This post doesn’t really belong on this blog since it has nothing to do with the intersection of science and faith. But recently I’ve been having interesting discussions over at Babel’s Dawn about the nature of language and its evolution in the human species. As in any such discussion, the topic of what aspects of language might be innate (meaning, ‘genetically programmed’) properties of the human species has arisen several times. Now, for those who don’t know, I’m a professional linguist, my subfield being generative syntactic theory. Even more specifically, I’m a theoretician who works within the Minimalist Program, a program for research built on the work of Noam Chomsky as well as others.. In minimalism, it is taken for granted that some parts of linguistic structure are innate. Discovering what those properties are (and what their sources might be) is a central part of minimalist goals.
Despite the consensus in my field, however, I have found that many non-linguists (and even linguists who are outside the subfield of formal grammar) have a strong resistance to the very idea of there being innate properties of language. Very often I think this resistance has its roots in an acute appreciation of the richness and cultural complexity of language. Anthropologists, literature experts, ethnologists, and linguists largely concerned with descriptivist goals (grammars, lexicographers, etc.) sometimes express the sentiment that formal linguistics (and syntactic theory in particular) is an effort to reduce the multi-faceted kaleidoscope of language to a simple system of computational rules. Anyone who knows the field, however, knows that this simply isn’t the case. Rather, syntactic theory focuses on a very specific set of principles – those that are responsible for the difference between well-formed and ill-formed sentences both in particular languages as well as in language universally. It is easy to see that these principles are largely isolable from cultural influence since grammaticality judgments often do not depend upon social or even discourse context.
Some of the principles responsible for sentence formation seem to have a universal character, being relevant for all human languages. It is these principles that Chomsky and those in the field he spawned have supposed to have an innate basis in the human genome. Most outside the field, however, have no idea what these principles are. Often those who argue for the implausibility of an innate language component are unaware of what linguists in fact claim is innate, making debate rather pointless.
Finally, to the point of the post: below I’m going to enumerate four principles for which there is a strong consensus as to their universality. That is, these are principles that almost no syntactician would deny are universal properties of human language. They are also four principles central to syntactic investigation and theorizing current in the field. After I enumerate them, I’ll say something more about the plausibility of their innate nature.
I. Hierarchical Structure. The relationship between words in a sentence in any language is not symmetric. Consider the sentences in (1) and (2). Note that John may precede himself in (1) and the sentence is grammatical. Reverse the order and the sentence is ungrammatical:
(1) John likes himself.
(2) *Himself likes John.
One might think this restriction is just about word order, but its easy to show this isn’t the case. Consider (3):
(3) I saw the picture of himself that John likes.
Here in (3) himself precedes John and the sentence is fine. What’s the difference? In (3) the phrase picture of himself is not in its original position which is as the grammatical object of the verb likes (in (3) the thing that John likes is the picture of himself). (3) is grammatical because John hierarchically dominates the original position of the phrase picture of himself before it is moved to its overt phonological position. All languages, even so-called ‘non-configurational’ languages, exhibit these hierarchical asymmetries.
II. Displacement/Movement. The data above also illustrate this principle. As I pointed out, the phrase picture of himself in (3) is interpreted in a position different from the one it is pronounced in. Displacement is also common in questions in many languages. In (4), the thing we are asking about, indicated by the word what, is actually the object of the verb read. What has ‘moved’ from its original position to the front of the sentence.
(4) What did John read?
All languages exhibit displacement in one form or another to varying degrees.
III. Recursion. Human language is a recursive system. This just means that we can combine phrases grammatically and iteratively to an infinite degree. Consider (5)
(5) John knows a man who lives in a house that has a new roof that was installed by my friend that used to date a woman that taught at the university that was attended by my brother that…
(5) is a simple sentence with an extremely long object that consists of an iterative sequence of subject relative clauses. One can easily see that this process could, in principle, continue forever without ever violating the principles of English grammar (though of course the capacities of individual memories come into play at some point, making such sentences unnatural and hard to follow). This is recursivity. Not only is it universal to human language, but it also seems to be unique to human language. No other species has been convincingly demonstrated able to detect or produce recursive patterns.
IV. Minimality. It is a surprising and universal fact about language that dependencies between two elements in a sentence cannot be interrupted by a third element of the same type. Consider the movement of question words to the front of the sentence like we saw back in (4). When more than one question word is present in the sentence, it is often the case that only one of them will move to the front. However, there are restrictions on which word it can be:
(6) Who do you think __ bought what?
(7) *What do you think who bought __?
The difference between (6) and (7) is that in (7) the question word who intervenes between the fronted what and its original position (indicated by the blank). In (6), however, no question word intervenes between who and its original position. (7) is said to be an instance of an intervention effect or minimality violation. The constraint doesn’t just apply to movement. Consider a language with object agreement such as Swahili. In this language, the verb shows agreement with objects that refer to humans. In (8), -mw- indicates agreement with the object mtoto. The sentence is ungrammatical without it.
(8) Juma a-li-mw-ona mtoto. ‘Juma saw the child.’
Now, consider a verb that takes two objects, the higher of which is non-human. Turns out it is impossible for the verb to agree with a human direct object in this configuration.
(9) *Juma a-li-m-pa dunia mtoto. ‘Juma gave the world a child.’
(10) Juma a-li-pa dunia mtoto. ‘Juma gave the world a child.’
The reason for (9)’s ungrammaticality is the same as the reason for (7)’s. In (9) dunia intervenes between the verb and the human direct object unlike in (8) where nothing intervenes between these two elements. Since its discovery, minimality has been found to play similar roles in a variety of constructions in dozens of languages. In fact, much of mimimalism is concerned with finding out just how much of linguistic structure minimality can explain.
All four of the above principles are central to current syntactic investigations and there is widespread consensus as to their universality. The question then becomes, what is the source of this universality? Does it come from the fact that all languages have a common ancestor? Unlikely, but even if that were the case, it would merely change the form of the question to why these particular properties have remained constant throughout language while many other properties of the original language have been lost. What about a shared construct of reality? Again, it is a possibility that must be considered, but if it is true, this certainly isn’t obvious. That is, there doesn’t seem to be an obvious reason why the necessities of interpreting reality or expressing internal thoughts would require properties like displacement or minimality or even hierarchical structure to be a part of language. We are left with the idea that the universality of these four properties (and possibly others) stems from genetic information that all humans share. This conclusion, however, isn’t the final one for linguists. Once we conclude some property is innate, the next question is whether that property is a principle specific to language only or whether it is a principle exapted from more general components of cognition. Recursion, for instance, seems to be not only a central part of language, but also counting, dancing, musical composition, and other cognitive activities. It is therefore unlikely that recursion is particular to language. Indeed, given the propensity for exaption that evolution generally exhibits, it is unlikely that any of the four principles above are wholly specific to language.
I know this has been long, but to those few who made it through, I hope it has served as an information source for a few of the basic claims modern syntacticians make for principles of language that are universal and plausibly innate. Discussion about the innateness of human language must deal with the specific arguments that center around specific claims about specific principles. It is useless to discuss the plausibility of innate linguistic principles philosophically since the questions about innateness are empirical question that can be (and have been) given plausible answers based on scientific inquiry. Any argument against innateness have to be similarly based on empirical investigations and claims.