If evolutionary contingencies select organisms based on what they can do, then brains and other evolved structures serve behavior. Behavior is an essential starting place for understanding those structures. But Christiansen & Chater (C&C), starting from a different direction, recommend the brain as starting place; they acknowledge an environmental role in the shaping of languages by language users, but offer precious little about details.
Central to C&C's position is Universal Grammar (UG), which resides in the brain. In their case, UG's particulars arise from patterns of language use rather than from innate sources, whether evolved or originating in some other way. We are not told how it got there, although C&C make a case for its emergence through interactions among language learners and users (perhaps a step backwards from Pinker & Bloom's [Reference Pinker and Bloom1990] BBS argument for language as an evolved system, but maybe a step toward acknowledging memetic roots in the evolution of particular languages). Neither are we told much about the properties of UG, which, as C&C acknowledge, have been a moving target since UG was initially proposed (e.g., Colapinto Reference Colapinto2007).
However UG is defined, and whatever its locus, either inside the head or extending into the language communities that according to C&C have engendered it, its import is primarily structural. It bears little on semantics (it doesn't really matter what people say as long as they do so grammatically) or on sentence function (the properties of imperatives don't really depend on what people do upon hearing them). In fact, probably C&C would not regard questions about function as even within the purview of linguistics: If questions are not about grammar, they are not about language (see also Jackendoff's [Reference Jackendoff2003] BBS treatment, which dealt with structural relations among syntactic, semantic, and phonetic systems but attended little to function).
The trouble is not just that such accounts leave out too many phenomena that should be of interest, but also that such limitations on the boundaries of inquiry may blind us to the interdependence of structure and function (A. C. Catania Reference Catania2000). It is as if we were interested in locomotion but looked only at particular interactions among muscles, examining coordinations as some muscles flexed while opposing ones relaxed and studying how movements at one joint depended on those at others. We could generate a grammar of locomotion, say, for the horse; and from it we might derive a systematic account of possible and impossible gaits. We might categorize different gaits (trots, canters, gallops), and perhaps could show that some not seen in nature (the rack) are teachable but others are not. Maybe we could go even further and show that some coordinations entering into our grammar of locomotion are pre-wired in the spinal cord (a Universal Grammar for Horses: UGH). But this would tell us nothing about where or when a horse might run, or with which gait, or whether in running it is following other horses or fleeing predators. The details of its locomotion, and thus the systems supporting it (e.g., muscles, tendons, bones, brain), evolved in the service of successful running.
These limitations apply even if we look only at horse locomotion. Suppose, however, that we sought a more general theory, applicable to all mammalian quadripeds or, to make it really universal, to all organisms. How should we adapt our grammar to the flippers or fins of aquatic animals? How about birds, and not just when they fly, but also when they walk or swim? And how about insects and arthropods, not to mention the vestigial limbs of snakes and our own bipedal walking? We can hardly deal with such differences without considering evolutionary contingencies, and the essence of such contingencies lies in the functions of locomotion. Were UGH adherents unwilling to extend UGH to these marvelous diversities, we might be sorely tempted to accuse them of a lack of imagination.
I argue, then, that C&C's position on language is comparable to the UGH grammarian's on locomotion, with different language communities substituting for our various locomoting species. In both cases, far too much is left out. Why so little about how particular vocabularies are acquired, or how speech interacts with other behavior? How might current environments shape linguistic units within an individual lifetime?
Space precludes an account of what we know about functional properties of language and ways in which contingencies shape the behavior of speakers (see A. C. Catania Reference Catania2003a; Reference Catania2006). But let us at least consider how the periphery might sculpt language units. Developing nervous systems divide into functional regions sometimes called compartments; of particular relevance is how boundaries form between them (e.g., Irvine & Rauskolb Reference Irvine and Rauskolb2001; Kiecker & Lumsden Reference Kiecker and Lumsden2005). For example, cortical tactile receptor areas in primates include boundaries between the projection areas for different fingers that can be visualized with appropriate stains. The boundaries may arise because adjacent cells on one finger are likely to be stimulated at the same time, whereas those on a neighboring finger, though projected close by on the cortex, are likely often to be stimulated separately. The number of compartments is determined during development by peripheral units (e.g., the separate rays in the nose of the star-nosed mole: K. C. Catania & Kaas Reference Catania and Kaas1997).
A long-standing puzzle is how exposure to linguistic environments can lead to discrimination among native phonetic units well before children begin to produce them. If projections from basilar membrane to auditory cortex have spatial properties that might correspond to phonetic boundaries, then a similar system could lead to separate b and p areas for children in an English-speaking environment and not for those in an Arabic-speaking environment, which lacks separate b and p consonant classes. The categorical perception emerging from such a process would certainly depend on the brain but could hardly be understood without a role for acoustic environments. Language units so created could then interact with the environment through other functional contingencies. And if it worked so for phonetic structure, should we not extend the account to syntactic and semantic structures?
If neural organization so depends on peripheral and environmental organization, plausible arguments follow about how environmental consistencies can sculpt boundaries and compartments corresponding to verbal units and to both structural and functional components of language (e.g., phonetic structure within given language communities, particular vocabularies, and perhaps even functional grammatical distinctions such as imperative versus declarative).
I have focused especially on directionality: To what extent do anatomies, environments and behavior drive or become driven by each other, and how might our understanding of language be illuminated by taking their interactions into account? In arguing that C&C have neglected the direction from behavior to brain, I have not maintained that things cannot go both ways.
Really, I do not doubt that the brain is important. But the brain is an adaptable organ, and it variously accommodates itself to the body and the broader environment within which it finds itself. When the brain is damaged in a stroke, one part of rehabilitation these days is to get the affected peripheral parts moving again: If the left arm is partially paralyzed, get the patient to use it by immobilizing the right arm. It is sometimes said that this treatment uses changes in behavior to change the brain, but the target of the therapy is, after all, the recovery of the behavior.
Both directions, from behavior to brain and from brain to behavior, are obviously significant. It would be astonishing to think that the same reciprocity does not exist for language. To the extent that C&C argue otherwise, they fail to make their case. It is premature to give up on evolutionary contingencies (A. C. Catania Reference Catania and Györi2001; Reference Catania, Lattal and Chase2003b); C&C can escape from a creationism in which language emerges from the brain, not a fully grown Venus but ready to be shaped into one, by taking into account how environments shape the brain.
