Get a site

Empirical Evidence in Support of Chomsky’s Theory of Generative Grammar?

Chomsky argues that there is a meta-grammar, a grammar-of-all-grammars, inherent to human language, which he calls “generative grammar”. His idea of a recursive function which describes all human language is theoretically intriguing, but empirically fairly readily falsified. In earlier posts I pointed to features of gramma unique to some languages, yet not found in others. Examples include, but are not limited to: declension of nouns, conjugation of verbs, presence, absence, and use of articles, prepositions or post-positions, nouns-as-verbs, compound-nouns, tonality, gender or gender neutrality, particles. Some of these features are not part of the English language, which is why you may never have heard of or considered them. Note that Chomsky appears only to speak English, French, and Hebrew, and that English is a bastard language; basically Germanic nouns for common items, a largely simplified Germanic grammar, with Latin terms for any advanced concepts. This is to say he lacks the empirical foundation to carry his argument.

Although I reject his generative grammar hypothesis I wish in this post to describe some language features which appear to me to be common to all languages I have studied, which is very few in comparison to the roughly 35,000 languages on Earth. However, my languages are representative of the three main trunks of Indo-European (Germanic, Romantic, Slavic) as well as two non IE langugages – Estonian and Mandarin.

One structure which appears fairly consistent accross languages is the use of the subject-verb-object SVO format for declarative and imperative sentences. In contrast, question sentences (interrogatives) only sometimes use the format VSO. Chinese for example relies on question particles such as “ma” and “ne” which allows the sentence to retain SVO form. Furthermore even within SVO there are wide variants on the word order: do indirect objects precede or follow direct objects? What about temporal elements and negation? Temporal elements, negation, and direct/indirect object placement vary between languges.

This post however caught my eye because it does confirm something intuitively common to me based on my language knowledge so far: the time-manner-place-object structure. In Mandarin sentences generally can be fit into this form:

This is still very far from Chomsky’s idea that there be a recursive grammar which describes all human grammar and that this generative grammar is inherent to the human, and only the human. However, SVO is consistent with my positions:
1) The human species suffered a near-extinction event some 150,000 years ago.
2) This near-extinction events resulted in a genetic bottleneck.
3) This genetic bottleneck was sufficiently narrow that humans at one point in history had only one language: proto-world.
4) Proto-world in turn evolved into Nostratic, and then Indo-European

These observations are consistent with genetics, mythology (e.g. tower of Babel), and the observed empirical facts of human language. They better explain the few common elements we can occasionally observe among human languages than Chomsky’s unproven hypothesis. When a hypothesis is empirically falsified we reject it, particularly when a more accurate hypothesis better fits the facts.

I point out the SVO structure partly in fairness to Chomsky, though even with SVO we quickly observe variances to contain nuance. Human language tends to use SVO for declarative sentences but readily admits exceptions. E.g.
“The ball I threw is now on the floor” OSVO
“That is the dog I walked”. OVS

In all I regard Chomsky’s generative grammar thesis as interesting more for what it says about scientific production than on its own terms: as a thesis it is untenable due to empirical falsification. As an example of the limits and possibilities of the scientific method it interesting.

Comments are closed.

Powered by WordPress. Designed by WooThemes