To give a more detailed insight into the ideals (and controversies) behind the founding of GLOW, we reproduce in these pages some of the discussions from the first issue of the GLOW Newsletter (September, 1978). Below is Jean-Claude Milner’s response to the GLOW manifesto.
Milner’s reply
The [news-]letter presenting the GLOW organisation contains in section B.1 (Concerning the object of inquiry) some propositions of general bearing. I do not agree with a certain number of them.
The doctrine of the letter can be summarized in three related points:
- (A) Linguistic theory must be interpreted in “realistic” terms; i.e., there is a reality [R1] of the language faculty and languages,2 which linguistic theory aims to express adequately.
- (B) The reality of the language faculty and languages has essential psychological substance.
- (C) A psychological reality is a specifiable state of a mental organ.3
The author of the letter formulates explicitly and justifies only (A); (B) and (C), on the contrary, are implict: they have to be reformulated, however, if the initial definitions of B.1 are to make any sense. In the eyes of the author of the letter, (B) and (C) are really a natural extension of (A) and their sole function is to clarify the interpretation of (A). This is why justifying (A) is allegedly sufficient to justify (B) and (C).
My general position is completely different: None of the propositions (A), (B), and (C) is trivial; on the other hand, even if (C) depends on (B) and (B) depends on (A), the inverse implication does not hold. Suppose one admits (A), one still has to prove (B), and suppose one admits (B), one still has to prove (C). The letter of GLOW does not say anything about these points and does not make any reference to where the required proofs could be found.4 This is at the very least a serious omission, but has even more regrettable consequences.
In order to make them clear, let us consider (A), (B) and (C) separately. Proposition (A) is treated correctly in the letter: (A) is effectively crucial to the interpretation of linguistic theory. It is equally exact that the decision to accept (A) cannot draw its justification from empirical data; it is about an epistemological choice, anterior to any treatment of data because it determines how such a treatment should be interpreted. I want to add at last that I myself make the same choice as the author of the letter: I accept (A).
Accepting (A) does not predetermine answers it would be given in the question of substance. On the contrary, it is the peculiarity of (B) and (C) that they voice hypotheses touching upon the substance where the values of R are defined. One can thus see in which sense (B) and (C) are absolutely not implied by (A).5 Moreover, it is not even evident that it is necessary, for the one who admits (A), that the question of substance should have any sense. From this point of view, (B) can be analysed in two sub-propositions:
- (Ba) One can and must attribute a certain substance to the values of R.
- (Bb) In the case of speaking, this substance is psychological.
Only those who subscribe to (Ba), can give a certain interest to propositions like (Bb). But (Ba) is non-trivial: one can perfectly imagine that a linguistic theory be comparable to Newtonian astronomy, and in consequence suppose a reality of its object, without voicing hypotheses about the substance of the latter.6 In any case, this would allow us to explain why two linguists can communicate in a precise and “realistic” way without agreeing on either (Bb) or (C).
The author of the letter implicitly admits (Ba). But one would wish for more consciousness that (Ba) does not stand by itself and above all, that nothing of the realistic character of linguistics is involved here. More precisely, a difference in opinion concerning (A) entails sensible differences between linguists: it is the notion of theory itself that is concerned here, as well as the notions of valid argumentation, of refutation, of confirmation etc. This, by the way, has been aptly stated in the recent history of the discipline. On the other hand, it is not easy to discern what consequences a difference concerning (Ba) would have, from the moment on that two linguistis adopt the same theory and accept (A).
Let us consider (Bb) for the moment. It should be even more clear that this is a non-trivial proposition. For there exist a wealth of types of substance other than psychological. It might be useful to recall here that Saussurian linguistics was the only one to propose somewhat sophistic conceptions of substance. Some of the questions it asked exhibit a certain pertinence in this regard: what is the substance of a daily train (e.g., the 8:47h train Paris — Amsterdam), admitting that each coach of this train could change every day? What is the substance of the dollar? And, if necessary to give a universal example: what is the substance of the prohibition of incest? Each of these questions allows for a different answer — which at least proves that the notion of substance of the language [faculty] is similar to one of the just evoked; but the contrary is not further evident. In any case, in the present state of our reasoning, no privilege attaches to (Bb): this proposition is not clearer, nor more “realistic”, nor more fertile, nor more reasonable, nor more founded than any other rivalling proposition, of the type: “language (speaking!) has the same substance as the prohibition of incest”.7
If one rejects (Bb), (C) is evidently deprived of any pertinence. I simply want to add this: in itself, (C) is nothing but a nominal definition, that is, a terminological decision. From this point of view, (C) is neither refutable nor confirmable. The only thing to resolve is if the decision to baptize an ensemble of psychological realities “mental organ” is a sensible decision.
From the point of view of clarity, nothing is more evident. If one gives “organ” its usual sense, it goes without saying that a mental organ is a contradiction in terms. If it is necessary, however, to give “organ” a new sense, the question is: which? The text of the letter remains silent about this point: another regretable omission. If, out of good will, one tries to fill this silence, one orients oneself towards a new definition of the term “organ” that does not seem to be sensibly different from the definition one would give terms like “structure”, “proess” or “apparatus”.
I would certainly not have the simplicity to underline the major differences that separate the language [faculty] from an organ in the usual sense, e.g. the heart. One could suppose that the author of the letter does not ignore them, and these [(differences)] are without doubt those that have lead to qualifying the term “organ” with a differential predicate “mental”. That this adjunction of this predicate is nothing but word-play, it is this I am not convinced of. But above all, the use of terms like “mental organ” is not justified because the terminological innovation suggests that the similarities of language [faculty] and heart are more numerous than the differences; once again, this is a non-tivial hypothesis. As long as it has not been proven — at the moment, it has not, and in my eyes, quite plausibly so —, the terminology of the letter does not even have the advantage of clarity and exactness.
Obviously, adopting (C) permits the definition of linguistics as a branch of the natural sciences, and especially of biology. But this result is only interesting if it permits a better definition of the methods effectively in use in the construction of a linguistic theory. However, until now, the modes of argumentation, of falsification, of invention used in the linguistics produced by Chomsky do not seem to have any particular similarity with what is in use in the natural sciences. The only discernable similarity is the one that would exist anyway between the sciences that want to be empirical and falsifiable — be they natural sciences or not.8
Concerning the belief that the value of (C) importantly permits the costruction of a vocabulary for linguistics homogenous with the one of the natural sciences, this is a fetishist opinion which I am refusing to attribute to the author of the letter. A science defines itself by its principles and its methods, but not its vocabulary.
In the guise of a conclusion, I will comment on a sentence from the GLOW letter: “the generative linguist regards the principles that determine the class of possible human grammars as a genetically based property of the human species” (p.1 -omitted9). One could consider whether the whole of section B.1 is only an elaboration of this sentence. Proposition (A) is one way to define the notion “generative linguist”; propositions (B) and (C) justify the conception of grammatical principles which is summarized here.
The cited sentence — let us call it (D) — allows only for two interpretations: either it expresses a logically necessary relation, of the sort “a generative linguist cannot negate (D) without contradiction”, or it describes a historical situation: “all generative linguists accept (D)”. As to the first interpretation, it follows from what has just been said that it is not true if (B) and (C) are logical consequences of (A). But there is nothing like that: in terms of pure logic, a linguist can well be generative and not adhere to (D). In particular, he can perfectly well assume that the universal principles that constrain the possible grammars are not essentially different from the prohibition of incest,which constrains a class of possible human societies, and which does not seem to be interpretable in a reasonable way as information inscribed in the genetic code of the species.
As to the second interpretation of (D), it is directly false: there is at least one generative linguist, namely me, who rejects (D).
One could certainly not expect from a letter of presentation that it develops proofs and justifications. Nevertheless, it should be normal that its author be sufficiently conscious about the non-trivial character of certain propositions in order to make clear that proofs or justifications would be necessary.
On the other hand, it is true that not all propositions require an empirical proof or justification. Thus, (A) receives its justification not from the fact that it has been empirically proven, but from the fact that two linguists who do not agree about (A) do not agree about the rest. But what is true of (A) is not true of (B), nor of (C), nor (D).
Finally, it is perhaps acceptable that the majority of generative linguists share a certain number of philosophical opinions, but they should be conscious about which these are. Now, in the absence of justification, (B), (C), and (D) are nothing but: opinions.It is thus not acceptable that these be presented like conditions for the membership of a group; — if at least the latter is to preserve a scientific character.
In brief, it is never good for anyone to give a reasoning the logic of which, how little it be, evokes the one of Lyssenko.
Jean-Claude Milner
— Translated by Anne Breitbarth, U.Tilburg
(Note: footnotes in italics are not in the original and have been added for clarification. -CT)
1 added for clarity, as he refers back to “R”. CT
2 This phrase is a translation of “langage et des langues”; we assume it to mean the human capacity for language and the various linguistic systems, e.g., French, Chinese, etc.
3 I do not ignore, obviously, that in most of the cases, the author of the letter paraphrases Chomsky or cites him directly. It therefore happens quite often that in criticizing the letter I am in fact criticizing Chomsky.
4 Very often, the reason for this is simple: these references do not exist. The texts of Chomsky on which the letter recurs generally do not conain anything other than the letter contains itself. Under these conditions, one cannot blame the author for not having given the necessary proofs, given a framework restricted like this. But one can blame him for having believed, and still believing, that they would be superfluous.
5 [The position of this footnote is not marked in the text. The translator suggests that it perhaps goes about here:] This follows from the text of the letter and the justification which is given from the “realistic” point of view (p.3).
6 Here, I am alluding to the use Newton made of gravitation. One will note that there are several ways to reject (Ba): one could claim that a respectable science is not obliged to give a substance to R, even if it could; one could claim that it is always obliged to do that, but is never able to; and one could claim that it neither can nor should do it.
History shows that all these opinions have been held, not without arguments: which confirms again that (Ba) is not evident. I want to add that these opinions could be held by representant of the same science, which should not necessarily hinder them, in that case, to work in it successfully. This confirms that (Ba) and even more so (Bb) does not necessarily have large consequences for the respectable science.
7 My personal position, which does not matter here, which does not interest anybody and which I am only mentioning for honesty, is the following: I accept (A), but neither (Ba) nor (Bb). From this, it follows that the propositions of linguistics are falsifiable, but not on the basis of evidence taken from the languages themselves. No falsification on the basis of psychological evidence (or biological, or which sort of non-linguistic evidence there is) is therefore acceptable to me. What strikes me is that this position is the one of all (o almost all) generative linguists, including those who accept (B) and (C). From this I conclude that (B) and (C) do not play any role in the construction of linguistic theory.
On the other hand, I am tempted to claim that effectively, language [faculty] is of the same substance as the prohibition of incest. But in my eyes, this thesis is no answer to (Ba) at all, it has not the same role as (Bb). It is on the contrary the very nature of linguistic theory to develop totally independently of this thesis: acting as if it did not exist.
8 Perhaps, the following implicit thesis should be added to the GLOW letter: “only the natural sciences are empirical and produce falsifiable propositions”. Obviously, the consequence would be that a science of the economy should be radically impossible, inasmuch as it is an empirical science. I do not have any opinon in this respect, but the consequence would seem daring, and maybe not to the liking of those who accept (C).
9 Apparently this is a note of the Newsletter editors, as the phrase doesn’t appear in the original letter by Koster et al.