Noam Chomsky held that the biology of the brain set limits of possibility on the structure of language. These were manifested in universal elements common yo all languages. Much of Chomsky’s work concerned the nature of these univerals.
Has this held up? Have such universal structural elements been found?
In the strong form of the principles and parameters theory— that humans come preinstalled with a the list of possible options for language (left versus right branching, prepositions versus postpositions, etc.), and language acquistion is just setting those options and vocabulary— I don’t think there’s any real evidence for it. Besides, it’s hard to come up with a genetic basis for that sort of setup, and the theory is hard to falsify anyway. In the weak form that there’s some genetic basis to language (versus general communication) acquisition that’s unique to humans and there are linguistic universal or near-universals (e.g., dual form implies a plural form; if singular and plural forms exist, then at least the plural is marked; etc.), it’s pretty hard to argue against. There are certainly observable neurological influences on language: the P400 and N600 spikes in response to semantics and syntax; damage to Wernicke’s and Broca’s areas and the corresponding different effects on fluency; the connection of FOXP2 to specific language impairments; and so on.
Beyond that, making the claim that the structure of the human brain limits language is like the Sapir-Whorf hypothesis: true only so far as it’s trivial, but unfalsifiable or simply ridiculous beyond that. There’s clear evidence for a neurological basis of language and for linguistic universals, but it’s hard to tie to the two together. If I were trying to look for one, I’d take a look at the other physical bounds on information content in speech. Humans have to articulate and disambiguate phonemes in speech, and that puts a limit on the number of phonemes, rate of speech, etc. Even without getting into the processing power of the brain, you have to be able to physically make and recognize those sounds despite background noise, poor dentition, etc. I don’t know what exactly it is, but there’s presumably some bound on #(distinct phonemes) #(phonemes per second), as well as each term separately. But that’s getting into Great Debates territory.
There was a story that ran on NPR today that made reference to one of Chomsky’s theories. I don’t remember which program, it might have been Here and Now or The World or The Takeaway. The story concerned an existing primitive tribe that speaks an uninflected language (I think that’s the term), something that (he said) Chomsky claimed was not possible. Then a scientist posited that even as far back as Homo Erectus, human throats evolved to make language sounds possible (probably phrased more scientifically correctly than that).
I was only half listening so I don’t remember a lot of details. If you’re interested I’ll try to find the program and it’s probably available on podcast or something.
eta: I found it, it was on The World, and you can listen to it here. The World : NPR
The language part came at the end.
I’m having trouble playing the podcast, but ‘uninflected’ in linguistics just means that words in the language undergo morphological changes depending on their grammatical function in the senstence. English is largely uninflected: verbs are marked for tense and number, nouns are marked for number, pronouns are marked for case, and that’s about it. Mandarin’s the canonical example of an isolating language: subjects don’t change when they become objects; nouns and verbs aren’t marked for number; adjectives don’t change according to the gender of the nouns they modify (because Mandarin doesn’t have gender); and so on. Japanese has some agglutinative features: atsu.i ‘hot’ becomes atsu.katta ‘was hot’, atsu.kuna.i ‘not hot’, and atsu.kuna.katta ‘wasn’t hot’, applying the regular past tense morphological change -i -> -katta to the negative form ‘atsukunai’. At the other extreme, you have languages like Latin, where verbs are marked for tense and number and various other things, nouns are marked for number and case (and gender), adjectives are marked for number and gender, and so on. It’s hard as a native English speaker for me to work up much enthusiasm for bothering with a complicated case system, but it’s not like learning Japanese is easier than Latin.
Er, don’t undergo etc. ‘Inflected’ means that they do change.
That podcast was presumably about Daniel Everett’s work on Pirahã:
Many linguists, including Everett, don’t accept Chomsky’s theories about universal grammar:
But these phonemes have evolved in most languages into a great deal more complexity than is necessary. I can understand a person whispering, almost perfectly, despite the abolition of quite a few voiced phonemes – the effect is simply a significantly larger number of homonyms than the many we already have. Hawaiians understand each other with only half the phonemes of indo-European languages.
So whatever the bound of distinct phonemes is, it is certainly far beyond what is necessary, maybe by an order of magnitude. I don’t know of there is a correlation between a language’s richness of phonemes and its utility for conveying information in a span of speech time. It would be interesting to find out.
Incidentally, Roderick Femm, what you probably mean is not an “uninflected language” but a language without recursion, since that (according to Everett) describes Pirahã.
A good book showing that the Sapir-Whorf hypothesis is wrong if asserted in any strong form is The Language Hoax: Why the World Looks the Same in Any Language by John H. McWhorter.
Coincidentally, I watched the movie “Arrival” last night, and that premise was basically built into the plot. Much of the difficulty in establishing communication with them stemmed from the fact that their alien minds worked very differently from ours, and that difference was reflected in their language.
That there’s some genetic basis to human language facility is trivially proven by the fact that we’re the only animal with anywhere near so rich a language. But there’s still a lot of variation in how that language facility manifests.
The closest thing I know of to a universal of human language is that “mama”, or something close to it, means “mother” in (nearly) all languages (and now for some of our resident linguists to provide exceptions).
Chomsky’s ideas are supported by two (related in fact) kinds of evidence. First is that children learn their native languages fairly quickly despite the fact that the evidence presented to them is woefully lacking and both ambiguous and contradictory (people make mistakes and talk to kids in a reduced version of the language). The second is more interesting in some ways. It is the transition from pidgin to creole. When people without a common language come together, they often create a pidgin. This is not a language since it largely lacks grammar; words borrowed from the speakers’ native tongues are just thrown together and somehow people communicate. But the kids who grow up in this environment seem to spontaneously create a real language out of this melange. It has a full grammar and, from what I have read, the created grammars seem to have a number of features in common, but I have forgotten what. Incidentally the language called Pidgin in New Guinea is actually a creole, whatever they call it.
As for the OP, I would say that it is difficult to say but I do accept at least some form of Chomsky’s hypothesis. Incidentally, AFAIK, the Chinese languages are uninflected. I certainly never read anything by Chomsky in which he suggested that every language had inflections. Certainly, every language has a grammar, but that is a different claim, one that could not be denied by anybody. Let me just add that my father always claimed that his native Yiddish (he learned English only when he went to school) had no grammar; they just threw words together. He was wrong, very wrong.
Can you point to a good resource that lays out some of these universals (and the structures/forms that are almost never found)?
[Well-put post, by the way, IMHO]
You can browse more than 2,000 of them at this page but they’re not all universals as such; for example, the first rule:
… has these counterexamples:
The first absolute seems to be rule 4:
Uh, needless to say, you need some knowledge of linguistics to understand these.
Try Comrie’s “Language Universals and Linguistic Typology.”
Thank you very much!