Please include the following for full credit:? ? 3 key points from the article ? ? 2 quotes that speak to you with an analysis

Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=hlld20

Language Learning and Development

ISSN: 1547-5441 (Print) 1547-3341 (Online) Journal homepage: https://www.tandfonline.com/loi/hlld20

Conversation and Language Acquisition: A Pragmatic Approach

Eve V. Clark

To cite this article: Eve V. Clark (2018) Conversation and Language Acquisition: A Pragmatic Approach, Language Learning and Development, 14:3, 170-185, DOI: 10.1080/15475441.2017.1340843

To link to this article: https://doi.org/10.1080/15475441.2017.1340843

Published online: 06 Oct 2017.

Submit your article to this journal

Article views: 1740

View related articles

View Crossmark data

Citing articles: 3 View citing articles

Conversation and Language Acquisition: A Pragmatic Approach Eve V. Clark

Department of Linguistics, Stanford University

ABSTRACT Children acquire language in conversation. This is where they are exposed to the community language by more expert speakers. This exposure is effectively governed by adult reliance on pragmatic principles in conversa- tion: Cooperation, Conventionality, and Contrast. All three play a central role in speakers’ use of language for communication in conversation. Exposure to language alone, however, is not enough for learning. Children need to practice what they hear, and take account of feedback on their usage. Research shows that adults offer feedback with considerable frequency when young children make errors, whether in pronunciation (phonology), in word-from (morphology), in word choice (lexicon), or in constructions (syntax). Adults also offer children new words for objects, actions, and relations. And, along with new labels for such categories, they also provide supplementary information about the referents of new words—information about parts, properties, characteristic sounds, motion, and function, as well as about related neighboring objects, actions, and relations. All this helps children build up and organize semantic domains as they learn more words and more language.

Conversation provides the major setting for learning a first language. It is where children find out how to use the forms of language they identify in the speech stream. As Roger Brown presciently noted (1968, p. 288):

“The changes produced in sentences as they move between persons in discourse may be the richest data for the discovery of grammar”

But identifying the forms oflanguageis onlythe firststep. AsJerome Bruner pointed out (1983, p.119), they must also learn how to use language:

“[W]hether human beings are lightly or heavily armored with innate capacities for lexicon-grammatical language, they still have to learn how to use language. That cannot be learned in vitro. The only way language use can be learned is by using it communicatively”

In this article, I review some critical findings on how children learn language through interaction in conversation.

Pragmatic principles in conversation

To use language communicatively, children rely on several pragmatic principles fundamental to adult language use, namely the Cooperative Principle, and the Principles of Conventionality and Contrast (see Clark, 1987, 1990; Grice, 1989). These principles underlie adult usage, and are central to communicative uses of language. They are important to the process of acquisition precisely because children acquire language within conversational settings.

CONTACT Eve V. Clark [email protected] Department of Linguistics, Stanford University, Margaret Jacks Hall, Building 460, Stanford, CA 94305-2150. Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/hlld. © 2018 Taylor & Francis Group, LLC

LANGUAGE LEARNING AND DEVELOPMENT 2018, VOL. 14, NO. 3, 170–185 https://doi.org/10.1080/15475441.2017.1340843

The Cooperative principle captures the fact that the goal speakers have when they speak guides their choices of how to say what they want to say, and what they expect their addressee to understand on each occasion. This principle can be characterized as follows:

“Make your contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.”

To be cooperative, adult speakers can generally be said to rely on the following maxims that characterize the Cooperative Principles (Grice, 1989) as they plan and make their contributions to a conversational exchange, maxims that are also in play of course for addressees:

(a) Quality: be truthful (don’t lie or make claims that are unwarranted) (b) Quantity: be informative (say what’s needed and no more) (c) Relevance: be relevant (stay on topic) (d) Manner: be brief and orderly (avoid obscurity and ambiguity)

These maxims each help in the planning of successive utterances, and in the smooth accumula- tion of common ground with addressees. This common ground consists of the information shared by all the participants in a conversation, and is contributed to, potentially, with each turn in an exchange. To make use of, and to add to, common ground, each speaker must keep track of what the others do and don’t know. Each participant can add to common ground with any contribution of new information. By adding new information, speakers add to common ground and, in doing this, accumulate further common ground during a conversational exchange. And addressees in turn ground new information by acknowledging it (see further H. Clark, 1996; E. Clark, 2001, 2015).

The Cooperative Principle itself depends on both Conventionality and Contrast. These principles capture some general assumptions that we all rely in communicating with language. Conventionality captures the fact that:

–“For certain meanings, speakers assume that there is a conventional form that should be used in the language community”

Conventionality captures the consistency within a speech community that offers speakers reliability in communication from one speaker to the next, and from one occasion to the next. For instance, a table is conventionally called a table in English, and an oak tree is called an oak, regardless of who in the community is talking. These are the conventional terms used to convey these particular mean- ings. Consistency over time in the use of such conventions within a community of speakers also allows for transmission across generations. Exposure to consistent usage of conventional terms allows children to learn those conventions and make use of them themselves.

Speakers rely on the conventional options in a language in order to be cooperative: if they don’t use the conventional terms and constructions expected in the community, other members of the community are unable to understand them.

The second principle here, the Principle of Contrast, depends on and complements Conventionality:

–“Speakers assume that any difference in form signals a difference in meaning”

Differences in form in the lexicon allow for extensive networks of both subtle and gross distinctions in meanings within a language. This holds both for distinctions introduced, say, among words marked for number, case, or gender, as well as by subtle diferences in distribution patterns for such near-synonyms as big and large among dimensional terms in English.

When speakers don’t use the expected conventional form, their addressees must then draw on any available information from the physical and conversational context in computing what the speaker’s intended meaning might actually be on that occasion (Clark & Clark, 1979). Conventionality and contrast work hand-in-hand in both adult language use and language acquisition. In acquisition, for

LANGUAGE LEARNING AND DEVELOPMENT 171

example, children early on act as if any difference in word form that they detect signals a “new meaning” for them to attend to and add to their vocabulary (see Clark, 1990, 1993). Notice that the Cooperative Principle works only because it relies on speakers’ use of conventions in language, and because terms and expressions within a language differ––that is, they contrast—in meaning.

These pragmatic principles for managing contributions to a conversational exchange demand a lot: Speakers must attend to and track what addressees do and don’t know, judge how appropriate their contributions are for the topic at hand, and tailor them to fit what the current addressee knows in the ongoing exchange. While children treat language as communicative from the start, they are not always good at making use of the maxims for Cooperation. They need experience in assessing what others already know, hence what is (or isn’t) relevant, and how much detail to provide in describing complex events (see, e.g., Clark & Kurumada, 2013; Harris, 1999; McTear, 1985; Siegal, 1997). Becoming more expert at conversation takes extensive practice in using language.

Children recognize early on that there are conventional words for objects and actions, and they ask for these, at first with gestures (e.g., Kelly, 2014; Kishimoto, Shizawa, Yasuda, Hinobayashi, & Minami, 2007; Olson & Masur, 2011). They elicit words from adults, and will reject offers of “wrong” words as young as age one, for example when an adult deliberately mislabels a shoe as an “apple” (Koenig & Echols, 2003). Such rejections show that even very young children already depend on conventionality. They also rely critically on contrast, assuming that the words they already know differ in meaning from unfamiliar, new, words (Clark, 1997, 2016).

Exposure to language in conversation

To acquire language, children have to be exposed to it. How much adult speech is addressed to children over the course of an hour, a day, a year? In a longitudinal study of families from different social classes in the U.S., Hart and Risley (1995) documented the amount of speech directed to children in terms of utterances and words per hour, as shown in Table 1.

They tracked families with daylong recordings at regular intervals for 2½ years, and found considerable consistency within families and within social class over time (r = .84). When such rates are extrapolated to a week, a year, and 4 years, the numbers accumulate, as shown in Table 2. These utterance and word counts, along with the extrapolations in Table 2, are based directly on the language spoken with the children; they exclude all speech between adults. The numbers here are also independent of how many adults there were in each household.

By tracking the amount of adult speech with the child in each family, Hart and Risley focused on the language most likely to offer children direct exposure to and experience of language use in everyday exchanges. Their estimates of how much language children experience in this way during their first 3 years of talking (from ages 1–4) are consistent with other studies that have estimated the amount of experience children have early on with language (e.g., Van de Weijer, 1998; for Dutch; Fernald, Perfors, & Marchman, 2006; and Weisleder & Fernald, 2013; for English; Hurtado,

Table 1. Amount of speech addressed to children by SES (based on Hart & Risley, 1995)

Social class Utterances/hour Words/hour

Professional 487 2,153 Middle/Lower 301 1,215 Welfare 178 616

Table 2. Exposure to language over time (based on Hart & Risley, 1995)

Social class One week One year Four years

Professional 215,000 11 million 44 million Mid/Lower SES 125,000 6 million 24 million Welfare 62,000 3 million 12 million

172 E. V. CLARK

Marchman, & Fernald, 2008; and Fernald, Marchman, & Hurtado, 2008; for Spanish). The amount of exposure children receive in the first few years has direct effects on their later development (see, e.g., Walker, Greenwood, Hart, & Carta, 1994; Hoff, 2003a; 2003b, 2010; Marchman & Fernald, 2008; also Fernald & Weisleder, 2015). Moreover, the more exposure children have to language in their first 3–4 years in the U.S., the better they do when they enter school.

Adults don’t talk with children in the same way that they talk with other adults. They cannot make the same assumptions when their interlocutors know very little, know few or no words, and have relatively little grasp of how to use the little language they have. Adults adjust their speech to accommodate to their children’s level: they make use of shorter utterances with pauses in between, they repeat their utterances, often several times, and they make use of distinction intonation and pitch patterns, designed in part to capture children’s attention (see, e.g., Broen, 1972; Gallaway & Richards, 1994; Snow & Ferguson, 1977). This accommodation to their children presumably helps children in processing what they hear and in making sense of it in context.

To engage children in conversation from the start, adults rely on joint attention, physical co- presence of the relevant objects and events, and on conversational co-presence with their use of familiar words and phrases. These three factors, of course, are central to any communicative exchanges between adults too. With children, though, they may take a bit more work to set up. Adults manage joint attention, for example, by directing one- and two-year-olds to attend to the object or event being talked about, with frequent reminders to maintain attention on the target. They rely in this on both speech (with attention-getters like Hey, Look, and See this?, also the child’s name) and gesture (with pointing, reaching towards, holding out, and showing the object of interest) as they encourage young children to attend (e.g., Deák, Walden, Kaiser, & Lewis, 2008; Estigarribia & Clark, 2007; Shimpi & Huttenlocher, 2007). This reliance by adults on joint attention with their children is not restricted to Western cultures: it is evident in adult-child interactions in Nigeria and in Mozambique (Childers, Vaughan, & Burquest, 2007; Mastin & Vogt, 2017), as well as elsewhere. Adults expect very young children to attend to the ongoing activity and talk, and adults manage their attention so they do so (see Chavajay & Rogoff, 1999).

In Western cultures, once children are attending, adults often elicit words from them, talk about what is happening, offer labels for unfamiliar objects and actions, and co-construct conversations in which they share utterances with their young children, at first by framing or scaffolding what the child can say, as in (1) and (2):

(1) Meredith (1;6, wanting to talk about a visit to the doctor): band-aid. Mother: Who gave you the band-aid? Meredith: nurse. Mother: Where did she put it? Meredith: arm. (Snow, 1978)

(2) D (1;6.11, being encouraged to tell Father about episode where Philip, aged 10, let out his budgerigar and it landed on D’s head) Mother: Did you see Philip’s bird? Can you tell Herb? D: head. head. head. Mother: What landed on your head? D: bird. [Clark, diary data]

In each case, the parent who knows about the relevant episode can provide a framing into which the child can insert single word utterances, and so “tell the story”. To do this, of course, the parent must share common ground with the child about the relevant event, or else such framing simply isn’t possible. This can be seen in what happened when Meredith tried to talk about her band-aid to a comparative stranger, just before the episode cited earlier in (1) (Snow, 1978).

(3) Meredith (1;6, talking to an unfamiliar adult): band-aid. Adult: Where’s your band-aid? Meredith: band-aid. Adult: Did you have a band-aid?

LANGUAGE LEARNING AND DEVELOPMENT 173

Meredith: band-aid. Adult: Did you fall down and hurt yourself?

Because this adult didn’t know about the relevant episode, the exchange initiated by the child went nowhere.

The co-construction of utterances that results from such framing helps children move from single-word utterances to sequences of single words, as in (4):

(4) Allison (1;6.21, after Mo suggests taking her coat off) Allison (points to her neck): up. up. Mother: What? Allison: neck. up. Mother: What do you want? What? Allison (points to zip and lifts up her chin): zip. zip. up. (Ochs, Schieffelin, & Platt, 1979)

And later on, to slightly more elaborate sequences as precursors to early constructions, much as in (5):

(5) G (2;7, pulls apart a large doll; small doll falls out): that! (holds up the small doll): that baby hide. again. (Ochs et al., 1979)

As children advance beyond their initial reliance on single words, they elaborate what they say by adding articles and demonstratives to nouns, as well as inflections to mark number, gender, and case, and, to their verbs, they add direct objects, locatives, and subjects, as they begin to build up verb paradigms and mark distinctions of person, number, and tense as well (see, e.g., Ochs et al., 1979; Rojas Nieto, 2011; Scollon, 1976; Veneziano & Clark, 2016; Veneziano, Sinclair, & Berthoud, 1990).

In achieving joint attention, adults rely on gaze, gesture, and language (Tomasello, 1995). For their part, infants attend to and track adult gaze from as young as 3 months. As they get older, they make more use of gaze in interaction (e.g., Brooks & Meltzoff, 2008; Johnson, Slaughter, & Carey, 1998; Slaughter & McConnell, 2003; Triesch, Teuscher, Deák, & Carlson, 2006). Gesture also plays a role in attracting and holding attention. Adults point at what they are talking about, whether whole objects or some part of an object (Clark & Estigarribia, 2011); and children may even give priority to an adult point over any words used when they are trying to identify the referent of an adult utterance (e.g., Grassmann & Tomasello, 2010). Gestures involve motion, and movement is a powerful attractor of attention at all ages.

To identify the referent of an adult’s referring expression may require a lot of work on the part of young children, especially if some of the words used are unfamiliar. How do young children manage? They make inferences from adult gaze and adult gesture in context, in order to locate any candidate referent that is in joint attention and physically present. They can add to this any information conveyed by words they already know. As they acquire more words and meanings, they can make more use of what the adult says. That is, children rely on the Cooperative Principle. As they make inferences about the speaker’s intended meaning on each occasion, they make use of anything else they know about candidate-referents in each setting. Another factor here is timing: if parents offer new labels when they can see that the one-year-old is attending to the target object or action, children should find it easier to assign a preliminary meaning for a new word-form in context (e.g., Axelson, Churchley, & Horst 2012; Yurovsky, Smith, & Yu, 2013).

Conversational co-presence plays a more prominent role as children get older and control both a larger vocabulary and more syntactic constructions. As they acquire more language, they can make more use of what adults are saying in order to identify the intended referent(s) on each occasion. The more words they already know, the more use children can make of conversational co-presence. This also makes them less dependent on the here-and-now in trying to understand adult speech.

Adults take up the topics children introduce; they provide scaffolding or framing (dependent on common ground with respect to the event involved) for them at the one-word stage; and they follow up on what their children want to talk about. By age 2;6, young children initiate at least half the

174 E. V. CLARK

conversational exchanges they have with their parents and others (e.g., Bloom, Margulis, & Tinker, 1996). Adults make use of vocabulary their children know, and introduce them to new words, linking these to words they already know. Along with such new words, adults often supply added information about their referents, and so further specify the inferences they license about the meanings involved (see Clark, 1998, 2001; Clark & Estigarribia, 2011; Clark & Wong, 2002). This is apparent in such exchanges as (6) and (7):

(6) Naomi (2;7.16): what is it? Father: Those are cobblestones. That’s a street made out of stones. [Sachs corpus/CHILDES]

(7) Child (2;11, looking at a book with mother) Mother: I don’t know if you know what that one is. Child: that’s a snake. Mother: It looks like a snake, doesn’t it? It’s called an eel. It’s like a snake only it lives in the water. (Gelman, Coley, Rosengren, Hartman, & Pappas, 1998)

Conversations about what is happening, what they are doing together, what they are playing at, and what they are reading together, are a primary source of information for children about forms and meanings in the language they are acquiring, and about how to use them.

Feedback on language use in conversation

Children make many errors, especially early on, as they start talking. Their errors are errors of both omission and commission. Since such errors can make young children hard to understand, adults often check up on what they mean: Errors of omission can make young children particularly hard to understand, for instance, when they produce one word alone, or a two-word combination, while errors of commission like the regularization of an irregular plural, as in mans for men, may be more interpretable yet still need to be checked on. And adults check up on their children’s intentions by reformulating their utterance to express what they apparently intended to say. The goal of adults and parents here is to make sure they have understood what their children intended to convey. This in turn allows adults to pursue conversations with their children without disruption.

Reformulations of erroneous child utterances are particularly important from a theoretical point of view because some researchers have claimed that children receive no negative feedback. This supposed absence of feedback, proclaimed as a belief about acquisition, was used as an argument to support the innateness of grammar, along with the belief that the language to which young children are exposed is impoverished. These positions were identified as “No negative evidence” (NNE) and “poverty of the stimulus” (PoS), and they elicited considerable debate (see further Pullum & Scholtz, 2002; Scholtz & Pullum, 2002). However, the tide here has turned as researchers examined the actual interactions between adult and child, studied the content of successive turns, and considered what might be informative for children acquiring a first language.

How much feedback do young children receive when they make an error? In an analysis of nearly 8,000 errors from 5 children, 2 acquiring French and three acquiring English, all recorded long- itudinally, Chouinard and Clark (2003) found that adult reformulations consistently offered children conventional versions of the target utterances immediately after the child error of omission or commission. These reformulations generally (70% of the time) took the form of a side sequence with rising intonation, thereby indicating that the adult was checking up on the child’s intention and needed confirmation (see Jefferson, 1972; Norrick, 1991), as in (8):

(8) Abe (2;6.4): Milk. Milk. ||Father: You want milk? ||Abe: Uh-huh. Father: Ok. Just a second and I’ll get you some. [Kuczaj Corpus/CHILDES]

LANGUAGE LEARNING AND DEVELOPMENT 175

In side-sequences like the one marked with || in (8), the child first makes an error of omission (as here) or of commission. Then the adult checks up on what the child intended by initiating a side sequence. The child responds to the checking-up in a third turn, affirming (occasionally rejecting) the interpretation offered, and the adult then continues with the exchange.

Adults also reformulate, particularly in following up errors of commission, with an embedded correction (about 30% of reformulations overall), as in (9):

(9) D (2;4.29, being carried): Don’t fall me downstairs! Father: Oh, I wouldn’t drop you downstairs. D: Don’t drop me downstairs. [Clark, diary data]

With embedded corrections (Jefferson, 1982), where the adult substitutes the conventional form for what the child got wrong (here, use of intransitive fall in lieu of transitive drop), in the turn right after the child’s utterance, children again often accept (or reject) the repair in their next turn, the third turn in such sequences. Overall, an important feature of reformulations is that they allow adults to check up on what children mean by presenting them with conventional versions of how to say whatever it is, without interrupting the flow of conversation.

Reformulations like these are common in middle-class Western families, but they are used less often as children learn more language, and consequently make fewer errors, as shown by the general trends over age in Figure 1. This is because what children intend when they speak becomes clearer with age to their adult interlocutors. What is important here is that reformulations follow errors right away, in the next turn. As a result, they give immediate feedback to children, by providing a conventional way to convey what the child had apparently intended.

Early on, adults reformulate between 45% and 60% of child errors, with statistically similar rates of reformulation for errors of pronunciation, inflections on words, word choice, and syntactic constructions in both English and French (see Chouinard & Clark, 2003), as shown in the summed data in Figure 2.

Adults check on errors of omission and commission with reformulations. For errors of omission, adult reformulations also disambiguate the meanings of homophones when children have omitted all relevant distinguishing information. For example, in French, adults systematically distinguish infinitives from past

Figure 1. Percentage of adult reformulations of child errors by age.

176 E. V. CLARK

participles in class-1 verbs: sauter “to jump” and sauté “jumped”, both pronounced /sote/. Adult speakers use modal verbs like pouvoir (can, be able to), vouloir (want to), and falloir (must) with infinitive verb forms to talk about actions that are anticipated and so have not yet occurred, as in Il peut sauter “he can jump”. But they use auxiliary verbs like avoir (have) and être (be) with past participles for talking about actions that have already happened, as in Il a sauté “he has jumped/he jumped”. When young children produce a verb form that could be either an infinitive or a past participle in French, adults generally reformulate the child form, using constructions that contain a modal verb for events that have not yet occurred, but constructions with an auxiliary verb for events that have already happened (Clark & de Marneffe, 2012). These options both contrast with adult uses of present tense verbs for ongoing (and occasionally for future) events. The pattern of adult verb form uses in such reformulations for one child is summarized in Table 3.

In short, adults rely on the Cooperative Principle as they reformulate children’s incomplete verb constructions where children have omitted the subject and modal or the subject and auxiliary verb in their utterance about an event. The reformulations adults offer in the next turn allow for a direct comparison with the child’s use of a bare (and hence ambiguous) verb form. These adult reformula- tions contain just the elements that children later make use of as they begin to build out on the left edge of bare verb forms in French (see Veneziano & Clark, 2016).

Reformulations in the form of conventional versions of what children appear to mean not only provide feedback when children make errors; they also add to the stock of linguistic forms children are exposed to, in specific contexts, to express particular intentions. Feedback supplements children’s exposure to language in an important way (see also Farrar, 1992; Saxton, Houston-Price, & Dawson, 2005; Strapp, Bleakney, Helmick, & Tonkovich, 2008).

Figure 2. Percent adult reformulations by child error-type.

Table 3. Percent …