Talk:Lucien Tesnière

Why the current association with generative grammar is false

User:Tjo3ya has asked me to explain why the following claim is false:

"Tesnière also argued vehemently that syntax is autonomous from morphology and semantics. This stance is similar to that of generative grammar, which takes syntax to be a separate component of the human faculty for language"

The proper question is why it should be true? Tesnière was clearly not anyhow involved in the generative enterprise, so, to begin with, the place of this discussion is not in the introduction. The proper question is how Tesnière's argument was different from Saussure, not Chomsky, and here we would take up the concept of antinomy between structural and linear structure. I will write a passage on this.

Secondly, as is less important from the perspective of Tesnière, I will comment the article "Bare phrase structure, label-less trees, and specifier-less syntax. Is Minimalism becoming a dependency grammar?" by Osborne, Putnam and Gross here.

The discussion cited above is somewhat misguided from the French structuralist as well as the generative perspective. Osborne et al. have compared Chomsky's Minimalist Program to Tesnière's dependency grammar although what they should have done is to open page 36 in Seuren 1998, Western Linguistics. This book as a whole may be taken as the history of the NP VP (NP) structure. You will see that the Speculative grammarian model (ca. 1300–1310 AD) is identical with 'Standard Theory'. The structure is best understood from the Aristotelian-mediaeval perspective, and as you will see, it survived with Wilhelm Wundt and was imported to the US by Leonard Bloomfield. Here, it becomes vital to understand that the Bloomfieldian position contradicts structuralism proper which is the interaction of meaning and form. Since the NP VP (NP) structure is not semantically motivated, American linguists would have to look for a different explanation for it. At this point you will need to click on language faculty, the idea referred to in the introduction of the current version of the Wikipedia article of Lucien Tesnière.

Coming back to Tesnière, you must realise that he was a semantician, and that he studied the relationship of syntax and semantics. This is what dependency grammar is all about. Tesnière's explanation for the "antinomy" between syntax and semantics entails that their organisation is quite different for reasons which are neither "logical nor psychological". Instead, what causes a conflict between syntax and semantics is the necessity to force a nonlinear semantic mapping into linear form.

This whole idea is logically and historically fully unrelated to generative grammar. We must consider the meaning of the word 'theory' from a structuralist proper and from a generative point. In generative grammar, each syntactic model is a "theory" of language because it is implied that the model describes a real cognitive structure, i.e. it is the theory of what the language faculty looks like. The structuralist (proper) opinion, on the other hand, is that such ideas are as relevant to modern understanding of language as the mediaeval speculative grammar is today. From this perspective, grammatical models are not taken to be real at all, but as mere analytical tools for linguistics. Structuralism certainly entails a theory of language, but this is not what Tesnière was working on. As the Sorbonne professor Jean Fourquet explains in his preface to the 1959 edition of Tesnière's Éléments de syntaxe structurale,

"Pendant que Tesnière s'attaquait à une oeuvre de réalisation immédiate, la linguistique structurale se développait autour de lui dans un autre sans, celui des constructions théoriques. Il n'a pas ignoré ce mouvement ... mais il n'a pas pensé, semble-t-il, qu'il fallût attendre la suprême mise au point d'une analyse ..."

Fourquet discusses the centricity of semantics for Tesnière's approach as follows:

"Le centre de la pensée syntaxique de Tesnière est la notion que l'on désigne imparfaitement par le terme de classe de mots. Il ne s'agit pas ici des parties du discours au sens ancient, considérées comme des espèces identifiables par l'observateur de leurs formes, mais par exemple du fait que donner une information sur un évènement à l'aide d'un verbe, de groupes à fonction substantivale (sujet, objet) et adverbiale (temps, lieu) implique une première activité conceptuelle ... Vu sous cet aspect, le structuralisme de Tesnière n'est pas un structuralisme formaliste."

Generative grammarians have developed a concept of linguistics which is only related to itself. To summarise, this concept is not related to Tesnière's idea, and what needs done is to 1) remove the foreign object from the introduction; 2) add an explanation of Tesnière's method as relates to his concept of antinomy between syntax and semantics; 3) eventually leave the bit at the end commenting Tesnière from a generative point. Thanks, Weidorje (talk) 09:28, 28 May 2020 (UTC)

I think the passage under dispute is being over interpreted. That passage does not state that Tesnière motivated Chomsky concerning the autonomy of syntax, nor does it state that Tesnière was somehow motivated by the tradition that motivated Chomsky's first works. It makes no claims at all in this regard. It merely points to the fact that both approaches emphasize the autonomy of syntax. In this regard, I am not opposed to the addition of a qualifier stating that this emphasis on the autonomy of syntax arose independently in both traditions. Such a qualification would then reduce the danger of misinterpretation in this regard. --Tjo3ya (talk) 13:21, 28 May 2020 (UTC)

User talk:Tjo3ya, I'm afraid you're not following at all. I will try to rephrase, translate the French passages, and then re-summarise. What you need to realise is that Chomsky and Tesnière are not talking about the same thing at all. We can use the terms 'autonomy of syntax' (Chomsky) vs. 'independence of syntax' (Tesnière), the latter being a reference to Tesnière's concept of 'antinomy between syntax and semantics'. These two are quite well discussed by the authors with no link inbetween.

Starting with Tesnière's analysis, it is fully motivated by semantics. By this I mean that when you do DG analysis, you figure out semantically which word refers to which, and it's fully an intellectual task. No dependency is taken as autonomous from semantics. This is contrasted with generative grammar. In Syntactic Structures, Chomsky argued that grammars are autonomous and independent of meaning in the sense that their primitives are not defined in semantic terms (p.17). That "should not, however, blind us to the fact that there are striking correspondences between the structures and elements that are discovered in formal, grammatical analysis and specific semantic functions" (p.101). In practice, most of the GG analysis makes sense semantically, but not the NP VP (NP) structure which is autonomous from semantics.

Here are translations of the French passages:

"While Tesnière undertook a task of immediate realisation, structural linguistics was developing around him in a different sense, in that of theoretical constructions. He did not ignore this movement ... but he did not believe in procrastinating uncovering an optimal way of analysis ...

In other words, Tesnière was an analyst, not a theorist.

"The core of Tesnière's syntactic thinkg is in the notion that the term 'parts of speech' had not been properly defined. It is not a question of age-old discussions where these are thought of as types identified by observing their forms, but for example from the fact that, giving information of an event by using a verb, groups of substantive functions (subject, object) and adverbial (time, place) implies primarily a conceptual activity ... From this perspective, Tesnière's structuralism was not formalist."

This means his analysis is fully semantic/conceptual. This is not what GG means with 'autonomy of syntax' or is it? I'm actually spending quite a lot of time explaining this to you, so maybe it's your turn to start arguing for your own case. In my opinion there is no over-interpretation. Tesnière is not related to GG, but also is he in no way related to the Chomskyan concept of "language faculty". This is wrong and needs to be removed, especially from the introduction.

To summarise: dependency grammar is a question of analysis which is not equated with theory. This is contrasted to GG which is a question of "theory" of language, although it is largely equated with analysis. For this reason (as well as any other) DG is not related to the concept of language faculty while GG is.

Also, DG is fully semantically motivated. This is in contrast with GG which is not taken to be semantically motivated (although, in practice, much of it is). These two approaches cannot possibly be equated or even called "similar" in any meaningful way. The intro will have to change. Weidorje (talk) 15:28, 28 May 2020 (UTC)

Thanks for the edit, User talk:Tjo3ya. As a result, the claim in the introduction is now clearer than before, and therefore even more obviously incorrect. — It's not you, it's the training in generative grammar ;) Otherwise you've done a good job with the article, but there's a conflict of interest regarding this point. On the other hand, it's actually interesting for the Tesnièrean view that his work is now getting attention from GG. I suggest we add two sections: one discussing the concept of 'antinomy', and another one for recent discussions in GG. Weidorje (talk) 07:16, 29 May 2020 (UTC)

I have read your comments and I agree that you are expending much effort to establish your point. However, I still think that the matter is being over-interpreted. If we were to ask your average Chomskyan syntactician if Chomsky's syntax views syntax as autonomous from morphology and semantics, the answer would likely be "Yes". That comports with Tesniere's vehement claims that syntax is independent of morphology and semantics.

The reason the statement belongs in the introduction is that it serves as a means of attracting interest. Chomsky's ideas and works are often the reference point for many modern syntacticians and grammarians. If what appears can be related to the dominant approach, then more are likely to continue reading the article.

The same is true of the article by Osborne et al. (2011) you mention. Osborne et al. chose the particular title of that article mainly as a means of capturing interest, not because they believe that dependency grammar and Chomsky's Minimalism have much in common.

All that stated, I suggest that you indeed add some clarifications. I think one good paragraph added to the Section "Autonomous syntax" could be appropriate. You could clarify there the extent to which Tesniere's understanding of autonomous syntax is distinct from the Chomskyan understanding of autonomous syntax. --Tjo3ya (talk) 07:55, 29 May 2020 (UTC)

Okay, I'll be looking at it given the time. The generative point is interesting as well, but the French structuralist view must of course be the primary aspect to do justice to the original which is independent and autonomous of Chomsky. GG is only one perspective among many, and the English Wikipedia has no commitment to it. Weidorje (talk) 10:06, 29 May 2020 (UTC)