6.7 Metalinguistic awareness

ization.25 While learning to speak a language, children become members of their society. In this perspective, language acquisition cannot be separated from language socialization – to be interpreted as socialization to language as well as socialization through language (Ochs and Schieffelin 1984). This new approach to language acquisition has established new theoretical and methodological standards for research. Thus, grammarians (e.g. Chomsky 1965) have typically assumed that the end goal of language acquisition is the production of competent speakers, but they have not analyzed the variety of meanings that such a concept might have for the members of a given group. Marjorie Goodwin (1990), for instance, argued that African American boys and girls develop a different kind of competence for verbally dealing with conflict (see chapter 9). Kulick (1992) showed that multilingual adult members of the Gapun community in Papua New Guinea socialize their children to become monolingual (in the local pidgin, Tok Pisin) even though they insist that they would like their children to also acquire the local vernacular (Taiap). In this case, as well as in the Haitian community in New York City studied by Schieffelin (1994), parents’ theory of what needs to be done to teach children to speak a given language and parents’ socializing practices are intimately related to their ideology of what is valuable and for what purpose. Reviewing these and other studies of language acquisition and socialization, Ochs and Schieffelin (1995: 91) have thus argued that grammatical development must be seen as an outcome of (1) the socially and culturally organized activities in which children participate regularly, and (2) the language(s) children are implicitly encouraged to acquire. These two dimensions of language development bring language acquisition studies back where they belong, namely, the context of children’s lives. This statement should not be interpreted as a permission to ignore biological and cognitive factors in language acquisition, but as an invitation to document progress in the acqusition of grammar in the context of existing interactional practices and existing ideologies of what it means to be a competent speaker in a given community.

6.7Metalinguistic awareness: from denotational meaning to pragmatics

Much of the progress that has been made in formal linguistic analysis in the last century is due to what the Russian linguist Roman Jakobson called the metalinguistic function of language, namely, the use of language to describe and analyze language (“a ‘cat’ is a three letter word,” “the German word for ‘development’ is ‘Entwicklung,’” etc.) (see chapter 9). This function is an essential part of the ability that native speakers – linguists included – use to isolate certain linguistic

25For a collection of child language studies in different languages that includes descriptions based on ethnographic work, see Slobin (1985a, 1985b, 1992).

199

Meaning in linguistic forms

forms and identify their meaning or function in discourse (either in an idealized, imagined context or in a specific one). It is the function of language that allows dictionaries and grammars to be written. The related concept of metalinguistic awareness is the knowledge that speakers have of their own language. Such knowledge is typically accessed through introspection and is considered by most linguists today an indispensable resource for linguistic analysis. There is no question that a great deal can be learned about a language by sitting down with native speakers and asking them questions about their language. This is often done with bilingual speakers to whom the linguist asks questions such as “How do you say — in your (native) language?,” “What does this (word/phrase/ sentence) mean?,” “Does this sound right?” “Does it make sense?” “Could it have a different meaning?”, “Is there another way of saying that same thing?” and so on. The same method can be used with monolingual speakers as researchers learn the few basic phrases and questions necessary to elicit a basic vocabulary first and then slowly build up a repertoire of increasingly complex grammatical patterns (e.g. compound words, complex phrases, sentences with embedded clauses). This procedure, based on native speakers’ intuitions and fieldworkers’ trained ear to transcribe what they hear, is a powerful instrument for gaining access to the grammatical patterns of any language. At the same time, much of the work in sociolinguistics and linguistic anthropology in the last three decades has shown that elicitation techniques alone can be problematic and should always be integrated with other methods, including hypotheses based on when an expression is used or how frequent it is in spontaneous speech. Some of the limits of elicitation techniques have to do with limits inherent in introspection as a guide to speakers’ knowledge of their use of language. William Labov’s (1966, 1972a) work on phonological variation, for instance, provides evidence of the failure of introspection to capture important regularities in the ways in which the same speaker changes pronunciation from one context to the next or the ways in which a community as a whole treats certain linguistic forms (by moving from one type of pronunciation to another or from a variable rule to a categorical rule). Speakers’ metalinguistic awareness does not include their ability to fully predict the type of variation their pronunciation exhibits in different social and cultural contexts.

As suggested by Silverstein (1981), the power of introspection and hence the reliability of metalinguistic awareness native speakers have may vary with the type of linguistic phenomena we are trying to describe. It seems relatively easy, for instance, for speakers to identify the referential meaning of a word when it names a concrete, visible object. In these cases, elicitation can be done by what philosophers call ostensive definition (“What is this?” “An apple.” “What does ‘foot’ mean?” “This.”).

200

6.7 Metalinguistic awareness

The same technique can be used for verbs or adjectives that describe activities or ways of being that can be represented by stereotypical movements or gestures. Thus, the meaning of the English verb walk can be illustrated by engaging in the act of walking and the adjective big can be explained by mimicking with the hands and arms something very large or tall. Things get more complicated, however, when we want to understand the meaning of words like intelligence, implication, responsibility. In these cases, we need to construct scenarios that can evoke such concepts. As anyone who has tried to teach a foreign language knows, we cannot always assume that a given gesture or even a whole anecdote will trigger the type of understanding we are looking for. When we get into constructing stereotypical scenes and characters, we are entering the realm of culture. How do we represent for instance the meaning of sad, happy, angry, offended, pleased? It depends on what conditions a given group of people see as associated with such states of mind. In some cultures, for instance, people might be characterized as sad when they are alone in their house, whereas in other cultures the same situation might be seen as a blessing!

It is even more difficult for native speakers to describe the relationship between linguistic forms and their pragmatic functions, that is, the use of speech forms to evoke or establish particular types of contexts, including the speaker’s stance or attitude, the social relations or relative status of the participants, and special attributes of particular individuals.

Extending Jakobson’s notion of metalinguistic function, Silverstein (1981, 1985b, 1993) introduced the term metapragmatic function for the use of language to describe such contextual aspects of speech-as-action (see chapter 7). He posited that the success one might have in getting access to a speaker’s metapragmatic awareness – that is, their ability to articulate the context for the use of certain linguistic expressions – was not random but tied to certain properties of the linguistic signs in questions. These properties include the referential quality of a linguistic form, that is, its ability to identify a particular referent (e.g. “the pronoun vous in this case refers to the speaker’s father”) and the relative creativity of the pragmatic sign, that is, the extent to which it presupposes the existence of what it refers to or instead helps establish it in the context.

According to Silverstein, when a linguistic sign seems to establish rather than presuppose a particular relationship, stance, or status, it becomes more difficult for native speakers to be aware of such functions. In other words, the more con- text-creating a linguistic sign is, the more difficult it is for native speakers to be conscious of its pragmatic force. I will illustrate this point with an example from my own work on Italian subject pronouns.

201

Meaning in linguistic forms

6.7.1The pragmatic meaning of pronouns

In Italian – as in many other languages, but not in English – the subject of a sentence with a finite verb does not need to be expressed for the sentence to be grammatical (a property of Italian syntax that has been called by generative grammarians “pronoun-drop”). Both (46) and (47) are perfectly acceptable Italian sentences and have the same denotational meaning, that is, they are considered true under the same conditions:

Grammarians who used native speakers’ intuitions (often their own) to study this phenomenon acknowleged that the presence of the subject pronoun in sentences like (46) is somewhat “marked” or special and suggested that it is due to emphasis or contrast.26 In other words, since Italian sentences do not need a full subject pronoun to be acceptable sentences and in fact most of the time do not have one in discourse, the presence of lui in sentences like (46) was interpreted either as an answer to a question about the identity of the subject (“who arrived at seven o’clock?”) or as a follow-up to a prior assertion (“nobody arrived at seven o’clock”). When I examined transcripts of Italian conversations, however, I found that the actual use of third person subject pronouns27 suggested the possibility of a different analysis. I discovered that, rather than indicating contrast or emphasis, pronouns like lei “she” and lui “he” tended to be used for main characters, that is, referents who were recurrently talked about and toward whom the speaker displayed interest and positive affect. The same pronouns were not used for minor characters, which were instead referred to by demonstratives like questo “this one” and quello “that one.” When a demonstrative was used for a main character, it co-occurred with negative evaluations, that is, it coincided with the speaker describing someone as incompetent or annoying. I concluded then that personal pronouns tended to be used for positive affect and demonstratives for negative affect (Duranti

26Summarizing the position of generative grammarians working on Italian, Haegeman (1994: 21) writes: “When no contrast or no special focus on the subject is needed the [subject] pronoun is absent.”

27I restricted the analysis to pronominal expressions referring to individuals who were not present.

202

6.7 Metalinguistic awareness

1984b). None of these pragmatic factors had been available to native speakers’ consciousness.

In light of Silverstein’s suggestions, one might speculate that the unavailability of this analysis to speakers’ intuitions is related to the fact that the property of being main character and the property of being presented in a positive light are considerably high on a scale of pragmatic creativity. These properties, in other words, are not independent of pronominal usage. They are, instead, established through such discourse resources as the type of pronominal reference that is used (no subject vs. pronominal subject) (Duranti 1991). Given linguistic anthropologists’ interest in the use of language as a resource for establishing particular institutional context and cultural practices, much of their work is likely to center around the more context-creating uses of language.

The more dependent on context a given expression, the more difficult it is to describe its functioning by just using intuitions about isolated sentences or made up stretches of discourse. Knowledge of grammar that includes linguistic structures as well as conditions for their use must thus be obtained by integrating elicitation and introspection with observation and documentation of language use. Recent work on metapragmatic awareness displayed during real-life interaction suggests that what might be difficult to evoke in elicitation contexts might instead be naturally produced during spontaneous interaction. Marco Jacquemet (1994) examined “pronominal violations” during court hearings involving the pentiti di Camorra, witnesses who had belonged to the criminal organization known as “Camorra” (the Neapolitan version of the Sicilian Mafia), had “repented” and decided to collaborate with the Justice Department. He showed that speakers involved in public confrontations in front of a judge often produce what he calls “metapragmatic attacks,” that is, inpromptu accusations based on the use of address forms by their interlocutor. These attacks typically take the form of complaints about the offence produced by the use of a T-form, that is, a second person singular form in place of a more formal V-form (Brown and Gilman 1960):

(48)(From Jacquemet 1994: 307, with slightly different transcription conventions, translation, and the addition of the features [T] and [V] to signal the address form being used by the speaker. Capitals for pronouns indicate that a full pronoun is being used instead of only the subject-verb agreement.)

203