8
Discourse analysis: speech and writing
8.1 Introductory
If you have worked through Part B of this book, you should by now be equipped to analyse the structure of English sentences. In this chapter and the next we aim to demonstrate how this knowledge of grammar can be applied to the analysis of DISCOURSE, or pieces of language which are bigger than a single sentence.
The term DISCOURSE applies to both spoken and written language (literary and- non-IiteraTy), rn"'taCT
any purpose. In Part B we used mostly invented sentences for analysis, and aid not pay much attention to the context in which they might have occurred. This was necessary in order to focus on the sentence as a unit of language structure, but now we shall expand our homonstoJookat ^mrik^nanguage jn.actual use^
In 1.3.3 we briefly examined the categories of language use which affect language variation for all language users. These categories were TENOR, MODE and DOMAIN. The category of MODE is particularly important because it is related to the distinction betweenspej^Iiim(Lwj|tag^ As we said in 1.3.3, mode 'has to do with the effects of the medium in which the language is transmitted'. The obvious distinction for English is between the auditory and visual medium, that is, between speech and writing.
In 1.1 we made the point that grammar is at least as much the study of speech as the study of writing. jyritjng,kjntrin^ Gorier* tlian~speech, buFeach performs different functions in society, uses different forms, and exhibits different linguistic characteristics. In this chapter we shall compare speech and writing in detail, and our comparison will be illustrated by the analysis of spoken and written discourse.
8.2 Speech and writing: which comes first?
In 1.1 we said that the written language is 'secondary toitsspokerjjform,
133
134
ENGLISH GRAMMAR FOR TODAY
DISCOURSE ANALYSIS: SPEECH AND WRITING
135
which developed first'. However, it should not be secondary in our consideration, for as we shall see, written and spoken language have different functions, different forms and different linguistic characteristics.
In the history of the human race, spoken language certainly canie before writing. Wej^w no evidence of thejj^tence of J^Wnjjsyjsteni _^jtaxJ^fadJ^gfe£g^QuL3ii^J^.whereas we assume that spoken language existed well before_then.J[h the history of individual societies, spoken language also pre-dates written language, and many languages spoken today have no written form. For the mdividual. too, spoken language comes first, since children learn to speak before they learn to write.
However, in societies which do have writing systems (such as the Roman alphabet, used for English and many other languages), the written language is very important from a social and educational point of view. It would be impossible to imagine our own society functioning as we know it without the advantage that writing gives (see 8.3). Literacy is closely associated with civilisation and education. It is no wonder, then, that the_a!ritten language usually has greater social prestige than the spoken language, and more official recognition. Speech is often evaluated socially according to its closeness to the written language, which explains why standard spoken English is probably closer to written English than is any other spoken variety. Written language js.o ft en vie wed_ a s more 'correct' than spoken language, and as more worthy of study. Also, from a legal point of view, written language takes precedence: a written contract, for example, is more binding than a verbal or spoken agreement ._Howeyer,Jr^^
speech and writing are different; we cannot say that one is superior to the other.
8.3   Functions of writing and speech
The social prestige of written language is probably derived from the added functions which a written variety can fulfil for a society.
Writing has the advantage of relative permanence^ wMch allows for record-keeping in _a_ form independent of the memories of those who keep the records. It also allows for communication over a great distance (by letters, newspapers, etc.), and to large numbers simultaneously (by publications of all kinds). The invention of the tape-recorder, the telephone and the radio have helped to overcome the limitations of the spoken language regarding timet distance and numbers, but these are relativelyjecent developments in human history^
Another advantage of written language is that it is not only permanent, but also visible. An important consequence of this is that it can be care-
fully planned and revised by the writer in a way that spoken language cannot. And for the reader, written language can be processed at leisure, with parts of it re-read and others omitted at will. This characteristic of written language promoted the development of literature, and intellectual development in general. Written language makes possible the creation of literary works of art in ways comparable with the creation of paintings or sculpture. It also promotes intellectual development by overcoming the limitations of human memory and allowing the storage of visually accessible knowledge.
Speech, of course, retains functions which writing will never be able to fulfil, such as quick, direct communication with immediate feedback from the addressee. Speech is particularly important in integrating an individual into a social group, and those who cannot speak, even though they may be able to write (e.g. deaf people), often experience severe social isolation. Speech is used far more than writing: speech is an everyday activity for almost everyone, whereas writing may not be.
-JliM-iMiPiLJS^^SiiES.^16 C0JnP^ementary m function, and we cannot say that one is morejmp^tarvtJSn the^ojhgr. Ideally we need to be able to use both appropriately as members of an English-speaking (and writing) society.
8.4   The form of speech and writing
As well as being different in functionjj^ejchjndjwritin^diffeH as a result of the difference of medium. F^aiJiresJiLsjieicJrijyi^ absent in Writing include rhythm, intonation and non-linguistic noises-such as sighs and laughter. Since speech is typically used in a face-to-face situation, it can also be accompanied by non-verbal communication such as gestures and facial expression. None of these features can easily be conveyed by conventional writing systems, and those wishing to represent them have had to devise special transcription systems. Writing, on the other hand, has several features which speech lacks, jnchjdjnjymncji tuation, paragraphing and the capitalisation of letters. Written language can be-spoken probably more easUy than srx)ken language can be written, but features of speech such as intonation have to be introduced by the speaker. Intonation can to some extent be conveyed by punctuation (especially commas, full stops and question-marks), but not completely. The intonation of the sentence I'll take a taxi to the station will differ according to whether the means of transport ('taxi') or the destination ('station') is the most important idea. This will be clear if you try reading the sentence aloud in different ways. The different meanings implied by differences of intonation would be difficult to convey in writing without changing the structure of the sentence.
136
ENGLISH GRAMMAR FOR TODAY
8.5   Linguistic characteristics of speech and writing
Having compared the functions and forms of speech and writing, let us now compare their linguistic characteristics. For the sake of clarity we shall outline the characteristics of 'typical' speech compared with 'typical' writing, though (as we shall see) there is actually some overlap between the two.
INEXPLICrrNESS. As we have said, speech is generally used in face-to-face situations, so that both the auditory and visual media are available. ATTl:esuR,*spe"e'cli can be much less explicit than writing, because (a) extra information is conveyed by 'body language^e.g. facial expressions^ gestures); (b)jtfoe immediate physical environment can be referred to. e.g. by pointing to objects or people; (c) shared knowledge of the participantsl^ ; anT(d)ln"
a conversionJ:here is an opportunity for feedback from the hearer, so that the message can be clarified or repeated. Speech tends to make frequent use of pronouns such as it, this and that, all of which reflect its inexplicitness. Consider the following imaginary conversation (inadequately represented in writing):
A. How did it go?
B. Not too bad. I'm just glad it's over.
A. Was it the last one?
B. Yes, for the time being.
Unless we were participants in this conversation, we could only guess at what it might be about, e.g. an examination or a tooth extraction;
LACK OF CLEAR SENTENCE BOUNDARIES. Related to inexplicitness. in speech is the absence of clearly defined units we can call sentences. In written language, a grammatical sentence as described in 2.2 is expected to begin with a capital letter and end with a full stop, and it is the acceptednorm for people to write in sentences. The same applies to scripted speech, like that of radio news bulletins. But in spontaneous speech, sentences are often difficult to delimit: they may simply be un-finished, because the Jaiowledge of the aMressee_ makes completion .unnecessary, or they maynot be discernable as unitsjgtah. The following is an excerpt from an actual conversation (dashes represent pauses):
and he was saying that erm - you can go to a nightclub in Birmingham — and watch Tony Bennet for about thirty bob -something like this a night with Tony Bennet - have a nice meal in very plushy surroundings very nice warm pleasant -
(Quoted in Crystal (1980 b), transcription simplified)
DISCOURSE ANALYSIS: SPEECH AND WRITING
137
Even with information about intonation, it is difficult to work out how the above could be divided into sentences. In particular, what is the status of the last group of words, very nice warm pleasant"! We cannot give a definite answer to this question, butjthas been suggested (by Crystal in the above-quoted article) that the clause may be a more appropriate unit for analysing speech than the sentence. In context, the absence of clear sentence boundaries does not mean that conversation is difficult to follow; it just shows that the conversation is organised in a different way from writing.
SIMPLE STRUCTURE. In general, speech is more simple in grammatical structure! than .writing. However, when we use terms like 'simple' and 'complex', we must be careful to explain what they mean. There is no one measure for complexity of structure, but the following measures, when combined, should be useful:
(a) Clause structure. How many elements do the clauses contain, and how many levels of subordination are there? A sentence with the structure [S P O], like (1), will be less complex than one with the structure [A SPOOA),like(2):
(1) [John read a book]
(2) [Last night John gave me a book [when he came home] ].
But the second sentence not only has more clause elements, it also has more subordination, since there is a subordinate adverbial clause, when he came home. The following sentence has even more subordination:
(3) [Last night John gave back the book [he had borrowed [when he had last seen me] ] ].
While the number of elements in the clause can be seen as 'horizontal'
complexity, subordination can be seen as 'vertical' complexity.
(b) Noun phrase structure. How many modifiers do they contain (horizontal complexity), and how many subordinate phrases (vertical complexity)? A noun phrase such as a book is clearly less complex than an interesting book about grammar on the table in the kitchen. The second noun phrase is horizontally complex; consisting of two pre-modifiers, a head, and two postmodifiers. It is also vertically complex because the prepositional phrase in the kitchen is contained in another -on the table in the kitchen - and this structure as a whole constitutes one of the postmodifiers of book.
(c) Where is vertical complexity located? In clauses, subordination at the beginning ('left-branching') seems to make for more complexity than embedding at the end ('right-branching'). This can be illustrated by drawing abbreviated tree diagrams of the following two sentences (Figures 8.1 and 8.2):
138 ENGLISH GRAMMAR FOR TODAY
(4) [The man [who is a friend of the woman [who lives in the house [which used to belong to us] ] ] came to see us yesterday].
(5) [Yesterday we saw the man [who is a friend of the woman [who lives in the house [which used to belong to us] ] ] ].
Most people would probably agree that the second sentence is easier to understand, and in that sense simpler. As we have shown, this can be explained in terms of the location of complexity.
Figure 8.1 Figure 8.2
Sentence (4) Sentence (5)
MCI MCI -__|    |__j    I     ^ j
i i
i i i
ii i
RC1 RC1
RC1 RC1
i i RC1 RC1
In noun phrases also, subordination towards the end of the phrase fends to be less complex than subordination at the beginning. So the noun phrasemy'sister's husband's brother's friend seems harder to understand than the friend of the brother of the husband of my sister (though neither is simple).
_Wehave thusoutlined ways of measuring complexity in discourse. In general, we can say that the greater the number of branches in the tree diagram of a sentence, the more complex the sentence will be. Also, naturaEy enough, there is a strong correlation between complexity and length (measured in number of words). Speech is less complex than writing because of the short time"available to produce and process it. lyntfiig,"on"the^{RFfKaMTcihbe re-drafted and re-read."
RJiPETITlVENESS. Because of the lack of permanence of speech, it is
DISCOURSE ANALYSIS: SPEECH AND WRITING 139
more repetitive than writing. Important information has to be repeated since the addressee cannot refer back to what has gone before. This is noticeable, for instance, in the amount of repetition that occurs in tele-vision commercials, and (for that matter) in normal conversation^ _
NORMAL NON-FLUENCY. This results from the unprepared nature of speech and refers to phenomena such as hesitation, unintended repetitions (e.g.//.. .), false starts, fillers (e.g.um, er),GRAMMATICAL BĽEND~s~äriď unfinished "sentences. A blend occurs where a sentence 'swaps horses' (see 11.7), beginning in one way and ending in another; for example, in Would you mind telling me what's the timel the sentence begins as an indirect question, but ends as a direct question. This is slightly different from a 'false start', where a sentence is broken off mid-way as a result of a change of mind; for example, You really ought - well do it your own way. These phenomena are edited out in written language, which consequently appears more fluent. We may also note the apparent fluency of fictional speech that appears in literature.
MONITORING AND INTERACTION FEATURES .These appear in speech, as a result of its use in dialogue, with a physically present addressee, rather than in monologue. MONITORING features indicate the speaker's awareness of the a(Mressee^"prěšCTcěliiaa"řělčUons; and mclude adverbs md'Idwrbl^
features invite the active participation of the addressee, as in questions, imperatives, second-person pronouns, etc. Writing, which is rarely used in dialogue, usually lacks these features.
INFORMALITY. The situations in which speech is used are generally less formal than those in which writing is used. Therefore, the linguistic characteristics of informality (see below in 9.2) generally appear in speech, while those reflecting formality appear in writing.
Table 8.1
	'TYPICAL SPEECH	'TYPICAL WRITING
1.	Inexplicitness \	Explicitness
2.	Lack of clear sentence 1	Clear sentence
	boundaries	boundaries
3.	Simple structure	More complex structure
4.	Repetitiveness	Non-repetitiveness
5.	Normal non-fluency	Fluency
6.	Monitoring features	No monitoring features
7.	Interaction features	No interaction features
8.	Features reflecting informality	Features reflecting formality
140
ENGLISH GRAMMAR FOR TODAY
DISCOURSE ANALYSIS: SPEECH AND WRITING
141
We can summarise the above discussion by listing the characteristics of 'typical' speech and writing (see Table 8.1, previous page).
However, although these are characteristics of 'typical' speech and writing, there is some overlap. For example, a less typical use of speech occurs on the telephone, jwhere the visualjmedium is notavailable, The result is that the language needs to be more explicit. Also, although speech is generally unprepared, it may be prepared for a lecture or debate, and we may expect more fluency as a result. Sometimes speech is pre^ pared word for word in advance, as in the script ofaplay oj^aJ:eJe^^on advertisement. This speech will also appear more fluent than normal.
Conversely .writing can sometimes display the characteristics of speech, as in a personal letter, which may have monitoring and interaction features (e.g. if you know what I mean; What do you think"!). Letters will also have features reflecting a lesser degree of formality than is typical for writing.
So instead of seeing written and spoken language as watertight subcategories of mode, we have to recognise that there is some overlap, depending onjhe ujejjfjthe language. We might think of mode of discourse as a continuum from 'typical' speech to 'typical' writing, with in-between examples. The continuum could be represented as below:
'Typical' speech
'Typical' writing
Conversation in a pub Seminar
Telephone conversation
Personal letter
Job interview
Radio discussion
Television advertisement
Lecture
Sermon
Script of a play Television news Newspaper Business letter This book
8.6   Analysis of spoken and written discourse
The above discussion should have provided the background necessary for an analysis of spoken and written discourse. For this purpose we shall use a transcript of part of an actual conversation about a summer holiday. We shall compare this with an imaginary version written in the form of a personal letter.
Since the following is a transcript of a conversation, there is no punctuation, but vertical lines indicate boundaries of the major units of intonation, and dashes indicate pauses. LaughteT is indicated thus -laughs - and the material in brackets indicates the responses of the person who is not speaking at the time. You may wish to (a) try reading the transcript aloud, and (b) write it out in conventional orthography, using capital letters and punctuation. This may help to identify for you some of the features which characterise it as speech: these features will be discussed below.
(1)    TRANSCRIPT OF CONVERSATION
Line
B.     so what how did you map out your day 1 you had your 1 breakfast in the kitchen | 2
A. we had our breakfast \ (laughs) in the kitchen | - and 3 then we sort of did what we liked 1 and er got ready to 4 go out | (m 1) we usually went out quite soon after 5 that | - erm the children were always up | at the crack of 6. dawn I (m |) with the farmer | - and they went in the 7 milking sheds 1 and helped him feed the pigs | and all 8 this ] you know we didn't see the children | - and er 9 then we used to go outj we - we had super weather | - 10 absolutely super | - and so we went to a beach j usually 11 for er but by about four o'clock it we were hot and we 12 had to come off the beach (m j m ) ) - so we'd generally 13 go for a tea somewhere ] just in case supper was delayed 14 you know j (laughs) laughs and then we'd get back | and 15 the children would go straight back on to the farm |. . . 16
(from Crystal and Davy, 1975, simplified transcription)
(2)    IMAGINARY LETTER VERSION
Line
Dear B, 1
I thought I would write and tell you about our 2
summer holiday, which we spent on a farm. 3
Every day, the children were up at the crack 4
of dawn with the farmer. They went to the milking 5
sheds with him and helped him feed the pigs, so that we 6
barely saw them at all. 7
Then we would have our breakfast in the 8
kitchen. After breakfast, we usually did what we Eked 9
for a short while, and then went out. 10
We had absolutely super weather, and so we 11
usually went to a beach. But by about four o'clock we 12
142
ENGLISH GRAMMAR FOR TODAY
DISCOURSE ANALYSIS: SPEECH AND WRITING 143
were hot and had to come off the beach. Then we'd 13
generally go and have tea somewhere just in case supper 14
was delayed. When we got back, the children would go 15
straight back on the farm ... 16
We shall analyse the above samples in terms of the characteristics listed, in Table 8.1, p. 139, as identifying 'typical' speech and 'typical' writing, and numbered 1-8. However, we should remember that while the conversation is close to 'typical' speech, the letter is some distance on the continuum (see previous section) from 'typical' writing.
1. The inexplicitness of (1) is reflected, for example, in that (line 6), which occurs at some distance from its presumed antecedent, breakfast (see 12.4). Breakfast is repeated for greater explicitness in (2), where there is no chance for the addressee to ask for immediate clarification. Also, and all this in(l) (lines8-9)is a vague, inexplicit reference to the farmer's tasks; it is eliminated in (2). The time sequence is made more explicit in the structure of the discourse of (2) than it is in (1), where events at dawn are mentioned after the description of breakfast. Lack of chronological reference to events is more common in speech than in writing, though it can occur in the latter.
2. If you have attempted a conventional orthographic transcription of (1), you may have had difficulty in deciding where the sentence boundaries should be. For example, do the occurrences of and indicate coordinated clauses, or new sentences? This is difficult to resolve for (1), whereas sentences are clearly marked by capital letters and full stops in (2).
3. The clause structure of (2) is generally more complex than that of (1), especially in its greater use of subordination where (1) uses coordination. Examples of subordinated clauses in (2) are which we spent on a farm (line 3), so that we barely saw them at all (lines 6-7), and when we got back (line 15). The last is embedded at the beginning of the main clause, which is less typical of speech than being embedded at the end. The noun phrase structure of the two samples, however, seems to be about equally complex.
4. The repetitiveness of speech is reflected in (1) in the paraphrasing of we usually went out quite soon after that (line 5) to then we used to go out (line 10), and in the repetition of the children (lines 6 and 9), super (lines 10' and 11) and beach (lines 11 and 13). In (2) there is some repetition, but less than in (1), and no paraphrasing.
5. The non-fluency features in (1) are er, erm, we - we (line 10), usually for er (lines 11-12), and it (line 12). The latter two are false starts, where the speaker changed her mind about what she was going to say. These features do not appear in (2) (except possibly as the deletion of errors in the original handwriting).
6. The monitoring features in (1) are sort of (line 4), you know (lines 9 and 15). There are none in (2), where there is no physically present addressee.
7. An obvious interaction feature in (1) is B's question, which is missing in (2). In (1) we also find responses from B while A is speaking, in the form of m and laughs. In line 15 A's laughter is in direct response to B's.
8. The informality of both (1) and (2) is reflected in the simple (non-parenthetical) structure and non-technical, accessible vocabulary.
In all, we can say that althoughboth (1) and (2) have the same speaker and addressee and cover the same topic, and although they are not too far apart on the continuum between 'typical' speech and 'typical' writing, (1) nevertheless exhibits more of the characteristics typical of speech, and (2) more of those typical of writing. One would expect to find even more characteristics of 'typical' writing in discourse which does not really have a spoken equivalent. See, for example, the legal discourse in Exercise 8 at the end of this chapter.
8.7 Conclusion
In this chapter we have exemplified the category of mode with special reference to spoken and written discourse. We hope to have shown that .speechJind writing are gene^aUy_cjcmpJejne^^ some overlap in their Jtoguistkjdharacteristics, depending on what they . are uiélOórj and" m~ what situation. In fact, the characteristics of spoken arid" written discourse can be better accounted for if we also consider the effect of the two other categories of use, TENOR and DOMAIN. These two categories will be dealt with in the next chapter.
Exercise
Exercise 8
Analyse- the following samples of discourse as in 8.6, paying special attention to the linguistic characteristics which reflect mode.
a. Carriage hereunder is subject to the rules and limitations relating to liability established by the Warsaw Convention unless such carriage is not 'international carriage' as defined by that Convention.
To the extent not in conflict with the foregoing carriage and other services performed by each carrier are subject to: (i) pro-
ttw sí&r!8&B-'J&ňi£2iSlĚtíSÍbĚL
144 ENGLISH GRAMMAR FOR TODAY
visions contained in this ticket, (ii) applicable tariffs, (iii) carrier's conditions of carriage and related regulations which are made part hereof (and are available on application at the offices of the carrier), except in transportation between a place in the United States or Canada and any place outside thereof to which tariffs in force in those countries apply.
(From 'Conditions of Contract', Laker Airways ticket)
b. B. well I remember Dave rang me up about this business | (yes |) of changing to family grouping | - and erm - er you know | it depends on so many things | really | but I have this friend of ours who lives er erm over the other side of Reading I you know |
A. oh yes |-
B. she erm - she teaches somewhere over that side | I don't quite know where | - but she's terribly against it | (is she | ) she's a far more experienced infant teacher than I am | you know | (yes | ) I mean I've only been doing infant teaching for a short while | - but she won't have anything to do with it | because she says | that it puts too much strain on the teacher |
B. I'm sure it does | . . .
(From Crystal and Davy, 1975, transcription simplified; conventions as explained for passage (1) on p. 141).
Discourse analysis: tenor and domain
9.1 Introductory
In the last chapter we looked closely at the effect of MODE on language, and we noticed in particular how speech and writing differ from one another. In this chapter we shall see how other kinds of language differ from one another. Just as we saw how people vary their language according to whether they are speaking or writing, we shall now see how they vary it according to factors such as who they are speaking to, in what situation, and what kind of activity the language is being used for.
This will involve us in investigating the effect oi the categories ol language use which we call TENOR and DOMAIN. Although we are considering the effect of each of the categories separately at first, we must recognise that they have a combined effect on any discourse. We shall illustrate this combined effect in 9.6.
In this chapter, as in the last, we shall be dealing with non-literary language. This is because the next chapter deals in particular with the analysis of literary discourse.
Ill
J WĚfí
|lÍT
1
si
9.2 Tenor
In 1.3.3 we said that tenor 'has to do with the relationship, b^tvregji_a speaker and the addressee(s) in a given situation, and is often characterised by greater or lesser formality'. To illustrate this, if the relationship between the speaker and addressee is distant and official, as in a legal document, the tenor will be FORMAL, whereas if it is close and intimate, as in a conversation between a husband and wife, the tenor will be INFORMAL . These two examples in fact represent extremes of formality, an^ it is more realistic to think pTascjfe
formal at one extreme to the most informal at the other. The tenor of a conversation between a solicitor and a client, lor example, would have an intermediate degree of formality.
Other scales relating to that of formality are those of POLITENESS
145