Freidemann Pulvermueller's "Neuroscience of Language" is an indispensible reference for anyone seeking more than cocktail-party knowledge about psychology, neurology, linguistics, and computational systems. Perhaps because it attempts to unite so many previously-unconnected fields in a concrete neural model of natural language (instead of in yet another abstract, mathematical model), the many technicalities of these fields unfold in a transparent, easy-to-follow order. The subtitle of Pulvermueller's book is not "Neuroscience for Dummies", but it succeeds in providing an integrated,lucid, instructive tour of the past 50 years of research in cognitive science.



Unfortunately, the book's subtitle is "On brain circuits of words and serial order", and when Pulvermueller finally turns to serial order in Chapter 8, I think his brain circuits start to malfunction. I say this because of an influential paper Karl Lashley wrote back in 1951, entitled, "On the problem of serial order in behavior". Briefly, the problem is that serial order isn't always serial (Lashley cited Spooner's toast "To our queer, old dean"). Unlike generations of pop pyschologists, Lashley's student, Noam Chomsky, understood this to be a devastating critique of behaviorism. The point about Spoonerisms is *not* that they are freaks of nature, but that they are ubiquitous in human behavior. Chomsky's prime example was the ubiquitous "transformation" of e.g. "John kissed Mary" into the passive voice "Mary was kissed by John", but almost every serial behavior in your human repertoire, from your route to work in the morning to the arpeggios you play on the piano, exhibits the serial non-seriality of Spoonerisms (or, more formally "metathesis"). Pulvermueller cites Lashley's paper, but he never really addresses metathesis. Instead, he builds his neural model out of "synfire chains", so that in the end the reader finds she has been given a superb 50-year tour of psycholinguistics only to wind up back in 1950, analyzing behavior in terms of neo-Skinnerian stimulus-response chains.



If you are interested in these issues, you should definitely get Pulvermueller's book, if only to understand how cognitive science became locked in this vicious circle of reasoning. To understand Spoonerisms, however, you should look at Loritz' "How the Brain Evolved Language". Unfortunately, Loritz follows the work of Stephen Grossberg, and that requires a really different way of thinking about thought--sort of like the difference between thinking in Newtonian terms and thinking in terms of relativity. I didn't really *get* Loritz (much less Grossberg) until I first read Jeff Hawkins' "On Intelligence". You might want to go this route, too.