Language and the Construction of Self: Developmental Reflections

Nancy Budwig

Clark University

Language and the Construction of Self: Developmental Reflections

Recent discussions have placed increased focus on the role of language in the construction of the self. While there appears to be increasing consensus that language plays a major role in this process, there has been little explicit discussion of how language is meant to be understood (see though Morgan, 1994). Some researchers have highlighted the extent to which language ought to be viewed as discursive action. On this account self is a fluid construct continually being constituted in face to face interaction. Others have placed more emphasis on notions of grammar, looking in particular at the role particular language systems (i.e. pronouns) play in the construction of self.

As I will argue in more detail below, the view one adopts towards language will necessarily have consequences for how one looks at ontogenesis. To say that semiotic systems such as language play a role in the child's construction of a socioculturally appropriate sense of self is to say little about the mechanisms by which this process takes place. And whether one views language as discursive action or in terms of grammatical systems indeed has consequences for the particulars of this process.

In the first part of this paper, I discuss some central ways language has been viewed focusing the review on social constructivist writings as well as those stemming from the study of human development. In the second part of this paper, I discuss data that leads to the reconsideration of aspects of the language - selfhood interface. I conclude by suggesting some future avenues of research.

Language and Its Role in Understanding the Construction of Self

Social Constructivist Positions on Self: Two Views of Language

Over the past years, there has been an emerging consensus among social constructivists that suggests the centrality of language to discussions of the construction of identity (see for instance Shotter & Gergen, 1989). This reflects the changing view of what language is and how it functions within everyday interactions. Rather than viewing language as a tool for representing an already given world, social constructivists have highlighted alternative functions of language. While there is consensus towards viewing language as action, and while many social constructivists share the view that, in fact, language is central to understanding identity, as Morgan (1994) has pointed out, within social constructivists' writings there is less agreement on what language is taken to be. Numerous terms including language, discourse, texts, narratives all are used in overlapping ways.

We will begin our discussion with a brief outline of two ways language has been treated in social constructivist writings. This discussion is not meant to be exhaustive but rather is meant to help pave the way for reflecting on two related issues - first, the relationship between social constructivist writings and research that focuses on children's self development; and second, the implications of alternative ways of viewing language for thinking about the claim that language plays a major role in the construction of self. In discussing the various ways language has been treated in social constructivist writings we will focus on a split between some who view language in terms of grammatical options and those who focus more on the notion of language as discursive action without special reference to linguistic forms.

Language as Grammar

One way in which social constructivists have viewed language has been in terms of typological options provided by language. The claim here is that language as a grammatical system provides a variety of ways for the speaking subject to insert self and other into discourse. Shotter (1989) has pointed out that ordinary language provides various mechanisms for situating agents in discourse. He discusses options such as voice contrasts and person to show that subtle variations in how speakers say somewhat similar content actually provide distinct positional fields for the subject. Shotter (1989) provides several examples of this position:

Ordinary language marks a number of important distinctions, to do with articulating the character of the situation in which an actual utterance is produced, in both the voice and the person of verbs. In the simple active voice, the subject of the verb, the agent, does something to someone or something other than or separate from itself. In the passive voice, the agent is de- emphasized and often goes unmentioned, so that an outcome can be described without it being necessary to indicate explicitly who or what was responsible for it. In other words, to talk in a different voice is not merely to say the same thing in a different manner or style: it is to represent, in one's way of speaking, the way in which the subject of the verb in one's utterance (which might of course be oneself) is actually involved in the process depicted by that verb - for instance, whether the subject is morally committed (or not) by its actions to those to whom its actions or statements are addressed (pp. 133-134).

Here with reference to voice alternation, and elsewhere with reference to other grammatical devices, Shotter supports the view that it is certain grammatical features of language that provide speakers with resources for various presentations of self.

Others adopting what I have referred to as a grammatical approach to language have focused more on the range of typological variation in the languages of the world arguing that such variation simultaneously provides speakers of those languages with distinct ways of constructing themselves and others. This comparative approach can be found, for instance, in the writings of Mülhäusler and Harré (1990):

In the Indo-European systems we have so far discussed, the first person is so used as to be indexical of speaker as morally responsible for his or her utterance. But one must be wary of taking this as a universal feature of the first person. One will go seriously astray by treating the Japanese first-person register as some kind of elaboration on the pronoun I. Not only are nearly all uses of the first-person indexical of the relevant 'me- group' rather than the speaker as an individual but the register includes a complex of clines of formality and informality that are group defined (p. 112).

It is not the case the Mülhäusler & Harré (1990) adopt a completely typological perspective. That is, they do see a sort of interaction between cultural ways and linguistic options presented by distinct grammatical systems as is reflected in the following statement:

Here we do have a stronger Whorfian effect. We shall try to show that there are distinctive senses of self identifiable in diverse cultures with languages that differ in just the dimension of indexicality of the first person ... We shall tackle this issue by looking both at the changing conception of self that has accompanied the social history of our own culture and at different contemporary grammars in which quite different shadows are cast by quite diverse grammatical systems (p. 18).

While these authors, then, are not denying other aspects might play a role in self construction, they have argued that language in general, and variation between languages in particular, provide speakers with distinct grammatical options for situating the subject.

Language as Discursive Action

There have been many ways in which language has been viewed in terms of discursive action, though for the present purposes we will review a few examples to highlight the basic point that there is little agreement among and, at times, within particular authors' writings regarding how language is to be understood. Davies and Harré (1990) summarize this view when discussing their notion of "positioning" by suggesting:

The view of language in which positioning is to be understood is the immanentist view expounded by Harris (1982), in which language exists only as concrete occasions of language in use. La langue is an intellectualizing myth - only la parole is psychologically and socially real. This position is developed in contrast to the linguistic tradition in which 'syntax', 'semantics' and 'pragmatics' are used in a way that implies an abstract realm of causally potent entities shaping actual speech (p. 43).

Similarly, Potter & Wetherell (1987) claim: "The concern is with language use: the way accounts are constructed and different functions" (p. 157).

While there are many differences between the ways different theorists have defined discourse and discursive action (see Morgan, 1994), to a certain extent such work differs from that described above in terms of method. Those adopting a discursive approach tend to carefully analyze actual conversation.

Just as grammar was said to offer the speaker options for constructing self and other, according to this perspective it is within the forum of conversational activity that speaker's find resources:

With positioning, the focus is on the way in which the discursive practices constitute the speakers and hearers in certain ways and yet at the same time is a resource through which speakers and hearers can negotiate new positions (Davies & Harré, 1990, p. 62)

Summary

Despite the differences in the research reviewed above in the discussion of the role of language in social constructivists accounts, one finds an important similarity among all the various positions in that each views language as providing more than a methodological tool for the researcher. Adopting a more controversial position, these authors have argued that language, in the various ways it is understood, plays a fundamental role in the actual construction of selfhood. Thus, language offers more than a tool for the researcher as a way to "underlying inner life"; it is claimed that through language speakers come to construct and deploy ever changing subjectivities.

While the above review holds together to the extent that the authors share the view that language plays a causal role in the construction of self, I have argued that there is quite a lot of diversity with regard to what specifically language is, and next to no discussion about how it is that language plays such a role. I have noted for instance that some have looked at language from the perspective of grammatical options found within (i.e. voice alternations, pronoun systems) and between languages. Others have focused on language in context, and as we have noted, while there is a lot of variation in terms of what this is taken to mean, the commonality between this group is the attempt to look at language as action without attention to "abstract" analyses of formal structures.

Developmental Approaches: Language as Tool, Language as Mechanism

The idea that language plays a role in developmental theorizing and research is not new. What has changed is the nature of the role language is thought to play. For quite a while, developmental psychologists have made use of language as a tool. That is, language is viewed in terms of what it offers the researcher in the form of a methodological tool. In contrast to this understanding of language as method, more recently there has been a growing amount of research that has explored language as mechanism, that is, the way language might provide the child with a wedge for constructing self. This shift in part reflects a general shift in gravity within psychology concerning the relationship between language and human functioning. While there is growing acceptance of assumptions related to the language as mechanism account, at the same time, a tremendous amount of developmental research continues to be conducted which adopts the view that language, at best, provides the researcher with a tool for understanding.

Examining research in the area of self development, we will turn now to a brief consideration of each of these approaches to language, starting with a brief discussion of language as a research tool, and then turning to a more complete consideration of language as mechanism for the child. The review of the language as mechanism approach will highlight the extent to which we are a long way from knowing how language might play a central role in children's self development.

Language as Method: Tool for the Researcher

Several different examples of research concerning self development illustrate the stance that language provides a method for the researcher to better understand the child's growing ability to conceive of self and others. Two kinds of examples will be reviewed here, first, studies examining self and other recognition; and second, studies of slightly older children which examines spontaneous verbalizations and answers to interview questions concerning children's growing sense of self and others.

Self recognition. The work of Lewis and Brooks-Gunn (1979) provides an excellent illustration of how language is thought to play a role in understanding children's growing abilities at recognizing self and other. As part of a larger study of children's knowledge of self, Lewis and Brooks-Gunn studied the spontaneous vocalizations of the infants in two studies of children ranging in age between 9 and 36 months of age. Both production and comprehension were studied. Study I focused on the production of verbal labels and assessed the frequency of verbal labels, the children's use of age- appropriate labels, and gender-related labels. While the specifics of the findings need not concern us here, what is relevant is the authors' beliefs that language provides the researcher with insight into underlying conceptual categories of the child. Note for instance the following statement by Lewis & Brooks-Gunn (1979):

The existential self or infant's knowledge of self as distinct from other may be inferred through the use of one's own name or a personal pronoun to label an unfamiliar picture of self and not to others. The categorical self or the categories used by infants to define themselves may be inferred through the social labels that they apply to persons varying on a number of dimensions (p. 162).

Such research nicely highlights the way in which developmental psychologists have appealed to the study of language as one way of assessing children's self development.

Self understanding and theory of mind. Going beyond verbal labels, developmental psychologists have used other sorts of language-dependent measures to find out about children's developing notions of self and other. In addition to making use of spontaneous verbalizations as indicators of underlying conceptual categories, another popular method has been to make use of interview techniques (see for instance Damon & Hart, 1988; Shatz, Wellman, & Silber, 1983; Wellman, 1990). In fact research in the area of "theory of mind" has relied so heavily on linguistic measures that some authors have explicitly addressed their concern with such an approach (see Wellman, 1990, pp. 203-205):

Concerns arise, therefore, because the methods I have used are all language dependent. ... There are two essential reasons for this reliance on language. First, in many cases there is no better alternative. Fortunately, and second, lack of better alternatives is due, in part, to the fact that language-dependent measures prove extremely revealing (p. 203).

Summary. Pulling together the various strands of research reviewed above, I would suggest that what holds such views together is that language provides a method that helps the researcher uncover children's budding conceptions of self and other across development. Note that it is not the case that these researchers suggest that language itself might provide a tool for the child to gain access to culturally appropriate notions of self.

Language as Mechanism: Tool for the Child

A second group of studies can be distinguished from those reviewed above to the extent that these researchers adopt the view that language not only plays a role in aiding the researcher, but the child as well. Here the argument is made that language itself plays a fundamental role in the child's coming to construct notions of self and other. While various researchers share the general belief that language plays a role in the child's construction of self, this position has been developed in quite different ways by different researchers. While language is viewed as mechanism, what language is taken to be varies according to different accounts. We will turn now to consider some of the similarities and differences in these positions.

Language as Significant Symbols. For Mead (1934) the development of self was necessarily dependent on the development of language:

... the language process is essential for the development of self. ... The self is something which has a development; it is not initially there, at birth, but arises in the process of social experience and activity, that is, develops in the given individual as a result of his relations to that process as a whole and to other individuals within that process (p.135).

Mead was clear that it was not interaction per se, but communication in the sense of "significant symbols" that contributes to the development of self:

The importance of what we term "communication" lies in the fact that it provides a form of behaviour in which the organism or the individual may become an object to himself (p. 138).

Language, then, is important in that it allows the individual to take the role of others and consequently perceive the self as object.

Language as Typological Variation. Yet another way researchers have begun to look at the impact of language on self development draws upon the work of Whorf (1956). Adopting a typological stance it has been suggested that the grammatical options provided to speakers of given languages influence their sense of self. This sort of view has been introduced into modern day developmental research by Lock (1980) who argues:

It is important to remember at this point that the child sees himself in relation to the way others see the world. Remember also that it is through language that the world is given its form (p. 191).

Lock goes on to discuss differences in grammatical structure of many distinct languages, noting the rich typological variation in linguistic structures and concludes by suggesting:

There are, thus, many linguistic realities for infants to be guided into; and consequently many potential conceptions of self (p. 192).

A very similar tie between grammaticalization of self in various languages and how this might impact on one's sense of self can also be found in Mülhäusler & Harré (1990) and Kerby (1991). For instance, Kerby (1991) summarizes:

On a narrative account, the self is to be construed not as a prelinguistic given that merely employs language, much as we might employ a tool, but rather as a product of language - what might be called the implied subject of self- referring utterances ... Of importance here is the way language prefigures a place for the subject in grammatical forms such as personal pronouns and adverbs of location (here, now, then, etc.) (pp. 4- 5).

What holds together this group of researchers is their belief that structural properties of language will impact on children's construal of self. The assumption is that children growing up in different language communities, would be likely to develop different senses of self to the extent to which language provides different ways of indexing oneself given differences in languages' typological make-up.

Language as Action. This third view emphasizes functional aspects of language and their role in indexing in face-to-face interaction how self is construed in different contexts. Such research, carried out under the heading of language socialization, argues that language is one of the major socialization tools (see Miller, Potts, Fung, Hoogstra, & Mintz, 1990; Ochs & Schieffelin, 1984; Schieffelin & Ochs, 1986). The functionally guided view of language is revealed in the following statement by Ochs & Schieffelin:

The process of becoming a competent member of society is realized to a large extent through language, by acquiring knowledge of its functions, social distribution, and interpretation in and across socially defined situations (p. 277).

Miller et al. (1990) have argued that face-to-face interactions with significant others provides an arena for the sharing of personal experiences and simultaneously provide an arena for the construction of self (p. 293).

While linguistic typological diversity is acknowledged among those advocating this view of language, it is the discursive aspects of language in use that provide children with mechanisms for self construction.

Summary. In this section, I have reviewed three perspectives that suggest that language provides an essential way in which children come to construct self. While all of these perspectives are in agreement that language plays a pivotal role, each view differs in the aspects of language that they see as relevant to this process. We have noted for instance that for Mead, it was the ability through language for the child to come to represent self as object. We have also seen an interest in the typological variation provided by speakers of diverse languages and the ways such variation might provide distinct conceptualizations of self and others. Finally, I have suggested that for another group of researchers it is through coming to learn to use language that children simultaneously acquire cultural ways of experiencing themselves, not so much because of typology but because of differences in the ways language is embedded in cultural ways of being.

Relating Language as Method and Language as Mechanism

Before turning in the remainder of this article to a consideration of how language might play a role in children's construction of self, I would like to raise three issues that arise when one examines findings from previous research concerning self development. First, I will address some differences between the social constructivists' views of language and that described in the previous section concerning language as mechanism in developmental theorizing. Second, I will suggest that until now much of the work has necessarily been dedicated towards the acceptance of the proposition that language can play a role in self construction, rather than outlining the specifics of this process. Finally, drawing upon a combined consideration of the developmental research reviewed above, I will suggest some necessary changes in assumptions about children in the language as mechanism approach that are necessary if we are to better understand how language could impact on children's construction of self across development.

Comparing Views of Language as Mechanism

What is interesting about comparing the modern work that has gone on within social constructivist thinking with that that has been carried out within developmental theorizing is that the social constructivists have more explicitly addressed one another's various positions about the language as mechanism position they are developing, and at the same time, one and the same author has addressed alternative construals of how language might play a role in self construction. In contrast, there has been little explicit comparison of this sort within writings of the developmental theorists - different theorists have worked with different ideas about what language is without much discussion of alternatives (see for some exceptions though those listed under view 3 above). As I will show below, it is essential to simultaneously consider what various perspectives of language can contribute, in particular how typological and functional distinctions in use might impact on children's constructions of self.

The Notion of Development

Looking through literature pertaining to the language as mechanism approach it becomes clear that most energy within such a position has been dedicated towards defining what language is or illustrating that children growing up with distinct typological systems or discursive communities have access to different kinds of language. What has yet to be unravelled is the role of distinct typologies or communities over developmental time. For instance, those adopting a typological perspective have suggested that "children" raised to speak different typological systems are provided with distinct linguistic options about self, but I know of no research that examines how such differences may differentially impact on the child across developmental time. Similarly, those working within the area of language socialization have made cross-cultural comparisons and have highlighted important differences in the discursive communities children are raised in, but it is not clear whether and how such differences impact on individual children over time. Until now, we have little in the way of longitudinal analyses to address such questions.

The Active Role of the Child

Researchers working within the language as method approach to self development have highlighted the active and creative role children play in their own development, and the fact that across developmental time children's notions of self and other often differ from adults and more developed others. This leaves open yet another question for those adopting the language as mechanism approach. Even if language offers a rich resource from which children can construct self and other, it remains to be understood how children's budding notions not only of themselves and others, but of language itself, interact with their use and understanding of language in context.

I will turn now to a set of studies I have conducted which begin to address these issues. It should be noted before doing so that this research was designed originally to better understand children's language development, rather than the role language plays in other domains of human development. I present the findings not because they answer the questions raised above in this section, but because of my belief that the sort of approach to the data offers a useful starting point for future work that hopefully will better address these questions. The work I will review focuses on children's and caregivers' indexing of self and other in spontaneous interactions.

Indexing Self and Other in Discourse: A Variation Analysis of American and German Caregiver-Child Interactions

Methodological Starting Points

Variation Analysis

The studies I will review below draw upon an approach to discourse that rarely is found in psychological studies drawing upon language data. This approach has been referred to by Schiffrin (1994) as variation analysis. This kind of discourse analysis, with its origins partially in linguistics, examines distributional patterns concerning alternative ways of saying what appears to be the same thing. The guiding assumption is that both linguistic and social factors can account for the variations in patterning.

What is particularly helpful about variation analysis as an approach to discourse for the current purpose of examining indexing of self and other in caregiver-child interactions is the simultaneous focus on linguistic structure and discourse function. This provides a way to integrate aspects of language which the review above suggests developmental psychologists interested in self development have previously considered distinctly. I have noted above that when researchers have appealed to language data in arguing that language either provides a tool for the researcher or a mechanism for the child, they rarely have looked at forms in context and have rather focused either on language structure or function. I will turn now to give a flavour of a series of studies I have conducted that makes use of variation analysis in examining first how children, and then, how caregivers index self and other in face-to face interactions.

Original Research Goals

As noted above, the studies I will draw upon were not originally designed to shed light on how language might play a role in understanding the construction of self. The studies of the children's language were part of a larger study of grammatical development in children. Previous research in the area of cognitive linguistics has indicated that all languages make use of morphosyntactic devices to index prototypical agency and various deviations (see DeLancey, 1984). Slobin's crosslinguistic research provided some evidence that young children link the use of grammatical devices to salient conceptual schemes and Slobin's discussion of Basic Child Grammar suggests that agentivity is one of the first schemes receiving grammatical treatment by children across the world (see Slobin, 1985). Drawing upon Slobin's notion of Basic Child Grammar and extending it in light of Ervin-Tripp's notion of control moves (Ervin-Tripp & Gordon, 1986), I hypothesized that children's earliest uses of pronominal and nominal self reference forms were guided by an inter-related set of semantic and pragmatic notions concerning agency and control. Looking at the development of grammar from a functionalist perspective (see Budwig, in press a for review), I aimed to show that the children's seemingly different ways to refer to themselves could better be understood as their attempts to situate themselves distinctly in activity frames.

About The Participants

In what follows I will draw upon two data sets collected between 1983 and 1992 in America and Germany. The American data stems from a longitudinal study I conducted of six white middle-class families in Northern California. The children, ranging between 20 and 32 months of age at the onset of the study, all attended a daycare center associated with a large university. Two girls and four boys were video-taped for a four month period, twice each month, either in dyadic play with a peer or with their caregiver. The play sessions included three activities introduced by the researcher: play with wooden blocks, play with manipulative toys, and looking through a photo-book which included pictures of the children and their peers engaged in activities at the daycare center.

The data from the German-speaking participants included longitudinal video recordings of three caregiver-child dyads (2 girls and 1 boy) growing up in and around (ex-East) Berlin just after unification. The dyads were studied longitudinally over a four month period in their home while playing with manipulative toys and building blocks. The children ranged from 20 - 30 months of age at the onset of the study. They all were just beginning to combine words in the month preceding the first video session.

After the video session, transcripts were made of all verbalizations of the caregiver and child. The German transcripts were prepared by a native speaker of German and all transcripts were entered into computer format using a modified version of the CHILDES Chat format (see MacWhinney, 1991).

Children's Indexing of Self and Other

The English-speaking Children

Preliminary analyses. A preliminary set of analyses led me to divide the children into two groups: ego- anchored children and nonego-anchored children. This distinction was based on three kinds of analyses: 1) an examination of mean length of utterance (MLU); 2) the distribution of reference to self and others in subject position; and 3) the range and distribution of self reference forms by the children. The three ego-anchored children (Grice, Jeffrey, and Megan) could be characterized as follows: 1) their MLUs were below 3.0 - that is they were in the earliest phases of combining words; 2) they primarily referred to self in subject position (i.e. 75% or more of their references to persons in subject position were references to self); and 3) these three children regularly relied on multiple forms of self reference, often in subject position in ways that deviated from adult usage (i.e. My open that, Me jump). In contrast, the nonego- anchored children (Eric, Keith, and Thomas) 1) had MLUs over 3.0; 2) more equally distributed their references to self and other; and 3) used a range of self reference forms, but primarily relied on the single form I and used self reference forms in conventional ways. With these preliminary findings in mind, we can turn now to summarize the results of the variation analyses of the two groups of children's use of self reference forms.

Ego-anchored children. Each of the three ego- anchored children drew upon several different forms to refer to themselves. Two forms which were used with great regularity by all children were I and My. All utterances with self reference forms were isolated and coded in terms of a multi-level coding scheme which examined various semantic and pragmatic factors related to the degree of agency and control expressed in clauses surrounding usage of these forms. The variation analyses support the conclusion that at a time before the ego-anchored children regularly referred to others, they temporarily borrowed self reference forms to contrast degree of agentivity and control expressed. I was found in clauses with stative verbs in which self was viewed as "experiencer" in assertions about the world. In contrast, the children employed My in utterances in which the child acted as prototypical agent, often by using language to bring about change.

The following example illustrates this contrast:

(1) Megan (MCH) and Mom (MMO) are playing with manipulative toys.

a. MCH: I want that one. (lifting container)

b. MMO: Oh you want that one, okay.

c. MCH: (tries to open container then says:) My open that.

d. MMM: What?

e. MCH: My open that Mommy. (giving container to MM)

f. MMO: Wanna open that?

g. MCH: Yeah (MM opens container)

In this example, the twenty-month-old child, Megan, makes use of two different self reference forms. The form I is used in line (1a) as she expresses her desire to play with a container which contains a nut. The mother's utterance in line (1b) and her lack of action support the conclusion that the mother had also contextualized the child's statement as an assertion of desire, in contrast to a request for the mother's assistance. Note though, that as Megan recognizes that she is unable to open the container on her own to get the nut out, she gives the container to her mother. What is interesting is the shift in the way she indexes herself in this utterance, switching in lines (1c) and (1e) to the more dynamic form My.

A comparative analysis of the ego-anchored children's use of these two forms revealed a cluster of inter-related semantic and pragmatic features that linked up with the use of such forms. As has been reported elsewhere in detail, distributional differences were found at the individual coding levels of semantic meaning and pragmatic function, and in the combined distributional analyses (see Budwig, 1989, 1990, in press a). For instance, comparing the use of I and My in subject position by the ego-anchored children we find that 100% of the instances coded as high in semantic agentivity and pragmatic control co-occurred with the use of My. In contrast, 82% of the usages that were coded as ranking low in both semantic agency and pragmatic control involved the use of I. Pulling together the analyses the ego-anchored children we find that they indexed themselves with several different forms, and that each form linked up with a distinct way of situating themselves in fields of human action. Of central importance to these children was marking self as prototypical agent, and various deviations, including, self as experiencer of mental states.

Nonego-anchored children. In contrast to the ego-anchored children, we have noted that the nonego-anchored children primarily indexed themselves with one form namely, I. As I have illustrated in greater detail elsewhere, the nonego-anchored children did not distinguish the use of I and My based on degree of agentivity and control. Although I was the preferred form for indexing self at the low end of the agentivity and pragmatic control scales (i.e. 94% of all such references involved I), this same form was also the preferred form at the high end of the scales, accounting for 86% of all usage of I and My by the nonego-anchored children. Thus the nonego-anchored children's usage of self reference forms differed from the ego-anchored children not only because such usage conforms to adult English (i.e. there were no instances of My and Me in subject position in ways that deviated from adult usage), but also to the extent that individual forms used to index self were not specifically linked to notions of agentivity and control.

While the nonego-anchored children did not contrastively employ self reference forms to take different stances on their own role in human action, it is interesting to note that about the time children give up such systems they develop other linguistic resources such as modal forms (i.e. gonna, needa, Wanna, hafta) and begin making use of voice contrasts (i.e. shifting between active, passive, and middle constructions). Such usage has been noted to link up with demarcating other deviations from prototypical agency, such as generic agency and contrasting motives concerning the origins of actions (see Budwig, 1990; Gee & Savasir, 1985) for further discussion.

Summary. Thus far, I have summarizing some analyses examining English-speaking children's indexing of self and other. Between the ages of two and three, I have noted that the group of children I studied underwent a shift in who they talked about and how they talked about them. The central shift can be summarized as follows. The children began by primarily referring to themselves and relied on a range of self reference forms to situate themselves in different ways. The slightly older nonego-anchored children referred to both self and other and did not contrastively employ self reference forms to adopt various stances of self, though I have suggested they make use of alternative linguistic resources to index different ways self and others can be situated in fields of action.

The German-speaking Children

Preliminary analyses. In contrast to the English-speaking children, the German children could not be divided neatly into two distinct groups based on the same qualifications. All three German children, like the ego- anchored children were at the earliest phase of combining words; all three had an MLU of under 3.0 and in fact two of the children had MLUs of under 2.0. While this made them appear to fit with the ego-anchored children described above, there were important differences with regard to the other two kinds of preliminary analyses. As noted above, the English- speaking children categorized as ego-anchored primarily referred to self as subject of their sentences. This was not the case for any of the three German children. The three German children referred to self approximately 50% of the time (range = 49% - 54%) across the four month study. So on this dimension, the three German children appeared more similar to the nonego-anchored children described above. With regard to the third set of preliminary analyses, namely the range and extent of deviations in self reference form usage, the German children looked both like the ego-anchored and nonego-anchored children. Like the ego-anchored children each child made use of a variety of pronominal and nominal forms to refer to self and other; and similar to the nonego-anchored children, no deviations of the sort noted for the English speaking children could be found (i.e. instances such as My cracked the eggs, Me jump). One final point worth mentioning is that although the two groups of children produced self and other references at about the same rate, the German group was far less likely to do so with pronominal forms (see Otto, 1994).

Variation analyses. Although there was some individual variation in the quantity of forms used and the rate of acquisition, there was an overall tendency for all the German children to link the use of self reference forms with particular functional contrasts. For instance the children used nominal forms ("Own Name" and "Other's Name") when referring to self and other as actor. The first person possessive pronoun meine was used in more dynamic utterances in which the child wanted to gain or maintain control of objects. Similarly the second person nominative pronoun du was used when commanding others to act in particular ways. The following examples illustrate these contrasts in the speech of one child, Katrin:

(2) Katrin and her mother are playing with the manipulative toys:

a. KCH: Katrin's, Katrin's, da Katrin's (taking items one by one out of toy bag) [>] [Katrin's, Katrin's, there Katrin's]

b. KMO: das [<] ist nicht Katrin's.

[that is not Katrin's]

c. KCH: meine!

[mine]

(3) Katrin and her mother are building with blocks:

a. KMO: Katrin, das [//] kuck mal, dann machst du das ja ganz doll kaputt.

[Katrin, look, then you will destroy it (referring to the block tower)]

b. KMO: # dann fallen ja alle Baustine raus.

[then all the blocks fall out]

c. KMO: # so (adjusting block tower)

[so]

d. MCH: meine Teurme!

[my tower]

In examples 2 & 3, the child is contrastively employing her own name or the possessive form to index two distinct stances. The nominal form appears in example 2(a) in acts of labelling objects, while in lines 2(c) and 3(d) the possessive form appears in acts where the child attempts to control objects. This contrast between a more volitional self and representational self has not only been noted in German by others (see Kolodziej, Deutsch, & Bittner, 1991) but also in English (see Deutsch & Budwig, 1983).

In contrast to the ego-anchored children, the German children index a generic perspective. That is, they often referred to their own actions with the impersonal form man [one] as in the example included below:

(4) Katrin and her mom are playing with a basket of pretend cooking toys.

a. KMO: Och, das ist nicht wie 'ne Kartoffel so weich, kuck mal, Nuesse sind ganz hart.

[Oh, look, that is not so soft as a potato, nuts are very hard.]

b. KCH: Loeffel Loeffel [>] kann man mal. (using a spoon to manipulate nuts)

[spoon, spoon one can a bit.]

c. KMO: Die [<] muss man +...

[this one must]

In this example and others like it, the child is de- emphasizing her own subjectivity and rather appears to be describing normative procedures of ways anyone doing what she is doing might proceed.

A third kind of contrast found in the German children's productions was one between references to others with nominal (i.e., other's name) and pronominal forms. As is revealed in Katrin's contrasts below, nominal forms appeared in referential contexts of making assertions about previous actions of the other which weren't under dispute. In contrast, Du (the second person nominative form) appeared in the context of control acts when the child attempted to order others to act in particular ways.

(5) Katrin (KCH) and her mother (KMO) are beginning to clean up blocks with one of the researchers (RES) nearby.

a. KCH: (Hands block to researcher)

b. RES: Dankeschoen.

[thanks]

c. KCH: Du einraemen!

[you clean up]

d. KMO: Das soll einraeumen der Onkel?

[The uncle should clean that up?]

e. KCH: Ja.

[Yes]

f. KMO: Na, und ihr raeumt die bieden hier noch ein,

ne?

[Hmm, and you clean the two of them in here again, Hmm?]

g. KCH: # Onkel raeumt, nicht?

[Uncle cleans, hmm?]

Note the switch in forms for referring to the researcher in this example. For instance, when Katrin hands the researcher a block with the expectation that he will help her clean up, he replies by saying "Thanks" pretending to interpret her gesture as a gift. Katrin replies with a command instructing the researcher on how to act using the Du form in line c. In contrast, in line g when the research assistant has complied with her request and the child reports to her mother that he is in fact cleaning up, Katrin uses the nominal form.

Summary. The German children, like the American children, drew upon a range of linguistic forms to index themselves and others and contrastively employed such forms in distinct activity contexts. In contrast to the American children, when the German children began to combine words they were equally likely to refer to themselves and others. Furthermore there were some differences in the kinds of contrasts that were indexed in the German and American children's talk. While the American children reserved the use of one form (my) for talk about self as intentional agent no such usage was found in the German children's talk. And the German children developed a way to index self in terms of impersonal agency in a way not noted for the American children. Thus while all children found ways to index various perspectives taken on self, what distinctions were drawn varied between the two groups of children.

Caregivers' Indexing of Self and Other

American Caregivers

Two sets of analyses were carried out with regard to the caregivers' indexing of self and other. First, we coded all instances of self reference made by the three caregivers of the ego-anchored children. Next, we examined all instances of self reference produced by the caregivers of the nonego- anchored children (see Budwig & Wiley, 1995, Quick, 1991). The primary question guiding our analyses was whether the two groups of children were receiving distinct input from their caregivers which might have influenced the distinct self and other reference systems the children constructed. At the most general level, we can report that the findings from the analyses of the ego-anchored and nonego-anchored children's caregivers were similar. That is, there were no apparent differences in the sort of input the two groups of children received. Such findings ran counter to our expectations for differences between the two groups of caregivers.

Our analyses also revealed a second set of unexpected findings. Based on what is known about English typology, we expected that the caregivers would primarily refer to themselves with the pronominal form I and we expected this form to appear with a variety of semantic predicates including both mental state verbs and action verbs. That is, given our knowledge of English grammar we did not anticipate the caregivers to restrict usage of I to a particular semantic cluster as was noted for the ego-anchored children. What was surprising was the finding that in fact the majority of all caregivers' uses of I linked up with references to self as experiencers in assertions. That is, when the caregivers indexed themselves with self reference forms, primarily they did so in utterances such as I like your little teapot. They rarely used self reference forms when referring to self as agent and when doing so it always was in the context of a permission request to the child, thereby down-playing their own agentivity. In sum, while the American caregivers never used pronominal forms in the non-standard ways noted above for the ego-anchored children, their use of I was nevertheless functionally restricted to a context similar to that reported for the ego-anchored children.

It is interesting to note that although the caregivers did not index their own actions with self reference forms, they were not sitting passively watching their children at play. Looking at the videotapes it became clear that the caregivers in fact did discursively refer to their own independent actions, but interestingly they tended to rely on a variety of other ways to do so. For instance, building their own block tower while the child looked on the caregivers said utterances like "We're building a big tower" or "Hey you're building a nice tower, aren't you". What is central here is that these caregivers have opted to discursively index their own actions either with joint reference forms or by linguistically constructing the child as agent. One possible interpretation of these findings seems to be that the mothers were actually down-playing their own agency in an attempt to index their child's budding agentivity (See Budwig, in press b for further discussion).

German Caregivers

The analyses carried out with the German caregivers' speech were similar to that described for the American caregivers. All instances of caregivers' indexing self with self reference forms were isolated and coded in terms of the semantic and pragmatic codes described above. The findings revealed quite distinct patterns of self reference when compared to the American caregivers. First, the German caregivers did not restrict the use of self reference forms to assertions about their mental states. There were numerous instances of the caregivers referring to themselves as agents of ongoing action sequences. Most frequently though when the caregivers referred to their own actions they simultaneously indexed actions that their children were carrying out or could carry out. In sum, one common way the German caregivers referred to themselves was in terms of joint agency. A second way in which the German caregivers frequently indexed their own actions was in terms of generic agency. That is, these caregivers were indexing their own actions in light of more general procedural or normative ways of doing with the term man.

Summary of Caregivers' Indexing

One way to get a sense of the distinct ways the American and German caregivers indexed self and other in the context of the play sessions is to compare instances from similar activities. One activity that took place in almost every session involved building together with the wooden blocks. Looking at such examples provides illustrations that the German and American children are getting very distinct sorts of input concerning the indexing of self and other.

(6) German child (GCH) and mother (GMO) have been building block towers.

a. GCH: Noch [/] noch einen Turm.

Another [/] another one tower.

b. GMO: Na, komm, wir machen zusammen einen ganz hohen.

Okay, come we (will) make together a very big (one).

c. GCH: (begins to build) das ...

this ...

d. GMO: Ja hier ...

Yes, here ...

e. GCH: Das nicht, nein? (showing a triangle)

This not, no?

f. GMO: Doch, das kann man auch nehmen.

Yes, one can also take that.

g. GCH: Das noch ...

This also.

h. GMO: Ja.

Yes.

i. GCH: Das noch ...

This also.

j. GMO: Ja. Jetzt kann man so was machen.

Yes. Now one can make something like this.

k. GMO: Kann man hier wider eins drauf.

Now one can again (place) one here.

(7) American child (ACH) and mother (AMO) are beginning to play with blocks.

a. ACH: My [/] my play with the # helicopter and you play w& with the blocks.

b. AMO: Okay. What should I build?

c. ACH: A [/] a tower.

d. AMO: A tower. Okay.

e. ACH: And [/] and you # and you build a bridge.

f. AMO: And a bridge too.

[pair continue to build for several turns]

g. AMO: How about +... Should we build a little # something for the helicopter?

h. ACH: no (whining) This not working (trying to adjust two blocks).

i. AMO: Where would you like to build it? # Can you get it? That's it. Okay. # You gonna make a big tall house?

j. ACH: No I want to. (implied negative)

k. AMO: What would you like to make?

l. ACH: I want +... (searching through blocks), I like +... # My like that (selecting a block).

In comparing examples (6) and (7), one finds illustrations of the sort of differences in indexing noted above. Example 6, stemming from one of the German caregiver-child dyads, highlights the joint and generic focus of such episodes. Note for instance in line 6 (b) that after the child proposes to build a block tower, the mother weaves this into a joint activity. And as the two begin building the dyad continually checks in with one another. Note as well the use of several generic agency markers by the caregiver in lines 6 (f), 6 (j), and 6 (k). In examples such as this the point seems to be one of describing procedural aspects of building prior negotiated structures rather than discussing and negotiating what each partner is doing.

The generic and joint focus in example (6) contrasts with the child-centred focus of example (7). In example (7), the American child, like the German child, has proposed to build a tower. First of all notice that in this instance, the child begins the interaction by claiming the toy helicopter and telling his mother to play with the building blocks. Also note in line 7 (b) that the mother questions the child as to what she should build. The sort of child-centred focus is found throughout the example. For instance, the mother seeks permission to act in particular ways thereby granting the child control of the flow. This mother also tries to establish a joint project (see line 7 (g)) but when this is rejected she turns to focus exclusively on what the child wishes to do in lines 7 (i) and 7 (k). This example, then, illustrates several of the features noted for the American caregivers' marking of self and other: 1) their infrequent use of indexing self in the context of action verbs, and if so, only in requests for permission to act in certain ways, thereby down playing their own agency; 2) use of joint reference forms when they are acting on their own; and 3) frequent indexing of the child as independent agent.

In summary the claim made here is that even when the play activities are quite similar (i.e. building block towers), the sort of input the children are getting is quite distinct. The German and American caregivers are indexing themselves and their children in distinct ways. Of course the question remains does such distinct input play a fundamental role in the sorts of linguistic systems the children create when indexing self and other. We will return to a consideration of this question in the final section of this paper.

Summary

In this section, I have reviewed findings from research examining how children and caregivers interacting within two different language communities index self and other in their spontaneous verbalizations in everyday activities. In doing so, I have made use of a form of discourse analysis referred to as variation analysis. Variation analysis examines distributional patterns of different ways of saying the same thing. Here we have paid particular attention to the ways the caregivers and their children referred to themselves. In summarizing these analyses I have argued that subtle variations in the way speakers employ such linguistic forms is indicative of various ways speakers situate self and other in ongoing discourse.

As for the children, we found some overlap in the ways the children referred to themselves. Both the English and the German children made use of a variety of forms to situate self distinctly in ongoing play frames. What differed though between children of different ages and children growing up within different language communities was the particular sort of stances that the children took. For instance, at first, the American children primarily referred to themselves, drawing upon a variety of linguistic devices to contrast between prototypical agentivity and certain deviations. At a time before the youngest American children regularly referred to others, they used first person pronominal forms often in ways that deviated from the target language and in that sense created systems that looked "error-like". The German children began indexing self and other already by the time they began combining words, and did not reveal the pattern found in the American children's data of going beyond the system in non- adult-like ways. That is, they made use of conventional resources to mark the sort of distinctions about self and other they found relevant. The German children, in contrast to the ego-anchored American children, were more interested in contrasting between their role and that of others as actors and what I have referred to as more normative or procedural stances with their early use of the impersonal construction man [one].

The caregivers from the two communities showed similarities and differences when comparisons were made between caregivers' indexing and child indexing, as well as when comparisons across were made across the two language communities. While the American caregivers did not produce linguistic distinctions identical to those of their children, certain similarities were noted. The American caregivers typically used the form I to index acts of assertions about their ongoing states. Even when used with action verbs, these instances appeared when requesting permission, and in general the American caregivers' indexing of self tended to down play their own agency. Other ways this was demonstrated was in their use of second person forms and joint references forms to refer to their own actions. In contrast, the German caregivers, like their children, did not tend to linguistically refer to themselves as prototypical agents. Two common ways of indexing self included indexing joint and generic stances. The German caregivers tended to relate their own actions to those of their child giving a sense of functioning as part of a team or they down played individual involvement of both themselves and their children by referring to general ways "one" does things without linguistically referring to concrete individuals.

Relating the summary of findings from the variation analyses reviewed above to our earlier discussion of language as a tool for the researcher several comments can be made. First, with regard to the children, the sorts of categories suggested to be drawn upon by the children go beyond findings of previous developmental research of the sort reviewed above. As I have argued elsewhere, variation analysis provides a valuable tool for researchers interested in children's understanding of various aspects of self and other in relation to human activity and could well provide developmental psychologists with added ways to focus on language data in studying self development (see Budwig, in press a; Budwig and Bamberg, in press for further discussion).

Examining the findings of the variation analyses from the standpoint of viewing language as mechanism in self construction we also find some promising beginnings. As was noted above, prior research has made the bold claim that language plays a central role in the child's construction of self, but we know little about this process. Drawing upon variation analysis and combining it with Gumperz's (1982) notion of contextualization cues - surface devices that play a central role in negotiating meaning in social contexts - could provide a beginning account of how children might make use of the input around them to construct ways of interpreting and producing their ongoing notions of self. Viewing the indexing of self and other as contextualization cues suggests ways speakers signal to one another how the flow of interaction is to be understood.

In sum, variation analysis is attractive for several reasons. First, it allows for the simultaneous analysis of linguistic structure and function. In addition, variation analysis provides not only a tool for the researcher, but also a way of beginning to understand how the child might make use of language in constructing self. While there are many valuable aspects to variation analysis, it is not without limitations. It should be clear that the sorts of analyses provided above tell us little about why caregivers and their children index self and other as they do. Several possibilities remain: for instance, it could be that differences in the typological options of German and English provide speakers with distinct options for linguistically indexing self and other or cultural differences in beliefs between speakers from Germany and America could differ in ways that influence speakers' indexing of self and other. The analyses provided above do not provide answers to such questions. In the final section of this paper, I will turn to a discussion of ongoing research which has been designed to address such issues.

Conclusions

At the outset of our discussion I examined the ways researchers adopting a social constructivist approach and developmental approach have viewed language and its role in the process of constructing personhood. In doing so I identified three issues that need to be addressed, namely 1) contrasting ways social constructivists and developmental researchers have viewed language as mechanism; 2) integrating the notion of development into the consideration of the role of language in self construction; and 3) how to take into account the active role of the child in accounts which rely on language as mechanism. I then went on to provide an illustration of one way in which language could be viewed both as a tool for the researcher and mechanism for the child based on an approach to discourse called variation analysis. In concluding, I would like to reflect on the three issues originally outlined from the point of view of the variation analyses I provided. I then will conclude this paper with a discussion of the need to tease apart typological possibility and discourse communities in future studies of children's construction of self.

Language as Method, Language as Mechanism Revisited

Contrasting Ways Researchers have Viewed Language as Mechanism

In reviewing various ways social constructivists and developmental researchers have viewed language as mechanism in the construction of self I noted above the extent to which language has been defined differently by various researchers at different points in time. Some have suggested that children construct different senses of self because language(s) provide structural options that become habitual categorizations of how to view self and other. In contrast to this sort of typological perspective has been a view that speakers draw upon language in providing distinct ways to situate self and other in terms of the functioning of language. This more social account is viewed in contrast to the grammatical focus of other research. One of the advantages of variation analysis as I have made use of it is that it allows the researcher to look simultaneously at both levels of language functioning. Drawing upon the work of Werner & Kaplan (1963/84, see also Berman & Slobin, 1994; Budwig, in press a) one can suggest that children attend to both levels to the extent that acquisition of new structures leads children to consider new functions, and the development of new functions leads children to seek new forms. It seems important therefore that any approach to the construction of self draws upon multiple and related levels of language functioning.

Integrating the Notion of Development into the Consideration of the Role of Language in Self Construction

A second theme that emerged in the discussion of social constructivist and developmental research provided at the outset of this article is that little work has explored how language might play a role over developmental time. What has yet to be unravelled is the role of distinct typologies and the impact of growing up in communities where language functioning differs. While such research has highlighted that children growing up speaking different languages are provided with distinct linguistic resources, we know little about how differences in such options play a role in real time.

The importance of this issue becomes clear when considering some of the findings from the variation analyses discussed above. In presenting the research concerning young American's indexing of self and other I noted two groups of children who drew upon the linguistic options offered by English in distinct ways. What is central for the present discussion are findings from the examinations of those children's caregivers' indexing of self and other. Although our analyses revealed almost identical input to the children in terms of how their caregivers referred to the children, at the different points in developmental time the children made distinct usage of such input. Thus, if we are to understand how it is language provides a mechanism for constructing self, it seems imperative to incorporate longitudinal analyses that more closely examine the relationship between the kinds of indexing children hear from their interactive partners and the different ways children make use of such indexing over developmental time.

The Active Role of the Child

As I have noted above, there is a sort of tension within the developmental research reviewed above that examines language as method and language as mechanism. It seems that those who adopted the language as method approach have highlighted the active and creative role that children play in constructing experience. Such research, though limited in its treatment of language to the extent that language is viewed as an added on feature to prior established conceptual categories, suggests that children's constructions of self may differ in important ways from those of their caregivers. Although those adopting the language as mechanism approach have suggested the importance of viewing children as active and creative participants, most all of the research carried out within such a framework has focused almost exclusively on a sort of unidirectional position of culture impacting via language on the child.

Reflecting on this tension in light of the variation analyses reviewed above, it becomes clear that children bring to their ongoing interactions notions of what language is and how it functions, and as the variation analyses point out, both the German and American children reinterpreted the relationship between language structure and language function in ways that made sense to them. The children acquiring English for instance did not simply assimilate the form-function patterns found in caregivers' indexing of self and other, but created totally novel meaning systems to mark notions such as prototypical agency (i.e. My build the tower). The comparisons I have presented between children growing up in different language communities highlights the need to better understand how children make use of discourse available to them. The one point that becomes clear is that children do not integrate discourse around them in any straightforward and passive way. What is needed is a better understanding of the similarities and variations to be found in children's indexing when examined in light of variations found in the systems they hear.

In sum, the review of prior theorizing and research on the role language plays in constructing self, as well as the summaries of the variation analyses I have conducted concerning language comparisons of children's and caregivers' indexing of self and other suggest several issues in need of further examination. We will turn now in the final section of this article to consider some ongoing work I am conducting to address another issue in need of further research.

Teasing Apart Typological Possibility and Discourse Community: Future Avenues of Research

In our review of the literature, we have noted that most frequently discussions of how language plays a role in self construction either focus on language structure or language function. Those focusing on language structure have suggested that languages provide distinct ways for speakers to situate themselves and others and that such distinctions could lead speakers of different languages to construct distinct selves (see for instance, Kerby, 1991; Lock, 1981). The implicit assumption behind this is that speakers of the same language, given their exposure to similar typologies, should be more similar to one another than speakers of languages that vary. In contrast, those who have focused on language functioning have not assumed the sort of within language similarity in that it has been suggested that language might function differently within distinct communities and settings (see for instance, Heath, 1983; Miller et al., 1990; Schieffelin & Ochs, 1986).

The variation analyses reviewed above pointed up the obvious but often forgotten point that what matters more to children's constructions is the language they hear, rather than what is grammatically possible. This became clear with regard to English to the extent that though grammatically possible in English to refer to self as agent, the caregivers studied, for various reasons, tended to index self by downplaying their own agency. In a related way, German seemed interesting to study because typologically it provided children with exposure to pronominal case marking that shared similarities with the novel systems constructed by the English-speaking children, though as was noted above, such pronominal contrasts were absent in the interactive contexts we studied. First of all, the caregivers rarely indexed self and other with pronominal forms, and second, they made use of constructions (such as the use of man noted above) that were not anticipated. Thus, while the options provided by distinct languages seem central to an understanding the role language might play, it becomes apparent that in working with the notion of language typology, research must be grounded in actual face to face verbalizations rather than assumptions about grammatical possibilities.

A second and related issue that becomes clear from the sort of comparisons made between German and English above is that we know very little about the causes of such differences. That is, why did the caregivers and children in the two languages look so different in the ways they indexed self and other? At this point answers must remain speculative because such work conflates two important sources of variation. First, the speakers had access to distinct typological systems providing them with different options for indexing self and other. Second, the German and American caregivers might have indexed self and other differently due to non-linguistic reasons, perhaps due to different communicative goals about the grounding of persons in discourse based on distinct cultural models. What this suggests is the need to tease apart crosslinguistic and cross-cultural variables. That is, we need to simultaneously study indexing of self and other both within and between language groups.

Along with others, I have begun to investigate within and between language variation in indexing of self and other in caregiver-child interaction, by focusing on discourse communities for which I imagine there to be differences in how adults construct notions of self and other (see Budwig & Wiley, 1995; Chaudhary & Budwig, 1995) even when the participants speak the same language. First analyses of caregivers' indexing reveal that not only do caregivers differ across language communities, but also children exposed to the same language are provided with distinct ways of indexing self and other. For instance, Lin (1993) reports on two groups of middle-class American caregivers who differed both in their beliefs about children in an interview, as well as their organization of conversational flow when interacting with their toddlers. In addition, as the following two examples stemming from these caregivers attempts to regulate their children reveal, the children are exposed to distinct input about indexing self and other:

(8) [Child (HCH) and Mother (HMO) are playing with a large box filled with dried rice and manipulative toys]

a. HCH: I wan(t) go in there. I go.

b. HMO: Um, I don't think ... I don't think that we can go in the box. Let me tell you. I think that kids that play with this just have to sit outside and play with ... I don't think they want you to go inside. Can we still have fun this way? Let's still have fun this way.

(9) [Child and Mother (CMO) are playing with same large box filled with dried rice and manipulative toys]

a. CMO: You gotta keep it in the bucket, alright? You gotta keep it inside okay? Keep it in there and you can play with the rice okay? You can't get it all over the floor because then you'll get buggies. Okay? Help mommy pick it up? Here put it in there. That's a good girl.

Note for instance that the mother in example (8) reveals several of the features of the American caregivers reviewed above. She tends to refer to herself with mental state verbs. And she shows several ways of down playing her own attempts to regulate her child's actions. In contrast, the mother in example (9) who has reported a less child-centred focus in the interview portion of the study uses quite a different style in trying to regulate her child. In contrast to the indirect manner of the mother in example (8), the mother in example (9) both uses the form "you" to tell her child what to do and several commands with null forms. It becomes clear from a reading of this and other examples that the two children are exposed to quite different ways of indexing self and other in discourse despite the fact that they both are learning the same language. While differences can be found in the ways the two groups of caregivers index self and child, we do not yet know whether and how such differences impact on the children's constructions of themselves. Only by teasing apart typological possibilities from discursive communities can we begin to better understand how language could play a role in children's construction of self.

In sum, I have argued that while there is a growing body of literature supporting the view that language plays an important role in children's construction of self, we actually know little about how this process takes place. While offering this as a critique, at the same time it appears that this can be viewed more positively, as an exciting shift. With enough of a critical mass in agreement that language is central to the construction of self, I have suggested the importance of moving on to better understand aspects of this process. Three aspects that I believe warrant future attention include 1) the simultaneous examination of language structure and language function over developmental time; 2) increased focus on ways language input impacts at various points in time; and 3) integrating the notion of the active role of the child in our understanding of how language plays a role in the ongoing construction of person talk.

References

Berman, R. & Slobin, D. (1994). Different ways of relating events in narrative: A crosslinguistic developmental study. Hillsdale, NJ: Lawrence Erlbaum Associates.

Bruner, J. (1990). Acts of meaning. Cambridge, MA.: Harvard University Press.

Budwig, N. (1989). The linguistic marking of agentivity and control in child language. Journal of Child Language, 16, 263-284.

Budwig, N. (1990). The linguistic marking of non-prototypical agency: An exploration into children's use of passives. Linguistics, 28, 1221-1252.

Budwig, N. (in press a). A developmental-functionalist approach to child language. Hillsdale, NJ: Lawrence Erlbaum Associates.

Budwig, N. (in press b). What influences children's patterning of forms and functions in early child language? In D. Slobin, J. Gerhardt, A. Kyratzis, & J. Guo (Eds.), Social interaction, social context, and language: Essays in honor of Susan Ervin-Tripp. Hillsdale, NJ: Lawrence Erlbaum Associates.

Budwig, N. & Bamberg, M. (in press). Language and its role in understanding intentional relations: Research tool or mechanism of development? Behavioral and Brain Sciences.

Budwig, N. & Wiley, A. (1995). Maternal input and children's special use of pronominal forms. (under review).

Chaudhary, N. & Budwig, N. (1995). Indexing child, self and others in caregivers' speech to young children in 10 Indian families. (under review).

Damon, W. & Hart, D. (1988). Self-understanding in childhood and adolescence. Cambridge: Cambridge University Press.

Davies, B. & Harré, R. (1990). Positioning: The discursive production of selves. Journal for the Theory of Social Behaviour, 20, 43-63. Deutsch, W. & Budwig, N. (1983). Form and function in the development of possessives. Papers and Reports on Child Language Development, 29, 36-42.

Ervin-Tripp, S. & Gordon, D. (1986). The development of requests. In R. Schiefelbusch (Ed.), Communicative competence: Assessment and intervention (pp. 61-95). Baltimore, MD: University Park Press.

Gee, J. & Savasir, I. (1985). On the use of "will" and "gonna": Towards a description of activity-types for child language. Discourse Processes, 8, 143-175.

Gumperz, J. (1982). Discourse strategies. Cambridge: Cambridge University Press.

Harré, R. & Gillett, G. (1994). The Discursive Mind. Thousand Oaks, CA.: Sage.

Harris, R. (1982). The language makers. London: Duckworth.

Heath, S. (1983). Ways with words: Language, life, and work in communities and classrooms. Cambridge: Cambridge University Press.

Kerby, A. (1991). Narrative and the self. Bloomington, IN: Indiana University Press.

Kolodziej, P., Deutsch, W. & Bittner, C. (1991). Das Selbst im Spiegel der Kindersprache (The self in the mirror of child language). Zeitschrift fur Entwicklungspsychologie und Paedagogische Psychologie, 23(1), 23-47.

Lewis, M. & Brooks-Gunn, J. (1979). Social cognition and the acquisition of self. New York: Plenum.

Lin, A. (1993). The child as conversational partner: The creation of participation roles as cultural activity. Unpublished doctoral dissertation, Clark University, Worcester, MA.

Lock, A. (1980). The guided reinvention of language. London: Academic Press.

MacWhinney, B. (1991). The CHILDES project: Tools for analyzing talk. Hillsdale, NJ: Lawrence Erlbaum Associates.

Mead, G. (1934). Mind, self, and society. Chicago: Chicago University Press.

Miller, P., Potts, R., Fung, H., Hoogstra, L., Mintz, J. (1990). Narrative practices and the social construction of self in childhood. American Ethnologist, 17, 292-311.

Morgan, M. (1994). Discourse, subjectivity (and bodies). Unpublished manuscript, Department of Psychology, Massey University, Palmerston North, New Zealand. Mülhäusler, P. & Harré, R. (1990). Pronouns and people. Oxford: Blackwell. Ochs, E. & Schieffelin, B. (1984). Language acquisition and socialization: Three developmental stories and their implications. In R. Shweder & R. Levine (Eds.). Culture theory: Essays on mind, self, and emotion (pp. 276-320). Cambridge: Cambridge University Press.

Otto, I. (1994). Linguistischer Egozentrismus bei deutschen und englischen Kindern. Staatsarbeit, Universität Keil, Germany.

Potter, J. & Wetherell, M. (1987). Discourse and social psychology. London: Sage.

Quick, H. (1991). Maternal input and children's use of self reference forms: A functional approach. Unpublished manuscript, Clark University, Worcester, MA.

Schieffelin, B. & Ochs, E. (Eds.). (1986). Language socialization across cultures. Cambridge: Cambridge University Press.

Schiffrin, D. (1994). Approaches to discourse. Oxford: Blackwell.

Shatz, M., Wellman, H., & Silber, S. (1983). The acquisition of mental verbs: A systematic investigation of first references to mental state. Cognition, 14, 301-321.

Shotter, J. (1989). Social accountability and the social construction of 'you'. In J. Shotter & K. Gergen (Eds.), Texts of identity. London: Sage Publications.

Shotter, J. & Gergen, K. (1989). (Eds.). Texts of identity. London: Sage Publications.

Slobin, D. (1985). Crosslinguistic evidence for the Language- Making Capacity. In D. Slobin (Ed.), The crosslinguistic study of language acquisition (Vol. 1, pp.1157-1256). Hillsdale, NJ: Lawrence Erlbaum Associates. Taylor, C. (1985). Philosophy and the human sciences: Philosophical papers. New York: Cambridge University Press.

Vygotsky, L. (1978). Mind in society: The development of higher psychological processes. M. Cole, V. John- Steiner, S. Scribner, & E. Souberman (Eds.). Cambridge, MA: Harvard University Press. Wellman, H. (1990). The child's theory of mind. Cambridge, MA: The MIT Press.

Werner, H. & Kaplan, B. (1963/84). Symbol formation. Hillsdale, NJ: Lawrence Erlbaum Press.

Wertsch, J. (1991). Voices of the mind: A sociocultural approach to mediated action. Cambridge, MA: Harvard University Press.


home Return to Virtual Faculty Discussion Page