Date: Thu, 5 Aug 2004 20:55:59 -0400 (EDT) From: Elsi Kaiser Subject: Pronouns: A Cross-Linguistic Study
Author: Darbhe Narayana Shankara Bhat Title: Pronouns: A Cross-Linguistic Study Series: Oxford Studies in Typology and Linguistic Theory Publisher: Oxford University Press Year: 2004
Elsi Kaiser, Center for Language Sciences, University of Rochester
[Another review of this book appears in issue 15.2197 -- Eds.]
INTRODUCTION In this monograph, Bhat provides a detailed typological study of pronominal elements, focusing primarily on their referential and morphological properties. He discusses examples from a wide range of languages and investigates not only personal pronouns but also demonstrative pronouns, indefinite pronouns, interrogative pronouns and relative pronouns. The book is divided into two parts called 'Pronouns' and 'Proforms.' This division, introduced in Chapter 1, reflects the basic claim underlying this monograph, namely that first and second person pronouns (which Bhat calls ''pronouns'') differ crucially from other pronouns (he calls them ''proforms'') and nouns. Bhat argues that the distinguishing characteristics can be derived from the main function of first and second person pronouns, namely denoting the speech roles of speaker and addressee, and not the individuals who carry out these roles. In contrast, proforms are used in a range of different functions for referring to general concepts such as location, time and manner. Since Bhat regards pronouns and proforms as fundamentally different in their functions and characteristics, he tackles them separately in Parts I and II of the monograph.
PART I In Part I, Bhat focuses on first, second and third person pronouns. He argues that first and second person pronouns share the property of being dissociated from their referents, and that this dissociation can be used to explain certain morpho-syntactic and ambiguity patterns. He also discusses the position of third person pronouns in the pronoun/proform dichotomy.
Part I starts with Chapter 2, which describes the dissociation of first and second person pronouns from their referents. Bhat claims that a number of traits that personal pronouns display - such as the inability to occur with modifiers or complements - can be derived from this dissociation. Modifiers and complements make the antecedent of a referring expression easier to find, but since first and second person pronouns denote speech roles (not actual individuals), this would lead to inefficiency. He analyzes apparent counterexamples as involving apposition. Bhat also attributes the lack of determiners with personal pronouns to the same dissociation. However, he points out that the occurrence of numeral modifiers with personal pronouns is not expected under his dissociation account. He uses data from Kannada to argue that apparent cases of numeral modification are in fact appositives. Nevertheless, Bhat concedes that there does not need to be an absolute dissociation between pronoun and referent for personal pronouns to function as speech-role denoting elements.
Chapter 3 focuses mainly on the referential ambiguities that arise with first, second and third person pronouns. Bhat argues that with first and second person pronouns, we are dealing with ambiguities between speech contexts. With the help of crosslinguistic examples, Bhat shows that logophoric pronouns are used to mark non-coreference of speech roles in different speech contexts, i.e. to distinguish between the exophoric speaker (the one who actually utters the entire sentence) and the endophoric speaker (the speaker of the reported speech). In contrast, with third person pronouns, we are dealing with linguistic expressions with endophoric or exophoric referents. Bhat states that ''third person pronouns can have their reference established either by the speech context, called exophoric reference, or by an expression occurring in the same sentence (or in one of the previous sentences), called endophoric reference'' (66). He presents data from Kannada illustrating the use of a special anaphoric pronoun to indicate when the antecedent is endophoric and not exophoric. Thus, according to Bhat, in the domain of first and second person pronouns, logophoric pronouns mark non-coreference, and in the domain of third- person pronouns, anaphoric pronouns mark coreference.
In Chapter 4, Bhat returns to his claim that personal pronouns are dissociated from their referents. He shows that personal pronouns can - seemingly in violation of their dissociated nature - occur with number and gender marking. However, Bhat claims these categories have special functions with personal pronouns, such as gender marking being used to encode social distinctions and number marking functioning as a 'conjoiner' between contrasting entities (e.g. the pronoun 'we' can conjoin/refer to multiple speech roles: speaker, addressee, other). In the last part of Chapter 4, Bhat discusses differences in case marking patterns between first and second person pronouns in comparison to proforms and nouns, and shows how the special case marking patterns for personal pronouns are related to their speech role functions.
Claiming that personal pronouns are distinct from proforms suggests that personal pronouns form a unified group. Chapter 5 addresses two conflicts inside the class of personal pronouns: (a) the speaker- addressee difference which separates first and second person pronouns and is reflected in the structure of dual pronouns, agreement patterns and prefix ordering, and (b) the dual requirement of being both dissociated from and indirectly associated with their referents. In most languages, personal pronouns are efficient 'shifters' between speech roles and do not contain information about specific individuals. However, as Bhat notes, there are some languages where pronouns carry information about their referents (e.g. gender, number) that cannot be related to the speech-role denoting function. This leads him to conclude that there are cases where markers of both dissociation from and association with the referent are tolerated.
Chapter 6 focuses on the position of third person pronouns in the personal pronouns/proforms dichotomy. Bhat suggests that languages can be divided into two typological groups: (a) three-person languages, where third person pronouns belong to the system of personal pronouns and (b) two-person languages, where third person pronouns belong to the system of proforms and are identical or derivationally related to demonstrative pronouns. In his typological study of 225 languages, 126 languages seem to be two-person languages (e.g. Hebrew, Hindi and Malayalam), and 99 can be analyzed as three-person languages (e.g. Indonesian, Finnish and Japanese; a full classification of all 225 languages is given in the book's appendix). This typological distinction, Bhat claims, is correlated with gender distinctions in third person pronouns: in his sample, two-person languages mark gender three times more often than three-person languages (39% vs. 13%). The division also correlates with deictic systems: three-person languages are approximately three times more likely to have person-oriented deictic systems (as opposed to distance-oriented deictic systems) than two- person languages. In the last part of this chapter, Bhat provides a brief summary of how different language families fit into the two- person/three-person typology.
PART II In Part II, Bhat investigates the internal structure and referential properties of proforms. He argues that proforms have a two-part structure consisting of a term that denotes a general concept and a pronominal element that denotes a specific function, and he investigates the nature of each part in detail. He also claims that proforms, in contrast to pronouns and definite nouns, require semantic - as opposed to pragmatic - identification. In the last section of Part II, he investigates the associations between indefinite, interrogative and relative pronouns.
The first chapter of Part II, Chapter 7, presents the claim that proforms have an internal two-part structure, namely (a) a term denoting a general concept (e.g. '-one' in 'someone') and (b) a pronominal element that denotes a specific function (e.g. 'some-' in 'someone', 'somewhere'). The general concepts that languages encode include person, thing, place and time. The functions encoded by the pronominal elements include deictic distinctions, interrogation and indefiniteness. Bhat points out that the existence of complex proforms and proforms lacking internal structure are not counterexamples for his dual-structure claim. According to him, complex proforms are created from two-part structures by the addition of extra morphology needed for disambiguation, and simplex proforms are the result of fusing via grammaticalization.
Chapter 8 looks in more depth at the two parts of a proform; the function-denoting pronominal element and the concept- denoting general term. Bhat shows that the functions of proforms (and the pronominal elements in them) can be divided into three groups: (a) demonstratives, (b) interrogative-indefinites and (c) relative anaphors. The distinctions made inside these main groups vary across languages, and there is also crosslinguistic variation in how these groups relate to each other (e.g. if two are collapsed in the sense of being represented by the same form). The other half of a proform, namely the general term, denotes some kind of a general concept (Chapter 7). Bhat suggests that the terms occurring in proforms can be grouped in four main categories: nominal, adjectival, verbal and adverbial, not all of which exist in all languages.
Chapter 9 focuses on the referential functions of proforms. The identification involved with proforms, Bhat claims, differs from the identification that occurs with definite nouns and third person pronouns. For the former, identification on an extra-linguistic level is required, whereas the latter only need identifiability on a linguistic level: The minimum requirement for a definite NP is coreference with a previously mentioned NP, but demonstratives (when used deictically) require visually uniquely identifiable antecedents. According to Bhat, linguistic-level identifiability is pragmatic and extra- linguistic identifiability semantic. An NP can have a referential or nonreferential interpretation, depending on the intention of the speaker. However, subsequent reference to this NP is with a definite NP or third person pronoun - i.e. linguistic-level identification is sufficient for use of these forms. Semantic identifiability is more 'substantial'. When a referent is introduced by a proform, it does not automatically become identifiable/referential; the addressee needs a sufficient 'basis for identification' from an extra-linguistic level in order for semantic reference to take place.
In Chapter 10, Bhat investigates the association that has been observed to hold between interrogative and indefinite pronouns. In many languages, these two sets of pronouns are identical (e.g. Lakhota) or derivationally related (e.g. Kannada). Bhat argues that the languages that exhibit this affinity do not have any interrogative pronouns, and that what we think of interrogative pronouns in constituent questions are actually just unmarked indefinite pronouns. According to him, these indefinite pronouns indicate lack of knowledge about that particular constituent. He suggests that the necessary 'question meaning' for constituent questions can be derived from other sources such as question particles (e.g. Lakhota), focus particles (e.g. Mangaranayi), focus structures (e.g. Kannada) and question intonation (e.g. Chinese, Vietnamese).
In Chapter 11, Bhat investigates how his analysis from Chapter 10 sheds light on three issues related to interrogatives and indefinites. First, he argues against the common view that indefinite pronouns are derived from interrogative pronouns, and instead develops an idea from Haspelmath (1997), namely that unmarked indefinite pronouns are turned into non-specific indefinites/universal pronouns by the addition of a conjunctive particle, and adding a disjunctive particle instead creates a specific indefinite. Second, Bhat notes that some indirect questions do not seem to involve any real interrogativity, e.g. ''I know who has gone home.'' In his opinion, if the interrogative pronouns in these contexts are unmarked indefinite pronouns, the lack of any interrogativity is not surprising. Thirdly, he addresses the affinity that exists in many languages between interrogative, relative and indefinite pronouns. If interrogatives are unmarked indefinites in these languages, the associations become simpler: we can just ask why there is a link between relative pronouns and indefinite pronouns. Bhat shows that this involves no contradiction, but does not provide a detailed analysis.
The last chapter, Chapter 12, is very brief, and consists of a short summary of Bhat's main claims. It is divided into four main sections, the first of which summarizes the pronoun-proform division that Bhat argues for. The second section concerns the distinction between conjunction and plurality discussed with respect to number marking in chapter 4, and the third section summarizes the differences between pragmatic and semantic identity and reference. The final section is brief overview of Chapters 10 and 11 on interrogation and indefiniteness.
EVALUATION This book is an important contribution to the typological study of pronouns and related elements. By presenting data from over 200 mostly non-Indo-European languages, Bhat provides a fascinating tour through the crosslinguistic complexity of pronominal elements, including personal, demonstrative, interrogative and relative pronouns. The breadth of the language sample allows him to discuss different groupings of languages and to point out a number of interesting typological dependencies and correlations that would not be apparent if one were looking at a more typologically homogeneous group of languages.
Of course, one of the drawbacks of such a broad typological study is the resulting lack of depth for any one particular language; there simply isn't enough space in a single book to discuss all the different systems in detail. However, the extensive bibliography (which is made up mainly of grammars of various languages) is a good resource for those who want to learn more about a particular language. In addition, the detailed appendix of the 225 languages and their positions in the two-person/three-person typology is very useful. In fact, the monograph is probably best suited for those who already have some familiarity with the morphological and referential properties of pronouns and other referring expressions, and who would like to broaden their crosslinguistic and typological knowledge. It makes a unique contribution to the field of pronoun research by bringing together and comparing information from so many different languages.
Since the book performs the dual task of presenting a lot of data in languages mostly unfamiliar to its readers, as well as Bhat's analysis of this data, in places it is rather densely written. Having more some summary charts or diagrams representing the crosslinguistic correlations and associations would have improved the readability of these section, especially the interesting but densely written Chapter 3. On the whole, though, most chapters are well- written, and the structure of the book mirrors the logic of Bhat's argumentation. The main claims that Bhat puts forth in the book, such as the pronoun- proform distinction, the speech-role denoting function of pronouns, and the indefinite-interrogative association, are presented in a very thorough manner, with a lot of supporting examples and discussion. However, in some places, such as the discussion of scope in Chapter 9 and the discussions of specific/nonspecific proforms and the semantics of questions in Chapter 11, more references to current theoretical linguistic literature would have been helpful in clarifying the relation between Bhat's claims and theoretical work on related topics. A few other points would also have benefited from more discussion. For example, it would have been helpful if, in the discussion of appositives in Chapter 2, Bhat had said more about his view of the distinction between appositives and modifiers. In addition, after reading Bhat's discussion of endophoric/exophoric reference with third person pronouns in Chapter 3, it was not clear to me whether he regards deictic gestures as necessary for exophoric reference, and whether he draws a distinction between nonlinguistic (i.e. not previously mentioned) referents and 'discourse' referents mentioned at some point in the preceding discourse.
This monograph raises a number of interesting questions for future research. For example, as the author himself points out, so-called bound pronoun languages (i.e. languages where clitics and verb agreement are used instead of free-standing pronominal forms) are not discussed in this monograph, but for a full understanding of pronouns, such languages also need to be investigated. An important avenue for future work thus consists of testing how bound pronoun languages fit into his approach. The detailed hypotheses and correlations that he outlines provide a good starting point.
Another interesting topic for future work is the relation between Bhat's approach and the existing research on the discourse-referential properties of pronouns, such as the well-known claims (mainly focusing on third person reference) that the most reduced referential forms (e.g. pronouns in English, null pro in languages that permit pro- drop) are used to refer to the most salient/accessible referents, and fuller referential forms (e.g. demonstratives, definite NPs, etc) are used for less accessible referents (Ariel 1990, Givon 1983, Gundel, Hedberg & Zacharski 1993, inter alia). It would be interesting to see how these findings relate to Bhat's hypothesis about the existence of two-person and three- person languages, and the different kinds of identifiability that he discusses. Another question that often crossed my mind while reading this monograph concerns historical change. Bhat discusses diachronic matters briefly in some chapters, but the typological correlations and associations he discusses raise many questions about language change that are unfortunately beyond the scope of this book, but may provide a fruitful direction for further research.
As a whole, Bhat's monograph represents an interesting and informative contribution to the typological study of pronouns.
Givon, T. 1983. Topic continuity in discourse: A quantitative cross- language study. Amsterdam: John Benjamins.
Gundel, J.K., Hedberg, N. & Zacharski, R. 1993. Cognitive status and the form of referring expressions in discourse. Language 69:274-307.
Haspelmath, M. 1997. Indefinite Pronouns. Oxford: Clarendon Press.
ABOUT THE REVIEWER:
ABOUT THE REVIEWER Elsi Kaiser is a post-doctoral fellow at the Center for Language Sciences at the University of Rochester. Her current research interests include anaphor resolution, the syntax/semantics/pragmatics interface, human sentence processing and Finnish linguistics. In her dissertation (2003), she used psycholinguistic experiments and corpus work to investigate reference resolution in Finnish, Dutch and Estonian.