dents even at undergraduate levels. It is characteristic of the book that it is not just concerned with linguistic modeling and theory, but also draws on a wide variety of sources, from genetics, psycholinguistics, neurobiology, anthropology, archeology, art, music, language acquisition, paleoanthropology, comparative biology, research into animal call systems, and experiments of teaching language to apes. One of B's definitive strengths is that he expresses himself so eloquently; another is that he is very knowledgeable about the generative theory and that he can refer to the relevant statements by Chomsky or any other proponent of the generative enterprise at any point to underpin his criticism. His overview of the theoretical developments is impressive, making the criticism so much more effective than it would have been otherwise. It is one of the framework's own experts who here raises his voice. Many readers will benefit from the historical overviews on crucial constructs of the generative theory, whether one agrees with B's approach or not. And the issues he raises go right to the heart of the theory. His own accounts of the chosen phenomena come across as a bit more sketchy, in my opinion. Although the arguments and data concerning long-distance dependencies may be convincing enough, there is clearly much more to be said about things like the left branch condition and the function of tense and deixis in human language. One may also-somewhat rightfully-accuse B of using certain notions that are poorly defined and explained; for example, his portmanteau effect (308) is used to explain certain differences between English and French, but is not in itself sufficiently explained, and it is also not clear that it derives from anything else-except maybe the arbitrariness of signs. But this principle is described in too hasty a manner to have much explanatory value as it stands in this context. Nevertheless, these minor objections do not subtract from the pleasure of reading this thought-provoking book. Although B is critical, he is in no way dismissive about what the generative enterprise has contributed to the field of linguistics, as he clearly acknowledges that the observations he discusses are in great part, 'if not entirely, due to the remarkable work of generativists' (318). However, 'putting the emphasis on the computational system made [the framework] discover important properties, but this overly formal bias is what prevents it from moving beyond repeatedly listing the facts by means of novel forms of diacritics, as we see in its culmination in cartography' (318). B clearly sees his own research program, the STL, as the natural continuance of the generative enterprise: 'The way out of this cycle is to appeal to the properties of the substances that underlie language. The STL program makes essential use of the perceptual and conceptual substances of language to motivate its theoretical primitives. The result is that the observations make sense and follow from undisputable primitives' (318).