does a number of problems at minimum expenditure.

> > But the events are already there in the universe of discourse. I know you > > don't like this axiom, but as long as we are arguing about Lojban, you are > > stuck with it. Suck it in, and make do. > > I thought you were saying that this was a necessity for a logical language.

What did I say that led you to think that? I am pretty sure I never claimed more than convenience -- in the Lojban context at that. Oh, and that it did seemm to catch something significant in the actual brehavior of words like "want."

> > As I said, suck it in and live with it. Or set out to change it, of course. > > There are some reasons to be unhappy with it, even though its advantages > > seem to override tham > > I haven't noticed any of the advantages yet.

That's because you are blind to the opaque context problems.

> > (and if we moved to the propositional form, even > > those disadvantages would disappear). > > Some of my arguments would indeed not work if this was about propositions > rather than events. (A whole other sort of complications would arise in that > case.) But the separation of objects and events that Lojban attempts > ("attempts" because it does not even carry it through consistently) is, in > my view, just pointless.

Well, I agree that Lojban doesn't follow through on this program,, but I see it as — where it is used — a clever piece of planning. What kinds of problems do you see arising if all the various abstractions were replaced by the corresponding propositions variously related to predicates. the objections about there being such things would largely disappear, since propositions (or even sentences) are much less objectionable things to have around than events (for some reason). And all of the advantages would remain (with perhaps a bit more work). The separation of objects and events (notice it also applies to properties and the rest of the abstract lot) would be pointless without the principle that abstractions are always available for quantifiers or reference; with that, however, the separation does a good job on opaque context problems.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Tue 22 of Aug., 2006 18:57 GMT On 8/18/06, John E Clifford wrote: > The separation of objects and events (notice it also applies to properties > and the rest of the abstract lot) would be pointless without the principle > that abstractions are always available for quantifiers or reference; with > that, however, the separation does a good job on opaque context problems.

The obvious separation for me is between objects and events on one side and properties, propositions, numbers, sets on the other side. Objects and events are things located in space-time, things with extension and duration. Properties, propositions, numbers and sets are not such things.

For me the availability of events for reference or quantification is as unproblematic as the availability of dogs, unicorns, Higgs' bosons, transfinite numbers, or anything else. It all depends on the domain of discourse, and nothing is in principle more available than anything else (though of course some things are in practice more frequently talked about than others). I don't see why abstractions would be more available than non-abstractions.

As for opaque context problems, I think it is useful to divide them into ontological and epistemological. The ontological ones don't present any difficulty once we see that objects "come in trees", as you put it for events, as much as events do. Figuring out at what point in the tree we are for a given utterance is no more problematic for objects than for events.

The epistemological problems (i.e. when one of the people mentioned in a cognitive predicate would not use the words which the speaker is using) don't seem to be solved by the use of abstractions. For example, let's say Mary asked me to tell you to take a book to her. I can say:

la meris djica lo nu do bevri ti my Mary wants you to take this to her.

or:

la meris djica ti Mary wants this.

Mary is not here, so in neither case would she use the word {ti} to indicate what she wants or what she wants you to do. Whether I use {ti} inside an abstraction or not seems to make no difference. Similarly with a description:

la meris cu djica lo nu do bevri lo va cukta poi cpana lo jubme ku'o my Mary wants you to take that book on the table to her.

la meris cu djica lo va cukta poi cpana lo jubme Mary wants that book on the table.

Again, in both cases the description is what I use to tell you what Mary wants. In neither case is that the description Mary would necessarily use. She may not know that the book she wants is on the table.

The Superman/Clark Kent case is mostly a red-herring, as far as I can tell, because in many contexts the two names do have different referents. In one case it is "the superhero when dressed in his tights", and in the other case it is "the superhero in his journalist disguise". They are very rarely used as co-referring expressions. In any case, these conundrums are not resolved by using abstractions because sometimes (most times?) we do want the referring terms inside an abstraction to be the speaker's, not those of someone else mentioned by the speaker.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 24 of Aug., 2006 13:35 GMT posts: 2388

> On 8/18/06, John E Clifford wrote: > > The separation of objects and events (notice it also applies to properties > > and the rest of the abstract lot) would be pointless without the principle > > that abstractions are always available for quantifiers or reference; with > > that, however, the separation does a good job on opaque context problems. > > The obvious separation for me is between objects and events on one > side and properties, propositions, numbers, sets on the other side. > Objects and events are things located in space-time, things with > extension and duration. Properties, propositions, numbers and sets > are not such things.

Events as such are not obviously spatio-temporal except that, when they exist they occurpy space and time. Of course, the same is true of objects and not of properties and the like. I suspect that this difference is why people are uncomfortable with saying all events are available. As usual, I suggest moving to propositions or even just sentences for the significant usage here.

> For me the availability of events for reference or quantification is as > unproblematic as the availability of dogs, unicorns, Higgs' bosons, > transfinite numbers, or anything else. It all depends on the domain of > discourse, and nothing is in principle more available than anything else > (though of course some things are in practice more frequently talked > about than others). I don't see why abstractions would be more available > than non-abstractions.

Well, the Lojban axiom is that events (and all abstractions) are more readily available. If you don't like this, work to change it. But, as noted, it has some advantages (not that you recognize them) and could be replaced by other, less objectionable, devices to the same effect.

> As for opaque context problems, I think it is useful to divide them into > ontological and epistemological. The ontological ones don't present any > difficulty once we see that objects "come in trees", as you put it for events, > as much as events do. Figuring out at what point in the tree we are for > a given utterance is no more problematic for objects than for events.

Well, objects don't seem to come in trees. There is no dog that subsumes other dogs; each dog is a unique individual without reference or implication to other dogs. The concept DOG does head a tree and has subordinate subsuming concepts alol the way down th individual dog concepts (if you want to admit such), but that is a different matter and one you have said is not relevant to the present problem (though it does provide a way of dealing with it — in which case we ought again say we are using it).

> The epistemological problems (i.e. when one of the people mentioned > in a cognitive predicate would not use the words which the speaker is using) > don't seem to be solved by the use of abstractions. For example, let's > say Mary asked me to tell you to take a book to her. I can say: > > la meris djica lo nu do bevri ti my > Mary wants you to take this to her. > > or: > > la meris djica ti > Mary wants this. > > Mary is not here, so in neither case would she use the word {ti} to indicate > what she wants or what she wants you to do. Whether I use {ti} inside an > abstraction or not seems to make no difference. Similarly with a description: > > la meris cu djica lo nu do bevri lo va cukta poi cpana lo jubme ku'o my > Mary wants you to take that book on the table to her. > > la meris cu djica lo va cukta poi cpana lo jubme > Mary wants that book on the table. > > Again, in both cases the description is what I use to tell you what Mary > wants. In neither case is that the description Mary would necessarily use. > She may not know that the book she wants is on the table. > As usual, the use of token reflexive examples tends to obscure the main issues. For one thing, "this" and the like (deictics) cutrs off one of the sources of problems: a pointed to thing definitely does exist in the relevant world. So the issue of generalization is not a problem. The issue of identity is more complex but even so the normal rule clearly fails sometimes and so isminvalid. What is interesting is that it also seems to work often, maybe even enough to be generally reliable, so the question is what are the parameters for successful use. It appears that one factor is whether the intenser is aware of the identity and able to make the connection, but this is neither necessary or sufficient. In the Mary case there is probably also some token-type ambiguity going on: does mMary want War and Peace and the book on the table is War and Peace? In that case, half od the condition for her wanting the book on the table is met, but not the part about"if the want were satisfied then she would have gotten the book on the table" in the token sense, rather than the type sense. I am not sure whether all apparent exceptions to the failure of LL fall into just these various kinds of confusions, but they may — which would simplify the parameter issue.

> The Superman/Clark Kent case is mostly a red-herring, as far as I can tell, > because in many contexts the two names do have different referents. In one > case it is "the superhero when dressed in his tights", and in the other case > it is "the superhero in his journalist disguise".

These are different senses, not different referents and the repetition of "the superhero" suggests that the referents are in fact the same.

>They are very rarely used > as co-referring expressions. In any case, these conundrums are not resolved > by using abstractions because sometimes (most times?) we do want the > referring terms inside an abstraction to be the speaker's, not those of someone > else mentioned by the speaker.

Yes, this is the source of the problem. Lois Lane does not believe that Superman and Clark Kent are the same individual, so she cannot accept the transition. I, reporting her wish, knowing better, might make the shift, but in doing so I misrepresent her mental state. If asked if she wanted Clark Kent, she would say "No" and surely she knows her own desires. She might even say this if she knew the identity. But generally, my representation of her wants are accurate only insofar as they correspond to hers — perhaps up to including her active beliefs. Wants take us into her want world, where the facts of the real world (in this case the world of this particular comic strip, not the real "real world") need not apply.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by JohnCowan on Thu 24 of Aug., 2006 13:35 GMT posts: 149 John E Clifford scripsit:

> Well, objects don't seem to come in trees. There is no dog that > subsumes other dogs; each dog is a unique individual without reference > or implication to other dogs.

Doesn't a dog subsume its legs and other body parts?

-- Even the best of friends cannot John Cowan attend each others' funeral. cowan@ccil.org --Kehlog Albran, The Profit http://www.ccil.org/~cowan


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 24 of Aug., 2006 13:42 GMT posts: 2388 Not in the relevant sense, that in which, for example, "Something happens" subsumes all events.


> John E Clifford scripsit: > > > Well, objects don't seem to come in trees. There is no dog that > > subsumes other dogs; each dog is a unique individual without reference > > or implication to other dogs. > > Doesn't a dog subsume its legs and other body parts? > > — > Even the best of friends cannot John Cowan > attend each others' funeral. cowan@ccil.org > --Kehlog Albran, The Profit http://www.ccil.org/~cowan > > > >


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 24 of Aug., 2006 14:02 GMT posts: 2388


> Not in the relevant sense, that in which, for example, "Something happens" subsumes all events.

And also, this subsumption would not help in the issue at hand, guaranteeing that the right thing (not thing part) existed in some subjunctive world.

> --- John Cowan wrote: > > > John E Clifford scripsit: > > > > > Well, objects don't seem to come in trees. There is no dog that > > > subsumes other dogs; each dog is a unique individual without reference > > > or implication to other dogs. > > > > Doesn't a dog subsume its legs and other body parts? > > > > — > > Even the best of friends cannot John Cowan > > attend each others' funeral. cowan@ccil.org > > --Kehlog Albran, The Profit http://www.ccil.org/~cowan > > > > > > > > > > > > >


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 24 of Aug., 2006 15:59 GMT On 8/24/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > The obvious separation for me is between objects and events on one > > side and properties, propositions, numbers, sets on the other side. > > Objects and events are things located in space-time, things with > > extension and duration. Properties, propositions, numbers and sets > > are not such things. > > Events as such are not obviously spatio-temporal except that, when > they exist they occurpy space and time. Of course, the same is true > of objects and not of properties and the like.

Right, that's more or less what I had in mind. Objects and events can, archetypically, have extension and duration, while properties, propositions, numbers, sets don't.

> I suspect > that this difference is why people are uncomfortable with saying all > events are available.

The funny thing is that in order to even say that and make sense, they need them to be available.

> As usual, I suggest moving to propositions or > even just sentences for the significant usage here.

Do you take {lu mi klama le zarci li'u} to refer to a sentence type, a sentence token, or either depending on the context?

> > For me the availability of events for reference or quantification is as > > unproblematic as the availability of dogs, unicorns, Higgs' bosons, > > transfinite numbers, or anything else. It all depends on the domain of > > discourse, and nothing is in principle more available than anything else > > (though of course some things are in practice more frequently talked > > about than others). I don't see why abstractions would be more available > > than non-abstractions. > > Well, the Lojban axiom is that events (and all abstractions) are more > readily available. If you don't like this, work to change it.

It sounds like a metalinguistic axiom. Given a Lojban utterance, if we use different metalinguistic interpretations but end up understanding each other in what we say, does it really matter that we use different metalinguistic theories of interpretation? Is it really necessary to have official metalinguistic axioms be part of what we call "Lojban"? It doesn't sound to me like something a user of Lojban needs to adhere to (or reject) in order to produce and understand "correct" Lojban utterances effectively.


> > As for opaque context problems, I think it is useful to divide them into > > ontological and epistemological. The ontological ones don't present any > > difficulty once we see that objects "come in trees", as you put it for events, > > as much as events do. Figuring out at what point in the tree we are for > > a given utterance is no more problematic for objects than for events. > > Well, objects don't seem to come in trees.

To me objects and events don't seem to differ in this respect.

> There is no dog that subsumes > other dogs; each dog is a unique individual without reference or implication > to other dogs.

Whereas there is some event of "Fido runs" that subsumes other events of "Fido runs"?

> The concept DOG does head a > tree and has subordinate subsuming concepts alol the way down th > individual dog concepts (if you want to admit such), but that is a different > matter and one you have said is not relevant to the present problem (though > it does provide a way of dealing with it — in which case we ought again > say we are using it).

All I say is that it works the same for dogs as for events. If what's relevant is a concept DOG, then what's relevant is a concept FIDO RUNNING. If what's relevant is Fido running, then what's relevant is dogs. {lo gerku} and {lo nu la fidos cu bajra} behave in the same way in all relevant aspects.


> As usual, the use of token reflexive examples tends to obscure the > main issues.

I don't understand what you mean by "token reflexive".

> For one thing, > "this" and the like (deictics) cutrs off one of the sources of problems: > a pointed to thing definitely does exist in the relevant world.

The question was whether we can use {ti} inside an abstraction even though the "intenser" would not use it.

> So the issue of generalization is not a problem. > The issue of identity is more complex but even so the normal rule clearly > fails sometimes and so isminvalid. What is interesting is that it also seems > to work often, maybe even enough to be generally reliable, so the question > is what are the parameters for successful use. It appears > that one factor is whether the intenser is aware of the identity and able to > make the connection, but this is neither necessary or sufficient. In the > Mary case there is probably also some token-type ambiguity going on: > does mMary want War and Peace and the book on the table is War and > Peace? In that case, half od the condition for her wanting the book on the > table is met, but not the part about"if the want were satisfied then she > would have gotten the book on the table" in the token sense, rather than > the type sense. I am not sure whether all apparent exceptions to the > failure of LL fall into just these various kinds of confusions, but they > may — which would simplify the parameter issue.

So you seem to be saying that token/type issues are not solved by the use of abstractions. The same issues arise whether we use an abstraction or not in the x2 of djica.

Suppose it's about a particular copy of War and Peace, one I had borrowed from Mary and now she wants it back. Can I say {la meris djica lo nu do bevri lo va cukta poi cpana lo jubme ku'o my} to you, even though Mary does not know that the book she wants is on the table?

> > The Superman/Clark Kent case is mostly a red-herring, as far as I can tell, > > because in many contexts the two names do have different referents. In one > > case it is "the superhero when dressed in his tights", and in the other case > > it is "the superhero in his journalist disguise". > > These are different senses, not different referents and the repetition of > "the superhero" suggests that the referents are in fact the same.

If "Superman" and "the superhero in his journalist disguise" are co-referential, why is "the superhero in his journalist disguise is flying" and "Superman is flying" seem to say such different things? Is the x1 of "is flying" an opaque context, so that co-referential terms are not interchangeable there?


> >They are very rarely used > > as co-referring expressions. In any case, these conundrums are not resolved > > by using abstractions because sometimes (most times?) we do want the > > referring terms inside an abstraction to be the speaker's, not those of someone > > else mentioned by the speaker. > > Yes, this is the source of the problem. Lois Lane does not believe > that Superman and Clark Kent are the same individual, so she cannot > accept the transition. I, reporting her wish, knowing better, might make > the shift, but in doing so I misrepresent her mental state. If asked if she > wanted Clark Kent, she would say "No" and surely she knows her own > desires. She might even say this if she knew the identity. But generally, > my representation of her wants are accurate only > insofar as they correspond to hers — perhaps up to including her active > beliefs. Wants take us into her want world, where the facts of the real > world (in this case the world of this particular > comic strip, not the real "real world") need not apply.

And whether we use an abstraction sumti or not seems to make little difference in our success or failure in representing her wants.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 24 of Aug., 2006 17:52 GMT posts: 2388

> On 8/24/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > The obvious separation for me is between objects and events on one > > > side and properties, propositions, numbers, sets on the other side. > > > Objects and events are things located in space-time, things with > > > extension and duration. Properties, propositions, numbers and sets > > > are not such things. > > > > Events as such are not obviously spatio-temporal except that, when > > they exist they occurpy space and time. Of course, the same is true > > of objects and not of properties and the like. > > Right, that's more or less what I had in mind. Objects and events can, > archetypically, have extension and duration, while properties, propositions, > numbers, sets don't. > > > I suspect > > that this difference is why people are uncomfortable with saying all > > events are available. > > The funny thing is that in order to even say that and make sense, they > need them to be available.

Well, I don't buy that, since this is just a case of predication, not reference (or at least may be).

> > As usual, I suggest moving to propositions or > > even just sentences for the significant usage here. > > Do you take {lu mi klama le zarci li'u} to refer to a sentence type, a sentence > token, or either depending on the context?

It has to be the last, although Lojban does not have a good way to make the distinction. Of course, if you use some sentence type, what you use is a token, so saying that sentences are available covers both cases.

> > > > For me the availability of events for reference or quantification is as > > > unproblematic as the availability of dogs, unicorns, Higgs' bosons, > > > transfinite numbers, or anything else. It all depends on the domain of > > > discourse, and nothing is in principle more available than anything else > > > (though of course some things are in practice more frequently talked > > > about than others). I don't see why abstractions would be more available > > > than non-abstractions. > > > > Well, the Lojban axiom is that events (and all abstractions) are more > > readily available. If you don't like this, work to change it. > > It sounds like a metalinguistic axiom. Given a Lojban utterance, if we > use different metalinguistic interpretations but end up understanding > each other in what we say, does it really matter that we use different > metalinguistic theories of interpretation? Is it really necessary to have > official metalinguistic axioms be part of what we call "Lojban"? It doesn't > sound to me like something a user of Lojban needs to adhere to (or > reject) in order to produce and understand "correct" Lojban utterances > effectively. > It is metalinguistic in the sense that it specifies the parameters of any universe of discourse. As far as I can tell, the reason for it is precisely to avoid the kinds of problems we have been talking about. In that sense, then, not having this axiom would generate a new set of problems (or rather, the old problems in a new set of cases). I suppose that the justification for having such a rule or axiom is ultimately going to be pragmatic and/or practical: does it contribute to the smooth flow of discourse (in this case taking logical features into consideration — something not always done). At the moment, it loks to me as though it does and I don't see another way to achieve the same result. On the other hand, it is a strange rule, so I would love to have an alternative — that didn't introduce even worse new problems.

> > > As for opaque context problems, I think it is useful to divide them into > > > ontological and epistemological. The ontological ones don't present any > > > difficulty once we see that objects "come in trees", as you put it for events, > > > as much as events do. Figuring out at what point in the tree we are for > > > a given utterance is no more problematic for objects than for events. > > > > Well, objects don't seem to come in trees. > > To me objects and events don't seem to differ in this respect. > > > There is no dog that subsumes > > other dogs; each dog is a unique individual without reference or implication > > to other dogs. > > Whereas there is some event of "Fido runs" that subsumes other events > of "Fido runs"?

Yes: "Fido runs" subsumes "Fido runs rapidly," "Fido runs slowing," "fido is running now." and so on indefinitely. Whenever any of the subsumed events occurs, so does the subsuming.

> > The concept DOG does head a > > tree and has subordinate subsuming concepts all the way down to > > individual dog concepts (if you want to admit such), but that is a different > > matter and one you have said is not relevant to the present problem (though > > it does provide a way of dealing with it — in which case we ought again > > say we are using it). > > All I say is that it works the same for dogs as for events. If what's relevant > is a concept DOG, then what's relevant is a concept FIDO RUNNING. If > what's relevant is Fido running, then what's relevant is dogs. {lo gerku} > and {lo nu la fidos cu bajra} behave in the same way in all relevant aspects.

Is there a dog that runs whenever any dog runs? Indeed, is there a dog that exists whenever any dog exists? To be sure, there is a problem, akin to token-type, for events (there is not for dogs), but it does not pose any serious problems here. The two espressions may be the same grammatically, but they are not semantically (metaphysically?).

> > > As usual, the use of token reflexive examples tends to obscure the > > main issues. > > I don't understand what you mean by "token reflexive".

A token reflexive expression is one which makes explicit reference to the utternace in which it occurs or the circumstance of that utterance. Typical cases are "I" and "you," "this," "here," "now" and so on.

> > For one thing, > > "this" and the like (deictics) cut off one of the sources of problems: > > a pointed to thing definitely does exist in the relevant world. > > The question was whether we can use {ti} inside an abstraction even > though the "intenser" would not use it.

And the answer is, sometime yes, sometimes no, leaving the question of what separates the two cases. I think the basic answer is "no," leaving the "yes' cases to be explained, Clearly the intenser can use them and sometimes, with existent things, they can be used if the things are identical. But there seem to be further conditions and I don't know what they are.

> > So the issue of generalization is not a problem. > > The issue of identity is more complex but even so the normal rule clearly > > fails sometimes and so is invalid. What is interesting is that it also seems > > to work often, maybe even enough to be generally reliable, so the question > > is what are the parameters for successful use. It appears > > that one factor is whether the intenser is aware of the identity and able to > > make the connection, but this is neither necessary or sufficient. In the > > Mary case there is probably also some token-type ambiguity going on: > > does mMary want War and Peace and the book on the table is War and > > Peace? In that case, half od the condition for her wanting the book on the > > table is met, but not the part about"if the want were satisfied then she > > would have gotten the book on the table" in the token sense, rather than > > the type sense. I am not sure whether all apparent exceptions to the > > failure of LL fall into just these various kinds of confusions, but they > > may — which would simplify the parameter issue. > > So you seem to be saying that token/type issues are not solved by the > use of abstractions. The same issues arise whether we use an abstraction > or not in the x2 of djica.

Actually, I wasn't saying anything about the use of abstraction, I was just trying to work through what seems to be the phenomena involved. The token-type problem can recur in abstractions but the identity problem apparently does not, since two events are not the same in either case.

> Suppose it's about a particular copy of War and Peace, one I had borrowed > from Mary and now she wants it back. Can I say {la meris djica lo nu do > bevri lo va cukta poi cpana lo jubme ku'o my} to you, even though Mary > does not know that the book she wants is on the table?

And does not necessarily have any wants with regard to me; she just want the book, however it gets to her. So that part does not carry over — unless she specified who was to bring it. But the part just about the book {la maris djica tu'a lo cukta poi cpana lo jubme} does seem to work with appropriate rigid designators (subjunctives are where those work, after all): if she were to get the book now on the table her want would be satisfied, and in this case, if her want were to be satisfied she would have gotten the book now on the table. I'm not sure this example clarifies much.Perhaps each case must be taken up separately.

> > > The Superman/Clark Kent case is mostly a red-herring, as far as I can tell, > > > because in many contexts the two names do have different referents. In one > > > case it is "the superhero when dressed in his tights", and in the other case > > > it is "the superhero in his journalist disguise". > > > > These are different senses, not different referents and the repetition of > > "the superhero" suggests that the referents are in fact the same. > > If "Superman" and "the superhero in his journalist disguise" are co-referential, > why is "the superhero in his journalist disguise is flying" and "Superman > is flying" seem to say such different things? Is the x1 of "is flying" an > opaque context, so that co-referential terms are not interchangeable there?

They both refer to the same thing but they get to that thing in different ways. Thi is the old question which led to the explication of the sense-reference distinction (the original was "How is "Hesperus = Phosphorus," if true, different from "Hesperus = Hesperus"? The direct correlate would be "The Evening Star is visible just before sunrise.") Even though the referent is the same, this particular way of referring to that object is at least incongruous in this context (when he is not in his journalist disguise) and possible worse: implying that he is wearing the disguise while flying. So, the two versions are not wrong semantically — that is, they are both true or both false — but the one is wrong pragmatically since it is incongruous and misleading. Of course, a person who does not know the identity would think the second one false, precisely because there is no one in a journalist disguise flying. He would simply be wrong, though understandably so.

> > > >They are very rarely used > > > as co-referring expressions. In any case, these conundrums are not resolved > > > by using abstractions because sometimes (most times?) we do want the > > > referring terms inside an abstraction to be the speaker's, not those of someone > > > else mentioned by the speaker. > > > > Yes, this is the source of the problem. Lois Lane does not believe > > that Superman and Clark Kent are the same individual, so she cannot > > accept the transition. I, reporting her wish, knowing better, might make > > the shift, but in doing so I misrepresent her mental state. If asked if she > > wanted Clark Kent, she would say "No" and surely she knows her own > > desires. She might even say this if she knew the identity. But generally, > > my representation of her wants are accurate only > > insofar as they correspond to hers — perhaps up to including her active > > beliefs. Wants take us into her want world, where the facts of the real > > world (in this case the world of this particular > > comic strip, not the real "real world") need not apply. > > And whether we use an abstraction sumti or not seems to make little > difference in our success or failure in representing her wants.

What exactly do abstractions do here? For generalizations we have the rule that every event -- and thus the one desired — is always available, so generalization goes through. For identity, we have that exchange of identical events goes through. The saving part of this is that events differing only in replacement of one object by another factually identical to it are not generally identical, That is, the event of Superman flying is not identical to the event of Clark Kent flying, even though Superman is identical to Clark Kent (and even though, in real world of this comic book, the same occurrence is covered by both). So the virtue of using events is that it saves LL and does so by indicating that replacement of one term of a factually identical pair by the other does not guarantee that the two resulting expressions are coreferential. The point is that abstractions always abstract (duh!) from the real situation (where the identity holds)to an alternate world (so to speak)where the factual identity cannot be assumed. This does not mean that the two event description do not represent the same event or equivalent ones with repect to some intensional act; it just means that we don't get from one to the other by LL, but rather by some other -- possibly valid, hopefully at least correct in thecase — means. Remember that (almost) the whole point of using abstractions is to save logical rules and it does eem to work. It is hard to see a way of doing it otherwise (well, we can go by places, but that is messy and unLojbanic). Of course, I think it also reflects a reality about our (in this case) wants, but that is not crucial.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 24 of Aug., 2006 20:00 GMT On 8/24/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > On 8/24/06, John E Clifford wrote: > > > I suspect > > > that this difference is why people are uncomfortable with saying all > > > events are available. > > > > The funny thing is that in order to even say that and make sense, they > > need them to be available. > > Well, I don't buy that, since this is just a case of predication, > not reference (or at least may be).

Even so, if this makes them uncomfortable:

ro da zo'u ganai da faursucta gi da selsnika'e "For every x, if x is an event abstraction, x is referrable."

then they would want to deny it:

su'o da zo'u ge da faursucta ginai da selsnika'e "There is some x such that x is an event abstraction and x is not referrable."

That someone would feel more comfortable with that strikes me as funny.


> > Do you take {lu mi klama le zarci li'u} to refer to a sentence type, a sentence > > token, or either depending on the context? > > It has to be the last, although Lojban does not have a good way to make > the distinction. Of course, if you use some sentence type, what you use > is a token, so saying that sentences are > available covers both cases.

And if that is good enough for sentences, why should it not be good enough for dogs?


> > > Well, the Lojban axiom is that events (and all abstractions) are more > > > readily available. If you don't like this, work to change it. > > > > It sounds like a metalinguistic axiom. Given a Lojban utterance, if we > > use different metalinguistic interpretations but end up understanding > > each other in what we say, does it really matter that we use different > > metalinguistic theories of interpretation? Is it really necessary to have > > official metalinguistic axioms be part of what we call "Lojban"? It doesn't > > sound to me like something a user of Lojban needs to adhere to (or > > reject) in order to produce and understand "correct" Lojban utterances > > effectively. > > > It is metalinguistic in the sense that it specifies the parameters of any > universe of discourse. As far as I can tell, the reason for it is precisely > to avoid the kinds of problems we have been talking about. In that sense, > then, not having this axiom would generate a new set of problems > (or rather, the old problems in a new set of cases). I suppose that the > justification for having such a rule or axiom is ultimately going to be > pragmatic and/or practical: does it contribute to the smooth flow of > discourse (in this case taking logical features into consideration -- > something not always done). At the moment, it loks to me as though > it does and I don't see another way to achieve the same result. On the > other hand, it is a strange rule, so I would love to have an > alternative — that didn't introduce even worse new problems.

It seems to me that extending the rule that you already allow for sentences and events to dogs would make things flow even more smoothly. No logical feature is broken with this extension, since there is nothing in Logic that distinguishes dogs from sentences or events.


> > > There is no dog that subsumes > > > other dogs; each dog is a unique individual without reference or implication > > > to other dogs. > > > > Whereas there is some event of "Fido runs" that subsumes other events > > of "Fido runs"? > > Yes: "Fido runs" subsumes "Fido runs rapidly," "Fido runs slowing," > "fido is running now." and so on indefinitely. Whenever any of the > subsumed events occurs, so does the subsuming.

So when I see that Fido runs rapidly, and when I see that Fido runs slowing, I'm seeing the subsuming event that Fido runs. But for some reason you don't allow this for dogs, so when I see Fido and when I see Spot, {mi viska lo gerku} cannot describe both situations with {lo gerku} having the same subsuming referent in both cases.

> > > The concept DOG does head a > > > tree and has subordinate subsuming concepts all the way down to > > > individual dog concepts (if you want to admit such), but that is a different > > > matter and one you have said is not relevant to the present problem (though > > > it does provide a way of dealing with it — in which case we ought again > > > say we are using it). > > > > All I say is that it works the same for dogs as for events. If what's relevant > > is a concept DOG, then what's relevant is a concept FIDO RUNNING. If > > what's relevant is Fido running, then what's relevant is dogs. {lo gerku} > > and {lo nu la fidos cu bajra} behave in the same way in all relevant aspects. > > Is there a dog that runs whenever any dog runs? Indeed, is there a dog > that exists whenever any dog exists?

Yes, whenever a dog runs, lo gerku runs. This is no different from saying that whenever an event of Fido running happens, lo nu la fidos cu bajra happens.

> To be sure, there is a problem, akin to token-type, for events (there is not > for dogs), but it does not pose any serious problems here. The two > espressions may be the same grammatically, but they are not semantically > (metaphysically?).

Why are they not the same semantically or metaphysically? By decree? If so, what is gained by decreeing it so?


> > The question was whether we can use {ti} inside an abstraction even > > though the "intenser" would not use it. > > And the answer is, sometime yes, sometimes no, leaving the question > of what separates the two cases. I think the basic answer is "no," leaving > the "yes' cases to be explained, Clearly the intenser can use them and > sometimes, with existent things, they can be used if the things are > identical. But there seem to be further conditions and I don't know what > they are.

I suspect it's the other way around. The basic case is that the words used inside an abstraction (unlike the case of quotation) are the speaker's, and it is only pragmatically that they sometimes acquire a quote-like quality in favor of the intenser.


> > > But generally, > > > my representation of her wants are accurate only > > > insofar as they correspond to hers — perhaps up to including her active > > > beliefs. Wants take us into her want world, where the facts of the real > > > world (in this case the world of this particular > > > comic strip, not the real "real world") need not apply. > > > > And whether we use an abstraction sumti or not seems to make little > > difference in our success or failure in representing her wants. > > What exactly do abstractions do here? For generalizations we have the > rule that every event — and thus the one desired — is always available, > so generalization goes through.

And there does not seem to be any reason why this rule should not also apply to objects.

> For identity, we have that exchange of identical events goes through.

Same as for objects.

> The saving part of this is that events differing only in replacement of one > object by another factually identical to it are not generally identical,

That sounds wrong, at least expressed that way. I don't see how you could replace something by itself and make a difference.

> That is, the event of Superman flying is not identical to the event of Clark > Kent flying, even though Superman is identical to Clark Kent (and even > though, in real world of this comic book, the same occurrence is covered > by both).

That sounds weird. It seems easier to have "Clark Kent" and "Superman" refer to different slices of the same character.

> So the virtue of using events is that it saves LL and does > so by indicating that replacement of one term of a factually identical pair > by the other does not guarantee that the two resulting expressions > are coreferential.

That does not save LL, as far as I can see. We have two indiscernible things (the event of Superman flying and the event of Clark Kent flying) that are not identical.


> The point is that abstractions > always abstract (duh!) from the real situation (where the identity holds) > to an alternate world (so to speak)where the factual identity cannot be > assumed. This does not mean that the two event description do not > represent the same event or equivalent ones with repect to some intensional > act; it just means that we don't get from one to the other by LL, but rather > by some other — possibly valid, hopefully at least correct in thecase -- > means. Remember that (almost) the whole point of using abstractions is > to save logical rules and it does eem to work.

Would you say that LL is a logical rule, or an ontological one?

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 25 of Aug., 2006 00:38 GMT posts: 2388

> On 8/24/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > On 8/24/06, John E Clifford wrote: > > > > I suspect > > > > that this difference is why people are uncomfortable with saying all > > > > events are available. > > > > > > The funny thing is that in order to even say that and make sense, they > > > need them to be available. > > > > Well, I don't buy that, since this is just a case of predication, > > not reference (or at least may be). > > Even so, if this makes them uncomfortable: > > ro da zo'u ganai da faursucta gi da selsnika'e > "For every x, if x is an event abstraction, x is referrable." > > then they would want to deny it: > > su'o da zo'u ge da faursucta ginai da selsnika'e > "There is some x such that x is an event abstraction and x is not referrable." > > That someone would feel more comfortable with that strikes me as funny.

Wel, as you have noted, the rule is metalinguistic. Of your odd cases (by you), the first is a tautology, since ro da selsnika'e. The other is therefore a contradiction. since the first is a tautology, I can't imagine that anyone is uncomfortable with it. The first consequently does not say what the axiom says — since it is true even if there are no events in the domain.

> > > > Do you take {lu mi klama le zarci li'u} to refer to a sentence type, a sentence > > > token, or either depending on the context? > > > > It has to be the last, although Lojban does not have a good way to make > > the distinction. Of course, if you use some sentence type, what you use > > is a token, so saying that sentences are > > available covers both cases. > > And if that is good enough for sentences, why should it not be good enough > for dogs?

Because dog types are not dogs, but sentence types are sentences. > > > > > Well, the Lojban axiom is that events (and all abstractions) are more > > > > readily available. If you don't like this, work to change it. > > > > > > It sounds like a metalinguistic axiom. Given a Lojban utterance, if we > > > use different metalinguistic interpretations but end up understanding > > > each other in what we say, does it really matter that we use different > > > metalinguistic theories of interpretation? Is it really necessary to have > > > official metalinguistic axioms be part of what we call "Lojban"? It doesn't > > > sound to me like something a user of Lojban needs to adhere to (or > > > reject) in order to produce and understand "correct" Lojban utterances > > > effectively. > > > > > It is metalinguistic in the sense that it specifies the parameters of any > > universe of discourse. As far as I can tell, the reason for it is precisely > > to avoid the kinds of problems we have been talking about. In that sense, > > then, not having this axiom would generate a new set of problems > > (or rather, the old problems in a new set of cases). I suppose that the > > justification for having such a rule or axiom is ultimately going to be > > pragmatic and/or practical: does it contribute to the smooth flow of > > discourse (in this case taking logical features into consideration -- > > something not always done). At the moment, it loks to me as though > > it does and I don't see another way to achieve the same result. On the > > other hand, it is a strange rule, so I would love to have an > > alternative — that didn't introduce even worse new problems. > > It seems to me that extending the rule that you already allow for > sentences and events to dogs would make things flow even more > smoothly. No logical feature is broken with this extension, since there > is nothing in Logic that distinguishes dogs from sentences or events. > Well, some distinctions have been noted. We are sure we can get a sentence (hence a proposition, property and so on) when we need one, even if we don't know exactly what one it is. There is no such guarantee with dog.Even if you add every possible dog, it is not guranteed that the one someone wants is in there, never mind the possibility of getting someone to agree to the extension. I suppose the axiom comes down to: "The other cannot refuse to extend the domain by a sentence."

> > > > There is no dog that subsumes > > > > other dogs; each dog is a unique individual without reference or implication > > > > to other dogs. > > > > > > Whereas there is some event of "Fido runs" that subsumes other events > > > of "Fido runs"? > > > > Yes: "Fido runs" subsumes "Fido runs rapidly," "Fido runs slowly," > > "Fido is running now." and so on indefinitely. Whenever any of the > > subsumed events occurs, so does the subsuming. > > So when I see that Fido runs rapidly, and when I see that Fido runs > slowly, I'm seeing the subsuming event that Fido runs.

Well, now you are getting into a problem area, since one sense of seeing events probably is intensional, so that — not with these cases I think — someone could see a subsumed event and not the subsuming. Of course, there is another sense of "see" where thios is not possible.

> But for some reason you don't allow this for dogs, so when I see > Fido and when I see Spot, {mi viska lo gerku} cannot describe both > situations with {lo gerku} having the same subsuming referent in both > cases.

Yes, "lo gerku" covers both cases, but it is not a dog. You can find an sentence which subsumes the tindividual sentemces, but you can't find a dog that subsumes the individual dogs. And it is the dog that we have to find.

> > > > The concept DOG does head a > > > > tree and has subordinate subsuming concepts all the way down to > > > > individual dog concepts (if you want to admit such), but that is a different > > > > matter and one you have said is not relevant to the present problem (though > > > > it does provide a way of dealing with it — in which case we ought again > > > > say we are using it). > > > > > > All I say is that it works the same for dogs as for events. If what's relevant > > > is a concept DOG, then what's relevant is a concept FIDO RUNNING. If > > > what's relevant is Fido running, then what's relevant is dogs. {lo gerku} > > > and {lo nu la fidos cu bajra} behave in the same way in all relevant aspects. > > > > Is there a dog that runs whenever any dog runs? Indeed, is there a dog > > that exists whenever any dog exists? > > Yes, whenever a dog runs, lo gerku runs. This is no different from saying that > whenever an event of Fido running happens, lo nu la fidos cu bajra happens.

Oh, how I wish I hadn't let my Latin slip; this argument is looking so Thirteenth Century that I want to bring out the whole panoply about proprietates terminorm. But I beeter stick to contemporary stuff. Your comment, while true, is not relevant to the point at hand. {lo gerko} is not the name of a dog, nor systematically a device to refer to a dog. {lo gerku cu gerku} is true only distributively — even when it refers to only one dog (which there is not generally any reasion to think it does), similarly for {lo gerku cu bajra}. On the other hand, {la fidos cu bajra} and {la fidos cu gerku} are personal/singular/collective predications. "Fido" refers to a dog; {lo gerku} refers to some number of dogs. And so on. But notice that {lo nu mi ponse lo gerku} and {lo nu mi ponse la fidos}, both refer to events in exactly the same way even though one subsumes the other. In the case of dogs, you don't have a thing that is both a dog and subsumes other dogs, you only have a pair of sentences one of which entails the other. for events you have both.


> > To be sure, there is a problem, akin to token-type, for events (there is not > > for dogs), but it does not pose any serious problems here. The two > > espressions may be the same grammatically, but they are not semantically > > (metaphysically?). > > Why are they not the same semantically or metaphysically? By decree? > If so, what is gained by decreeing it so? > The only decree is in deciding that a real difference in logical performance is wotht remarking, but that decision is probably out of our hands, given that we are constructing a logical language.


> > > The question was whether we can use {ti} inside an abstraction even > > > though the "intenser" would not use it. > > > > And the answer is, sometime yes, sometimes no, leaving the question > > of what separates the two cases. I think the basic answer is "no," leaving > > the "yes' cases to be explained, Clearly the intenser can use them and > > sometimes, with existent things, they can be used if the things are > > identical. But there seem to be further conditions and I don't know what > > they are. > > I suspect it's the other way around. The basic case is that the words used > inside an abstraction (unlike the case of quotation) are the speaker's, > and it is only pragmatically that they sometimes acquire a quote-like quality > in favor of the intenser.

By "the speaker," I gather you mean the person who reports on someone else's intensional state, the one who says "Mary dreamed ..." or "Mary wants... ." It is of course his words, but he is accurately reporting only if the report accurately reflects that state. If he says Lois want to sleep with Clark Kent when she wants to sleep with Superman, he is not doing this. Nor is he when she wants War and Peace and he reports she wants that book there (which is, indeed, War and Peace)-- or, rather, it takes a special understanding of the words to make it come out right. The special meaning may be pragmatically arranged for — and indeed all the successful cases may be pragmatically saved from inaccuaracy by things like "She couldn't know even that there was a book there, so referring to what she wants as "that book there" must be based on some other criteria than her literal want" and so on. Whicheveris the basic answer — and ther may not be a basic answer — the other needs explaining and, of course, following the intenser's thought is easy to explain — it is just accuaracy, a goal always assumed.


> > > > > But generally, > > > > my representation of her wants are accurate only > > > > insofar as they correspond to hers — perhaps up to including her active > > > > beliefs. Wants take us into her want world, where the facts of the real > > > > world (in this case the world of this particular > > > > comic strip, not the real "real world") need not apply. > > > > > > And whether we use an abstraction sumti or not seems to make little > > > difference in our success or failure in representing her wants. > > > > What exactly do abstractions do here? For generalizations we have the > > rule that every event — and thus the one desired — is always available, > > so generalization goes through. > > And there does not seem to be any reason why this rule should not also > apply to objects.

Wello, I have given you a list of reasons why evrents are different from ordinary things. The fact that you don't like these differences or that you don't consider them important does not change the fact that there are these differences and that they are the basis for the distinction. Freely entering new things violates the conventions of conversation and also doesn't solve any problems (without adding a number of further conditions on what are added, conditions which are even less appealing than the axiom — at least in its propositional form.

> > For identity, we have that exchange of identical events goes through. > > Same as for objects.

No: Superman and Clark Kent.

> > The saving part of this is that events differing only in replacement of one > > object by another factually identical to it are not generally identical, > > That sounds wrong, at least expressed that way. I don't see how you could > replace something by itself and make a difference.

Identity is always about expressions, as you well know, i.e., it is a claim of coreferntiality somewhat oddly expressed. So we are saying that in an expression referring to an event, if we replace one referring expression by another which is in fact coreferential, the result is not usually (or at least not guaranteed to be) a reference to the same event. The factual identity does not carry over to the event world (in the real world, it will typically refer to the same actual event, but come at it from different angles, like the Morning Star and the Evening star).

> > That is, the event of Superman flying is not identical to the event of Clark > > Kent flying, even though Superman is identical to Clark Kent (and even > > though, in real world of this comic book, the same occurrence is covered > > by both). > > That sounds weird. It seems easier to have "Clark Kent" and "Superman" > refer to different slices of the same character.

They refer to the same character period. A slice of one is a slice of the other. To be sure, that character may be acting as one persona rather than another at various times, but that hardly means that they slices in those times are slices of different wholes (discontinuous people, for one thing, are probably not allowed outside theology and maybe abnormal psych).

> > So the virtue of using events is that it saves LL and does > > so by indicating that replacement of one term of a factually identical pair > > by the other does not guarantee that the two resulting expressions > > are coreferential. > > That does not save LL, as far as I can see. We have two indiscernible things > (the event of Superman flying and the event of Clark Kent flying) that are not > identical.

But they are discernible; that is the whole point. Lois Lane has no trouble discerning them (though, admittedly, she has never seen one of them). It is only our accidental knowledge of a factual identity that even tempts us to think they are the same. Well, that and the fact that because of it, they refer to the same actual occurrence in this world. But that doesn't hold generally, so the dientity is not there. > > > The point is that abstractions > > always abstract (duh!) from the real situation (where the identity holds) > > to an alternate world (so to speak)where the factual identity cannot be > > assumed. This does not mean that the two event description do not > > represent the same event or equivalent ones with repect to some intensional > > act; it just means that we don't get from one to the other by LL, but rather > > by some other — possibly valid, hopefully at least correct in thecase -- > > means. Remember that (almost) the whole point of using abstractions is > > to save logical rules and it does eem to work. > > Would you say that LL is a logical rule, or an ontological one?

It is a logical rule; that is why preserving it is important for a logical language. I am not quite sure what an ontological rule is, so I don't see the relevance of the question. As always, the problem comes from the subjunctive nature of some cases, which is assumed for abstractions.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 25 of Aug., 2006 14:03 GMT On 8/24/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > On 8/24/06, John E Clifford wrote: > > > --- Jorge Llambas wrote: > > > > On 8/24/06, John E Clifford wrote: > > > > > I suspect > > > > > that this difference is why people are uncomfortable with saying all > > > > > events are available. > > > > > > > > The funny thing is that in order to even say that and make sense, they > > > > need them to be available. > > > > > > Well, I don't buy that, since this is just a case of predication, > > > not reference (or at least may be). > > > > Even so, if this makes them uncomfortable: > > > > ro da zo'u ganai da faursucta gi da selsnika'e > > "For every x, if x is an event abstraction, x is referrable." > > > > then they would want to deny it: > > > > su'o da zo'u ge da faursucta ginai da selsnika'e > > "There is some x such that x is an event abstraction and x is not referrable." > > > > That someone would feel more comfortable with that strikes me as funny. > > Wel, as you have noted, the rule is metalinguistic.

Right, so these people are not uncomfortable in having all events available for them to refer to in the metalanguage. They are uncomfortable with the availability of all events to someone using the object language. Still funny.

> Of your odd cases (by you), the first is a > tautology, since ro da selsnika'e. The other is therefore a contradiction. > since the first is a tautology, I can't imagine that anyone is uncomfortable > with it. The first consequently does not say what the axiom says — since > it is true even if there are no events in the domain.

Right, the maxim is not about what's generally available the metalinguist, but only about what's available to the users of the object language. These people feel uncomfortable in letting the users of the object language having too many things at their disposal.

> > > > Do you take {lu mi klama le zarci li'u} to refer to a sentence type, a sentence > > > > token, or either depending on the context? > > > > > > It has to be the last, although Lojban does not have a good way to make > > > the distinction. Of course, if you use some sentence type, what you use > > > is a token, so saying that sentences are > > > available covers both cases. > > > > And if that is good enough for sentences, why should it not be good enough > > for dogs? > > Because dog types are not dogs, but sentence types are sentences.

Why not allow dog types to be dogs, if you allow sentence types to be sentences? That doesn't seem logical.


> > It seems to me that extending the rule that you already allow for > > sentences and events to dogs would make things flow even more > > smoothly. No logical feature is broken with this extension, since there > > is nothing in Logic that distinguishes dogs from sentences or events. > > > Well, some distinctions have been noted. We are sure we can get a > sentence (hence a proposition, property and so on) when we need one, > even if we don't know exactly what one it is. There is no > such guarantee with dog.

That's not a rule from logic, it's an external stipulation that you impose.

> Even if you add every possible dog, it is not guranteed that the one > someone wants is in there, never mind the possibility of getting someone > to agree to the extension.

Same for events. If it's possible to want an impossible dog, it's equally possible to want an impossible event.

> I suppose the axiom comes down to: "The other cannot refuse to extend > the domain by a sentence."

But what's the rationale for introducing such an axiom? Or rather, if you do introduce it for sentences, why can't it be generalized to every kind of thing?


> > > > > There is no dog that subsumes > > > > > other dogs; each dog is a unique individual without reference or implication > > > > > to other dogs. > > > > > > > > Whereas there is some event of "Fido runs" that subsumes other events > > > > of "Fido runs"? > > > > > > Yes: "Fido runs" subsumes "Fido runs rapidly," "Fido runs slowly," > > > "Fido is running now." and so on indefinitely. Whenever any of the > > > subsumed events occurs, so does the subsuming. > > > > So when I see that Fido runs rapidly, and when I see that Fido runs > > slowly, I'm seeing the subsuming event that Fido runs. > > Well, now you are getting into a problem area, since one sense of seeing > events probably is intensional, so that — not with these cases I think -- > someone could see a subsumed event and not the subsuming. Of course, > there is another sense of "see" where thios is not possible.

I was talking of the ordinary sense of "see", yes. If it works for events, why couldn't it work for dogs?

> > But for some reason you don't allow this for dogs, so when I see > > Fido and when I see Spot, {mi viska lo gerku} cannot describe both > > situations with {lo gerku} having the same subsuming referent in both > > cases. > > Yes, "lo gerku" covers both cases, but it is not a dog. You can find an > sentence which subsumes the tindividual sentemces, but you can't find > a dog that subsumes the individual dogs. And it is the dog that we have > to find.

If you can find an event that subsumes both events, then you can find a dog that subsumes both dogs.

Even for sentences it works that way: {mi ciska lu mi klama le zarci li'u le mi papri .i do ciska lu mi klama le zarci li'u le do papri}. We both wrote the same sentence-type, but each of us wrote his own sentence-token. Where can you find the sentence type that subsumes both sentence tokens?


> > Yes, whenever a dog runs, lo gerku runs. This is no different from saying that > > whenever an event of Fido running happens, lo nu la fidos cu bajra happens. > > Oh, how I wish I hadn't let my Latin slip; this argument is looking so > Thirteenth Century that I want to bring out the whole panoply about > proprietates terminorm. But I beeter stick to > contemporary stuff. Your comment, while true, is not relevant to the point > at hand. {lo gerko} is not the name of a dog, nor systematically a device to > refer to a dog. {lo gerku cu gerku} is > true only distributively — even when it refers to only one dog (which there is > not generally any > reasion to think it does), similarly for {lo gerku cu bajra}. On the other hand, > {la fidos cu > bajra} and {la fidos cu gerku} are personal/singular/collective predications. > "Fido" refers to a dog; {lo gerku} refers to some number of dogs. And so on. > But notice that {lo nu mi ponse lo gerku} and {lo nu mi ponse la fidos}, both > refer to events in exactly the same way even though one > subsumes the other. In the case of dogs, you don't have a thing that is both > a dog and subsumes other dogs, you only have a pair of sentences one of > which entails the other. for events you have both.

But there is nothing in Logic that leads one to make that distinction. It's just your preferred ontology that's behind it.

> > > To be sure, there is a problem, akin to token-type, for events (there is not > > > for dogs), but it does not pose any serious problems here. The two > > > espressions may be the same grammatically, but they are not semantically > > > (metaphysically?). > > > > Why are they not the same semantically or metaphysically? By decree? > > If so, what is gained by decreeing it so? > > > The only decree is in deciding that a real difference in logical performance > is wotht remarking, but that decision is probably out of our hands, given that > we are constructing a logical language.

There is nothing in logical performance that requires making dogs different from events and sentences in the relevant sense.


> > > What exactly do abstractions do here? For generalizations we have the > > > rule that every event — and thus the one desired — is always available, > > > so generalization goes through. > > > > And there does not seem to be any reason why this rule should not also > > apply to objects. > > Wello, I have given you a list of reasons why evrents are different from > ordinary things.

A list of one: events are always available for reference and ordinary things aren't. But that's not a logical reason, it's an imposed axiom.

> The fact that you don't like these differences or that you don't consider > them important does not change the fact that there are these differences > and that they are the basis for the distinction.

Is there any other reason besides the axiom that events and sentences are always available for reference while ordinary objects are not? (Forgive me if I'm asking you to repeat something you arlready said, but if there is a list of reasons, having them listed in succint form for reference would be helpful.)


> > > The saving part of this is that events differing only in replacement of one > > > object by another factually identical to it are not generally identical, > > > > That sounds wrong, at least expressed that way. I don't see how you could > > replace something by itself and make a difference. > > Identity is always about expressions, as you well know, i.e., it is a claim > of coreferntiality somewhat oddly expressed.

You seem to be saying that the expression {ko'a du ko'e} does not express what it appears to express, i.e. something about the referents of {ko'a} and {ko'e}, but rather that it expresses something about the referring terms {ko'a} and {ko'e}, namely that they are coreferential.

> So we are saying that in an > expression referring to an event, if we replace one referring expression by > another which is in fact coreferential, the result is not usually (or at least > not guaranteed to be) a reference to the same event. The factual identity > does not carry over to the event world (in the real world, it will typically refer > to the same actual event, but come at it from different angles, like the > Morning Star and the Evening star).

So typically, but not always, {ko'a du ko'e} entails {lo nu ko'a broda cu du lo nu ko'e broda}?

How do you say "there are two different things such that I saw that one chased the other" if you cannot rely on the {du} used outside of the event carrying over inside?


> > Would you say that LL is a logical rule, or an ontological one? > > It is a logical rule; that is why preserving it is important for a logical > language. I am not quite sure what an ontological rule is, so I don't > see the relevance of the question.

Would you say this article gives a fair exposition of the principle: ?

If the principle fails in the quantum domain, then either the principle is not a logical principle, or the quantum domain is illogical.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 25 of Aug., 2006 20:39 GMT posts: 2388

> On 8/24/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > On 8/24/06, John E Clifford wrote: > > > > --- Jorge Llambías wrote: > > > > > On 8/24/06, John E Clifford wrote: > > > > > > I suspect > > > > > > that this difference is why people are uncomfortable with saying all > > > > > > events are available. > > > > > > > > > > The funny thing is that in order to even say that and make sense, they > > > > > need them to be available. > > > > > > > > Well, I don't buy that, since this is just a case of predication, > > > > not reference (or at least may be). > > > > > > Even so, if this makes them uncomfortable: > > > > > > ro da zo'u ganai da faursucta gi da selsnika'e > > > "For every x, if x is an event abstraction, x is referrable." > > > > > > then they would want to deny it: > > > > > > su'o da zo'u ge da faursucta ginai da selsnika'e > > > "There is some x such that x is an event abstraction and x is not referrable." > > > > > > That someone would feel more comfortable with that strikes me as funny. > > > > Wel, as you have noted, the rule is metalinguistic. > > Right, so these people are not uncomfortable in having all events available > for them to refer to in the metalanguage. They are uncomfortable with the > availability of all events to someone using the object language. Still funny.

Sorry, I miss the joke.

> > Of your odd cases (by you), the first is a > > tautology, since ro da selsnika'e. The other is therefore a contradiction. > > since the first is a tautology, I can't imagine that anyone is uncomfortable > > with it. The first consequently does not say what the axiom says — since > > it is true even if there are no events in the domain. > > Right, the maxim is not about what's generally available the metalinguist, > but only about what's available to the users of the object language. > These people feel uncomfortable in letting the users of the object language > having too many things at their disposal.

Oddly, people usually don't seem to have a problem with having events and the like freely available — until you present them with the general principle. You can get away with whipping a proposition or an event into the domain without anything like the kind of coment you get when you try to introduce a unicorn, say — usually without any coment at all. I suppose this is due in part to the sense that events and propositions doen't commit you to as much as objects do, whether they are in the domain or not doesn't make much practical difference, and also to the fact that the device for generating these semilinguistic are clearly always available whereas there are no devices for generating concrete objects. there are problably other factors as well.

> > > > > Do you take {lu mi klama le zarci li'u} to refer to a sentence type, a sentence > > > > > token, or either depending on the context? > > > > > > > > It has to be the last, although Lojban does not have a good way to make > > > > the distinction. Of course, if you use some sentence type, what you use > > > > is a token, so saying that sentences are > > > > available covers both cases. > > > > > > And if that is good enough for sentences, why should it not be good enough > > > for dogs? > > > > Because dog types are not dogs, but sentence types are sentences. > > Why not allow dog types to be dogs, if you allow sentence types to be > sentences? That doesn't seem logical.

It is not a matter of what is allowed, but of what is the case. A dog type is an abstraction, not something with four legs and a tail. Sentences sre already abstractions, a certain kind of structure and that structure occurs at every level. Calling cockerspanielness a dog is even more illogical. > > > > It seems to me that extending the rule that you already allow for > > > sentences and events to dogs would make things flow even more > > > smoothly. No logical feature is broken with this extension, since there > > > is nothing in Logic that distinguishes dogs from sentences or events.

Nothing in Logic, perhaps, bu8t quite a bit in fact (and, hence, I suspect quite a bit in Logic as well, but we haven't done much on the logic of dogs).

> > Well, some distinctions have been noted. We are sure we can get a > > sentence (hence a proposition, property and so on) when we need one, > > even if we don't know exactly what one it is. There is no > > such guarantee with dog. > > That's not a rule from logic, it's an external stipulation that you impose.

I didn't impose it (nor claim that it was from Logic), it is just the way things are. Do you know a device that will guarantee that you get any dog, even if you don't yet know anything about it?

> > Even if you add every possible dog, it is not guranteed that the one > > someone wants is in there, never mind the possibility of getting someone > > to agree to the extension. > > Same for events. If it's possible to want an impossible dog, it's equally > possible to want an impossible event.

But there is nothing to prevent the generating device from producing an impossible event. There is no device for generating any dog.

> > I suppose the axiom comes down to: "The other cannot refuse to extend > > the domain by a sentence." > > But what's the rationale for introducing such an axiom? Or rather, if you > do introduce it for sentences, why can't it be generalized to every kind > of thing?

Because abstractions are different in a relevant way (generating device, for example — but see the others mentioned) from other sorts of things. And, of course, adding it for other sorts of things would not solve problems and would introduce new ones, while introducing it for abstractions resolves old problem and introduces no new ones. > > > > > > > There is no dog that subsumes > > > > > > other dogs; each dog is a unique individual without reference or implication > > > > > > to other dogs. > > > > > > > > > > Whereas there is some event of "Fido runs" that subsumes other events > > > > > of "Fido runs"? > > > > > > > > Yes: "Fido runs" subsumes "Fido runs rapidly," "Fido runs slowly," > > > > "Fido is running now." and so on indefinitely. Whenever any of the > > > > subsumed events occurs, so does the subsuming. > > > > > > So when I see that Fido runs rapidly, and when I see that Fido runs > > > slowly, I'm seeing the subsuming event that Fido runs. > > > > Well, now you are getting into a problem area, since one sense of seeing > > events probably is intensional, so that — not with these cases I think -- > > someone could see a subsumed event and not the subsuming. Of course, > > there is another sense of "see" where this is not possible. > > I was talking of the ordinary sense of "see", yes. If it works for events, why > couldn't it work for dogs?

What dog is it that I see when I see Fido and then spot and then Rex? It is none of those and not any other either.

> > > But for some reason you don't allow this for dogs, so when I see > > > Fido and when I see Spot, {mi viska lo gerku} cannot describe both > > > situations with {lo gerku} having the same subsuming referent in both > > > cases. > > > > Yes, "lo gerku" covers both cases, but it is not a dog. You can find an > > sentence which subsumes the individual sentences, but you can't find > > a dog that subsumes the individual dogs. And it is the dog that we have > > to find. > > If you can find an event that subsumes both events, then you can find a > dog that subsumes both dogs.

I have been wondering for some time whether you were just obtuse or were deliberately sophitical. With this line, which would be strange ina nay case and too strange for someone who just didfn't understand what is going on, I have to go for sophistical. That aside, I will again ask you what dog it is. You can cover both events by saying "a dog" but that is not the name of a dog, it is just an indefinte way to refer to dogs, and being indefinite can refer to one dog at one time and another at another. To put it another way, "a dog" is predicated of only distributively, not personally.

> Even for sentences it works that way: {mi ciska lu mi klama le zarci li'u > le mi papri .i do ciska lu mi klama le zarci li'u le do papri}. We both wrote > the same sentence-type, but each of us wrote his own sentence-token. > Where can you find the sentence type that subsumes both sentence > tokens?

Hey, you just gave it: {mi klama le zarci} taken as referring to a type. Doing token-type overtly is a bit of a pain, since someone who like to prolong arguments can say "Oh, that's just another token" to which the reply always comes down to "No, it is a template which gives the pattern for all tokens of this type (under sutitable rules)."

> > > > Yes, whenever a dog runs, lo gerku runs. This is no different from saying that > > > whenever an event of Fido running happens, lo nu la fidos cu bajra happens. > > > > Oh, how I wish I hadn't let my Latin slip; this argument is looking so > > Thirteenth Century that I want to bring out the whole panoply about > > proprietates terminorm. But I beeter stick to > > contemporary stuff. Your comment, while true, is not relevant to the point > > at hand. {lo gerko} is not the name of a dog, nor systematically a device to > > refer to a dog. {lo gerku cu gerku} is > > true only distributively — even when it refers to only one dog (which there is > > not generally any > > reasion to think it does), similarly for {lo gerku cu bajra}. On the other hand, > > {la fidos cu > > bajra} and {la fidos cu gerku} are personal/singular/collective predications. > > "Fido" refers to a dog; {lo gerku} refers to some number of dogs. And so on. > > But notice that {lo nu mi ponse lo gerku} and {lo nu mi ponse la fidos}, both > > refer to events in exactly the same way even though one > > subsumes the other. In the case of dogs, you don't have a thing that is both > > a dog and subsumes other dogs, you only have a pair of sentences one of > > which entails the other. For events you have both. > > But there is nothing in Logic that leads one to make that distinction. It's > just your preferred ontology that's behind it.

Which distinction (there are several in the previous paragraph)? None of them is about Logic per se (although several were developed by logicians to deal with logical problems). For the fundamental one about the difference between dogs and events, that is just the way things are, although I suppose we could chanse down some reasons why they are that way (the uniqueness of objects vs. the hierarchies of abtractions, say). My ontology has nothing to do with it (unless you mean the ontology forced by raw facts is somehow mine). I don't see the point of this remark, unless it is to suggest thatthings might actually be otherwise, in which case some evidence to that might be called for. But, since I take it you are not insane but are merely posing as an objecter for sophistical or neral annoyance purposes, I will not make the call.

> > > > To be sure, there is a problem, akin to token-type, for events (there is not > > > > for dogs), but it does not pose any serious problems here. The two > > > > espressions may be the same grammatically, but they are not semantically > > > > (metaphysically?). > > > > > > Why are they not the same semantically or metaphysically? By decree? > > > If so, what is gained by decreeing it so? > > > > > The only decree is in deciding that a real difference in logical performance > > is worth remarking, but that decision is probably out of our hands, given that > > we are constructing a logical language. > > There is nothing in logical performance that requires making dogs different > from events and sentences in the relevant sense.

We have just spent several rounds discussing what in logic calls for these differences (admittedly without the guarantee that it has to be this way but in the absence of any suggestion of another way to do it). Further, we have the structural differences betweeen dogs and events that permit the suggested differences in treatment. Do you have a workable alternative to propose (noting that what you jokingly — I assume — have proposed doesn't work or even move in the right direction)?

> > > > > What exactly do abstractions do here? For generalizations we have the > > > > rule that every event — and thus the one desired — is always available, > > > > so generalization goes through. > > > > > > And there does not seem to be any reason why this rule should not also > > > apply to objects. > > > > Well, I have given you a list of reasons why events are different from > > ordinary things. > > A list of one: events are always available for reference and ordinary things > aren't. But that's not a logical reason, it's an imposed axiom.

No, that is a move that is made possible by the other differences listed. (Not a very good rhetorical ploy that, sicne it is too eadsy to check).

> > The fact that you don't like these differences or that you don't consider > > them important does not change the fact that there are these differences > > and that they are the basis for the distinction. > > Is there any other reason besides the axiom that events and sentences are > always available for reference while ordinary objects are not? (Forgive me if > I'm asking you to repeat something you arlready said, but if there is a list > of reasons, having them listed in succint form for reference would be helpful.)

That was not the question you seemed to be asking. To this question the answers are first, that in fact they are treated this way in actual conversation and this is possible because of the ways they differ from ordinary objects (generated,abstract, hierarchical, and so on).

> > > > > The saving part of this is that events differing only in replacement of one > > > > object by another factually identical to it are not generally identical, > > > > > > That sounds wrong, at least expressed that way. I don't see how you could > > > replace something by itself and make a difference. > > > > Identity is always about expressions, as you well know, i.e., it is a claim > > of coreferntiality somewhat oddly expressed. > > You seem to be saying that the expression {ko'a du ko'e} does not express > what it appears to express, i.e. something about the referents of {ko'a} and > {ko'e}, but rather that it expresses something about the referring terms {ko'a} > and {ko'e}, namely that they are coreferential.

I am not sure what the difference is that you are trying to express here. {ko'a du ko'e} is true just in case the thing referred to by {ko'a} is the same thing as that referred to by {ko'e} or just in case ko'a is the same as ko'e. The first is a little crearer than the second and also seems tro be more an explanation that a repetition. Otherwise, what is your point. The problems with identity arise because of the terms involved and so mentioning those terms seems a relevant way to get into a discussion of those problems.

> > So we are saying that in an > > expression referring to an event, if we replace one referring expression by > > another which is in fact coreferential, the result is not usually (or at least > > not guaranteed to be) a reference to the same event. The factual identity > > does not carry over to the event world (in the real world, it will typically refer > > to the same actual event, but come at it from different angles, like the > > Morning Star and the Evening star). > > So typically, but not always, {ko'a du ko'e} entails > {lo nu ko'a broda cu du lo nu ko'e broda}?

well, there are a lot of conditions to lay down on any case here, but for now let's say that, given the fuzziness of "typically", yes.

> How do you say "there are two different things such that I saw that one chased > the other" if you cannot rely on the {du} used outside of the event carrying > over inside?

The problem is, remember, tyically moving from inside to outside. Once the two tjhings are established, they carry over to the event ("rigid designators" as Cowan noted — though it is not always easy to tell when a expression is one). Here, fo course, the fact that we are dealing with seein also helps. > > > > Would you say that LL is a logical rule, or an ontological one? > > > > It is a logical rule; that is why preserving it is important for a logical > > language. I am not quite sure what an ontological rule is, so I don't > > see the relevance of the question. > > Would you say this article gives a fair exposition of the principle: > ? > > If the principle fails in the quantum domain, then either the principle is not > a logical principle, or the quantum domain is illogical. But this is about the Identity of Indiscenable, not Leibniz's Law (I wondered about the relevance of your comment about two indiscernaible things that were not identical, but, since the example given wasn't of things of that sort, I didn't ask about it). Leibniz's Law is the Principle of the Indiscernibility of Identicals (given identicals, they have all properties the same), the other one is a bit more metaphysical, though it is a theorem of many systems (set theory, for example) where the parameters — what counts as a property, say — are restricted. There is also a gneral "proof": if two things are indiscernible then if one has a property the other does,also but x (say) has the proeprty of being identical with x, so y (which is indiscernible from x) must be identical with x as well. Everybody senses this proof is fallacious somehow but noone has a really convincing case against it. Leibniz has it easy; for him "is identical with x" is not a property.

Back to a question from a while back about when factual identities work and whne they don't. Suppose we have the woman wanting a book and we say "She wants the book on the table." When is this an accurate report of her situation. 1. She say "I want the book on the table," being basically indifferent to what the book is but using this as the way to get a book and content to abiode with whatever it brings her. Clearly our report is accurate in this case, both sides of the definition work out. 2. she says the above thinking that the book on the table is War and Peace, which is what she wants, but the book is in fact Alice In Wonderland. Clearly we have got it wrong since the definition works neither way. But it is her fault, so we can't be condemned. 3. She say "I want War and Peace," meaning any copy of that text, and the book on the table is in fact a copy of War and Peace and we say "She want the book on the table" meaning that particular copy. We have clearly misrepresented her, since, while if she should get that book shwe would be satisfied, she might bre satisfied without getting that particular book. 4. As 3 but now we mean "a token of the book type of which the book on the table is a token." Now we have it right, I think. 5. She means her autographed copy of W&P and that is the book on the table. Again, we have it right I think. 6. The book on the table is not her copy of W&P but is a copy of that. In either sense, this is wrong. Somehow the notion of generating a rigid designator (both she and we)seems involved here but the exact rule eludes me. In the meantime, we can fall back on the defintion to see whaat work.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 25 of Aug., 2006 22:19 GMT On 8/25/06, John E Clifford wrote: > --- Jorge Llambías wrote: > > > Right, so these people are not uncomfortable in having all events available > > for them to refer to in the metalanguage. They are uncomfortable with the > > availability of all events to someone using the object language. Still funny. > > Sorry, I miss the joke.

The object language is Lojban, the metalanguage we are using to discuss it is English. These people have a problem with Lojbanists having events at their disposal for reference, but they themselves feel no discomfort in making reference to events in order to set up the metalinguistic axiom, and presumably they take for granted that the events are at their disposal in the language they are using. So someone investigating this behaviour from a meta-meta-linguistic perspective might find it funny. I do anyway.


> > Right, the maxim is not about what's generally available the metalinguist, > > but only about what's available to the users of the object language. > > These people feel uncomfortable in letting the users of the object language > > having too many things at their disposal. > > Oddly, people usually don't seem to have a problem with having events > and the like freely available — until you present them with the general > principle.

Odd to you, funny to me. More or less the same thing.

> You can get away with whipping a > proposition or an event into the domain without anything like the kind of > coment you get when you > try to introduce a unicorn, say — usually without any coment at all.

What kind of comment do you get when you try to introduce a unicorn (in a relevant context)? I really don't see why trying to introduce a fictional creature such as unicorns in a relevant way would raise more eyebrows than trying to introduce a fictional event in a relevant way. It seems to me that both can be done with equal ease. If you claim to have seen a unicorn that will raise some comment, but the same will happen if you claim to have seen a unicorn crossing the street. And in a context in which a unicorn crossing the street will go unchallenged, so will a unicorn.

> > Why not allow dog types to be dogs, if you allow sentence types to be > > sentences? That doesn't seem logical. > > It is not a matter of what is allowed, but of what is the case. A dog type > is an abstraction, not something with four legs and a tail.

If a dog type does not have four legs and a tail, then it is not much of a dog type. Just as if a sentence type does not have words it is not much of a sentence type.


> > > > It seems to me that extending the rule that you already allow for > > > > sentences and events to dogs would make things flow even more > > > > smoothly. No logical feature is broken with this extension, since there > > > > is nothing in Logic that distinguishes dogs from sentences or events. > > Nothing in Logic, perhaps, bu8t quite a bit in fact (and, hence, I suspect > quite a bit in Logic as well, but we haven't done much on the logic of dogs).

There are indeed quite a bit of things that distinguish dogs form sentences and events: dogs have legs and sentences don't, sentences have words and dogs don't, events have participants and dogs don't, etc. But only logical differences would warrant having different logical laws apply to them.


> > > Even if you add every possible dog, it is not guranteed that the one > > > someone wants is in there, never mind the possibility of getting someone > > > to agree to the extension. > > > > Same for events. If it's possible to want an impossible dog, it's equally > > possible to want an impossible event. > > But there is nothing to prevent the generating device from producing > an impossible event. There is no device for generating any dog.

What is the generating device that produces impossible events but not impossible dogs?


> > Would you say this article gives a fair exposition of the principle: > > ? > > > > If the principle fails in the quantum domain, then either the principle is not > > a logical principle, or the quantum domain is illogical. > But this is about the Identity of Indiscenable, not Leibniz's Law

" This is often referred to as 'Leibniz's Law' and is typically understood to mean that no two objects have exactly the same properties."

"The converse of the Principle, x=y → ∀F(Fx ↔ Fy), is called the Indiscernibility of Identicals. Sometimes the conjunction of both principles, rather than the Principle by itself, is known as Leibniz's Law."

> Leibniz's Law is the Principle of the Indiscernibility of Identicals

That's not what the cited article says, but I'll take your word for it.

> There is also a gneral "proof": if two things are indiscernible then if one > has a property the other does, also but x (say) has the proeprty of being > identical with x, so y (which is indiscernible from x) must > be identical with x as well. Everybody senses this proof is fallacious > somehow but noone has a really convincing case against it.

From what the article says, that seems to rely on the property of "thisness" or "haecceity", {lo ka ce'u du ko'a}, which not everyone accepts as a true property.

> Leibniz has it easy; for him "is identical with x" is not a property.

Did Leibniz formulate all three principles? (Identity of Indiscernibles, Indiscernibility of Identicals, and their conjunction?)

This article seems relevantly interesting too:

mu'o mi'e xorxes

Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Sat 26 of Aug., 2006 14:36 GMT posts: 2388

> On 8/25/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > Right, so these people are not uncomfortable in having all events available > > > for them to refer to in the metalanguage. They are uncomfortable with the > > > availability of all events to someone using the object language. Still funny. > > > > Sorry, I miss the joke. > > The object language is Lojban, the metalanguage we are using to discuss > it is English. These people have a problem with Lojbanists having events at > their disposal for reference, but they themselves feel no discomfort in > making reference to events in order to set up the metalinguistic axiom, > and presumably they take for granted that the events are at their disposal in > the language they are using. So someone investigating this behaviour from > a meta-meta-linguistic perspective might find it funny. I do anyway.

Ah! Does this mean that you have no problem with events being freely available in Lojban? Or does this just mean that you have no problem with absolutely anything being freely available. It appears that by your lights, any time we comment on the availability of any sort of thing, we are making use of that thing in the language we are using. Thus, the situation with events would not be peculiar to you and so the advantage of having it is lost. Unfortunately, having the other things to enter at will does not solve problems in the way that having events does.

> > > Right, the maxim is not about what's generally available the metalinguist, > > > but only about what's available to the users of the object language. > > > These people feel uncomfortable in letting the users of the object language > > > having too many things at their disposal. > > > > Oddly, people usually don't seem to have a problem with having events > > and the like freely available — until you present them with the general > > principle. > > Odd to you, funny to me. More or less the same thing. > > > You can get away with whipping a > > proposition or an event into the domain without anything like the kind of > > comment you get when you > > try to introduce a unicorn, say — usually without any coment at all. > > What kind of comment do you get when you try to introduce a unicorn > (in a relevant context)? I really don't see why trying to introduce a fictional > creature such as unicorns in a relevant way would raise more eyebrows > than trying to introduce a fictional event in a relevant way. It seems to me > that both can be done with equal ease. If you claim to have seen a unicorn > that will raise some comment, but the same will happen if you claim to have > seen a unicorn crossing the street. And in a context in which a unicorn > crossing the street will go unchallenged, so will a unicorn.

I suppose the point is that events are never fictional (well, they don't exist, perhaps, but they always are). But if you talk about seeing that raises another issue, since things seen (in the usual sense) have to exist. Claiming that something exists is rather more venturesome than just using them as values for variables. In this latter sue, people seem to have no problem with events and the like, but do with unicorns. It might be, of course, that there is some notion of existence in the back of all that, that is, that events exist even when they don't occur (and properties even when they have no bearers and so on)(and then it is occurring that seeing requires of events seen). Given that people do not do well at the the distinction between being and existing, it is hard to decide just where the differential behaviors come from. The axiom apparently reflects merely the fact of the phenomenon, without pursuing its ground. Notice that the cases we are interessted in do not entail existence (or occurrence) and, indeed, often entail the opposite.

> > > Why not allow dog types to be dogs, if you allow sentence types to be > > > sentences? That doesn't seem logical. > > > > It is not a matter of what is allowed, but of what is the case. A dog type > > is an abstraction, not something with four legs and a tail. > > If a dog type does not have four legs and a tail, then it is not much > of a dog type. Just as if a sentence type does not have words it is > not much of a sentence type.

Well, a dog type has four legs and a tail in a very different way than a dog (token, if you will) does. And, indeed, the legs and tail it has are aslos very different, being, I suppose, leg types and tail type. The basic problem is that a dog type is an abstraction, a dog is not. A sentence type (containing word types) is an abstraction, too, but then so is a sentence token in the relevant sense, If we do call something written out or spoken(in one place at one time, and so on) a sentence token (though it can be argued that it is just a way of presenting a sentence, not a sentence itself), then the notion of sentence at issue is relatively a type. As I have said, if you want to referent of {lo gerku} to be a dog type rather than dogs, a consistent program could be worked outr. But 1) you have in the past said things that imply that that is not what you mean and that you mean for {lo gerku} to refer just to dogs and 2) this approach wouldf be very unLojbanic, shifting from concreta to abstracta without a mark to show it and brining abstracta in in exact paralellism with expressions that clearly are about concreta and are said to be marked off only by specificity (I don't hink you can hold that lack of specificity is just necessarily abstractness).


> > > > > It seems to me that extending the rule that you already allow for > > > > > sentences and events to dogs would make things flow even more > > > > > smoothly. No logical feature is broken with this extension, since there > > > > > is nothing in Logic that distinguishes dogs from sentences or events. > > > > Nothing in Logic, perhaps, bu8t quite a bit in fact (and, hence, I suspect > > quite a bit in Logic as well, but we haven't done much on the logic of dogs). > > There are indeed quite a bit of things that distinguish dogs from sentences > and events: dogs have legs and sentences don't, sentences have > words and dogs don't, events have participants and dogs don't, etc. But > only logical differences would warrant having different logical laws apply > to them.

I don't recall suggesting that different logical laws apply. In fact, the point of all of this is to find a way where the usual logical laws apply in places where they do not under one way of doing things. If there were genuinely different logics then the fact that some laws don't apply would be understandable (well, given the generality of the laws involved, maybe not, but at least there would be some groundwork for making the case for their not applying).

> > > > > Even if you add every possible dog, it is not guranteed that the one > > > > someone wants is in there, never mind the possibility of getting someone > > > > to agree to the extension. > > > > > > Same for events. If it's possible to want an impossible dog, it's equally > > > possible to want an impossible event. > > > > But there is nothing to prevent the generating device from producing > > an impossible event. There is no device for generating any dog. > > What is the generating device that produces impossible events but > not impossible dogs?

Well, we begin with the device for generating sentences and then take the generic event creater from sentence "that.... is true." It is not problem at all to generate sentences that are contradictory or otherwise necessarily false, so ... . There is no way to generate dogs at all. You can dream one up and add it (given cooperation) to the domain, but you can't be sure you get them all in — and particularly that you get the one you want in.

But,as noted before, a lot of this is beside the point at issue. There is no adding of dogs and the like involved. Sentences of the sort we are concerned with are evaluated in the basic domain, not in some extended one. What extension there is occurs only in the subjunctive domains, which disappear eimmediately the sentence is evaluated. to be sure, we can — after the evaluation -- bring some of those things into the basic domain, we can talk about the dog I want once it is established that I want a dog, but that dog play no role in the basic domain in the evaluation of the sentence. The difference involved here is that, while that dog may already be in the basic domain, there is no guarantee, while there is one for events. I don't see any way to guarantee that the dog you want is one already in the domain (indeed, I know it often is not and so would reject any device that seemed to do that), so the most you can do here is attack the notion that events are always in the domain. Your attack seems to be that, if we allow events we should allow dogs. Then, since I clearly reject allowing dogs, I shouldn't allow events either (this is not obviously the way you are arguing — I think you are actually arguing that we should dogs). In any case, I reject the connection: events seem to me different enough from dogs that we could (as we indeed do) allow events to be always available while not extending the same courtesy to dogs (or any other concreta) but to all abstracta. Notice, this is a very different issue (though I think we have occasionally mixed them up) from what we can ADD to a domain at a given point. To be sure, one consequent of the present situation is that we never have to add an event, only call attention to it, but we do have to add a concretum.

> > > > Would you say this article gives a fair exposition of the principle: > > > ? > > > > > > If the principle fails in the quantum domain, then either the principle is not > > > a logical principle, or the quantum domain is illogical. > > But this is about the Identity of Indiscenable, not Leibniz's Law > > " This is often referred to as 'Leibniz's Law' and is typically understood to > mean that no two objects have exactly the same properties." > > "The converse of the Principle, x=y → ∀F(Fx ↔ Fy), is called the > Indiscernibility > of Identicals. Sometimes the conjunction of both principles, rather than > the Principle by itself, is known as Leibniz's Law." > > > Leibniz's Law is the Principle of the Indiscernibility of Identicals > > That's not what the cited article says, but I'll take your word for it.

Sorry; I was using it in the sense that I usually use it in, as the name of a rule of inference, like Modus Ponendo Ponens, say. the the looser philosophical usage, the name can be applied to either and is probably more generaly used for the first (although calling that a law seems stretching it a bit — I like "Principle" better).

> > There is also a general "proof": if two things are indiscernible then if one > > has a property the other does, also; but x (say) has the proeprty of being > > identical with x, so y (which is indiscernible from x) must > > be identical with x as well. Everybody senses this proof is fallacious > > somehow but no one has a really convincing case against it. > > From what the article says, that seems to rely on the property of > "thisness" or "haecceity", {lo ka ce'u du ko'a}, which not everyone > accepts as a true property.

One reading of haeceity is just the the open identity formula, as you note. Another has to do with individual concepts, which even more people have trouble with.

> > Leibniz has it easy; for him "is identical with x" is not a property. > > Did Leibniz formulate all three principles? (Identity of Indiscernibles, > Indiscernibility of Identicals, and their conjunction?)

Well, he did state both separately; I am not sure he ever brought the two together explicitly.

> This article seems relevantly interesting too: >



Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 30 of Aug., 2006 18:19 GMT On 8/26/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > What is the generating device that produces impossible events but > > not impossible dogs? > > Well, we begin with the device for generating sentences and then > take the generic event creater from sentence "that.... is true." It is not > problem at all to generate sentences that are contradictory or otherwise > necessarily false, so ... . There is no way to generate dogs at all.

I'm not sure what you mean by "the generic event creater".

It seems to me that in the weak sense that events can be created from expressions, then dogs can be created from expressions as well. In the strong sense that dogs cannot be created from expressions, neither can events. I don't see any intermediate sense in which events, but not dogs, can be created from expressions.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Wed 30 of Aug., 2006 22:18 GMT posts: 2388

> On 8/26/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > What is the generating device that produces impossible events but > > > not impossible dogs? > > > > Well, we begin with the device for generating sentences and then > > take the generic event creater from sentence "that.... is true." It is not > > problem at all to generate sentences that are contradictory or otherwise > > necessarily false, so ... . There is no way to generate dogs at all. > > I'm not sure what you mean by "the generic event creater".

Just that: S's being true is an event for every sentence S. It is arguable that every event is, in fact, just one of these. What (other dogs aside) creates dogs?

> It seems to me that in the weak sense that events can be created from > expressions, then dogs can be created from expressions as well. In the > strong sense that dogs cannot be created from expressions, neither can > events. I don't see any intermediate sense in which events, but not dogs, > can be created from expressions.

I am not sure just what you mean by the strong and the weak senses. I can think of one possibility: in the weak sense, what is created is not events but descriptions of events and that is all this proposal does. It is equally easy to create descriptions of dogs. In the strong sense, what is created is actual events, which seem no easier to create than actual dogs. If this is what you mean, I think the situation is somewhere between. Events are not created, they just are (that axiom again, but also our linguistic practice in any language *that I know of). What we need in the present situation is to draw these events into our discussion *their bare existence does very little if we get beyond quantification, as we seem to do in most interesting cases. And we can always summon the one we want because we can always find the right sentence ("can" is a very generous sense — we may, in fact, always screw up). Dogs don't always exist so we cannot be sure that we can — in a given context — find the description that will summon one to our need. We can come up with a description of a dog, but no dog need fit that description.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 31 of Aug., 2006 19:20 GMT On 8/30/06, John E Clifford wrote: > Events are not created, they just are (that axiom again, but also our > linguistic practice in any language *that I know of).

In no language that I know of there seems to be such a distinction made between events and objects.

As for the metalinguistic axiom, in order to express it in English (about Lojban) you need it to be inoperative in English. You need to be able to refer, in English, to certain things that are not available for reference in Lojban. It is not clear why we would want to burden Lojban with this. It would mean Lojban cannot be used as the metalanguage to explain Lojban, because you would need to violate the axiom of Lojban in order to state the axiom of Lojban in Lojban.

> What we > need in the present situation is to draw these events into our discussion > *their bare existence does very little if we get beyond quantification, as > we seem to do in most interesting cases. And we can always summon > the one we want because we can always find the right sentence ("can" is a > very generous sense — we may, in fact, always screw up). Dogs don't > always exist so we cannot be sure that we can — in a given context — find > the description that will summon one to our need. We can come up with > a description of a dog, but no dog need fit that description.

Why would it be more difficult to find an expression to summon a dog than one to summon an event? "The dog I want" seems like the perfect expression to summon the dog I want. That doesn't of course create the dog as existing in the real world, but neither does "the event I want" create an event as existing in the real world, in both cases the expression summons the required referent into the domain of discourse.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 31 of Aug., 2006 20:01 GMT Let's consider the sentence {mi djica lo xanto}. (I'm tired of dogs, so I'm switching to elephants.)

How could we interpret it? Possible candidates:

(1) It is uninterpretable, or else trivially false, because the gi'uste says {djica} takes an event in x2 and elephants are not events.

(2) It means something similar to {mi djica lo nu lo xanto cu zasti}, so it expresses a wish that is trivially satisfied because elephants do exist.

(3) It means something similar to {mi djica lo nu da (cavi) xanto}, "I want that there be an elephant (around here now)", which is a wish that may or may not be eventually satisfied.

(4) Something else.

>From the point of view of usefulness, (1) is utterly useless, (2) is almost useless (we normally don't express wants that are trivially satisfied), and (3) is the useful one.

So what about from the point of view of "Logic", consistency, etc. Is there any problem with {mi djica lo xanto} having an interpretation along the lines of (3)? (Notice that I'm not saying this is an exact paraphrase, just an approximation that gives the general idea of what it means to put an object in the x2 of {djica} in terms of what it means with an event there.)

The objections to having an object in x2 of {djica} are that this somehow causes problems with the axiom of existential generalization or with the intersubstitution of identicals. But neither of these axioms really cause any problems.

>From {mi djica lo xanto} it follows that {mi djica su'o da}, i.e. there is something that I want. It even follows that {mi djica su'o da noi xanto}, "I want something, which happens to be an elephant" and even {mi djica su'o da poi xanto}, "I want something that is an elephant", (although these last two are not derived from existential generalization only). This does not require that the thing I want exists, or that it is a certain one among a group of more than one elephants. There is no need for {lo xanto} to have more than a single referent in the current domain of discourse, and this is no different from what happens with events. If I say {mi djica lo nu da xanto} I'm not saying that among many events of there being an elephant, there is a certain one such that that very one is the one I happen to want.

As for intersubstitution of identicals, let's first consider the case of events:

la lois djica lo nu la superman ly cinba "Lois wants Superman kissing her."

la lois na djica lo nu la klark kent ly cinba "Lois does not want Clark Kent kissing her."

Let's say both those sentences are true. That means {lo nu la superman ly cinba} and {lo nu la klark kent ly cinba} refer to different events. But what happens if Superman/Clark does kiss Lois? Then:

lo nu la superman ly cinba ca'a fasnu Superman kissing her does happen.

lo nu la klark kent ly cinba ca'a fasnu Clark Kent kissing her does happen.

both seem to be true. And both {lo nu la superman ly cinba} and {lo nu la klark kent ly cinba} would seem to refer to the same occurrence. (There weren't two separate kissing occurrences.)

So {lo nu la superman ly cinba} and {lo nu la klark kent ly cinba} are different events but the same occurrence. Is there a problem with this? No, as long as we don't take the two expressions to refer to the same logical object.

And what about objects now?

la superman cu cinba la lois

la klark kent cu cinba la lois

Buth would seem to be true together, because Superman and Clark Kent are the same person. But:

la lois cu djica la superman

la lois na djica la klark kent

would also seem to be both true. How come, if Superman and Clark are the same person? That's because Superman and Clark are different personas. Just as one same occurrence can subsume two different events, one same person can subsume two different personas. As long as {la superman} and {la klark kent} refer to different logical objects, there is no problem in {la lois cu djica la superman} and {la lois cu djica la klark kent} having different truth values. That depends on the interpretation of the referring terms, but it is no different for object than for events.

See for a nice article on relative identity, which nicely dissolves all the problems with identity conundrums.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 01 of Sep., 2006 00:39 GMT posts: 2388

> On 8/30/06, John E Clifford wrote: > > Events are not created, they just are (that axiom again, but also our > > linguistic practice in any language *that I know of). > > In no language that I know of there seems to be such a distinction made > between events and objects.

Ther clearly is a difference if the domain is the "real world"one. Most people who I have seen writing about it seem to think that the distinction is there also for any domain. That is, we can call attention and quantify over abstract objects without changing the domain. So long as it holds in one world, then devices that depend upon it not holding are invalid, even if usually effective. I am not hung up on this, since, as noted before, it does not touch on the issue actually at hand, namely opaque contexts. There since the evaluation has to be before the thing is mentioned (or, rather, in a different world than the one referred to) it is enough that events are always there even without being added.

> As for the metalinguistic axiom, in order to express it in English (about > Lojban) you need it to be inoperative in English. You need to be able to refer, > in English, to certain things that are not available for reference in Lojban. > It is not clear why we would want to burden Lojban with this. It would mean > Lojban cannot be used as the metalanguage to explain Lojban, because > you would need to violate the axiom of Lojban in order to state the axiom > of Lojban in Lojban.

I don't follow this at all. Why must the axiom not apply to English s well as Lojban so that it can be expressed for Lojban? The axiom, after all, says that abstracts are available, not that they are not available. That dogs are not similarly available is clearly the case in at least one domain, which is enough for the invalidity claim. I happen to think (with others) that this applies rather generally, but I can see how you might take the claimed decisive cases as not proving the point.

> > What we > > need in the present situation is to draw these events into our discussion > > *their bare existence does very little if we get beyond quantification, as > > we seem to do in most interesting cases. And we can always summon > > the one we want because we can always find the right sentence ("can" is a > > very generous sense — we may, in fact, always screw up). Dogs don't > > always exist so we cannot be sure that we can — in a given context — find > > the description that will summon one to our need. We can come up with > > a description of a dog, but no dog need fit that description. > > Why would it be more difficult to find an expression to summon a dog > than one to summon an event? "The dog I want" seems like the perfect > expression to summon the dog I want. That doesn't of course create the > dog as existing in the real world, but neither does "the event I want" create > an event as existing in the real world, in both cases the expression > summons the required referent into the domain of discourse.

The point is that the summoning of an event always gets what it asks for, the summoning of a dog may not. The dog fails in at least the "real world" domain, i.e., one where everything in the domains must exist. Of course, this assumes that an event existing is different from it occurring, but no similar dodge seems to work for dogs.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 01 of Sep., 2006 15:26 GMT On 8/31/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > On 8/30/06, John E Clifford wrote: > > > Events are not created, they just are (that axiom again, but also our > > > linguistic practice in any language *that I know of). > > > > In no language that I know of there seems to be such a distinction made > > between events and objects. > > Ther clearly is a difference if the domain is the "real world"one. Most people > who I have seen writing about it seem to think that the distinction is there > also for any domain.

Would you say this gives a fair account of the general understanding of what events are?

> That is, we can > call attention and quantify over abstract objects without changing the > domain.

When you say events are abstract objects, you mean this from a metalinguistic perspective, right? Someone who sees a dog running would not claim that they are seeing an abstract object. Someone who is hit by a brick falling on their foot would not say that some abstraction caused their pain. The dog running and the brick falling on the foot are very concrete events. Even when Lois says she wants Superman kissing her, she would not say she wants some abstraction, she would say that she wants something concrete. In the metalinguistic sense that events can be said to be abstractions, anything can be said to be an abstraction.


> > As for the metalinguistic axiom, in order to express it in English (about > > Lojban) you need it to be inoperative in English. You need to be able to refer, > > in English, to certain things that are not available for reference in Lojban. > > It is not clear why we would want to burden Lojban with this. It would mean > > Lojban cannot be used as the metalanguage to explain Lojban, because > > you would need to violate the axiom of Lojban in order to state the axiom > > of Lojban in Lojban. > > I don't follow this at all. Why must the axiom not apply to English s well > as Lojban so that it can be expressed for Lojban? The axiom, after all, says > that abstracts are available, not that they are not available.

Right, but unless the axiom is the trivial "abstracts are available simply because everything is available" then for the axiom to say something interesting there have to be other things that are not available.

> That dogs are not similarly available is clearly the case in at least one > domain, which is enough for the invalidity claim.

If there are domains where dogs are not available (say in a context where variables can only take numbers as values) then clearly there are also domains where events are not available. This does not show a distinction between dogs and events.


> The point is that the summoning of an event always gets what it asks > for, the summoning of a dog may not. The dog fails in at least the > "real world" domain, i.e., one where everything in the domains must > exist. Of course, this assumes that an event existing is different from > it occurring, but no similar dodge seems to work for dogs.

The summoning of an event fails to get what it asks for in the "number" domain, i.e. one where everything in the domain must be a number. Also it does not get what it wants in the "animal" domain, i.e. one where everything must be an animal. This does not point to a distinction between dogs and events.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by JohnCowan on Fri 01 of Sep., 2006 16:18 GMT posts: 149 Jorge Llamb?as scripsit:

> When you say events are abstract objects, you mean this from a > Someone who sees a dog running > would not claim that they are seeing an abstract object. Someone > who is hit by a brick falling on their foot would not say that some > abstraction caused their pain. The dog running and the brick falling > on the foot are very concrete events.

No, I think natlang syntax is betraying you. I don't see the event of a dog running (lo nu lo gerku cu bajra); what I see is a dog, which is at present running (lo gerku noi bajra). Ditto for a "brick falling". OTOH I can think about (the event of) a dog running.

> Even when Lois says she wants > Superman kissing her, she would not say she wants some > abstraction, she would say that she wants something concrete.

I do say she wants an abstraction, and that all objects of desire are abstractions, though they may be abstractions created from concrete things. An event is a state of affairs, and what we desire is that a certain hypothetical state of affairs become actual.

-- On the Semantic Web, it's too hard to prove John Cowan cowan@ccil.org you're not a dog. --Bill de hOra http://www.ccil.org/~cowan


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 01 of Sep., 2006 16:23 GMT On 9/1/06, John Cowan wrote:

> No, I think natlang syntax is betraying you. I don't see the event of > a dog running (lo nu lo gerku cu bajra); what I see is a dog, which is > at present running (lo gerku noi bajra). Ditto for a "brick falling". > OTOH I can think about (the event of) a dog running.

You don't say {mi viska lo nu lo gerku cu bajra} and {lo nu lo bliku cu farlu lo mi jamfu cu xrani mi}?


> > Even when Lois says she wants > > Superman kissing her, she would not say she wants some > > abstraction, she would say that she wants something concrete. > > I do say she wants an abstraction, and that all objects of desire are > abstractions, though they may be abstractions created from concrete > things. An event is a state of affairs, and what we desire is > that a certain hypothetical state of affairs become actual.

Then what you are saying is that the x2 of {djica} has to be a {du'u}, never a {nu}.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by JohnCowan on Fri 01 of Sep., 2006 16:45 GMT posts: 149 Jorge Llamb?as scripsit:

> You don't say {mi viska lo nu lo gerku cu bajra} and > {lo nu lo bliku cu farlu lo mi jamfu cu xrani mi}?

No. Both of those sound utterly bizarre to me.

> >An event is a state of affairs, and what we desire is > >that a certain hypothetical state of affairs become actual. > > Then what you are saying is that the x2 of {djica} has to be a {du'u}, > never a {nu}.

No, that would be different. That would be saying that what we desire is for a certain proposition to be true. That's not unreasonable, just not the way we have currently defined "djica". (I wouldn't be opposed to adding "(nu/du'u)" to the gismu list entry.)

-- John Cowan http://www.ccil.org/~cowan Raffiniert ist der Herrgott, aber boshaft ist er nicht. --Albert Einstein


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 01 of Sep., 2006 16:45 GMT posts: 2388

> Let's consider the sentence {mi djica lo xanto}. (I'm tired of dogs, > so I'm switching to elephants.) > > How could we interpret it? Possible candidates: > > (1) It is uninterpretable, or else trivially false, because the gi'uste > says {djica} takes an event in x2 and elephants are not events.

I would take this to be the claim that there are some elephants you want to have (making "have" about as vague as possible). That is, it breaks out of the opaque context. I am not sure how closely that {lo xanto} has to be restricted, though there must be some restrictions on it.

> (2) It means something similar to {mi djica lo nu lo xanto cu zasti}, > so it expresses a wish that is trivially satisfied because elephants > do exist.

There seem to be cases where this is all that is meant by "I want an x." That cannot be what is meant by {mi djica x} when there are xs (as I pointed out at the beginning of all this) and so in this reading {mi djica lo x} is always false.

> (3) It means something similar to {mi djica lo nu da (cavi) xanto}, > "I want that there be an elephant (around here now)", which is a wish > that may or may not be eventually satisfied.

I wouldn't want to be that specific about it, since I may want the thing as far away as possible and I may want it in some special relationship (summed up by "I have"). Yes, that is probably what was intended by by {mi djica lo xanto}. The point is that it doesn't work but rather collpses back to my version 1. It does so because of the general rule that says {Flo xanto} entails (Ex:x elephant) Fx {da poi xanto zo'u mi djica da}. So, since what is wanted always turns out to be some vaguely specified event involving the thing mentioned, the word list recommends that we say as much overtly (Logic likes that approach). This also allows us to be as precise as we need be abvout what that event is. And it gets around the various opaque problems in every situation.

> (4) Something else. > > >From the point of view of usefulness, (1) is utterly useless, (2) is almost > useless (we normally don't express wants that are trivially satisfied), > and (3) is the useful one. > > So what about from the point of view of "Logic", consistency, etc. Is there > any problem with {mi djica lo xanto} having an interpretation along the > lines of (3)? (Notice that I'm not saying this is an exact paraphrase, just > an approximation that gives the general idea of what it means to put an > object in the x2 of {djica} in terms of what it means with an event there.)

None whatsoever except, as noted, the proclivity for having all one's cards on the table and for not violating logical laws.

> The objections to having an object in x2 of {djica} are that this somehow > causes problems with the axiom of existential generalization or with > the intersubstitution of identicals. But neither of these axioms really > cause any problems.

They certainly do. There is no unicorn you want when you truthfully say (mi djica lo pavyseljirna}-- or at least there need not be. Further, when you truthfully say that, you do not want a centaur though unicorns are exactly centaurs. This kind of error does seem to me to be a problem. Logic doesn't like inferences which are invalid yet laws of logic. The problem arise only because we do not mark the opaque places somehow — it doesn't matter how. The use of event expressions rather than object expressions in these places does the job and also increases accuracy somewhat (though {tu'a} isn't much of an improvement). If we leave just the thing expressions, we have to mark the places by lists, which is the worst way to do it (that I can think of).

> >From {mi djica lo xanto} it follows that {mi djica su'o da}, i.e. there > is something that I want. It even follows that {mi djica su'o da noi xanto}, > "I want something, which happens to be an elephant" and even > {mi djica su'o da poi xanto}, "I want something that is an elephant", > (although these last two are not derived from existential generalization > only).

Actually, none of these inferences are valid, since the want claim ight be true but each and every one of the other false. That is, if there are no elephants or none that you want in particular, there isn't something you want and certainly nothing something that is an elephant, whether incidentally or essentially. I think that you thik that saying {lo xanto} sticks an elephant into the domain, but that is simply not the case: the elephants go into some subjunctive domain, not the one we are talking in.

> This does not require that the thing I want exists, or that it is > a certain one among a group of more than one elephants. There is no > need for {lo xanto} to have more than a single referent in the current > domain of discourse, and this is no different from what happens with > events.

The point here eludes me. Yes, particular generalization holds in your wish world — it holds in every world, after all. But that has nothing to do with the issue, which is not about the wish world but about the world in which you say {mi djica ...}. That I want one does not require ANY elephants to be in the world where you want one; it says nothing about that world, except that you have a want. The elephants come into things only in evaluating the claim that you such a wish and that evaluation projects into two sets of worlds, one defined by your getting what you want and the other by having that want assuaged. If what you want is an elephant, then the first of those worlds will definitely contain elephants and, if it is true that you want an elephant, so will the second, None of this says anything about what there is in the original world, the one where you have the want (notice that you don't have it in these subjunctive worlds). So, if you now take the thing you want and project it into that original world, you have no guarantee that it will be there. It has, of course, been introduced into a domain — but not that one, only the subjunctive ones.


> If I say {mi djica lo nu da xanto} I'm not saying that among many > events of there being an elephant, there is a certain one such that > that very one is the one I happen to want.

I think we are getting muddled here. There certainly is an event that I want, namely the event of there being an elephant. Now, there are probably a lot of other events involving elephants that I want, but they are aren't mentioned. What I do not want is a particular occurrence of that event, any occurrence will do. But I have said nothing about occurrences of events here, just events tout court. Events can be individuated separately from their occurrences (and, indeed, must be since two seaparate events can have the same occurrence). Dogs are apparently inidviduated only by their occurrences. So, while I can pick out an event from the blue and know that it is available, I can't do that with a dog: "dog" refers to concreta, "event" to abstracta (whose occurrences are conreta generally).


> As for intersubstitution of identicals, let's first consider the case of events: > > la lois djica lo nu la superman ly cinba > "Lois wants Superman kissing her." > > la lois na djica lo nu la klark kent ly cinba > "Lois does not want Clark Kent kissing her." > > Let's say both those sentences are true. That means {lo nu la superman > ly cinba} and {lo nu la klark kent ly cinba} refer to different events. But > what happens if Superman/Clark does kiss Lois? Then: > > lo nu la superman ly cinba ca'a fasnu > Superman kissing her does happen. > > lo nu la klark kent ly cinba ca'a fasnu > Clark Kent kissing her does happen. > > both seem to be true. And both {lo nu la superman ly cinba} and > {lo nu la klark kent ly cinba} would seem to refer to the same > occurrence. (There weren't two separate kissing occurrences. > So {lo nu la superman ly cinba} and {lo nu la klark kent ly cinba} > are different events but the same occurrence. Is there a problem > with this? No, as long as we don't take the two expressions to > refer to the same logical object.

Yes, this all works fine. What is a logical object? The event? But we have already said (and demonstrated) that the two events are different.

> And what about objects now? > > la superman cu cinba la lois > > la klark kent cu cinba la lois > > Both would seem to be true together, because Superman and Clark Kent > are the same person. But: > > la lois cu djica la superman > > la lois na djica la klark kent > > would also seem to be both true. How come, if Superman and Clark are > the same person? That's because Superman and Clark are different personas. > Just as one same occurrence can subsume two different events, one same > person can subsume two different personas. As long as {la superman} and > {la klark kent} refer to different logical objects, there is no problem in > {la lois cu djica la superman} and {la lois cu djica la klark kent} having > different truth values.

What are the logical objects here? What you call personas appear to be what would standardly be called senses or individual concepts, intensional entities both. The problem now comes down to this. {la lois djica la superman} has (apparently) the same logical form as {la lois viska la superman}. In the latter, {la superman} refers to the thing Superman, a gross physical object in this case. In the former, for everything to work out right, {la superman} has to refer to ^Superman, the individual concept unique to Superman. How do we know when the difference occurs? We can memorize (or keep a list) of all the places where the reference of a rweferring phrase is its sense rather than its referent. Or we can arrange things so that the shift doesn't happen at all. Or we can have a mark that warns of the shift. Clearly, one of the latter two approaches is better than the first. And they actually come very close to being the same thing in practice: in Lojban, we write {tu'a lo xanto} rather than {lo xanto} in the problematic places, so they are marked. But technically, what we have done is always put there a term that refers to an intensional object, which is ipso facto unaffected by merely factual identities and which always is available.

That depends on the interpretation of the > referring terms, > but it is no different for object than for events.

Except, of course, that in one case we overtly refer to an intensional object, in the other we do so only covertly and wihtout warning (except for the word used, e.g., {djica}). > See for a nice > article on relative identity, which nicely dissolves all the problems > with identity conundrums.

Well, assuming what you say is a reasonable summary (and, since it is essentially correct, I'll take it that it is), then I do not see why you keep fighting the Lojban approach which overtly takes advantage of these facts rather than sweeping them under the rug. I don't feature you as a fan of the memory-or-list approach, but doing nothing at all seems irresponsible. And, once we hit upon a program, it seems we should carry it all the way through, as Lojban has not yet done.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 01 of Sep., 2006 17:08 GMT On 9/1/06, John Cowan wrote: > Jorge Llamb?as scripsit: > > > You don't say {mi viska lo nu lo gerku cu bajra} and > > {lo nu lo bliku cu farlu lo mi jamfu cu xrani mi}? > > No. Both of those sound utterly bizarre to me.

The gi'uste wants the x1 of xrani to be an event. (And if I recall correctly, you have something like {ko viska lo su'u ci smacu cu bajra} in CLL.)

What about {daspo}. Can abstractions destroy things? Or {rinka}, can abstractions cause things to happen?

> > >An event is a state of affairs, and what we desire is > > >that a certain hypothetical state of affairs become actual. > > > > Then what you are saying is that the x2 of {djica} has to be a {du'u}, > > never a {nu}. > > No, that would be different. That would be saying that what we desire > is for a certain proposition to be true. That's not unreasonable, just > not the way we have currently defined "djica". (I wouldn't be opposed > to adding "(nu/du'u)" to the gismu list entry.)

If what we desire is that a certain state of affairs become actual, then {mi'o djica lo du'u da ca'a fasnu}. If what we desire is for a certain proposition to be true, then {mi'o djica lo du'u da ca'a jetnu}.

A state of affairs that becomes actual does not remain an abstraction, except perhaps in a metalinguistic sense. A state of affairs that becomes actual must occupy a region of space-time and is therefore not abstract in the sense of being outside of space-time.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 01 of Sep., 2006 20:39 GMT On 9/1/06, John E Clifford wrote: > > I think that you thik that saying {lo xanto} sticks an elephant into > the domain, but that is simply not the case: the elephants go into > some subjunctive domain, not the one we are talking in.

Indeed I do think that saying {lo xanto} sticks elephants into the domain of discourse.

I don't think speakers talk _in_ a domain of discourse. A domain of discourse is the set of things one talks _about_. A speaker can very easily talk about things that don't exist where the speaker exists.


> So, if you now take the > thing you want and project it into that original world, you have no > guarantee that it will be there. It has, of course, been introduced into a > domain — but not that one, only the subjunctive ones.

You seem to be identifying the domain of discourse (the set of things one talks about in a given discourse) with the world where the discourse takes place. For me those are very different things.

> > If I say {mi djica lo nu da xanto} I'm not saying that among many > > events of there being an elephant, there is a certain one such that > > that very one is the one I happen to want. > > I think we are getting muddled here. There certainly is an event that > I want, namely the event of there being an elephant.

Exactly. *The* event of there being an elephant. Not some particular one among all the events of there being an elephant. The one that subsumes them.

> Now, there are probably a lot of other events involving elephants that I > want, but they are aren't mentioned. What I do not want is a particular > occurrence of that event, any occurrence will do.

Exactly.

> But I have said nothing about occurrences of events here, just events > tout court.

Right, but the point is that the expression {lo nu da xanto} could very well refer to occurrences too, given the right context, as in {mi viska re lo nu da xanto}.

> Events can be individuated separately from their occurrences (and, indeed, > must be since two seaparate events can have the same occurrence).

As long as one individual can subsume another one, yes.

> Dogs are apparently inidviduated only by their occurrences.

Dogs have a subsumption tree as much as events do. Dogs can be individuated separately from their stages, for example. Young Fido and Old Fido are the same dog, even though they are different logical objects. Fido subsumes them both.

> So, while I can pick out an event from the blue and know that it is available, > I can't do that with a dog: "dog" refers to concreta, "event" to abstracta > (whose occurrences are conreta generally).

That's a metalinguistic classification that doesn't seem to offer any advantage. If {lo nu bajra} can refer to concreta, just like {lo gerku}, then {lo gerku} can refer to abstracta, just like {lo nu bajra}. As long as we keep in mind that "concreta" and "abstracta" are metalinguistic labels, not predicates of the object language, there is no problem. And the sumti expression used is never by itself sufficient to make the classification, neither in the case of events nor in the case of dogs. We need to interpret it according to the context.

> > So {lo nu la superman ly cinba} and {lo nu la klark kent ly cinba} > > are different events but the same occurrence. Is there a problem > > with this? No, as long as we don't take the two expressions to > > refer to the same logical object. > > Yes, this all works fine. What is a logical object? The event? But > we have already said (and demonstrated) that the two events are > different.

In this particular example, the two events are two different logical objects, yes, even though they are subsumed by the same occurrence.

> > As long as {la superman} and > > {la klark kent} refer to different logical objects, there is no problem in > > {la lois cu djica la superman} and {la lois cu djica la klark kent} having > > different truth values. > > What are the logical objects here? What you call personas appear to be > what would standardly be called senses or individual concepts, intensional > entities both.

You mentioned personas a couple of posts ago, and I think it fits well with the relative identity idea. A and B can be the same person and at the same time A and B are different personas.

So we can say that Supernan and Clark are the same person but different personas, and then, since they are different personas they are perforce different logical objects, and there is no problem in Lois wanting one and not the other. Lois may think that what she wants is a particular person, in any guise it may care to take and with no personas involved, but we know that what she really wants is a particular persona, since we know that the person won't satisfy her unless it appears as the persona she wants. She doesn't know that this person often appears as different personas to her, so she would make no distinction between person and persona.

> The problem now comes down to > this. {la lois djica la superman} has (apparently) the same logical form > as {la lois viska la superman}.

Yes. Just as {la lois djica lo nu la superman ly cinba} has the same logical form as {la lois ganse lo nu la superman ly cinba}.

> In the latter, {la superman} refers to the thing Superman, a gross physical > object in this case. In the former, for everything to work out right, > {la superman} has to refer to ^Superman, the individual concept unique > to Superman.

The persona Superman, yes. Just as {lo nu la superman ly cinba} has to refer to an occurrence in one case and to a particular subsumed event in the other. The same distinction occurs for objects as for events.

> How do we know when the difference occurs?

We need to examine the context of what is being said. There is no other way.

> We can memorize (or keep a list) of all the places where the reference > of a rweferring phrase is its sense rather than its referent.

That wouldn't help, because the same place of a predicate may be occupied by different nodes of the subsumption tree in different contexts. (The subsumption tree does not have only two points either, it's not just a sense and a referent, or a type and a token, or an individual and a stage, or ..., but it's an indefinitely long chain of branches.)

> Or we can arrange things so that the shift doesn't happen at all. Or we can > have a mark that warns of the shift. Clearly, one of the latter two approaches > is better than the first. And they actually come very close to being the > same thing in practice: in Lojban, we write {tu'a lo xanto} rather than {lo xanto} > in the problematic places, so they are marked.

A single mark does not tell you at which point of the indefinitely extended subsumption tree you are at. Does {tu'a lo xanto} refer to an event type, or to an event token of that event type, for example?

> > See for a nice > > article on relative identity, which nicely dissolves all the problems > > with identity conundrums. > > Well, assuming what you say is a reasonable summary (and, since it > is essentially correct, I'll take it that it is), then I do not see why you > keep fighting the Lojban approach which overtly takes advantage of these > facts rather than sweeping them under the rug.

I thought it was you doing the sweeping under the rug. The problem you set out to solve for {lo xanto} resurfaces identically for {lo fasnu}, since there are no logical differences between {lo xanto} and {lo fasnu}.

If {djica lo fasnu} and {viska lo fasnu} cause no problems, (and we seem to agree that they don't cause problems), then neither will {djica lo xanto} and {viska lo xanto}, because from the point of view of formal logic we don't care about the specific meaning of {xanto} or {fasnu}.

If {lo fasnu} can refer to an event that subsumes different events, so can {lo xanto} refer to an elephant type that subsumes different elephant tokens.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 01 of Sep., 2006 23:16 GMT posts: 2388

> On 9/1/06, John E Clifford wrote: > > > > I think that you thik that saying {lo xanto} sticks an elephant into > > the domain, but that is simply not the case: the elephants go into > > some subjunctive domain, not the one we are talking in. > > Indeed I do think that saying {lo xanto} sticks elephants into the > domain of discourse. > > I don't think speakers talk _in_ a domain of discourse. A domain > of discourse is the set of things one talks _about_. A speaker can very > easily talk about things that don't exist where the speaker exists.

Existence doesn't enter, as we have said over and over. What is of interest is what is available for quantification. Things buried away in subjunctive worlds are not. They can — in a separate act — be brought into an ongoing domain, but mere mention in not always enough — especually in these very indirect ways.

> > So, if you now take the > > thing you want and project it into that original world, you have no > > guarantee that it will be there. It has, of course, been introduced into a > > domain — but not that one, only the subjunctive ones. > > You seem to be identifying the domain of discourse (the set of things > one talks about in a given discourse) with the world where the discourse > takes place. For me those are very different things.

I guess I am trying to figure out just what (if any one thing) you do mean by domain of discourse. I take it to be all the things one can quantify over or immediately refer to in a conversation. The point about the things one says one wants is tht, although they are referred to, that reference is not immediate but mediated (through subjunction) and so doe not enter the actual domain of discourse. On the other hand, no one can object or refuse to allow the usage in a want claim, because that reference is determinative of another domain of discourse briefly visited. If I try to go on with that reference outside that narrow bound, then I have to get the consent of my fellow conversants. Hence the non-peculiarity of "I want a unicorn even though there aren't any." I think you want every referring expression, regardless of context, to refer, to guarantee that there is something in the domain to which it refers, but that just leads to paradoxes and fails to explain actual linguistic usage.

> > > If I say {mi djica lo nu da xanto} I'm not saying that among many > > > events of there being an elephant, there is a certain one such that > > > that very one is the one I happen to want. > > > > I think we are getting muddled here. There certainly is an event that > > I want, namely the event of there being an elephant. > > Exactly. *The* event of there being an elephant. Not some particular one > among all the events of there being an elephant. The one that subsumes > them.

Yes. So?

> > Now, there are probably a lot of other events involving elephants that I > > want, but they are aren't mentioned. What I do not want is a particular > > occurrence of that event, any occurrence will do. > > Exactly. > > > But I have said nothing about occurrences of events here, just events > > tout court. > > Right, but the point is that the expression {lo nu da xanto} could very > well refer to occurrences too, given the right context, as in > {mi viska re lo nu da xanto}.

Yes. There is a systematic ambiguity in Lojban, which no one has really bothered to sort out since it doesn't generally cause problems. I am not sure it is causing one now either. What is the point of your remark?

> > Events can be individuated separately from their occurrences (and, indeed, > > must be since two separate events can have the same occurrence). > > As long as one individual can subsume another one, yes.

Even without that. Consider that two people may well describe an occurrence in totally different terms and yet both describe it correctly (the American Civil War has a large variety of different decriptions, for example, not all of them even compatible with some others).

> > Dogs are apparently inidviduated only by their occurrences. > > Dogs have a subsumption tree as much as events do. Dogs can > be individuated separately from their stages, for example. Young Fido > and Old Fido are the same dog, even though they are different logical > objects. Fido subsumes them both.

I am not sure just what you mean by a logical object. We can, for various purposes consider /Fido as the sum of his spatiotemporal parts. For other purposes, that would be a bad move. And others are probably indifferent. I guess I don't see your point here or what it has to do with events. I suppose we could consider an event as being somehow the sum of its parts — and sometimes we do: in processes we talk about the various stages, for activities we talk about the bit that is repeated over and over to make the activity, and so on. But that doesn't seem to deal with the issue at hand, whether there is an event that I want when I want an event, to which the answer is clearly yes — in just the way that there is not a dog I want when I want a dog (any one will do).

> > So, while I can pick out an event from the blue and know that it is available, > > I can't do that with a dog: "dog" refers to concreta, "event" to abstracta > > (whose occurrences are conreta generally). > > That's a metalinguistic classification that doesn't seem to offer any > advantage. If {lo nu bajra} can refer to concreta, just like {lo gerku}, then > {lo gerku} can refer to abstracta, just like {lo nu bajra}. As long as we keep > in mind that "concreta" and "abstracta" are metalinguistic labels, not > predicates of the object language, there is no problem. And the sumti > expression used is never by itself sufficient to make the classification, > neither in the case of events nor in the case of dogs. We need to > interpret it according to the context.

Well, no. Admittedly, there is an ambiguity about event expressions which allows them to refer both to abstract objects (and there is no problem in saying so in Lojban or in English, it is not metalinguistic except in the sense that it is talking about talking) and occurrences, which are a bit more concrete (there is a work-around on this but I find it more muddling than helpful and I don't think it is needed anyhow). But {lo gerku} only refers to dogs, not to any abstract thing (Dogkind or Dogness or whatever). Now, to be sure, we could totally restructure the semantics of Lojban so that {lo gerku} did just refer to Dogdom (or Urdog even), but you have declined that offer (quite wisely, on the whole)in the past. The end result would scarcely be noticable in the language in use but the semantics not would have this curious two-track scheme: abstracts for bare {lo broda}, concrete things for everything else — except, of course, the things already marked as being abstracts. As I have said repeatedly, this move seems very much against the tendency in Lojban to take everything in terms of discrete individuals (even the work for water is a count noun, after all). And, so far as I can tell, the only thing you gain by it is that {mi djica lo pavyseljirna} doesn't screw up the logical part of "logical language" — which can be dealt with (and better) by other means.

> > > So {lo nu la superman ly cinba} and {lo nu la klark kent ly cinba} > > > are different events but the same occurrence. Is there a problem > > > with this? No, as long as we don't take the two expressions to > > > refer to the same logical object. > > > > Yes, this all works fine. What is a logical object? The event? But > > we have already said (and demonstrated) that the two events are > > different. > > In this particular example, the two events are two different logical > objects, yes, even though they are subsumed by the same occurrence.

I wouldn't say that the two events are subsumed under the same occurrence; I would say that the occurrence is subsumed under both events. the events are clearly higher up the tree than the occurrence.

> > > As long as {la superman} and > > > {la klark kent} refer to different logical objects, there is no problem in > > > {la lois cu djica la superman} and {la lois cu djica la klark kent} having > > > different truth values. > > > > What are the logical objects here? What you call personas appear to be > > what would standardly be called senses or individual concepts, intensional > > entities both. > > You mentioned personas a couple of posts ago, and I think it fits well > with the relative identity idea. A and B can be the same person and at > the same time A and B are different personas. > > So we can say that Supernan and Clark are the same person but different > personas, and then, since they are different personas they are perforce > different logical objects, and there is no problem in Lois wanting one and > not the other. Lois may think that what she wants is a particular person, in > any guise it may care to take and with no personas involved, but we know > that what she really wants is a particular persona, since we know that the > person won't satisfy her unless it appears as the persona she wants. She > doesn't know that this person often appears as different personas to her, so > she would make no distinction between person and persona.

Yes, that about sums it up, although "persona," while apt in this case, misses out on the fact that we are here talking about intensional objects, with all that that entails (though the term is introduced here exactly to explain the way such objects do work and why that both is not a problem and why it might seem to be one).

> > The problem now comes down to > > this. {la lois djica la superman} has (apparently) the same logical form > > as {la lois viska la superman}. > > Yes. Just as {la lois djica lo nu la superman ly cinba} has the same logical > form as {la lois ganse lo nu la superman ly cinba}.

But the point is that the logical form in the second case really is the same; both expressions refer to the event, wheereas in the first pair one occurrence of {la superman} refers to the object Superman (otherwise Clark Kent) while the other refers to the persona Superman (distinct from the persona Clark Kent). The ambiguity in the second case (if we insist that there is one) is not between the referent of the expression and its sense but between two possible referents. The beauty of the event arguments is that now there are no opaque places, just opaque objects in perfectly normal places. So, the sense/reference ambiguity never occurs (and, if Cowan had his way, neither would the event/occurrence, apparently).

> > In the latter, {la superman} refers to the thing Superman, a gross physical > > object in this case. In the former, for everything to work out right, > > {la superman} has to refer to ^Superman, the individual concept unique > > to Superman. > > The persona Superman, yes. Just as {lo nu la superman ly cinba} has to > refer to an occurrence in one case and to a particular subsumed event in the > other. The same distinction occurs for objects as for events.

Well, the two cases are somewhat analogous (although that the event description HAS TO refer to an occurrence seems unsupported — it is just easier to allow the ambiguity than to go through all the folderol about events being presented by — or through — occurrences). Analogous, but not the same, for one is semantically motivated, the other is only factually so.

> > How do we know when the difference occurs? > > We need to examine the context of what is being said. There is no other > way.

Well, part of the point of using events was so that we did not have to examine the context, so that the difference was on its face. Now, it turns out that we have to do this word by word -- even place by place, which is both tedious and unreliable. Better to prevent the problem arising at all. If that means that we should also not use event language in one place or the other where we do, that is OK, too: which context shall we drop and what shall we substitute (I would drop them {djica} and substitute propositions, but that is a predilection of mine that I don't expect others to share).

> > We can memorize (or keep a list) of all the places where the reference > > of a referring phrase is its sense rather than its referent. > > That wouldn't help, because the same place of a predicate may be occupied > by different nodes of the subsumption tree in different contexts. (The > subsumption tree does not have only two points either, it's not just a sense > and a referent, or a type and a token, or an individual and a stage, or ..., > but it's an indefinitely long chain of branches.)

Pretty much the choice is between sense and reference in the general opaque case and between event and occurrence in that case. so, what we have to do is use the right thing at the right time: there are going to be cases of wanting where we can refer by referent: "I want that elephant right there" for example. And there will be times when in places where we expect regular referents, we need to use the other form. And the same will be true for events and occurrences, I'm sure. So, the labels on places are not requirements, merely guidelines: be careful here to say just what you want to say. By the way, it really is just sense and reference; the other things, token and type, and the like are from a different sort of tree (well, sense and reference aren't from trees at all, they are just two aspects of any expression. You can talk about the sense of the sense, but that is a constant and so not very interesting.


> > Or we can arrange things so that the shift doesn't happen at all. Or we can > > have a mark that warns of the shift. Clearly, one of the latter two approaches > > is better than the first. And they actually come very close to being the > > same thing in practice: in Lojban, we write {tu'a lo xanto} rather than {lo xanto} > > in the problematic places, so they are marked. > > A single mark does not tell you at which point of the indefinitely extended > subsumption tree you are at. Does {tu'a lo xanto} refer to an event type, or to > an event token of that event type, for example?

I am not sure whether type/token as such really applies to events (except possibly as the difference between events and occurrences. Each event, wherever it is on the tree is unique. It may be lower or higher and a subsumption tree, but what is below it is not a token of it nor is it a token of some higher event. At worst it is a species of a higher genus and genus to the lower species. Now, officially, {tu'a lo xanto} is a very high event on the tree: "something happens involving an elephant or elephants" with the admonition to fill it out more exactly "in the obvious way." So, the intended event is probably quite far down under the explicit one. And the explicit one will generally not work, since there are all sorts of things involving elephants such that should they occur, my want would not be satisfied and such that I could be satisfied without their occurring. Pragmatically, I think this is usually enough (don't say more than you need to) and it dodges the logical bullet nicely.

> > > See for a nice > > > article on relative identity, which nicely dissolves all the problems > > > with identity conundrums. > > > > Well, assuming what you say is a reasonable summary (and, since it > > is essentially correct, I'll take it that it is), then I do not see why you > > keep fighting the Lojban approach which overtly takes advantage of these > > facts rather than sweeping them under the rug. > > I thought it was you doing the sweeping under the rug. The problem you > set out to solve for {lo xanto} resurfaces identically for {lo fasnu}, > since there > are no logical differences between {lo xanto} and {lo fasnu}.

Of course it does, it is about the effect of a given place on what is in that place, the shift from extension to intension; that is why we want to get rid of those places. But we can only do that if the approach is uniform, else we are back to — if we still have a place where an ordinary reference to a thing has to be taken as to its sense, then replacing such reference with intensional objects won't work(indeed, it will be worse, since now {lo nu... } is not about the event but about the sense of the expression and that will get the wrong thing).

> If {djica lo fasnu} and {viska lo fasnu} cause no problems, (and we seem to > agree that they don't cause problems),

I'm not sure where you got the idea that we agree on this. They behave exactly like the {lo xanto} case. That is, the place will still shift the meaning from referent to sense.

> then neither will {djica lo xanto} and > {viska lo xanto}, because from the point of view of formal logic we don't care > about the specific meaning of {xanto} or {fasnu}.

Quite true; we care about the form, which, as you have set things up involves the specific meanings of {djica} and {viska}. We want to get away from that and really only be concerned with the form.

> If {lo fasnu} can refer to an event that subsumes different events, so can > {lo xanto} refer to an elephant type that subsumes different elephant tokens.

The analogy doesn't get off the ground, for two reasons 1)the interesting case is not {lo fasnu...} but a referring expression starting {lo nu ...} (we go by form, not content — or are trying to) and 2) within such desirable expressions the subsumption relation is different (certainly in content, maybe in form as well) from that between tokens and types.

Once we no longer have any opaque context is handled by referring to an abstract object rather than a concrete one, we can get rid of all opaque places. Until we get to that goal, we are going to get opaquwe places with all the problems they entail.

Oh yes, an {lo xanto} doesn't refer to an elephant type (if the token/type distinction is meaningful when talking about elephants); it refers to elephants.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Sat 02 of Sep., 2006 21:39 GMT posts: 2388 Enough already! While I still think that events are significantly different from kinds, I concede that xorxes will always be able to come up with a decent analogical argument to the contrary. So I concede that the whole problem of intensional contexts could be handled as well using kinds as events (indeed, since I have in the past laid out how this would work, this part I have never denied). That being said, I still think that the way to deal with the problem of intensional contexts (and, yes, I still think there is a problem – as xorxes ought by now) is – in the case of {djica} and the like – to use event descriptions (for other predicates other abstracta will be appropriate: propositions or properties and maybe some of the rarer forms; happily {tu’a} is imprecise on which abstractum is involved). I do so for the following reasons (so far). 1. It is the way Lojban has chosen to go, even though it has not been very thorough about it and even though Lojbanists have pretty regularly not followed the recommended pattern.

2. It appears that most languages, when looked at with an eye to this issue, use events or something very like. It is certainly the case that languages regularly quantify over events directly (i.e., not “for some event …,” just “anything” or “something” clearly referring to events e.g., they happen). And logicians surely do (insofar as there is a consensus on this or any issue).

3. We know what an event is and how to form an event description; we are not sure what sort of things kinds are: they may be just properties or they may be sui generis. They may event not be abstract at all, in the usual sense of not being concrete (both these lines have been pursued but I forget what the upshot was, aside from that they all worked somehow). So, do we need a new abstractor for kinds (and where would we get it?) or is the present set enough?

4. No alternative proposal has actually been offered. Xorxes has regularly pointed out that the supposed advantages of events are also available for kinds and the like, but he has not actually proposed that the problem should be dealt with in this way. Notice, he has not dealt with the issues in the previous section, surely a first step in such a proposal. He has instead continued to use ordinary descriptions in this context and talked as though they had or might have a different reading.

5. Most importantly, *something* needs to be done: a logical language cannot have intensional contexts unmarked, since then certain logical laws do not apply. The Lojban solution is to do away with intensional places but allow intensional descriptions. The rules of logic apply to them in an explicit way, so that even the appearance of the old problems disappears. However, so long as there is one unmarked intensional place, the program fails: since we have no formal way of knowing that the place is still intensional, we can get failures of major laws – not something a “logical language” wants to do.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 06 of Sep., 2006 16:19 GMT On 9/2/06, John E Clifford wrote: > Enough already! > While I still think that events are significantly different from kinds, I concede > that xorxes will always be able to come up with a decent analogical argument > to the contrary.

My argument does not compare events with kinds. My argument compares events with objects, (and kinds of events with kinds of objects, or event types with object types, etc.).

> So I concede that the whole problem of intensional contexts could be > handled as well using kinds as events (indeed, since I have in the past > laid out how this would work, this part I have never denied). That being > said, I still think that the way to deal with the problem of intensional > contexts (and, yes, I still think there is a problem as xorxes ought by > now) is in the case of {djica} and the like to use event descriptions (for > other predicates other abstracta will be appropriate: > propositions or properties and maybe some of the rarer forms; happily > {tu'a} is imprecise on which abstractum is involved). I do so for the > following reasons (so far). > 1. It is the way Lojban has chosen to go, even though it has not been > very thorough about it and even though Lojbanists have pretty regularly > not followed the recommended pattern.

This reason can only persuade someone who is convinced that this chosen path is not a cul-de-sac. If one thinks that Lojban has taken a wrong turn in this regard, then following this path is not very appealing. Especially knowing that Lojban has not even made this choice in a thorough and consistent way.

> 2. It appears that most languages, when looked at with an eye to this > issue, use events or something very like. It is certainly the case that > languages regularly quantify over events directly (i.e., not "for some > event ," just "anything" or "something" clearly referring to events > e.g., they happen). And logicians surely do (insofar as there is a > consensus on this or any issue).

That languages can and do use events this way has not been under dispute, as far as I can tell. I am certainly not proposing banishing events from Lojban. It is also pretty clear that languages use objects in pretty much the same way that they use events.

> 3. We know what an event is and how to form an event description; > we are not sure what sort of things kinds are: they may be just properties > or they may be sui generis. They may event not be abstract at all, in the > usual sense of not being concrete (both these lines have been pursued but > I forget what the upshot was, aside from that they all worked somehow). So, > do we need a new abstractor for kinds (and where would we get it?) or is > the present set enough?

We know what an event is as much as we know what an object is. The relevant comparison is not between events and kinds, but between events and objects. If events are "abstract" in some sense, how can they be visible? If there is no problem with events being visible and being wantable, then there is no problem with objects being visible and being wantable.

> 4. No alternative proposal has actually been offered. Xorxes has > regularly pointed out that the supposed advantages of events are also > available for kinds and the like, but he has not actually proposed that > the problem should be dealt with in this way. Notice, he has not dealt > with the issues in the previous section, surely a first step in such a > proposal. He has instead continued to use ordinary descriptions in this > context and talked as though they had or might have a different reading.

You misrepresent my position. I have not pointed out that the supposed advantages of events are also available for kinds. I have pointed out that whatever is available for events is also available for objects. For the event-only solution to work, we are already making use of kinds/types anyway. We are already using events that subsume other events. And since there is no problem in using this subsumption tree with events, the exact same structure works for objects.

> 5. Most importantly, *something* needs to be done: a logical language > cannot have intensional contexts unmarked, since then certain logical laws > do not apply. The Lojban solution is to do away with intensional places but > allow intensional descriptions. The rules of logic apply to them in an explicit > way, so that even the appearance of the old problems disappears. However, > so long as there is one unmarked intensional place, the program fails: since > we have no formal way of knowing that the place is still intensional, we can > get failures of major laws not something a "logical language" wants to do.

All logical laws seem to work, both for events and for objects. All the problems with identity seem to arise from misconstruing relative identity as absolute identity. But this is not something that affects objects any differently than it affects events. Two events can be the same F and different Gs as much as two objects can be the same F and different Gs. Taking two events that are the same F as identical just because they are the same F leads to the same conundrums as taking two objects that are the same F as identical just because they are the same F.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Wed 06 of Sep., 2006 23:14 GMT posts: 2388

> On 9/2/06, John E Clifford wrote: > > Enough already! > > While I still think that events are significantly different from kinds, I concede > > that xorxes will always be able to come up with a decent analogical argument > > to the contrary. > > My argument does not compare events with kinds. My argument compares > events with objects, (and kinds of events with kinds of objects, or event types > with object types, etc.).

Well, I hope you are wrong about what your comparison is between, because the analogies (indeed, they are too close to be that — identities) that work are between events and thing types and between things and occurrences. Of course, we get things muddleed a bit since we tend to call occurrences events and thus make the token type fallacy. I don't think we are so apt to make the same mistake between things and thing types, but we do do so occasionally.

> > So I concede that the whole problem of intensional contexts could be > > handled as well using kinds as events (indeed, since I have in the past > > laid out how this would work, this part I have never denied). That being > > said, I still think that the way to deal with the problem of intensional > > contexts (and, yes, I still think there is a problem – as xorxes ought by > > now) is – in the case of {djica} and the like – to use event descriptions (for > > other predicates other abstracta will be appropriate: > > propositions or properties and maybe some of the rarer forms; happily > > {tu'a} is imprecise on which abstractum is involved). I do so for the > > following reasons (so far). > > 1. It is the way Lojban has chosen to go, even though it has not been > > very thorough about it and even though Lojbanists have pretty regularly > > not followed the recommended pattern. > > This reason can only persuade someone who is convinced that this chosen > path is not a cul-de-sac. If one thinks that Lojban has taken a wrong turn in > this regard, then following this path is not very appealing. Especially > knowing that Lojban has not even made this choice in a thorough and > consistent way.

Well, I don't think it is a wrong turn. I think that it was a bright move and the only problem is that we didn't carry it through. I see I didn't put down one major reason for using events here: it seems to be psychologically correct. Whenever you ask someone what it means to want something, he always comes up with an event (more or less specific), not a thing. And, of course, events fit the truth test very nicely, things only with more fill-in (mainly of events).

> > 2. It appears that most languages, when looked at with an eye to this > > issue, use events or something very like. It is certainly the case that > > languages regularly quantify over events directly (i.e., not "for some > > event …," just "anything" or "something" clearly referring to events > > e.g., they happen). And logicians surely do (insofar as there is a > > consensus on this or any issue). > > That languages can and do use events this way has not been under dispute, > as far as I can tell. I am certainly not proposing banishing events from > Lojban. It is also pretty clear that languages use objects in pretty much > the same way that they use events.

No, only object types — objects themselves prove to be fairly intractable, certainly in the real world domains, but in the extedended ones as well. > > > 3. We know what an event is and how to form an event description; > > we are not sure what sort of things kinds are: they may be just properties > > or they may be sui generis. They may event not be abstract at all, in the > > usual sense of not being concrete (both these lines have been pursued but > > I forget what the upshot was, aside from that they all worked somehow). So, > > do we need a new abstractor for kinds (and where would we get it?) or is > > the present set enough? > > We know what an event is as much as we know what an object is. The > relevant comparison is not between events and kinds, but between events > and objects. If events are "abstract" in some sense, how can they be > visible? If there is no problem with events being visible and being wantable, > then there is no problem with objects being visible and being wantable.

You seem to be talking about occurrences here. Events become visible as occurrences, just as object types become visible as objects.

> > 4. No alternative proposal has actually been offered. Xorxes has > > regularly pointed out that the supposed advantages of events are also > > available for kinds and the like, but he has not actually proposed that > > the problem should be dealt with in this way. Notice, he has not dealt > > with the issues in the previous section, surely a first step in such a > > proposal. He has instead continued to use ordinary descriptions in this > > context and talked as though they had or might have a different reading. > > You misrepresent my position. I have not pointed out that the supposed > advantages of events are also available for kinds. I have pointed out that > whatever is available for events is also available for objects. For the > event-only solution to work, we are already making use of kinds/types > anyway. We are already using events that subsume other events. And since > there is no problem in using this subsumption tree with events, the exact > same structure works for objects.


No object (that is not itself aa abstract) subsumes anything else. The nearest that we can come to that — for objects and occurrences — is the relation berween an objects (as extended in space and time) and its spatio-temporal parts, and that — while it has the same mathematical structure -- is a rather different relation semantically.

> > 5. Most importantly, *something* needs to be done: a logical language > > cannot have intensional contexts unmarked, since then certain logical laws > > do not apply. The Lojban solution is to do away with intensional places but > > allow intensional descriptions. The rules of logic apply to them in an explicit > > way, so that even the appearance of the old problems disappears. However, > > so long as there is one unmarked intensional place, the program fails: since > > we have no formal way of knowing that the place is still intensional, we can > > get failures of major laws – not something a "logical language" wants to do. > > All logical laws seem to work, both for events and for objects. All the > problems with identity seem to arise from misconstruing relative identity > as absolute identity. But this is not something that affects objects any > differently than it affects events. Two events can be the same F and different > Gs as much as two objects can be the same F and different Gs. Taking > two events that are the same F as identical just because they are the same > F leads to the same conundrums as taking two objects that are the same F > as identical just because they are the same F.

I agree that the problem with identity is misconstruing relative identity with absolute identity. The problem is that relative identity is often appropriate and to be used, but occasionally is not. We need some easy to tell which of these situations we are in — or to thwart misuse altogether. The event solution (and, admittedly, the object kind solution) does this latter. I am not sure what you mean by your Fs and Gs, but it seems to be about occurrences rather than events. And, of course, the conditions for identity of events is different from that for occurrences — and is always absolute (I think), The problem comes from suing relative identity inside an event description — in the Lojban pattern — but that is already disallowed.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 07 of Sep., 2006 12:58 GMT On 9/6/06, John E Clifford wrote: > > Of course, we get things muddleed a bit since we tend to call > occurrences events and thus make the token type fallacy.

How do you avoid the token type fallacy in Lojban? If {lo nu mi klama le zarci} refers to an event, how do you refer to one of its occurrences?

(My answer is that we don't normally distinguish type from token in Lojban, but I'd be interested to know what your answer is.)

> I see I didn't put down one major reason for using events here: it seems > to be psychologically correct. Whenever you ask someone what it > means to want something, he always comes up with an event (more or > less specific), not a thing.

Whenever you ask someone what it means to see/hear/move something, he always comes up with an event (more or less specific), not a thing.

That doesn't seem to help much.

> And, of course, events fit the truth test very > nicely, things only with more fill-in (mainly of events).

What is the truth test?

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 07 of Sep., 2006 13:39 GMT posts: 2388


> On 9/6/06, John E Clifford wrote: > > > > Of course, we get things muddleed a bit since we tend to call > > occurrences events and thus make the token type fallacy. > > How do you avoid the token type fallacy in Lojban? If {lo nu mi > klama le zarci} refers to an event, how do you refer to one of > its occurrences? > > (My answer is that we don't normally distinguish type from token in > Lojban, but I'd be interested to know what your answer is.)

I agree that we don't often distinguish types and tokens in Lojban (or English, for that matter). In English we can often indicate the token case with indexicals ("this" and the like — but that may be only a relative token, not yet necessarily an object) and we can always say "ultimate token" and the like in Philosophy talk. I don't know just how to do in explicitly in Lojban -- what is good word for "token" (I assume "type" is something like {ckaji}). Of course, locating an object or occurrence in space and time is probably the surest way of doing the ultimate tokening (that is what the indexicals do, after all). So narrowly tensed sentences work for events occurrences, at least some times.

> > I see I didn't put down one major reason for using events here: it seems > > to be psychologically correct. Whenever you ask someone what it > > means to want something, he always comes up with an event (more or > > less specific), not a thing. > > Whenever you ask someone what it means to see/hear/move something, > he always comes up with an event (more or less specific), not a thing. > > That doesn't seem to help much.

Interesting, if true — I haven't checked with others and my introspection is not as confident as it is about wishes and the like.

> > And, of course, events fit the truth test very > > nicely, things only with more fill-in (mainly of events). > > What is the truth test?

The one we have been using all along: "I want x" is true just in case I have a psychological lack such that, if *** x were to occur, that lack would be filled, and, if that lack were to be filled, *** X would have occurred.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by JohnCowan on Thu 07 of Sep., 2006 14:05 GMT posts: 149 John E Clifford scripsit:

> I agree that we don't often distinguish types and tokens in Lojban > (or English, for that matter). In English we can often indicate the > token case with indexicals ("this" and the like — but that may be only > a relative token, not yet necessarily an object) and we can always say > "ultimate token" and the like in Philosophy talk. I don't know just > how to do in explicitly in Lojban — what is good word for "token" > (I assume "type" is something like {ckaji}).

Probably "sevzi", the word for "instance" or "avatar".

Type-token talk is flawed anyhow, as Quine points out in the appropriate entry of _Quiddities_ (I don't have it to hand, but it's probably the obvious one), because of the riddle of compound types. Types should be compounded of types, as tokens of tokens, and it's clear that there are two tokens of "the" in the token "the cat sat on the mat", but how many types of "the" are there in the type "the cat sat on the mat", two or one? Either answer gets you into obvious difficulties.

-- John Cowan http://ccil.org/~cowan Micropayment advocates mistakenly believe that efficient allocation of resources is the purpose of markets. Efficiency is a byproduct of market systems, not their goal. The reasons markets work are not because users have embraced efficiency but because markets are the best place to allow users to maximize their preferences, and very often their preferences are not for conservation of cheap resources. --Clay Shirkey


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 07 of Sep., 2006 14:06 GMT On 9/7/06, John E Clifford wrote: > > I agree that we don't often distinguish types and tokens in Lojban > (or English, for that matter).

Great! Then our only disagreement seems to be that you consider this problematic in the case of objects and unproblematic in the case of events, whereas I consider it equally unproblematic in both cases.

> In English we can often indicate the token case with indexicals ("this" and > the like — but that may be only a relative token, not yet necessarily an > object) and we can always say "ultimate token" and the like in Philosophy > talk.

And "ultimate tokens" are context dependent anyway. What counts as an ultimate token in one context might disintegrate into finer subtokens in another context.

> I don't know just how to do in explicitly in Lojban -- > what is good word for "token" (I assume "type" is something like {ckaji}).

I would use {mupli} for the token-type relationship, but bearing in mind that nothing is a token in an absolute sense, or a type in an absolute sense.

> Of course, locating an object or occurrence in space and time is probably > the surest way of doing the ultimate tokening (that is what the indexicals > do, after all). So narrowly tensed sentences work for events > occurrences, at least some times.

I don't have a problem even with indexicals for types. As in "what will you have? That, what she's having" (i.e. the type, not the token).


> > What is the truth test? > > The one we have been using all along: "I want x" is true just in case I have > a psychological lack such that, if *** x were to occur, that lack would be > filled, and, if that lack were to be filled, *** X would have occurred.

I think the satisfaction condition for wanting an event is a bit more involved than just the event occurring. It is also required that the wanter realizes that the wanted event has occured (if Clark kisses Lois, then the event Lois wants, Superman kissing her, did occur, but because she didn't realize that the event she wants occured, her want was not satisfied.)

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 07 of Sep., 2006 14:38 GMT posts: 2388 Yeah, token-type is probably not the ideal concept set to use, although it is handy for pointing to the issues involved and — if not pushed too hard — does the job. To be sure, the distinction is always relative (or almost so), but there does come a point of a significant shift: the two tokens of "the" in the following (on my screen now — though the same would apply to those on your screen when you first read them)"the cat is on the mat" are untypy in a way that all the tokens and types above them are not. That is, it is hard to say what they might be types to. In a printed form, that endures, unlike the electronic forms, I suppose we can talk about the temporal subsegments of a token's existence, but this seems odd to call "token" rather than "part" or the like. We don't usually call the beginning midddle and end of an occurrence tokens of the occurrence, only parts (even for activities, where the parts are pretty much of the same type as the whole). I think that, in getting bogged down in potential complexities of any particular sort of analysis of what is going on here, we are apt to miss the important distinction that is involved, however hard it may be to make it in a foolproof way.


> John E Clifford scripsit: > > > I agree that we don't often distinguish types and tokens in Lojban > > (or English, for that matter). In English we can often indicate the > > token case with indexicals ("this" and the like — but that may be only > > a relative token, not yet necessarily an object) and we can always say > > "ultimate token" and the like in Philosophy talk. I don't know just > > how to do in explicitly in Lojban — what is good word for "token" > > (I assume "type" is something like {ckaji}). > > Probably "sevzi", the word for "instance" or "avatar". > > Type-token talk is flawed anyhow, as Quine points out in the appropriate > entry of _Quiddities_ (I don't have it to hand, but it's probably the > obvious one), because of the riddle of compound types. Types should be > compounded of types, as tokens of tokens, and it's clear that there are > two tokens of "the" in the token "the cat sat on the mat", but how many > types of "the" are there in the type "the cat sat on the mat", two or one? > Either answer gets you into obvious difficulties. > > — > John Cowan http://ccil.org/~cowan > Micropayment advocates mistakenly believe that efficient allocation of > resources is the purpose of markets. Efficiency is a byproduct of market > systems, not their goal. The reasons markets work are not because users > have embraced efficiency but because markets are the best place to allow > users to maximize their preferences, and very often their preferences are > not for conservation of cheap resources. --Clay Shirkey > > > >


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 07 of Sep., 2006 14:48 GMT On 9/7/06, John Cowan wrote:

> Type-token talk is flawed anyhow, as Quine points out in the appropriate > entry of _Quiddities_ (I don't have it to hand, but it's probably the > obvious one), because of the riddle of compound types. Types should be > compounded of types, as tokens of tokens, and it's clear that there are > two tokens of "the" in the token "the cat sat on the mat", but how many > types of "the" are there in the type "the cat sat on the mat", two or one? > Either answer gets you into obvious difficulties.

A two way absolute distinction is flawed, but type-token is a relative distinction. Nothing is a type or a token in an absolute sense. Dogs (type) have four legs. Are those legs types or tokens? They are at one level four tokens of the type leg, but at another level each leg is a leg type (the left front leg type, the right front leg type, etc.)

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 07 of Sep., 2006 18:18 GMT posts: 2388


> On 9/7/06, John E Clifford wrote: > > > > I agree that we don't often distinguish types and tokens in Lojban > > (or English, for that matter). > > Great! Then our only disagreement seems to be that you consider this > problematic in the case of objects and unproblematic in the case of events, > whereas I consider it equally unproblematic in both cases.

I'm not sure in what sense problematic. Cowan has just pointed out that the whole token-type theory fails at a certain level of stress. I have no more problem with the distinction (or relative positioning) for objects than for events. I suspect that where we differ is that I do think that there are ultimate tokens in both cases and that what you would take as further tokens I would consider to be in a different relation — parts, say.

> > In English we can often indicate the token case with indexicals ("this" and > > the like — but that may be only a relative token, not yet necessarily an > > object) and we can always say "ultimate token" and the like in Philosophy > > talk. > > And "ultimate tokens" are context dependent anyway. What counts as > an ultimate token in one context might disintegrate into finer subtokens in > another context.

Well I think this is our disagreement (or the one that has to do with tokens and types anyhow). I think that a real physical rabbit is an ultimate token, You may have to deal in a structurally similar way with its parts — spatial or temporal — but they are parts, not tokens. The rabit, on the other hand, is a tken of rabbitness and a vareity of other types more and less general (some of wehich are themselves tokens of higher types. I think it misses a significant difference to treat the parts of a rabbit in the same way as the token of a type. I know the analogies can be carried on indefinitely, but that does not seem to me to discount a central fact of difference.


> > I don't know just how to do in explicitly in Lojban -- > > what is good word for "token" (I assume "type" is something like {ckaji}). > > I would use {mupli} for the token-type relationship, but bearing in mind that > nothing is a token in an absolute sense, or a type in an absolute sense.

Again, we just disagree on this. Rabbithood is a type, albeit also a token of animalhood, say. But a particular rabbit ids only a token (of an array of types).

> > Of course, locating an object or occurrence in space and time is probably > > the surest way of doing the ultimate tokening (that is what the indexicals > > do, after all). So narrowly tensed sentences work for events > > occurrences, at least some times. > > I don't have a problem even with indexicals for types. As in "what > will you have? > That, what she's having" (i.e. the type, not the token).

Right, indexicals are not infallible guides to toeknness, though they do generally get us down the tree to lower (more nearly purely token) slots.

> > > > What is the truth test? > > > > The one we have been using all along: "I want x" is true just in case I have > > a psychological lack such that, if *** x were to occur, that lack would be > > filled, and, if that lack were to be filled, *** X would have occurred. > > I think the satisfaction condition for wanting an event is a bit more involved > than just the event occurring. It is also required that the wanter > realizes that > the wanted event has occured (if Clark kisses Lois, then the event Lois wants, > Superman kissing her, did occur, but because she didn't realize that the > event she wants occured, her want was not satisfied.)

Yes, the standard formula leaves out the bit about the person being aware of the event in question. There is a long argument about that — leaving it out does mean that the definition is sometimes wrong; but putting it in would open the possibility to getting it wrong as well, since just believing that the event had occurred would become enough to satisfy the conditions. I am, admittedly, not quite sure why this is wrong, why thinking you have what you want is not as good a justification for saying you want it as actually having it is. But those who deal with these things seem to think it screws up some metaphysical (it doesn't really seem to be semantic) principle. I think, on balance, that I would actually go — for the definition at least — with the thoroughly subjective format (but then, I think that all of these abstractions are really propositions, as events become on this reading). So, "a wants ***b" is "there is a psychological lack which a feels and which is such that, if a were to believe that ***b occurred, the lack would be filled, and, if that lack were to be filled, a would have come to believe that ***b had occurred" In terms of the issue of where reference is in the sentence, this actually buries it yet another level away from the surface, the evaluating domain.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Thu 07 of Sep., 2006 19:19 GMT On 9/7/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > Then our only disagreement seems to be that you consider this > > problematic in the case of objects and unproblematic in the case of events, > > whereas I consider it equally unproblematic in both cases. > > I'm not sure in what sense problematic.

In the sense that you have no problem with using one and the same expression to refer to events type or token, but you do have a problem with using one and the same expression to refer to objects type or token.

> Cowan has just pointed out that the whole token-type > theory fails at a certain level of stress. I have no more problem with > the distinction (or relative positioning) for objects than for events. I suspect > that where we differ is that I do think that there are ultimate tokens in both > cases and that what you would take as further tokens I would consider to > be in a different relation — parts, say.

We differ about that too, probably. But for the puroposes of this discussion (i.e. the "shove it into an event" solution to intensional contexts), the first disagreement seems like the relevant one.

I would not hold that the part-whole relationship is equivalent to the token-type one. But I would take parts in some contexts to function as tokens. For example:

> I think that a real physical rabbit is an ultimate token, You may have to deal > in a structurally similar way with its parts — spatial or temporal — but they > are parts, not tokens.

Suppose R is a rabbit, and R' is the part of R consisting of R minus one hair. I would say R and R' are both one and the same rabbit, it would be strange to say that R' is not a rabbit, and I would hesitate saying that either is an ultimate token.

> > I think the satisfaction condition for wanting an event is a bit more involved > > than just the event occurring. It is also required that the wanter > > realizes that the wanted event has occured > > Yes, the standard formula leaves out the bit about the person being aware > of the event in question. There is a long argument about that — leaving it > out does mean that the definition is sometimes wrong; but putting it in > would open the possibility to getting it wrong as well, since > just believing that the event had occurred would become enough to satisfy > the conditions.

Why couldn't the condition be that both the event happens _and_ the wanter is aware that it happens?

This brings up an interesting parallel:

Wanting an object is satisfied by a token of the object existing and the wanter having it available in some sense.

Wanting an event is satisfied by a token of the event existing (i.e. ocurring) and the wanter having it available in some sense (i.e. being aware of it).

> I am, > admittedly, not quite sure why this is wrong, why thinking you have what > you want is not as good a justification for saying you want it as actually > having it is. But those who deal with these > things seem to think it screws up some metaphysical (it doesn't really > seem to be semantic) > principle. I think, on balance, that I would actually go — for the definition > at least — with the thoroughly subjective format (but then, I think that all > of these abstractions are really propositions, as events become on this > reading).

"Wanting that ..." certainly seems to belong with the "knowing that ...", "believeing that ...", "assuming that ...", "finding out that ...", etc. bunch, i.e. those that require {du'u} in Lojban. "Wanting X" belongs with the "thinking about X", "imagining X", "talking about X" etc. bunch. I take {djica} as defined to be in the latter category. In any case, both would seem to be acceptable predicates for Lojban.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Thu 07 of Sep., 2006 22:07 GMT posts: 2388

> On 9/7/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > Then our only disagreement seems to be that you consider this > > > problematic in the case of objects and unproblematic in the case of events, > > > whereas I consider it equally unproblematic in both cases. > > > > I'm not sure in what sense problematic. > > In the sense that you have no problem with using one and the same > expression to refer to events type or token, but you do have a problem > with using one and the same expression to refer to objects type or token.

Well, actually, now that I have gotten clear on the distinction, I feel the same qualm about using the same term in the case of events. That is to say, for careful discussion. In just normal vulgar speech I figure that context etc. will take care of what is going one and, if not, then later correction in the conversation will preven seriously bad things from happening. Of course, I think that the distinction needs to be made in Lojban on a more regular basis, though it is not clear just what the limits are (clearly I think that wants and other opaque contexts are cases for the distinction).

> > > Cowan has just pointed out that the whole token-type > > theory fails at a certain level of stress. I have no more problem with > > the distinction (or relative positioning) for objects than for events. I suspect > > that where we differ is that I do think that there are ultimate tokens in both > > cases and that what you would take as further tokens I would consider to > > be in a different relation — parts, say. > > We differ about that too, probably. But for the purposes of this discussion > (i.e. the "shove it into an event" solution to intensional contexts), the first > disagreement seems like the relevant one. To be careful, remember that this is not a "shove it all into events" proposal but rather, a "always use the appropriate abstraction" proposal (well, not exactly a proposal since it is the defacto position, labeit poorly executed).

> I would not hold that the part-whole relationship is equivalent to the > token-type one. But I would take parts in some contexts to function > as tokens. For example: > > > I think that a real physical rabbit is an ultimate token, You may have to deal > > in a structurally similar way with its parts — spatial or temporal — but they > > are parts, not tokens. > > Suppose R is a rabbit, and R' is the part of R consisting of R minus one hair. > I would say R and R' are both one and the same rabbit, it would be strange > to say that R' is not a rabbit, and I would hesitate saying that either is an > ultimate token.

I miss the point here. Of course they are the same rabbit — that is how you set it up, after all, But it would be the same rabbit in any case (I think the technical term used to be genidentity). So what is your point here. The rabbit is a token, each of its parts are its part, not tokens of it as a type.

> > > I think the satisfaction condition for wanting an event is a bit more involved > > > than just the event occurring. It is also required that the wanter > > > realizes that the wanted event has occured > > > > Yes, the standard formula leaves out the bit about the person being aware > > of the event in question. There is a long argument about that — leaving it > > out does mean that the definition is sometimes wrong; but putting it in > > would open the possibility to getting it wrong as well, since > > just believing that the event had occurred would become enough to satisfy > > the conditions. > > Why couldn't the condition be that both the event happens _and_ the > wanter is aware that it happens?

There were reasons — which I did not find convincing — convincing for thinking that the actual occurence drops out: the wanter has merely to believe — not know — that the event occurred for the lack to be filled (it was a mental lack, after all). Of course, he may discover that the event has not occurred and the lack reappear, but that surely does not mean that he did not want it before.

> This brings up an interesting parallel: > > Wanting an object is satisfied by a token of the object existing and the > wanter having it available in some sense. > > Wanting an event is satisfied by a token of the event existing (i.e. ocurring) > and the wanter having it available in some sense (i.e. being aware of it).

Nice, but of course in the first one it is also required that the wanter be aware of the object being available (this is no doubt part of the availability, so separating it off in one case but not the other makes parallelism less clear. The point in favor of events is that there is an event required in both cases and there is something to be said for saying what that event is, rather than having to figure it out from context. Of course, we can work with either events (occurrence types)or object types and (given how vague the events typically are) there is little technically to tell them apart. But that assumes that, in the object case, reference is made to object types, not to objects.

> > I am, > > admittedly, not quite sure why this is wrong, why thinking you have what > > you want is not as good a justification for saying you want it as actually > > having it is. But those who deal with these > > things seem to think it screws up some metaphysical (it doesn't really > > seem to be semantic) > > principle. I think, on balance, that I would actually go — for the definition > > at least — with the thoroughly subjective format (but then, I think that all > > of these abstractions are really propositions, as events become on this > > reading). > > "Wanting that ..." certainly seems to belong with the "knowing that ...", > "believeing that ...", "assuming that ...", "finding out that ...", etc. bunch, > i.e. those that require {du'u} in Lojban. "Wanting X" belongs with the > "thinking about X", "imagining X", "talking about X" etc. bunch. I take > {djica} as defined to be in the latter category. In any case, both would > seem to be acceptable predicates for Lojban.

Well, I am not sure about categorization. I have no problem dealing with {djica} with any abstraction, though I think that it is useful to keep various categories separate if there turn out to be really different categories, as I am sure there are. Wanting x and wanting that x both seem to be subjunctive and counter factual. Believing that x and that lot are prettry clearly not counterfactual (you can believe what is true, after all, but you cannot want what you believe you have). Thinking about x and the like are precisely the area where Lojban fails to mark opaque contexts, so bringing them in is not terribly useful at the moment (by the way, I think that imaging is a different sort of thing from thinking about and is more like picturing,, but I haven't done the sort of work needed to really sort all these out). That at least some of these arose from malglico copying of English subject raising just makes the point the clearer — and harder to understand why we went that weay, since I know that the problem was noted at the time).


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by JohnCowan on Fri 08 of Sep., 2006 14:21 GMT posts: 149 Jorge Llamb?as scripsit:

> A two way absolute distinction is flawed, but type-token is a relative > distinction. Nothing is a type or a token in an absolute sense.

I don't understand this. Something can be a type only if it has a class of tokens that it represents; a specific elephant (Jumbo, say; http://en.wikipedia.org/wiki/Jumbo ) is not a type of anything, because he does not represent anything, non-metaphorically speaking.

-- "The serene chaos that is Courage, and the phenomenon cowan@ccil.org of Unopened Consciousness have been known to the John Cowan Great World eons longer than Extaboulism." "Why is that?" the woman inquired. "Because I just made that word up", the Master said wisely. --Kehlog Albran, The Profit http://www.ccil.org/~cowan


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 08 of Sep., 2006 15:09 GMT On 9/7/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > In the sense that you have no problem with using one and the same > > expression to refer to events type or token, but you do have a problem > > with using one and the same expression to refer to objects type or token. > > Well, actually, now that I have gotten clear on the distinction, I feel the > same qualm about using the same term in the case of events.

OK, that at least makes some more sense to me.

Even though Lojban makes no structural distinction between types and tokens, in many cases it is still pragmatically straightforward to tell which is meant. For example, since a type is a singleton, any quantified term has to be at least Griceanly interpreted as involving tokens, since quantifying over a singleton is pointless. And given that we have available, through the use of quantifiers, a fairly reliable indication that tokenness is involved, the absence of any quantifier is at least suggestive of typeness, or at least that the type-token distinction is not of interest.

> That is to say, for careful discussion. In just normal > vulgar speech I figure that context etc. will take care of what is going one > and, if not, then later correction in the conversation will preven seriously > bad things from happening. Of course, I think that the distinction needs > to be made in Lojban on a more regular basis, though it is not > clear just what the limits are (clearly I think that wants and other > opaque contexts are cases for the distinction).

But in the current language we don't have any structural way of making the distinction. (We can always make the distinction by explaining out what we mean, of course, but presumably you are talking about doing it within the grammar.)

> > But for the purposes of this discussion > > (i.e. the "shove it into an event" solution to intensional contexts), the first > > disagreement seems like the relevant one. > To be careful, remember that this is not a "shove it all into events" proposal > but rather, a "always use the appropriate abstraction" proposal (well, not > exactly a proposal since it is the defacto position, labeit poorly executed).

(Maybe you mean the de jure position, since surely it is not the de facto one. It is not even clear that it is the de jure one either. Maybe the position that reflects your understanding of the original intent.)

Anyway, what I have been discussing all along in this thread is the shove-it-into-an-event solution, as presented for example for {djica}. My point has been that any problem that objects may have in intensional contexts are replicated for events, because events have the same potential type-token subsumption tree that objects have.

If other entities with no tokens, such as propositions, are used then matters might be different, but I'm comparing objects with events and I'm saying that shoving objects into events doesn't escape the intensional context issues, it only transfers them from the objects to the events.


> Wanting x and wanting that x both > seem to be subjunctive and counter factual. Believing that x and that lot > are prettry clearly not counterfactual (you can believe what is true, after all, > but you cannot want what you believe you have).

If you may want what you already have (not knowing that you have it), then that would mean wanting is not necessarily counter factual, wouldn't it? Is factuality defined by what really is, or by what the wanter believes to be?

> Thinking about x and the like are precisely the area where Lojban fails > to mark opaque contexts, so bringing them in is not terribly useful at > the moment

By requiring events in the x2 of {djica} (as opposed to du'u), Lojban "fails" to mark the opaque context of {djica} as much as it does for {pensi}, because Lojban cannot distinguish event types from event tokens anymore than it can distinguish object types from object tokens.

> (by the way, I think that imaging is a different sort of thing > from thinking about and is more like picturing,, but I haven't done the > sort of work needed to really sort all these out). That at least some of > these arose from malglico copying of English subject raising just makes > the point the clearer — and harder to understand why we went that weay, > since I know that the problem was noted at the time).

That's assuming that there is a problem with not making the token-type distinction obligatory to begin with.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 08 of Sep., 2006 15:31 GMT On 9/8/06, John Cowan wrote: > Jorge Llamb?as scripsit: > > > A two way absolute distinction is flawed, but type-token is a relative > > distinction. Nothing is a type or a token in an absolute sense. > > I don't understand this. Something can be a type only if it has a > class of tokens that it represents; a specific elephant (Jumbo, say; > http://en.wikipedia.org/wiki/Jumbo ) is not a type of anything, > because he does not represent anything, non-metaphorically speaking.

Well, do you at least accept that any type can be a token of some other type? This is really all that is needed to escape the Quinean paradox. Then there is no problem with the components of a type being at the same time tokens of some type and types with their own tokens.

As for ultimate tokens, I would not have a problem saying (for some purposes at least) that for example the stages of Jumbo are tokens of the type Jumbo. For some purposes it is useful to consider baby Jumbo and grown-up Jumbo as two different objects, but both are also Jumbo.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 08 of Sep., 2006 16:10 GMT posts: 2388

> On 9/7/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > In the sense that you have no problem with using one and the same > > > expression to refer to events type or token, but you do have a problem > > > with using one and the same expression to refer to objects type or token. > > > > Well, actually, now that I have gotten clear on the distinction, I feel the > > same qualm about using the same term in the case of events. > > OK, that at least makes some more sense to me. > > Even though Lojban makes no structural distinction between types and > tokens, in many cases it is still pragmatically straightforward to tell which > is meant. For example, since a type is a singleton, any quantified term > has to be at least Griceanly interpreted as involving tokens, since > quantifying over a singleton is pointless. And given that we have > available, through the use of quantifiers, a fairly reliable indication that > tokenness is involved, the absence of any quantifier is at least suggestive > of typeness, or at least that the type-token distinction is not of interest.

I am not too happy with leaving the token/type distinction always up to context in the unquantified cases. And, indeed, I would always take something like {lo gerko} as about tokens (as you do on at least some occasions). But that leaves types unaccounted for, partly becaue we are not quite sure what a type is. Of course, different types of types (or types of different sorts of things) may themselves be different sorts of things: a particular-book is a structure of sentence types, presumably, while dog-type is something else — although I suppose it might be a structure of dog-part types. That latter seems to get us into a regress with the added problem of defining structures, so I'd as soon not go there. So, if we want to make a token/type distinction at least sometimes overt we need a device — and I don't think just always using whatever the Lojban is for "token" and "type" is quite what is wanted (it is too long, for one thing). Note also that quantifiers get only relative tokens, they ,may point to further types if that is where we start: three event types of winning a race, for example, still need not be occurrences.

> > That is to say, for careful discussion. In just normal > > vulgar speech I figure that context etc. will take care of what is going one > > and, if not, then later correction in the conversation will preven seriously > > bad things from happening. Of course, I think that the distinction needs > > to be made in Lojban on a more regular basis, though it is not > > clear just what the limits are (clearly I think that wants and other > > opaque contexts are cases for the distinction). > > But in the current language we don't have any structural way of > making the distinction. (We can always make the distinction by > explaining out what we mean, of course, but presumably you are > talking about doing it within the grammar.)

Yes, that is a central problem (if types are not some already recognized sort of thing in the abstraction list, NU?)

> > > But for the purposes of this discussion > > > (i.e. the "shove it into an event" solution to intensional contexts), the first > > > disagreement seems like the relevant one. > > To be careful, remember that this is not a "shove it all into events" proposal > > but rather, a "always use the appropriate abstraction" proposal (well, not > > exactly a proposal since it is the defacto position, labeit poorly executed). > > (Maybe you mean the de jure position, since surely it is not the de facto > one. It is not even clear that it is the de jure one either. Maybe the position > that reflects your understanding of the original intent.)

Well, it is the position of the wordlist — baselined and all, and not yet changed. that it was not carried through very well and is regularly not followed seem to me to merely be evidence of the pervasiveness of malglicitude, since the errors always follow English usage rather than some other source.

> Anyway, what I have been discussing all along in this thread is the > shove-it-into-an-event solution, as presented for example for {djica}. My > point has been that any problem that objects may have in intensional > contexts are replicated for events, because events have the same potential > type-token subsumption tree that objects have.

Except, I would (and always did more or less without noticing it) take events as types and objects (when referred to by {lo broda} expressions, at least) as about tokens. Also, we need to be careful here: if the program of event (etc.) were carrid through completely, there would be no cases of predicates with opaque place; all opacity would be shifted to the general principle that the insides of abstractions are opaque — which is a quite reasonable view to take (abstractions being transsworld or what have you and the problems with opacity being world-bound). As I have said before, if you want to use object types instead of event (types), this will work, too. BUT the object types must be marked as such, and by all coherent accounts to date {lo broda} is not so marked. Notice that, once the principle is established, not much change in appearance is involved: {tu'a lo gerku} is vague at least about what abstraction is involved — it might be dog types as well as events.

> If other entities with no tokens, such as propositions, are used > then matters might be different, but I'm comparing objects with events and > I'm saying that shoving objects into events doesn't escape the intensional > context issues, it only transfers them from the objects to the events.

Are you sure that propositions have no tokens? It seems to me that they tree up in much the same way and come to final token in a sentence (which is a token of both a proposition and a sentence type). But that may be controversial. The objects for events doesn't work, for me, because, for me, events are types and objects are ultimate tokens. My, things are getting clearer after only ten years or so at this.

> > > Wanting x and wanting that x both > > seem to be subjunctive and counter factual. Believing that x and that lot > > are prettry clearly not counterfactual (you can believe what is true, after all, > > but you cannot want what you believe you have). > > If you may want what you already have (not knowing that you have it), then > that would mean wanting is not necessarily counter factual, wouldn't it? Is > factuality defined by what really is, or by what the wanter believes to be?

Yeah, that is what I finally came down for, although I am not wildly enthusiatic about it -- except in my "it's all propositions" mode.

> > Thinking about x and the like are precisely the area where Lojban fails > > to mark opaque contexts, so bringing them in is not terribly useful at > > the moment > > By requiring events in the x2 of {djica} (as opposed to du'u), Lojban "fails" > to mark the opaque context of {djica} as much as it does for {pensi}, because > Lojban cannot distinguish event types from event tokens anymore than it > can distinguish object types from object tokens.

No, it doesn't mark the place because, if the proposal were carried through, the place would not be opaque. The opacity would be shifted to the abstraction, and all the usual laws would apply at that place (but for the abstraction of course: there is an event that you wnat and two identical events could be intersubstitued (but it is hard to get two identical events). I take it that being an abstraction is a sufficient condition for being a type.

> > (by the way, I think that imaging is a different sort of thing > > from thinking about and is more like picturing,, but I haven't done the > > sort of work needed to really sort all these out). That at least some of > > these arose from malglico copying of English subject raising just makes > > the point the clearer — and harder to understand why we went that way, > > since I know that the problem was noted at the time). > > That's assuming that there is a problem with not making the token-type > distinction obligatory to begin with.

Well, whether that is the best way to describe the problem certainly was not obvious back then; we did it in terms of opaque contexts and just simply missed some (or played some sort of game about the opacity of the thought carrying over to the raised subject — which was then unmarked). I am not sure that token/type is the only — or even best — way to describe the problem and its solution. However, it is the way that had led to the present clarification, so let's continue to use if for a while and see what results.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 08 of Sep., 2006 16:15 GMT posts: 2388

> On 9/8/06, John Cowan wrote: > > Jorge Llamb?as scripsit: > > > > > A two way absolute distinction is flawed, but type-token is a relative > > > distinction. Nothing is a type or a token in an absolute sense. > > > > I don't understand this. Something can be a type only if it has a > > class of tokens that it represents; a specific elephant (Jumbo, say; > > http://en.wikipedia.org/wiki/Jumbo ) is not a type of anything, > > because he does not represent anything, non-metaphorically speaking. > > Well, do you at least accept that any type can be a token of some > other type? This is really all that is needed to escape the Quinean > paradox. Then there is no problem with the components of a type being > at the same time tokens of some type and types with their own > tokens.

This is surely the case, though I am not sure it solves Quine's paradox (which I am also not sure is really a paradox — always a problem with Quine).

> As for ultimate tokens, I would not have a problem saying (for some > purposes at least) that for example the stages of Jumbo are tokens > of the type Jumbo. For some purposes it is useful to consider baby > Jumbo and grown-up Jumbo as two different objects, but both are > also Jumbo.

I don't see why this relation has to be called a token/type one — stages and parts seem clearer ways of putting, for all the formal similarities. Of course, I think types are abstract while parts and stages involve (the very unabstract) spatio-temporal continuities.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 08 of Sep., 2006 18:54 GMT On 9/8/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > > But in the current language we don't have any structural way of > > making the distinction. (We can always make the distinction by > > explaining out what we mean, of course, but presumably you are > > talking about doing it within the grammar.) > > Yes, that is a central problem (if types are not some already recognized > sort of thing in the abstraction list, NU?)

What is wanted here is something that for a given predicate {broda}, will return something that fills the x1 of broda. There is no NU such that {lo NU (ce'u) broda cu broda} is always the case, except for And's {poi'i}. But, while {poi'i} is useful because it allows more flexibility in selecting the required place, we already have {lo} to do what it does directly.

> > Anyway, what I have been discussing all along in this thread is the > > shove-it-into-an-event solution, as presented for example for {djica}. My > > point has been that any problem that objects may have in intensional > > contexts are replicated for events, because events have the same potential > > type-token subsumption tree that objects have. > > Except, I would (and always did more or less without noticing it) take events > as types and objects (when referred to by {lo broda} expressions, at least) > as about tokens.

If there is no problem with event types being visible, occurring in the world, and generally having causal effects, then it is hard to see why would there be a problem with object types being visible, existing in the world and generally having causal effects.


> BUT > the object types must be marked as such, and by all coherent accounts > to date {lo broda} is not so marked.

If the object types must be marked as such, then so must the event types. If event types can be wanted and be visible, so can object types. If an unmarked {lo broda} is incoherent (and I don't think it is) then an unmarked {lo nu broda} would be equally incoherent. Indeed they are both cases of {lo }, and what the form of the selbri is should be irrelevant.


> The objects for events doesn't work, for me, > because, for me, events are types and objects are ultimate tokens. My, > things are getting clearer after only ten years or so at this.

What exactly does it mean to say that events are types? For something to be an event-type, it has to have event-tokens for which the event-type is the type, doesn't it?

> I take it that being an abstraction is a sufficient condition for being a type.

Does that mean the same as "being an abstraction is a sufficient condition for having tokens"? Does the number five, for example, have tokens?

It appears that the notion of "abstraction" is one of the most ill-defined in all of this. This doesn't seemt to help much.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 08 of Sep., 2006 20:08 GMT posts: 2388

> On 9/8/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > > But in the current language we don't have any structural way of > > > making the distinction. (We can always make the distinction by > > > explaining out what we mean, of course, but presumably you are > > > talking about doing it within the grammar.) > > > > Yes, that is a central problem (if types are not some already recognized > > sort of thing in the abstraction list, NU?) > > What is wanted here is something that for a given predicate {broda}, will > return something that fills the x1 of broda. There is no NU such that > {lo NU (ce'u) broda cu broda} is always the case, except for And's {poi'i}. > But, while {poi'i} is useful because it allows more flexibility in selecting > the required place, we already have {lo} to do what it does directly.

This brings us to the other thing that Lojban doesn't have but should (with, in this case, the excuse that we didn't learn about it until much later): the distributive collective distinction separate from {lo/loi}. At least I think that is what is wanted, since I can't really understand what you are saying here. There isn't anything that will satisfy the {X broda cu broda} requirement with guaranteed truth and without special modalities understood. {lo broda cu broda} literally is false when there are no brodas; it is usually used (of at all) understood to be under a definitional modality, a shorthand way of talking about essential features of brodaness (a boring one in this case). And nothing else is literally a broda except lo (ro) broda. Now brodatype is a broda in a sense like the distributive one: all of its ultimate tokens are brodas, definitionally (this would be a somewhat more direct way of saying what {lo broda cu broda} usually means. But at least occasionally we need to mark this different sort of predication or we will start thinging that brodatype is one of its ultimate tokens — notoriously not a good thing to do. I don't remember the details of And's {poi'i}, but if we need a marker for types, we probably have to go off into that word space to get it.

> > > Anyway, what I have been discussing all along in this thread is the > > > shove-it-into-an-event solution, as presented for example for {djica}. My > > > point has been that any problem that objects may have in intensional > > > contexts are replicated for events, because events have the same potential > > > type-token subsumption tree that objects have. > > > > Except, I would (and always did more or less without noticing it) take events > > as types and objects (when referred to by {lo broda} expressions, at least) > > as about tokens. > > If there is no problem with event types being visible, occurring in the world, > and generally having causal effects, then it is hard to see why would there be > a problem with object types being visible, existing in the world and generally > having causal effects.

OK, so I am not as good as I like to think about keeping events and occurrences separate: what is visible and actually causal, as opposed to causally related, are occurrences. just as what are visible, etc. for objects are the objects themselves not the types. Well, there is no problem with events and object types existing in the world.

> > > BUT > > the object types must be marked as such, and by all coherent accounts > > to date {lo broda} is not so marked. > > If the object types must be marked as such, then so must the event types. They already are; that is what {nu} does. That we occasionally misuse it (though the misuse seems to be more in talking about it thanin using it in Lojban) doesn't means it doesn't have that function. That is, events are types; occurrences are tokens.

> If event types can be wanted and be visible, so can object types.

Very true, but they aren't visible and are at present wanted either, though that latter presents no problems that I can see.

> If an > unmarked {lo broda} is incoherent (and I don't think it is) then an unmarked > {lo nu broda} would be equally incoherent. Indeed they are both cases of > {lo }, and what the form of the selbri is should be irrelevant.

Neither is a problem: one refers to dogs and the other refers to events (types). There is only a problem if you use {lo broda} to refer to both types and ultimate tokens, as you surely seem to be doing. > > > The objects for events doesn't work, for me, > > because, for me, events are types and objects are ultimate tokens. My, > > things are getting clearer after only ten years or so at this. > > What exactly does it mean to say that events are types? For something to > be an event-type, it has to have event-tokens for which the event-type is the > type, doesn't it?

Yes, butr out of habit (and the way Lojban is set up), event tokens are called occurrences to avoid confusion (and constantly repeating "-type" and "-token"). It would be nice to have a word for object types generally, say "kinds."

> > I take it that being an abstraction is a sufficient condition for being a type. > > Does that mean the same as "being an abstraction is a sufficient condition > for having tokens"? Does the number five, for example, have tokens?

Yup; any pentad.(Admittedly, i is a bit harder to argue for.)

> It appears that the notion of "abstraction" is one of the most > ill-defined in all > of this. This doesn't > seemt to help much.

Yeah, it is probably worse than token/type (having been around longer and so gotten more said about it). Happily, we aren't into the metaphysical problems — which make up most of the argument. I am not concerned really about what the various kinds of abstractions are — you can eventually make do with just about anything; I am concerned with how they affect grammar and semantics. So, given my heritage and this interest, I am more or less content with a Montagovian explication, mainly in terms of functions from worlds to their contents (not that I particularly believe in worlds either, but it is a handy schematism). In Lojban, an abstraction is something referred to using an abstraction operator (I don't think they are all in NU), which is why, if we are going to have types as significant items, they need an operator (or to be assigned to an existing one — in fact all of the abstractions may serve as types for some things, I think). (Neat that the site where the discussion is is named for Plato, since he got the abstraction ball rolling — though it was around, less clearly, before.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 08 of Sep., 2006 21:14 GMT On 9/8/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > On 9/8/06, John E Clifford wrote: > > > > Yes, that is a central problem (if types are not some already recognized > > > sort of thing in the abstraction list, NU?) > > > > What is wanted here is something that for a given predicate {broda}, will > > return something that fills the x1 of broda. There is no NU such that > > {lo NU (ce'u) broda cu broda} is always the case, except for And's {poi'i}. > > But, while {poi'i} is useful because it allows more flexibility in selecting > > the required place, we already have {lo} to do what it does directly. > > This brings us to the other thing that Lojban doesn't have but should > (with, in this case, the excuse that we didn't learn about it until much later): > the distributive collective distinction separate from {lo/loi}. At least I think > that is what is wanted, since I can't really understand what you are saying > here.

I'm saying that one minimum requirement that any X that purports to refer to the broda-type must satisfy is that {X cu broda} holds. There is no official NU that meets that requirement.

(Long ago, before xorlo, I used to use {lo'e broda} for this, but having to constantly explain that this was meant as the broda-type and not as a typical broda was a pain. {lo broda} normally requires no such explaining.)

> I don't remember the details of And's {poi'i}, but if we need a marker for > types, we probably have to go off into that word space to get it.

Here:


> > If there is no problem with event types being visible, occurring in the world, > > and generally having causal effects, then it is hard to see why would there be > > a problem with object types being visible, existing in the world and generally > > having causal effects. > > OK, so I am not as good as I like to think about keeping events and > occurrences separate: what is visible and actually causal, as opposed > to causally related, are occurrences. just as what are > visible, etc. for objects are the objects themselves not the types.

The point is that predicate places that take {lo nu ...} are mostly of the kind that take occurrences. The ones such as x2 of djica with a preference for event-types are in the minority.

> > If the object types must be marked as such, then so must the event types. > > They already are; that is what {nu} does.

If that's the case, then you would need some marker for the majority of uses of {nu} that would have to be unmarked for type.

> That we occasionally misuse it (though the misuse seems > to be more in talking about it thanin using it in Lojban) doesn't > means it doesn't have that function. That is, events are types; > occurrences are tokens.

And if {lo nu ...} can't be used to refer to occurrences, what would we use to refer to occurrences?

> > If an > > unmarked {lo broda} is incoherent (and I don't think it is) then an unmarked > > {lo nu broda} would be equally incoherent. Indeed they are both cases of > > {lo }, and what the form of the selbri is should be irrelevant. > > Neither is a problem: one refers to dogs and the other refers to events > (types). There is only a problem if you use {lo broda} to refer to both > types and ultimate tokens, as you surely seem to be doing.

I am certainly using both {lo broda} and {lo nu broda} to refer indifferently to types or tokens. You are doing it only with {lo nu broda}. Why is there a problem in what I do but not in what you do?

> In Lojban, an abstraction is something referred to using an abstraction > operator (I don't think they are all in NU), which is why, if we are going > to have types as significant items, they need an operator (or to be > assigned to an existing one — in fact all of the abstractions may serve > as types for some things, I think).

I'll leave that to those interested in having the distinction be of significance. For my part, I am content with dispensing with ultimate tokens altogether. If it works for events, there is no reason for it not to work for objects.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 08 of Sep., 2006 23:42 GMT posts: 2388

> On 9/8/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > On 9/8/06, John E Clifford wrote: > > > > > > Yes, that is a central problem (if types are not some already recognized > > > > sort of thing in the abstraction list, NU?) > > > > > > What is wanted here is something that for a given predicate {broda}, will > > > return something that fills the x1 of broda. There is no NU such that > > > {lo NU (ce'u) broda cu broda} is always the case, except for And's {poi'i}. > > > But, while {poi'i} is useful because it allows more flexibility in selecting > > > the required place, we already have {lo} to do what it does directly. > > > > This brings us to the other thing that Lojban doesn't have but should > > (with, in this case, the excuse that we didn't learn about it until much later): > > the distributive collective distinction separate from {lo/loi}. At least I think > > that is what is wanted, since I can't really understand what you are saying > > here. > > I'm saying that one minimum requirement that any X that purports to > refer to the broda-type must satisfy is that {X cu broda} holds. There is > no official NU that meets that requirement.

But, a broda type is generally not a broda, for all sorts of reason. To say it is has and expect that what you say is true, therefore, to be some non-literal reading (not even collective v distributive if you are really after types). I take it to be definitional, not that that helps much since Lojban has no mark for that, unless {ca'e} can be pressed into service (there are a couple other words in the rhetorical categories that might more usefully — in addition -- function as modals.

> (Long ago, before xorlo, I used to use {lo'e broda} for this, but having to > constantly explain that this was meant as the broda-type and not as > a typical broda was a pain. {lo broda} normally requires no such explaining.) > > > I don't remember the details of And's {poi'i}, but if we need a marker for > > types, we probably have to go off into that word space to get it. > > Here: > > > > > If there is no problem with event types being visible, occurring in the world, > > > and generally having causal effects, then it is hard to see why would there be > > > a problem with object types being visible, existing in the world and generally > > > having causal effects. > > > > OK, so I am not as good as I like to think about keeping events and > > occurrences separate: what is visible and actually causal, as opposed > > to causally related, are occurrences. just as what are > > visible, etc. for objects are the objects themselves not the types. > > The point is that predicate places that take {lo nu ...} are mostly > of the kind that take occurrences. The ones such as x2 of djica with a > preference for event-types are in the minority.

That isn't problematic, that I can see, since a relation with an occurrence is also a (more attenuated) relation with a (indeed, several) type. It might be tidier to do it another way, but this will work (and it probably cuts off some of the forgotten opaque cases as well. By the way, how sure are you of your statistics: a look at the first 30 cases on the wordlist showed 16 clearly type, 8 clearly occurrence and 6 with no obvious preference. Of course this is hardly a random sample but there doesn't seem to be any reason to think the distribution is affected by alphabetic place.


> > > If the object types must be marked as such, then so must the event types. > > > > They already are; that is what {nu} does. > > If that's the case, then you would need some marker for the majority of > uses of {nu} that would have to be unmarked for type.

That is a problem, of course, if we are going to stick with the token/type distinction. It is a nice complement to the problem of marking object types.

> > That we occasionally misuse it (though the misuse seems > > to be more in talking about it thanin using it in Lojban) doesn't > > means it doesn't have that function. That is, events are types; > > occurrences are tokens. > > And if {lo nu ...} can't be used to refer to occurrences, what would > we use to refer to occurrences?

Well, what do you use to refer to types? There is a parity of problems here. Of course, I suppose that what we will continue to do as long as no problem arises is use the most convenient form and let context decide. But that won't do as a permanent solution.

> > > If an > > > unmarked {lo broda} is incoherent (and I don't think it is) then an unmarked > > > {lo nu broda} would be equally incoherent. Indeed they are both cases of > > > {lo }, and what the form of the selbri is should be irrelevant. > > > > Neither is a problem: one refers to dogs and the other refers to events > > (types). There is only a problem if you use {lo broda} to refer to both > > types and ultimate tokens, as you surely seem to be doing. > > I am certainly using both {lo broda} and {lo nu broda} to refer indifferently > to types or tokens. You are doing it only with {lo nu broda}. Why > is there a problem in what I do but not in what you do?

Well, first of all, I have the problem, too — or rather its mirror image. And then there is the fact that at the crucial point — opaque contexts — what I use is primarily the right thing (type) and yours primarily the wrong thing (token). You may *mean* it for a type, but it looks to the logician like a token — with the usual bad results.

> > In Lojban, an abstraction is something referred to using an abstraction > > operator (I don't think they are all in NU), which is why, if we are going > > to have types as significant items, they need an operator (or to be > > assigned to an existing one — in fact all of the abstractions may serve > > as types for some things, I think). > > I'll leave that to those interested in having the distinction be of > significance. > For my part, I am content with dispensing with ultimate tokens altogether. > If it works for events, there is no reason for it not to work for objects. But as you note, it doesn't work for events except accidentally.

By the way, another reason why events are better than object types is that they can handle {le broda} and {la brod} without any adaptation. It is at least not obvious that there is a type behind either of these expressions to claim to be using. If there is, however, it will be different enougfhto require some additional apparatus. >


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Mon 11 of Sep., 2006 15:04 GMT On 9/8/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > I'm saying that one minimum requirement that any X that purports to > > refer to the broda-type must satisfy is that {X cu broda} holds. There is > > no official NU that meets that requirement. > > But, a broda type is generally not a broda, for all sorts of reason.

And similarly a broda token is generally not the broda.

Given the object language predicate {broda}, it can truthfully apply in the object language both to what we would call broda tokens in the metalanguage and to what we would call the broda type in the metalanguage. In the object language, the distinction is not generally made (until you want to make it, of course, in which case you use the appropriate predicate).

We can't force the distinction to always be made. If we eliminate the predicate {broda} from the object language in favour of two separate predicates {broda-type} and {broda-token}, all we have done is duplicate the number of predicates in the object language but we have not made any advances, because now the new predicate {broda-token} will have its own type and its tokens from a metalinguistic perspective, and so will the new predicate {broda-type}. We can try to fix this by duplicating the predicates of the object language again, introducing {broda-token-token}, {broda-token-type}, {broda-type-token}, {broda-type-type} but then this process will never end.


> > The point is that predicate places that take {lo nu ...} are mostly > > of the kind that take occurrences. The ones such as x2 of djica with a > > preference for event-types are in the minority. > > That isn't problematic, that I can see, since a relation with an occurrence > is also a (more attenuated) relation with a (indeed, several) type.

I can agree with you about that. Where we disagree is in that you think this is only the case for events, whereas I think this is the case for anything else as well.

> By the way, how sure are you of your statistics: a look at the first 30 > cases on the wordlist showed 16 clearly type, 8 clearly occurrence > and 6 with no obvious preference.

Let's see:

type-prone: bajra-4, bandu-3, banxa-3, banzu-2, batke-3, bende-4, bilga-2, bilni-3, binra-3, binra-4, bongu-2, bredi-2, cabra-2, cafne-1, cedra-2, certu-2

token-prone: badri-2, balvi-1, balvi-2, bandu-1, bapli-2, basna-3, bebna-2, betri-1, bradi-3, cabna-1, cabna-2, cacra-1, catra-3, cfari-1

Maybe a slight advantage to the type-prones then, although many in either group could easily shift. But in any case, the important point is that sometimes we do want the tokens.


> > > That we occasionally misuse it (though the misuse seems > > > to be more in talking about it thanin using it in Lojban) doesn't > > > means it doesn't have that function. That is, events are types; > > > occurrences are tokens. > > > > And if {lo nu ...} can't be used to refer to occurrences, what would > > we use to refer to occurrences? > > Well, what do you use to refer to types?

I use {lo broda}. I make no grammatical distinction (in the object language) between the metalinguistic types and tokens.

> There is a parity of problems here. Of course, I > suppose that what we will continue to do as long as no problem arises > is use the most convenient form and let context decide. But that won't > do as a permanent solution.

We only have one form to use in the object language, so there isn't a most convenient form to choose. What context will determine is the interpretation that you have to give to that form in the metalanguage if you decide to have tokens and type in the metalanguage.

> > I am certainly using both {lo broda} and {lo nu broda} to refer indifferently > > to types or tokens. You are doing it only with {lo nu broda}. Why > > is there a problem in what I do but not in what you do? > > Well, first of all, I have the problem, too — or rather its mirror image. And > then there is the fact that at the crucial point — opaque contexts — what I > use is primarily the right thing (type) and yours primarily the wrong thing > (token).

No, mine is not primarily token (nor primarily type). There is no primacy of token or type from my perspective.

> You may *mean* it for a type, but it looks to > the logician like a token — with the usual bad results.

Well, the logician should take a better look then, because there is nothing in Logic that requires it to be interpreted as a token.


> By the way, another reason why events are better than object types > is that they can handle {le broda} and {la brod} without any adaptation. > It is at least not obvious that there is a type behind either of these > expressions to claim to be using. If there is, however, it will be > different enougfhto require some additional apparatus.

I'm sure there is. For example:

"The flag of the United Kingdom of Great Britain and Northern Ireland is sometimes called the Union Jack. This red, white, and blue flag was first used in 1801." that's a type usage of "this red, white, and blue flag".

"This summer the Union Jack was flying in every street here." that's a type usage of the name "the Union Jack".

"Union Jacks were everywhere: at Wembley, in the streets, at parties, and festooned around pubs and clubs." that is (or could be taken as) a token usage of the name "Union Jack".

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Mon 11 of Sep., 2006 17:18 GMT posts: 2388

> On 9/8/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > I'm saying that one minimum requirement that any X that purports to > > > refer to the broda-type must satisfy is that {X cu broda} holds. There is > > > no official NU that meets that requirement. > > > > But, a broda type is generally not a broda, for all sorts of reason. > > And similarly a broda token is generally not the broda.

What is the force of "the" here? A broda token is a broda, but there will usually be others. > Given the object language predicate {broda}, it can truthfully apply in > the object language both to what we would call broda tokens in the > metalanguage and to what we would call the broda type in the metalanguage. > In the object language, the distinction is not generally made (until you want > to make it, of course, in which case you use the appropriate predicate).

Well, no. Of courese the predicate {broda} doesn't apply to anything in the metalanguage, since it is not a metalanguage predicate. However, in the metalanguage, we can say that it applies to (has as its referents) just brodas (ultimate tokens). Broda types don't enter in, they are not in the extension of {broda}. Whether a broda type is a sense of {broda} or a set (of one sort or another) of brodas, or the Urgoo out of which all brodas are sliced or something else, it is not assigned (as a general rule)to the extension of {broda}. The fact that the distiction is not regularly made in the object language means that cases where what we say is meant to be about broda types rather than brodas have to be understood in some special way — as a different relation between term and predicate than the usual being among the referents. So we need a way to mark this shift — even if we don't always use it. We can use an abstraction marker to indicate that the term refers to types, or we can mark the predication as being different (as collective and distributive might be marked) or we can put the whole sentence under a modal to mark the shift, or we could introduce a new (or reuse an old) gadri. Any of these (and there are problably others) will work just fine and all appear to be optional but readily available. We might also introduce new prediates that applied directly to types (or a general scheme for creating them existing ones"broda-type" from "broda" for example), both for classifying them (as the example) and for saying things about them.


> We can't force the distinction to always be made.

Well, we could try, but the odds against success are enormous.

> If we eliminate the > predicate {broda} from the object language in favour of two separate > predicates {broda-type} and {broda-token},

As noted, {broda-token} is alrady in the language,, namely {broda}. So, we are merely adding {broda-type).

> all we have done is duplicate > the number of predicates in the object language but we have not made > any advances,

We we have explcitly separated token talk from type talk, which is a considerable advance -- though this may nopt be the most efficient way to do it.

> because now the new predicate {broda-token} will have its > own type and its tokens from a metalinguistic perspective, and so will the > new predicate {broda-type}.

Well, yes, {broda-token}, i.e., {broda}, will continue to refere (somewhat more indirectly perhaps) to brodas and there will now be broda-token types, which might even (but don't obviously have to be) different from broda-types. Similarly, there is not reason th think that broda-type tokens will be different from broda-types. So all this will dd will be broda-type types, which, I suspect, were already a possibility in the original scheme (Since most metalanguages don't have types — in this sense — as a category, it is a little hard to say what would be in a metathory which contained them. Perhaps, if you would layout what that metaptheory would be like it would help a bit.)


> We can try to fix this by duplicating the > predicates of the object language again, introducing {broda-token-token}, > {broda-token-type}, {broda-type-token}, {broda-type-type} but then this > process will never end.

This just seems garbled to me. If any of this will work at all, then {broda} refers to brodas (ultimate tokens}. In some cases it refers to each of them individually, in others to some collections of them. It does not refer to the types at all (except, of course, for predicates like "is a type" and so on). Indeed, since it is not clear what types in general are, it is not clear where — or whether — they fit into the general scheme of things. At best, a type is like a giant collective, of which the predicate may be predicated distributively. More likely, it is aomw aort of abstract entity, akin to a sense, so that sentences in which , for example, {lo broda} refer to the type have to be considered in material mode, about the relations among senses or other abstract entities. It is clearly in this latter way that something like {lo broda cu broda} is ALWAYS true. Of course, if there are broda, then saying {lo broda cu broda} is true but pretty pointless (it is not too pointed even in the material mode, unlike a definition, say). Lojban doesn't have good devices for making the material mode (or a whole bunch of otheres, come to that), so, if we are going to allow the usual ambiguity betweeen token and type, we have to rely on context or rather complex rewriting to resolve the ambiguity when a problem arises. Since this ambiguity is pervasive, it would seem a good idea to get a device for dealing with it from the get-go and one simple enought to use all the time when we are aware of the difference and it might make a difference. One such device would be another abstractor for types (assuming they are not just properties, say); another would be a mark of yet another kind of predication (although we don't have marks for the kinds we recognize already yet); a third is to use a new gadri (or to finally find a use for some old one that has fallen into dissuetude). And finally we can get a modal that points to abstract operators at work. Finally, as for the infintie regress, I don't see it — why won't broda-token-tokens be just brodas again? > > > > The point is that predicate places that take {lo nu ...} are mostly > > > of the kind that take occurrences. The ones such as x2 of djica with a > > > preference for event-types are in the minority. > > > > That isn't problematic, that I can see, since a relation with an occurrence > > is also a (more attenuated) relation with a (indeed, several) type. > > I can agree with you about that. Where we disagree is in that you think this > is only the case for events, whereas I think this is the case for anything else > as well.

No, I think that every token (for all practical purposes) is a token of many types. And so, I suppose, every predicate is a relation not merely with tokens but also with those types, albeit attentuated. I am afraid I have lost the point here, since yo seem to admit the crucial part, that {djica}, for example, favors types over tokens. (We will skip over the Montague point that all predicates apply to sense first of all and onlyderivatively to tokens - referents, since that is at a very deep level not at issue here).

> > By the way, how sure are you of your statistics: a look at the first 30 > > cases on the wordlist showed 16 clearly type, 8 clearly occurrence > > and 6 with no obvious preference. > > Let's see: > > type-prone: bajra-4, bandu-3, banxa-3, banzu-2, batke-3, bende-4, bilga-2, > bilni-3, binra-3, binra-4, bongu-2, bredi-2, cabra-2, cafne-1, cedra-2, certu-2 > > token-prone: badri-2, balvi-1, balvi-2, bandu-1, bapli-2, basna-3, bebna-2, > betri-1, bradi-3, cabna-1, cabna-2, cacra-1, catra-3, cfari-1

We disagree on a couiple of thse, but I expect that is just evidence that they could go either way.

> Maybe a slight advantage to the type-prones then, although many in either > group could easily shift. But in any case, the important point is that > sometimes we do want the tokens.

Of course, which is why I have a problem, since I lack an explicit (and compact — I can, of course get there eventually) way of specifying an occurrence of an event. > > > > > That we occasionally misuse it (though the misuse seems > > > > to be more in talking about it than in using it in Lojban) doesn't > > > > means it doesn't have that function. That is, events are types; > > > > occurrences are tokens. > > > > > > And if {lo nu ...} can't be used to refer to occurrences, what would > > > we use to refer to occurrences? > > > > Well, what do you use to refer to types? > > I use {lo broda}. I make no grammatical distinction (in the object language) > between the metalinguistic types and tokens.

Well, types and token are not just in the metalanguage; they are what we talk about in the object language and so we need different ways of talking about them at least sometimes — since what we can (and do) say about them is different and we can't always do it using the same sorts of words.

> > There is a parity of problems here. Of course, I > > suppose that what we will continue to do as long as no problem arises > > is use the most convenient form and let context decide. But that won't > > do as a permanent solution. > > We only have one form to use in the object language, so there isn't a most > convenient form to choose. What context will determine is the interpretation > that you have to give to that form in the metalanguage if you decide to have > tokens and type in the metalanguage.

Well, we have only one *convenient* form to use in the object language, in Lojban. we have other devices already — and can propose still more — to use. The point is that we need to come up with some devices to use regularly — and to get in the habit of using in known tricky places.

> > > I am certainly using both {lo broda} and {lo nu broda} to refer indifferently > > > to types or tokens. You are doing it only with {lo nu broda}. Why > > > is there a problem in what I do but not in what you do? > > > > Well, first of all, I have the problem, too — or rather its mirror image. And > > then there is the fact that at the crucial point — opaque contexts — what I > > use is primarily the right thing (type) and yours primarily the wrong thing > > (token). > > No, mine is not primarily token (nor primarily type). There is no primacy of > token or type from my perspective.

Then you haven't been dealing with Lojban (probably for a long time). It is perfectly clear that {lo broda} refers primarily to brodas, the ultimate tokens of broda-type. In fact, if it did not, we dould not get the whole token-type program off the ground. {lo} is after all, merely the insopecific version of {le} and {le} clearly refers to individual objects, not to object types (since there is no reason to believe there even is an object type corresponding to a {le} phrase.

> > You may *mean* it for a type, but it looks to > > the logician like a token — with the usual bad results. > > Well, the logician should take a better look then, because there is nothing > in Logic that requires it to be interpreted as a token.

Actually, there is — terms refer to objectds, i.e., ultimate tokens. But this requirement is not from Logic but from Lojban (a language you really ought to get to know). > > > By the way, another reason why events are better than object types > > is that they can handle {le broda} and {la brod} without any adaptation. > > It is at least not obvious that there is a type behind either of these > > expressions to claim to be using. If there is, however, it will be > > different enougfh to require some additional apparatus. > > I'm sure there is. For example: > > "The flag of the United Kingdom of Great Britain and Northern Ireland > is sometimes called the Union Jack. This red, white, and blue flag > was first used in 1801." > that's a type usage of "this red, white, and blue flag".

Certainly is; matches the first type use.

> "This summer the Union Jack was flying in every street here." > that's a type usage of the name "the Union Jack".

Yup.

> "Union Jacks were everywhere: at Wembley, in the streets, at parties, > and festooned around pubs and clubs." > that is (or could be taken as) a token usage of the name "Union Jack".

Well, it may be token relative to the earlier ones, but is still not an ultimate token. But, of course, none of these are on the point, which is about using event types rather than object types in {djica2), remember. And, of course, in the {le} case, we have no guarantee that what is called the type here really is a type for the things referred to the referents of {le broda} my not be tokens of a broda-type, after all. Nor is it perfectly clear that what we have here are cases of {le} altogether.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Mon 11 of Sep., 2006 20:18 GMT On 9/11/06, John E Clifford wrote: > Of courese the predicate {broda} doesn't apply to anything in > the metalanguage, since it is not a metalanguage predicate. However, > in the metalanguage, we can say that it applies to (has as its referents) > just brodas (ultimate tokens).

Since I don't have context independent ultimate tokens in my metalanguage, that doesn't help me much. For example, I can't tell without context whether the referent of {lo lerfu} are letter types or letter tokens. In fact, the very distinction between letter types and letter tokens makes little sense without a context, because anything will be a token of something or other and (in my view at least) anything can be a type of other things.


> > If we eliminate the > > predicate {broda} from the object language in favour of two separate > > predicates {broda-type} and {broda-token}, > > As noted, {broda-token} is alrady in the language,, namely {broda}. So, > we are merely adding {broda-type).

But for some predicates, such as {nu bajra}, you want it restricted to what in the metalanguage we would call the broda-type.


> (Since most metalanguages don't have types — in this > sense — as a category, it is a little hard to say what would be in a > metathory which contained them. Perhaps, if you would layout what > that metaptheory would be like it would help a bit.)

The metalanguage we are using now is English, isn't it? I take it we understand what we mean by the distinction between "this text contains 300 word tokens" and "this text contains 300 word types" in English. I'm saying that the Lojban {le dei selsku cu vasru cinono valsi} could mean either, we need context to tell which one it means. And we could also introduce {valsi sevzi} and {valsi se sevzi} (or whatever predicate is best for "type" and "token" in Lojban) to make matters more clear, but {lo valsi} by itself doesn't tell us which one it is. (Just as "word" in English doesn't either.)


> > > > The point is that predicate places that take {lo nu ...} are mostly > > > > of the kind that take occurrences. The ones such as x2 of djica with a > > > > preference for event-types are in the minority. > > > > > > That isn't problematic, that I can see, since a relation with an occurrence > > > is also a (more attenuated) relation with a (indeed, several) type. > > > > I can agree with you about that. Where we disagree is in that you think this > > is only the case for events, whereas I think this is the case for anything else > > as well. > > No, I think that every token (for all practical purposes) is a token of many > types. And so, I suppose, every predicate is a relation not merely with > tokens but also with those types, albeit attentuated. I am afraid I have lost > the point here, since yo seem to admit the crucial part, > that {djica}, for example, favors types over tokens.

I certainly agree that {djica} favors types over tokens. Where we disagree apparently is in what {lo broda} favors.

> > > > And if {lo nu ...} can't be used to refer to occurrences, what would > > > > we use to refer to occurrences? > > > > > > Well, what do you use to refer to types? > > > > I use {lo broda}. I make no grammatical distinction (in the object language) > > between the metalinguistic types and tokens. > > Well, types and token are not just in the metalanguage; they are what we > talk about in the object language and so we need different ways of talking > about them at least sometimes — since what we can (and do) say about > them is different and we can't always do it using the same sorts of words.

I have no problem with that, given the "at least sometimes". For those times where we do need to distinguish, a predicate meaning "x1 is a token of type x2" would be useful.


> It is perfectly clear that > {lo broda} refers primarily to brodas, the ultimate tokens of broda-type. > In fact, if it did not, we dould not get the whole token-type program off > the ground. {lo} is after all, merely the insopecific version of {le} and {le} > clearly refers to individual objects, not to object types (since there is no > reason to believe there even is an object type corresponding to a {le} phrase.

It is not clear to me that there is anything wrong with, for example:

{zo klama cu vasru mu lerfu}

or with

{rexa da lerfu le lojbo}

which use different token levels of {lerfu}.


> > > You may *mean* it for a type, but it looks to > > > the logician like a token — with the usual bad results. > > > > Well, the logician should take a better look then, because there is nothing > > in Logic that requires it to be interpreted as a token. > > Actually, there is — terms refer to objectds, i.e., ultimate tokens. But > this requirement is not from Logic but from Lojban (a language you really > ought to get to know).

Lojban, as far as I understand, does not come with a fixed ontology. It is quite flexible in that regard, as any general purpose language has to be.

> > > By the way, another reason why events are better than object types > > > is that they can handle {le broda} and {la brod} without any adaptation. > > > It is at least not obvious that there is a type behind either of these > > > expressions to claim to be using. If there is, however, it will be > > > different enougfh to require some additional apparatus. > > > > I'm sure there is. For example: > > > > "The flag of the United Kingdom of Great Britain and Northern Ireland > > is sometimes called the Union Jack. This red, white, and blue flag > > was first used in 1801." > > that's a type usage of "this red, white, and blue flag". > > Certainly is; matches the first type use. > > > "This summer the Union Jack was flying in every street here." > > that's a type usage of the name "the Union Jack". > > Yup. > > > "Union Jacks were everywhere: at Wembley, in the streets, at parties, > > and festooned around pubs and clubs." > > that is (or could be taken as) a token usage of the name "Union Jack". > > Well, it may be token relative to the earlier ones, but is still not an ultimate > token.

What would count as an ultimate token flag?

> But, of course, none of these are on the point, which is about using event > types rather than object types in {djica2), remember. And, of course, in > the {le} case, we have no guarantee that what is called the type here really > is a type for the things referred to the referents of {le broda} my not be > tokens of a broda-type, after all. Nor is it perfectly clear that what we have > here are cases of {le} altogether.

These examples were to show that, at least in English, names and definite descriptions can be used to refer to types. It is not clear to me why {le} and {la} could not be so used.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Tue 12 of Sep., 2006 01:38 GMT posts: 2388

> On 9/11/06, John E Clifford wrote: > > Of courese the predicate {broda} doesn't apply to anything in > > the metalanguage, since it is not a metalanguage predicate. However, > > in the metalanguage, we can say that it applies to (has as its referents) > > just brodas (ultimate tokens). > > Since I don't have context independent ultimate tokens in my metalanguage, > that doesn't help me much. For example, I can't tell without context > whether the referent of {lo lerfu} are letter types or letter tokens. In fact, > the very distinction between letter types and letter tokens makes little > sense without a context, because anything will be a token of something > or other and (in my view at least) anything can be a type of other things. > Well, we just speak different languages here (and have very different metalanguages as well. I am not sure whether my metalanguage has types and tokens at all, but if it does , it clearly has ultimate tokens (typically spatio-temporally discrete and continuous objects). Indeed, since the token/type distinction is from uses involving linguistic expressions it seem always to require ultimate tokens (particular marks of carbon on a particular piece of paper, spatio-temporally located, etc.). For the linguistic usage, types can just a readily be sets of ultimate tokens as any other interpretation (none works perfectly). Then the relation between one type and another lower one, is different from that between a type and a token (inclusion v membership — though we could use L-sets and collapse those together, too). I get the impression that what you want your token/type distinction to do cannot be reduced to sets (of either sort). So, to prevent confusion at least (and also because this attempt to generalize the notion is getting hopelessly messy), I suggest we drop the toke/type talk and try to figure out what we really want. I am fairly sure that the token/type distinction — in anything like its classic form — will not cover all the cases we want to deal with — either the opaque contexts or the cases where we shift back and forth between reference to specifics and more general claims. In fact, looking at general claims might be useful, since what is going on there is fairly clear: a simple sentence is used to make a complex claim about a whole class of objects. The claim can be translateds only with difficulty (and some uncertainty) back to a claim about individual objects, although the theory is that such atranslation could always be done eventually. These general claims often take a form indistinguishable from claims about particular individuals — at best distinguished in English by words like "generally" and, so far as I can tell, not distinguished in Lojban in any systematic way at all. Thus, the case of not distinguishing individual from generality is part of a general pattern (which we knew). In the one case, at least, the solution is to recognize the complexity of the underlying situation (say the behavior of countless anonymous individuals over vast spans of time) and simply indicate somehw that that is what is going one, rather than a fact about a delimited bunch of individuals in a narrow space-time area. (Or, of course, we could mark that it is the latter that is going one, not the former. I once suggested using both {su'a} and {su'anai} for this job, allowing ambiguous cases to remain). This does not obviously solve the opaque context cases, which yield nicely to the usual abstractions — events and properties and propositions. Leaving these cases ambiguous presents more serious problems, of course, largely because we do want to sometimes take most of these places as also transparent.


> > > If we eliminate the > > > predicate {broda} from the object language in favour of two separate > > > predicates {broda-type} and {broda-token}, > > > > As noted, {broda-token} is alrady in the language,, namely {broda}. So, > > we are merely adding {broda-type). > > But for some predicates, such as {nu bajra}, you want it restricted to > what in the metalanguage we would call the broda-type. > We would call it that in the object language, too, if we could, but our object language makes this hard (not conceptually, just lexically). Well, as I said, the token/type stuff is getting unduly messy, so lets say what is clearly the case, namely that {nu bajra} refers to abstract entities, events — but that is obvious from the fact that it is an abstract predicate. To get down to the occurrences from which this is abstracted takes a different move, but it is not a type to token move. {nu bajra} does not refer to a set, if for no other reason than that we currently have no way to refer to its members. (that problem being solved, we could go a long way with {nu bajra} being about a set of occurrences, but I think ultimately it would not quite work).

> > (Since most metalanguages don't have types — in this > > sense — as a category, it is a little hard to say what would be in a > > metathory which contained them. Perhaps, if you would lay out what > > that metaptheory would be like it would help a bit.) > > The metalanguage we are using now is English, isn't it? I take it we > understand what we mean by the distinction between "this text contains > 300 word tokens" and "this text contains 300 word types" in English. > I'm saying that the Lojban {le dei selsku cu vasru cinono valsi} could > mean either, we need context to tell which one it means. And we could > also introduce {valsi sevzi} and {valsi se sevzi} (or whatever predicate > is best for "type" and "token" in Lojban) to make matters more clear, > but {lo valsi} by itself doesn't tell us which one it is. (Just as "word" in > English doesn't either.) > Well, if English is our metalanguage (I would hope it was somewhat enriched), then, while I can tell, what is meant by "word token" and "word type", I am less than clear about what is meant by "dog token" (apparently not just a dog) and "dog type." A lot of what holds for linguistic objects (woords, letters, sentences, etc.) does not generalize in obvious ways to other kinds of things and what does gneralize is bettr dealt with under more traditional categories, where the moves have been pretty wewll worked out. Using a linguistic example doesn't help much, of course -- whilwe they are clear, it is not clear what we are to take away from that to the dog case (or elephant). What is clear, and related in a general sort of way, is that {lo xantu cu citka lo srasu} may be making a particular claim about some particular elephants and some particular clump of grass, or a general claim about elephants and grass or a middlesized claim about particular elephants and grass in general or (less plausibly in this case) a general claim about elphants and a particular clump of grass. Ordinarily, we can sort out which by context (as with the comment about the last alternative). When we cannot immediately, further developments may do it for us. But it would be nice to have a way that would always work when we need it and one small enough to be used when some worry about confusion occurred to the speaker without unduly straining the conversation with what might turn out to have been unneccessary. (Obviously, the {su'anai} won't do because of the intermediate cases). But when this is done, the problem of opaque contexts remains — except that we know how to solve that already — we just have to actually use it. > > > > > > The point is that predicate places that take {lo nu ...} are mostly > > > > > of the kind that take occurrences. The ones such as x2 of djica with a > > > > > preference for event-types are in the minority. > > > > > > > > That isn't problematic, that I can see, since a relation with an occurrence > > > > is also a (more attenuated) relation with a (indeed, several) type. > > > > > > I can agree with you about that. Where we disagree is in that you think this > > > is only the case for events, whereas I think this is the case for anything else > > > as well. > > > > No, I think that every token (for all practical purposes) is a token of many > > types. And so, I suppose, every predicate is a relation not merely with > > tokens but also with those types, albeit attentuated. I am afraid I have lost > > the point here, since you seem to admit the crucial part, > > that {djica}, for example, favors types over tokens. > > I certainly agree that {djica} favors types over tokens. Where we disagree > apparently is in what {lo broda} favors.

But that cannot be an issue: {lo broda} is about objects. I am not sure what it would mean to say it was about types though translating from the earlier I suppose it means that it is about properties, which is clearly wrong. What seems to be at issue is that {lo broda} is a natural form to use in general claims, in a way that {su'o broda} is not (though they are equivalent in particularities). Now, generalities are simple ways of saying complex things about objects, like statiistics and such notions as "typical." The general {lo xanto cu citka lo srasu} is about the referents of {lo xanto} (and {lo srasu} too) over large space and long periods of time, so not the particular {lo xanto} of a given occasion (in this way, you could argue that {lo xanto} here referred to its sense, the individual concept, which would fit right in with taking generality to be a modal concept, but misses the relation — vague as it is — to normal quantified sentences). In particular, though, it is not about some odd kind of object, a general elephant, anymore than a statistical statement is about an object, the average elephant, say. But this seems tto be what you are claiming. although you may just be insisting that here we need to recognize the modality of generality and so have {lo xantu} referring to its sense, which would, to be sure, tie up with the opaque cases, where that is definitely what is going on (or, at least, that is the best explanation within the ontology of Lojban — which does not have clear individual concepts, alas). But individual concepts don't seem to me to be easily gotten to by analogy with types.

> > > > > And if {lo nu ...} can't be used to refer to occurrences, what would > > > > > we use to refer to occurrences? > > > > > > > > Well, what do you use to refer to types? > > > > > > I use {lo broda}. I make no grammatical distinction (in the object language) > > > between the metalinguistic types and tokens. > > > > Well, types and token are not just in the metalanguage; they are what we > > talk about in the object language and so we need different ways of talking > > about them at least sometimes — since what we can (and do) say about > > them is different and we can't always do it using the same sorts of words. > > I have no problem with that, given the "at least sometimes". For those times > where we do need to distinguish, a predicate meaning "x1 is a token > of type x2" would be useful. > As noted, I think there are better — i.e., more general and involving fewer changes in the ontology of Lojban — solutions. This one would require that we make some general sense of the token/type distinction and this seems either to lead to a sucking vortex or to fall back on othr distinctions which are already allowed for and better understood (or, at leaast, to concepts which fit easily into theexisting framework).

> > It is perfectly clear that > > {lo broda} refers primarily to brodas, the ultimate tokens of broda-type. > > In fact, if it did not, we dould not get the whole token-type program off > > the ground. {lo} is after all, merely the inspecific version of {le} and {le} > > clearly refers to individual objects, not to object types (since there is no > > reason to believe there even is an object type corresponding to a {le} phrase. > > It is not clear to me that there is anything wrong with, for example: > > {zo klama cu vasru mu lerfu} > > or with > > {rexa da lerfu le lojbo} > > which use different token levels of {lerfu}.

Well, these are linguistic examples again and so don't help much. They are obviously OK and might usefully be disambiguated by token/type — or level of type maybe here — markers. That doesn't say much about elephants, however. (I though there were only 24, but I lose count sometimes.) > > > > > You may *mean* it for a type, but it looks to > > > > the logician like a token — with the usual bad results. > > > > > > Well, the logician should take a better look then, because there is nothing > > > in Logic that requires it to be interpreted as a token. > > > > Actually, there is — terms refer to objects, i.e., ultimate tokens. But > > this requirement is not from Logic but from Lojban (a language you really > > ought to get to know). > > Lojban, as far as I understand, does not come with a fixed ontology. > It is quite flexible in that regard, as any general purpose language > has to be.

Well, it has concrete objects and an array of abstract one — none of which is overtly marked as a token or a type so far as I can find: events of various sorts, properties, propositions and various functions allied to these, numbers and mathematical functions of a variety of sorts, and so on. Nor is there any real need for types that I can see — whatever they do can be done with what we have on hand (with sets, for example).

> > > > By the way, another reason why events are better than object types > > > > is that they can handle {le broda} and {la brod} without any adaptation. > > > > It is at least not obvious that there is a type behind either of these > > > > expressions to claim to be using. If there is, however, it will be > > > > different enougfh to require some additional apparatus. > > > > > > I'm sure there is. For example: > > > > > > "The flag of the United Kingdom of Great Britain and Northern Ireland > > > is sometimes called the Union Jack. This red, white, and blue flag > > > was first used in 1801." > > > that's a type usage of "this red, white, and blue flag". > > > > Certainly is; matches the first type use. > > > > > "This summer the Union Jack was flying in every street here." > > > that's a type usage of the name "the Union Jack". > > > > Yup. > > > > > "Union Jacks were everywhere: at Wembley, in the streets, at parties, > > > and festooned around pubs and clubs." > > > that is (or could be taken as) a token usage of the name "Union Jack". > > > > Well, it may be token relative to the earlier ones, but is still not an ultimate > > token. > > What would count as an ultimate token flag?

A particular piece of cloth existing over a particular space-time area.

> > But, of course, none of these are on the point, which is about using event > > types rather than object types in {djica2), remember. And, of course, in > > the {le} case, we have no guarantee that what is called the type here really > > is a type for the things referred to: the referents of {le broda} may not be > > tokens of a broda-type, after all. Nor is it perfectly clear that what we have > > here are cases of {le} altogether. > > These examples were to show that, at least in English, names and definite > descriptions can be used to refer to types. It is not clear to me why {le} > and {la} could not be so used.

I didn't say they couldn't be, I said it was not obvious that they could be. You have presented a case — on the usual comparison between English and Lojban constructions (but these are always suspect — that is why there is formal logic after all) — that there are cases where we can do it. To be sure, the examples are dangerously close linguistic cases. This may, of course, be setting up a slippery slope argument to the generalization of the token/type distinction. Or it may be covering over some very different distinction all together.

The ultimate point is that we can do everything we need to do without a separate token/type, using just familiar objects we already have (or that are part of the pattern already in place). Doing this also shows up some thing that talking about tokens and types (and saying some very strange things about them, it seems to me) leave unsaid — the nature of generality, for example.

I suppose this is not much more than that I don't think there are types and that you have not given me any reason to change my mind on that. I do know that there is no reason to introduce types into Lojban and so I think we ought not do it. There is a reason to provide in Lojban a way for sorting out the various uses of expressions — or rather the various sorts of sentences in which expressions occur — in a convenient explicit way, leaving for ordinary use the expressions we use now. And it may turn out that the way to do this looks like a token/type way. In that case, you can call it token/type as long as you don't mean it literally (although aside from being literally about literals, it is not clear what taking it literally would mean).


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Tue 12 of Sep., 2006 16:08 GMT On 9/11/06, John E Clifford wrote: > So, to prevent confusion at least (and also because this attempt to > generalize the notion is getting hopelessly messy), I suggest we drop > the toke/type talk and try to figure out what we really want.

(1a) mi viska lo nu do zutse lo xanto

(1b) mi djica lo nu do zutse lo xanto

(2a) mi viska lo xanto

(2b) mi djica lo xanto

Whatever the relationship that exists between {lo nu do zutse lo xanto} as it appears in (1a) and as it appears in (1b) is, I maintain that it is the same relationship that exists between {lo xanto} as it appears in (2a) and as it appears in (2b). It doesn't make a difference what name we give to that relationship as long as we are in agreement that it is the same relationship.

If we agree that the two are related in the same way, then the claim that shoving an object into an event "solves" anything is just not tenable. All the shoving does is shift whatever "problem" there was for objects to events.


> Thus, the case of not distinguishing individual from generality is part of a > general pattern (which we knew). In the one case, at least, the solution is > to recognize the complexity of the underlying situation (say the behavior > of countless anonymous individuals over vast spans of time) and simply > indicate somehw that that is what is going one, rather than a fact about a > delimited bunch of individuals in a narrow space-time area. ... > This does not obviously solve the opaque context cases, which yield nicely > to the usual abstractions — events and properties and propositions.

Opaque contexts do not "yield nicely" to events if they don't yield equally nicely to objects, as there is no relevant distinction between events and objects to which they could yield.

> Well, as I said, the token/type stuff is getting unduly messy, so lets say > what is clearly the case, namely that {nu bajra} refers to abstract entities, > events — but that is obvious from the fact that it is an abstract predicate.

In what way is it obvious that events are more abstract than objects? Is it because NUs are called "abstractors" in Lojban? Would that make {lo se li'i bajra} for example an obviously abstract entity?

> To get down to the occurrences from which this is abstracted takes a > different move, but it is not a type to token move. {nu bajra} does not refer > to a set, if for no other reason than that we currently have no way to refer > to its members. (that problem being solved, we could go a long way with > {nu bajra} being about a set of occurrences, but I think ultimately it would > not quite work).

Then it would seem that this "solution" to opaque contexts has broken more than it has fixed.


> Well, if English is our metalanguage (I would hope it was somewhat > enriched), then, while I can tell, what is meant by "word token" and > "word type", I am less than clear about what is meant by "dog token" > (apparently not just a dog) and "dog type." A lot of what holds for > linguistic objects (woords, letters, sentences, etc.) does not generalize > in obvious ways to other kinds of things and what does gneralize is bettr > dealt with under more traditional categories, where the moves have been > pretty wewll worked out.

You could use flags as a stepping stone. There the distinction is still obvious and they are not linguistic objects. BTW, it is not just me that has no problem in seeing how this generalizes to dogs and stuff. From

"Outside of philosophy and linguistics, scientists often quantify over types in their theories and refer to them by means of singular terms. When, for example, we read that the "Spirit Bear" is a rare white bear that lives in rain forests along the British Columbia coast, we know that no particular bear is rare, but rather a type of bear. When we are told that these Kermode bears "have a mutation in the gene for the melanocortin 1 receptor" (The Washington Post 9/24/01 A16) we know that it is not a token mutation, token gene and token receptor being referred to, but a type. It is even more evident that a type is being referred to when it is claimed that "all men carry the same Y chromosome. This one and only Y has the same sequence of DNA units in every man alive except for the occasional mutation that has cropped up every thousand years" (The New York Times, Nicholas Wade 5/27/03). ... The preceding paragraph contains singular terms that (apparently) refer to types. An even more telling commitment to types are the frequent quantifications over them. ... It was claimed (in the sixties), for example, that "there are thirty particles, yet all but the electron, neutrino, photon, graviton, and proton are unstable." Artifactual types (the Volvo 850 GLT, the Dell Latitude D610 laptop) easily lend themselves to reference also. In chess we are told that accepting the Queen's Gambit with 2dc has been known since 1512, but Black must be careful in this openingthe pawn snatch is too risky. Type-talk is ubiquitous."


> although you may just be insisting that here we need to recognize the > modality of generality and so have {lo xantu} referring to its sense, which > would, to be sure, tie up with the opaque cases, where that is definitely > what is going on (or, at least, that is the best explanation within the > ontology of Lojban — which does not have clear individual concepts, alas). > But individual concepts don't seem to me to be easily gotten to by analogy > with types.

Calling them (metalinguistically) types does not seem outrageous to me, but I'm not particularly interested in defending this or that label.


> As noted, I think there are better — i.e., more general and involving fewer > changes in the ontology of Lojban — solutions. This one would require that > we make some general sense of the token/type distinction and this seems > either to lead to a sucking vortex or to fall back on othr distinctions which > are already allowed for and better understood (or, at leaast, to concepts > which fit easily into theexisting framework).

If there are several solutions, there doesn't seem to be any reason not to let a thousand flowers bloom. My point in all this discussion is simply that putting an object inside an event does not solve what it purports to solve, it only shifts the perceived problem from objects to events.


> > {rexa da lerfu le lojbo} > (I though there were only 24, but I lose count sometimes.)

(It depends on whether {.}, {'} and {,} count as lerfu or not.)

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Tue 12 of Sep., 2006 18:24 GMT posts: 2388

> On 9/11/06, John E Clifford wrote: > > So, to prevent confusion at least (and also because this attempt to > > generalize the notion is getting hopelessly messy), I suggest we drop > > the toke/type talk and try to figure out what we really want. > > (1a) mi viska lo nu do zutse lo xanto > > (1b) mi djica lo nu do zutse lo xanto > > (2a) mi viska lo xanto > > (2b) mi djica lo xanto > > Whatever the relationship that exists between {lo nu do zutse lo xanto} as > it appears in (1a) and as it appears in (1b) is, I maintain that it is the > same relationship that exists between {lo xanto} as it appears in (2a) and > as it appears in (2b). It doesn't make a difference what name we give > to that relationship as long as we are in agreement that it is the same > relationship.

Well, let's see. In 1a. a term which looks to refer to intensional objects, events, is being used to refer to a particular occurrence that manifests that intension (I assume it is one occurrence, it might be — without any change in the general comments, refer generally to a wide range of such occurrences). In 1b. the same expression is used to refer to intensional objects, events, just as it is supposed to. In 2a, an expression that appears to refer to elephants does (either some particular bunch of them on a partiuclar ocasasion to to a wider range in general). In 2b., the same expression now refers to an intensional object of some sort — apparently a property or (whatever this may be) a kind or type or species. So, the similarity between the two seems to be that we (some of us some of the time at least)use the same expressions sometimes for extensional objects and sometimes intensional one. The expressions used in this dual way are in some cases the normally intensional ones, in others the normally extensional ones. We rely on context -- largely (in these cases at least) the predicates to which they attach — to resolve the ambiguity. If this is your point, then we agree about these cases.

> If we agree that the two are related in the same way, then the claim that > shoving an object into an event "solves" anything is just not tenable. All > the shoving does is shift whatever "problem" there was for objects to events.

I'm not clear about what you mean by "in the same way" I suppose it is (since otherwise they seem exactly opposite of one another) that in one the reference is to extensional objects and the other to intensional ones. But that doesn't lead to anything like what you claim. If you move from referring to extensional object to referring to intensional one, the problems disappear more or less by definition (the problem arising from treating intensional objects as though they were extensional). The difference that makes a difference is whether we refer to the intensional objects by terms that are explicitly meant to so refer or not (and correspondingly for extensional one, of course). Having the apparent referents be the ones intended just prevents trying the wrong moves, since it doesn't even look as thought they might apply. Of course, that works the othe way as well: using an intensional term to refer to extensional objects means we might miss rules that do apply (but don't appear to). Failing these shifts we need to keep handy a list of places where it is intensions that are meant (I am assuming that extensions are the norm)-- which is, of course, what we dso now (and sometimes forget and screw up).

> > > Thus, the case of not distinguishing individual from generality is part of a > > general pattern (which we knew). In the one case, at least, the solution is > > to recognize the complexity of the underlying situation (say the behavior > > of countless anonymous individuals over vast spans of time) and simply > > indicate somehow that that is what is going one, rather than a fact about a > > delimited bunch of individuals in a narrow space-time area. > ... > > This does not obviously solve the opaque context cases, which yield nicely > > to the usual abstractions — events and properties and propositions. > > Opaque contexts do not "yield nicely" to events if they don't yield equally > nicely to objects, as there is no relevant distinction between events and > objects to which they could yield.

The relevant difference is precisely the problems of opaque context. I think you are involved in a use/mention confusion here. {lo xanto} refers to elephants in isolation, but in {djica2}, if the problems of opaque context are not to plague us, it refers to some intensional object that has to do with elephant: the property of being an elephant as a first guess. Now, the fact that you use a word that normal stands for elephants and thay you can avoid the opaque problems while doing so, does not mean that you can avoid the opaque problems while referring to elephants: the words are the same, the referents different as determined by context. And 1a has the reverse problem -- referring to an extension by a term thst looks to be for an intension. What there is no relevant difference between {lo xanto} and {lo nu do zuts lo xanto} about is that both are used both ways. But that does not mean that they refer to the same thing when used in the different ways (obviously, from the way the ways are specified).

> > Well, as I said, the token/type stuff is getting unduly messy, so lets say > > what is clearly the case, namely that {nu bajra} refers to abstract entities, > > events — but that is obvious from the fact that it is an abstract predicate. > > In what way is it obvious that events are more abstract than objects? > Is it because NUs are called "abstractors" in Lojban? Would that make > {lo se li'i bajra} for example an obviously abstract entity?

Well, yes, there is {li'i}, an "abstractor" that goes to the maximally concrete, which may be why no one has used it worth mentioning since the paraplegic who proposed it originally. But, OK, I take back the "obvious," but it is still a pretty good reason (though, as you point out, conversion can come back to concrete fairly easily).

> > To get down to the occurrences from which this is abstracted takes a > > different move, but it is not a type to token move. {nu bajra} does not refer > > to a set, if for no other reason than that we currently have no way to refer > > to its members. (that problem being solved, we could go a long way with > > {nu bajra} being about a set of occurrences, but I think ultimately it would > > not quite work). > > Then it would seem that this "solution" to opaque contexts has broken > more than it has fixed.

But this is preciselly not a solution to the opaque context problem but only an attempt to salvage some sense from the token/type as it was being extended. the solution to the opaque context problem remains goinf over to intensional objects (which sets are not — I am not sure about types). > > > Well, if English is our metalanguage (I would hope it was somewhat > > enriched), then, while I can tell, what is meant by "word token" and > > "word type", I am less than clear about what is meant by "dog token" > > (apparently not just a dog) and "dog type." A lot of what holds for > > linguistic objects (words, letters, sentences, etc.) does not generalize > > in obvious ways to other kinds of things and what does generalize is better > > dealt with under more traditional categories, where the moves have been > > pretty well worked out. > > You could use flags as a stepping stone. There the distinction is still > obvious and they are not linguistic objects. BTW, it is not just me that > has no problem in seeing how this generalizes to dogs and stuff. From > :

That site seems to have type/token and property/object (and so probably sense and reference) fairly well jumbled together. It is evidence that either someone sees what all this means or else that someone else is thoroughly confused by the terminology. In any case, I don't see that this distinction, even if it can be made to make sense, has anything very useful to say about the Lojban situation, and it seems to introduce a new bunch of objects which are either useless as they stand or redundant. The flag case is useful in that it is not linguistic but is very like linguistic (conventional sign, importance of physical shape) but all the cases seem to be pretty clearly just matters of degree of generality and so serve as a nice step from normal generality to application of that notion to linguistic items — getting rid of token/type at home.


> "Outside of philosophy and linguistics, scientists often quantify over types > in their theories and refer to them by means of singular terms. When, for > example, we read that the "Spirit Bear" is a rare white bear that lives in rain > forests along the British Columbia coast, we know that no particular bear is > rare, but rather a type of bear. When we are told that these Kermode bears > "have a mutation in the gene for the melanocortin 1 receptor" (The Washington > Post 9/24/01 A16) we know that it is not a token mutation, token gene and > token receptor being referred to, but a type. It is even more evident that a > type is being referred to when it is claimed that "all men carry the same Y > chromosome…. This one and only Y has the same sequence of DNA units > in every man alive except for the occasional mutation that has cropped up > every thousand years" (The New York Times, Nicholas Wade 5/27/03). > ... > The preceding paragraph contains singular terms that (apparently) refer to > types. An even more telling commitment to types are the frequent > quantifications over them. ... It was claimed (in the sixties), for example, > that "there are thirty particles, yet all but the electron, neutrino, photon, > graviton, and proton are unstable." Artifactual types (the Volvo 850 GLT, > the Dell Latitude D610 laptop) easily lend themselves to reference also. In > chess we are told that accepting the Queen's Gambit with 2…dc has been > known since 1512, but Black must be careful in this opening—the pawn > snatch is too risky. Type-talk is ubiquitous."

Ubiquity doesn't need to imply literal correctness. Type talk may even be convenient (or it may just be about sets or some such thing). But none of this forces us to take types seriously. Nor does anything else, since type-talk can always be unpacked in a different way as talk about objects (ultimate tokens), about properties, and othe abstract entities that we already are stuck with. It might be that, were we designing Lojban now, we would have built in a different set of categories, in cluding types perhaps, but were not and we didn't. So, let's do a good job with what we have, rather than trying to first introduce new things, then try to do a job with them.

> > > although you may just be insisting that here we need to recognize the > > modality of generality and so have {lo xantu} referring to its sense, which > > would, to be sure, tie up with the opaque cases, where that is definitely > > what is going on (or, at least, that is the best explanation within the > > ontology of Lojban — which does not have clear individual concepts, alas). > > But individual concepts don't seem to me to be easily gotten to by analogy > > with types. > > Calling them (metalinguistically) types does not seem outrageous to me, but > I'm not particularly interested in defending this or that label. > > > > As noted, I think there are better — i.e., more general and involving fewer > > changes in the ontology of Lojban — solutions. This one would require that > > we make some general sense of the token/type distinction and this seems > > either to lead to a sucking vortex or to fall back on other distinctions which > > are already allowed for and better understood (or, at least, to concepts > > which fit easily into the existing framework). > > If there are several solutions, there doesn't seem to be any reason not to > let a thousand flowers bloom. My point in all this discussion is simply that > putting an object inside an event does not solve what it purports to solve, it > only shifts the perceived problem from objects to events.

I really don't think you have been paying attention. It is not "putting an object inside an event" that solves the problem. What solves the problem is using explicitly intensional terms at places where intensional objects are referred to. There are reasons for thinking that events are good intensional objects to refer to explicitly in {djica2}, but, so far as solving the problem goes, any other intesnional object and its appropriate term would do as well. What get referred to at that place is an intensional object in any case, even if it is referred to by some unquestionably extensional term. It is just that, from the point of the "logical language," which means, inter alia, that inferences bear their forms on the faces of the Lojban representations of them, it is better that the intensional object referred to there be referred to by an intensional term. (And, by parity of reasonsing, at least when reasoning is involved, that extensional terms be used in extensional places.) This move also has the advantage that we can then distinguish easily those cases where a normally intensional place takes a genuinely extensional object — as when there is a particular elephant I want, for example.


> > > > {rexa da lerfu le lojbo} > > (I though there were only 24, but I lose count sometimes.) > > (It depends on whether {.}, {'} and {,} count as lerfu or not.) Ah! I didn't count . and ,


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Tue 12 of Sep., 2006 20:18 GMT On 9/12/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > > > (1a) mi viska lo nu do zutse lo xanto > > (1b) mi djica lo nu do zutse lo xanto > > (2a) mi viska lo xanto > > (2b) mi djica lo xanto > > Well, let's see. In 1a. a term which looks to refer to intensional > objects, events, is being used to refer to a particular occurrence that > manifests that intension

How do you tell that {lo nu do zutse lo xanto} looks to refer to intensional objects? Would for example {da poi nu do zutse lo xanto} also look like that? Or {lo fasnu no'u lo nu do zutse lo xanto}? Or {lo fasnu poi mi ca pensi ke'a}? Or {ko'a}, if it has been previously assigned to {lo nu do zutse lo xanto}? Is it a particular form of the selbri converted by {lo} that makes it look as referring to intensional objects, or is it the meaning of that selbri? Is the predicate {nunbajra} for example to be treated the same way as the predicate {nu bajra}?

> The expressions used in this dual way are in some cases > the normally intensional ones, in others the normally extensional ones.

That's where we disagree. In your interpretation, there are some expressions that are normally intensional (and abnormally extensional?) and others that are normally extensional (and abnormally intensional?). In my interpretation there is no such distinction between expressions.


> It might be that, were we designing Lojban now, we would have built > in a different set of categories, in cluding types perhaps, but were not > and we didn't. So, let's do a good job with what we have, rather than > trying to first introduce new things, then try to do a job with them.

It seems to me that it is your interpretation that will require the introduction of something new, since you think it's necessary to mark the different uses of {lo nu ...} for example.


> > My point in all this discussion is simply that > > putting an object inside an event does not solve what it purports to solve, it > > only shifts the perceived problem from objects to events. > > I really don't think you have been paying attention. It is not "putting an > object inside an event" that solves the problem. What solves the problem > is using explicitly intensional terms at places where intensional objects > are referred to.

But defining event descriptions as explicitly intensional creates the problem of what to do when we want to refer to them extensionally.


> There are reasons for thinking that events are good > intensional objects to refer to explicitly in {djica2}, but, so far as solving > the problem goes, any other intesnional object and its appropriate term > would do as well.

I don't see any reason to treat events differently from objects. Again I don't seem to be particularly alone in this:

<distinction between objects and events [Quine 1960]. Rather, they would regard the relevant distinction as one of degree: both objects and events would be species of the same "material inhabitant of space-time" genus>>

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Tue 12 of Sep., 2006 21:38 GMT posts: 2388

> On 9/12/06, John E Clifford wrote: > > --- Jorge Llambías wrote: > > > > > > (1a) mi viska lo nu do zutse lo xanto > > > (1b) mi djica lo nu do zutse lo xanto > > > (2a) mi viska lo xanto > > > (2b) mi djica lo xanto > > > > Well, let's see. In 1a. a term which looks to refer to intensional > > objects, events, is being used to refer to a particular occurrence that > > manifests that intension > > How do you tell that {lo nu do zutse lo xanto} looks to refer to intensional > objects? Would for example {da poi nu do zutse lo xanto} also look like > that? Or {lo fasnu no'u lo nu do zutse lo xanto}? Or {lo fasnu poi mi ca > pensi ke'a}? Or {ko'a}, if it has been previously assigned to {lo nu do zutse > lo xanto}? Is it a particular form of the selbri converted by {lo} that makes > it look as referring to intensional objects, or is it the meaning of > that selbri? > Is the predicate {nunbajra} for example to be treated the same way as the > predicate {nu bajra}?

Well, I would have said it was because of the abstractor, but, as you have pointed out, that is unreliable. so now I suppose I have to say that I know about {nu} at least, I am not, obviously claiming that these expressions always (and perhaps not even usually) refer to intensional objeects. only that that is what they are specifically designed to do. Whether that extends to other expressions involving {nu} has to be worked out as we go along. I suppose it is surer in the case of {da poi nu...} which seems to require even more strongly than {lo} that the referent meet the predicate's requirements. {fasnu} is a puzzle, since I am not at all sure what goes into {fasnu1} and event or the occurrence of an event — but it would probably be a {lo nu...} in either case (for lack of a better in one case). I was actually kinda thinking of things like {numbajra} to take up the event/occurrence slack on the occurrence side, though I haven't thought the details through.

> > The expressions used in this dual way are in some cases > > the normally intensional ones, in others the normally extensional ones. > > That's where we disagree. In your interpretation, there are some > expressions that are normally intensional (and abnormally extensional?) > and others that are normally extensional (and abnormally intensional?). > In my interpretation there is no such distinction between expressions.

This seems to fly in the face of facts. Does {lo ka ...} ever refer to anything but a property? or {lo du'u...} a proposition? {lo nu} seems odd in this respect. Of course, that is because of the predicates involved, but that was what I was going on. Grammatically, of course, there is nothing to choose between them, but some predicatees just require that what satisfies them is intensional.So, taking the predicate involved into account, we can make the distinction, which is what I did. > > > It might be that, were we designing Lojban now, we would have built > > in a different set of categories, including types perhaps, but we're not > > and we didn't. So, let's do a good job with what we have, rather than > > trying to first introduce new things, then try to do a job with them. > > It seems to me that it is your interpretation that will require the > introduction of something new, since you think it's necessary to mark > the different uses of {lo nu ...} for example.

Yes, there is no handy device for indicating occurrences and so certain bits of precision are not so readily available, (there are also a slough of missing maodals), but these will be needed whatever the system is — they are genuine gaps in the language, which we have papered over informally so far.

> > > > My point in all this discussion is simply that > > > putting an object inside an event does not solve what it purports to solve, it > > > only shifts the perceived problem from objects to events. > > > > I really don't think you have been paying attention. It is not "putting an > > object inside an event" that solves the problem. What solves the problem > > is using explicitly intensional terms at places where intensional objects > > are referred to. > > But defining event descriptions as explicitly intensional creates > the problem of what to do when we want to refer to them extensionally.

So it does, so that is a problem. But, notice I don't really have a lot of choice here. I am going to need both an occurrence expression and an event one, so, if I say that {lo nu...} is for occurrences, I need a new one for events. (And, of coursem, using {lo nu} in {djica2} no longer would work). so the conservative approach is to keep {lo nu...} for events and find a new occurrence word.

> > > There are reasons for thinking that events are good > > intensional objects to refer to explicitly in {djica2}, but, so far as solving > > the problem goes, any other intesnional object and its appropriate term > > would do as well. > > I don't see any reason to treat events differently from objects. Again > I don't seem to be particularly alone in this: > : > > <> distinction between objects and events Quine 1960. Rather, they would > regard the relevant distinction as one of degree: both objects and events > would be species of the same "material inhabitant of space-time" genus>>

Yes, I agree with all of that on the metaphysical level: objects are just more or less long events and events are just particularly amorphous objects. These are often useful in metaphysics. But we aren't doing metaphysics exactly here, we are trying to get a coherent semantics for a language which does distinguish events (off from other objects, if you will). Now is it your point that {lo nu ...} expressions ought to be for occurrences rather than for the intensional analogs? That might be a better way to do things, but it is apparently not how Lojban has done it. the problems raised do not seem sufficient to lead us to go back and change that initial position. So, within that context, we keep {lo nu...} for intensional events and so on.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Wed 13 of Sep., 2006 14:52 GMT On 9/12/06, John E Clifford wrote: > > Well, I would have said it was because of the abstractor, but, as > you have pointed out, that is unreliable. so now I suppose I have to > say that I know about {nu} at least, I am not, obviously claiming that > these expressions always (and perhaps not even usually) refer to > intensional objeects. only that that is what they are specifically > designed to do.

I'm not sure that's so clearly the case. Before this thread, I don't think anybody had made the argument that NU abstractions referred to intensional objects. Nor is it clear that any other NU abstractor than {nu} would ever have to. {lo ka ce'u broda}, for example, has a single referent, a property, and that property is the whole of its extension, so there is no need for intensional objects there. The same for {lo du'u broda}, which has always the same referent, a proposition, and that proposition is all there is to its extension. The only abstractor that would seem to be required to have this odd behaviour of referring to "intensional objects" is {nu}, and that only sometimes, not in every one of its uses. So I am not really convinced that it was designed to refer to intensional objects.

> Whether that extends to other expressions involving {nu} has to be worked > out as we go along. I suppose it is surer in the case of {da poi nu...} > which seems to require even more strongly than {lo} that the referent > meet the predicate's requirements.

What seems odd to me is saying that intensionality is a property of the referent rather than of the referring term, i.e. that the predicate {nu broda} would mean something like "x1 is an intensional object, an event of brodaing".

> > > The expressions used in this dual way are in some cases > > > the normally intensional ones, in others the normally extensional ones. > > > > That's where we disagree. In your interpretation, there are some > > expressions that are normally intensional (and abnormally extensional?) > > and others that are normally extensional (and abnormally intensional?). > > In my interpretation there is no such distinction between expressions. > > This seems to fly in the face of facts. Does {lo ka ...} ever refer to anything > but a property? or {lo du'u...} a proposition?

Never. Their extensions are always a property and a proposition respectively. I think what you are driving at is that in {ka broda}, {du'u broda} or {nu broda} the predicate {broda} is used intensionally. But what we are concerned with is not the inner predicate {broda}, we are concerned with the full predicates {ka broda}, {du'u broda}, and {nu broda}. Those are the predicates acted on by {lo}, and those are the predicates that will lead us to our referent. And those predicates do not (should not) behave any differently than any other predicate of the language.


> > > > My point in all this discussion is simply that > > > > putting an object inside an event does not solve what it purports > > > > to solve, it only shifts the perceived problem from objects to events. > > > > > > I really don't think you have been paying attention. It is not "putting an > > > object inside an event" that solves the problem. What solves the problem > > > is using explicitly intensional terms at places where intensional objects > > > are referred to. > > > > But defining event descriptions as explicitly intensional creates > > the problem of what to do when we want to refer to them extensionally. > > So it does, so that is a problem. But, notice I don't really have a lot of > choice here. I am going to need both an occurrence expression and an > event one, so, if I say that {lo nu...} is for occurrences, I need a new one > for events.

If you are going to distinguish those two, you will also need to distinguish word type from word token, flag type from flag token, and so on. This is not something special about events.

> (And, of coursem, using {lo nu} in {djica2} no longer would work). > so the conservative approach is to keep {lo nu...} for events and find a > new occurrence word.

But then it can hardly be claimed that the shoving of objects into events was a nice way of solving anything.


> > : > > <> > distinction between objects and events Quine 1960. Rather, they would > > regard the relevant distinction as one of degree: both objects and events > > would be species of the same "material inhabitant of space-time" genus>> > > Yes, I agree with all of that on the metaphysical level: objects are just more > or less long events and events are just particularly amorphous objects. > These are often useful in metaphysics. But we aren't doing metaphysics > exactly here, we are trying to get a coherent semantics for a language > which does distinguish events (off from other objects, if you will). Now is it > your point that {lo nu ...} expressions ought to be for occurrences rather than > for the intensional analogs?

My point is that {lo nu ...} expressions ought to be no different than any other {lo } expression. Since I have no problem in using the same expression for types and for tokens, depending on context, I don't really want the distinction you want to make between events and occurrences to be obligatory. (It can always be made as required, of course, but by explaining it, not by some grammatical marker. This is because the distinction is not really binary but rather it encompasses a whole tree, so a single marker would not be of much use. For me there are no absolute type and absolute token, but rather something is relatively a token of something else.)

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by pycyn on Fri 15 of Sep., 2006 15:58 GMT posts: 2388

> On 9/12/06, John E Clifford wrote: > > > > Well, I would have said it was because of the abstractor, but, as > > you have pointed out, that is unreliable. so now I suppose I have to > > say that I know about {nu} at least, I am not, obviously claiming that > > these expressions always (and perhaps not even usually) refer to > > intensional objeects. only that that is what they are specifically > > designed to do. > > I'm not sure that's so clearly the case. Before this thread, I don't think > anybody had made the argument that NU abstractions referred to > intensional objects. Nor is it clear that any other NU abstractor than > {nu} would ever have to. {lo ka ce'u broda}, for example, has a single > referent, a property, and that property is the whole of its extension, so > there is no need for intensional objects there. The same for {lo du'u broda}, > which has always the same referent, a proposition, and that proposition > is all there is to its extension. The only abstractor that would seem to > be required to have this odd behaviour of referring to "intensional objects" > is {nu}, and that only sometimes, not in every one of its uses. So I am not > really convinced that it was designed to refer to intensional objects.

Well, I do see that the case for {nu} pointing to an intenssional object is rather circumstantial. So drop that, relieving the need to find something for occurrences and leaving only the genral problem of moving from extensional to intensional objects. The rest of this paragraaph puzzles me: properties and propositions are paradigm intensional objects, not extensions at all but rather what determines the extension of the expression in question — predicates or sentences as the case might be. I also worry a little about whether the referent of {lo ka broda} is a single property or the whole family of subsumed properties, the various ways of being a broda. Both positions have some virtues and probably some problems, but the occasions where there would be a significant difference are few — in the object language at least. If properties and propositions are extensional (a set and a truth value or, at best, a situation), what is left to be intensional? I suppose you will say "Nothing" but that leaves us with problems about opaque contexts again.

> > Whether that extends to other expressions involving {nu} has to be worked > > out as we go along. I suppose it is surer in the case of {da poi nu...} > > which seems to require even more strongly than {lo} that the referent > > meet the predicate's requirements. > > What seems odd to me is saying that intensionality is a property > of the referent rather than of the referring term, i.e. that the predicate > {nu broda} would mean something like "x1 is an intensional object, > an event of brodaing".

Well, I don't suppose that "is an intensioanl object" would be part of the usual definition; that it was an intensional object would be apparent from what it did.

> > > > The expressions used in this dual way are in some cases > > > > the normally intensional ones, in others the normally extensional ones. > > > > > > That's where we disagree. In your interpretation, there are some > > > expressions that are normally intensional (and abnormally extensional?) > > > and others that are normally extensional (and abnormally intensional?). > > > In my interpretation there is no such distinction between expressions. > > > > This seems to fly in the face of facts. Does {lo ka ...} ever refer to anything > > but a property? or {lo du'u...} a proposition? > > Never. Their extensions are always a property and a proposition > respectively. I think what you are driving at is that in {ka broda}, > {du'u broda} or {nu broda} the predicate {broda} is used intensionally. But > what we are concerned with is not the inner predicate {broda}, we are > concerned with the full predicates {ka broda}, {du'u broda}, and {nu broda}. > Those are the predicates acted on by {lo}, and those are the predicates > that will lead us to our referent. And those predicates do not (should not) > behave any differently than any other predicate of the language.

Well, I gues the point is that their referents (I will skip {nu}) are different sorts of things (in standard model theory, they are functions from worlds to sets and truth values or situations, depending just where one stands on such issue — "standard" is a bit of a misnomer sometimes. It is precisely the whole predicate, {ka broda} or {du'u ti broda}, that refers to such objects (or rather {lo ...} refers to such objects and the full predicates are true only of such objects). The objects are the senses of the inner predicates (as opposed to some set (or bunch) of brodas which is that predicate's extension). Even if you don't buy into model theory of this sort, the property of being a broda is clearly a different thing from just being a broda, which is what the extension would appear to be. These predicates don't behave — in one sense — any different from others: the extension of {ka broda is just the property (or properties) of being a broda and {lo} picks these (this) out. So, {lo ka broda} is an extesnional expression in one sense; it is intensional only in the sensse that what it refers to is an intensional object. "intesional expression" is shorthand for "expression that refers to an intensional object." How were you using it (or since you didn't use it, how were you taking me to be using it)?

> > > > > My point in all this discussion is simply that > > > > > putting an object inside an event does not solve what it purports > > > > > to solve, it only shifts the perceived problem from objects to events. > > > > > > > > I really don't think you have been paying attention. It is not "putting an > > > > object inside an event" that solves the problem. What solves the problem > > > > is using explicitly intensional terms at places where intensional objects > > > > are referred to. > > > > > > But defining event descriptions as explicitly intensional creates > > > the problem of what to do when we want to refer to them extensionally. > > > > So it does, so that is a problem. But, notice I don't really have a lot of > > choice here. I am going to need both an occurrence expression and an > > event one, so, if I say that {lo nu...} is for occurrences, I need a new one > > for events. > > If you are going to distinguish those two, you will also need to distinguish > word type from word token, flag type from flag token, and so on. This is > not something special about events.


I don't see that this follows. I don't see the connection between intensions and types at all (except, of course, that if types are going to help with the opaque problems, they will be intensional). Indeed, outside of linguistic items I don't see the use of types at all (and I think the linguistic cases can be dealt with without types as well).

> > (And, of course, using {lo nu} in {djica2} no longer would work). > > so the conservative approach is to keep {lo nu...} for events and find a > > new occurrence word. > > But then it can hardly be claimed that the shoving of objects into events > was a nice way of solving anything.

I still think that working with the intensions of events (of occurrences) has some virtues over working with object intensions (properties or individual concepts), but the usual event sense used, {tu'a lo broda} is so amorphous as to not really be significantly better than the other. In fact, were all of this to get worked into Lojban (I suppose LoCCan 3, really) I would vote for {tu'a} as the word for up (^), the function for getting to senses, rather than an abbreviation of the sort now used.

> > > > : > > > <> > > distinction between objects and events Quine 1960. Rather, they would > > > regard the relevant distinction as one of degree: both objects and events > > > would be species of the same "material inhabitant of space-time" genus>> > > > > Yes, I agree with all of that on the metaphysical level: objects are just more > > or less long events and events are just particularly amorphous objects. > > These are often useful in metaphysics. But we aren't doing metaphysics > > exactly here, we are trying to get a coherent semantics for a language > > which does distinguish events (off from other objects, if you will). Now is it > > your point that {lo nu ...} expressions ought to be for occurrences rather than > > for the intensional analogs? > > My point is that {lo nu ...} expressions ought to be no different than any other > {lo } expression. Since I have no problem in using the same > expression for types and for tokens, depending on context, I don't really > want the distinction you want to make between events and occurrences > to be obligatory. (It can always be made as required, of course, but by > explaining it, not by some grammatical marker. This is because the distinction > is not really binary but rather it encompasses a whole tree, so a single marker > would not be of much use. For me there are no absolute type and absolute > token, but rather something is relatively a token of something else.)

The intension-extension distinction is binary. Now within intensions and within extensions there are all sorts of tree structures having to do with the meaning or the membership of items of each sort. And I suppose you can technically talk about the intension and extension of either sort of thing, but these quickly gets to be pointless (and the second stage on the usual theory, somewhat -- but not muc — later on others) and is not at all treelike.

Notice, that, if we really need the token type distinction, we can — as you say — no single marker will help (even though I think there are absolute tokens, if you start to talk that way) and we can always spell out what we mean as much as need be. In particular, if {lo nu...} refers to occurrences, we can use it in various modalities, whether we bother to mention that or not -- or, indeed, given Lojban's weakness in this area, even can mention it. I don't see that token/type can do much more and introduces a messy metaphysics (or a redundant one).


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definitions

Posted by Anonymous on Fri 15 of Sep., 2006 21:38 GMT On 9/15/06, John E Clifford wrote: > --- Jorge Llambas wrote: > > On 9/12/06, John E Clifford wrote:

I moved this discussion to jboske@yahoogroups.com as requested by the jatna (in case anyone else is reading :-)

mi'e xorxes

Earlier

Posted by stevo on Thu 10 of Aug., 2006 21:53 GMT posts: 381 In a message dated 8/10/2006 1:56:45 PM Central Standard Time, jjllambias@gmail.com writes:


> Of course. In that case, "Bob" will be the name of a li-abstraction, > there is nothing wrong with that. You can also have li-abstractions > as values of variables, and of course {lo namcu} refers to a li-abstraction > too. The point is that anything that fills a place that requires a > li-abstraction > has to refer to a li-abstraction. This is not about the form that the sumti > there can have but about the type of value that the sumti there can refer to. >

What's a "li-abstraction"? Isn't "li" a gadri, like "le" and "lo", and not an abstraction like "nu" and "za'i"?

The abstraction closest in meaning to "li" that I can see is "ni", the amount abstraction.

stevo


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by Anonymous on Thu 10 of Aug., 2006 22:43 GMT On 8/10/06, MorphemeAddict@wmconnect.com wrote: > What's a "li-abstraction"? Isn't "li" a gadri, like "le" and "lo", and not > an abstraction like "nu" and "za'i"?

While it's not an "abstraction" as we usually use the term in Lojban, any number can also be an "abstraction".

The term we should probably be using is "sumti restriction".

-- Theodore Reed (treed/bancus) www.surreality.us


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by Anonymous on Fri 11 of Aug., 2006 13:03 GMT On 8/10/06, MorphemeAddict@wmconnect.com wrote: > What's a "li-abstraction"?

Mainly numbers, but there are other things such as dates and clock-times that are li-abstractions in Lojban too.

> Isn't "li" a gadri, like "le" and "lo",

Yes. Not exactly alike because they are in different selma'o, but they are all gadri. All gadri can be used to refer to abstractions.

>and not > an abstraction like "nu" and "za'i"?

{nu} and {za'i} are usually classed as abstractors in Lojban, i.e. members of selma'o NU. (They may be abstractions too in the sense that any word could be considered an abstraction. Perhaps we should also include text-abstractions as another sumti class to consider.)

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by stevo on Thu 17 of Aug., 2006 18:12 GMT posts: 381 In a message dated 8/17/2006 9:35:01 AM Central Standard Time, jjllambias@gmail.com writes:


> > First please correct — as I have done — the first claim > > to read "If you want a dog, you are in luck because dogs exist." > > Why should dogs existing be enough to satisfy my want of a dog? > Do you think that existence of X is always a sufficient condition for > satisfaction of a want of X? If you don't think that (and I know you don't > think that) then why say that the fact that dogs exist means I'm in luck? >

You'd really be out of luck if you wanted a dog and they didn't exist. So the fact that they exist means that at least part of the necessary conditions are fulfilled. It's hardly a sufficient condition though.

stevo


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by JohnCowan on Thu 17 of Aug., 2006 18:32 GMT posts: 149 MorphemeAddict@wmconnect.com scripsit:

> You'd really be out of luck if you wanted a dog and they didn't exist. So > the fact that they exist means that at least part of the necessary conditions > are fulfilled. It's hardly a sufficient condition though.

Quite right. Such situations are commonplace, however; you may want a pair of (your) clean black socks, but then you find out they're all in the wash.

-- Normally I can handle panic attacks on my own; John Cowan but panic is, at the moment, a way of life. http://www.ccil.org/~cowan


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by pycyn on Fri 18 of Aug., 2006 00:27 GMT posts: 2388

> In a message dated 8/17/2006 9:35:01 AM Central Standard Time, > jjllambias@gmail.com writes: > > > > > First please correct — as I have done — the first claim > > > to read "If you want a dog, you are in luck because dogs exist." > > > > Why should dogs existing be enough to satisfy my want of a dog? > > Do you think that existence of X is always a sufficient condition for > > satisfaction of a want of X? If you don't think that (and I know you don't > > think that) then why say that the fact that dogs exist means I'm in luck? > > > > You'd really be out of luck if you wanted a dog and they didn't exist. So > the fact that they exist means that at least part of the necessary conditions > are fulfilled. It's hardly a sufficient condition though.

This is the point: since all you want is just a dog — nothing special about it — an existing dog oughtto be sufficient. Since it is, you want something more, regardless of what you say.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by pycyn on Fri 18 of Aug., 2006 00:32 GMT posts: 2388

> MorphemeAddict@wmconnect.com scripsit: > > > You'd really be out of luck if you wanted a dog and they didn't exist. So > > the fact that they exist means that at least part of the necessary conditions > > are fulfilled. It's hardly a sufficient condition though. > > Quite right. Such situations are commonplace, however; you may want a pair of > (your) clean black socks, but then you find out they're all in the wash. > That is, what you wanted was to have your clean black socks available to wear, which you shortened somewhat in stating your want.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by JohnCowan on Fri 18 of Aug., 2006 01:46 GMT posts: 149 John E Clifford scripsit:

> > Quite right. Such situations are commonplace, however; you may want > > a pair of (your) clean black socks, but then you find out they're > > all in the wash. > > > That is, what you wanted was to have your clean black socks available > to wear, which you shortened somewhat in stating your want.

My point was that if all your black socks are dirty, then the desire for *clean* black socks is unsatisfiable *for the same reason* that the desire for a unicorn is unsatisfiable, though less permanently so.

-- John Cowan cowan@ccil.org http://ccil.org/~cowan Nobody expects the RESTifarian Inquisition! Our chief weapon is surprise ... surprise and tedium ... tedium and surprise .... Our two weapons are tedium and surprise ... and ruthless disregard for unpleasant facts.... Our three weapons are tedium, surprise, and ruthless disregard ... and an almost fanatical devotion to Roy Fielding....


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by pycyn on Fri 18 of Aug., 2006 14:46 GMT posts: 2388 This discussion has - as usual, unfortunately — gotten off the point and into a vatriety of irrelevancies or downright errors.

The issue was whether to have those little "preferred argument type" parentheses on some places. The general point was that, while not obligatory, the indications were useful, espexcially when 1) they were required by the obvious sense of the predicate and 2) when the English idiom was markedly different. The latter case, I maintained was especially important where following the English idiom led to results that went against logic, that is appeared to allow moves that were invalid. In keeping with that I recommended putting thee suggestions in even more places where following English led to these kinds of error.

That led to a discussion of whether the usually valid arguemnts were in fact invalid in these cases. Xorxes, with his usual ingenuity presented a number of cases in which both premise and conslusion were true and and explanation in terms of how universes of discourse changed over the course of a conversation. And I tried to show both that the supposed cases where the rules were not invalidated were not genuinely such cases and that the universe of discourse did not change in the way proposed.

Both of these discussions turn out to be irrelevant to the issue. The heart of the issue is not how universes change but in what universe (or stage of the universe) a sentences is evaluated. For example, is the sentence "I want a unicorn" evaluated before or after the universe haw been expanded to contain the unicorn I want. The answer is pretty clearly "Before," since we want the claim to be sometimes true and, in the universe as expanded it is never the case that I want a unicorn since the expansion is exactly to get me the unicorn I want. Once that issue is settled the moves to make the aruments appear OK fade: in the universe before expansion, while it may be true that I want a unicorn, it is not guaranteed that there is a unicorn I want or any unicorn at all. Similarly, even if, in the universe in which the dream is described rather than the universe of the dream, any unicorn is a centaur and any centaur a unicorn (because there are neither), it does not follow that, if I dream of a unicorn, I dream of a centaur. And the evaluation of both claims takes place in the unexpanded universe, since in the expanded universe I am not dreaming or at least not of the thing that defines that universe. Now, it is the case that some pairs of premise and conclusion can be plausibly argued to be both true. But this, of course, does not show that the argument form is valid. On the other hand, just one case of true premise and false conclusion show that the form is invalid and there seems to be such cases readily to hand.

So, it does seem to be useful to use a different idiom from the English in those places, to avoid te appearnce that the rules applied. The Lojban suggestion has been to use abstractions. With abstractions in those places the rules go through without a hitch (although the identity of abstractions is not always what one might like). Further, as xorxes has demonstrated, the satisfaction of a want and the like always involve some abstraction realized and we might as well say what that abstraction is rather than leaving it entirely to context.


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by Anonymous on Fri 18 of Aug., 2006 18:46 GMT On 8/18/06, John E Clifford wrote: > The heart of the issue is not how universes change but in what universe > (or stage of the universe) a sentences is evaluated.

Here you start mixing worlds, where things take place, with domain of discourse, which has referents as its members. One issue is how the domain of discourse changes as the discourse progresses, and another issue is in what worlds do the members of the domain of discourse exist. In the above sentence you start using "universe" for "domain of discourse" and then switch to using "universe" for "world".

> For example, is the sentence "I want a unicorn" evaluated before or > after the universe haw been expanded to contain the unicorn I want.

It's evaluated after the domain of discourse has been expanded to contain unicorns as a member. There is never any "the unicorn I want" at stake, because the reference is to unicorns in general, not to any particular unicorn. (There is a reading in which it means "I want a certain unicorn", but presumably that is not the intended reading here.)

> The answer is pretty clearly "Before," since we want the > claim to be sometimes true and, in the universe as expanded it is never > the case that I want a unicorn since the expansion is exactly to get me > the unicorn I want.

Here you are talking about expanding the world where things exist rather than the domain of discourse which contains the referents of referring terms. That's odd, because utterances don't expand the world in that way. (The only sense I can think of in which an utterance expands a world is in that before the utterance was made the utterance did not exist in the world, and after it is made, the utterance does exist. Perhaps performatives have the power of expanding the world in some sense too.) A claim is not true or false in a domain of discourse. It is true or false in a world, which is not expanded in any relevant sense by uttering "I want a unicorn". The world in this case is the real world, and in the real world I exist and unicorns don't exist. The predicate "want" is true, in the real world, if the referent of "I", which exists in the world, and the referent of "unicorns", which happens not to exist in the world, are related by the predicate "want". There is no need for the two referents being related by a predicate to exist in the world where the utterance is made. In plenty of cases they are even required to exist in different worlds.

> Once that issue is settled > the moves to make the aruments appear OK fade: in the universe before > expansion, while it may be true that I want a unicorn, it is not guaranteed > that there is a unicorn I want or any unicorn at all.

But the world is not relevantly expanded by the utterance. Only the domain of discourse may be expanded. And the claim is never evaluated in a domain of discourse, it is evaluated in a world. We don't need any guarantee that any referent exist in any world. Existence in a world, if required at all for the evaluation, is something that must be evaluated independently of the membership of the referent in the domain of discourse.

> Similarly, even if, in the universe in which the dream is described rather > than the universe of the dream, any unicorn is a centaur and any centaur > a unicorn (because there are neither), it does not follow that, if I dream > of a unicorn, I dream of a centaur.

In the world in which the dream is described, even if no unicorn exists and no centaur exists, a unicorn is not a centaur. For "a = b" to be true in a world it is not necessary that a and b exist in that world. And saying that all things that don't exist in a given world are thereby the same thing in that world seems just wrong. All that is required for "a = b" to be true in a world is that "a" and "b" have the same referent in the domain of discourse of the utterance, which may include (and usually does include) things that don't exist in the world in which the utterance is made.

mu'o mi'e xorxes


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Def...

Posted by pycyn on Fri 18 of Aug., 2006 21:12 GMT posts: 2388

> On 8/18/06, John E Clifford wrote: > > The heart of the issue is not how universes change but in what universe > > (or stage of the universe) a sentences is evaluated. > > Here you start mixing worlds, where things take place, with domain of > discourse, which has referents as its members. One issue is how the > domain of discourse changes as the discourse progresses, and another > issue is in what worlds do the members of the domain of discourse exist. > In the above sentence you start using "universe" for "domain of discourse" > and then switch to using "universe" for "world".

No, I meant universe of discourse throughout. One of the main points of ahving a universe of discourse is to determine whether sentences are true or false in it. And it is always in some such universe that sentences are evaluated — even if the universe is just "the real world" (or the universe associated with it, if you will).

> > > For example, is the sentence "I want a unicorn" evaluated before or > > after the universe haw been expanded to contain the unicorn I want. > > It's evaluated after the domain of discourse has been expanded to contain > unicorns as a member. There is never any "the unicorn I want" at stake, > because the reference is to unicorns in general, not to any particular > unicorn. (There is a reading in which it means "I want a certain unicorn", > but presumably that is not the intended reading here.)

Well, no. The whole thing about unicorns comes in exactly in evaluating the sentence. "I want a unicorn" is true just in case I have a psychological lack such that, if I were to have a unicorn (suppose) the lack would disappear. So, yes, the unicorn I want does enter in, or at least a unicorn that I get does. That unicorn is, however, in another universe, counterfactual to this one, one in which unicorns exist and I get one. But it comes up on for evaluating the want claim in this universe. Once that is settled, the unicorn plays no role in this universe. Or need not -- we can always bring it in, if we want 9and the others agree).


> > The answer is pretty clearly "Before," since we want the > > claim to be sometimes true and, in the universe as expanded it is never > > the case that I want a unicorn since the expansion is exactly to get me > > the unicorn I want. > > Here you are talking about expanding the world where things exist rather > than the domain of discourse which contains the referents of referring terms.

Ther is nothing here about existing, only about being in a universe. The point is that the referring term "a unicorn" (if indeed it is a referring term) refers to a thing in a different universe from the one in which the want claim is evaluated (indeed, one that come into consideration just to evaluate that claim.

> That's odd, because utterances don't expand the world in that way.

Well, we agree about something at least.

> (The > only sense I can think of in which an utterance expands a world is in that > before the utterance was made the utterance did not exist in the world, and > after it is made, the utterance does exist. Perhaps performatives have the > power of expanding the world in some sense too.)

They do: they create rights and duties at least.

> A claim is not true > or false in a domain of discourse.

That is the only place it can be.

It is true or false in a world, which is

I take it "world" here means the extension of the predicate "exists" or some such thing in some universe.

> not expanded in any relevant sense by uttering "I want a unicorn". > The world in this case is the real world, and in the real world I exist > and unicorns don't exist. The predicate "want" is true, in the real world, > if the referent of "I", which exists in the world, and the referent of > "unicorns", > which happens not to exist in the world, are related by the predicate "want". > There is no need for the two referents being related by a predicate to exist > in the world where the utterance is made. In plenty of cases they are even > required to exist in different worlds.

To my astonishment, I agree with almost everything you say. It is irrelevant, of course, but it is nice to have some concord. Yes, mentioning things do not call them into existence nor do a wanter and a wanted have to exist in the same world (that is part of the reason there is a problem with expressing wants). But, of course, I didn't say anything against that — indeed, I didn't say anything about that.

I think, from what you have said, that you think that sentences aren't evaluated in universe of discourse. But they have to be. A universe of dsiscourse is a full interpretation and it has to be, if descriptions are use, since we need to know the extensions of all the predicates and the identifications of all the objects and so on. Otherwise, adding a unicorm=n makes no sense: it is either a new thing or an old one, but in any case it has to be something that is in the extesnion iof unicorn. And, of course, the extensions of all the predicates and the identities of all the objects is exactly what is needed to evaluate a sentence. I think that you think a universe of discourse is just a bunch of things with perhaps labels attached, but that is a totally inadequate notion for even minimal use (keeping track of what is being talked about say). And if you get more, you get an evaluator of sentences.

> > Once that issue is settled > > the moves to make the aruments appear OK fade: in the universe before > > expansion, while it may be true that I want a unicorn, it is not guaranteed > > that there is a unicorn I want or any unicorn at all. > > But the world is not relevantly expanded by the utterance. Only the domain > of discourse may be expanded.


But notice that I was talking about the universe, the domain, not about the world at all.

> And the claim is never evaluated in a domain > of discourse, it is evaluated in a world.

It has to be evaluated in a domain if at all. There is no rwequirement in this sort of evaluation that the things involved exist, only that they are available for fariables and referrers.

>We don't need any guarantee that > any referent exist in any world. Existence in a world, if required at all > for the evaluation, is something that must be evaluated independently > of the membership of the referent in the domain of discourse.

Well, actually, presumably what exists and what does not is also decided in the domain, since we can certaily make reference to "an existing water buffalo" (or unicorn for that matter)and so we needs must have that much of an evaluation done.

> > Similarly, even if, in the universe in which the dream is described rather > > than the universe of the dream, any unicorn is a centaur and any centaur > > a unicorn (because there are neither), it does not follow that, if I dream > > of a unicorn, I dream of a centaur. > > In the world in which the dream is described, even if no unicorn exists > and no centaur exists, a unicorn is not a centaur. For "a = b" to be true > in a world it is not necessary that a and b exist in that world.

Or, as this case points out, even for them to be in the universe of discourse. I am not sure just what the counterfactual for dreams is like (there are several interesting but no convincing stories) but I am pretty sure that some such line is involved, that is that recourse is made to some counterfactual universe.

> And saying > that all things that don't exist in a given world are thereby the same thing > in that world seems just wrong. All that is required for "a = b" to be true in > a world is that "a" and "b" have the same referent in the domain of discourse > of the utterance, which may include (and usually does include) things that > don't exist in the world in which the utterance is made.

Well, since the claim was about universes, I suppose that this just makes my point: in a universe in which there are no unicorns and no centaurs, the universal claim "Everything is a unicorn just in case it is a centaur" clearly holds and that is what is needed for identity.

Posted by stevo on Wed 09 of Aug., 2006 03:43 GMT posts: 381 In a message dated 8/8/2006 7:36:35 PM Central Standard Time, wikidiscuss@lojban.org (rlpowell) writes:

> All paranthetical or square bracketed remarks in gismu definitions are > descriptive rather than proscriptive.

Do you mean "prescriptive"? "Proscribed" means 'banned'.

stevo


Score: 0.00 Vote: 1 2 3 4 5 top of page Reply

Edit  Delete  Report this post	

BPFK gismu Section: Parenthetical Remarks in Brivla Definit...

rlpowell Posted by rlpowell on Wed 09 of Aug., 2006 04:58 GMT posts: 14214 On Tue, Aug 08, 2006 at 11:49:13PM -0400, MorphemeAddict@wmconnect.com wrote: > In a message dated 8/8/2006 7:36:35 PM Central Standard Time, > wikidiscuss@lojban.org (rlpowell) writes: > > > All paranthetical or square bracketed remarks in gismu > > definitions are descriptive rather than proscriptive. > > Do you mean "prescriptive"?

Yes.

-Robin