Why agree? Why move? Unifying agreement-based and discourse-configurational languages (review)

The main title of the book is brilliant: short, snappy, and certain to arouse curiosity as it promises to answer two of the most fundamental questions in current generative syntax (the subtitle could have been left out). Does the book answer the questions? Yes, it does, although, not surprisingly, there are further questions raised by the answers, which I return to later. In the process of elucidating these two big questions, Shigeru Miyagawa discusses and develops interesting, well-informed, and for the most part convincing analyses of a range of phenomena in a number of languages. Japanese figures prominently in the discussion, but Finnish, Kilega, Kinande, Chinese, and Old Japanese also provide important pieces in the jigsaw puzzle that M assembles.

The book is a ‘Linguistic inquiry monograph’, and as such is not too long, easy to carry with you, well produced (although the Finnish examples could have been better proofread), and affordable. The series still uses endnotes instead of footnotes, which is irritating if you want to read the notes as they come. The trick is to just ignore the notes until the end of the chapter, which worked well in the case of this book. M’s presentation is lucid, never more formal than required to get the point across. Each chapter starts with an extensive synopsis that is then further articulated, making his reasoning always easy to follow.

Ch. 1 is titled ‘Why agree?’ (1–29). Giving away the answer here feels a bit like revealing the murderer in a whodunnit, but I will do it nonetheless (or this review will remain very short): agreement occurs to establish a functional relation. By ‘functional relation’ M means relations such as subject-predicate, theme-rheme, and focus-presupposition. Agreement is a means of forming such a relation at a distance, for instance, between finite T and an argument in VP. Agreement is typically accompanied by movement, placing the constituent agreed with in the specifier position of the agreeing head. Thus, the answer to ‘Why agree?’ is also part of the answer to ‘Why move?’ (the title of Ch. 2). According to M, the reason for the movement is not that agreement presupposes a spec-head relation (as in Chomsky 1993), but instead, the movement is required to ‘keep a record of functional relations for semantic and information-structure interpretation’ (33). Agreement is described formally in terms of Chomsky’s (2001) probe-goal valuation relation: a probe with an uninterpretable feature is valued by a goal with a matching interpretable feature. This is often reflected in phonetic form (PF) by agreement and case morphology. Since uninterpretable features get deleted subsequent to spell-out, however, there is no record of the relation in logical form (LF), unless there is also movement. The principle is called probe-goal union (PGU): a goal moves to unite with its probe.

The deeper reason why we want functional relations, according to M, is that it greatly increases the expressivity of language. He does not actually discuss what human language would look like without these relations, but as far as I understand, we would be restricted to referring/pointing to individuals and events, perhaps only in our immediate surroundings, without being able to say anything about them. There would be only vPs, no TPs or CPs.

One of the leading ideas in this book is that all languages have Agree, accompanied by movement, but Agree is not always a matter of φ-feature valuation, resulting in morphological agreement. Another form of Agree employs topic/focus features. This important idea is introduced in Ch. 1 and elaborated in subsequent chapters. M does not claim that languages do one or the other. In fact, he argues at some length that (Mandarin) Chinese, commonly taken to be a topic-prominent language par excellence, has person features in T involved in agreement, albeit without morphological effects, and that even Japanese has some φ-feature-based agreement...