196 EXTENSION: LINGUISTIC READINGS At first sight, it would appear that glottalisation is indeed beginning to invade the speech of Cardiff working-class females; the average scores for the group as a whole are considerably higher in 1990 than in the two earlier recordings. But there is actually more involved. Tables D1.4 and D1.5 present the percentage of glottalised and other variants of word-final /t/ in the interview style of each of the four working-class girls. There is an apparently inconsistent pattern inasmuch as two informants (Carol and Judy) show a considerable glottalisation increase, whereas the others (Julie and Marilyn) do not. However, social mobility and personal ambition (cf. Douglas-Cowie 1978) must be taken into account if the developing pattern is to prove liable to expla- nation. To do so involves abandoning for the moment quantitative research techniques and going beyond the statistical data to consider the personalities and backgrounds of the individuals concerned. All four subjects spent their childhood and teenage years in Ely (a sprawling coun- cil house estate on the west side of Cardiff notorious for its image of criminality, high unemployment and social deprivation) and all were classified as working class on the basis of their father’s occupation. All four informants had low glottalisation scores both in their first interviews as children (aged approximately 10) and in their teens (aged about 14). From then on, however, a dichotomy emerges, coinciding with widen- ing differences in lifestyles and career prospects. In their interviews, two of the subjects gave considerable evidence of ambitious career plans and determination to improve their living circumstances. Both had indeed, in terms of their own or their Beverley Collins and Inger Mees Table D1.4 Percentage of [t], [v, vt] and ø for pre-consonantal word-final /t/ (interview style) at three points in time: four subjects individually scored Word-final /t/ % [t] % [ V , V t] % ø N tokens Carol 1976 55.2 20.7 24.1 29 1981 47.2 7.6 45.3 53 1990 27.8 37.5 34.7 72 Judy 1976 40.0 20.0 40.0 10 1981 31.8 0.0 68.2 22 1990 14.3 74.3 11.4 35 Julie 1976 24.0 8.0 68.0 25 1981 13.5 11.5 75.0 52 1990 47.6 11.9 40.5 42 Marilyn 1976 37.9 6.9 55.2 29 1981 34.2 10.5 55.3 38 1990 22.9 2.9 74.3 35 GLOTTALISATION IN CARDIFF 197 husband’s occupations, already transferred to a higher socio-economic class. A quote from Judy’s interview will serve as an example: It was a very good job – and if I’d been the type of person – who was prepared to just have a nice cushy little number and just stay there – I could have stayed there for the rest of my life – but er – I – found that it wasn’t fulfilling enough – that – you know – I – I felt I could do something more. Judy is referring to the fact that she already had one successful career behind her as a dental technician and at the time of the interview was almost within sight of achiev- ing her ambition of becoming a fully qualified chiropodist. The other young woman, Carol, had established herself in a career in office accounting while her husband was a qualified chartered accountant; the couple had moved into a new house in a leafy suburb to the north of Cardiff. The remaining two subjects had experienced less in the way of good fortune. Marilyn, who was not in paid employment, was living in temporary rented accommodation in an inner-city area with her young daughter and her lorry driver partner. Julie was a lone parent bringing up her young son. She had remained in Ely, managing to find work as a part-time waitress/barmaid. The tone and content of both of their interviews contrast sharply with those of the two ambitious subjects. There is little evidence of long-term plans to change lifestyle or improve their socio-economic circumstances; Julie’s references to her restaurant duties are indicative of the attitudes of both women: Beverley Collins and Inger Mees Table D1.5 Percentage of [t], [v, vt] and ø for pre-pausal word-final /t/ (interview style) at three points in time: four subjects individually scored Word-final /t/ % [t] % [ V , V t] % ø N tokens Carol 1976 53.3 26.7 20.0 15 1981 38.1 19.0 42.3 21 1990 15.0 60.0 25.0 20 Judy 1976 14.3 14.3 71.4 7 1981 0.0 50.0 50.0 20 1990 10.0 85.0 5.0 20 Julie 1976 30.8 15.4 53.8 13 1981 0.0 20.0 80.0 15 1990 31.8 13.6 54.5 22 Marilyn 1976 11.1 38.9 50.0 18 1981 29.4 0.0 70.5 17 1990 50.0 20.0 30.0 10 198 EXTENSION: LINGUISTIC READINGS I’m just glad I got this now – you know – I’m on my feet again now – I’m working or whatever – a bit of pocket money I call it . . . you haven’t got to bother with the till – they write it down on – what’s on the menu – you know – what they’re having with their meals – so you just pour the drinks – so you haven’t got to work out anything – which is good. This difference in general attitude is reflected in the distribution of the scores for the word-final /t/ variants so that by the time of the 1990 round of interviews (all by now were in their early twenties) the young women also show great divergence in their glottalisation scores: Carol and Judy exhibit a striking increase in glottalisation while the scores of Julie and Marilyn remain consistently low. The changes made by Carol and Judy to their pronunciation reflect a desire to gain higher socioeconomic status – a desire that has led them to differentiate their speech from that of their peers. [. . .] We are left with two questions to consider. Why is Cardiff peculiarly socially sensitive to glottalisation features as compared with the rest of South Wales? And why should what appears at first sight to be simply a spread of such features to working-class speech need more analysis? The significance of glottalisation in our study appears to be related to the wider question of the status of varieties of English in the Cardiff region. Here, as Coupland (1988: 40, 46–51) has pointed out, a perception of accent types exists which is quite different from that found elsewhere in South Wales. Cardiff does not have any prestigious form of Welsh English as an icon. Despite living in the capital city, Cardiffians in general, and working-class Cardiffians in particular, do not enthusi- astically participate in the general revival of Welsh loyalty, nor do they consider themselves part of the new Welsh ethnicity which has been noted by some observers. The 1997 Referendum underlined the relationship between accent divisions and group identity, when districts included in the CE dialect area, i.e. Cardiff, Newport and the Vale of Glamorgan, voted in large numbers against the Welsh Assembly (ironically, it was nevertheless eventually decided to locate the Assembly in Cardiff itself). Most speakers of Welsh English (Types 1 and 2, including GSWE, in our classification above) would appear to regard Welsh accented speech with approval. But, in Cardiff, broad GSWE varieties are regarded with opprobrium. Welsh speech features such as clear /l/, steady-state vowels in goat and face, and ‘lilting’ intona- tion are (in exaggerated form) frequently the butt of ridicule. In working-class Cardiff communities this negativism is likely to be far greater than that reserved for broad Cardiff accents. On the other hand, RP, near-RP and accents of similar type are regarded in Cardiff in much the same way as in England, and do not suffer the extreme nega- tive reactions which have been reported for Type 1 and 2 areas of Wales (Coupland 1988: 51). The attitude of the media is suggestive here: on Cardiff local radio, the presenters overwhelmingly have RP (natural or acquired) accents or mild British regional accents without any Welsh associations. Glottalisation may be attractive to ambitious CE speakers because it represents, at subconscious level, a move away from local Welsh accent characteristics towards more sophisticated and fashionable speech. While we have ourselves shown that glot- talisation actually has considerable time-depth in RP (Collins and Mees 1996), there is no denying the extent of its recent expansion, perhaps under the influence of what has been dubbed ‘Estuary English’ (Rosewarne 1984). Nowadays, glottalisation is Beverley Collins and Inger Mees GLOTTALISATION IN CARDIFF 199 associated with London life, metropolitan fashions and trend-setting attitudes; it is to be heard from royalty as well as rock stars. It is not difficult therefore to account for the fact that in Cardiff and the adjacent area, where many inhabitants appear to look away from Wales towards England, glottal stop is a prestige feature. Apart from glottal- isation, it is perhaps not mere chance that there are other recent linguistic changes in CE, e.g. the change of face and goat to obvious diphthongal glides and the rounding of price, that can be associated with movement away from Welsh-accented speech and towards south-eastern English varieties. Since Labov (1966: 495–6) it has frequently been confirmed that solidly working- class speakers are relatively secure in their speech norms, and make the least effort to change. So it is not surprising that our two non-ambitious subjects have not altered their speech habits. The ambitious working-class women, on the other hand, are eager to strive towards a new class status and to acquire the trappings that go with it. The upwardly mobile Cardiff females can be seen to acquire RP-style glottalisations together with a professional career, a suburban house and a well-qualified partner. Those who lack such aims are also likely to lack glottalised forms. What superficially appears to be a straightforward linguistic change in our real- time study – the spread of a new feature to a South Wales working-class community – turns out to be rather more complex in at least two ways. Firstly, account must be taken of local linguistic attitudes and group identities. Secondly, the working class must not be considered as one amorphous whole, but it must be recognised that persons may shift social class or have ambition to do so. Their linguistic behaviour needs to be analysed accordingly. [. . .] Issues to consider q Is there a phonetic feature in the accent of your own speech community that signals an aspiration towards being identified with another, different speech community? Do you yourself adopt such features in order to be identified with a different group? Ask other people around you what they think of your accent (see also A9 and B9). q Social class is a notoriously difficult concept to define and socio-phoneticians have adopted different ways for defining social class in their studies of phonetic speech features. What are the advantages and disadvantages of taking Collins and Mees’ approach of defining social class on the basis of father’s occupation? Other researchers have classified the social class of their informants by asking the informants themselves and taking this ‘self-definition’ as their social class grouping (Milroy 1980). What advantages and disadvantages can you think of for this alternative approach? How else could social class be defined in order to consider the role it plays in relation to the production of different phonetic variants? You should also read D9 to inform your thinking on this issue, as it presents alternatives to a macro social class approach. q Consider the comments made by Collins and Mees in the final paragraph in the passage above. Do you agree that the situation is more complex as a result of local attitudes and language loyalty (or disloyalty)? What about the social class stra- tification model that is applied both here and commonly in sociolinguistic studies? Are there any other complexities that might need to be taken into account? Beverley Collins and Inger Mees 200 EXTENSION: LINGUISTIC READINGS THE SEARCH FOR UNITS OF MEANING The first part of this book (strands 1 to 5) is organised across the linguistic rank scale from morphology up to discourse. However, this categorisation is merely an analyt- ical convenience: every piece of language consists of elements at most or all of these levels all at once. Furthermore, the categories of sounds, words, phrases, clauses, and so on, are not necessarily easily contained or simple to define. For almost every definition of a word, for example, a counter-example or difficult case can usually be imagined. In this extract, John Sinclair argues for a phrasal approach to the definition of a lexical item (a ‘word’). His procedure draws on the use of a concordancing program run on a large corpus of examples of English usage, and he is particularly careful to proceed step by step through his method and argument. John Sinclair (reprinted from Trust the Text: Language, Corpus and Discourse (2004), London: Routledge) The starting point of the description of meaning in language is the word. This is one of two primitives in language form, the other being the sentence. The sentence is the unit that aligns grammar and discourse, and the word is the unit that aligns gram- mar and vocabulary. [. . .] A glance at any dictionary will confirm the status of the word as the primary unit of lexical meaning. A dictionary lists the words of a language and alongside each one provides an account of the meaning or meanings. Since the common words of a lan- guage typically can have several meanings, these are usually listed in separate para- graphs [. . .] The model is clear – words are the units of language but are prone to multiple ambiguity. The phenomenon attracts a great deal of academic activity, because it has to be accounted for. Most of the explanations are historical, and show the way word forms can coalesce in time, and meanings can specialize and diverge. Theories of meaning arise, with concepts such as ‘core meaning’ (Carter 1987), and scientific experiments are conducted with the aim of providing evidence to support the theories. Dictionaries, however, also show that the equation ‘word = unit of meaning’, while reliable in general, has to be qualified in a few cases. Compounds, for example, typi- cally consist of two words, each of which has an independent existence, but together they make a meaning that is different from the normal putting together of their indi- vidual meanings. Blackbird is the usual example; a blackbird is a black bird, but not all black birds are blackbirds. In addition, the bigger dictionaries often include a few paragraphs at the end of the entry where a number of idiomatic phrases are listed, with explanations to show that these also claim a meaning in combination that they do not have in simple concatenation [. . .]. The low prominence of these features, and the almost total absence of provision for them in the grammar, makes it clear that they are considered as marginal phenomena, almost aberrations, exceptions that prove the rules. I say ‘almost total absence’ because although the traditional parsing and analysis was quite pure in this respect, the business of language teaching has brought into promin- ence one type of combination in English that is so common it cannot really be ignored. D2 John Sinclair THE SEARCH FOR UNITS OF MEANING 201 This is the phrasal verb, the verb plus particle that conjures up an unpredictable meaning; the scourge of the learner. The structure does not fit the model, neither semantically nor grammatically, because a single meaning-selection straddles a major structural boundary. As a result, dictionaries for the learner usually make special provision for phrasal verbs, and grammars for learners make apologies for their very existence. Besides compounds and phrasal verbs we can mention idioms, fixed phrases, vari- able phrases, clichés, proverbs, and many technical terms and much jargon, as exam- ples of recognized patterns where the independence of the word is compromised in some way. In conventional descriptions of a language, whether lexical or grammatical, they are tucked away, well off-centre. They seem to be anarchic, individual, unstable, one-off items that just do not fit into a tidy description. Unlike phrases and clauses which fit together in Chinese boxes with labelled bracketing, these spill out all over the place, fit no hierarchical place, and relate in mysterious ways to word meaning. Sometimes the criterion given for identifying phrasal verbs, idioms, etc., is that the meaning is not the same as the sum of the meaning of the constituent words. [. . .] a none of the words may appear to contribute directly to the meaning of the expres- sion (bear on = be relevant to) b some may, while others may not (to beat someone up) c each still seems to mean what it normally means (the rain beats down). This last type is usually called a collocation, a frequent co-occurrence of words; it does not have a profound effect on the individual meanings of the words, but there is usually at least a slight effect on the meaning, if only to select or confirm the meaning appropriate to the collocation, which may not be the most common meaning. So in ‘the rain beats down’, the meaning of ‘beat’ is ‘[to hit] hard, usually several times or continuously for a period’ (Cobuild 1995). It is thus clear that there are many cases in texts where the independence of the choice of words is compromised, because other patterns cut across them and con- strain them. [. . .] One hypothesis, to be explored [here], is that the notion of a linguistic item can be extended, at least for English, so that units of meaning are expected to be largely phrasal. [. . .] The idea of a word carrying meaning on its own would be relegated to the margins of linguistic interest, in the enumeration of flora and fauna for example. Part of the supporting argument for this hypothesis is that words cannot remain perpetually independent in their patterning unless they are either very rare or spe- cially protected (for example by being technical terms, if indeed that status offers the protection that is often claimed for it). Otherwise, they begin to retain traces of repeated events in their usage, and expectations of events such as collocations arise. This leads to greater regularity of collocation and this in turn offers a platform for specializa- tion of meaning, for example in compounds. Beyond compounds we can see lexical phrases form, phrases which have to be taken as wholes in their contexts for their dis- tinctive meaning to emerge, but which are prone to variation. [. . .] In considering the corpus data, we shall begin in an area of patterning that on intuitional grounds should be relevant – the area of very frequent collocations, idioms, fixed phrases and the like. If we are to find evidence of extended units of John Sinclair . prestigious form of Welsh English as an icon. Despite living in the capital city, Cardiffians in general, and working-class Cardiffians in particular, do not enthusi- astically participate in the general. corpus of examples of English usage, and he is particularly careful to proceed step by step through his method and argument. John Sinclair (reprinted from Trust the Text: Language, Corpus and. her young daughter and her lorry driver partner. Julie was a lone parent bringing up her young son. She had remained in Ely, managing to find work as a part- time waitress/barmaid. The tone and