Freudenthal's critique of Loglan

From Lojban
Jump to navigation Jump to search

Read pdf online.

by H. Freudenthal.

James Cooke Brown, Loglan, a logical language. The Loglan Institute, Gainesville, Florida, 1966, Mimeo­graphed. vii, 222 pp.

Inventing a language is a tedious business. What the author experienced as flashes of thought, will not be recognized by the reader as such. Of course the author always believes that his system works because as soon as it does not, he is ready to revise it. With the same right the reader, less committed, is allowed to be more pessimistic. The referee is left with the hardest task. It would be a silly enterprise to quote out of the inexhaustible amount of details, but it would not be easy to find enough major points from which something like a total view should be possible. The only attitude I can imagine is criticism, not to belittle the effort but to show that it is taken seriously. In fact a review on a language invention which is not utterly critical, must be either superficial or non-serious.

Loglan should be a logical language according to the author.

Clearly he does not mean the word ‘logical’ in a technical sense, but rather in the vague and illogical sense it has in ordinary speech. To a quite modest degree the author has been influenced by logistics; nowhere this influence has been profound. This, probably, is the most severe criticism against his design.

Without doubt the author borrowed from logistics the idea to deal with 3. 11 content words (that is all except structure words) as predicates, and to state clearly how many free places a particular predicate contains and in which order they are arranged. Also from logistics he took the requirement that all subjects be free or bound variables.

There are a few reasons why he did not succeed in realizing this program. First, he did not clarify the logical status of the so-called modifiers (adjectives added to a noun, adverbs added to an adjective or an adverb, and so on). Do they mean that a predicate has hidden free places which are not accounted for in the official number? Are modifiers to be considered as predicates or as variables or, perhaps, as structure words?

Secondly, though naming variables is regulated by more effective rules than in natural languages, it still suffers from the usual this- that-rule - ‘this’ referring to the last mentioned subject and ‘that’ to the one but last.

Thirdly, in dealing with variables, there is no indication that the author has understood the basic character of binding procedures, whether these are bindings by quantifier, article binding, inter­rogative binding or demonstrative binding (compare the reviewer’s Lincos). Notwithstanding many improvements in details, the binding techniques are as implicit, as weak and as logical as in natural languages.

The problem of modifiers, mentioned as point one, can only be solved in the frame of a satisfactory binding technique. In many other cases the lack of insight into binding techniques has led to absurd solutions. According to the author a predicate A always means ‘capable of being A’, unless it is modified by a tense word (such as ‘now’ for the present tense), which is not counted as a free place of A. So 'burns’ means ‘flammable’, and to express 'is burning’ one has to say ‘now burns’. (It is not clear how ‘was flammable’ and ‘will be flammable' has to be translated. ) A correct solution would be to recognize a time variable in every realistic predicate and to create good techniques for different kinds of binding. Obviously the author's solution is greatly influenced by English idiom - often the present tense, if confronted with the gerund construction, means a capability instead of an actual event.

The presentation of Loglan suffers from dogmatism. It is evident that at every step the language builder has had to choose between two or more possibilities. In important cases one would like to know why the author chose one rather than another, but such motives are never explained. For instance, why are all adjectivic predicates assumed to mean comparatives? I cannot imagine any reasonable answer. Why do modifiers precede the modified word? Because tape worm English (pretty small girls school) suggests this quite illogical construction? Of course such a rule cannot be maintained to the bitter end. So the author has to introduce structure words which allow the modifier to be placed after the modified word

Such ‘conversion' connectives which indicate the permutation of free places, are, in general, a healthy idea. So are the spoken punctuations (e. g. in ‘pretty small girls school’) though even in punctuation the author could have learned more from logistics. Much care is bestowed on a phonetic system which clearly indicates the separation of clauses into words and the belonging of a word to a word class (structure words, content words, proper nouns and so on).

The presentation suffers not only from a lack of reasoning, but still more from the total absence of connected texts. I would have liked to understand the author’s technique of building abstractions which is exposed twice (p. 38 and p. 71). Abstraction is a complex problem which cannot completely be solved by the formal means of logistics though some knowledge of logistics can be quite a help. On the other hand understanding the less formal part of abstraction presupposes long illustrative texts, which simply are absent.

There is no glossary of structure words. The glossary of the 1000 odd content words may be a source of many questions, in particular the question how these notions have been chosen. One is struck by a lot of rare and involved notions whereas the most fundamental and most primitive ones are lacking.

The present pamphlet is called a preprint edition. It is to be hoped that obviously wrong decisions will be revised in the definitive print.

University of Utrecht

H. Freudenthal