Milroy and Milroy: Belfast – Change and Variation in an urban vernacular

Milroy and Milroy: Belfast – Change and Variation in an Urban Vernacular

Structure

1. Introduction

1.1. Starting Points

1.2. Networks

2. Findings - Quantification and Analysis of Variables

2.1.Variable 1 (th)

2.2.Variable 2 ( )

2.3.Variable 3 (a)

2.4.Variable 4 ( )

2.5.Variable 5 (ai)

3. General Conclusions

1.Introduction

1.1.Starting Points

-concerned with different kinds of variability observed in the study of Belfast working-class speech + information of data so far been analysed

-three geographically differentiated working-class communities in comparable inner city areas

-initial interest: rise of an urban vernacular speech in relatively young industrial cities

-geographical differences reflect a historical difference (communities as a whole are of different ages)

-concentrated on working-class speech to illuminate the social and attitudinal reasons for the persistently non-standard nature of urban vernacular phonology

- observations tend to justify their hypothesis (based on history and ethnography of the city) that Belfast exhibits a relatively early stage in the development of an urban vernacular

-> It appears that since we are dealing with a relatively early stage of an urban vernacular, we are likely to find clear binary choice systems (probably resulting from dialect contact) in which one item in the pair is recessive and more often found in relatively casual styles or particular groups of speakers. (also neutralisation of distinctions or obliteration of whole phonemic classes are likely to be found)

-changes may or may not move in the direction of what is normally considered to be the standard

1.2.Networks

-Investigation of three communities in the inner city: Protestant area in Ballymacarrett (East), Protestant Hammer area of the Shankill Road (West) and Catholic Clonard area of the Falls Road (West) -> all decayed working-class areas with a high incidence of unemployment and other kinds of social malaise

-some interviews extended to three hours or more, moving from one person to another in the family or group until the quota sample of sixteen persons in each area was filled

-> each quota sample consisted of eight young people (aged 18-25) and eight middle-aged people (aged 42-55), equally divided between males and females

-during research: evidence that the concept of network is of the greatest importance in interpreting results

-dense close-knit networks = extended visiting, corner-hanging and a homogeneous traditional form of employment within the area (reinforce dense multiplex network ties, enhancing male solidarity + segregation of sex roles)

-of their 3 communities only Ballymacarrett is provided with traditional, locally based employment (shipyard), other areas were based on the (now declined) linen industry which left many men unemployed or travel out to work

-the Hammer is in the process of redevelopment -> consequently, the distribution of the linguistic variables in terms of age and sex is much less regular and symmetrical in the Hammer than elsewhere

-degree to which individuals approximate to a vernacular speech norm seems to correlate to the extent to which they participate in close-knit networks -> a close-knit group tends to be homogeneous (loose-knit group is more variable)

-close-knit group can impose behavioural norms of all kinds on its members (like dress, political attitudes or moral standards)

-they made a quantitative statement of relative standardisation of phonology by examining the (a) variable in the speech of their most corrected informant and comparing it with others

-they made a quantitative statement of relative standardisation of phonology by examining the (a) variable in the speech of their most corrected informant and comparing it with others

-MC speakers tend to converge on (a) with a little variation in vowel-height or backing

-score/indicy: score of zero assigned for (a), two points for front-raised (æ) and for back ( ) and four points for ( ) or closer and also for back-raised and rounded ( )

-MC speaker scores about 0,5 whereas a vernacular speaker scores 2,0 or more

-most corrected speaker was a Clonard woman of mixed Protestant-Catholic parentage with an employed husband -> score of 0,45

-speakers with close-knit networks showed most markedly the typical allophonic variation of the vernacular and scored 2,0 or more -> all possible deviations from standardised (a) were exhibited

2.Findings – Quantification and Analysis of Variables

-quantified data for 7 variables: (th), ( ), (a), ( ), ( ), (t) and (ai) -> Milroy and Milroy reported about 5 (th), ( ), (a), ( ) and (ai)

-(th) and ( ) as simple binaries, scored as percentages of occurrence or non-occurrence of medial (ð) or of ( )

-tokens listed in columns representing degrees of frontness, centrality or backness and height (for vowels) as well as presence of dentality, glottalisation etc. (for consonants)

-tokens are recoverable = valid generalisations about allophonic constraints can be made, also environments that are particularly variable as well as innovation in one particular sex age group

2.1. Variable 1 (th)

- Vernacular speakers variably delete the interdental voiced fricative intervocalically (together, mother, other …) -> avoided by people outside the subculture

- (th) not as conscious symbol of the vernacular for those within the community

- in formal reading style nearly 100% occurrence of (ð), nobody remarked on (ð) deletion in discussions of sleng (Belfast vernacular)

- very clear sex difference -> in none of the communities do the male and female scores overlap -> men who delete least, still delete more than women who delete most

- no evidence of shift taking place

- extremely symmetrical character of (th) data, lack of any evidence of change

2.2.Variable 2

-alternative realisations in Belfast vernacular

-Clonard has a reversed sex pattern: older women rather than men who make extensive use of ( ) -> has conscious vernacular prestige

-two Hammer men remarked that they used less sleng as they got older

-in contrast, Clonard women show a high incidence of this variant which carries covert prestige -> young Clonard women appear to introduce changes in the direction of higher ranked communities

-not witnessing a large-scale phonological shift

-tendency of young men to perceive ( ) variant as carrying vernacular prestige will lead to permanent systematic linguistic change

2.3.Variable 3 (a)

-vernacular vowel may be considerably fronted and raised before velar consonants (bag [bg]), variably backed and backraised in other environments often with some rounding or centring glide (man [m: n)

-front-raising appears to be regressive + indications that trend towards backing is in progress

-some kind of sex grading in all areas; Ballymacarrett stable sociolinguistic market and clear stylistic differentiation of (a)

It appears that an innovation originating with a stable, highly ranked community within the WC is being spread into another community by a homogeneous group, who simultaneously adopt it as a stylistic marker.

The Clonard (a) norm (with little backing) is relatively close to MC norm.

A change in the form of a tendency to back (a) is spreading through the community, from East to West and highest to lowest ranked WC community and phonetically from front to back.

2.4. Variable 4 (ɛ )

- in sleng, low-mid vowel ( ɛ) is realised in monosyllables closed by a voiceless stop as a slightly retracted vowel near (æ) = -T

- before all other consonants, it is realised as a mid-vowel tending to glide to centre and considerably lengthened

- liquid or nasal + voiceless obstruent –ns, -nt, -ls, -lt = -NT

- gradual tendency to raising correction before –NT is a binary choice system in which either (æ) or the long mid-vowel may be chosen

- great majority of di polysyllables prefer low, short vowel (æ) in almost all environments, whereas (ɛ ) is regular in monosyllables except before –T, -NT

- complex low-mid alternation is markedly sex graded throughout the city

- most men = low vowel (æ); females = tendency to transfer to long mid class

- monosyllables: -T, -NT environments low, short; other environments: mid, long -> ALL di polysyllables low, short

2.5. Variable 5 (ai)

- raised and fronted into (ɛ ~ ei) all environments in some parts of Ulster

- vernacular vowel is characteristically (ei), for some speakers higher and almost monophthongal before obstruents

- three variants: (ei), ( i), ( ) Index: 0=( ), 1=( ), 2=(ei)

- in all three communities casual style is distinguished from more monitored styles by marked fronting and raising

- variation appears to derive from a dialect contact situation -> Mid and West Ulster have a relatively high front nucleus for (ai), whereas Scots have Aitken´s Law

- nucleus of (ai) ranges from high to low and speaker´s perception of what is appropriate, is variable

- variable rule – variable application of Aitken´s Law -> variation may partly account for the less clear stylistic distinctions in Ballymacarrett because this community is not only Protestant but also of Scots heritage

3. General Conclusions

- offer a full description of Belfast working-class speech, its variations and speakers differential evaluations of such variants

- data suggest that changes are taking place in so far as both general and restricted shifts are revealed by age, sex and areal patternings

- very great variability in the vowels -> (a) system ranges mid-front to mid-back, the ( ɛ) system from low-front (short) to mid-front (long), (ai) system ranges from high-front to fully low

- gradual shifts are taking place in degrees of fronting, backing, and raising at least in the case of (a) and (ai)

- dental (t) is receding within general classes and sometimes between two phonemes (u and ʌ)

- general tendency to raise a rather low vowel has worked through a number of environments and now reached environment –NT and disyllables where binary alternation patterned by sex, age and style is now observed

-> Where stylistic variation is observable, the identity of a stigmatised variant is usually clearer than that of any prestige target.

- no evidence that speakers from core working-class areas use middle-class speech as a model in any real sense -> on the contrary, in test it was shown that speakers have difficulty in recognising and identifying middle-class accents

- diachronic model of dialect contact offers a more powerful general explanatory framework for the kind of variation so far observed

Universität Leipzig

Seminar: Language Variation and Change

Dozent: Prof. Dr. Buchstaller

Referentin: Anne Grimmling

02.05.2012