Chapter 4 Spoken texts In the last two chapters we have been arguing that a text is more than simply a random collection of sentences. We have shown that the parts of a text are interconnected (ie a text is cohesive) and that a text makes some kind of overall sense (ie it is coherent). But the discussion has been limited to written texts only. It's now time to re-assert the fact that texts are not only written, but are also spoken. (Remember that, initially, we defined a text as 'a continuous piece of spoken or written language' (p. 6).) There are a number of reasons for emphasizing the spoken dimension. For a start, language originates in speech, both historically and in terms of an individual's own linguistic development. And most day-to-day language use is spoken. Moreover, from a teaching point of view, most (but not all) learners of a second language are keen to acquire at least a measure of oral fluency, so the study of spoken discourse ought to hold some interest for programme designers, coursebook writers, examiners and, not least, teachers themselves. And finally, there are a number of ways in which the boundary between spoken and written language is rather blurred, which suggests that to discuss one apart from the other may distort our understanding of how speakers and writers create and interpret text. Discovery activity 4.1 Differences between written and spoken texts In the following conversational extract48, four Australian women, all related, are talking about the fact that one of them (Greta) will move into her parents' home while renovations are being done on her own. This leads into a discussion about the way computers become obsolete so quickly and mention is made of a scheme whereby old computers are re-cycled. The speakers are: Joan (74), her daughters, Greta (47) and Claire (41), and her daughter-in-law, Alice (38), who is married to Joan's son, Philip. (The sign L indicates an overlap, ie where one speaker starts speaking before another has finished.) Skim the extract and note at least five features that characterize the text as being an instance of spoken language: 4.1 Computers 1Greta And urn I'll take the I'll take the computer over because I've got my work stuff on computer so. Dad can play with the Internet or something. 2 Claire I think I should give urn Philip my computer and I'll keep the laptop I love it. [laughs] ■ 3 Greta I had, 11 took it into work to see whether they could load you know work stuff up on it. The guy sort of looked at me and said how old is this? And it's about four years old but of course you know in computer terms that's... 4 Alice Ancient. 5 Greta Ancient. So. 63 chapter 4 Spoken texts Oh I've got a laptop that's I don't know how older than Apple. You can't even write Lwhat's you know, the idea that you sort of, people Lbut it's not even worth fixing. Just throw It out. You know. |_By the time they come out by the time they come out they've already improved them. You know. It's just extraordinary. Well there is a big programme that ah people buying up computers you know we we but what d'you do with them you just chuck 'em out. Mmm. No. And there's a... a they're sort of doing... recycling them for use in you know underprivileged areas and third world countries and stuff. Or give them to your relatives. [chuckles] Well. Yeah. I mean um. But it's just it's you know like al! these analogue phones. In in in about four months' time you know they're gone. Loh don't talk to me about analogue phones. Have you got an analogue phone? No no Commentary ■ ■ ■ You probably noted a number of obvious features of spoken language that are not usually present in written language, such as hesitations, false starts, repetitions and incomplete utterances. You may also have noted the frequent use of expressions like know, well, oh and mmm, as well as several instances of vague language such as sort of,... or something, ... and stuff. Perhaps you picked up on the informal register, as in the guy and you just chuck 'em out. And you may even have noticed how many sentences and clauses are connected by the relatively simple connectors and, but and because. What you won't have noticed, but what would of course have been present in the original conversation, is the use of stress and intonation to signal, among other things, what information is given and what information is new. ■ We can categorize these various features of spoken language according to whether they derive principally from a) its spontaneity, b) its interactivity, c) its interpersonality or d) its need to achieve coherence. We'll deal with each of these characteristics in turn. Spontaneity Most speech is produced 'on-line', that is to say in real time and with little or no time for much forward planning. This accounts for a number of characteristic features which we will call performance features. These include • filled pauses: I should give um Philip my computer • repetitions: In in in about four months' time • false starts and backtracking: And there's a... a they're sort of doing... recycling them • incomplete utterances: What's you know, the idea that you sort of, people Another effect of the spontaneity of speech production is its 'one-clause-or-phrase-at-a-time' construction. Rather than being built up in sentence-length 6 Alice 7 Joan 8 Alice 9 Greta 1U Claire 11 Alice 12 Claire 13 Joan 14 Greta 15 Claire 16 Alice 17 Greta 18 Alice 54 Chapter 4 Spoken texts units, speech tends to be produced in smaller 'runs', each run representing a unit of meaning. These runs are tacked on to each other, rather than being embedded inside larger units (as is typically the case with formal writing) and are often linked by the highly frequent conjunctions and, but and so. So, Greta's turn 3 is constructed out of these elements: The guy + sort of + looked at me + and said 4- how old is this? + And it's about four years old + but + of course + you know + in computer terms + that's... ancient This cumulative construction principle tolerates the addition of sentence 'slots' that in written language would be considered ungrammatical. For example, when Greta says I'll take the computer over because I've got my work stuff on computer so., the final so occupies the sentence 'tail': a place reserved for some kind of comment on, or qualification of, what has been said. (The same speaker uses the same tail in turn 5: Ancient. So.) Common tail-slot-fillers include question tags (isn't he?, didn't they?), adverbials that convey speaker attitude (actually, really, quite frankly), vagueness expressions (and that sort of thing) and topic clarifiers, as in It needs a bit of a prod that fire. A similar slot is available at the head of utterances, which is typically used to 'announce' the topic of the utterance that follows, so that Alice might have said: Analogue phones. Don't talk to me about analogue phones. To make on-line production even easier, many of its individual runs consist of 'chunks'. Chunks are multi-word units that behave as if they were single words and typically consist of short formulaic routines that are stored and retrieved in their entirety. In the above extract sort of, of course and you know are typical chunks, but so also is in computer terms, where computer fills a variable slot in the frame in X terms. Likewise, about Xyears old is another example of a frame with a variable slot. Of course, this segmentation into bite-sized chunks not only makes production easier, but it makes processing on the part of the listener easier too. This is a fact that is sometimes forgotten when materials writers write texts for listening practice that are constructed out of sentence-length units, rather than clause- or phrase-length ones. In a well-intentioned attempt to 'tidy up' spoken language, they may, in fact, be making it harder to process. Interactivity Talk of the type represented by the conversation about computers is clearly interactive: the speakers interact by taking turns to speak, keeping silent when others are speaking, interrupting at times and signalling their agreement or amusement by grunts, laughs and chuckles. An obvious instance of this interactivity is the asking and answering of questions, as in: 17 Greta Have you got an analogue phone? 18 Alice No no Speakers also acknowledge their audience by asking and answering their own (rhetorical) questions: 10 Claire ...but what d'you do with them you just chuck 'em out. Another interactive device is the back-channelling Xbax. listeners do in order to register that they are following the speaker's drift, as in: 65 apter 4 Spoken texts 10 Claire ...but what d'you cfo with them you just chuck 'em out. 11 Alice Mtnm. No. Of course, with so many speakers jockeying to have a turn, it's not all plain sailing, and this is reflected in the number of interruptions and overlapping turns which the transcription attempts to capture: 6 Alice You can't even write 7 Joan Lwhat's you know, the idea that you sort of, people 8 Alice Lbut it's not even worth fixing. Just throw it out. You know. 9 Greta LBy the time they come out by the time In order to manage the cut-and-thrust of interactive talk as smoothly as possible, speakers use a number of linguistic devices, called discourse markers, to signal their intentions and to show how what they are going to say, or have just said, is connected to what went before or what is coming up. Discourse markers are not unlike the gestures and devices that drivers make use of in order to indicate their intentions in heavy traffic. Take, for example, turn 15: 15 Claire Well. Yeah. I mean um. But it's just it's you know like all these analogue phones. The speaker/driver starts the turn/manoeuvre by using the markers/indicators well and yeah to provide a fairly non-committal response to what has just been said (Joan's light-hearted suggestion of giving old computers to one's relatives rather than to some charity). This is followed by the marker I mean that signals that some kind of clarification is going to follow; then few? signals that this clarification perhaps contrasts with what has gone before (Joan's flippant remark) and you know appeals to the shared knowledge of the other speakers as a new topic is introduced (analogue phones). These discourse markers signpost the shifts and turns in the on-going interactive progress of the talk. They also have a connecting function identical to that of the cohesive devices that we looked at in written texts. They connect the elements within speaker utterances and across them. Notice, too, that they frequently take the form of multi-word units, or chunks: / mean..., you know... What is not so obvious from the written transcription of this conversation is the way that changes in pitch and emphasis, ie intonation, also serve to signpost the direction and interconnections of the talk. So when Claire says like all these analogue phones the words analogue phones are emphasized by means of a marked step up in pitch, conferring on them the status of new information and a new topic of conversation. Interpersonality Conversation is not simply the exchange of information, but has a strong interpersonal function. That is, it serves to establish and maintain group solidarity. Casual conversation is often punctuated by laughter, or at least chuckles. (Interestingly, people who are speaking tend to laugh more than people who are listening.) Even when speakers disagree, they do it in such a way as not to threaten the face of other speakers. Hence, speakers use hedges, such as yeah but, in order to blunt the force of a disagreement. Or they use vague language in order not to sound too assertive and opinionated: Chapter 4 Spoken texts Claire ... they're sort of doing... recycling them for use in you know underprivileged areas and third world countries and stuff. And they are constantly referring to shared knowledge and appealing for agreement through the use of markers like you know, question tags (such as isn't it? don't you?) and rising intonation. They also demonstrate empathy by completing and repeating each other's utterances, as in this example: 3 Greta ... but of course you know in computer terms that's... 4 Alice Ancient. 5 Greta Ancient. These attempts to harmonize the joint construction of talk are further demonstrated in the frequent use of exaggeration and strongly evaluative language. The choice of ancient rather than simply old is a case in point. The use of evaluative language acts as a means of flagging the speaker's attitude to what is being said, in order to minimize the chance of misunderstanding and also to bring the talk into line with the views of the other speakers: 9 Greta ... by the time they come out they've already improved them. You know. It's just extraordinary. And, a little later in the conversation, Claire and Alice, who have been discussing the uses that mobile phones are put to, comment: Alice Oh they're amazing. Claire Unbelievable. This on-going evaluation of talk can also take the form of swearing and the use of expletives, a distinctive feature of some conversational registers. A (relatively mild) instance occurs a little later in the conversation: Greta It is is just quite extraordinary what is on that on that bloody Internet. Absolutely extraordinary. When conversationalists fail to demonstrate 'high involvement' in any of these ways, they risk being considered cold or even hostile. The character Joe in the British TV comedy The Royle Family is almost catatonically uncooperative49: Dad [struggling for conversation] Antiques Roadshow's on in a minute. Joe Oh aye? Dad Barbara's mother's down, why don't you nip through and say hello? Joe No. [pause] Dad So life treating you all right, is it? Joe Can't complain, [pause] Nice bit of cake. Dave Did you hear the thunder last night? Joe No. Dave Slept right through it then? Joe Must have done. When Joe finally leaves, Dad comments: Bloody heil, he's hard work, ain't he? 67 ipter4 Spoken texts Coherence Conversation that is stripped of those features that result from its spontaneity and those that signal its interactive and interpersonal character, reads fairly flat on the page - not unlike written language. Here, for example, is the beginning of text 4.1 stripped down: Greta I'll take the computer over because I've got my work stuff on computer. Dad can play with the Internet. Claire I should give Philip my computer and I'll keep the laptop. I love it. Greta I took it into work to see whether they could load work stuff up on it. The guy looked at me and said how old is this? It's about four years old but in computer terms that's ancient. Alice I've got a laptop that's olderthan Apple. You can't write but it's not worth fixing. Throw it out. But there is no doubt that, even when stripped of the characteristic features of spoken language, the text makes sense. That is, it is coherent. (In fact, without the distraction of all the urns and ahs, it is possibly more coherent, which is one reason, perhaps, why fictional dialogue, as in novels or drama, usually dispenses with these features. It is also one reason, perhaps, why textbook dialogues are like they are: transcribed conversations, such as 4.1, are notoriously difficult to make sense of - an issue we will return to shortly.) But unlike written text, whose coherence is principally the responsibility of one person, the writer, coherence in the computer conversation is a collaborative enterprise. The speakers co-operate to ensure that what they say is relevant to what has been said before and to the overall purpose of the talk. Failure to co-operate in this joint work is likely to result in conversational breakdown, and the resultant lack of relevance is often a characteristic of the speech of mental patients, as in this instance50: Doctor What is your name? Patient Well, let's say you might have thought you had something from before, but you haven't got it anymore. Doctor I'm going to call you Dean. A less pathological example of conversational incoherence is this transcript of an on-line chat exchange, in which the initiator ('ten') doesn't realize that he is chatting not with a human being, but with a computer program (AOLiza51) that has been designed to simulate conversational interaction: 4.2 ten AOLiza ten AOLiza ten AOLiza ten AOLiza ten AOLiza ten AOLiza how you been? have you asked anyone else? no, should I? why 'no'? i don't know, like to get it from the source, what's up? does that trouble you? does what trouble me? are such questions much on your mind? sorry, missed the last, could u resend? i've told you that apologies are not required. ok. so how are you? why do you ask? Chapter 4 Spoken texts ten just wondering, how's the job? the family? stuff like that. AOLiza have you asked anyone else? ten about what? The 'conversation' continues in this rather bizarre fashion for another 35 moves, before 'ten' says: ten i don't know, i feel like i'm having a conversation with myself. AOLiza don't you really know? ten what are you talking about? AOLiza what comes to mind when you ask that? ten i'm just trying to figure out what this is all about AOLiza do you enjoy being just trying to figure out what this is all about? ten what's that supposed to mean? AOLiza why do you ask? ten good bye AOLiza i'm not sure i understand you fully, ten that makes two of us Discovery activity 4.2 Coherence in talk What is it about the preceding 'conversation' that threatens its coherence and raises doubts in 'ten's' mind? At the same time, how is it that AOLiza does in fact succeed in keeping 'ten' chatting on-line for a total of 70 turns? Commentary ■ ■ ■ There are a number of ways that AOLiza's responses might set alarm bells ringing. For a start, he/she (/it?) doesn't follow the 'script', that is, the conventional way of conducting certain routine exchanges. In the initial greeting, for example, AOLiza departs from the usual script by following the question how you been? with another question, have you asked anyone else? Moreover, several of AOLiza's questions seem to have no obvious connection to what preceded them, as in: ten ...what's up? AOLiza does that trouble you? On the other hand, some of AOLiza's questions do seem to connect with the preceding comment and this is no doubt one reason why 'ten' was persuaded to keep chatting. For example: ten sorry, missed the last, could u resend? AOLiza i've told you that apologies are not required. Because AOLiza has been programmed to respond to the word sorry by making an explicit reference to apologies^ some degree of relevance can be inferred. Likewise, when 'ten' comments: ten i'm just trying to figure out what this is all about AOLiza incorporates his utterance into its own response (albeit somewhat awkwardly): AOLiza do you enjoy being just trying to figure out what this is all about? This technique conveys at least the illusion of relevance and is sufficient incentive for 'ten' to keep on trying. ■ 69 r 4 Spoken texts Relevance In his lectures on logic and conversation, the philosopher Grice proposed a number of conversational maxims (or rules) to which speakers adhere and without which conversation would simply break down. One of these maxims concerns the relation of the speaker's utterance to the 'accepted purpose or direction of the talk-exchange'. Grice sums up this conversational maxim as: Be relevant. This means that, unless given explicit indications to the contrary, speakers assume that each other's utterances both relate to a mutually agreed topic and follow on from one another. Even where the relevance is not explicit, we will attempt to infer it. One very obvious way that speakers signal the relevance of what they are saying is by repeating all or part of what previous speakers have said. AOLiza's technique - of incorporating bits of the other speaker's utterance into its responses - is well attested in studies of the development of child language, as in Adult You do that one Child Now 1 do that one Indeed, amongst themselves, children seem to be able to sustain long conversations that consist entirely of repeating or slightly modifying one another's utterances. In this transcript of two three-year-old twin boys interacting52, the 'conversation' is sustained almost purely through the repetition of three words: Bl you silly you silly B2 no Toby's silly Bl you silly B2 no you silly no not, you silly Bl you silly B2 no not no silly Bl no silly B2 no no you silly you silly (etc, for another 36 turns) Similar, though less prolonged, 'incorporation sequences' are common in adult conversation. Later on in the computer conversation cited earlier, for example, the following sequence occurs: Claire No no I want to buy I've got an analogue I've got an analogue. Greta |_No no she's wants she's got an analogue. She's got an analogue. Alice Oh you've got an analogue as well? Alice and Greta 'echo' Claire's phrase got an analogue in such a way that the exchange is tightly bound together and maximum cohesion is ensured. Repetition serves at least two functions: it binds utterances together, thereby enhancing the sense that speakers are being relevant. It also creates a sense that all participants are in harmony - that they are 'singing to the same hymn sheet' - and thereby supports conversation's interpersonal function. One researcher of conversational repetition concluded, 'Repetition is a resource by which conversationalists together create a discourse, a relationship and a world.'53 Lexical repetition is a good indicator that the speakers are all 'speaking to topic'. This notion of topic is a key one in terms of conferring coherence on talk. It is the Chapter 4 Spoken texts lack of a consistent topic that characterizes the AOLiza chat, such that, at one point, 'ten' comments, 'i'm just trying to figure out what this is all about'. In the computer conversation (text 4.1), there is no doubt as to what the conversation is about. Not only are the words computer and laptop repeated, but also the phrase work stuff and different permutations of o/Jand ancient, all in just six turns: Greta And um I'll take the I'll take the computer over because i've got mv work stuff on computer so. Dad can play with the Internet or something. Claire I think I should give um Philip my computer and I'll keep the laptop I love it. [laughs] Greta I had, 11 took it into work to see whether they could load you know work stuff up on it. The guy sort of looked at me and said how old is this? And it's about four years o|d but of course you know in computer terms that's... Alice Ancient. Greta Ancient. So. Alice Oh I've got a laptop that's I don't know how older than Apple. Here it is clear - just from the words that are repeated - that the speakers are 'speaking to topic'. Of course, topics can change, and there is a good example of that in turn 15 of the same conversation when Claire says: Claire it's you know like all these analogue phones which prompts Alice to resp ond: Alice Oh don't talk to me about analogue phones which, in fact, signals the beginning of a long sequence where they talk about nothing else! This is a good example of how topic shift is co-operatively managed (and another instance, incidentally, of how repetition across turns reinforces the sense of shared purpose). Alongside direct repetition, there are various forms of indirect repetition that also serve to maintain topic consistency and to bind talk together. One of these is the use of lexical chains, such as other words relating to the theme of computers: Internet, Apple, load up; and the use of referring expressions, as in these instances of the pronouns it, they and them: Greta I took ]t into work to see whether they could load you know work stuff up on it. ... by the time thev come out they've already improved them As in written text, conjuncts, such as so, and, but, or, make connections within and across utterances: Claire And there's a... a they're sort of doing... recycling them for use in you know underprivileged areas and third world countries and stuff. Joan Or give them to your relatives. And we have already seen how discourse markers, such as oh and well, signal the speaker's intentions as to the direction that the conversation is taking. All these features, then, contribute to the overall coherence of the talk. 71 pter 4 Spoken texts Macrostructure But there is another, more top-down, way that talk is imbued with sense and that is the way it conforms to certain fairly predictable organizational sequences, or macrostructures, which extend over several turns. We saw how AOLiza failed to adhere to the standard greetings script: ten how you been? AOLiza have you asked anyone else? Predictable two-way exchanges, such as greetings, or saying thank-you, are called adjacency pairs: Al Hi! A2 Hi there. Bl Thanks for that. B2 You're welcome. Three-part exchanges are characteristic of a lot of classroom talk, where they are called IRF (initiate - respond - follow up) exchanges, as in this example: Teacher (initiates) What is the capital of Peru? Student (responds) Lima. Teacher {follows up) Good. In the following joke, the student fails to recognize the script: Teacher What's the protective outer layer of a tree called, Tom? Tom I don't know. Teacher Bark, Tom. Bark! Tom Woof, woof! Even longer predictable sequences characterize transactional talk - that is, talk whose purpose it is to achieve the exchange of goods or information. Discovery activity 4.3 Service encounters Here is a transactional dialogue from a coursebook54. How could you describe its macrostructure? To what extent do you think it is a typical example? 4.3 Assistant Yes? Riaz Could I have a packet of aspirins, please? Assistant Here you are. Anything else? Riaz Have you got any toothbrushes?. Assistant Yes, these are five pounds, and those are seven pounds fifty. Riaz One of those, please. How much is that? Assistant Five pounds eighty for the aspirins and seven pounds fifty for the toothbrush. That's thirteen pounds thirty, please. Riaz Here you are. Thanks a lot. Chapter 4 Spoken texts Commentary ■ ■ ■ The dialogue embodies a number of features of what are called service encounters, such as: Assistant Yes? sale initiation Riaz Could I have a packet of aspirins, please? sale request Assistant Here you are. Anything else? sale compliance Riaz Assistant Have you got any toothbrushes? Yes, these are five pounds and those are seven pounds fifty. sale enquiry Riaz One of those, please. sale request Riaz Assistant How much is that? Five pounds eighty for the aspirins and seven pounds fifty for the toothbrush. That's thirteen pounds thirty, please. sale Riaz Here you are. purchase Thanks a lot. purchase closure What is slightly unusual about the dialogue is that there is no second sale compliance, ie after One of those, please. Research suggests that the sequence sale request and sale compliance form obligatory elements in service encounters and the seller would say something like Here you are or There you go plus Anything else? or Will that be all? Also missing (but not obligatory) is an opening and a closing. These typically consist of greetings, like Good morning and formulaic parting shots like Have a nice day. Depending on how well the seller and buyer know one another, openings and closings may be quite extended and chatty, as in this example, recorded at a supermarket check-out in New Zealand55: [0 = Operator; C = Customer] 0 Good morning. C .. Morning. 0 How are you? C I'm fine thanks. 0 You look well. You look nice. C ... had - had ten days in hospital. 0 Oh, did you? You feeling better? C I've had a new hip put in. 0 Oh, well good for you. As - is - you going well with it? C Yep. 0 Super. C Down to - ah - one crutch. 0 Good for you. C On my right side. Tell me, the Sheba petfood. You've got beef cuts, beef and kidney, but no turkey in... ■ 73 pter 4 Spoken texts Opening and closing As with service encounters, conversation between friends also has its openings and closings. A famous example of the latter is the prolonged closing attributed to the heir to British throne, in a secretly recorded mobile phone conversation with his girlfriend, part of which went like this56: He Don't want to say goodbye. She Neither do I, but we must get some sleep. Bye. He Bye, darling. She Love you. He Bye. She Hopefully taik to you in the morning. He Please. She Bye.! do love you. He Night. She Night. He Night. She Love you forever. He Night. She G'bye. Bye my darling. He Night. and so on for another 24 turns! Story sequences It might seem, though, that apart from openings and closings, casual conversation has no structure at all. However, researchers have identified several organizational features of casual conversation that suggest that it does in fact have predictable macrostructures. One of these features is the regular occurrence of story sequences, story being defined very generally as to include: • a temporal location • specification of participants • a sequence of events • evaluation So, in the conversation about computers quoted earlier, Greta's short account of her interaction with the technician at work constitutes a rudimentary story: Greta I had, 11 took it into work to see whether they could load you know work stuff up on it. The guy sort of looked at me and said how old is this? And it's about four years old but of course you know in computer terms that's... Alice Ancient. Greta Ancient. So. The temporal location is simply in the past, no further specification being given. The participants include herself and the technician. The sequence of events is captured in the main finite verbs: took it in, looked at me, said... The evaluation expresses the speaker's attitude to the story and underscores the point of the story. The use of the word ancient deliberately exaggerates the age of the computer and accounts for the technician's look and question. The point of the story would have been lost if the speaker had simply said: Chapter 4 Spoken texts The guy sort of looked at me and said how old is this? And I said it's about four years old. So. In fact, it could be said that the evaluation (or moral) of this little story continues all the way to the point where the topic of analogue phones is introduced. That is, the story triggers a commentary on the wastefulness of built-in obsolescence and what might be done about it. Speech-in- action Another kind of talk - apart from stories and their commentaries - is what is called speech-in-action. This is talk that focuses on the immediate environment of the speakers, including whatever the speakers themselves are doing. An example occurs a little later in the same conversation from which the extract about computers comes: Greta That fire's died. I've got to get this fire going. Claire Er. Joan You've got the ah bellows there. Claire Good on you brown owl. [laughs] Alice [laughs] Greta I'll just a little bit of extra thing. Can someone just go and get a get a Claire A log? Notice that speech-in-action has a high proportion of deictic language, ie expressions that refer directly (or' point') to the people, things or activities in the immediate environment. For example: that fire... this fire... you... there. And the verbs go and get imply both movement away from the place where the speakers are and movement back to it. Casual conversation, then, consists largely of alternating sequences of story-plus-commentary and speech-in-action, all framed by openings and closings. This kind of organization comprises the loose macrostructure of conversation. Like the rules of a game, it is a shared familiarity with this overall structure, along with a commitment to conversation's largely interpersonal purpose, that provide a safe and predictable framework within which speakers can appear to be spontaneous and creative. Discovery activity 4.4 Features of speech To summarize, here is another extract from the same conversation that Computers comes from. Can you identify ... ... features that owe to its spontaneous construction? ... features that owe to its interactive nature? ... features that serve its interpersonal function? ... features that lend it coherence? 75 er 4 Spoken texts 4.4 Gas barbecue 1 Claire 2 Joan 3 Claire 4 Joan 5 Claire 6 Joan 7 Greta 8 Claire 9 Greta 10 Claire [general "Alice 12 Joan 13 Claire 14 15 Greta Claire I think you should get the gas connected. Oh look. I'll have to wait until Adrian [her husband] passes on to the next stage before I do that. Why? He can't abide the thought of gas. I don't LWhy? He must have been LYou should've heard him you should've heard him carry on the day that we were having the barbie at my place and he said to me oh Greta he said... he said that's gas isn't it your your barbecue. And I said yes. And he said... but everything will taste of gas. [laughs] I said what like just like yours taste of barbie of lighter fluid. Yes. laughter] Like all the best restaurants in the world. Oh there you know cooking with gas is the LOh ! can't stand having to cook on electric stoves. |_No. Hotplates. Oh they're dreadful. Commentary ■ ■ ■ Evidence of spontaneity: • repetitions and false starts, eg I said what like just like yours taste of barbie of lighterfluid. • clause-by-clause, and phrase-by-phrase, construction, eg and he said to me + ohjane-v he said + he said + that's gas + isn'tit + youryourbarbecue. • formulaic language ('chunks'), eg You should've heard him; ... can't abide the thought of...; and he said... and I said... Evidence of interactivity: • questions, e.g. Why? • back-chanelling, eg Greta: / said what like just like yours taste of barbie of lighter fluid. Claire: Yes. • interruptions and overlapping turns, as in turns 12 and 13, for example • use of discourse markers, primarily oh. Interpersonal features: • laughter • hedging language, eg I think you should get the gas connected... • discourse markers that appeal to the listener, e.g. oh look, you know • evaluative language, e.g. I can't stand... they're dreadful Chapter 4 Spoken texts Coherence features: topic consistency, signalled by • lexical repetition, eg gas, barbie/barbecue, cook/cooking • lexical chains, eg gas —> barbecue —> lighter fluid —> cooking —» electric stoves —> hotplates • referring expressions, eg Adrian. ..he-- - him\ hotplates... they... • substitution, eg before I do that (ie before I get gas connected) • linkers, eg And I said yes. And he said... macrostructures: • adjacency pairs, specifically question and answer, as in turns 3 and 4 • story structure (turns 7 to 11), including temporal location, participants, events and evaluation. ■ Classroom applications How is spoken language dealt with in the language classroom? Are the features that we have identified adequately represented in classroom materials? Should they be? What kinds of texts and activities are best suited for the development of competence in this area? Discovery activity 4.5 Representation of spoken language Let's start by looking at the way spoken language is typically represented in ELT materials. To what extent, for example, do coursebook dialogues reflect the language features of the Computers extract? Here, for example, is an invented - but I think fairly typical - dialogue on the theme of computers written in 'TEFL-ese'. How does it differ from the Computers conversation (text 4.1)? 4.5 A B A B A B A B A B A B A Kevin, have you got a minute? Sure. How can I help you? Can you have a look at my computer? What's the matter with it? It keeps crashing. How long have you had it? It's only four years old. That's very old for a computer. Can you fix it? I can fix it, but you should get a new one. What can I do with the old one? You could give it to chanty. That's a good idea. 77 iter 4 Spoken texts Commentary ■ ■ ■ On the plus side, the contrived dialogue provides a number of useful expressions, such as Have you got a minute? How can I help you? What's the matter with it? That's a good idea. It also embeds some representative examples of frequently targeted grammar items, such as the modal verbs can, should, could, without these being overly obtrusive or unnatural. In this sense, it lends itself to a variety of classroom applications. What is notably absent from the dialogue, however, is any evidence of the spontaneous and collaborative construction of talk. The turns, all of roughly equal length, alternate with machine-like precision, with no overlaps, interruptions, false starts or repetitions. There are no pauses, filled or unfilled, no incomplete or ungrammatical utterances, and therefore nothing to reflect the consequences of on-line production in real-time. From the point of view of coherence, though, this might not be a bad thing. As we saw above, talk which has been 'tidied up' like this is easier to read on the page. But it is not necessarily easier to listen to: without the pauses, repetitions and false starts, the informational load becomes much more concentrated, and therefore more difficult to process aurally, especially for learners. It's like listening to the spoken equivalent of text messages. As noted earlier, materials writers who write simplified dialogues in the interests of making things easier for learners may sometimes be making them more difficult. Also, language models that lack the characteristic 'stops and starts' (called disfluencies) of spoken language may convey a misleading message to learners (and some teachers), ie that such disfluencies are mistakes and are to be avoided. It may come as a relief to learners to discover that native speakers, too, are tongue-tied at times. There is a good case, then, for exposing them to naturally-occurring, spontaneous (ie unscripted) examples of spoken language. One significant difference between texts 4.1 and 4.4 is that the latter is more transactional than interpersonal. That is to say, it is concerned primarily with 'getting tilings done', rather than with the maintenance of interpersonal relationships (as in Computers). On the whole, casual chat of an interpersonal kind is under-represented in ELT materials, in favour of the more utilitarian exchanges associated with obtaining goods, services and information. There may be good reasons for this, such as the fact that, for many learners, 'getting things done' will be their primary objective in using English. Nevertheless, even 'getting things done' involves a degree of interpersonal 'grooming* and the kind of language associated with social grooming perhaps deserves inclusion, even in the most mundane exchanges. If learners are given no guidance as to how to oil the social cogs, they may end up sounding like the neighbour Joe, in the Royle Family extract on page 67. ■ And guidance need not involve a great song and dance. The invented coursebook dialogue (text 4.4) could be simply and easily improved by the addition of some vague language, some evaluative language, some appeal-type discourse markers, attitude-type tails and backchannel devices. For example: B What's the matter with it? A It keeps sort of crashing. You know. B I see. How long have you had it? A It's only about four years old or so. B Wow. I mean, that's incredibly old for a computer, actually. Chapter 4 Spoken texts It could, of course, be argued that coursebook dialogues are a genre apart (we'll be looking more closely at genre in the next chapter) and that their function is less to replicate the features of spoken language than to contextualize targeted language items in typical but intelligible contexts. Besides, many of the more idiomatic or regionally localized features of spoken language may be of little use or interest to learners whose objective is English as an internationallanguage. What's more, if learners are being exposed to a regular diet of spoken language in their classrooms, through, for example, their interactions with their teacher, then there may be less onus on the coursebook to provide realistic models. Unfortunately, as it happens, this is not always the case. Discovery activity 4.6 Teacher-learner interaction Here, for example, is an extract of actual teacher-learner interaction. In what ways is it similar to - and different from - naturally-occurring conversation of the type exemplified in the Computers extract? 4.6 Phil Collins T OK, look at the last text on the sheet that Cathy gave you OK?... What's it about? ... the last text. 51 The last text... T Who's it about? 52 It's about Phil Collins'life. T Yeah. It's about Phil Collins... erm ... what does Phil Collins do? 52 ... singer SI ... plays drums I think T He's a singer and he...? 53 Plays drums T He's a singer and he plays the drums so he's a...? 54 Drummer, he's a drummer. T OK. Does he sing well? Does he sing well? Is he a good singer? Ss Yes [laughter]... Commentary ■ ■ ■ It should be obvious, from this small extract, that the 'conversation' is very onesided, with the teacher asking all the questions and the students answering them, using the three-part IRF framework that we looked at on page 72. T Who's it about? (initiate) S2 It's about Phi! Collins' life. (respond) T Yeah. (follow up) Several of the questions are in fact less questions than oral gap-fills: T He's a singer and he...? 79 ter4 Spoken texts Moreover, the teacher's questions are all display questions, ie questions that require the learners to display knowledge that is already known by the teacher. Such questions, apart from being rare in naturally-occurring conversations, usually require only one- or two-word answers and therefore provide the respondent with little conversational 'rope' to play with. Compare them with real questions, such as Who's your favourite singer? or Do you play an instrument? ■ In short, the spontaneous, (two-way) interactional and interpersonal features of conversation are almost totally absent from this kind of discourse. This does not disqualify it as a form of discourse in its own right: this kind of'teacher talk' has a long tradition and it serves a very useful pedagogical purpose. But it can no way be considered a valid model for - or practice of - casual conversation. Hence, learners who are exposed only to this kind of interaction might emerge less than fully prepared for the realities of fast-moving, interactive, chat. And it is often 'conversation' that learners nominate when asked what it is they most hope to improve on. How, then, are learners to acquire conversational skills? Discovery activity 4.7 Conversational skills Look at these two activities. Both are targeted at the development of conversational competence. How do they compare? jnversation rm groups of three to five. rite the following topics on cards: ist weekend recent film good restaurant ome sports news n animal story ake more cards by adding two or three yie topics of your choice. .uffle the cards and place them face wn in front of you. tie student takes a card, reads it aloud d the group discusses the topic. Tien the group feels that there is nothing Dre to say about the topic, the next card picked up, and so on. Conversation ^} Listen to this recording of a real conversation and answer these questions: How many speakers are there? Which speaker speaks the most? Which speaker speaks the least? How many turns does this speaker have? Which speaker interrupts another speaker? One speaker tells a story. Put the stages of the story in order: significance of the story time and place solution problem event characters Chapter 4 Spoken texts Commentary ■ ■ ■ The first extract adopts a deep-end approach to the development of speaking skills, plunging the learners into a speaking task to see how they cope. It assumes that many of the characteristic features of spoken language are transferable across languages and that what learners chiefly need is classroom opportunities to activate these features. An over-reliance on deep-end activities, however, ignores the fact that learners may need to know how certain conversational routines, such as introducing a new topic, are realized in English. Expressions like by the way or speaking of which do not necessarily translate directly from the learner's first language and may have to be taught. The second extract, on the other hand, takes a more analytical approach to the teaching of speaking skills, treating speaking as the accumulation of a number of discrete sub-skills - a sort of walk-before-you-can-run approach. What this approach tends to ignore is that many of the speaking sub-skills that are targeted, such as turn-taking and story-telling, are already part of the learner's speaking competence in their first language. They don't need to know that speakers take turns or that stories are staged, so much as how turn-taking and story-telling are realized in English, plus opportunities to practise this skill. ■ So, to sum up, the deep-end approach perhaps credits too much knowledge on the part of the learner and the walk-before-you-can-run approach assumes too little. Some sort of compromise might be in order. Discovery activity 4.8 Task-based learning How does the following activity combine elements of a 'deep-end' approach and of a 'walk-before-you-can-run' approach? Conversation In groups of three, choose one of these Now, listen to a recording of some topics and have a short conversation about people talking about these topics. it. At any point, any student can then a Number the topics in the order that they are introduce another one of the topics. mentioned. fast weekend b Write down any expressions that speakers use a recent film in order to a good restaurant -introduce a new topic some sports news -return to an old topic. an animal story c Study the transcript of the conversation and note down any other useful expressions that you would like to practise. Now, repeat part 1 of this activity, but this time try and use some of the expressions you noted in part 2. 81 iter 4 Spoken texts Commentary ■ ■ ■ This sequence combines both the deep-end plunge and the focused analysis of data. By inviting learners to choose and incorporate features they themselves have noticed in the data, the approach allows for the fact that students will be coming at the task with different abilities and styles. And by both starting and ending with the performance of a speaking task, the approach fits into a task-based learning framework. ■ Classroom applications Combining a deep-end approach with an analytical one is relatively easily achieved simply by selecting and sequencing activities of each type. Thus, in the task you have just looked at, the sequence begins with a deep-end activity (have a short conversation) and is followed by an analytic one (study the transcript). This, in turn, leads back to a re-run of the deep-end activity. Building up lesson sequences that alternate between these two poles ensures that learners are getting the best of both worlds. Other possible activity combinations include: Dialogue re-writing • Listen to a naturally-occurring conversation. • Study a transcript of the conversation, noting particular features, eg backchannelling. • Take a coursebook-type dialogue that lacks these features and re-write it in order to incorporate them. • Rehearse it and then act it out. Transcribing • Have a conversation or discussion in pairs or small groups and record it. • Transcribe the recording, or part of it. • With reference to the transcript, discuss ways in which the conversation could be improved, • Re-do the conversation, incorporating the improvements. Task repetition Learners should be able to move around to do the following activity, either milling-that is, freely moving around the class - or organized into a dyadic circle, that is, one half of the class seated in an inner circle facing the other half, who are seated in an outer circle. At a cue from the teacher, learners change partners. In the dyadic circle they do this by all the learners in one of the circles moving around one place. • In pairs, learners have a conversation about a pre-selected topic, eg last weekend, or a problem that is worrying me. • They change partners and repeat the task, adjusting to the new partner. • They change partners again, and so on. Here, there is no explicit analysis stage, but the effect of repeating the same task allows learners, over successive rehearsals, to pay less attention to what they are saying and more attention to how they are saying it. They may also be picking up useful words and expressions from the other learners that they talk to. So the analysis is working at an implicit, incidental level, but need be no less effective for that. Chapter 4 Spoken texts Conclusion Just as written texts are more than a collection of random sentences, so too are spoken texts more than a collection of random utterances. While many of the distinctive features of spoken language derive from its spontaneous and collaborative construction, this should not be allowed to mask the fact that speakers are as intent on achieving coherence as are writers. Even 'text' so apparently formless and purposeless as casual conversation has shape, structure and sense. Moreover, its sense is co-constructed, each speaker building on what other speakers have said and always keeping the other speakers 'in mind'. The mechanisms which give talk its structure, such as turn-taking, openings and closings, adjacency pairs and macrostructures, and the requirement to make contributions relevant, are common to all cultures and languages, so probably do not to be taught again 'from scratch'. What learners need, on the other hand, is practice at transferring these skills into their second language. It may help them do this if they have exposure to models of spoken language that more closely approximate spontaneous, collaborative talk than is often the case. This may simply involve maximizing opportunities for talk in the classroom. One of the saddest things I heard a student say was, 'Our teacher is very good, but she doesn't talk to us.' 83