Myths, magic, and poorly drawn battle lines: Commentary on Evans 2014

Evans’s The language myth: Why language is not an instinct is a not-so-subtle attempt to counterweigh Steven Pinker’s (1994) The language instinct. For more than two decades now, opponents of Pinker’s view of language have lamented the fact that introductory students (and nonspecialists in general) do not have access to a well-written, accessible, yet opposing, view of how language works. Enter The language myth (TLM), which reads like a manifesto out of the School of Cognitive Linguistics. In that respect, the book is a welcome addition to the fray, since there is, after all, significant room for discussion on just about every major point within the nativist program (dubbed, by E, the ‘language-as-instinct hypothesis’).

The basic premise of the book is that the language-as-instinct hypothesis consists of several related myths, which, when combined, provide the veneer of a believable theory. E deconstructs the overall hypothesis into six myths and attempts to debunk each of them individually. The manner in which this is organized and presented is both superbly readable and, unfortunately, fatally flawed. While it sounds like a good idea to present succinct ‘myths’, followed by a methodical take-down of each myth, this need for succinctness and ease of reading means that E is forced to simplify some very complicated and controversial facts. Debates that span decades and that have consumed thousands of printed journal pages are summarized into mere sentences, often in a way that is most pernicious for the language-as-instinct camp. Ultimately, the descriptions of the ‘myths’ are such gross misrepresentations of the generativist field, such cartoonish caricatures of the mainstream of generativist thinking, that it makes the debunking of said myths rather hollow.

Let us begin with a relatively minor example, followed by more serious objections below. The first myth addressed by E is that human language is fundamentally distinct from other forms of animal communication. E’s position is that animal communication forms a continuum, with human language the bookend on the complex side of the continuum. The way the argument is made is rather simple: Charles Hockett’s design features of language are presented, and then it is shown that various forms of animal communication share these features.

E’s point is that for every feature of language, some form of animal communication exists that shares that particular design feature. For example, for the design feature ‘displacement’, E points to some evidence that chimps and other primates are able to refer to objects that are no longer in view, thus displaying this feature (55). E goes down the list of features in this way (oddly, omitting arbitrariness and discreteness), showing that some form of animal communication shares each individual design feature. But of course, as any instructor of an Introduction to Language course will tell you, Hockett’s point was that human language has all of these design features, and there are no other forms of animal communication that have all of the features. This is a very basic point, but one that underscores the idea that in this book, the portrait of the generativist approach is often not an accurate depiction, but rather a crude sketch, designed to be easily debunked.

Consider next the chapter on language acquisition (‘Is language innate?’). The chapter begins with a summary of the reasoning behind the nativist approach, which, once again, is thoroughly simplified and inaccurate (in fact, of the roughly eighty references in this chapter, there are only three to empirical work from a generativist perspective—Crain & Nakayama 1987 and perhaps Bickerton 1981, 1984—with the remaining seventy-five or so references overwhelmingly to usage-based empirical research). As an example of how damaging this is, consider E’s presentation of the argument from the poverty of stimulus (101): ‘The input children receive is impoverished, in the sense that language exposure is partial, to say the least; and there is no negative evidence to warn them off incorrect grammatical patterns’. This, according to E, predicts that errors should abound, and E suggests word order as an example...