Not compelling: Commentary on Evans 2014

Evans has written a book that contains all of the key ingredients of an introductory textbook on linguistics. It covers most of the subjects one would typically expect, and the reader encounters most of the big names any introductory text in linguistics usually mentions. The difference is that E couches everything in a pungently anti-nativist rhetoric. The empirical evidence presented is systematically taken to imply that tenets formulated by Chomsky and his followers, notably Steven Pinker, are incorrect. According to the author, everything we have learned about language—its structure, evolution, history, variability, acquisition, and representation/processing in the brain—since Chomsky famously proposed that language is an innate faculty of the human mind actually serves to disprove that thesis.

E’s alternative to what he calls the language-as-instinct point of view is that language is an outgrowth of cultural intelligence—a specifically human trait—and that all of its properties are rooted in patterns of communicative use. The tone of the book is thus polemical. It is also occasionally tendentious, and the author does not eschew derogatory qualifications of the school of thought he is attacking or its representatives, suggesting that advocates of linguistic nativism are not objective and not sensible. Consider the following quotations:

The findings I’ve been discussing add up to the body of evidence that any objective person would find compelling.

(57)

… does any sensible person really believe that language could plausibly have evolved for anything other than communication?

(258)

Even the title, which includes the term ‘myth’ as a qualification for a scientific theory, reflects the book’s polemical stance.

Setting aside these caveats, we can ask: does E have a case? Let me try to answer this by considering the points he makes about primary language acquisition. E does two things: he argues that the nativist account of primary language acquisition is not supported by the evidence, and he puts forward an alternative, usage-based account, arguing that this is adequate and complete. A major theme in E’s critique of the universal grammar hypothesis is that the predictions it makes for the process of language acquisition are incorrect. On this count, I can go along with E. The available evidence does not show that acquisition of properties of a native language grammar is instantaneous and uniform across the board, and the trajectories that we see in young children’s natural language development do not establish that a mechanism such as parameter setting is psychologically real. It is important, however, to remember that the original ideas about language acquisition in the Chomskyan framework were motivated by formal considerations; the acquisition of a grammar was approached as a purely computational problem. Psycholinguists subsequently took the computational theory as an idealized version of a theory at the level of implementation and tested its predictions. In fact, much of the evidence E considers to contradict the nativist approach results from empirical work on child language development propelled by the very proposals made by Chomsky and his followers. Science progresses.

E claims that the field of developmental psycholinguistics today has a complete and firmly grounded answer to the question of how children acquire their native language, one that avoids any association with language-specific innate abilities. He argues that two generic skills are sufficient for young children to attain full competence in their native language, without an innate universal grammar: pattern recognition and cultural intelligence.

Pattern recognition refers to children’s ability to extract statistical patterns from sequences of linguistic units. A classic example of this capacity is provided by a study (Saffran, Aslin, & Newport 1996) that shows that eight-month-old infants exploit transitional probabilities to segment words from a continuous stream of syllables. Over the years, extensive research has made clear that this ability is not restricted to strings of linguistic units; children are also sensitive to the statistical structure of musical tone sequences, series of visual patterns, and much more. Moreover, many nonhuman species show similar statistical sensitivities, and so the underlying mechanism would seem to be neither specifically linguistic nor specifically human.

Cultural intelligence ‘predisposes us to being pro...