Current tendencies in colloquial London speech

Table of Contents

1. Introduction

2. Grammaticalisation

3. Teenage Languae
3.1 COLT – The Bergen Corpus of London Teenage Language
3.2 Features of teenage Language

4. Tendencies of adult language
4.1 Johnny Vaughan's Morning Show
4.2 Findings
4.2.1 Glottalisation
4.2.2 Negative concord
4.2.3 Reported speech
4.2.4 Question tags
4.2.5 Omission of have
4.2.6 gonna, gotta, wanna
4.2.7 Functions of like

5. Results

6. Conclusion

Bibliography

Appendix

1. Introduction

The language of adolescents is considered to play a key role in grammatical change as it is supposed to deviate the most from standard norms of speech, as it shows the most innovations. Stenström, too, agrees with this view: “[Youth language] is where linguistic innovations and new trends tend to make their first appearance, some of them even entering and becoming part of the standard language“ (1997a: 146). Although not all innovations of everyday speech do eventually become the norm of standard speech, the reverse is true: Practically all norms of standard speech did at one time in history begin as deviative innovations.

So youth talk is a big candidate for being a primary source for language change. The question, why languages change has been asked for centuries (cf. Aitchison: 133). While earlier theories of language change assumed that it primarily happens through the errors in the acquisition process of children, which are henceforth kept, it was shown later that this plays only a minor role in language change. The first one to note this was Roman Jakobson for sound changes: “[D]ie Sprechweise der Kinder kann zwar zur Quelle oder zum Medium einer Sprachveränderung werden, doch maßgebend für die letztere bleibt die Nachfrage des Sprachgebildes nach der bezüglichen Mutation“ (Jakobson: 16). So the language of younger speakers may be the motive of language change, but not the deeper cause.

Since not all children show the same innovations when acquiring their mother tongue, peculiarities of a single child will only eventually become part of adult speech if other people adopt it. If this is not the case, the children will either stick to this peculiarity with no particular consequence for the speech community as a whole, or they will – more likely – eventually adopt the norm variant.

The speech variant of the next generation up the ladder is a better candidate for being the source of language change, as the peculiarities of teenagers' speech are not peculiar features of individual speakers, but they exist by convention, which means that at least the speakers of one group must have already agreed upon this specific feature. This makes it a lot easier for any innovation to carry on into adult speech.

Furthermore, it is more likely for youth language features to become standard than that of other varieties because on the one hand, all youths will eventually become adults, while on the other hand, youth language is not a variety restricted to a very specific situation but by definition the default way of speaking for teenagers, so to say.

So unlike other varieties, youth language does apply to a great part of the speech community in general. Therefore, it should be very easy for features of youth language to carry on into adult language by just not being dropped by their speakers.

However, this assumption of teenage language being the main source for language change is of course not self-evident. It is therefore of high interest to have a further look whether such a proposition is really true or not. Therefore, it is good idea to have a closer look at adolescents' as well as adults' everyday speech and compare them with regards to their linguistic innovations. If the above assumption is true, then linguistic innovations should appear later in adult language than in teenagers' language.

This is what will be done in this work. After a short excursus on important aspects of the theory of grammaticalisation, we will have a close look at adolescents' speech and its peculiarities through some works on the COLT study of 1993. We will then examine a small corpus on adults' colloquial language in 2007. In the end, we will compare the results to see what this tells us about language change and the role of teenage language therein.

We will mainly be focused on grammatical, i. e. syntactical or morphological aspects, respectively. Phonetic aspects are for the most part excluded except for the phenomenon called t-glottalisation, which is easy to identify and also very interesting in the context of a work on London English.

2. Grammaticalisation

When it comes to the grammatical aspects of a language, language change is in great parts the process of grammaticalisation. So if we want to have a closer look at grammatical changes in English, we need to clarify the notion of grammaticalisation first.

Generally, grammaticalisation is the process of a lexical item becoming a grammatical one. According to Hopper & Traugott, grammaticalisation processes typically bear a number of the following features (cf. 2 f.):

a) A change of meaning of the item appears in only a very local context.
b) The change is possible because the new meaning is – at least in the context in which it first appears – is already implied in the original meaning.
c) A reanalysis of the different members of a grammatical construction with regard to their specific roles within this construction takes place.
d) This reanalysis is only discoverable, if the new function is generalised, i. e. the item in question is used in a context incompatible with the original meaning.
e) After the reanalysis has taken place, the item being grammaticalised may undergo other changes, for example phonological reduction.
f) Various stages of grammaticalisation coexist.
g) The coexistence of various stages tends to constrain the applicability of newer meanings.
h) The grammaticalised item has had a rather general meaning in the first place.
i) While old meanings – or rather parts of the bundle of semantical features – of the item in question get lost in the process of grammaticalisation, new ones may be acquired.

Such processes typically lead to a chain of grammaticalisation from lexical word to grammatical word to clitic to inflectional affix (cf. Hopper / Traugott: 7), which eventually leads to the loss of the item.^[1] This does not mean, however, that a grammaticalised item, a gram, as Bybee calls it (cf. Bybee: 6), is destined to take the full path of grammaticalisation. The process may simply stop at some point.

The fact that grammaticalisation only applies to linguistic items which already have a very general meaning also implies that it does not have to be lexical items that are grammaticalised. It may well be that an already existing gram resumes its path of grammaticalisation or that it further grammaticalises on a totally different path.^[2]

So it is safe to say that grammaticalisation is going on if a bundle of the above mentioned features apply. As we will see, this is the case with some of the the nonstandard features of colloquial English discussed here.^[3]

3. Teenage language

Anna-Brita Stenström compared the deviations, i. e. nonstandard grammatical features, she had found in the COLT data with ones that appeared in a study in Reading done in the 1980s, thus comparing the speech of adolescents over a time range of about ten years. Interestingly, although both corpora show a long list of features that are generally regarded as nonstandard, these two lists were actually very much alike. Out of the 24 features Stenström lists for the COLT data, only three do not appear in the Reading corpus, namely an extension of the use of one as an indefinite pronoun, the use of dunnit and the use of “wunnit as an invariant tag“ complementing innit, which exists in both corpora (cf. Stenström, 1997b: 142). On the other hand, two features of the Reading data do not appear in the COLT data, i. e. the use of what as a relative pronoun referring to people^[4] and the use of worse instead of bad.

These results imply that maybe youth language is in fact not more innovative than any other type of language, at least not with regard to grammatical aspects. On the lexical level, this is a whole different matter.^[5]

3.1 COLT – The Bergen Corpus of London Teenage Language

The most extensive study on Teenage language is the COLT project done by the University of Bergen in 1993. The corpus consists of about half a million words recorded in 50 hours of tape recording. The subjects are 31 teenagers between 13 and 17 years living in a variety of London boroughs. It is also part of the BNC, the British National Corpus (cf. http://torvald.aksis.uib.no/colt/).

The usefulness of the COLT study is reflected in the large number of works relying on this corpus. Since it does not make very much sense in this small work to reinvent the wheel and search the corpus all over again, we will rely on some of these works regarding some innovative aspects of youth language, especially by Anna-Brita Stenström, who is responsible for the project at Bergen University (cf. http://torvald.aksis.uib.no/colt/).

3.2 Features of teenage language

In her 1997b work on nonstandard grammatical features of London teenagers, Stenström compares the results of the COLT project with that of an earlier study from Cheshire on teenagers in Reading. Stenström notes that “[a] direct comparison between Cheshire's and my findings is not possible“ (Stenström, 1997b: 142). One cannot derive from this any conclusions about a possible development of teenage language between 1982 (Cheshire) and 1993 (COLT), because the two studies do not only differ in the time of their recordings, but also in the region and age range and social class of their subjects – Cheshire's recruits were lower class adolescents between 9 and 17 from Reading. So if there is a big difference between the two studies, this may be due to other factors than just time. However, the opposite is true. Only three of the nonstandard grammatical features shown in the COLT data cannot be found in the Reading study. Reversely, only two of the Reading features do not appear in the COLT data. This is remarkable in that it indicates that there is not too much of a change going on in the grammar of teenagers' speech over time.

The nonstandard grammatical features shown in the COLT corpus are:

– negative concord for lexical verbs as well as auxiliaries
– irregular past tense or participle forms
– past participle for present perfect (i.e. omission of auxiliary have)
– use of ain't, in't, in
– innit and wunnit as invariant tags
– use of dunnit
– irregular tense in conditional sentences
– never as negative preterite marker
– multiple negation
– them as demonstrative pronoun
– nonstandard reflexives
– irregular use of one
– what instead of which
– intrusive s
– nouns of measurement in singular
– simple for complex prepositions
– complex for simple prepositions
– no marking of adverbs
– regularisation of comparatives
– double comparison

(cf. Stenström, 1997b: 142)

New in the sense that these features do not appear in the Cheshire study are only the use of wunnit and dunnit and the nonstandard use of the pronoun one.

Nonstandard use of one refers to expressions such as:^[6]

(1) What cardigan? Your one ?

(2) Yeah but my ones, my platforms are out replacing the pronoun forms of the possessives, e. g. yours or mine. Also after genitives:

(3) my Dad said yeah well my one, no my Dad's one was

An interesting phenomenon is the nonstandard use of what replacing which. While in the COLT data, this replacement may only happen if the antecedent is nonpersonal, there is no such constraint (cf. Stenström, 1997b: 142). So instead of spreading, a nonstandard feature seems to have diminished in this particular case.

Still another case is the use of “invariant tags“ as Stenström calls them (142). In the Reading study, only innit appears. So it seems to be the oldest of such tags.^[7] The reason why I find the term invariant tags a little bit unfortunate is that it is exactly the COLT corpus that suggests that innit is in fact not invariant, as are corresponding tags in other languages, as for example oder in German. Invariant is supposed to suggest that innit, for example, is not only a shortening of isn't it but stands for other forms as well, e. g. isn't he, aren't they or aren't you. Therefore, dunnit is not termed invariant, as it does only appear as a replacement for doesn't it in the COLT data. Nevertheless, the appearance of other tags accompanying innit shows that it is not really invariant, but there might be a new, reduced system developing.

These three tags show different stages of grammaticalisation. While dunnit is just a reduced form of doesn't it:

(4) ...just shows your ignorance dunnit really?

Wunnit, although still restricted to third person,^[8] represents two different tenses, future and past:

(5) That was ages ago though, wunnit

(6) ...be a good night tonight, wunnit

So both wasn't it and won't it melt in this tag. An increasing use of this tag, however, means a restriction of innit to other tenses, especially present tense. This might evolve into a system of alternation between innit and wunnit. In this case, it is clearly misleading to talk about invariant tags.

However, innit is clearly the most grammaticalised of these “invariant“ tags in that it occurs instead of many more standard tags than just isn't it, e.g. isn't he

(7) That man is smart innit ?

or doesn't she

(8) She love her chocolate, innit ?

Even, or maybe especially, quite deviant tags such as were they are replaced by innit

(9) Sam and Fern weren't there innit?

Innit even extends to situations in which standard English does not allow tags at all, e. g. commands:

(10) Don't beat me! Innit ?

In sharp contrast to this wide ranged usage of innit in the early 1990s, the situation of ten years earlier seems to be quite different. Although innit is clearly used in the Reading data, it seems not to have been very noteworthy: “The form innit is only referred to in passing“ (Stenström, 1997b: 143). In the Reading data “only be, have and occasionally do re followed by innit“ (Stenström, 1997a: 145). So in this comparatively short time, innit seems to have gone a long way on its path to grammaticalisation.

Stenström also examines differences between classes. Therefore, she compares the data of two speakers each from the upper middle class and the lower working class. As one might expect, the more stigmatised nonstandard features such as negative concord, the use of ain't as well as them as a demonstrative and multiple negation and especially the use of innit appear with a much higher frequency in the data of the lower working class recruits. However, this does not mean that they are totally absent from the speech of upper middle class teenagers, in fact, all of the aforementioned features do appear at least once in the higher middle class data.

There is another nonstandard grammatical feature that is interestingly not mentioned by Stenström, although it appears in Spillius's article (1996) and Stenström suggests that all “[t]he tendencies reported in Independent on Sunday [...] have been observed in the Bergen Corpus of London Teenage Language“ (Stenström, 1997b: 141). This is the extended use of like. Spillius mentions it:

The traditional British sentential links such as "you know", or the use of "sort of" as three dots in mid-sentence, have been commonly replaced by "like", as in: "I'm like, interested in reading, but not like - that much", or in reported speech: "She said to me, `you idiot', and I'm like, `what did you call me?'"

(Spillius: 1)

According to Spillius, “[t]his can infuriate parents“ (1). What Spillius in his article describes quite vaguely as “three dots in mid-sentence“ is actually a variety usages of like as a pragmatic marker which, according to Gisle Andersen, can be traced back to one core function:

[I]t is a pragmatic marker of non-identical resemblance. The utterance may involve a non-identical resemblance between a linguistically encoded concept and a concept that figures in the speaker's thought, in which case like is capable of facilitating pragmatic inferential processes of enrichment and loosening.

(Andersen: 35)

One very important usage of these is that as a marker of reported speech, which marks the following not only as reported – i. e. not exactly what was said word by word – but as interpretive, meaning that what was really said had the same effect as it would have if the original speaker had said the following.

[...]

^[1] This chain of grammaticalisation bears a striking similarity to the circle of language types: isolating>agglutinating>inflecting>isolating.

^[2] For example the German preposition an, which has further grammaticalised into a superlative marker am. It is also involved in yet another grammaticalisation process into a marker of various aktionsarten, e.g. am Kochen sein, ans Arbeiten bringen.

^[3] See chapters 4.2.3, 4.2.4 and 4.2.7.

^[4] A nonpersonal relative pronoun what, though also nonstandard, does appear in both corpora.

^[5] For a closer look on that discussion, see e. g. Hudson.

^[6] If not mentioned otherwise, all examples in this chapter are taken from Stenström, 1997b.

^[7] Although Stenström seems to suggest this, we cannot say that this tag did already exist in the 1960s, this impression is simply due to a misprint. The article by Alex Spinnius dated on “Sunday, 27 March 1966“ (Stenström, 1997b: 141) is of course from 1997b rather than from 1966. On the contrary, Stenström (1997a) notes that “innit does not occur at all in the London Lund Corpus“ of Spoken English (LLC) from the 1960s (139, 141).

^[8] Unfortunately, Stenström does not mention whether this is supposed to be restricted to only it, to just third person singular or to all instances of third person.