Why Wittgenstein Needs Chomsky

The limits of my language mean the limits of my world.

Ludwig Wittgenstein
Pieter Bruegel the Elder, "The Tower of Babel" (1563)

Why Wittgenstein Needs Chomsky

Language vs. the Language Faculty

The famous quote above from the Tractatus Logico-Philosophicus just means that the categories we use to make sense of the world are part of us, and not part of the world. The boundaries that we draw around classes of objects – this is a chair, and that is not – are largely pragmatic, and not absolute. Extending this to the construction of new categories, and we get the Ricoeurian philosophy of metaphor which has largely been borne out by the cognitive science: the only way to generate new concepts is to bootstrap them on top of existing concepts – that is, to use a metaphor. Conversely, as Hayek argued in The Sensory Order, “An event of an entirely new kind which has never occurred before” – i.e. for which no metaphor could be constructed – “could not be perceived at all.”

It’s unfortunate that Wittgenstein has been characterized as a philosopher of language and wrote his aphorisms about language, because it suggests the plausibility of something like the Sapir-Whorf hypothesis, that the structure of thought depends on the structure of the language you speak.

Big claims were originally made on behalf of Sapir-Whorf. The Hopi language supposedly did not grammaticalize time (this was later shown to be incorrect), so Hopi speakers find it difficult to think of the future. Or, Chinese speakers are unable to perceive fine differences between blue and green because the language uses the same word for both.

There is a consensus now that such effects do not exist, though much more modest claims have been made more recently on its behalf, such as the fact that a German speaker and a Spanish speaker will tend to assign different genders to a personified fork, as the word is gendered differently in the two languages. In any case, this is a difference of association, not of perception.

So what of Wittgenstein? The strong claims seem plausible on Wittgensteinian grounds. If language determines the categories with which one approaches the world, why shouldn’t differences in grammar or vocabulary shape perception?

The problem is that everything Wittgenstein said about language pertains more properly to the language faculty. This is the idea of Chomskyan universal grammar. The structure of actually spoken languages is not unbounded; it cannot take any form and thus determine perception as an independent force. In fact it could not conceivably be so unbounded. If new concepts are generated by analogy to existing concepts, then at least some ur-concepts, and rules for relating new concepts to them, must be innate. There must be some basic and innate symbol-manipulation capacity that governs both language and perception, independently of the particular language spoken.

Wittgenstein is therefore not wrong, per se, but the practical relevance of his point is rather circumscribed. How did millennia of premodern philosophers get along as if concepts inhered in the things of the external world, anyway? If it were simply because they spoke the same language and had the same concepts, how were Western philosophers able to incorporate elements of Chinese philosophy following the establishment of trade routes between Europe and China? It was not a function of shared language at all, but of a shared perceptual and conceptual apparatus.

Then there are more mundane differences between languages. I once heard a lecture from a former commander in Afghanistan who remarked on the fact that English has a vocabulary an order of magnitude larger than Dari. Because of its heritage from both Saxon and French, it can make fine semantic distinctions between close synonyms, which makes technical writing difficult to translate.

At this point it’s useful to distinguish between two components of language mentioned earlier: grammar and vocabulary. Grammar is more or less fixed for life by age 12, and flows from a more basic perceptual/conceptual apparatus, but is not itself that apparatus. Vocabulary, on the other hand, though constrained in its phonology by grammatical rules, is basically semantically unbounded, so long as there exists a sufficiently similar analog for new concepts to bootstrap onto. The fact that Dari probably has no separate words for ‘conception’ and ‘understanding’ (a technical distinction imported into English from German), or that Polish has no separate words for sex and gender, is not the cause of the different philosophical outlooks of Afghans and Poles, but a result. Similarly, it is not the lack of a word for “annular combustor” that holds back the Afghan aerospace industry. If either distinction or concept suddenly became salient, there would be no difficulty in borrowing or constructing new terms.

The same is true for other grammatical differences with social implications. Japanese’s grammaticalization of honorifics is more likely the result of a hierarchical society than an independent force perpetuating it. The fact that many hunter-gatherer societies have closed numeral systems is not a cause of scientific backwardness, but a result of it. Indeed, the fact that many hunter-gatherer tribes do readily borrow open-ended numerical systems from neighboring societies, even if they have little need for it themselves, shows that the issue is not one of their world being limited by the language they speak.


So What About Pirahã?

Which brings us to Pirahã, the go-to single example for people who want to own Chomsky. In no small part this is because Daniel Everett, the ethnolinguist who popularized it, sold it this way. Here, indeed, is a language that (supposedly) lacks recursion, the ability to take chunks of concepts and create a new concept with them – an ability that Chomskyans have identified as one of the fundamental mental operations underlying language and (therefore) human conceptual sophistication. If a tribe grows up speaking a language lacking recursion, perhaps this really does limit their world in profound ways.

Everett’s interpretation of the data is not uncontroversial. Perhaps, like the grand claims once made for Sapir-Whorf, these too will fall apart under scrutiny. But let us suppose the facts are exactly as Everett has interpreted them. What does it prove?

Well, it shows that Universal Grammar is not necessarily universal. But so what? Despite the name, universality was never the point. The point, rather, was innateness. Is it implausible that the language faculty could decay in a small and isolated tribal population who had no need of its more abstract tricks? It certainly does not prove that the spoken language is exerting an independent retarding force on the conceptual sophistication of the Pirahã.

The fact that the language faculty is a core part of human perception and cognition is uncontroversial, though there is still disagreement over what exactly constitutes that faculty. The fact that this faculty is generative of language, and not generated by language, is suggested by events like the spontaneous development of Nicaraguan sign language or the creolization of pidgin languages. Wittgenstein has much of value to say, and his characteristic “just counterintuitive enough to be plausible” aphorisms stick very effectively to the brain. But he must be interpreted as speaking of the language faculty, and not of particular languages. If he is not to be put in the service of defunct linguistic ideas, Wittgenstein needs Chomsky.


LanguageLinguisticsPhilosophyDaniel EverettF.A. HayekLudwig WittgensteinNoam Chomsky


Facebook Twitter Reddit Threads


  • 1

    Jim C

    Nov 19, 2017 at 20:52 | Reply

    We must be drawing from the ether… Just presented on Wittgenstein and the structure of mind and society at SEA today.

    For Wittgenstein, language functions in coordinating action. The name “language game” seems to imply this.

    • 1.1

      Cameron Harwick

      Nov 19, 2017 at 21:15

      Neat! Yeah, I had been thinking about our conversations about language games after reading another silly Pirahã article recently. Send me what you’ve written too if it’s in a shareable state, I’ll be interested to read it.

Leave a Reply

More Content