I’m a fellow of wide interests, but on this blog I try to be single minded and ask So what? of every thing I report. So what does this or that bit of news have to tell us about the origins of language? Mainly, C&C’s new book argues in favor of the position I took a few years back in my own book about how language began (Babel’s Dawn). Our ancestors became community minded enough to be willing to share knowledge, and turned to using their babbling sounds in a meaningful way. Others were able to repeat and use the same sounds. Over time the speech became more elaborate, but retained its learnability for the simple reason that any expressions that were too hard to learn disappeared from use. Also the brain developed improved capacity for paying attention, combining perceptions, expanding working memory, etc. I think Christiansen and Chater would agree with that account, although I don’t think they focus so much on the rise of communal dependence for survival and they are not much interested in the biological side of the story.
One important point they contribute to the narrative is the “now or never” matter. Oral languages must be understood as they are spoken. There is no time for the kind of processing required for the hierarchical trees used by generative grammarians. Systems developed for computers also tend to use some form of hold-and-parse method in which the sentence is taken whole and subjected to various analytical procedures. Humans, however, cannot hold sentences of any length and must interpret as the sentence arrives. Presumably, today’s language users can understand more complex utterances than could be handled by the first speakers—after all, there is no reason to assert a priori that the first speakers were more capable than today’s chimpanzees and bonobos. We know from experience that modern apes can use some brief sign language, but not more than that.
One handicap facing the authors is that they offer no way of formally processing language as it comes into one’s ears. Instead, they argue their case against the supposed objections of Chomsky and company. On the one hand, it is nice to see that they do have arguments in defense of the idea that recursion can flow from usage and does not demand some kind of innate brain module, but on the other hand this stuff is a distraction. The strength of Chomsky’s appeal is less in the inevitability of his positions (which tend to be strange and counter-intuitive) and more in the fact that he has a full system. By now the once promising Chomskyan paradigm leads onto ever more bewildering propositions, but it is still a paradigm. C&C can show that the old approach produces evolutionarily impossible statements, but the history of science proves that it is hard to beat a system with mere objections and counter-examples. You need a new system.
The closest C&C come to a formal process is in their statement that we “chunk and pass” what we hear. That is, we bundle strings of words into meaningful wholes and pass these units on to a “higher level of linguistic representation.” [p. 100] So, as I gather, sounds are chunked into phonemes which are chunked into words which are chunked into phrases, etc. The authors provide a nice description of the principles of chunking and passing, but have not formalized the process so that we can see how it works in detail with one or more languages.
This is the point where I jump in and say I know the details of how it works. As a direct result of the work done on this blog I have been posting peer-reviewed papers (here and here) on what I call Attention-Based Syntax, and ABS assumes a chunking process. Now I plan to help myself to the passing portion of C&C’s theory. The authors write about “eager processing.” I call it attentive processing. Take a sentence more or less at random from the literary canon: The whole of Gaul is divided into three parts. Then we can mark its chunks: [The whole of Gaul] [is divided] [into three parts] or is it [The whole of Gaul] [is divided into] [three parts]? It’s nice to have a problem at the beginning. Answering it forces more inquiry. We can also ask how the listener knows where to break the chunks. In short, we have created something that raises and frames questions for further research, something that is not troubled by the impossibilities of the Chomskyan system.
So I am proposing to wed some of C&C’s general ideas with my detailed system. I don’t really expect C&C to join the matrimonial process—if only because they have sworn fidelity to a less specific constructivist approach—but I’m helping myself to the parts of their theory that seem useful and I will see how far I can run with it.