In Proceedings from the 24th Annual BU Conference on Language Development. Cascadilla Press, Somerville, MA, 2000. Early Communication: Beyond Speech-Act Theory Anna Papafragou University of Pennsylvania 1. Introduction For the past two decades, speech-act theory has been one of the basic tools for studying pragmatics from both a theoretical and an experimental perspective.
In this paper, I want to discuss certain aspects of the theory with respect to data from early communication in children. My aim will be to show that some of the central assumptions of the speech-act model of utterance comprehension need to be rethought.In the second part of the paper, I will outline a different pragmatic approach to verbal understanding and present a preliminary application of this approach to the developmental data. Let me start with a brief reminder of the basic tenets of the speech-act approach. According to standard speech-act assumptions, understanding utterances is a matter of knowing the rules according to which the utterances have been produced. Rules for producing utterances are rules for performing speech acts (warning, advising, requesting, promising, threatening etc.
. The speech act or acts performed in uttering a sentence are in general a function of the meaning of the sentence (the literal force hypothesis – see Searle 1969): (1) a. b. c. You are going to dance. Are you going to dance? Dance. The utterances in (1) are examples of the canonical (i. e.
literal and direct) illocutionary forces of the three basic sentence types. The declarative in (1a) is used to perform an assertion, the interrogative in (1b) a request for information and the imperative in (1c) a request for action.It is often the case, however, that utterances are used to convey illocutionary forces which differ from their canonically specified ones. For instance, declaratives and interrogatives are frequently used to perform requests for action: (2) a. b. You will get up right at this minute.
Will you get up? The utterances in (2) are cases of indirect speech acts. The recovery of such speech acts is taken to proceed along broadly Gricean lines. Interestingly, speech-act theorists propose that the recovery of a class of indirect speech acts is compressed by precedent.Consider the examples: (3) a. b. c.
d. e. Can you lend me that book? Could you lend me that book? Would you be kind enough to lend me that book? Would it be too much trouble for you to lend me that book? Would you be willing to lend me that book? According to what is known as the standardisation thesis, the forms in (3) have become standardised for the (indirect) performance of the speech-act of request. Standardisation is defined as follows: A certain linguistic form (such as Can you_? ) is standardly used to perform a speech act F in a community G if and only if: i.
It is mutually believed in G that generally when a member of G utters T, his illocutionary intent is to F; and ii. Generally when a member of G utters T in a context in which it would be infelicitous to utter T with (merely) its literally determined force, his illocutionary intent is to F. (adapted from Bach & Harnish 1979: 195) It is assumed that standardisation (through repeated usage) gives rise to shortcircuited implicatures, thereby facilitating the comprehension of a great number of indirect speech acts (see also Searle’s 1975 conventionalisation thesis).The main motivation for using utterances with indirect illocutionary forces is politeness: in (3), the speaker is ‘asking without asking’ to borrow the book. Several concerns about the main claims of the speech-act approach have been expressed in the linguistic and philosophical literature. In particular, the association of sentences with literal and direct illocutionary forces, the connection between indirectness and politeness, and the way standardisation affects utterance comprehension have been the targets of much criticism (see Levinson 1983, Tsohatzidis 1994, Papafragou 2000).In exploring the implications of the speech-act approach for the study of the development of communicative skills, I will show that the same areas create problems in dealing with the acquisitional data. 1 For reasons of consistency, I will focus primarily on the speech act of requesting.
2. The developmental story: some problems 2. 1 The ‘literal force’ hypothesis The main question which speech-act theory poses for developmental research is: How do children acquire the ability to assign illocutionary forces to utterances of different surface forms?Notice that, according to the literal force hypothesis, form-force mappings draw on different resources in direct and indirect speech acts: in the former case, the illocutionary force of the utterance can be read off from sentential type, while in the latter, it depends on inferential processes which operate after the literal content of the utterance has been derived. The literal force hypothesis thus gives rise to the prediction that the processing of indirect speech acts, being a two-stage process, should present greater difficulties for the language-learning child than the processing of direct speech acts.
There is ample evidence that, in contrast to this prediction, children are capable of producing and understanding a variety of form-force mappings from early stages in development. As far as production is concerned, data from ErvinTripp (1977) and Dore (1977) show that preschoolers have at their disposal a range of both direct and indirect forms for requesting. Garvey (1975), using spontaneous speech data from dyads of children between 3;5-5;7, found that indirect requests were both performed and understood by younger subjects (although indirectness increased by age).Read & Cherry (1978) elicited directives from children aged 2;6, 3;6 and 4;6; they report that the younger population had as many means of expressing requests as the older groups. From the point of view of comprehension, there is observational and anecdotal evidence (Dore 1977, Ervin-Tripp 1977) which suggests that even 3year-olds can appropriately respond to indirect requests such as hints (I’m cold, It’s noisy in here). In another study, Shatz (1978) reports that 2-year-olds respond as appropriately to action requests phrased as question directives as to those phrased as imperatives. More robust evidence comes from experimental data such as Carrell’s (1981), who reports that a wide variety of requests was understood by children aged 4 to 7 years.
Reeder’s (1980) work demonstrates that children at 2;6 can choose different paraphrases of an utterance such as (4) depending on whether the context encourages a request or an offer interpretation as in (5): (4) Would you like to play on the train? (5) a. b. I want you to play on the train. (request) I’ll let you play on the train. (offer) Particularly compelling support for the early understanding of indirect equests comes from attempts by children in Garvey’s study to justify their noncompliance with a request in terms of inability to perform the action, lack of willingness, lack of obligation to comply etc. ; these responses show that children have grasped the conditions under which requests are legitimately made: (6) S: Here, do that. Do the rest of that, okay? H: I can’t. (7) S: Wait for the snake to come.
H: I don’t need the snake. It is, of course, generally true that imperative directives emerge earlier and are used more frequently in child speech compared to, e. g. question directives or hints; they also give rise to fewer misunderstandings on the part of the child. This, however, is no corroboration of the literal force hypothesis, but can be easily accommodated by any account which recognises that certain form-force mappings involve greater inferential complexity than others (see section 4 below).
2. 2 Routinisation and short-circuiting Several researchers assume that children base their initial understanding and production of speech acts on pairings of linguistic strings and the typical/ recurrent interpretations such strings achieve in context.Such ‘situated’ interpretations, it is claimed, come about through ‘social routinisation’ and may be the precursors of illocutionary standardisation. Here are some examples from the speech of 2-year-olds (from Newcombe & Zaslow 1981): (8) a. b. (9) a. b. Where’s raisins? Where’s a stick? I can’t close him.
I can’t do myself that white. There are three reasons to suspect that such routinisation, if it exists, is much more limited than it is usually thought. First, typical string-interpretation pairings themselves presuppose a degree of non-trivial inferencing.In order to arrive at such pairings, the child needs to extract from the situation the relevant dimensions for the interpretation of the utterance, and this cannot be done by association alone.
Second, even very young children may reformulate their requests when they fail to obtain a response (Bates 1976, Becker 1982): this provides evidence that they have more than an automated string-interpretation pairing available. Third, there are data which suggest that young children’s conversational contributions are sensitive to the demands of the communicative situation.For instance, toddlers have been noted to use the following forms with adults but not with peers (Read & Cherry 1978) – presumably because peers are less likely to attend to their desires and needs: (10) I want/need my cup.
Another example is the case of a 2-year-old girl studied by Lawson (1967, cited by Ervin-Tripp 1977). With peers, the girl used imperatives almost exclusively, while with adults she used desire statements of the type in (10) or question directives.In addition, she differentiated between children her age and slightly older children, using only permission and question directives to 4-year-olds and only a few imperatives with please to 3-year-olds.
These examples show an awareness of the appropriateness of certain forms for use in certain contexts which goes well beyond the command of ‘situated’ interpretations. In any case, the ‘situated interpretations’ in (8)-(9) do not seem to be equivalent to classic cases of standardisation such as (3).It is doubtful that standardisation as defined in section 1 plays a role in the development of pragmatic abilities. In Shatz & McCloskey’s (1984) study with 2-year-olds, as many children responded with yes/no plus action to ‘standardised’ request forms (e. g.
Can you put the ball in the truck? ) as did children to less standardised forms (May you put the ball in the truck? ): this is contrary to the expectation that standardised forms should produce fewer acknowledgements of their literal content, since their indirect interpretation is assumed to be streamlined. . 3 Indirectness and politeness There is a fair amount of evidence showing that the relationship between indirectness and politeness is not as straightforward as assumed by speech-act models.
Increased indirectness does not always signify increased politeness, as the following example shows: (11) Would you be so kind as to clean the mess in your room? As a result, studies of politeness have had to resort to more fine-grained analyses of what ‘counts as’ polite in given contexts.In the field of development, it has been shown that young children show some sensitivity to social status, age and other factors which may affect the selection of appropriate linguistic means in various conversational situations (Becker 1982). It is worth mentioning that the three aspects of the speech-act model I have concentrated on have also been questioned in the experimental literature on adult sentence processing. It has been shown that literal or direct speech-act interpretations do not invariably enjoy priority over non-literal or indirect interpretations (Gibbs 1983).It has also been demonstrated that standardisation and politeness ratings vary with context, rather than being associated with specific linguistic constructions (Gibbs 1981). As far as the acquisitional picture is concerned, the conclusion from the previous pages seems to agree with Levinson’s (1983: 282) observation: 3 However, despite much use of the terms speech act and performative, this recent work on language acquisition does not really support the importance of the concept of speech act at all; rather it emphasizes the essential roles hat communicative intention, utterance function and the interactive context play in the acquisition of language.
In the remainder of the paper, I want to introduce a different account of how utterances are understood in context which departs radically from speech-act assumptions. I will argue that this account is better equipped to provide a framework for advances in children’s early communicative ability. 3. An alternative pragmatic account One way of avoiding the problems faced by traditional speech-act theories is to abandon the assumption that sentences have literal forces.This move eliminates the distinction between direct and indirect speech acts, and leaves the theorist with a rather general problem: namely, how to pair utterances with illocutionary forces in context. On this account, mood indicators such as declarative, interrogative or imperative word order, intonation, particles and so on, do not encode primarily information about speech acts but instead have a general semantic content. Depending on contextual conditions, this semantic content may give rise to what has been felt to be a variety of illocutionary forces (Sperber & Wilson 1986/1995).
Let me use imperatives to briefly sketch how such an account might work: (12) a. b. c.
d. e. f. g. h. i. Get me my glasses. Stand up.
Try harder. Be happy. Take a break. Have a chocolate. Be careful. Give me your purse.
Come on, fight. Adopting a proposal by Wilson & Sperber (1993), I assume that imperatives semantically indicate that the propositional form of the utterance describes a state of affairs which is desirable (from someone’s point of view) and potential. Additional contextual considerations determine from whose point of view this state of affairs is considered desirable and, together with ragmatically available assumptions about the status, authority, etc. of the conversation participants, they contribute to different contextual readings of imperative utterances.
Consider (12a). Since getting her glasses is manifestly4 a state of affairs which the speaker considers desirable from her own point of view, and the hearer is manifestly in a position to realise this state of affairs, the utterance, on its most natural interpretation, is a request for action.The interpretation of (12b) is similar; assuming that the speaker has authority over the hearer, the utterance will be understood as an order. In the next example, trying harder is probably desirable (or at least beneficial) from the hearer’s point of view: consequently, (12c) will be understood as conveying advice. In (12d), being happy is considered desirable from the point of view of the speaker (and beneficial for the addressee); moreover, it is mutually manifest that whether the hearer is happy or not lies beyond the interlocutors’ control.
These assumptions are responsible for the ‘wish’ reading of (12d). Following similar lines of reasoning, we can explain why (12e) is interpretable as granting permission, (12f) as making an offer/suggestion, (12g) as issuing a warning, (12h) as raising a threat, and (12i) as communicating a dare. This analysis explains the variety of functions exhibited by sentence types without assuming that most of these functions need to override some ‘literal force’. It also avoids the problem of having to specify necessary and sufficient conditions for, e. g. questionhood, but accounts for the range of uses to which interrogatives etc. are put in terms of fine-grained context differences.
One may object that the reliable cross-linguistic appearance of the three basic sentence types (declaratives, interrogatives and imperatives) provides some support for canonical form-force mappings. One way of capturing such regularities in this system is to define three generic speech acts, SAYING THAT, TELLING TO, and ASKING WHETHER, in terms of the truth-conditional information encoded by the declaratives, imperatives and interrogatives respectively (Wilson & Sperber 1993).What have been considered more specific illocutionary forces can then be shown to derive through pragmatic inferencing from these underspecified speechact descriptions: for instance, offers, requests, wishes, and so on, may all arise contextually as instantiations of the TELLING TO proto-description. In this framework, politeness is no longer connected to ‘indirectness’: whether a linguistic form is judged to be polite or not depends on whether it conforms to certain culturally determined expectations about linguistic behaviour under different conditions.Furthermore, the place of standardisation needs to be rethought: since there are no indirect speech acts proper, whose comprehension is routed through a direct (or ‘literal’) force, the need for standardised inference is less pressing. A way of retaining whatever is valid in the intuitions underlying standardisation is to assume that members of different linguistic communities possess a set of higher-order representations specifying the use of linguistic representations (setting expectations of appropriateness, style etc. depending on context). These higher-order representations set expectations for the use of specific expressions in context and therefore affect the processing of such expressions.
The picture which emerges is very different from the original speech-act models. In fact, there are reasons to rethink the speech-act model even more radically. In particular, one may question the fundamental assumption that utterance comprehension proceeds in terms of speech-act recognition.As Sperber & Wilson (1986/1995) have argued, it is one thing for the theorist to classify speech acts as a way of capturing different intuitions about utterance readings; it’s another to attribute these classifications to the interlocutors and assume that they have to be performed in order for utterances to be understood. As S point out, communication proceeds through the recognition on the part of the addressee that the communicator intends to make manifest her intention to ake certain assumptions (more) manifest to the addressee.
The precise set of assumptions which the addressee recovers depends on complex considerations involving the amount of anticipated cognitive effects as well as the amount of processing effort incurred in deriving those effects. S claim that, although ordinary speech-act descriptions may be part of what the speaker intended to make manifest through her utterance, they are not necessarily so; consequently, they are not invariably recovered during utterance comprehension. 13) is presumably not part of the assumptions which were intentionally made manifest through (12d), but the theorist’s paraphrase of the range of cognitive effects which (12d) was meant to cause: (13) The speaker wishes the hearer to be happy. 4. Developmental implications Much work on the development of pragmatic abilities carried out within the speech-act framework has been devoted to a descriptive enterprise; its aim was to classify children’s early utterances in terms of speech acts and to collect facts about children’s success in different communicative situations.There are good reasons to believe that a pragmatic account of the kind presented in the previous section may prove better suited to the descriptive task of capturing children’s communicative skills. Furthermore, and more crucially, this account may contribute to a truly explanatory model of the mechanisms underlying early communicative performance. A first advantage of the proposed approach over speech-act analyses comes from a shift in emphasis: rather than classifying utterances in terms of speechact labels, this account seeks to establish how utterances succeed in making manifest a range of assumptions.
Speakers intend to make manifest to their audience a number of assumptions through the use of a single utterance; moreover, the degree of manifestness of these communicated assumptions typically varies (a limiting case are poetic metaphors which convey a set of increasingly weak implicatures; Sperber & Wilson 1986/1995). This has concrete implications for developmental research. Consider the following example, uttered by a 4-year-old who wants to play ‘airplane’ (from Gordon & Ervin-Tripp 1986: 319): (14) Okay, everybody, the airplane’s starting. On a standard speech-act account, such examples are treated as (indirect) requests.In fact, the utterance may be analysed as intended to achieve a range of effects: informing the audience that the game is about to start, communicating excitement, as well as recruiting players. One might try to capture this flexibility in speech-act terms by assuming that each utterance is used to perform a variety of speech acts.
This possibility is not sufficiently explored in most speech-act frameworks, and is virtually ignored in the speech-act based acquisition literature. In any case, even this option would not accommodate the fact that the assumptions conveyed by an utterance may vary in strength.In the second place, the proposed account offers a natural way of dealing with a related class of cases in which an utterance becomes a stimulus for achieving second-order effects. Several authors have noted that children exploit the assumptions concerning social status, authority etc. , on which utterance comprehension draws. For instance, Mitchell-Kernan & Kernan (1977) report uses of imperative requests in the speech of 7- to 12-year-olds which aim not at compliance but at testing or manipulating status relationships.
The following example is uttered by a 7-year-old to an 11-year-old girl: (15) Bring your li’l self here.In terms of the present account, apart from communicating a request, (15) succeeds in making manifest the assumption that the speaker rejects the hearer’s authority. This assumption causes a range of further cognitive effects, since it contradicts other assumptions in the mutual cognitive environment of the communicators (according to which age correlates with authority).
Such secondorder effects – which are obviously very important for social reasons – can be naturally accommodated within a model which recognises that utterances can be used to make manifest a variety of assumptions on different levels.A further desirable consequence of the proposed approach is that it avoids the problems of the literal force hypothesis. To recall an earlier example: it has commonly been observed that utterances with want and need are among the very first means of expressing requests in child language and appear around the same time as imperatives (during the second year).
This is puzzling from the perspective of the literal force hypothesis (according to which these are indirect forms of requesting) but can be explained by the alternative approach I have taken.On this approach, the degree of manifestness of a request depends (among other things) on the explicitness of the request, i. e. on whether the requested/desirable action is specified in the explicitly communicated content of the utterance.
From the point of view of explicitness, early imperatives typically fall together with want-need statements. There is plenty of evidence that what I have called ‘explicitness’ characterises the majority of children’s early requests (Ervin-Tripp 1977, Becker 1982). Interestingly, it is only around the age of 4 or 5 that children first start sing requests in which the desired object/course of action is implied (cf. Ervin-Tripp’s 1977 example He made sand go into my eyes). As the recent literature has demonstrated (Astington, Harris & Olson 1989), at this age children’s ability to handle representations of other people’s mental states including intentions, beliefs etc. , shows great advances compared to previous ages. As is obvious from the above examples, reference to children’s mentalising is crucially implicated in a realistic developmental account of how children come to master different aspects of form-force mappings.There are many interesting ways in which a pragmatic account of utterance comprehension can interact with research on theory of mind and perspective-taking.
One area which seems to be particularly promising is pre-linguistic communication. Following Sperber & Wilson (1986/1995), I assume that ostensive stimuli other than utterances may be used to make manifest to an audience a set of assumptions: moreover, the process of interpreting these stimuli in the two cases is subject to similar constraints (e. g. expectation of cost-efficient derivation of cognitive effects).
Consider the following description of how an infant aged 14 months uses an extended repertory of means to communicate with adults (from Lock 1980: 95-6): Mother enters the room holding a cup of tea. Paul turns from his playpen in her direction and obviously sees it. He cries vestigially and so attracts his mother’s attention; immediately, he points toward her and smacks his lips concurrently. Mother: No, you can’t have this one, it’s Andy’s.
Mother gives me (the observer) the cup of tea, and I put it on the mantelpiece to cool. Paul crawls across to me and grasps my knees.I turn to look at him; he looks toward the mantelpiece and points, turns back to me, continues to point, and smacks his lips. In this example, the infant uses pointing, gaze alternation, vestigial crying and lipsmacking as ostensive stimuli to convey something close to a request. One of the problems for developmental pragmatics within the speech-act framework was that the speech-act analysis could not extend straightforwardly to nonlinguistic communication: nonverbal speech acts do not seem to fit into a continuum with verbal speech acts (pace Bates 1976).The alternative pragmatic starting point proposed in section 3 escapes this problem by allowing more finegrained analyses of how stimuli are mapped onto communicative (or protocommunicative) purposes. Furthermore, it allows us to draw useful links to current research investigating infants’ preliminary attempts to grasp intentionality in others.
5. Conclusion In order to become successful communicators, individuals need to understand how different utterance forms can be the vehicles of different communicative intentions.Among the things which a communicator needs to master in order to correctly map an utterance onto its intended interpretation are at least the following: firstly, the linguistic resources required to assign syntactic and semantic structure to the utterance; secondly, an advanced metarepresentational device handling the attribution of mental states; thirdly, a system of social concepts involving status, authority, etc. ; finally, a set of higher-order representations specifying how linguistic forms are appropriately used in specific contexts.From the point of view of pragmatic theory, the challenge is to describe and motivate the workings of such a system in adults. From the point of view of developmental theory, the challenge is to provide an account of how young communicators acquire a gradual understanding of this system; furthermore, any developmental account needs to be responsive to the fact that very young children already seem attentive to significant dimensions of form-force mappings. In this paper, I have argued that standard speech-act theory cannot do duty as (or be the basis of) a plausible developmental account.I have also proposed a pragmatic account based on Sperber & Wilson (1986/1995), and argued that it offers a more promising starting point in dealing with the developmental data.
Endnotes * Thanks to Kent Bach, Lila Gleitman, Neil Smith, Deirdre Wilson and the members of the Language Acquisition and Processing group at IRCS (UPenn) for helpful feedback. 1. One may object that speech-act theory was not meant as a psychologically plausible account of verbal understanding but as a philosophical analysis of the kind of knowledge and reasoning required in order to comprehend utterances.However, this is neither the intention of speech-act theorists (see, e. g. , Chapter 11 in Bach & Harnish 1979 titled ‘The speech act schema and psychology’), nor the way the theory was interpreted by researchers in pragmatics, sentence processing and language acquisition. 2.
Shatz (1978) presents some evidence that the accurate responses in her sample are attributable to an ‘action bias’. 3. For previous critical discussions of the speech-act framework from a developmental perspective, see Shatz & McCloskey (1984), Gordon & ErvinTripp (1984), Becker (1982). 4.An assumption is manifest to an individual at a given time if she is capable of representing it and treating it as (probably) true.
When they are manifest to both speaker and hearer and assumed by these individuals to be manifest to both, assumptions are mutually manifest. References Astington, J. , P. Harris & D. Olson, eds.
(1989) Developing Theories of Mind, Cambridge University Press, New York. Bach, K. & R.
Harnish (1979) Linguistic Communication and Speech Acts, MIT Press, Cambridge, Mass. & London. Bates, E. (1976) Language and Context, Academic Press, New York. Becker, J. 1982) “Children’s Strategic Use of Requests to Mark and Manipulate Social Status”. In S.
Kuczaj (ed. ), Language Development, Vol. 2: Language, Thought and Culture, Erlbaum, Hillsdale, 1-35. Carrell, P.
(1981) “Children’s Understanding of Indirect Requests: Comparing Child and Adult Comprehension”, Journal of Child Language 8, 320-345. Dore, J. (1977) “Children’s Illocutionary Acts”. In R.
Freedle (ed. ), Discourse Production and Comprehension, Erlbaum, Hillsdale. Ervin-Tripp, S. (1977) “Wait for Me, Roller Skate! ” In S. Ervin-Tripp & C. Mitchell-Kernan (eds. ), Child Discourse, Academic Press, New York, 165188.
Garvey, C. 1975) “Requests and Responses in Children’s Speech”, Journal of Child Language 2, 41-63. Gibbs, R. (1981) “Your Wish Is My Command: Convention and Context in Interpreting Indirect Requests”, Journal of Verbal Learning and Verbal Behaviour 20, 431-444. Gibbs, R. (1983) “Do People Always Process the Literal Meanings of Indirect Requests? ” Journal of Experimental Psychology: Learning, Memory, and Cognition 9, 524-533. Gordon, D.
& S. Ervin-Tripp (1984) “The Structure of Children’s Requests”. In R. Schiefelbusch & J.
Pickar (eds. ), The Acquisition of Communicative Competence, University Park Press, Baltimore, 295-321.Levinson, S. (1983) Pragmatics, Cambridge University Press, Cambridge UK. Lock, A. (1980) The Guided Reinvention of Language, Academic Press, New York.
Mitchell-Kernan, C. & K. Kernan (1977) “Pragmatics of Directive Choice Among Children”. In S. Ervin-Tripp & C. Mitchell-Kernan (eds. ) Child Discourse, Academic Press, New York, 189-208.
Newcombe, N. & M. Zaslow (1981) “Do 21/2-Year-Olds Hint? A Study of Directive Forms in the Speech Of 21/2-Year-Old Children to Adults”, Discourse Processes 4, 239-252. Papafragou, A. (2000) “Indirect requests: A rethink”. Ms. /Under revision. Read, B.
& L.Cherry (1978) “Preschool Children’s Production of Directive Forms”, Discourse Processes 1, 233-245. Reeder, K. (1980) “The Emergence of Illocutionary Skills”, Journal of Child Language 7, 13-28.
Searle, J. (1969) Speech Acts, Cambridge University Press, Cambridge UK. Searle, J. (1975) “Indirect Speech Acts”. In P. Cole & J. Morgan (eds. ), Syntax and Semantics 3: Speech acts, Academic Press, New York, 59-82.
. Shatz, M. (1978) “On the Development of Communicative Understandings: An Early Strategy for Interpreting and Responding to Messages”, Cognitive Psychology 10, 271-301. Shatz, M. & L.McCloskey (1984) “Answering Appropriately: A Developmental Perspective on Conversational Knowledge”. In S.
Kuczaj (ed. ), Discourse Development, Springer, New York, 19-36. Sperber, D. & D. Wilson (1986) Relevance: Communication and Cognition, Blackwell, Oxford.
2nd ed. 1995. Tsohatzidis, S. , ed. (1994) Foundations of Speech Act Theory, Routledge, London.
Wilson, D. & D. Sperber (1993) “Linguistic Form and Relevance”, Lingua 90, 125.
Anna Papafragou Early communication: Beyond speech-act theory In this paper, I discuss certain aspects of speech-act theory with respect to data from preschoolers’ speech.I consider, in particular, the association of sentences with literal and direct illocutionary forces, the way illocutionary standardisation is assumed to affect utterance comprehension, and the connection between indirectness and politeness: I conclude that these aspects of the speech-act model of utterance comprehension do not offer a satisfactory framework for dealing with early communication in children. In the second part of the paper, I outline a different pragmatic approach to verbal understanding based on Sperber & Wilson (1986/1995), and argue that it offers a better starting point for accounting for the developmental data.