Размещено на http://www.allbest.ru/

Introduction

A speech act is a minimal functional unit in human communication. Just as a word (refusal) is the smallest free form found in language and a morpheme is the smallest unit of language that carries information about meaning (-al in refuse-al makes it a noun), the basic unit of communication is a speech act (the speech act of refusal).

According to Austin's theory, what we say has three kinds of meaning:

propositional meaning - the literal meaning of what is said

It's hot in here.

illocutionary meaning - the social function of what is said

«It's hot in here» could be:

an indirect request for someone to open the window

an indirect refusal to close the window because someone is cold

a complaint implying that someone should know better than to keep the windows closed (expressed emphatically)

perlocutionary meaning - the effect of what is said

«It's hot in here» could result in someone opening the windows

Based on Austin's, and Searle's theory, Cohen identifies five categories of speech acts based on the functions assigned to them.

Representatives Directives Expressives Comissives Declaratives

assertions suggestions apologies promises decrees

claims requests complaint threats declarations

reports commands thanks offers

Speech act theory attempts to explain how speakers use language to accomplish intended actions and how hearers infer intended meaning form what is said. Although speech act studies are now considered a sub-discipline of cross-cultural pragmatics, they actually take their origin in the philosophy of language.

It was for too long the assumption of philosophers that the business of a «statement» can only be to «describe» some state of affairs, or to «state some fact», which it must do either truly or falsely. But now in recent years, many things, which would once have been accepted without question as «statements» by both philosophers and grammarians have been scrutinized with new care. It has come to be commonly held that many utterances which look like statements are either not intended at all, or only intended in part, to record or impart straight forward information about the facts.

Philosophers like Austin, Grice, and Searle offered basic insight into this new theory of linguistic communication based on the assumption that «the minimal units of human communication are not linguistic expressions, but rather the performance of certain kinds of acts, such as making statements, asking questions, giving directions, apologizing, thanking, and so on». Austin defines the performance of uttering words with a consequential purpose as «the performance of a locutionary act, and the study of utterances thus far and in these respects the study of locutions, or of the full units of speech». These units of speech are not tokens of the symbol or word or sentence but rather units of linguistic communication and it is «the production of the token in the performance of the speech act that constitutes the basic unit of linguistic communication». According to Austin's theory, these functional units of communication have prepositional or locutionary meaning (the literal meaning of the utterance), illocutionary meaning (the social function of the utterance), and perlocutionary force (the effect produced by the utterance in a given context).

Speech acts have been claimed by some to operate by universal pragmatic principles (Austin, Searle, Brown & Levinson). Others have shown them to vary in conceptualization and verbalization across cultures and languages. Although this debate has generated over three decades of research, only the last 15 years marked a shift from an intuitively based approach to an empirically based one, which «has focused on the perception and production of speech acts by learners of a second or foreign language (in the most cases, English as a second or foreign language, i.e., ESL and EFL) at varying stages of language proficiency and in different social interactions». Blum Kulka et. al., argue that there is a strong need to complement theoretical studies of speech acts with empirical studies, based on speech acts produced by native speakers of individual languages in strictly defined contexts.

The illocutionary choices embraced by individual languages reflect what Gumperz calls «cultural logic». Consider the following passage:

The fact that two speakers whose sentences are quite grammatical can differ radically in their interpretation of each other's verbal strategies indicates that conversational management does rest on linguistic knowledge. But to find out what that knowledge is we must abandon the existing views of communication which draw a basic distinction between cultural or social knowledge on the one hand and linguistic signaling processes on the other.

Differences in «cultural logic» embodied in individual languages involve the implementation of various linguistic mechanisms. As numerous studies have shown, these mechanisms are rather culture-specific and may cause breakdowns in inter-ethnic communication. Such communication breakdowns are largely due to a language transfer at the sociocultural level where cultural differences play a part in selecting among the potential strategies for realizing a given speech act. Hence the need to make the instruction of speech acts an instrumental component of every ESL/ EFL curriculum [2, 89].

When second language learners engage in conversations with native speakers, difficulties may arise due to their lack of mastery of the conversational norms involved in the production of speech acts. Such conversational difficulties may in turn cause breakdowns in interethnic communication. When the nonnative speakers violate speech act realization patterns typically used by native speakers of a target language, they often suffer the perennial risk of inadvertently violating conversational (and politeness) norms thereby forfeiting their claims to being treated by their interactants as social equals.

Communication difficulties result when conversationalists do not share the same knowledge of the subtle rules governing conversation. Scarcella ascribes high frequency of such difficulties to the fact that «nonnative speakers, when conversing, often transfer the conversational rules of their first language into the second». Scarcella provides the following example. (Bracketing indicates interruptions.)

1) speaker A: Mary's invited us to lunch. Do you wanna go?

2) speaker B: Sure. [I'm not busy right now. [Why not?

3) speaker A: [Good [I'll come by in about thirty minutes

4) speaker B: Think we oughta bring [anything?

5) speaker A: [No, but I'll bring some wine anyway.

In this exchange, the native speaker B inaccurately concluded that the nonnative speaker A is rude since like many Americans, he regards interruptions as impolite.

Rather than associate rudeness with A's linguistic behavior, however, B associates rudeness with A herself. B's reasoning might be as follows: A interrupts; interruptions are rude; therefore, A is rude. Such reasoning is unfortunate for A, who comes from Iran where interruptions may be associated with friendliness, indicating the conversationalist's active involvement in the interaction. Learners who repeatedly experience conversational difficulties tend to cut themselves from speakers of the target community, not only withdrawing from them socially, but psychologically as well. «Psychological distance» or a «high filter» might be related to a number of factors, including culture shock and cultural stress». All these factors ignite a cycle that eventually hinders second language acquisition.

First, the learners experience conversational difficulties.

Next, they become «clannish», clinging to their own group.

This limits their interaction with members of the target culture and increases solidarity with their own cultural group.

That, in turn, creates social distance between themselves and the target group.

The end result is that the second language acquisition is hindered since they don't receive the input necessary for their language development.

Cohen claims that the fact that speech acts reflect somewhat routinized language behavior helps learning in the sense that much of what is said is predictable. For example, Wolfson & Manes, have found that adjectives nice or good (e.g., «That's a nice shirt you're wearing» or «it was a good talk you gave») are used almost half the time when complimenting in English and beautiful, pretty, and great make up another 15 percent [3, 32].

Yet despite the routinized nature of speech acts, there are still various strategies to choose form - depending on the sociocultural context - and often a variety of possible language forms for realizing these strategies, especially in the case of speech acts with four or more possible semantic formulas such as apologies and complaints. Target language learners may tend to respond the way they would in their native language and culture and find that their utterances are not at all appropriate for the target language and culture situation.

At present, there is an increasing number of studies dealing with teaching speech act behavior in an ESL/ EFL classroom. Olshtein and Cohen, for instance, conducted a study of apologies made by EFL learners in Israel who were taught a set of lessons on the strategies used by native English speakers to apologize. They found that situational features can indeed be taught in the foreign language classroom. Whereas before these apology lessons, the nonnative speakers' apologies differed from the native English speakers', after instruction, learners selected strategies, which were more native-like.

Scarcella provides second language instructors with a number of guidelines intended to reduce negative consequences of communication difficulties and increase the learners' conversational competence through improving their motivation:

Stress the advantages of conversing like a native speaker.

Stress that it is not necessary to converse perfectly to communicate in the second language.

Impress upon learners that they should not be overly concerned with communication difficulties.

Help students accept communication difficulties as normal.

Provide students with information about communication difficulties.

Do not expect students to develop the conversational skills needed to overcome all communication difficulties.

Provide communicative feedback regarding student success in conveying meaning and accomplishing communicative objectives.

Teach students strategies to help them overcome communication difficulties in the real world.

1.2 The evolution of the speech acts

The idea of tracing speech acts in historical contexts was discussed in the inaugural issue of the Journal of Historical Pragmatics, with the core issue raised by a polemical question: "Is a Diachronic Speech Act Theory Possible?". In fact, studies in this area had already been conducted. Back in the 1970s, Schlieben-Lange and Weydt discussed the historicity of speech acts. Lebsanft provided a close analysis of greetings in Old French; in English, Arnovick's article series on illocutionary histories of selected speech acts started in 1994/1995, and her monograph on the same topic came out in 1999 [4, 49].

The inaugural issue of the Journal of Historical Pragmatics contained a host of diachronic speech act studies including Schrott on questions in Spanish, Culpeper and Semino on speech acts related to witch hunts in Early Modern England, and Jucker and Taavitsainen on insults in the history of English. Within the past seven years, historical pragmatics has become an established field in historical linguistic study and the number of diachronic speech act studies has grown considerably, even to the extent that it has become of equal importance as (or at least a rival to) the study of linguistic processes. Recent studies include Jucker, Kohnen, Alonso Almeida and Cabrera-Abreu, Busse, Pakkala-Weckstrom, Milfull, Archer, Grzega and Taavitsainen and Jucker. This volume represents a new stage of diachronic speech act research, but there are, of course, still many gaps, and the long lines of diachronic development are only just emerging.

Considering the fundamental importance of speech act values, the dearth of knowledge is even more surprising. While we know a lot about the development of sounds and sound patterns, and the structure of words, phrases and sentences, we still know very little about how speakers used words and sentences to communicate. Did earlier speakers of English use the same repertoire of speech acts that we use today? Did they use them in the same way? How did they signal speech act values and how did they negotiate them in cases of uncertainty?

We feel that speech act analysis could serve as a ground-breaker towards a pragmatic history of the English language, although we only have flashes of past practices and a more detailed and comprehensive picture is not yet possible. In the quotation cited at the beginning of this chapter, Bertuccelli Papi uses the analogy of a mountain that can be climbed from various points of departure and with various types of equipment. In the same way, on our quest for a pragmatic history of English through diachronic speech act analysis, we are on the largely uncharted slope of the pragmatic side of the mountain.

Researchers can take different starting points when they investigate the history of individual speech acts or individual classes of speech acts and they can use different tools, but they always have to be aware of the slippery and treacherous nature of their endeavour.

Historical speech act studies can take one of two forms. Diachronic speech act analysis aims at charting the manifestations of speech acts through various periods by comparing two or more synchronic descriptions, but synchronic descriptions, dealing with one period only, are also possible. Language history presupposes a longer time line, and thus it is the former type of analysis that comes into focus in this volume.

It may be useful to start by comparing some key issues in historical linguistics, and look at how other, better-charted fields of language history and the better-established subfield of historical pragmatics, i.e. linguistic processes, deal with these key issues, and then go on to consider the case of speech act analysis in historical pragmatics. Historical speech act studies have several pitfalls, and different types of problems are encountered.

Historical pragmatics is situated at the crossroads between pragmatics and historical linguistics, but the overlap is by no means complete. Language change is the core area where historical pragmatics meets historical linguistics, as studies on linguistic processes are central to both. Historical linguistics aims at adding to our historical knowledge about languages and language families, trying to explain and interpret language history, whereas historical pragmatics focuses on the meaning-making processes in past contexts. The latter studies show how meaning is negotiated and how more is conveyed than is said, and it takes language users into account. The area of overlap includes language change, but there are different emphases and different motivations become foregrounded. Traditionally, linguistic processes are discussed with phonogological, morphological, syntactic, and semantic motivations, but some pragmatic aspects may demand other tools. In general, the status of pragmatics in linguistic change is not a straightforward matter. Lass acknowledges the difficulties and concludes that the area is best avoided, touching upon it only in passing. Away from the common core there are other differences besides emphasis.

Speech act studies cannot be squarely placed in the overlapping area of language change. For example, one of the corner-stones of historical linguistics, underlying most studies in the field, is the Uniformitarian principle. This has been explicitly discussed in connection with phonological reconstruction and sociohistorical linguistics. There are several different formulations of the theory ranging from strong claims to observations which take a safer stance and approach the variationist principle (for a discussion, see Taavitsainen and Fitzmaurice). In her groundbreaking book on sociohistorical linguistics, Romaine takes the Uniformitarian principle as her point of departure, but in a somewhat modified form. She claims that "the linguistic forces which operate today and are observable around us are not unlike those which have operated in the past". Brinton discusses the principles from the point of view of grammaticalisation and other linguistic processes, such as pragmaticalisation, lexicalisation, and idiomaticisation, and states that the underlying assumption is that pragmatic meaning works uniformly over periods and societies. The key to interpretation is the context of utterance. There is even a more general formulation of the Uniformitarian principle: "no linguistic state of affairs (structure, inventory, process, etc.) can have been the case only in the past". This claim is interesting as even pragmatic phenomena like speech acts seem to repeat the basic patterns in slightly modified forms over the course of history. Ritual insults provide a case in point: they occur in Old English in Beowulf, in Old Icelandic sagas, in Middle Dutch romances, in the Finnish national epic the Kalevala, in health guides in Early Modern England, among London teenagers and black youths in New York, to name but a few manifestations that provide flashes of something that seems deeply rooted in human behaviour. But the question of whether this is valid more generally, in other types of speech acts, remains to be answered in the future. Inherently polite speech acts can be more sensitive to changes of fashion and cultural variation (cf. the contrary meanings of gestures in different cultures), but, at the same time, this principle must hold at some level. People and human behaviour cannot have changed so much in the course of years, decades, and centuries [6, 81].

From the present point of view it is of interest that the Uniformitarian principle was formulated in the nineteenth century and underlies several fields of study in that period. If we go back to philology with its emphasis on contextual interpretations, we find nineteenth-century scholars pondering how the nature and uses of speech must have been the same throughout the history of language. The basic distinction in modern pragmatics between conventional and conversational implicature applies to historical interpretations as well: some pragmatic inferences are valid in multiple contexts while implicatures of conversational contexts may not be replicable. Jokes and insults have been mentioned in connection to this; unless the hearer is familiar with the culture and the context in which the speech act occurs, s/he may not be able to understand it.

Other instances of verbal communication may be problematic as well. Meaning- making processes are sensitive to context and the meaning of an utterance may be completely different in different contexts, for example, the sentence "Your hair is so long?" could be an expression of several speech acts with different meanings, depending on the context. It could be an indirect command "have your hair cut", an insult or a compliment, or just a neutral statement. In spoken language the tone of voice and the intonation often make the intended meanings and illocutions explicit and help in the interpretation, but extralinguistic cues are lost in the written mode, and we have to rely on other means of interpretation. In some genres, like fiction, narrators' comments are sometimes present and make the intended meanings explicit.

Meanings are negotiated, and we can make inferences by examining utterances in their context, taking various factors into account. In speech act studies we look at social action through fragments: instances of an activity type. In this activity, the context gives us clues on how to understand and interpret the speech act. The frame of the action and the response are important (cf. cognitive approaches). The context influences how we understand what we see, we may perceive a physical object from different perspectives by moving around it, weighing human activity in a local environment, with changes according to the angle of view. For example, a typical compliment formula can have the reverse meaning if the context so requires (see Jucker et al. this volume). In addition to local environments, various layers of context can be discerned, and we can move from more concrete to more abstract levels. In this way we may view speech acts from various perspectives and visualise them in a multidimensional space [7, 63].

Speech acts are fuzzy concepts that show both diachronic and synchronic variation. In a multidimensional pragmatic space, speech acts can be analysed in relation to neighbouring speech acts, to their changing cultural groundings, and to ways in which they are realised. The fuzziness of speech acts requires a prototype approach; individual instances vary in their degree of conformity to their prototypical manifestations and sometimes the group identity is only vague. Our model allows a great deal of both diachronic and synchronic variation, and neighbouring speech acts can be defined in relation to one another in the pragmatic space, in analogy with the semantic field theory which views word meanings in relation to neighbouring words. The multidimensional space allows conceptualisations of the different aspects and characteristics of speech acts. The model should be discussed, developed and tested further to see whether such a way of visualising speech act realisations can provide a model for diachronic speech act analysis.

In our previous work we sketched the multidimensional space of the speech act of insults in the following way:

ritual, rule governed - creative

typified - ad hoc

truth-conditional - performative

conventional - particular

ludic - aggressive

intentional - unintentional

irony - sincerity

Formal level:

Semantics:

Context dependence:

Speaker attitude:

Reaction:

reaction in kind - denial, violence, silence

The point of departure is the description of the speech act with its distinctive, obligatory defining criteria, e.g. insults must contain a predication about the target, and this predication has to be perceived as disparaging. What people considered insulting is a matter of culture to a large extent, and with our speech-act verb survey of verbal aggression we achieved an ethnographic view of what was considered insulting. In this way, speech act studies can give us insights from within a society into its norms and values, from the point of view of people living and acting in the culture under investigation. We included non-intended insults in the category as well, but illocutions may count and an insult may be intended even if not perceived as such, so it is possible to draw the lines differently. Insults and compliments are alike in many respects, and the presence of irony or banter can reverse the meanings completely [8, 76].

Although the above framework was drafted for insults and verbal aggression, it can be applied to other speech acts and speech act categories. Of the Searlean categories, several fit nicely into the framework and highlight different aspects of it. Expressive speech acts are often considered the most elusive and difficult category, expressing "the psychological state specified in the sincerity condition about a state of affairs specified in the propositional content". Besides insults, apologies, compliments, thanks and greetings are some of the most important speech acts of the expressive category, and they have received attention both in the earlier studies and in this volume. For example, the dimension of irony versus sincerity becomes prominent with compliments. Apologies are expressed in routinised, perhaps even ritual and rule-governed forms, though creative instances can also be found. Commissives, in turn, highlight the problems of truth-conditional speech acts, and perhaps the same dimension is important for assertives as well. For directives, speaker attitudes are of special concern, with politeness issues at the forefront. Context and second turns need to be taken into account in all categories, and the dimension of typified/routine versus ad hoc is of special interest for the inventories of speech acts as well as for the methods of speech act retrieval from computerised corpora. This dimension (i.e. typified - ad hoc) is also genre-specific, e.g. in Middle English saints' lives, insults are structurally important by providing the turning-points of the plot, and thus the dimension of ritual versus creative is highlighted.

If we view diachronic speech act analysis as a form of contrastive analysis, historical speech act analysis is similar in some respects to contrastive analysis, but there are differences as well. The analysis relies on the identification of similar speech functions in different cultures so that we have tertium comparationes that remain constant across space or time. However, an important difference is found in the fact that contrastive analysis compares the realisations of a specific speech act in different cultures in two or more disparate contexts. In contrast, continuity is found between one stage of language and another in historical speech act analysis. There is a linear development from older stages of language to more recent phases. What this kind of development can tell us about politeness, for instance, is most interesting.

The speech acts that have received most attention are generally those that constitute face-threatening acts, e.g. requests, apologies, complaints, thanking. Because of this close affinity, it is relevant to discuss the issue of culture-specificity versus universal validity of Brown and Levinson's politeness theory based on face wants, "the want of every 'competent adult member' that his actions be unimpeded by others". It has been claimed that this key issue does not hold in all cultures and that the universal significance of the theory is a gross mistake. Instead, they want to go unimpeded is a fundamentally cultural script, quintessentially English, reflecting one of the corner-stones of the Anglo-American culture.

Apologies, for instance, can be seen as face-threats to the speaker's own positive face because they acknowledge an offence or a perceived offence for which the speaker is responsible. In Modern English, apologies tend to be routinized expressions of remorse and regret. In particular, the phrase I am sorry or just sorry focuses on the speaker's emotion. In Early Modern English apologies were less routinized and more often involved phrases such as pardon me or excuse me, which ask for the addressee's forgiveness and thus also constitute a face-threat to the addressee's negative face [9, 112].

Requests in Middle English, to take another example, were often made as directives indicating the speaker's desire. By the end of the Middle English period this was no longer seen as appropriate. The wishes of the speaker alone were no longer seen as sufficient justification for a request, and the co-operation of the addressee was seen as an important factor giving rise to indirect requests indicating negative politeness, i.e. a consideration for, or at least a token acknowledgment of, the addressee's wish not to be imposed upon. Towards the end of the Early Modern period directives became more conventionalized through the use of expressions or "politeness markers" such as pray, I beseech you, do me a favour or, later still, please.

Thus, the diachronic development of speech acts can reveal the formation of cultural scripts. There are differences between speech acts in their relation to the notion of face, as understood by Brown and Levinson. For example

Orders and requests can easily threaten rapport, because they affect our autonomy, freedom of choice, and freedom from imposition, and this threatens our sense of equity rights (our entitlement to considerate treatment).

In compliments, the connection with French court culture and cultural influences from outside are clear.

In a historical perspective, politeness in English society, as witnessed by written texts of the past, especially fiction and drama, has moved from positive politeness culture to negative politeness. However, in the form of address terms in private letters, there is also evidence for the opposite development, i.e. in the direction of more positive politeness. Their Corpus of Early English Correspondence ranging from the fifteenth to the seventeenth century revealed a development from a predominance of negatively polite honorifics to a predominance of positively polite nicknames and terms of endearment. One important reason for this development might be the growing privacy of personal letters in the centuries under investigation. This development, therefore, does not provide counter-evidence for the more general trend towards the negative politeness culture described above.

As always, research methods and research questions are closely linked. The available research methods narrow down the questions that we can ask, and specific research questions require specific research methods. The first studies of speech acts were carried out by philosophers. Austin and Searle used philosophical methods to think about speech acts. They tried to find out what it means to use language for certain effects, and they reflected on the specific conditions that have to be met if speakers are to successfully christen a ship, make a bet, ask a question or give a piece of advice. Such methods are clearly unsuitable for historical investigations. While we may reach fairly reliable answers when we ask ourselves what it means to issue a command or to make a promise, we cannot ask such questions about earlier stages of the language. However, in the meantime, the range of research methods used in speech act analysis has seen considerable expansion.

In addition to the philosophical approaches, experimental methods and, later, corpus-based methods have been developed. Blum-Kulka et al., for instance, developed discourse completion tasks that were used in cross-cultural speech act research. Informants for a range of different languages were asked to fill in blank spaces in constructed dialogues in order to get specific and carefully controlled realisations of requests or apologies. Trosborg added role-plays and role enactments to the inventory of speech act research methods. Participants were asked to improvise scenes in which requests, complaints or apologies were likely to figure prominently. In the case of role enactments, participants played roles that they were familiar with from their own daily lives. In the case of role-plays they also acted out unfamiliar roles, for instance, a participant who, in his or her normal life is a student, may play the role of a professor. However, such experimental methods are also unsuitable for historical investigations [10, 22].

Historical investigations depend on corpus-based methodologies, and these were developed rather late in the history of speech act research. We can distinguish between manual searches of small-scale corpora and computerized searches of large corpora. Both can be applied to historical data, but the latter depend on large-scale computer- readable corpora and appropriate search algorithms, both of which have only become available relatively recently. One of the first large-scale computer-based investigations of a speech act in Present-day English was carried out by Deutschmann, who investigated apologies in the one-hundred million words of the British National Corpus.

Manual methods have a longer history. Lotscher, for instance, investigated the use of swear words and insults in the history of Swiss German, and Lebsanft investigated greetings in Old French. In recent years, similar investigations in English have proliferated, see e.g. Hughes and McEnery on swearing, Arnovick, Jucker and Taavitsainen on insults, Busse on requests, Milfull on advice in Middle Scots, or Arnovick and Pakkala-Weckstrom on promises in Chaucer's Franklin's Tale.

Manual methods either focus on a fairly small corpus, e.g. Pakkala-Weckstrom (this volume), who concentrates on promises in Chaucer's work or Busse (this volume), who concentrates on directives in one particular play, i.e. Shakespeare's King Lear, or else they are necessarily eclectic. In Jucker and Taavitsainen, for instance, we investigated insults from Old English to Present-day English by picking out relevant examples in the history of English on the basis of our own reading and of the relevant literature. This is called "illustrative eclecticism" [11, 47].

Computerized searches for specific speech acts can only be undertaken if the speech act tends to occur in routinized forms, with recurrent phrases and or with standard Illocutionary Force Indicating Devices (IFIDs). Deutschmann, for instance, analyzed apologies that tend to occur in routinized forms or with the IFID sorry. In some cases, specific speech acts can also be realized with the speech act verb itself used performatively, as in the case of "I hereby apologize". Thus the investigation is based on a selection of representative features throughout the history of English or at least some part of the history of English. This is called "structured eclecticism".

2. Compliment as the part of speech acts

2.1 The definition of the compliment

Compliments are contradictory speech acts: they enhance the face of the addressee while they put her in a potentially awkward position. Thus a compliment is at the same time a face-enhancing act (FEA) and a face-threatening act (FTA). In response, compliment recipients must either agree to the compliment in some way or other and thereby violate the modesty maxim, or they must disagree and thereby violate the maxim of agreement. The sincerity of speaker illocutions, e.g. whether a routine compliment or a sarcastic comment, may be problematic, and specifications for the right interpretation often accompany the speech act. These mechanisms have been researched in various languages and contrastive studies are frequent. Compliments are gendered speech acts: according to empirical studies women give and receive more compliments than men, and the topics are different. From a historical point of view compliments have not received much attention. A notable exception is Beetz, who has studied compliments in the history of German.

In this paper, we shall take the first steps towards a speech act history of compliments in English by outlining their development from a more general sense to the more specific known as compliments today. We shall discuss the claims made on the basis of Present-day compliments about gender issues, about women being more prone to give and receive compliments and about the topics of compliments being gender- specific.

We shall investigate whether, and how, predecessors of Present-day compliment can be found in our corpus material from the Early Modern period to the early twentieth century, and we shall assess the contexts of such compliments, including compliment responses. The historical perspective, with an analytic grid of gender, can indicate how conventional the features of gender-specific compliments are. It has also been noticed in recent literature that politeness is not conceived in the same way in all English-speaking cultures in our present-day world; the focus of this article is on past cultures in Britain, which adds an interesting angle to the discussion and which sets social practices in relation to social issues of women's position in past and present cultures. Our second aim is to further the methodology of corpus-aided speech act research. Our "ethnographic" method for studying compliments of the past builds on our earlier studies, especially the insights gained with speech act labels of verbal aggression. Our hypothesis is that speech act labels allow an ethnographic view of how speakers describe, classify and evaluate speech acts. Compliments are particularly intriguing as critical or even negative speaker illocutions may be disguised in seemingly positive utterances [12, 79].

Every speaker is endowed with positive and negative face, i.e. the wish to be appreciated and liked and the wish to exercise one's own free will and to be unimpeded by others. These face wants have often been criticized as being culture-specific. We do not take a strong view on this issue, but we see it as highly likely that different cultures differ greatly in the relative weight that they give to positive and to negative face wants. Our use of these categories in our analyses does not imply a claim that these notions are universal but - as we hope to show - they nevertheless continue to be useful theoretical concepts, not only for Present-day English, but also for earlier stages of English. Furthermore, politeness has already proved a more diversified and dynamic notion than usually recognized; we hope to take up some additional issues about the gender-based approach.

Many verbal and non-verbal actions of interlocutors have an impact on the positive and the negative face of both the speaker and the addressee. Speech acts, such as requests and apologies, have long been analyzed as FTAs of the addressee's negative face and the speaker's positive face respectively. In a request, the speaker imposes on the addressee's negative face by asking him or her to do something that he or she might otherwise not have done, and in an apology the speaker admits having perpetrated an offence that is serious enough to require such an apology. However, in addition to acts that are inherently face threatening, there are also actions that enhance the face wants of the addressee and perhaps of the speaker herself. Kerbrat-Orecchioni suggested the term "face-flattering acts" (FFAs) for such speech acts. We think that this is an important theoretical insight, as we will show below, but we propose the term "face-enhancing act", because of its more neutral connotation. Flattery implies excessive and insincere praise. We want to use the term in a more neutral sense to refer to speech acts that inherently add to the positive or negative face of the addressee or the speaker.

Holmes argues that women tend to be linguistically more polite than men and connects this behavior with positive politeness, building up solidarity and friendliness. According to her, gender is a determining factor and compliments are more common among women: women receive more compliments, mostly from other women, so that compliments can be considered indexical of positive politeness and feminine strategies in New Zealand data. In American culture, compliments form an essential part of the norms of women's behavior; complimenting becomes a heightened verbal activity of American girls at one stage of their development, and it is not only positive politeness but also sarcastic compliments that play an essential part in this process, casting doubt on the overwhelming interpretations of compliments as belonging to positive politeness. Tannen investigates an extract of work-place discourse between female office workers, where the complimenting behavior clearly reflects the status and power differences between the interlocutors.

In studies on present-day compliments, various pragmatic research methods have been applied and discussed in the literature. Manes and Wolfson argue that only the ethnographic method is reliable for collecting compliments. By ethnographic method they understand field methods like the diary method or the participant observation method. Together with their students they collected 686 examples of compliments as they occurred in "real situations", i.e. in everyday interactions. Their claim of capturing authentic language use by this method is valid in so far as participant observation does not distort the behavior of the people under scrutiny. In general, the earlier literature seldom takes the overtones of compliments into account, although sarcastic compliments are common [13, 69].

The same approach and method of collecting compliments was used by Holmes to investigate complimenting behavior of speakers in New Zealand. Together with her students, she collected 484 compliments or rather compliment exchanges, because she and her students were careful to also collect the responses that were given to these compliments. The exact words used in both the compliment itself and the response to it were taken down as soon and as accurately as possible, with relevant contextual details. She claims that compliments fulfill very different functions in men's and women's speech: "women tend to use compliments as positively affective speech acts whereas men more often perceive them as face-threatening acts". The topics of compliments are different, as women receive compliments about their appearance while men receive more compliments about their pos-sessions. Holmes's more recent assessment on gender differences and politeness is based on the same empirical research.

The most comprehensive study on American compliments consists of a corpus of 1,062 compliment events, including both compliments and responses to them. The focus of the article is different from Holmes's study, and more attention is paid to the responses. In general, compliments from men were accepted, while compliments paid by women received varied responses. Twelve different response types are distinguished: appreciation token, acceptance, praise upgrade, comment history, reassignment, return, scale down, question, disagreement, qualification, no acknowledgement and request interpretation.

In addition to the diary method, discourse completion tasks have been used as the data collecting method. In such a task, informants, usually students at the university of the researcher, are given descriptions of sample situations and they are asked to provide what they perceive to be a natural compliment in the given situation. An example of this method is provided by Chen, who used a discourse completion task to investigate cultural differences in compliment responses between speakers of Chinese and of American English. Respondents had to react to four slots for different responses, in four different situations, with compliments that praised the looks, clothes (a nice sweater), achievements (the performance in a presentation) and possessions (e.g. a Rolex watch) of the addressee. The discourse completion task is a somewhat simpler and more artificial method of data collecting, used mostly for contrastive studies, but to our knowledge it has not been applied to a gender study. If we apply Holm- es's observations on compliment topics and gender to the above situations, looks and clothes would stereotypically be female topics and achievements and possessions male talk. Chen claims that most compliments paid in daily life are about these four topics. More recently, Schneider and Schneider have extended this line of research and used the same discourse completion task with speakers of Irish English and of German. This method, although it works for contrastive studies, simplifies real life situations and ignores all more refined nuances [14, 72].

Language communities are said to react differently to compliments and the use and the status of compliments vary. Larger sequences of turns reveal the attitudes. Speakers of American English or New Zealand English are said to react with thanks or acceptances, while the Japanese are said to reject or downplay compliments. In some cultures third turns occur as well, as it is important to upgrade the compliment and stress its sincerity, e.g. by adding "and this is not a compliment" or "and I really mean it". The purpose of compliments is clarified by such a statement: the prime motivation for the compliment is to give a declaration of "truth", the positive evaluation is claimed to be objective and not only for making the other person feel good.

Compliments have also been considered from the feminist point of view. Patterned conversational moves can coerce people into gendered roles and enforce them. For example, a compliment offered to a woman on her appearance is part of the linguistic practice by which women are judged by their looks, while men are complimented on their actions and judged by their accomplishments. Comments on women's looks can easily be understood as patronising, and in work situations they can efficiently downplay a woman's professional contribution.

Some languages are claimed to have more routinized compliments, while others are said to be more creative; this may be true of different historical periods of the same language as well. Manes and Wolfson claim that "one of the most striking features of compliments in American English is their almost total lack of originality", and Holmes supports this view: "Compliments are remarkably formulaic speech acts in that a very small number of lexical items and syntactic patterns account for the great majority of them." Of the 686 compliments collected by Manes and Wolfson and their co-researchers, the pattern that something is/looks/seems (really) good/ beautiful/some other positive adjective, accounts for more than half of all the compliments, and the three most common patterns account for 85 per cent of all. Thus, according to the earlier literature, Present-day compliments are highly formulaic. According to Lewandowska-Tomaszczyk, Polish compliments follow very similar patterns and are equally formulaic. In a recent survey, Holmes summarizes the earlier research results on the narrow range of adjectives and syntactic patterns. In her own corpus, nice, good, beautiful, lovely and wonderful, were the core adjectives, whereas Wolfson's American corpus had nice, good, beautiful, pretty and great. Likewise, the verbs that predominated were few: like, love, enjoy and admire. Syntactically, women's compliments were found to be more complex, while men tended to revert to the minimal pattern, e.g. "great shoes!". (Jucker, Schneider, Taavitsainen and Breustedt, this volume, test the claim about the formulaic nature of compliments and whether it can fruitfully be made use of in corpus linguistic methodology) [15, 39].