5. CERESOS EVALUADOS EN BAJA CALIFORNIA
5.4 CERESO EL HONGO
The lexicographer still has a tendency to consider the occurrence of a word in print the chief or sole criterion for its inclusion in the dictionary; the gram-marian rarely ventures beyond the safe confines of the sentence, a unit that is of doubtful value in the description of casual speech.
(Svartvik, 1980, p. 167) This is not the place to rehearse the distinctions made in many sources between spoken and written language (for accounts, see Stubbs, 1980, Ch. 2;
McCarthy and Carter, 1994, Ch. 1). The main aim is to draw attention to the fact that it is written discourse with which we are mostly concerned in this section, and in the book as a whole (although Sections 4.6 and 4.7 also report briefly on research into lexis in spoken discourse). In this section we examine with reference to lexis some general differences between speech and writing.
In one of few studies in this area, Ure (1971) undertook research into lexical density in spoken and written texts. An important reason for studying lexical density is that spoken and written texts differ in their relative degrees of lexical density. Put very generally, it is the case that there are more lexical words relative to grammatical words in written text compared with spoken texts. Ure studied 34 spoken texts and 30 written texts with a total length of 21,000 words in each medium. She found that the spoken texts had, on average, a lexical density of under 40 per cent; the written texts had, on average, a lexical density of over 40 per cent (ranging, in fact, from 36 to 57 per cent). One main reason for this difference is that written texts can carry a higher information load than spoken texts. Written texts are permanent and Lexis and discourse 97
can be re-read, if necessary. Spoken texts are more ephemeral, are less planned, and rely more on an immediate physical context for their inter-pretation. Spoken texts must also be more predictable. Since lexical words are generally less predictable than grammatical words it is reasonable to expect that written texts have a higher proportion of lexical words. The lexical density of a text is normally calculated in percentage terms by the following formula:
Lexical Density= number of separate words
total number of words in the text ×100
For example, in a text of a total of 60 words with the presence of 33 separate lexical words, the lexical density would be 55 per cent.
Other studies of lexis along the continua between spoken and written texts include research reported in Crystal (1980a) into the relative distribution of adverbial phrases. These are very common in spoken discourse, especially in casual conversation. It is widely recognized that adverbials are an indeterminate, or at least mixed, class in standard grammar. They can have extralinguistic reference and describe occurrences in the real world (e.g. He walked home quickly); they can convey attitudes (e.g. Fortunately, they lost the match); and they can also act as connectors between one part of a discourse and another (e.g. Anyway, I am sure he will come).
Crystal discusses spoken texts in which adverbials occur in 59 per cent of clauses. Ochs (1979), who distinguishes between ‘planned discourse’ and
‘unplanned discourse’ (roughly, written and spoken), points out the higher density of nouns in spoken discourse. Brown (1982), though she cites no data, supports the general conclusions of Ochs and Crystal. Halliday (1989), however, argues that nominalization is a feature of written discourse.
Other studies have examined the discourse functions of certain items which are more common in spoken than in written discourse. These include items such as well, OK, right, anyway, now and so, which have been given the suitably vague term ‘particle’ largely because such items have an uncertain status within the categories of standard grammar. Such items complicate matters further in that they do not have propositional content and, when occurring in utterance-initial position in spoken discourse, for example, do not have ready counterparts in conventional written text. Their function is often to indicate a boundary between what has gone before and a new stage in the discourse (see Sinclair and Coulthard, 1975, for their function in marking boundaries in classroom discourse). A useful collective term for these items is discourse markers. Some detailed studies of particular discourse markers have been made (see Schiffrin, 1988). For example, well has been quite extensively studied (see Svartvik, 1980; Owen, 1981; Stubbs, 1983, pp. 68–70) for its role in interaction where it has a variable function: it can indicate a break in the discourse; it can preface a modification of assumptions made in the immediately preceding discourse; it can serve to begin to close down a topic
or conversation; or it can signal a change from one topic to another. Lexical items such as this need to be studied in ways that do not rely on traditional grammatical categories which are, in any case, based on written discourse.
(See Owen, 1985 for an analysis of anyway.)
Another feature of lexis which is again more endemic in spoken rather than written discourse, that of vague language. Vague language is an inherent property of language; indeed, arguably, all lexical items have a vague denota-tional range. We are interested here in lexical markers of such vagueness, although we must again recognize that interpretation of degrees of vagueness depends on contextual factors such as whether the discourse is a formal or informal one. Informal spoken contexts usually produce the highest degrees of vagueness.
Examples of vagueness include ‘number approximations’ such as about, approximately, or so, (a)round and or (occurring between two numbers), and what Channell (1993) terms ‘vague category identifiers’ such as and things, anything/something like that, and or something. Examples are:
(1) I’ve bought some oranges and things.
(2) There are about twenty or so people in the room.
(3) Let’s examine two or three examples.
(4) It’s called a piston valve or something like that.
Vague language produces specific effects, and primary among them is a detachment on the part of the producer from the absolute truth of the propositions asserted. But lexical vagueness can also signal a lack of knowledge or a failure to find the required words; and it can convey simply a judgement that, in a certain context, too great a degree of precision would be out of place, or would not be understood by an interlocutor. In (3), for example, specification of a precise number would indicate a markedly systematic and rigorous approach. The presence of number approximation suggests a more informal and relaxed procedure.
Stubbs (1986b) examines further categories of lexical vagueness which he points to as obvious surface markers of detachment. The existence of items such as so-called, so to speak and quote unquote enable language users to distance themselves, often ironically and satirically, from what they are saying. For example:
This Government’s so-called education policy militates against innovation.
In this example, the presence of so-called suspends any commitment to what the words refer to and marks the utterance attitudinally. By contrast, Stubbs points out, items exist which allow a commitment to what is said or which signal to an interlocutor that what follows is to be taken ‘seriously’. Examples would be: strictly speaking, to be accurate and technically. The existence of all Lexis and discourse 99
such lexical markers indicates the extent to which point of view can be encoded, lexically. Reference to ‘what is said’, ‘interlocutor’ and ‘utterance’
also indicates a view that examples of vague language occur more frequently and extensively in spoken discourses, although, ‘technically’, this claim would need to be more extensively attested by reference to corpora of spoken and written data.