jbocre: Bear goo

melting gummy bear

Bear goo has become an iconic phrase in connection with xorlo. The background of it is that there is no consensus on whether or not {lo cribe} can refer to bear goo. Arguments for both sides have been brought forth, and it tends to come down to philosophy. It also turns out that the question of bear goo is a broader one than xorlo. Below is a casual conversation about bear goo, which was held at the end of July 2013 on the #lojban IRC channel on freenode.

Day 1:

<selpa`i> Btw, I would like to point out that the reason I kept saying "bear goo aside" when we were talking about xorlo was because I knew you were strongly opposed to the idea and I didn't want the discussion to revolve around something I felt was not central to the current point under discussion, and not because I thought bear goo was unthinkable.

<selpa`i> doi la latro`a :)

<Ilmen> just for refreshing my memories, what is the difference (now) between {PA broda} and {lo PA broda}?

<selpa`i> pa da poi ke'a broda vs zo'e noi broda gi'e zilkancu li PA lo broda

<Ilmen> .i je'e .ui ki'e la selpa'i

<latro`a> also, that zilkancu equation is a slight problem

<latro`a> since if you count off by units of the whole {lo broda} ball, you will always get 1

<selpa`i> Surely, in a ready-made universe, bear goo must seem terrifying.

<latro`a> that is, making that equation makes sense requires a little bit of circularity, in artificially defining {zilkancu} to be distributive in its 3rd argument

<latro`a> I still think bear goo is a dumb idea in the first place, but again, xorlo comes down to what {zo'e} is and what broda'ing means, not really what lo means

<latro`a> if people come to a consensus that bear goo meets the definition of {cribe}, then the fact that it doesn't meet the definition of "bear" is superfluous

<latro`a> the difference is that I completely reject that "anything to do with bears" i.e. {zo'e ne lo ka cribe} matches {lo cribe}

<selpa`i> I understand. You view the universe as a set of predefined individuals, simply put. Not everyone does that, but it's common in math and classical logic.

<latro`a> that's orthogonal to what I just said

<latro`a> it's also a bit deceptive; I wouldn't say it's true or false, however

<selpa`i> Not really. You want to define a priori all the things that cribe and give them a single type.

<selpa`i> I think.

<latro`a> not at all

<latro`a> this is simpler than that: I just don't think bear goo {cribe}s

<latro`a> there's no a priori universal classification of {cribe} here

<latro`a> I just never thought bear goo actually {cribe}'d, and that it was silly to say it did

<selpa`i> There is, but it might be subtle. You're saying the node that cribe is on ends at bear goo.

<latro`a> there are still valid questions about whether things do: does a bear corpse {cribe}? how long after death does it stay that way?

<latro`a> again, no

<latro`a> there's a gray area

<latro`a> bear goo is just way past it

<selpa`i> But it sounds like you want to define a priori how far cribe can go in either direction, and have that be absolute for every time and place.

<latro`a> nope

<selpa`i> Not?

<latro`a> I just don't think bear goo is ever there

<latro`a> the gray area moves

<latro`a> bear goo is just beyond the boundary, and by a long shot

<latro`a> so it's never caught

<latro`a> really, we agree on more than you think when I say that "I reject bear goo"

<selpa`i> And you cannot imagine any context where it moves far enough to encapsulate goo?

<latro`a> indeed

<latro`a> that's not a damn bear, that's all there is to it

<latro`a> once you've turned it into goo, it came from a bear but isn't one

<selpa`i> But then the limits of your cribe are more rigid than mine, and obviously this goes for all broda.

<latro`a> I think the difference is fairly marginal

<latro`a> and again, bear goo doesn't really have anything to do with it, it doesn't change anything, it just changes the predicate

<selpa`i> I could construe a context where cribe is enough/appropriate to identify/distinguish between different animals having been in some place, or maybe when following a trail.

<latro`a> xorlo didn't have anything to do with it

<latro`a> then you're still not talking about a bear's physical presence at this moment

<latro`a> (actually, I'm not even really sure what you were going for with that comment)

<selpa`i> Let's say you and I are out in the jungle and following a trail.

<selpa`i> We come across some goo.

<selpa`i> We're trying to find Bear.

<selpa`i> But there are other animals there too

<selpa`i> So cribe or not.

<latro`a> if you said {xu ta cribe} I would say {ki'a}

<latro`a> seriously

<latro`a> not even {ua nai}

<latro`a> (there's unfortunately not a thing between, where you say you understand the syntax but don't understand the basic semantic assumptions)

<selpa`i> Okay. Now let's go the other direction, towards the root node.

<selpa`i> Say in said jungle, there live different animals: bears, birds, and tigers (or whatever).

<selpa`i> They are rare, so we see them rarely if at all. You and I take different paths through the jungle and meet again at the end of the day.

<selpa`i> You ask me "how many animals did you see?"

<selpa`i> I say: "Two. But I didn't see Tiger."

<selpa`i> Maybe I could add "I came across a big flock of birds this noon"

<selpa`i> I'd assume you to ki'a or something again.

<selpa`i> Obviously, you say, I saw more than two animals if there was a flock of birds

<latro`a> correct

<selpa`i> But then!

<latro`a> that one would be more like {na'i}, however, as it isn't just "wtf?", it's "that's not consistent"

<selpa`i> Okay, but I think it makes sense to say I saw only two animals; the bear and the bird.

<latro`a> I disagree

<selpa`i> I know!

<latro`a> you saw two {danlu gunma}, one of which is a singleton

<selpa`i> That's why I am saying you are tending towards a ready-made universe.

<latro`a> not really

<latro`a> it's dynamic

<latro`a> there's fluctuation from context and so forth

<selpa`i> It doesn't seem very dynamic.

<latro`a> but there's also some things that are just blatant, like a flock of birds not being an animal

<selpa`i> To you it surely seems that way. :)

<latro`a> but this is just a question of definitions to me, still: "made up of animals" and "is an animal" are completely different predicates, surely

<selpa`i> I don't see where your view is dynamic.

<latro`a> suppose you have a live bear

<latro`a> it goes about its life, throughout its life it's a bear

<latro`a> it dies

<latro`a> immediately after it dies, I still identify it as a bear

<selpa`i> Dynamic to me would at least imply that i can call a flock of birds a single bird.

<selpa`i> Since, at the beginning of our journey, went out to find three animals: bears, tigers, birds.

<latro`a> sorry, can I finish my example?

<selpa`i> Please.

<latro`a> was distracted for a sec

<latro`a> so the bear dies, and immediately after it dies, I identify it as a bear: it has bear teeth, bear claws, bear fur, bear shape, etc.

<latro`a> as it decays, it starts looking less like a bear

<latro`a> its fur falls out, its teeth and claws decay

<latro`a> eventually its skin is removed

<latro`a> etc.

<latro`a> much later, it is composted and incorporated into the soil

<latro`a> by the time it is in the soil, it is most definitely not a bear

<latro`a> but it is not determinate a priori when exactly it stopped being a bear

<selpa`i> But it's decided a priori that being in the soil is when it stops.

<latro`a> that is, there's a period when it was definitely a bear, extending throughout its life and through some of the aftermath of its death, and a period when it is definitely not a bear, long after its death, but there is a period where it is merely "bearish", and whether "bear" actually applies to it depends on other factors

<latro`a> I would say yes, but that has more to do with "bear" than anything else

<selpa`i> How exactly is bear goo different from a decayed, toothless, bald bear?

<latro`a> na'i; you haven't said how decayed it is, and even in the state of decay it's a gray, contextual area

<selpa`i> If a truck *just* ran over it, everyone who is present does know it's a bear.

<latro`a> you're asking me to make a universal statement about something that I was just saying wasn't universal

<Visirus> Was a bear

<latro`a> was a bear, correct

<latro`a> the transition needn't be gradual like the decay in the forest

<selpa`i> No, but you just said after its death it's still a bear?

<latro`a> it can be abrupt

<latro`a> it's still a bear because it's recognizable by its current features as such

<selpa`i> It's both dynamic and not dynamic at once it seems.

<latro`a> in different senses yes

<Visirus> It's a corpse, more logically

<selpa`i> Why are those features only of a visual nature?

<latro`a> not just visual

<selpa`i> A bear that got squashed and remains in the same place is recognizable by being in the same spot.

<latro`a> that's using external information

<selpa`i> Is this forbidden a priori?

<latro`a> that's a good question

<latro`a> I don't have a good answer, but my initial reaction is to say yes

<Rnuomer> since I started learning lojban I've been thinking everything in terms of verbs

<latro`a> put it this way, if you can't imagine walking up to the scene with no information whatsoever and saying {ta [ca] cribe}, it's not a bear

<selpa`i> The ready-made view is *very* strongly prevalent on IRC nowadays, that's why I imagine it's hard to see the other view.

<latro`a> I still don't like this "ready-made" description

<Rnuomer> so I'd think "Is the thing bear-ing?"

<latro`a> it's inaccurate

<Rnuomer> if so, it's a bear

<latro`a> I don't know what a better term is, but only part of the system is static

<latro`a> much of it is dynamic

<latro`a> because as I said, there's no definite transition point where it stopped being a bear

<latro`a> on the other hand, there is a region that is definitely bear and a region that is definitely non-bear

<latro`a> but the middle is gray, fluid, and indeterminate

<selpa`i> There are some absolute classifications you are applying on the universe and then use them once and for all, even if some things are dynamic, you have just shown that some things are not, for example the limits of cribe seem rather clear, and a flock of birds is never a bird.

<latro`a> the limits of cribe aren't

<latro`a> but yes, a flock of birds is never a bird

<selpa`i> How can you say that if {lo cipni} is Bird? Then you must forbid that interpretation, which is a very ready-madeist (sorry) thing to do.

<latro`a> hrm, I need to play with this issue

<latro`a> that last point is a good one

<latro`a> my internal resolution comes from my previous interpretation, which is more self-consistent than my current, somewhat wishy-washy one

<latro`a> which is to say that {lo za'u cipni cu cipni} is a distributive statement

<latro`a> but this is incompatible with the {sruri lo dinju gi'e krixa} perspective

<latro`a> (which I still don't actually like, but for more practical than philosophical reasons)

<latro`a> when I say "internal resolution", I mean the answer that manifests before I've had to compare perspectives etc.

<Rnuomer> I don't suppose my idea makes any sense =:x

<latro`a> it's between the two, Rnuomer

<latro`a> I prefer to think of selbri as noun-verbs, and consider the best brivla to have place structures that are neither truly nounish nor truly verbish

<Rnuomer> I think a selbri is more a verb tho

<latro`a> note that in lojban, the issue you're talking about wouldn't happen, selpa'i

<latro`a> you would've said {mi viska lo cipni}

<latro`a> not "I saw a bird"

<latro`a> and I wouldn't have concluded it was singular

<latro`a> where we perhaps run into trouble is {mi viska lo pa cipni}

<latro`a> if a flock of birds {cipni}, then {lo pa cipni} is ambiguous as to whether it is actually a flock or not

<Rnuomer> seeing as sumti are definitely nouns

<Rnuomer> (right?)

<latro`a> despite seemingly being explicitly singular

<latro`a> sorta; in english nouns are themselves content-words

<latro`a> in lojban the only sumti that are content-words are KOhA

<latro`a> cf. "dog" vs. "gerku"

<selpa`i> Had a phone call.

<Rnuomer> and sumti with LE + selbri are "something that [selbri]s"

<Rnuomer> pe'i

<latro`a> my notion of noun-verb is a bit vague; the point is that it has to do with both a state of being and a state of action

<latro`a> nounish selbri are about states of being; verbish selbri are about states of action; noun-verbish selbri build in both, describing what something is via what it does and vice versa

<selpa`i> < latro`a> note that in lojban, the issue you're talking about wouldn't happen, selpa'i -- I think it would. Consider {mi viska [lo] ci danlu}.

<latro`a> can we jump down a little?

<latro`a> because I think we already hit the heart of the issue

<latro`a> namely

<latro`a> can {lo *pa* cipni} be a flock?

<Rnuomer> what is the difference between lo and loi then?

<latro`a> loi is explicitly non-distributive

<latro`a> lo is explicitly not explicit about distributivity

<latro`a> (nor about whether distributivity even makes sense, if there are no quantifiers present)

<xalbo> I don't feel comfortable with {lo pa cipni} being a flock. I'm ok with saying of a flock that it {cipni}. I have not yet reconciled this...

<selpa`i> Yes, {lo pa cipni} can be a flock, or conversely a flock can be a cipni pa mei

<selpa`i> In my view.

<latro`a> that's philosophically robust but pragmatically awful, pe'i

<Rnuomer> well

<Rnuomer> is the flock birding together as one unit?

<Rnuomer> or do they each individually bird, as a mass?

<Rnuomer> a mass of birding things?

<Rnuomer> (shush me if I'm being dumb though =:x)

<latro`a> selpa'i's view is that the answer is "both", I think

<selpa`i> The fact that you consider it pragmatically awful when it is the cognitive/natural language approach is surprising.

<latro`a> my reason that it is pragmatically awful is that there is literally no way to make it explicit that you're talking about "one bird" in this framework

<xalbo> I think each of them birds separately, and so we have more than one thing that birds.

<latro`a> that's how I would think of it as well

<latro`a> {lo za'u cipni cu cipni} and {lo pa cipni cu cipni} are different types of statements to me

<Rnuomer> I think the issue is we haven't defined what birding is

<latro`a> that's what I was saying above with cribe: this is more about what the predicate means than what lo means

<selpa`i> One of the main issues is that this is about whether or not it *can* be defined.

<latro`a> true

<latro`a> there's also a question of local definition vs. global definition, if you claim that any definition at all works

<Rnuomer> I'd use like a checklist sort of thing

<latro`a> I don't claim global definition

<latro`a> but I do claim local definition

<selpa`i> In a ready-made universe, it would be. In the couterpart model, it would be considered impossible, because of an inifinity of nodes.

<Rnuomer> does it have feathers? check; does it chirp? check; etc.

<Rnuomer> oversimplifying but the idea is that

<latro`a> the issue is when it fails some attributes but clearly satisfies others

<Rnuomer> then I'd think there's a difference between typical & nessessary traits of a bird

<latro`a> similar to what I was describing, yeah

<latro`a> the problem is that you then have a rabbit hole

<Rnuomer> =:3

<xalbo> I tend toward prototype logic. I have in my head an image of the prototype "Bird", and things either fall into the empirical cluster in thingspace that triggers that concept, or they don't.

<xalbo> Things at the edges get iffy, and then I back up and start having to talk about what they actually are.

<Rnuomer> perhaps we can pick an easier example

<latro`a> we've written down three metaphors for the same thing

<selpa`i> I can introduce some other points, like "We all have the same Furby". (but let's hear out Rnuomer)

<Rnuomer> e.g. flying

<Rnuomer> I mean "what does it take to qualify as "flying""

<Rnuomer> we can clearly say that someone standing on the ground is in fact not flying

<Rnuomer> relative to the ground, anyway

<xalbo> .ie

<Rnuomer> so one of the conditions of "flying" would be uhh

<Rnuomer> not... standing on the ground?

<selpa`i> I child that's being held up into the air might exclaim "look mommy i'm flying!"

<Rnuomer> would he be, though?

<latro`a> (my view: {lo verba cu lifri lo ka vofli kei gi'e nai vofli}

<latro`a> but that's somewhat orthogonal to the general topic)

<xalbo> Yes, but I think that child would not be speaking truly. That same child might then say "Look, I'm a kitten" while scampering on all fours.

<Rnuomer> if the only thing in the checklist is "not on ground" then we'd call that flying

<selpa`i> And it will say "Look, I'm taller than you" while standing on a stage.

<Rnuomer> however, there are probably more things to test for for "flying"

<latro`a> interestingly

<latro`a> that one actually works in lojban

<latro`a> and not nearly as well in english

<latro`a> {mi galtu je nai clani zmadu do}

<Rnuomer> so you'd need enough items on the list to define what "flying" is

<latro`a> the problem with such a list is that the list elements have lists

<xalbo> I'd say that x flies iff x is in an atmosphere in a gravity well, and supported by the atmosphere and not by any solid object.

<latro`a> eventually something is primitive

<xalbo> Interestingly, vofli2 makes balloons not qualify, though my mental model of "flying" fits them.

<Rnuomer> can something fly through space, tho?

<xalbo> Under that model, no.

<latro`a> we internalize it as such, but the physics are actually completely unrelated

<xalbo> (Which means it doesn't match my use of the word "fly" either. Damn.)

<Rnuomer> also

<Rnuomer> I have a teddy bear on my bed, can we say that it is bear-ing?

<xalbo> I contend that it neither bears, nor {cribe}.

<Rnuomer> we could call it "le cribe" though

<latro`a> I just had a slight weird math-epiphany

<latro`a> {le} is unrelated to whether it actually bears

<Rnuomer> le blanu cribe

<latro`a> the epiphany was a neat metaphor

<selpa`i> I would postulate that the majority of branches indeed lack terminal nodes.

<xalbo> Does a bear cribe in the woods?

<latro`a> for this linguistic discussion along with a concept from probability

<latro`a> anyone care to hear it? I can give an intuitive description of the math

<latro`a> it'll take about a paragraph

<xalbo> Will it fit in the margin? Do tell.

<latro`a> you can imagine, without having to go through all the math, a process of diffusion in a force field

<latro`a> that is, a particle moves around randomly in space, but depending on its position in space it may be pushed more in one direction or another

<latro`a> you can now imagine labeling two distinguished regions A and B; A definitely has some property and B definitely doesn't

<latro`a> (the physical example is a chemical system, where A is definitely reactants and B is definitely products)

<latro`a> this diffusion system induces a function called a committor, which is the probability of getting to B before going back to A, from each point x

<latro`a> the committor is a "reaction coordinate", in the sense that as it increases, the system is "more B-ish", and as it decreases, the system is more "A-ish"

<latro`a> going back to the force field for a second, in physical examples the force field is the gradient of some energy, that is, the system tries to decrease in energy for the most part

<latro`a> now that we have an energy, we can talk about temperature; specifically, in these systems the committor depends strongly on the temperature

<latro`a> when temperature is low, the energy is the dominant contributor, the system stays away from high energy areas, and the committor abruptly goes from near 0 to near 1 as you pass over an energy barrier

<latro`a> when temperature is high, the energy is a less important contributor, the system goes pretty much everywhere, and the system gradually transitions from near 0 to near 1

<xalbo> temperature, in this case, is the amount of randomness in the motion of the particles?

<latro`a> right

<xalbo> je'e do'u continue

<latro`a> the application here is to consider A as "~P", B as "P", and T as a fuzziness parameter: P is more or less fuzzy depending on the size of T

<Visirus> I like this metaphor

<latro`a> if T is very low, P is essentially sharp; there's a very small "gray area" where P "is debatable", and otherwise everything's crisp

<Visirus> Very much

<latro`a> and the reverse when T is high

<Visirus> It's an inverse proportional relationship

<Visirus> It's like saying, vagueness vs precise meaning.

<latro`a> you can stretch the metaphor a tad further, and imagine the diffusion as your mind going about its process of figuring out whether to assign a given input x to A or to b

<latro`a> *B

<latro`a> for "low T" or an input near A or B, it's a quick process which almost always has the same outcome

<latro`a> for "high T" or input near the dividing surface, it can be a gradual process, and you sometimes conclude A, sometimes B

<Visirus> Why only A and B?

<Visirus> There can be other options

<latro`a> it could be n-ary

<Visirus> Yup

<Visirus> lojban therefore represents a sort of most probably logical instead of perfectly logical

<latro`a> but unless the predicates depend on one another you could probably call that diffusion in several separate binary systems at once

<latro`a> and yes, perfect logic requires perfect definitions

<Rnuomer> so in the syntax, there's no real difference between "lo ractu" and "lo gleki ractu?"

<latro`a> there's a tanru-parse in the second one

<latro`a> at top level there's not, at mid-level you can distinguish

<Visirus> But otherwise, no. It could be the T is high enough to mean either.

<Rnuomer> but the truth value conditions are the same?

<Visirus> Why not?

<Rnuomer> the seltau doesn't matter, right?

<Rnuomer> or do I understand wrong =:x

<Visirus> The seltau is telling you the area of T that it is more probable to be

<Visirus> Narrowing

<Visirus> lo mlatu includes lo cladu mlatu then

<Visirus> Additionally, imagine the T of a gismu being centered on it and the seltau narrowing the field. Then it's hierarchical.

<Visirus> Therefore you can say lo cribe goo and lo goo cribe and they don't mean the same thing.

<Visirus> Since lo goo cribe is an entirely different T than lo cribe goo, you can't refer to bear goo as just lo cribe. It's a different logical subsection

Day 2:

<Visirus> Expanding the field analogy a little bit, imagine something, say some goo on the ground.

<Visirus> Now, this goo is primarily a type of goo

<Visirus> or something or other having to do with goo

<Visirus> It is possible that it's from a bear

<Visirus> But, the goo itself is primarily a different fuzziness region, T, than bear

<Visirus> "bear goo" would then therefore lie outside the T of bear

<Visirus> And cannot be termed lo cribe

<Visirus> It's like electrons with different orbitals.

<latro`a> I'm not sure why you're using the letter T btw

<latro`a> T in the analogy was temperature, not a region of description space

<Visirus> That's what this is.

<Visirus> Probabilistic space.

<Visirus> Such as, an electron is most likely at point A but can be contained anywhere within region T

<Visirus> And is, and is not, simultaneously at all such points.

<Visirus> lo mlatu may or may not be lo cladu mlatu

<latro`a> I know, I'm talking about the symbol T

<latro`a> it's taken already

<Visirus> Fine, call it Pspace

<Visirus> lol

<latro`a> that's also taken, lol

<latro`a> albeit by computer scientists

<latro`a> but really, T is an important aspect of this, because it has to do with how sharply you care to delineate regions of the description space

<Visirus> Not at all

<latro`a> "hot" discussion is metaphorical, fluid, open, informal; "cold" discussion is rigid, crisp, logical

<Visirus> It's just incredibly unlikely that lo mlatu is lo gerku so it's negligible to the point of practically being 0

<latro`a> I think we're talking about different things...

<Visirus> Yes.

<Visirus> I'm talking about a probabilistic model.

<latro`a> I know, but the model naturally includes a temperature

<latro`a> the volatility, informality, whatever you want to call it is intrinsic

<latro`a> sometimes we want to have crisp, clean definitions; other times we don't

<Visirus> This is more akin to a quantum field theory I think.

<latro`a> that would also have a temperature once you go to the thermodynamic limit

<latro`a> I'm saying that there's not a fixed probability measure

<latro`a> when you're joking around among friends, terms blur and mix more freely than when you're in a courtroom

<Visirus> So, you see, you can define mlatu as being fully defined at whatever A is in English.

<latro`a> that's contradictory

<Visirus> Wait

<Visirus> But

<latro`a> English is subject to this same probabilistic interpretation, if not more

<Visirus> It can exist anywhere within the field of lo mlatu

<latro`a> you can't grab a natlang to use as a base

<Visirus> Ok

<Visirus> mlatu

<Visirus> the A point of it

<Visirus> the end node

<Visirus> It doesn't matter the language

<Visirus> In lojban, lo mlatu includes all lo seltau mlatu

<latro`a> (as an aside, quantum is not purely probabilistic; if that were the case, transitions between observable states would be impossible)

<latro`a> but yes; broda is a less crisp region of description space than brode broda

<Visirus> Well, quantum tunneling is what electrons do to jump energy levels.

<Visirus> Yes

<Visirus> Now,

<latro`a> that's not a particle effect

<latro`a> it's a wave effect

<latro`a> which is why it's not pure probabliity

<Visirus> It's probabilistic is the point.

<latro`a> (also, not every quantum transition is a tunneling process)

<latro`a> (tunneling is a rather specific type of process where a nonclassical transition occurs)

<Visirus> You're pointing out irrelevancies.

<latro`a> s/nonclassical/classically forbidden

<latro`a> sorry, my remark was just an aside that you replied to :)

<Visirus> Ok

<Visirus> mlatu is defined at whatever point A may be

<latro`a> I would interpret a given selbri as itself being a potential

<latro`a> in this model

<Visirus> If something has a high probability of lying within the lo brode mlatu space, it's a lo mlatu

<latro`a> it's not "defined at a point", instead it's a potential on the whole space

<Visirus> But mlatu itself is defined at a point.

<latro`a> perhaps, perhaps not

<latro`a> depends on if you claim that there is a crisp region at all

<latro`a> with mlatu in particular there probably is, but with other selbri this may not be so obvious

<Visirus> Under this, if something is observably primarily something and you call it that, then you can't take out the tertau

<Visirus> er

<Visirus> seltau

<Visirus> lo goo cribe

<latro`a> sure; seltau tighten the potential

<Visirus> Yes

<latro`a> but a different tertau gives you a different potential altogether

<latro`a> with different structure

<Visirus> YES

<Visirus> My solution to the bear goo problem.

<latro`a> it's not entirely a solution, because you have to get people to agree that the goo is or isn't a bear

<Visirus> It can be a bear type of goo

<latro`a> it can also be a goo bear

<Visirus> but if you look at goo and call it a bear, you'd better have a damned good reason

<latro`a> which is the whole problem

<latro`a> yes

<latro`a> but xorlo basically suggests that the reasons don't have to be as good as you might exepct

<Visirus> Without explanation, you cannot change the potential

<latro`a> *expect

<latro`a> given context

<latro`a> well, it changes itself

<latro`a> that's the difficult part

<Visirus> Because one would expect something to lie within a certain potential

<latro`a> T goes up and down with context, and terms even shift in their meaning, which changes the potential

<Visirus> If you change it all willy nilly like, they'll be, obviously, confused.

<latro`a> sure

<latro`a> on the other hand, if you define {lo broda} as {zo'e ne lo* ka broda}, where "lo*" is a magic thing that makes a ka like we normally use it, then it's not confusing

<Visirus> So, one must always use the most obvious potential based on as little outside context, unless it's already given that both parties know such context.

<latro`a> since bear goo does in fact have something to do with being a bear, even if it isn't itself *actually* a bear

<latro`a> eh, that doesn't exactly fix it, though, because we don't talk about the potential directly

<Visirus> You'd have to think about it.

<latro`a> consider selpa'i's example from yesterday

<Visirus> If I know it's bear goo but you don't, it's almost intentionally confusing to call it lo cribe

<latro`a> if a flock of birds {cipni}, and you see one flock of birds, then you saw {lo pa cipni}, even if the flock had 10 birds in it

<latro`a> and yes, of course there's deceptiveness, that didn't need a probabilistic interpretation to be concluded :)

<Visirus> The probabilistic interpretation makes so much sense though imo

<latro`a> it helps, yes

<latro`a> but really the end point here is "be communicative"

<latro`a> which doesn't need any formalism whatsoever

<Visirus> A computer could use the probabilistic engine to determine better translations for ideas natlang <-> lojban

<latro`a> if one person thinks {cribe} means "living bears" and the other thinks it means "anything having to do with bears"

<latro`a> then they're not being communicative

<Visirus> The person thinking living bears is wrong then

<Visirus> because that's a seltau

<latro`a> not...exactly

<latro`a> I'm using english as metalanguage here

<latro`a> so don't gloss {cribe} as "bear"

<latro`a> you can restructure the description space so that {cribe} is "living bears" and {cribe morsi} is "bear corpses"

<Visirus> a dead bear has a potential of being called lo cribe and lo xadni.

<latro`a> that depends on the structure of the description space being used by the person

<Visirus> The potentials are so close though, because of the nature of the vagueness of the thing, that it's a choice.

<latro`a> you draw that conclusion from natlang interpretation more than anything else mio

<latro`a> *imo

<latro`a> there's no particular reason why bear corpses *must* be bears

<Visirus> The potential for the thing you're naming

<latro`a> is the one that you have in your mind

<Visirus> It has potential to be other things

<latro`a> not theirs

<latro`a> that's the whole problem

<latro`a> one person's potential may rise sharply when you pass over into the "dead" region

<latro`a> the other's may not

<Visirus> Yes, consider all, or as many as possible, and determine the most likely based on as little context as possible. Only immediate observables.

<latro`a> if you can

<latro`a> the problem is that this formalism doesn't help you perform that "figuring it out" process

<Visirus> A computer could use it to better translate things

<latro`a> maybe; they have to have information about attributes that make things more bearish or less bearish

<Visirus> Using a sort of tag cloud format

<latro`a> which ultimately comes down more to something like Rnuomer's checklist, much different from xalbo's "prototype" model

<Visirus> Go on...

<latro`a> which for a computer would be more neural network: when presented with a bear-candidate, what fires? how does this compare to something that we definitely call a bear?

<latro`a> for example, to me a living bear is more bearish than a bear corpse

<Visirus> Yes

<latro`a> even a fresh one

<latro`a> I'd still call a fresh bear corpse a bear

<latro`a> but my potential has gone up by that point

<latro`a> then as it decays it goes up further, and sometime before the point where I can't even tell it was a bear, the potential is so high that it's not worth thinking about

<Visirus> Then you can't call it a bear

<latro`a> by that point, sure

<latro`a> I'm describing my potential, though; others' potentials are different

<Visirus> The area in between is the fuzzy

<latro`a> I think selpa`i's potential rises less sharply as the bear dies

<latro`a> based on our discussion

<latro`a> in general I think selpa'i is "hotter" than I am, in this formalism

<Visirus> If the world were the movie Equilibrium, this would be no issue.

<latro`a> na slabu

<Visirus> You can't remove all the uncertainty, but you can diminish most of it. Definitions need to be specific, or people may speak with the knowledge that no matter what, they'll never be able to completely remove the fuzziness from the meaning.

<Visirus> Meh

<Visirus> .i mi xagji

<selpa`i> To me it's very difficult to priorly define a personal scale of potential, as everything is highly sensitive to context; the psychology isn't static throughout time. Making up a scale here and now is to some extent futile (or requires a *lot* of imagination and foresight) as the universe "collapes" time and again and needs to be re-differentiated each time.

<latro`a> one nice thing about this model to me is that the actual *potential* changes much more slowly than T

<latro`a> at least for me

<Visirus> Remove context

<selpa`i> Impossible.

<latro`a> I may fluctuate in how much I care about the boundaries between concepts

<selpa`i> And undesirable at least for me.

<latro`a> but the boundaries themselves (in the sense of the potential, not sharply delineated regions) move slowly

<latro`a> for example

<latro`a> the fact that a bear corpse is less bearish than a living bear

<latro`a> is an invariant for me

<latro`a> the idea that a bear corpse *is a bear*

<latro`a> is not

<latro`a> the probability is always lower, but it could be a difference of 1 vs. 0.9 or 1 vs. 0.5

<latro`a> the tricky thing about all this is that there is SOME effective nonexistence of context

<latro`a> er

<latro`a> nonrelevance I guess

<latro`a> if there weren't we would never be able to communicate

<selpa`i> Or we grow up learning our language in a context.

<selpa`i> Which might be the same thing effectively.

<latro`a> not....exactly

<latro`a> context isn't relevant if it's constant

<selpa`i> That's basically what I said (meant).

<latro`a> I'm saying that there are some basic assumptions that are so absurdly hard to break that it doesn't matter, or at least so it seems

<latro`a> if there weren't, we wouldn't be able to depend on those assumptions to communicate

<latro`a> a blunt example: "assumption" does not mean "fish", ever

<Ilmen> lo se sruma / lo finpe

<latro`a> perhaps one sensible assumption is that the potential is finite on a bounded region, where the bounds are invariant

<latro`a> that is, there are some things that might in a bizarre context be bears, but aren't ruled out a priori

<latro`a> and some things that are usually bears, and some things that are always bears

<latro`a> and then everything else is never ever a bear-

<latro`a> on a more practical note

<selpa`i> Could your model be called a dynamic-range-but-definitely-always-some-endpoint Ready-Madeist view?

<selpa`i> Since your dead bear scale is flexible-ish, but always has some endpoint.

<selpa`i> And this would fit with your idea of there being things that can never ever cribe

<latro`a> especially with something that's not even done

<latro`a> "perhaps one sensible assumption"

<latro`a> I didn't postulate anything >.>

<selpa`i> No, sorry, I didn't even refer to your last idea

<latro`a> then there's absolutely nothing to get the static endpoints from

<latro`a> because prior to that I'd only said that there are "practical static endpoints", which means it's not in the model at all

<latro`a> just a consequence

<latro`a> so, no, don't call it that

<latro`a> anyway

<latro`a> on a more practical note

<selpa`i> I vaguely remember you saying that at some point, a cribe stops cribe'ing absolutely (though not in those words).

<latro`a> eh, I try to avoid fatci and its english counterparts

<selpa`i> Even if you didn't settle on anything.

<selpa`i> I'm just trying to comment on those points

<latro`a> I'm not sure whether that postulate should be built into the model, or if you should instead have an unbounded potential that just tails off for most predicates

<latro`a> so don't quote me on it being built in

<latro`a> because it's not

<latro`a> ANYWAY

<latro`a> been trying to change the subject for 5 minutes

<latro`a> I'd like to try and work out the {lo du'u mi viska pa loi za'u cipni cu nibli lo du'u mi viska pa lo cipni} thing

<latro`a> "I see a flock of {cipni}; a flock of {cipni} {cipni}'s; therefore I see one {cipni} (namely, the flock)"

<latro`a> provided {loi za'u cipni cu cipni}, everything else passes through

<latro`a> I should have said "one flock", however, not "a flock"

<selpa`i> Interesting, you seem to be taking this whole thing from a whole 'nother angle.

<selpa`i> This is an entirely different dimension of the "I see one bird" thing.

<selpa`i> In my example, it was about slicing up the universe in such a way that a flock of birds (all of a single species probably) are seen and described as a single bird, because in that particular universe the distinction between those individual flock members doesn't exist.

<selpa`i> There literally is only one bird there.

<selpa`i> This is what happens in a non-ready-made universe. The universe always starts out as a big clump, and can be sliced up in infinitely many ways, and then stuff happens post-differentiation.

<selpa`i> In a ready-made view, the universe gets sliced up once and never collapes again.

<selpa`i> In non-RM, it always goes back to a clump.

<latro`a> I know

<latro`a> but *even in this view*

<latro`a> you can have it that single birds {cipni} and flocks of birds {cipni}

<latro`a> then take a bunch of single birds, put them in a group

<latro`a> say that the group {cipni}

<latro`a> see the group

<latro`a> and now say that you saw only one thing that {cipni}

<selpa`i> lo pa tadni cu sruri lo dinju

<latro`a> indeed

<selpa`i> It's a good point.

<selpa`i> It's a somewhat related, but really quite distinct phenomenon, not really hinging on any ready-made talk.

<latro`a> it definitely doesn't require any ready-made hypotheses

<selpa`i> Right.

<latro`a> all it requires is that in a given context you accept that a group made up of brodas is a broda

<selpa`i> With {loi} things seem a bit unsettled, but you can do this with just {lo}.

<latro`a> with {loi} it depends a bit more on the predicate, arguably

<selpa`i> {loi} having the problem of possibly adding properties (or removing) from the single broda

<latro`a> but I would be inclined to agree with it for {cipni} and such

<latro`a> at least, naively

<latro`a> this "gotcha" makes me hesitant, but if I hadn't considered it, I would have no issue with {loi cipni cu cipni}

<selpa`i> lo ci cipni cu cipni .i pa lo cipni cu go'i

<selpa`i> why not pa cipni cu go'i

<latro`a> uhh

<latro`a> avoid go'i, please

<latro`a> because attempting to answer your question confused me

<latro`a> you replaced the only sumti that was filled

<latro`a> so it wasn't clear whether {go'i} was actually just {cipni} or "the previous sentence's cipni"

<latro`a> in idiomatic lojban it'd be the former if all the sumti were replaced

<latro`a> at any rate, {pa da cipni} definitely doesn't happen

<latro`a> but you could group the universe such that {pa da cu cipni gi'e gunma}

<latro`a> that's the problem with masses, the speaker is free to build and dismantle them

<latro`a> also, these outer quantifiers play differently with "cognitive" predicates vs. "noncognitive" predicates

<latro`a> for example, if I see a flock of 10 birds, {pa cipni cu zvati} is false, but {mi viska pa cipni} can be true

<latro`a> if I see the flock but can't pick out individual birds

<latro`a> (maybe make it 1000 birds)

<selpa`i> You can look at the flock, see individual birds, and still claim that pa cipni cu zvati (because extra birds don't add a count to how many different birds you perceive), that's the example I explained earlier.

<selpa`i> It's most simple to imagine (I think) if you let all the flock be eagles, then an eagle more or less doesn't change that there is just one bird, namely the eagle.

<latro`a> that's deceptive at best

<latro`a> you observe {pa cipni cu zvati}, but you're wrong, because the components are also birds

<latro`a> a problem is that in fact a very large number of {cipni} are present

<latro`a> supposing there's 10 birds in a flock present, then you have the 10 singletons, the 45 pairs, the 120 triples, etc.

<latro`a> so many hundreds of {cipni} are "present"

<latro`a> because every subgroup *exists*, even if not every subgroup *matters*

<latro`a> in fact an even larger number of subgroups *exist* when you start allowing for bird goo; for example, a whole bird+another bird's liver is perhaps a {cipni} too

<selpa`i> See, this is why I keep thinking that you are RM-ist. And this is not in any way meant in a bad way, it's simply a different perspective. Either you can't perceive the universe non-RM, or you just find it horrible. Which is it?

<selpa`i> Because I am trying to explain that they don't, in a way, exist.

<ksf> rm?

<ksf> and latro`a is completely right from a set-theoretical POV btw.

<selpa`i> Of course.

<ksf> ...assuming that birds are distinguishable, though.

<selpa`i> Math is usually ready-made.

<latro`a> I can understand it, but I don't see the problem in regrouping in a non-RM setting

<latro`a> 1) there are 10 birds, as we understand it in english

<selpa`i> Can you imagine there being a universe where number doesn't exist?

<latro`a> 2) groups of birds are birds

<latro`a> er

<latro`a> groups of birds are {cipni}

<latro`a> conclusion: >1000 {cipni} exist

<latro`a> and while I can imagine it, it's sufficiently impractical that I don't really care to bother

<latro`a> xorlo isn't worth sacrificing outer quantifiers as a concept for

<latro`a> nothing is, really

<latro`a> we need them to be communicative

<latro`a> there's a point when I stop caring about the philosophy of all this because it's so far down the rabbit hole that it doesn't mean anything anymore

<latro`a> I feel the same about most of the attempts that have been made at formalizing subjunctivity

<latro`a> anything that involves outer quantifiers not meaning what they should mean is so far down the rabbit hole that it's gone to china and back 1000 times already

<latro`a> going back to my example

<latro`a> supposing there are 10 birds the way we mean it in english

<selpa`i> What they should mean? They still do what they do, namely they quantify over something. They don't tell you what the domain of discourse is, or about cardinality, but why should they?

<latro`a> why are there only 10 {cipni}, if we acknowledge that groups of birds are birds

<latro`a> the problem is that you can't change the domain of discourse

<latro`a> so we have to have a sane one

<selpa`i> I can't change it?

<latro`a> there's no explicit way to set it, no

<latro`a> not in lojban

<selpa`i> So?

<selpa`i> There is always one.

<latro`a> that means you need a sane one

<selpa`i> Isn't "sane" extremely subjective?

<latro`a> yes, hence the whole probability discussion

<ksf> as soon as you equate singletons and sets you get every imaginable kind of decidability problem.

<latro`a> but one in which "there are 10 birds" means "there are 10 possible regroupings of birds" is not sane

<latro`a> period

<ksf> "sane" isn't subjective when what you're saying triggers the halting problem.

<latro`a> and I don't see why, even in a non-RM setting, the speaker shouldn't be allowed to freely regroup things

<ksf> what's the problem with using cmima, anyway?

<latro`a> if you can regroup things and also can express that there are 10 birds on a branch in the sense that english means, then you're going to have to have that a flock of birds isn't a bird

<latro`a> sets are awkward as hell for a lot of reasons

<latro`a> they don't actually do anything other than cmima

<latro`a> and se mei

<latro`a> they *encode* other things

<latro`a> but that's indirect

ksf says: If you want to speak about permutations of sets of birds, speak about permutations of sets of birds, not birds.

<selpa`i> On IRC, as noted, RM is the prevalent mode.

<latro`a> the issue is that we've acknowledged that {lo za'u broda cu broda} and that {lo za'u broda cu [do something that they can only do as a group]}

<latro`a> and that this is at the same level of predication

<latro`a> (I've proposed having distinct CU for different levels of predication more than once)

<latro`a> (I've never liked masses-as-sumti)

<latro`a> (it's orthogonal to the real issue, which is how the plural entity behaves)

<selpa`i> (if a mass has other properties, than it is good to have it as a sumti)

<latro`a> that's one way to look at it

<latro`a> but there's a different one

<latro`a> namely that it has different properties *in different senses of "have property"*

<ksf> ...you don't have to equate sets and single birds for that to work. you just have to have a cardinality function that's defined over the intersection of `Set a` and `a`

<latro`a> a plural group of students has a collective property of surrounding the building and an individual property of screaming

<latro`a> the big gap that this opens is that it may not always make sense to assume that a mass is openable

<latro`a> but then you just respond to an individual statement with {na'i} and all is well

<ksf> if you want to do some mathematical woo-hoo to get formulas to look nicer you can define "cipni" to be "set of birds with cardinality 1".

<latro`a> you're not going to be able to do this with sets

<latro`a> you can go ahead and give up on that

<tsani> Hm.

<tsani> I've read all the backscroll and I have an idea.

ksf thinks his definition is mathematically sane and not actually different from what latro`a wants

<tsani> It seems like we can preserve the formal definitions of the gadri proposal, even if we throw "loi broda cu broda" being true out the window.

<latro`a> ksf: not really; there are fundamental obstacles to actually building this sort of thing up from set theory

<latro`a> in particular the old "what are the members of pi?" question

<latro`a> set theory is terrible

<latro`a> it's an awkward foundation that is really only nice because it shows that you don't have to assume that much to get off the ground

<ksf> oh, I was talking about sets as a data type. I'm into type theory, myself.

<latro`a> and then godel showed that you don't even get off the ground anyway

<tsani> If we consider the definition of the inner quantifier, {lo PA broda} = {zo'e noi ke'a broda zi'e noi zilkancu li PA lo broda}, it turns out that inner quantifiers just need to "quantify" over instances of {lo broda}

<tsani> Then, we say that {lo broda} can just *produce* any type.

<latro`a> we actually talked about that yesterday

<latro`a> and noted that it's disastrous

<tsani> The Gadri Proposal therefore implies selpa'i's earlier statements.

<latro`a> for the exact reasons we were talking about

<selpa`i> tsani: What do you mean by type?

<ksf> http://en.wikipedia.org/wiki/Intuitionistic_type_theory

<tsani> In a way you'd phrase it, 'lo doesn't have a type'.

<latro`a> ksf: yes yes, I know what you're going for, but that doesn't work either

<selpa`i> If you're saying that {lo} can stand in for {lo'i}, then I'd disagree.

<latro`a> language is more complicated than a robust formalism

<tsani> selpa`i: that doesn't really matter right now.

<selpa`i> Okay, "no type" doesn't sound like "producing any type"

<selpa`i> I'm really only trying to understand you.

<latro`a> you're speaking from different formalisms

<latro`a> as in, your protest to "can produce any type" is relative to an untyped formalism

<latro`a> afaict

<tsani> You reject types throughout Lojban, don't you, selpa'i, so I'm surprised that you're not just agreeing.

<selpa`i> Let me explain.

<selpa`i> I can make it quite concrete: I don't consider {lo ro ninmu} in {do melrai lo ro ninmu} to "stand in" for {lo'i ro ninmu}.

<selpa`i> Instead, my definition of {traji} doesn't use a set.

<latro`a> (btw, ksf, I don't want to be rude, but I'd appreciate if you stopped with the set theory/type theory discussion; it's been beaten to death by this community already, and it's long since accepted that it's not adequate)

<tsani> Yeah, that much I know.

<tsani> We've all basically thrown lo'i out the window.

<ksf> well, at least type theory has a chance of describing predicate logic sanely, but I understand.

<latro`a> predicate logic isn't adequate, either

<ksf> and there goes the myth :)

<latro`a> that's part of the point of this discussion

<latro`a> (that's also long since accepted, afaict)

<latro`a> (so yeah, not adding anything here)

<tsani> If {lo broda} is explicitly not explicit about distributivity (in other words, it can produce individuals or collectives) then the zilkancu equation of the gadri proposal lets {lo pa cipni} be a flock of birds *without saying that a flock of birds is a bird*.

<tsani> (That's the main difference that I'm pointing out.)

<latro`a> I still question whether that's what was actually intended

<selpa`i> tsani: Depends on what you mean by types again. I don't reject the idea that a du'u cannot dacti, so clearly I agree with types, so what exactly do you mean? If you mean that I find rigid-typing a bit inconvenient (i.e. sumti places being super limited in what they can take), then yes, I do.

<latro`a> in the zilkancu equation

<tsani> I agree that it may have been unintentional on xorxes's part, but to be honest, and being somewhat acquainted with his ideas by proxy of selpa'i, then I'm inclined to think that this awkward consequence was intentional.

<latro`a> also

<latro`a> an obvious terrifying corollary

<latro`a> {lo ci cipni} can be 3 arbitrarily nested bird groups

<tsani> yup

<latro`a> it could even just be three different ways of breaking up all of the birds into groups

<ksf> what about ditching distributivity in favour of polymorphism?

<tsani> Also, that being said, {lo broda} can produce things (in one context) that do not {broda} (in another context).

<latro`a> that's "explicitly not explicit about distributivity"

<latro`a> already done

To be continued?...