Very good intro to a fairly advanced (but exciting) topic in linguistics,
This review is from: The Atoms Of Language: The Mind's Hidden Rules of Grammar (Paperback)
he Atoms of Language by Mark Baker
One way of looking at this book is that it deals with what Mr. Baker calls the "Navajo Code Talker" paradox (that's basically the author's engaging way of introducing his subject): languages on this planet are very much alike (within limits, any human can learn any language) but also very different. The best example of this is the use of the "Navajo Code Talkers" during WWII: those american citizens were enrolled to translate important messages into their own language (Navajo) for communication on the battle front. The expectation was that Navajo being a hard language, Japanese code breakers would find it difficult to "decrypt" the messages. It appears the initiative was highly successful as the Japanese never managed to "break the code", showing that languages can be very different. But the fact that it is possible to translate back and forth between Navajo and English also shows that languages are not completely non-commensurable. So we have a bit of a paradox.
Mr. Baker uses throughout the book a chemical analogy: there are basic ingredients in the human psyche which he calls the atoms of language. What we observe in nature are the much more complicated analogues of molecules, with many atoms put together and interacting in interesting and not always predictable ways. I'm honestly not sure the analogy is particularly compelling, but the author does not take it so far that it becomes annoying.
So what are Mr. Baker's atoms? They're a bit abstract, even as far as atoms go, since they are really parameters for languages' grammars. If you've got a bit of a computer science mind, this might make some intuitive sense: if I want to produce procedurally an "object" that represents a grammar, what parameters do I need to specify to have a complete description of the grammar? If you're not into this kind of thinking the author will do quite a bit of fairly competent handholding to get you to the point where you should understand what he's saying.
Now what's so interesting about all this? First of all, it appears that parameters are not set randomly. There are certain combinations of parameters that are basically non-sequitur. That's not something that would have been obvious in advance, but it's equally well something that's not particularly easy to interpret. The author actually acknowledges that we don't have the final word on this topic.
In summary, this book provides a fairly pedagogical introduction to a rather advanced current research topic. I'm not entirely convinced that the atom and chemistry analogy route chose by the author was the best way to introduce the subject, but at the end of the day I must acknowledge he gets his point across.