Person Reference in Interaction: Linguistic, Cultural and Social Perspectives

by N. J. Enfield, Tanya Stivers

Overview

How do we refer to people in everyday conversation? No matter the language or culture, we must choose from a range of options: full name ('Robert Smith'), reduced name ('Bob'), description ('tall guy'), kin term ('my son') etc. Our choices reflect how we know that person in context, and allow us to take a particular perspective on them. This book brings together a team of leading linguists, sociologists and anthropologists to show that there is more to person reference than meets the eye. Drawing on video-recorded, everyday interactions in nine languages, it examines the fascinating ways in which we exploit person reference for social and cultural purposes, and reveals the underlying principles of person reference across cultures from the Americas to Asia to the South Pacific. Combining rich ethnographic detail with cross-linguistic generalizations, it will be welcomed by researchers and graduate students interested in the relationship between language and culture.

Product Details

ISBN-13:	9780521872454
Publisher:	Cambridge University Press
Publication date:	04/26/2007
Series:	Language Culture and Cognition , #7
Pages:	370
Product dimensions:	5.98(w) x 9.02(h) x 0.94(d)

About the Author

Nick Enfield is a scientific staff member in the Language and Cognition Group of the Max Planck Institute for Psycholinguistics, Nijmegen.

Tanya Stivers is a scientific staff member in the Language and Cognition Group of the Max Planck Institute for Psycholinguistics, Nijmegen.

Read an Excerpt

Cambridge University Press
978-0-521-87245-4 - PERSON REFERENCE IN INTERACTION - Edited by N. J. Enfield and Tanya Stivers
Excerpt

1     Person reference in interaction

---

       Tanya Strivers, N. J. Enfield and Stephen C. Levinson

1.1    Introduction

Person reference is a subject that stands at a central intersection between the various behavioural sciences. How persons are classified and individuated lies at the heart of social theory; how different cultures do so has preoccupied anthropology; how we recognize them from face and voice is much investigated in psychology and the cognitive neurosciences; how we refer to persons has been a central topic in philosophy; and the grammatical machinery involved in tracking protagonists in discourse is an important topic in linguistics. Yet, despite the fact that person reference has this centrality, the empirical study of person reference in natural conversation – the central genre of language use – has been curiously neglected, particularly from a cross-cultural perspective that might throw much light on the relation between culture, social structure and language use.

This volume attempts to fill this gap. Each chapter looks at person referencein a specific language and culture, as reflected in everyday language use attempting to understand unmarked versus marked usage primarily with respect to initial third-person references but also in subsequent and in first-person references. We see quite quickly that how people refer to individuals in interaction is amazingly varied. There are different name formats: simple first name Laurie, first-name-plus-surname Serena Edwards, title-plus-surname Missus Hallman. There are kin titles like Mommy or Granny. There are more complex possessed kin terms in which kinship references triangulate through someone else: Suzanne’s husband, Laurie’s dad. There are descriptions like that guy who does those c’mmercials. There are names and descriptions combined into units – Laurie from our class, Silly Dan – displaying both expansion (e.g., adding a descriptor from our class or silly) and contraction (e.g., reduction of the baptismal name Daniel to Dan). Despite this range of expressions, the chapters collected here show that the domain is still highly rule-governed and orderly. By taking a cross-linguistic perspective, we are rapidly led into the specifics of cultural principles for categorizing and naming persons, and the cultural preoccupations that may highlight one or other of these principles, and bias actual use. At the same time though these studies lend strong support to universal principles that govern this domain, which thus suggest some fundamental shared features of human social organization and principles governing social interaction in general.

In this introduction, we first sketch some of the background that makes this subject so central to philosophy, cognitive science, sociology, anthropology, linguistics and the study of social interaction. We also review some of the specific background in conversation analysis, which has informed and lies close to each of the studies reported in the chapters. Finally, we review some of the general findings that emerge, concentrating on the universal tendencies that are clearly discernable.

1.2    The background: person-identification and reference in cognitive science, philosophy, anthropology and linguistics

1.2.1    Identifying and categorizing individuals

There are many reasons for thinking that reference to persons is a fundamental phenomenon at the intersection between language and social structure. In the sociological dimension, all higher forms of sociality rely on distinguishing individuals so that they can be assigned distinctive social roles. In addition to distinguishing individuals (and not all social systems do – for example, sheep do, but ants do not),¹ social systems can work both with the assignment of individuals to absolute categories (worker bees vs. queens, for example) and relational ones (mother vs. offspring, senior vs. junior). These different principles, which long antedate the arrival of humans on the planet, are reflected in linguistic practices with names (Mary, Ramu), roles (child, postman) and relational terms (uncle, daughter, leader) cross-culturally reflected in the languages described in this book.

Given the deep phylogenetic basis for being able to correctly identify individual people, it is not surprising that cognitive neuroscience research has revealed two discrete brain mechanisms for face versus voice recognition (Belin, Zatorre and Ahad 2002; Sergent, Ohta and MacDonald 1992; von Kriegstein, Eger, Kleinschmidt and Giraud 2003). Moreover, these two neurological areas, while specialized, are coupled so that when someone hears a familiar voice, they readily access the person’s face (von Kriegstein, Kleinschmidt, Sterzer and Giraud 2005).

1.2.2    From individuation to reference: names and descriptions

The human innovation, of course, is language, which introduces what Hockett (1960) called the design feature of displacement – talking about individuals who are not here now. Communication also presupposes speakers and addressees in potentially different knowledge states (otherwise, why communicate?), and with different relations to the referent, and thus introduces triangulation between speaker, addressee and referent. This will play a large role in this book (see especially Haviland, this volume, for discussion of this triangle).

The speaker’s problem is to find a referring expression that will identify, for the addressee, the very individual in mind. Languages offer essentially two fundamental ways to do this, through names and through descriptions. As a first approximation, names (like George Washington) are typically non-compositional (or at least, successful reference has little to do with any such compositionality), and reference is achieved by a direct conventional link between the individual and the name, while descriptions are compositional and the whole has a descriptive content that picks out the individual in mind (cf. The first president of the USA).

The dichotomy between names and descriptions seems to show up firmly in the psychology of naming. Just like a glimpse of a face may give us instant person recognition, so names may tap directly into a specialized person register.² However, such advantages are countered by signal disadvantages. Names are difficult to remember and vulnerable to loss during brain injury or aging (for a review, see Valentine, Brennan and Brédart 1996). People routinely have more difficulty retrieving proper names than they do retrieving semantic information (e.g., a person’s occupation) or naming objects (Brennan, Baguley, Bright and Bruce 1990; Burke, MacKay, Worthley and Wade 1991; Hanley and Cowell 1988; Hay, Young and Ellis 1991; McWeeny, Young, Hay and Ellis 1987). People typically take longer to retrieve familiar names than related semantic information (Johnston and Bruce 1990; Young, McWeeny, Ellis and Hay 1986).³ And, people’s abilities to remember other people’s proper names are more vulnerable to damage (e.g., by attrition in old age) than related semantic information about those people (Brédart, Brennen and Valentine 1997; Milders, Deelman and Berg 1999). Proper names also take longer to retrieve than other semantic information among older adults (Maylor and Valentine 1992).

The most obvious explanation may be the correct one: names, by virtue of their special, direct link with their referents, bypass the web of semantic notions and all the connections they have with one another. Retrieving a semantic notion is like pulling any one of the threads in the web, which eventually will lead, via other concepts, to the specific one sought after, while in contrast retrieving a name offers no such redundancy or multiplicity of routes – there’s just one thread linking the name with the referent. Of course we also associate other properties with the referent, and so psychologists have debated whether these two kinds of knowledge run in serial (Bruce and Young 1986) or in parallel during retrieval (Abdel Rahman and Sommer 2004; Burton and Bruce 1992; and see also Schweinberger, Burton and Kelly 2001). They have also wondered whether it is the uniqueness of the referents or the lack of semantic content in names that is responsible for the retrieval difficulties (Brédart, Valentine, Calder and Gassi 1995; Burton and Bruce 1992).

The dichotomy between names and descriptions, however, can be questioned, at least in part. In philosophy, the dominant view, influentially argued by Kripke (1972), is that indeed names have a special status: Essentially a name is hooked to a referent not by a meaning that picks out the referent, but by a historical – causal chain of events – there was a ‘baptism’ as it were, and then an historical sequence of referring actions that traded on that original act (see also the historical range of views assembled in Ludlow 1997).

Searle (1997[1958]) makes the following point: ‘Suppose we ask, “Why do we have proper names at all?” Obviously, to refer to individuals. “Yes, but descriptions could do that for us.” But only at the cost of specifying identity conditions every time reference is made’ (p. 591). Searle strikes to the core of a theoretical argument for why names work differently to descriptions in the conversationally grounded theory of person reference that motivates this volume’s comparative work. When we describe a person, we commit to selecting some features and not others as constituting ‘the description’. Names give us a way to refer by specifically AVOIDING committing to one or another description of the referent:

(T)he uniqueness and immense pragmatic convenience of proper names in our language lie precisely in the fact that they enable us to refer publicly to objects without being forced to raise issues and come to agreement on what descriptive characteristics exactly constitute the identity of the object. They function not as descriptions, but as pegs on which to hang descriptions. Thus the looseness of the criteria for proper names is a necessary condition for isolating the referring function from the describing function of language. (Searle 1997[1958]: 591)

The Searlian view perhaps goes some way to explaining why, despite the cognitive problems associated with proper names, we use them so extensively.

Another way to partially erode the distinction between proper names and descriptions is to note that cross-culturally the picture may be more clouded. The giving of proper names, in a liberal sense, seems to be universal (e.g., see Bodenhorn and vom Bruch 2006; Mithun 1984; Tooker and Conklin 1984). However, in many cultures, personal names do not have the properties we normally associate with them – for example, they may not be nouns, but verbs or even whole sentences that thus carry plenty of descriptive content; they may not be freely chosen but strictly inherited (in which case they might be more akin to names for natural kinds than to names with a Kripkean baptism); they may be considered private and never used; and, most out of kilter with the Anglo notion of a proper name, they may not be fixed but endlessly changing. Even when a name looks like the same kind of thing cross-culturally, it is possible that it is understood strictly descriptively in one culture and strictly causally – historically in the other (Machery, Mallon, Nichols and Stich 2004). Personal names will be universal only under a wide Wittgensteinian ‘family resemblance’ notion.

Since the naming practices of other cultures inevitably play a role in the chapters of this book, it is worthwhile saying a little more about the observable diversity. On the question of descriptive content, a Mohawk name (an inflected verb) like Aronhianónhnha ‘He watches the sky’ is clearly replete with compositional semantic content (Mithun 1984). Lévi-Strauss (1966) discusses a wide range of practices showing how names can convey substantial information much like descriptions do, for example, names may convey the state of mind of the mother at the time of birth (the Lugbara of Uganda) or the totem of the individual (Aranda of Australia), or even something about his or her place of residence (e.g., Yurok of California). From a semantic point of view, though, this information may play little role in reference – it is arguably connotation not denotation (i.e., it is made available rather than explicitly offered). On the other hand, when the baptismal rights are so restricted that from the name we know the social category (e.g., the clan of the father), these restrictions can play a role in circumscribing possible reference. In many of the chapters that follow, naming systems thus serve to designate the category membership of the bearers. (As this chapter was being written, a news story on the war in Iraq reported thousands of people across the country having their names changed by deed poll to avoid becoming targets of attacks and reprisals because of the religious transparency of their names. The example demonstrates how consequential the information given off by a name can be.)

In addition to differences in practice for bestowing names, some societies make relatively little use of personal names. Bird-David (1995: 73–4) describes the Nayaka as using kin terms or just two sex-linked names in childhood, followed by the use of frequently changing nicknames in adolescence (cf. Sidnell, this volume), and then kin terms almost exclusively in adulthood. Others make use of names that change through the life cycle, and which within each stage may proliferate through wordplay (see Rosaldo 1982 on the Ilongot). In many societies, the use of names is hedged in with restrictions. For example, Mayali true (‘bush’) names are hidden private property (Evans, nd). While some societies (e.g., the Yurok) avoid or forbid the use of the same name for two individuals (Lévi-Strauss 1966: 189), others set up quasi-magical relations between namesakes. Australian Northern Territory namesakes are in a taboo relationship (cf. Levinson 2005 and this volume for a discussion of Yélî Dnye namesakes). Throughout the tribal world, sharing a name may be taken to indicate a sharing of essence. And using a name may be circumscribed with social constraints, like using another’s personal belongings. Such restrictions can lead to the use of alternate referring expressions (Levinson 2005). They may also lead to culture-specific differences in preferred practices of person reference.

Names, we have suggested, get part of their utility from the Searlian avoidance of descriptive content. They may also offer a hot line to the person-identification system so rapidly accessed by the visual face-recognition system. But when names cannot be used, or other factors intervene (discussed shortly), either relational terms or non-relational descriptors come into play. Prime among the relational terms are kin terms. A huge amount of anthropological investigation has gone into understanding the range of kinship systems, their relationship to inheritance, marriage and demeanour, and to the kin term systems that express them (cf., Fox 1967; Keesing 1975; Lévi-Strauss 1969; Parkin and Stone 2004). In contrast, relatively little work (but see Bloch 1971; Luong 1984; Zeitlyn 1993) has gone into understanding the actual use of kin terms in interaction, and this work has emphasized how, especially in small communities, there are usually multiple competing kinship relations between the propositus and the referent.⁴ The choice therefore becomes strategic not only between say his daughter’s husband and his son-in-law but also between Ben’s son-in-law and my cousin. The strategic perspective is very much in line with the chapters in this book, where a central issue is why some particular mode of reference rather than another has been chosen.

Non-relational descriptions of course enjoin this strategic point of view: There is always an indefinite number of ways by which a thing or person can be referred to. How children learn that the same thing is at stake from different points of view is a puzzle (Brown 1958). The choice between the neighbour opposite, the man who drives the Porsche, the bastard who yells at my kids is clearly going to be occasioned by what we are talking about, to whom we are talking and what we are trying to accomplish with the utterance. Sacks (1992) pointed out what an extraordinary resource for sociological analysis is provided by the category terms that surface in these descriptions. He pointed out too that they tend to come in contrastive sets (mother, father, child; doctor, patient; teacher, student) that are implicitly articulated in discourse. Curiously, neither sociologists nor anthropologists have capitalized on this implicit ethnosociology.

One might expect that linguistics had a great deal to say about such descriptions. But linguistics has been preoccupied with other aspects of person reference. Firstly, it has focused on the grammatical category of person (Siewierska 2004), reflected in pronoun systems, verbal inflections and more obliquely in constraints on many linguistic operations. Here, of course, the crucial parameters of speaker, addressee and other are focal. Secondly, linguistic work has concentrated not on first mention, as when a new referent is introduced, but on the extensive grammatical machinery for handling subsequent reference or anaphora. There is a huge literature on pronouns, zeros, reflexives and reciprocals, and also on tracking protagonists in narrative, where many languages have such exotic specialisms as fourth persons, switch-reference or logophoric pronouns (see Huang 2000). Even studies devoted to natural usage (e.g., Chafe 1980; Fox 1987) mostly pay scant attention to the form of initial mention of referents including even names. Downing (1996: 95) remarks ‘proper names have gone largely unremarked in the literature on referential choice’. She argues that information structure, what is given, what is new, what is presupposed, plays a crucial role in such choice.

In this book, the prime focus is on initial reference to third persons, and thus on the choice between name, kin term or other relator, and description. If the focus had been on second-person rather than third-person reference, there would be a vast sociolinguistic and anthropological literature to draw on (see e.g., Brown and Gilman 1960; Ervin-Tripp 1986 [1972]). Instead, this volume sails into, if not uncharted waters, at least uncrowded seas. We turn now to the areas of central concern to this book.

1.3    Key notions for the empirical study of initial person reference

In this section, we introduce two sets of concepts that will prove useful in understanding the chapters included here. The first concerns marked and unmarked choices of referring expressions, and the second, a set of principles for organizing initial reference that have come out of work on the conversational organization of English.

1.3.1    Achieving reference: the logic of linguistic formulation

Throughout the review above we have noted the perennial issue of the speaker’s choice among multiple alternative means of referring, first at the level of which mode to employ (names, relators or descriptions), then at the level of which particular form (e.g., which name) to use. To these should be added non-linguistic means, particularly pointing gestures, which especially in small-scale communities can play an important role, as mentioned in some of the chapters that follow. Many of the chapters (e.g., Brown, Haviland and Levinson) refer to the critical role that co-speech gesture plays in person reference. All of the new contributions to this book are based on video-recorded data, in which we observe hand-pointing to play a role in person reference (cf. Enfield, Kita and de Ruiter 2005; Haviland 1993; Levinson 2005; Schegloff 1984). However, the chapters concentrate on linguistic resources – a systematic treatment of pointing in person reference awaits further work.

What dictates a speaker’s choice of referring expression? Contributors to this volume primarily adopt an interactional, social view of reference, where what matters are the actions being undertaken by interlocutors standing in specific social relations to one another in a social context. But to achieve reference, and thereby the associated social actions, speakers and hearers need reliable mechanisms for informational alignment, so that reference actually succeeds. This is not to deny that it may be important on occasion to keep reference vague (see Levinson, this volume and Garde 2003). A key mechanism is the distinction between unmarked manners of formulation and the marked nature of departures from these defaults, along with the special interpretations that these exceptional departures invite. We shall first outline what we mean by marked, as a way of bringing out some basic pragmatic principles of interpretation. We then focus on the informatics of person reference itself.

1.3.1.1    Marked and unmarked   For any recurrent type of coordination problem conventionally solved by the use of language, there should be an unmarked way to formulate it. In other words, if it is the kind of thing you need to say regularly, there will be a standard way to say it (Brown 1958). Correspondingly, saying it in some other way is marked. In one type of markedness, two items differ with respect to the presence of some extra specification. A semantically marked item has some extra semantic specification (e.g., Dutch hengst ‘non-castrated male horse’ vs. paard ‘horse’). A formally marked item has some extra explicitly distinct formal specification ( parent’s brother vs.uncle). Distinct from these is PRAGMATIC markedness, by which an item is unexpected or less usual in some context than a possible alternative (e.g., automobile vs. car in everyday conversation) (Ervin-Tripp 1986 [1972]; Levinson 2000).⁵

As an illustration from the realm of person reference, consider the second-person singular pronouns in Dutch: informal jij and formal u. Suppose that of these two jij is semantically bare, meaning simply ‘you’, and u has some additional semantic specification that accounts for its polite, formal, deferent, distant meaning (cf. Wierzbicka 1992: 319–24). Despite being formally/semantically unmarked, jij may on occasion be the pragmatically marked item. That is, using jij for ‘you’ in contexts where u is appropriate for reasons of politeness (e.g., in a service encounter) may be taken to index a choice NOT to use u, thus giving rise to an implication of disrespect. In such cases, pragmatic markedness is defined neither purely in terms of the linguistic nor the ethnographic system, but rather in terms of more locally defined contextual expectations.

It is critical to clarify what a claim of pragmatically unmarked or ‘default’ entails. Formal and semantic markedness are defined by properties of the linguistic system, and are therefore stable independent of usage context (e.g., in English, plural is formally marked by -s; singular is not formally marked). Pragmatic markedness, on the other hand, is by definition sensitive to social situational usage. The value of a particular type of formulation cannot be said to be marked or unmarked ‘for the language’ if this is taken to mean ‘unmarked across the full range of contextual settings in which that language may be used’. When contributors to this volume speak of unmarked or default manner of formulation for a given language/culture, this refers to a specific subset of contexts, in particular those that are characteristic of the kind of maximally informal, self-organizing mode of conversation among intimates that incorporates the fewest constraints on interaction, and which forms the type of ordinary setting from which the data are drawn. (See Haviland’s chapter for critical discussion.) Different defaults may apply in more constrained settings such as court proceedings, meetings and rituals of various kinds.

When a listener encounters a pragmatically unmarked formulation of person reference (e.g., ‘John’ in Where’s John?) he/she will not normally reflect on the selected manner of formulation, and as Schegloff (1996a) puts it, ‘nothing but referring is being done’ (but cf. discussion in Enfield’s chapter). On the other hand, when a listener encounters a pragmatically marked formulation of person reference (e.g., Where’s His majesty?), two questions arise in the listener’s mind. First, ‘Why is the speaker not formulating this reference in the normal, unmarked way?’ A generic answer is that the speaker is saying something other than the usual (Grice 1989) or, for person reference, ‘Because the speaker wants to do more than just achieve reference to the person’ (Schegloff 1996a). A second question for the hearer is: ‘Why is the speaker formulating this reference in THIS way?’ (i.e., specifically as His majesty and not any one of a million other possible departures from the unmarked). There is no generic answer to this second question. An analysis must look to the specifics of the formulation, where overt clues in the formulation itself should reveal what ‘more’ is being done (see chapters by Levinson, Oh and Stivers, this volume).

1.3.2    Principles of person-reference emerging from the study of english conversation

Backgrounded or even absent in much of the research discussed so far is a concern with the social action that is under way by virtue of some mechanism of person reference being selected and employed. This point has been explicitly acknowledged and promoted in anthropology (e.g., Bloch 1971), although it was first combined with detailed structural analysis of the moment-by-moment particulars of face-to-face interaction by Sacks and Schegloff in the early 1970s (following the work of Garfinkel and Goffman). Sacks discussed the concept of a ‘recognition-type description’ for places and objects in story telling. Speakers use this type of description in an attempt to secure a display of recognition from their interlocutor (Sacks 1992, vol. 2, p. 180). In Sacks’ example, a speaker refers to a location as ‘the main entrance there where the silver is an’ all the (gifts an’ things)’. This way of referring is specifically designed to secure some indication from the recipient that she knows the place.

Schegloff (1972) examined conversational reference to places, showing that people select from among alternative expressions in ways that are sensitive to the respective locations of the conversation participants, the social action being undertaken by the utterance in its context, and the identity of the recipient of the utterance (see Sacks, Schegloff and Jefferson 1974; Schegloff 1972 regarding recipient design more generally). These issues were critical to the development of Sacks and Schegloff’s (1979) account of person reference (reprinted as Chapter in this volume). Instead of being considered distinct on formal grounds, kinship terms, descriptions, and names were considered together as types of person reference because they were all used by speakers as means of, at the very least, achieving recognition of a person in the course of performing some social action (e.g., announcing news, complaining).

Sacks and Schegloff treated person reference as a systematic domain with its own structure and proposed two organizing principles for determining how person reference should be formulated: (1) a preference for using

---

© Cambridge University Press

Table of Contents

1. Person reference in interaction Tanya Stivers, N. J. Enfield and Stephen C. Levinson; Part I. Person Reference as a System: 2. Two preferences in the organization of reference to persons in conversation and their interaction (1979) Harvey Sacks and Emanuel A. Schegloff; 3. Optimizing person reference - evidence from repair on Rossel Island Stephen C. Levinson; 4. Alternative recognitionals in person reference Tanya Stivers; 5. Meanings of the unmarked: why 'default' person reference does more than just refer N. J. Enfield; Part II. The Person Reference System in Operation: 6. Conveying who you are: the presentation of self, strictly speaking Emanuel A. Schegloff; 7. Person reference in Yucatec Maya William F. Hanks; 8. Principles of person reference in TΔl Penelope Brown; 9. Non-initial person reference in Korean: choosing between quasi-pronouns Sun-Young Oh; 10. Person reference in Tzotzil gossip: referring dupliciter John B. Haviland; Part III. The Person Reference System in Trouble: 11. Intersubjectivity and progressivity in person (and place) reference John Heritage; 12. Repairing person reference in a small Caribbean community Jack Sidnell; 13. Reference and 'reference dangereuse' to persons in Kilivila: an overview and case study Gunter Senft.