5.1.1 Language games

The linguistic turn in philosophy was initiated in the late nineteenth century by the German mathematician Gottlob Frege (1848-1925) who considered the logical shortcomings and imperfections of language a hindrance for precise arithmetic thinking. Due to inferences in language, he “found the inadequacy of language to be an obstacle; no matter how unwieldy the expressions I was ready to accept, I was less and less able, as the relations became more and more complex, to attain the precision that my purpose required” (Frege 1967: 5–6; emphasis added).

Consequently, Frege took to examining language from a mathematical-logical point of view as an integral part of his larger scientific project of logicism. Frege’s intention was to reconstruct language in a logical manner for mathematical and philosophical purposes because ordinary language contained too many potentially misleading superficial similarities to be used for precise logic. This led him to deduce that sentences denote, logically speaking, either true or false meaning (referring to statements, thoughts, and ideas which he understood as objects in their role in logic), an assumption that was difficult to prove, given the fact that words only point to meanings but do not contain them and that language is a sociocultural tool.

Some decades later, Ludwig Wittgenstein (1889-1951) turned his attention to ordinary language which Frege had dismissed as logically deficient. In his work Philosophical Investigations (1953), Wittgenstein acknowledges the vagueness and lack of precision of ordinary language, yet attempts to provide meticulous analyses of the grammar of ordinary language in order to describe both the implications inherent in the actual use of language and the external conditions under which use of language makes sense. This includes the necessity to portray tacit knowledge which we normally take for granted. Language must be understood properly, that is, with all its potential shortcomings, because of its development into the form of language that obviously is viable for optimal interaction and construction of self, Other, and others. In direct contrast to Frege, he suggests: “When I talk about language (words, sentences, etc.) I must speak the language of every day. Is this language somehow too coarse and material for what we want to say? Then how is another one to be constructed?” (Wittgenstein 1953: § 120; emphasis in the original). In a similar approach to Bakhtin, Wittgenstein considers language, not as an isolated system of signs, but as a system deeply embedded in social use. Hence, he firmly locates the meaning of a sign or word in its actual use by stating that “the meaning of a word is its use in the language” (Wittgenstein 1953: §43), and: “Every sign by itself seems dead. What gives it life? – In use it is alive” (Wittgenstein 1953: § 432). The meaning of a word and of a sign is derived from its use by other people. Our principal point of reference for meaning is not private experience or a dictionary entry, but each other: “So one might say: the ostensive definition explains the use – the meaning – of the word when the overall role of the word in language is clear. Thus if I know that someone means to explain a colour-word to me the ostensive definition ‘That is called ”sepia“’ will help me to understand the word” (Wittgenstein 1953: § 30).

According to Wittgenstein, the actual use of language is governed by the rules of language games (Sprachspiele) which are in turn derived from the social practice of a speech community in a historical dimension. Hence, language and social performance are intricately interwoven in the language game. Wittgenstein suggests that different language games have their own distinctive grammars, just as different games (for instance, chess, soccer, badminton, cricket, etc.) have their own rules. And just as ordinary games have no mega-rules which can be applied to all games, language games cannot be regulated by some sort of master-game; rather, they are irreducible and have their own rules, just like ordinary games. Although language games are obviously determined by their rules, it would be a mistake, according to Wittgenstein, to think of rule-following as a fundamental structure. To follow a rule, be it in the games of language or chess, is just custom and tradition (cf. Wittgenstein 1953: § 199). However, when playing any game, including a language game, one can make a mistake by applying a rule wrongly. Hence Wittgenstein infers a paradox: “no course of action could be determined by a rule, because every course of action can be made out to accord with the rule” (Wittgenstein 1953: § 201). Rules cannot be identified by instructions; rather, they can be identified through the common practices of those who participate in the game, which include reference to rule-books and precedents (cf. Wittgenstein 1953: § 199).

For the use of language in social contexts, this means that the rules which characterize the concepts expressed are those manifested by the sociocultural practices of the speakers or writers engaged in the language game. In fact, we can only participate in social practice when we have a certain mastery of the language game. With complete mastery remaining unachievable, the levels of mastery or competence vary from person to person. Wittgenstein (1953: § 31) explains that knowing what the figure of the king is in chess does, not automatically implies that one knows the rules of the game of chess, that is, how the figure of the king is moved properly. Also, the material and shape of the figure itself is irrelevant as long as all players agree that it actually represents the king, even if it is represented by just a match or a cork. If one equates the figure of the king in chess with that of a word in language, as Wittgenstein (1953: § 108) does, the need for learning or acquiring the rules of the language game in the process of playing the game is essential. For children growing up into a speech community, this is an incidental and tacit process which is part of their socialization. But if one is learning a second language, it is obvious that the other language game has to be learned, including the rules, by using the second language, or playing the other game, not according to the L1 rules but by applying the L2 game rules.

However, there are important differences, as no two games follow the same rules. Therefore, it makes no sense to refer to the rules of the first language game when learning a L2; ideally, the L2 learner must learn the other language in its pragmatic and social context from the perspective of tradition and use, as applied by the players of the other language game. The level of understanding and mastery of the second language game, that is, the appropriate use of relevant terms and concepts, the understanding of rules and their explanation, and the application of appropriate register in Discourses and genres, becomes evident in the actual sociolinguistic practices as a (second) language user by (inter-)acting in the L2 cultural and social contexts. Wherever immersion in the L2 society and culture is impossible, for example, in institutionalized L2 learning, a rich learning environment with regard to a multiplicity of learning opportunities has to be provided to mimic the L2 environment. However, this does not imply that the L2 classroom is by definition deficient compared to the L2 immersion. The L2 classroom can provide a didactically and methodologically more purposeful approach to the L2 learning of a particular group of L2 learners (in terms of age, interests, levels of engagement, etc.) by orienting the learning efforts to their specific zones of proximal development (cf. Section 9.5). It can provide a structured and secure space for learners to explore and reflect upon alternative realities, which are accessed through the L2, and blended with internalized cultural traditions, values, and worldviews.

5.1.2 Speech act theory

Wittgenstein’s emphasis on the grammar of ordinary language games has generated a broader approach to analyze ordinary language. Linguists and language philosophers such as Karl Bühler, Roman Jakobson, Charles Morris, John Austin, and John R. Searle approached the analysis of spoken communication from a philosophical angle critical of logical positivist philosophers, such as Frege, Moore, Russell, and others who viewed everyday language as somewhat deficient. The British language philosopher John L. Austin, for instance, observes that everyday language is an extremely efficient and smooth medium of communication for ordinary people. Therefore, he argues, analysts should not strive to rid everyday language of its imperfections, but should focus rather on understanding how people manage to communicate efficiently with ordinary language, since it is used successfully by ordinary people to cope with a myriad of situations in everyday life. Austin focuses on ordinary language, not only to provide an alternative to positivism, but also because he understands ordinary language as a conceptual legacy with socio-historical and phylogenetic dimensions. “[O]ur common stock of words embodies all the distinctions men have found worth drawing, and the connexions they have found worth marking in the lifetime of many generations” (Austin 1946: 129-130, cited in Thomas 1995: 31). This means that ordinary language carries with it conceptualizations and distinctions, such as ordering, inviting, or requesting, which have developed over generations. Therefore, they must be important to the users of that language in a social dimension.

From these reflections, Austin concludes that language is not reducible to semantic function but that it has an important additional function, namely the performative dimension. When someone says I hereby declare you man and wife, then this person not only speaks, but at the same time performs the social, and legally binding, action of marrying two people; of course, not everyone can utter these words with the intended effect, only those who have been institutionally legitimized to carry out these actions can do so. Thus, there must be certain conditions that have to be fulfilled if the performative force of a speech act can be socially successful. However, sometimes speech acts result in the failure of their performative intent, which Austin describes as these speech acts being “infelicitous” (Austin 1962:14). Therefore, there have to exist certain felicity conditions for speech acts to be successful. When examining social conventions supporting performatives, it is obvious that there is a gradient between performative actions that are highly institutionalized (for example, a judge pronouncing a verdict), through to less formal acts of everyday life such as greeting, warning, thanking, etc.

If speech acts are treated as social actions, then there cannot be abstract truth-conditions isolated from social contexts. Hence, Austin’s best-known book carries the programmatic title How to Do Things with Words (Austin 1962), emphasizing the capacity of speech acts to create social realities. Thus, Austin is interested, not in analyzing continuous discourse, but in the analysis of isolated speech acts. The speech act is differentiated into three forces, the locutionary (actual uttering of meaningful words), the illocutionary (intention or force behind words) and perlocutionary forces (effect of illocution on hearer) (cf. Searle 1969: 22-25). The notion of an illocutionary force of speech acts refers to the concept that every utterance carries with it the speaker’s intent to achieve a particular purpose, and that this can be analyzed by interpreting the utterance.

The American philosopher John R. Searle developed aspects of Austin’s work further by attempting to systemize it. In what can be seen as an expansion of Wittgenstein’s notion of language games, he proposes a more detailed classification of the variety of major categories of speech acts and extends the rules to the include verbs such as requesting, asserting, questioning, thanking, advising, warning, greeting, congratulating. He also tries to extend the analysis of speech acts into the immediate social context in which it is produced. Searle (1976:10–16) suggests that, while there is a myriad of language-specific speech acts, all speech acts can be categorized into five main types: (1) representatives (committing the speaker to the truth of the expressed proposition; for example, asserting or concluding, as in John smokes a lot); (2) directives (attempts by the speaker to get the addressee to do something, for example, requesting or questioning, as in Get out of this room!); (3) commissives (committing the speaker to some future course of action, for example, promising, offering, threatening, as in I promise to pay you tomorrow); (4) expressives (expressing a psychological state, for example, thanking, apologizing, welcoming, congratulating, as in Congratulations on your promotion), and (5) declaratives (effecting immediate changes in the institutional state of affairs and tending to rely on elaborate extralinguistic institutions, for example, excommunicating, christening, marrying, declaring war, firing from employment, etc.).

Searle uses a mix of criteria to establish these five types, including the speech act’s illocutionary force (the purpose of the act), its “fit” in the world (relationship between language and external world), the psychological state of the speaker (speaker’s state of mind), and the content of the act (propositional content of act). He also further develops Austin’s felicity conditions into a classification of conditions underlying a successful speech act. Searle (1969: 66-67) distinguishes between preparatory, propositional, sincerity, and essential conditions for a speech act and explicates these conditions for different types of speech acts, for instance, for questioning (cf. Searle 1969: 66) where S is the speaker, H the hearer, and P the proposition expressed in the speech act:

1. Preparatory 1: S does not know the answer for a yes/no question, does not know whether P is true or false, for an elicitative or WH-question, does not know the missing information.
2. Preparatory 2: It is not obvious to either S or H that H will provide the information at the time without being asked.
3. Propositional: Any proposition or propositional function.
4. Sincerity: S wants this information.
5. Essential: The question counts as an attempt to elicit this information from H.

Useful as this prototypical characterization may be, it certainly is not applicable to all types of questions. For example, rhetorical questions, ironic remarks, questions by a teacher in the classroom, or by a lawyer in court, do not follow these conditions for questions.

Although speech act theory transformed the analysis of language with the fundamental insight that use of language must be functionally motivated, rather than formally defined, it has also been criticized in several respects. For example, there is no clear relationship between the discourse function (illocutionary force) and the grammatical form (type of clause) (cf. Eggins and Slade 1997: 40). Austin’s performative hypothesis, namely that only performative verbs could be used to perform actions, is untenable because of the lack of formal criteria for distinguishing performative verbs from other verbs, and hence many acts are performed “using language where it would be impossible, very odd or very unusual to use a performative verb” (Thomas 1995: 46). For instance, it would be very unusual to explicitly announce that one is insulting someone by uttering the phrase I (hereby) insult you! The speech acts of insulting, hinting, boasting, etc., are typically performed without using a performative verb. And finally, Searle focuses on the description of speech act verbs, although he claims to provide the rules for complete speech acts (cf. Thomas 1995: 99). Thomas arrives at a conclusion which could be applied to most attempts of linguistics to find rules for interpersonal conversation: “The whole approach to describing speech acts in terms of rules was misconceived. Within linguistics there is a powerful push towards formalization; formalisms give an impression of intellectual vigour which, when applied to most areas of pragmatics, has proved to be almost entirely illusory” (Thomas 1995:107).

In their analysis of speech acts, Austin and Searle applied a philosophical – rather than linguistic – perspective in that they focused on the interpretation – rather than the production – of utterances in discourse, thus differing from Bakhtin’s approach. Despite the criticism that could be leveled against their work in terms of the classification of speech acts, it contributed to the rise of pragmalinguistics, which in turn resulted in the communicative approach of second language teaching and learning methodology. The communicative approach tries to link speech (albeit frequently formulaic speech) to definable socio-pragmatic contexts, such as at the restaurant or in the post office, in order to prepare L2 learners for the communicative demands they will face when encountering these situations in the other speech community.

5.1.3 The cooperative principle and maxims of conversation

Paul Grice (1975) developed ordinary language philosophy further by closely examining the implications of utterances. By identifying a distinction between conversational implicature and literal meaning, Grice made connections to the work of language philosophers who are concerned with truth-conditions. A lot of meaning can be intentionally implied in saying something. However, how one says it and what one does not say can be more important than the message itself. For example, by saying The car looks black, the speaker implies that he or she doubts that the car in question is indeed black. If he or she thought it were black, he or she would have said The car is black. This point raises the question of appearance and reality which was largely overlooked by language philosophers of the truth-conditional school since they did not make this distinction. Therefore, they did not discuss the presupposition that, if something looks black, it actually is black. If a speaker wants to avoid or retrospectively cancel implications without incoherence, he or she could simply say that The car looks black, and I have no doubt but that it is black. Hence, implications are an important feature of conversation, and Grice (1975; 1979) has developed a sophisticated account of conversational implicature in terms of principles and maxims – rather than rules – in order to account for the interpretations of interlocutors in a conversation. The use of principles have in this context an advantage over rules; they are more flexible since they can be invoked to different degrees, and are thus more adequate for the analysis of dynamic linguistic contexts, such as spoken interaction: “Rules are all or nothing, principles are more or less. Rules are exclusive, principles can co-occur. Rules are constitutive, principles are regulative. Rules are definite, principles are probalistic. Rules are conventional, principles are motivated” (Thomas 1995:107).

Thus, Grice’s approach to using principles for analyzing communication promises to avoid the misconception of forcing socio-functional categories and structures into mechanistic sets of rules. Grice assumes that there seem to be sufficient regularities in the inference-forming behavior of interlocutors to exploit this by implying something in verbal interaction, rather than stating it explicitly. At the center of Grice’s pragmatic approach to conversation analysis is the co-operative principle, a kind of tacit agreement by interactants to cooperate in communication, which states that in conversational interaction, the participants work on the assumption that a certain set of rules is in operation (for example, genre, narrative, Discourse), unless they receive indications to the contrary. Hence, Grice advises participants in interaction, “Make your conversational contributions 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” (Grice 1975: 45). This general co-operative principle is then operationalized into four maxims of conversation which Grice (1975: 45-46) takes to govern all meaningful interaction (and therefore can be learned and used by L2 learners):

Grice assumes that in conversation the interlocutors are communicating under the mutual assumption that the utterances are made along a number of parameters, namely that the truth is being told and the background knowledge of interactants is taken into consideration in utterances. Grice suggests that these assumptions are generally adhered to in communication, and that they have a guiding influence on the conversation. However, this is not necessarily the case, as the speaker might intentionally tell lies (hence violating the first maxim), or he or she may report on irrelevant events or experiences in order to avoid silence, and the listener may – or may not – realize this.

In his theoretical approach, however, Grice takes for granted that interlocutors operate on the assumption that, as a rule, the maxims will be observed. He also assumes, as Fairclough (1995: 46) points out, that conversations occur co-operatively between interlocutors as equals, thus ignoring potential inequalities in the power relationship between them. Additionally, as Frow (2005: 79) criticizes, Grice understands speech as purely rational exchange of information, thereby ignoring other potential functions, such as sarcastic, rhetorical, or ironic. Furthermore, Grice’s examples of utterances consist almost exclusively of short sentences or exchanges that are isolated from the broader communicative and social context, thus limiting the sense-making potential of certain kinds of immediate and pragmatic information. And lastly, Grice (1979) characterizes a speaker’s meaning as an overt intention to cause a certain cognitive effect on the part of the listener (cf. Chapters 5 & 6). This implies that the speaker’s meaning is an intention or mental state. However, “the mental states of others cannot be simply perceived or decoded, but must be inferred from their behaviour, together with background information” (Sperber and Wilson 2008: 87). This includes the subjective perception of unintended signals on the part of the addressee which the speaker sent out (in the perception of the listener) and to which the addressee attaches meaning.

This implies that Grice’s model provides no valid understanding of the complex processes by which meanings, values, emotions, and truth-effects are shaped and constrained by genre, narrative, or Discourse. Consequently, the general principles (cooperation; relevance) and the more specific maxims are too sweeping to explain how particular kinds of information are implied in certain contexts (cf. Frow 2005: 80). Hence Gricean pragmatics, although providing useful tools for heuristic purposes, imply an idealized concept of conversations as homogenous, conflict-free, and equal sites of verbal interaction that blatantly ignores the potential for conflict, chaos, difference, and asymmetry in power inherent in interaction. Thus, its value for L2 learners is ultimately rather limited; however, it has the potential to focus the awareness of L2 learners on the complexities of interaction, and the relevance of the concept of implication for interaction can clearly be learned in this context.

5.2.1 Communicative competence and contextualization cues

The notion of dynamic cultural, social, and pragmatic contexts for communication was elaborated upon by sociolinguists and ethnographers who devised analytical concepts, such as contextualization cues. The sociolinguist Dell Hymes and the linguistic anthropologist John J. Gumperz have been leading this field of research into the use of language in the social contexts of everyday life. The basic assumption of this approach to research is contained in this statement:

On this basis of the social and ethnographic situation of speech production, Chomsky’s (1965: 3) notion of linguistic competence has been criticized as too narrow because it operates with the notion of an ideal speaker or hearer acting within a rule-governed linguistic system in an artificial space, cut off from social, cultural, discursive, and pragmatic contexts. Chomsky purposely restricts his analysis to the linguistic system: “[A] grammar is an account of competence. It (...) attempts to account for the ability of a speaker to understand an arbitrary sentence of his language (...). [I]f it is a linguistic grammar it aims to discover and exhibit the mechanisms that make this achievement possible” (Chomsky 1974: 73). In response to the limitations of Chomsky’s notion of linguistic competence, Hymes (1972b) developed the broader notion of communicative competence which takes the social and pragmatic context of situated speech into consideration. David Crystal defines communicative competence as the

The relevance of this notion of communicative competence lies in the fact that the interlocutors are not only seen as more or less competent speakers of a language, but as social actors interacting with others within complex communities which are organized by a variety of social institutions and who, as such, interact through a network of intersecting sets of expectations, beliefs, desires, and values which are structured by generic configurations such as genre, narrative, and Discourse, and generated by the blending of conceptual spaces. Communicative competence is not so much an intrapersonal property (as Chomsky’s notion of linguistic competence would suggest) but a dynamic interpersonal construct that can only be examined by means of overt performance between speakers in the process of communication, as embedded in the broader Discursive, social, and cultural context.

However, communicative competence is difficult to analyze with theoretical models and difficult to teach and learn in second language instruction, as it is tied to the specific communicative situation, the interactants and the force of interaction, tacit sociocultural knowledge, time, space, etc. Communicative competence contains not only linguistic competence in Chomsky’s sense, but, according to Savignon (1983), also discourse competence (the ability to connect sentences in stretches of discourse and to form a series of meaningful utterances), sociolinguistic competence (“an understanding of the social context in which language is used: the roles of the participants, the information they share, and the function of the interaction” [Savignon 1983: 37]), and strategic competence (strategies for maintaining conversation even in difficult circumstances, e.g., when an interlocutor does “not know how to express a point, leading to paraphrase, circumlocution, repetition, hesitation, avoidance, and guessing, as well as shifts in register and style” [Savignon 1983: 40-41]). Bachmann (1990) introduces further competences, for instance “organizational competence” and “pragmatic competence” (Bachmann 1990: 87). However, this model also has its deficiencies. For instance, it fails to include notions such as the voice or force of the utterance, while the previous linguistic, social, and cultural development of the subject in terms of acquiring competences which may be deficient in some aspects, is neglected. Due to the difficulties of defining a sufficient set of sub-competences to comprehensively capture the complex processes at play in interaction, the more general term communicative competence seems better suited to emphasizing the fact that there are social, cultural, pragmatic, and subjective forces at work in interpersonal communication. These are typically too specific to allow for the development of a comprehensive set of rules, or even principles, which can explain these complex forces and processes.

The linguist John Gumperz took on the challenge of explaining the more tacit configurations by focusing on subtle contexts influencing the process of constructing meaning in conversations, such as prosody. He approached this terrain of linguistic research from an anthropological perspective by analyzing these features of conversation with interactants from different sociocultural backgrounds (including interethnic groups such as British and Indian speakers of English living in Britain, cf. Gumperz 1982). Applying this cross-cultural approach, Gumperz could demonstrate that discursive features such as intonation can be interpreted by participants in the conversation quite differently. For instance, while some interactants may interpret the intonation of certain utterances as rude or aggressive behavior, the identical intonational pattern in conversation might be interpreted by others as indicating consideration and deference. The interpretation of the same conversational units in terms of intonation is therefore also dependent on the sociocultural background of the interactants, which may guide their understanding of the content and context of the situation. Gumperz concludes that our understanding of a communicative situation is to some extent guided by contextualization cues which he defines as “any feature of linguistic form that contributes to the signalling of contextual presuppositions” (Gumperz 1982: 131).

The contextualization cues in discourse are not the same for every individual participating in conversation; they are shaped by tacit cultural background knowledge which is applied to the understanding of the conversation. “What we perceive and retain in our mind is a function of our culturally determined predisposition to perceive and assimilate” (Gumperz 1982: 4). On that basis, idiomatically competent participants in conversation can use expressions and characteristics of expressions such as voice, pitch, and prosody to intentionally create certain contexts of communication and subtexts of meaning which then have to be inferred by the receiver, for example, by using exaggerated impetus in uttering certain words or phrases. Speakers and listeners use contextualization cues in order to make certain cognitive, social, and discursive schemata available to the interactants for agreeable understanding. Gumperz refers to this implicit and constructive cooperative element in interaction when he states, “Do not verbalize explicitly, what the conversation is about, rely on the listeners’ ability to use this background knowledge. If he is a friend, he will guess what is wanted” (Gumperz 1982: 138-139). This maxim can only be exercised in recurrent pattern combinations of contextualization cues, effortlessly establishing a shared understanding of the situation and developing a shared pool of knowledge, despite the fact that there is no guarantee for successful communication since so many factors are influencing the construction of meaning in conversation.

These foundational differences in cultural background make successful intercultural communication extremely difficult, since contextualization cues are either absent or inadequate for the appropriate interpretation of context. Contextualization cues can, however, in broadening Gumperz’ 1982 definition, also be non-verbal, for example, clothes, certain moves, or non-linguistic utterances, such as laughter or silence. However, this fact complicates the solving of the “crossword puzzle” (Levinson 1995: 238) of conversation, especially if one takes into consideration possible instances where something is perceived, but not intended to be meant, or on the contrary, where something is implied as a contextualization cue, but not inferred as a cue for conversation.

If one considers the close link between language and culture, the notion of communicative competence can be expanded to include cultural elements to a greater extent than the notion of sociolinguistic competence would suggest. Cultural elements are usually tacit in language use; they do not rise to consciousness unless they become problematic in interaction. This is frequently the case when interaction takes place across languages and cultures. Consequently, the notion of intercultural competence has emerged.⁵³ The concept of intercultural competence is much broader than those of linguistic or communicative competence because “the term goes beyond aspects of skill and performance, embracing deeper notions of disposition, intention, motive and personal identity” (Fleming 2009: 3). However, the notion of intercultural competence is not in conflict with that of communicative competence. Rather, it underlies the communicative elements with (inter-)cultural dimensions of subjective construal in the process of L2 learning, as is reflected by the term “intercultural communicative competence” (Byram 1997: 32).

5.2.2 Analyzing discourse

A less comprehensive, yet more focused approach on isolated and limited aspects of interaction is provided by Conversation Analysis (CA), Discourse Analysis (DA), and Critical Discourse Analysis (CDA). These approaches do not try to develop an overall model of linguistic and sociocultural dynamics, as Wittgenstein, Austin, Searle, and Grice attempted, but intentionally limit their focus of analysis to micro-aspects of conversation with the goal of arriving at rules and principles of the communicative micro-structure in its social context. For this reason, they may be of value to second language teaching and learning methodologies; they have the potential to provide the L2 learner with rules and principles of how to appropriately engage in L2 interactions. While speech act theory applies a philosophical approach to the interpretation of units of conversation and Grice approaches the analysis of conversation from a pragmatic angle, CA, DA, and CDA are much narrower in their approach by trying to (re-)construct a framework for the microanalysis of the linguistic structure of conversational exchange, also taking the pragmatic context of the discourse into consideration. This context is not devoid of social value because it is itself a semiotic construct, for example, register, genre, or Discourse.

Conversation analysis, as a methodological research approach, focuses on empirically-based accounts of observable conversational behaviors of interlocutors which are embedded in social action in general, and talk-in-interaction in particular. Conversation is understood to be an organized, socially structured event. CA aims to analyze how interlocutors behave and orient their use of language to certain behavioral practices, as they co-construct ongoing conversation in real time. Conversational partners respond in spontaneous and creative, yet typically highly structured ways to the continuously evolving text and context of the interaction. The activities of getting attention, taking the floor, topic nomination, topic development, topic termination, turn-taking, topic clarification, repair, topic shifting, avoidance, and interruptions are all important research areas in CA. In focusing on these mechanisms inherent in conversation, the social context of analysis is restricted to the immediate contextual influences on communicative behavior. The emphasis of CA is clearly on the analysis of talk-in-interaction without resorting to contextualization.

CA basically differentiates between two main types of talk-in-interaction: ordinary conversation and institutional talk. Ordinary conversation refers to speech exchanges such as everyday chitchat that occurs between friends, neighbors, and acquaintances, which mainly serves to maintain the social fabric between these interlocutors. Utterances in ordinary conversation often appear in “adjacency pairs,” for example, questions and answers in formulae of greetings and responses (“Hi, how are you?” “Fine, thanks, and you?”). Certain responses are socially expected or preferred, while others are considered to be inappropriate. By paying close attention to these features, a lot can be learned about cultural norms and social relationships. In contrast to ordinary conversation, institutional talk tends to be more formalized, for instance, classroom talk, press conferences, doctor-patient conversations, courtroom interactions, etc. Institutional talk is strongly guided by the dominant Discourse, it is oriented towards a certain outcome, and it positions the participants in relation to one another. Knowledge of the principles, linguistic and behavioral, of institutional talk is part of the language competence of every adult native speaker of a language; however, in L2 curricula, they remain largely sidelined. The reason for this ignorance may be the implicit assumption that institutional talk and its guiding Discourse are identical across cultures. Obviously, this assumption, while possibly sustainable for certain aspects of Discourses across closely interrelated cultures, is plainly wrong in a broader context.

Conversation analysis, which has developed from ethnomethodological approaches (e.g., Garfinkel 1967), originally endeavored to discover and construct an analytical grid which could explain certain micro-structural patterns in conversation. One of the very early proponents of CA, Harvey Sacks, defined the ultimate goal of conversational analytical methodology as follows: “What one ought to seek is to build an apparatus which will provide for how it is that any activities, which members do in such a way as to be recognizable as such to other members, are done, and done recognizably” (Sacks 1972: 332). Several decades later, however, such an apparatus still does not exist, and should it ever be constructed, the question has to be asked what its exact purpose should be, if it is not tied in with the broader social context of verbal interaction, as it takes place and the cultural background of the interlocutors which provides, among other information, the contextualization cues.

Conversation analysis, however, has been very successful in developing a sustainable theoretical framework for some micro-structural patterns of conversation from an analytical perspective which does not treat “participant orientations, relevancies, and intersubjectivity (...) as states of mind that somehow lurk behind the interaction, but as local and sequential accomplishments that must be grounded in empirically observable conversational conduct” (Markee and Kasper 2004: 495). This approach has been partially successful for analyzing classroom interaction as a specific mode of institutional talk which is characterized by an unequal power system in terms of interaction between teachers and learners, where teachers have a privileged position to assign topics and the right to speak to learners. Furthermore, teachers have the right to evaluate the quality of pupils’ contributions to the emerging interaction through initiating or providing repairs in terms of corrections and to initiate and end question-answer sequences. By providing micro-analyses of conversational turns, both in terms of co-construction of meaning and in terms of learners’ construction of social membership categories and identities on an emerging interactive basis, CA can contribute to a deeper understanding of the intersubjective construction of meaning in the L2 classroom. The advantage of such an emic approach is the use of “a grounded, social, and discoursal perspective on language, rather than an idealized, cognitive, sentence-level understanding of what language is” (Markee and Kasper 2004: 495). This approach has the potential of analyzing the conversational construction of meaning and identities in the intersubjective space which is spread out between the interactants, in terms of turn-taking, sequence organization, and repair organization of talk in classroom interaction. Thus, cognition is seen not as subjectively internalized and therefore inaccessible competence, but as socially distributed and observable behavior. In his attempt to develop a learning tracking methodology for CA for second language acquisition (SLA), Markee (2008) tried to analyze how L2 learners deploy these intersubjective resources “that underlie all talk-in-interaction, combined with the co-occurrent organization of eye gaze and embodied actions – (...) to co-construct with their locally enacted, progressively more accurate, fluent, and complex interactional repertoires in the L2” (Markee 2008: 406). While this research has excellent potential to analyze “the interactional architecture of discursive practices” (Markee and Kasper 2004: 495) which may be focused on a particular grammatical, morphological, lexical, pragmatic, cultural, or other L2-related items, it is of limited value for the analyses of broader processes of L2 learning and L2 acquisition because it does not relate to an explicit theory of language learning. It may account for the intersubjective learning of a particular L2 item by learners in a particular location and at a specific time, and these local insights might be combined in the long term to develop a CA-for-SLA theory, as Markee (2008) suggests. For the moment, however, this theory is just at the very beginning of being developed, and for the foreseeable future it cannot claim validity; it is also questionable if this emic (i.e., participant-relevant) CA-for-SLA research can transfer results to other social, institutional, spatial, language and culture-related contexts to lay the foundation for a L2 learning theory.

If CA is already extremely difficult to use for the development of a L2 learning methodology, it is even more so for developing a theoretical framework for analyzing ordinary conversation which can be located in a huge range of settings with even more factors influencing the interaction and the interlocutors (in their conversational behavior). Due to the complex and potentially chaotic nature of the course of actual verbal interaction, conversation analysis has been criticized for three major drawbacks: it has a “lack of systemic analytical categories” as well as applying a “fragmentary focus” and providing a “mechanistic interpretation of conversation” (Eggins and Slade 1997: 31). Eggins and Slade (1997) suggest that, due to its focus on isolated excerpts of speech,

Thus, CA is not in a position to provide a comprehensive analysis of conversation with inclusive meso-levels of structure and context, paying attention to all influencing factors. In view of these drawbacks of conversation analysis, Eggins and Slade (1997: 32) conclude that “a shift of orientation away from conversation as a form of social interaction that is incidentally verbal, and towards conversation as linguistic interaction that is fundamentally social” is needed for a more comprehensive analysis of conversation. This shift towards social and cultural aspects of conversation, however, begs the question as to how these potentially very wide-ranging social aspects can be analyzed in a more general, rule-generating manner. Hence, it is difficult to see what CA can contribute to L2 teaching and learning methodologies in terms of useful guidelines transcending micro-structure that is, of how to approach teaching and learning appropriate communicative behavior in L2 situations.

Discourse analysis in the narrower sense tries to examine the relationship between the forms and functions of language by analyzing both the structure and the functions of spoken and written interaction in more rigidly structured varieties. ⁵⁴ DA refers “to all forms of spoken interaction, formal and informal, and written texts of all kinds. So when we talk of ‘discourse analysis’ we mean analysis of any of these forms of discourse” (Potter and Wetherell 1987: 7). DA examines language as it is used in structured units larger than the sentence, for instance, discourses such as job interviews, classroom lessons, or doctor-patient interaction. Thus, discourse analysis focuses on inter-sentential discourse relations and on the structure of the discursive exchange which is seen as a basic unit of conversation. However, without taking the pragmatic contexts of discourse into consideration, communication can be very ambiguous. Conversely, conversation is marked by the exchanges between the interlocutors; the utterances of one participant are usually followed and built upon by the utterances of other participants. Therefore, DA promises to analyze actual discourse by paying attention to both the internal structure of the discourse, including the psychological influences of interlocutors, and the pragmatic (or external) context influencing discourse.

However, since the context refers both to very complex psychological influences and highly structured situations, Levinson (1983: 287) criticizes DA for being based on “premature formalization” of structures and categories mainly derived from very restricted forms of formal discourse, such as classroom interaction, and subsequently transferred to spoken interaction in general. Bakhtinian notions of heteroglossia and polyphony, for example, are not considered in this approach. Thus, discourse analysis is also not in a position to provide comprehensive and reliable tools for analyzing spoken interaction. Therefore, it is also of no great value for L2 teaching and learning, at least not in the beginning and at intermediate levels.

Critical Discourse Analysis takes a broader approach to analyzing spoken interaction in trying to take all influencing factors into consideration, including the social and cultural setting and power relations among the interlocutors. CDA understands written and spoken discourse as a form of social practice which not only reflects social relationships but actively constitutes and organizes them. Therefore, it tries to analyze the interrelation of linguistic means and concrete discursive acts on the one hand, and the interplay of discursive practice and political, social, and institutional “reality” on the other. The focus of interest for these critical analyses lies in examining the power over the discourse and the power within the discourse.

Hence, CDA views discourse analysis as a “transdisciplinary activity” (Fairclough 1997: 4). It recognizes that language in spoken interaction is never neutral, that it always depends on a multitude of factors, including the motivation, intentions, and objectives of the interlocutors, the structure of the interaction and the processes of construing meaning; it is also structured to a large extent by social surroundings in terms of social domains, Discourses, genres, and institutional framework. According to Fairclough (1992: 63; emphasis added), discourse is “a form of social practice, rather than a purely individual activity or a reflex of situational variables.” Fairclough (1995) elaborates on the consequences of this view by suggesting that, “Viewing language use as social practice implies, first, that it is a mode of action (Austin, 1962; Levinson, 1983) and, secondly, that it is always a socially and historically situated mode of action, in a dialectical relationship with other facets of ‘the social life’ (its ‘social context’) – it is socially shaped, but it is also socially shaping, or constitutive” (Fairclough 1995: 131; emphasis in the original).

CDA views micro-actions, such as verbal interaction, and macro-social structures as inextricably linked because the latter influence the former and every action on the micro-level contributes to the reproduction of “macro structures” (Fairclough 1995: 34).

Thus, Fairclough (1995: 35; emphasis in the original) criticizes the CA approach, because “it makes little sense to study verbal interactions as if they were unconnected with social structures.” One of the problems Fairclough identifies with approaches such as CA or Gricean pragmatics is that they construct conversation as speech events between consciously operating, independent, rational, social actors, who are co-operating in conversation in order to achieve certain goals through homogenous interactions. Such an idealized view, however, ignores the multi-faceted situational, cultural, and social influences on the situation, as well as the interdependence of views and the intersubjectivity of the interlocutors in their efforts to co-construct meaning, which might be done from positions of difference and conflict, rather than cooperation. Therefore, Fairclough suggests that there must be a “knowledge base” of four components which constitutes interlocutors’ background knowledge of the structures and orderly progress of interactions. These four components consist of the interlocutors’ mutual knowledge of language codes, of the principles or norms of language use, of the situation, and of the world (cf. Fairclough 1995: 33). Since this background knowledge is never shared to the exact same degree by the interlocutors, and since they all have different personal and discursive histories, there can be neither an immediate nor an identical understanding of the topics discussed in conversation. It is this difference, in its various degrees, based on only broadly shared patterns of language, society, and culture that brings spoken interaction and written texts to life. Difference also provides the motivation for interaction in which it is the motor for the collaborative effort to co-construct meaning on subjective, group, and social levels. Traditional approaches to analyzing verbal interaction typically fall short of embracing this wider sphere of influence; rather, they are concerned with the analysis of language in its usage at a micro-level.

Some of the shortcomings of CA and CDA were exposed in the late 1990s by two of the leading representatives of these approaches: Emanuel Schlegoff, a theoretician in the CA tradition, and Margaret Wetherell, one of the proponents of CDA. Both highlight the shortcomings of the opposite approach, albeit from the perspective of the rival position. Each accuses the rival approach of “theoretical imperialism” because it allegedly ignores the voice of the subjects of their enquiries and focuses instead on their respective theoretical framework in analyzing the conversation and discourse respectively. Schlegoff alleges that critical discourse analysis,

Margaret Wetherell responds to this critique by pointing out that conversation analysis practices theoretical imperialism in its own way:

This debate between Schlegoff and Wetherell highlights a problem common to all theoretical approaches of analyzing intersubjective interaction. The factors and configurations influencing the interactive behavior of the interlocutors are so manifold and complex (for instance, psychological factors, or the history of intersubjective conversations) that a generalized theoretical framework for analyzing these influences must always fall short of adequately capturing them in their complexity and interplay.

In a broader sense, there are also other relevant aspects for the analysis of interaction which are not sufficiently taken into consideration by CA, DA, or CDA, for instance, nonverbal aspects which may be important for the ongoing interaction. These nonverbal aspects are particularly relevant for language functions where social contact is important, emphasizing not so much what is communicated, but how it is communicated. The means by which this can be conveyed include body language, eye contact, physical distance or proximity, gestures, artifacts (for example, clothes or makeup), and olfactory aspects (such as perspiration or perfumes). These nonverbal aspects of communication can be very subtle, and are frequently subconsciously deployed. In addition, they are highly culture-specific, frequently also specific to sociolects (age groups) and dialects (regional variations), and therefore cannot be deployed universally.

Although the theoretical categorization of language-in-use has certainly helped to make some sense of the messy and dynamic reality of the structures of conversation, some of these analytical approaches have sacrificed the potential richness of meaning in interaction for the sake of a tidy system of rules. At first glance, these rules may be useful for L2 learners, for example, with regard to sociocultural norms of indirectness versus directness of speaking. For instance, in Germany meaning is sometimes conveyed much more directly in conversation than in England; while in Germany it is perfectly acceptable to deploy a direct no in conversation, in England a more indirect utterance such as I am not sure would be more appropriate to use, although both utterances ultimately carry the same meaning. These socioculturally situated and subjectively deployed traditions can be analyzed in CA, DA, or CDA, and can subsequently be used as principles to be mediated in L2 pedagogy.

At second glance, however, there seems to be a tendency in these analytical approaches to discourse that Schlegoff (1997: 167) describes as “theoretical imperialism,” that is, the tendency of the theorist to impose relatively narrow rules on potentially rich conversation (in terms of direction, invocation of social status, shared memories, etc.). These rules may be of some use to L2 pedagogy, but if they fall short of capturing all influencing factors on conversation they may be misconceived. Therefore, it seems necessary to reverse the risk of theoretical imperialism in the analysis of ongoing processes in the construction of meaning in interactive processes. Rather than trying to establish a theoretical framework and then applying it to real life conversations, it would be more viable to focus on the broader situational realization of speech acts or communication during the construction of meaning and then trying to establish a comprehensive understanding of it for the purpose of L2 learning. This approach focuses on momentarily relevant factors influencing particular communicative moves without claiming to construct de-contetxualized theoretical rules for the conduct of discourse.