Person (review)

Cambridge University Press’s red series (‘Cambridge textbooks in linguistics’) has included a number of titles that provide a crosslinguistically and theoretically informed overview of a grammatical category, for example, Blake 1994 on case, Lyons 1999 on definiteness, and Corbett 2000 on number. Similarly, Anna Siewierska’s Person is not really a textbook, but rather a survey book that covers almost everything there is to say about personal pronouns and other person markers: morphophonological subtypes and syntactic functions of person forms (Ch. 2), the structure of person paradigms (Ch. 3), person agreement (Ch. 4), the conditions for using third-person forms in discourse and in sentences (Ch. 5), honorific distinctions in person forms (Ch. 6), and diachronic change in person forms (Ch. 7). This book is even more strongly crosslinguistically oriented than others in the same series: the index lists over 600 languages, and many of the book’s generalizations are based on a database of 402 languages that the author put together over the years together with Dik Bakker. So in addition to providing an overview of what other linguists have contributed to our knowledge of person forms in the world’s languages, the book constitutes a milestone of typological research in its own right, summarizing and extending the results of earlier publications by the author (e.g. Siewierska 1998, 2003). The book is written very accessibly, not presupposing any familiarity with advanced terminology (or formalisms). This does not mean that it is easy to read: it has examples from unfamiliar languages on every page, and it is densely packed with information that needs to be digested. The summaries and points of discussion in this review can give only a very partial impression of its rich contents.

Medieval Latin school grammar makes a sharp distinction between ‘personal pronouns’ (like ego, tu) on the one hand, and person-number conjugation forms (video ‘I see’, vides ‘you see’, etc.) on the other, and this distinction is still with us. S lumps these together as ‘person forms’ (or, synonymously, ‘person markers’), because person affixes often function much like free personal pronouns (and vice versa), and it seems difficult to draw a principled line between person-denoting pronouns and person-denoting nouns like Thai phǒm ‘I’ (which can be modified by a demonstrative). I think this terminological choice is felicitous, because the term ‘pronoun’ did not come down to us with a precise crosslinguistically applicable definition. However, S could have spent a little more time on defining ‘person’, because Benveniste (1956) was of course right that only first- and second-person forms denote discourse roles. S speaks of ‘the discourse roles of speaker, addressee and third party’ (2), although a page earlier she restricts the notion of discourse role to the first and the second person. The justification for including all three persons in this book is that it yields a much more interesting picture (8), not that ‘person’ is a pragmatically coherent category, and this could have been more frankly admitted. Another question arising from the definition of ‘person’ is whether all languages have person forms. S seems to presuppose that the answer is yes (she mentions Madurese, with two person forms, as a minimal system), but it would be easy to imagine a language where overt referents are always descriptions or names, and some languages in Southeast Asia (where the speaker and the hearer are often referred to by name) may in fact come close to this situation.

Ch. 2 discusses ‘morphophonological form’ and syntactic functions of person forms. By ‘morphophonological form’ S refers to the distinction between independent and dependent person forms, with the latter subdivided into weak, clitic, bound (i.e. expressed morphologically), and zero. She discusses and exemplifies the various criteria that have been employed for distinguishing the subtypes of person forms in different languages, concluding that the universal applicability of these subtypes is ‘by no means unproblematic’ (40). She suggests that the distinction is not discrete but rather gradual, but since the language-particular differences are typically discrete, a...