Sandra Chung's book, The Design of Agreement: Evidence from Chamorro, is not for the faint-hearted. As a comment on the jacket suggests, there is material for both "specialists" (presumably theoretical syntacticians) and "nonspecialists interested in Chamorro or language in general." Twenty years of fieldwork and even more years of research in theoretical syntax yields a rich mixture. Specialists must be prepared to enter a very detailed world of Chamorro data where slight variations in the morphology point to large underlying differences in constructions (whether or not a wh-phrase moves through or skips over a Spec position), restrictions on pronouns interact with coordination and the licensing of negative concord in such a way that subjects act structurally high and low simultaneously at S-structure. Possibly one can get by just reading the prose, skipping the examples, accepting the generalizations, and simply concentrate on the conclusions and the account, but then much of the brilliance of the fieldwork presented would be missed. Further, without taking the exact form of the data seriously, one loses out on some of the beauty of Chamorro (why is it that the same morphemes are used for [-finite] and for wh-agreement, as pointed out by Haïk [ 1990] and Dukes [ 1992]? Why is it that examples of subjacency violations for topics are agents of relative clauses within existential constructions?)—all elements that lead to further questions, which may lead to further discoveries. [End Page 170]
For the nonspecialist, the task is probably even harder and maybe the waste is greater (though this may be unfair, because I would count myself as belonging to the group above). The book is certainly full of wonderful observations on Chamorro. Anyone working on Chamorro or a related language (or someone interested in languages in general perhaps) will be educated and entertained by the Chamorro facts. By the end of the book, however, the facts are used mainly to support very intricate interrelations of the theory—in particular, the Principles and Parameters theory of Chomsky ( 1981) and subsequent work—(Operator-C agreement, and wh-agreement and what it tells us about adjunct typologies, adjunct movement, and "trace-sprouting" at LF). As they are presented, the data are answers to questions raised by the theory and trying to read for only the data might feel a bit like reading a series of answers with the related questions suppressed.
Doing a review article on such a work is filled with possible pitfalls. Both the data structure and the theory structure are so tightly constructed that playing with any part of it will be sure to have an effect elsewhere and/or rely crucially on a bit of data not presented in the book. Further, both the theory and the data touch on so many modules of the grammar that to pick out any focused topic necessarily leaves other equally important parts unmentioned. In fact, the problem is not finding something from the book that is interesting enough to explore, the problem is picking one of the many possibilities. I've chosen to pass this choice on to the reader. Rather than a more traditional review article, and perhaps as a nod to Chung's own interest in creating problem sets as evidenced by her workbook on GB syntax problems, I present a set of possible paper topics on Chamorro, and some related languages—in particular Malagasy.
The topics represent a range in terms of size, detail, and theoretical and data content. Not being an expert on Chamorro, I have many questions about the Chamorro data, the answers for which might make a paper topic disappear or lead it in a different direction. Being more of a theory specialist, I have many questions about the theoretical side of the accounts—generally tending toward the possibility of alternative accounts. I acknowledge that some of the proposed alternatives are technology heavy, but the purpose is to be forced to think through what sort of data might help choose between...