Segregation and Variation in 1970s Belfast: A Study of Phonological Variation in Belfast Urban Vernacular.

Sarah McLoughlin, © 2007.

This article will examine English as it was spoken in inner-city Belfast in the 1970s, just after the commencement of the Troubles in 1969. Sporadic outbursts of intercommunity hostility occurred before this date, but the severity of the post-1969 conflict led to increased segregation, as people moved out of ‘mixed’ areas and into clearly demarcated communities. Did this diminished contact between the ethnic communities of Belfast have an effect on the phonology of inner-city speakers? I will address the question of how variation in phonology is linked to social variation, for example, differences in age, sex, and location in the city. I will focus on the inner-city, and will refer to the variety of English spoken there as Belfast Urban Vernacular (BUV), to distinguish it from more standardised varieties spoken in the outer-city and suburbs.

Why urban dialectology?

Why is it so important, when discussing dialect in the UK, to include the cities? The dialectologist is often imagined as someone who records rapidly-disappearing rural forms of language for posterity, the very forms that are being eroded by the spread of the cities. However, the cities themselves quickly develop their own dialects, and although the types of English spoken in urban centres are often among the most stigmatised varieties, even among their own users, urban or suburban dwellers account for 80% of the population of England (J. Milroy 1981: 4). The figure is lower for Northern Ireland, but still significant. According to the last census, carried out in 2001, 34% of the country’s population of 1.6 million live in the Greater Belfast area. The sheer force of numbers makes studying the dialect of Belfast integral to the study of Northern Irish English. It is also easier to study the spread of phonological variants in an area of high population density than it would be in, for example, a rural area. In a city environment, members from different groups may be placed in close spatial proximity to each other, whereas in rural settings they choose to live in different towns and villages. They are also more mobile, and thus less dependent on the norms of their immediate community to shape their language practices. There are however, tightly-knit communities established in some inner-city areas whose members are not mobile, and in this case members will be much more strongly influenced by the linguistic norms of their immediate network, rather than the institutional pressures to use ‘correct’ English. This is a point which I will explore in more detail with reference to the ‘social network theory’ developed using data from 1970s Belfast.

Walt Wolfram and Natalie Schilling Estes argue that all language is pre-programmed for change, including phonological change. There is no reason why the Great Vowel Shift should be regarded any differently from the phonological variation in BUV. The only difference is that one is a completed change and one is in progress (91). These changes spread more rapidly in cities (J Milroy 1981: 5), and are thus easier to observe. The phonological variation occurring in urban Englishes is often a movement towards some supra-local form, i.e. dialect levelling, however this does not mean a movement towards the national standard. For example in their study of the phonology of Tyneside (Newcastle) English, Dominic Watt and Lesley Milroy conclude that the dialect is moving towards a ‘northern standard’. However, this standard is very different from RP, as is ‘Estuary English’, which is becoming a levelled dialect in the south-east of England (Foulkes and Docherty 43). This observation can also be applied to Belfast, which is sufficiently geographically distant from the RP-speaking areas of England that the supposed ‘national standard’ is not regarded as a prestige form.

Why Belfast?

This mental and geographical isolation of Belfast from England provides an opportunity to study changes which are occurring from inside the community, rather than as an attempt to emulate RP. This was reinforced by the effects of sectarian violence. During the period under discussion, there was very little movement into the inner-city communities from outside, as they were regarded as too volatile. Immigration, and immigrant communities, a usual feature of larger cities, and one which can have a linguistic influence on the local dialect, were almost non-existent. In general, Belfast dialect is stigmatised, regarded as inferior and incorrect, and it is the standard variety of Northern Irish English that is taught in schools, associated with the middle and upper classes, and presented as the key to social and economic advancement. BUV has persisted despite these pressures. The static nature of these communities, clearly demarcated, with little movement in or out, makes it much easier to isolate phonological changes and pinpoint where they might be coming from.

The linguistic studies of BUV in the 1970s were carried out at a time when ethnicity was becoming much more of a concern in Belfast. As tensions between the Protestant and Catholic communities grew, it became more necessary than ever to display ‘allegiance’, to know who was part of your community and who was an ‘outsider’. Wolfram and Schilling-Estes claim that periods like this, when language users are concerned with issues of nationality and community, they will become much more aware of their use of language and how this differs from others outside their community. Awareness of dialect difference may be connected with “a growing self- or group- awareness. Thus, members of a particular social group may seize upon language differences as part of consciousness raising” (17). In 1970s Belfast, language variations that had been mere regional markers began to be reformulated as markers of ethnicity. Belfast in this period provides an example of how an urban dialect can persist and undergo change, despite being stigmatised and largely isolated from outside contact.

Ethnic self-identification of BUV speakers

First it may be useful to clarify exactly what I mean in this article by ‘ethnicity’. We are used to regarding ‘ethnicity’ as nearly synonymous with ‘race’, but this meaning is redundant when discussing a country in which 99% of inhabitants identified themselves as ‘white’ in the last census. Kevin McCafferty provides a definition of ethnicity that can be applied to a Northern Irish context in his book, Ethnicity and Language Change, where he quotes from the sociologist T.H. Eriksen:

Ethnicity….should be taken to mean the systematic and enduring social reproduction of basic classificatory differences between categories of people who perceive each other as being culturally discrete (quoted in McCafferty 71).

In other words, if the Protestant community and the Catholic community perceive themselves as being separate and different from each other in terms of religion, politics, national allegiances and cultural practices, and continually reinforce and reinscribe these differences, then they are different ethnicities. Ethnicity, in a Northern Irish context, refers to this complex web of self-definition and group identification, rather than a simple divide based on religious practice.

The patterns of ethnic division in Belfast are largely a result of waves of immigration by different groups, from different areas, at different times. It may seem facetious to discuss the events of several hundred years ago in relation to linguistic behaviour in Belfast in the last thirty years, but as Jonathan Bardon points out in his history of Belfast, the community boundaries have hardly changed since they first developed. Voting patterns are still shaped by the original settlement and immigration patterns, so it seems reasonable to assume that linguistic behaviour in different areas of Belfast will still be strongly influenced by where the people in those areas originally came from (Bardon 12).

Origins of BUV speakers

Belfast was founded at the mouth of the River Lagan by English and Scottish settlers in the early seventeenth century. The crown saw plantation as a means to discourage further rebellion and to diminish the possibility that Catholic Spain might use Ireland as a base to attack England. By the 1660s, Belfast was the most important port in Ulster (Bardon 18). Across the river was the town of Ballymacarrett, which would eventually be absorbed into Belfast City (becoming part of East Belfast). The town was peaceful compared to the rest of the region, as Catholics were a tiny minority in Belfast, too few and too poor to be considered a threat.

This situation changed with the population explosion of the city in the industrial revolution. Belfast became a centre for industry, first for linen weaving, then cotton, then shipbuilding and engineering. The inhabitants of the surrounding countryside were drawn in by the possibility of work in the city and the famine and destitution of the countryside. The population soared from 19,000 in 1801 to 70,447 in 1841 and the city pushed westwards. By 1840, a third of the population was Catholic and tension grew between the two communities as the city grew overcrowded. Segregated working-class districts began to develop in the inner-city as the more prosperous inhabitants of Belfast migrated to the suburbs. The penal laws preventing Catholics from owning large amounts of land mean that the ‘prosperous’ were almost always Protestant. This left the inner-city districts to be populated by the remaining, poorer Protestants and Catholics. BUV as a dialect is a class-specific accent: most of its speakers are working-class.

Belfast developed into two areas, divided by the River Lagan: West Belfast, which was predominantly Catholic and East Belfast (containing the former town of Ballymacarrett), which was predominantly Protestant. It is useful to bear these origins and ethnic divisions in mind as we examine the scholarship published about BUV. Sectarian divisions do not produce variant features, but they may reinforce them. East Belfast speech was heavily influenced by the dialect of the area most of its inhabitants came from, which was the Ulster-Scots of the mainly Protestant counties, Antrim and Down (see map below). The Ulster-Scots dialect was a descendent of the dialect spoken by Scottish settlers, which had been influenced by the surrounding Ulster English. In fact, the geographical position of the city, sandwiched between these two Ulster-Scots areas, would suggest that BUV should show much heavier Ulster-Scots influence than it does. The immigration patterns of the nineteenth-century are the reason it is more like the speech of Mid-Ulster. In this period, most people living in the inner city were recent immigrants (in the 1890s the population grew by a third, 60% of this was due to immigration) mainly from the Lagan Valley area (J. Milroy 1981: 24). As a result, the speech of West Belfast is heavily-influenced by the speech of the Lagan Valley, which in turn was a linguistic descendent of the dialects of English settlers who planted that area, rather than the Scottish. John Harris notes that,

The influx of Catholics from the mid-nineteenth century onwards from south and west Ulster, where the predominant non-Irish influence was English rather than Scots….and an increase in pressures towards standardisation…militated against the maintenance of strongly non-standard Scots forms (Harris 140).

So, to summarise: the phonology of West Belfast, which is a majority Catholic area, is influenced by the dialects of the English who settled in the Lagan Valley, and the phonology of East Belfast, which is a majority Protestant area, is influenced by the dialects of the Scottish who settled in Counties Antrim and Down. However, it is also important to note that ‘more influenced by Ulster-Scots’ and ‘more influenced by Lagan Valley’ are labels to denote small differences in what are very similar forms of language which would sound almost identical to a non-speaker of NIE. It is perhaps more accurate to describe BUV as a continuum along which these features fall, rather than two clear ‘varieties’. Speakers will be aware of all the variants, but will choose to use those which most clearly represent their background or allegiance. As James Milroy puts it,

Dialect is a property of the community, and every native speaker has roughly the same kind of access to it and roughly the same knowledge about it….the speakers..know how use the resources of variation available to them, and they use them for many purposes, including the marking of varying social roles and functions (J Milroy 1992: 75).

In Belfast, these functions include age, sex, location within the city (i.e. West/East) and ethnicity.

The Milroy Study: Communities

James and Lesley Milroy carried out a number of influential studies of Belfast English, the first in 1977 in the inner-city, and a supplementary study in 1980 in two outer-city (and lower-middle class rather than working class) communities. I am focusing on this study as, although thirty years old, it was the last major study of BUV phonology. It is the 1977 study with which I am mostly concerned here, as this focused on BUV rather than the outer-city. Many other linguists have used the data from this fieldwork for their research (for example John Harris or Ellen Douglas-Cowie – see Bibliography), but none have written on it as extensively as the Milroys. In this study, they gathered data from three working-class Belfast districts. Two were in West Belfast (the Clonard - Catholic), and the Hammer - Protestant), and one was in East Belfast (Ballymacarrett - Protestant). Their hypotheses were that the tightly-knit, segregated nature of these inner-city communities would place great pressure on their members to conform to vernacular linguistic norms rather than aspiring to standard usage.

In order to foreground the sociolinguistic conclusions of this study, it is necessary first to outline the ethnic boundaries and social conditions in the communities in question. As a Protestant district in a Protestant part of the city, Ballymacarrett is unproblematic. The situation in West Belfast is more complex. The Clonard is situated on the Falls Road, and the Hammer is situated on the Shankill Road. Both roads extend outwards from the city centre, and are only “a few hundred yards apart” at the inner-city end (J Milroy 1981: 41). At distances this minimal, we would expect there to be little difference between the speech of the two communities, but it is important to remember that the communities are segregated, both physically and culturally. Fredrick Boal’s article, ‘Territoriality on the Shankill-Falls Divide, Belfast’, provides a snapshot of segregation in the area in 1967-8.

Boal notes that these communities are separated not only residentially (which is not unusual as “segregation on the basis of both economic and ethnic characteristics is a feature of most cities”, Boal 59) but also by ‘activity’, that is, as well as living separately members do not interact with each other. It is perfectly possible to live in one of these communities and never encounter members of ‘the other side’, except in neutral areas such as the shopping districts of the city centre. Boal’s study was carried out prior to the outbreak of the ‘troubles’ in 1969, so it is likely this segregation has intensified, with people moving out of previously mixed areas. One physical barrier that was not present at the time of Boal’s original study was the “Peace Line”. Barbed wire fences began to be constructed between the two communities after violent disturbances in 1969. Today these have been replaced by brick or corrugated iron barriers.

The "Peace Line", separating the Catholic and Protestant areas of West Belfast. Photo by Kiernan Joliat, 02 Oct 2004, downloaded 18 Dec 2006 (http://www.pbase.com/sonasgael/image/34627230). Photo governed by Creative Commons Licence.

Methodology of the Study

Previous projects designed to gather Irish vernacular speech had been criticised for their methods. One of these was the Tape-Recorded Survey of Hiberno-English speech (1973-80). Much of this data consists of exchanges between one fieldworker and one informant, in which the fieldworker holds the power, controls the exchanges, sets the agenda and chooses the topics of conversation (Kirk 68). The informant views the fieldworker as an authority figure, and thus self-correction towards standard grammar and phonology (in order to seem ‘correct’) is likely. If self-correction occurs, many instances of vernacular or dialect speech (which is what the study set out to record) will not be evident, and a skewed picture of vernacular speech will be produced.

In designing the methodology of their projects, the Milroys set out to address these problems, and minimise the effect of the fieldworker as much as possible. The researchers relied on mutual acquaintance introductions to gain access to and move through a community over an extended period. This instilled some measure of trust and familiarity on the part of the informants and minimised the self-correction of speech which might have taken place in the presence of a researcher who was regarded as a complete outsider to the community. It also avoided the ‘pre-selection’ of more ‘respectable’ (i.e., standardised) speakers that might have occurred had the researchers been introduced through institutional channels (e.g. by teachers, clergymen or social workers) (J Milroy 1981: 90).

The study consisted of a combination of interviews, reading of word-lists, and recording of unprompted discussion among informants. This was intended to capture both formal, interview style (closer to the standard) and a more spontaneous style (tended to be more vernacular). These were labelled ‘IS’ (Interview Style) and ‘SS’ (spontaneous style). Group discussion was intended to encourage each informant to speak as he or she normally would in the presence of the others, i.e. in the vernacular, and to diminish the “perception of the interview as a speech event subject to clear rules” (L. Milroy 1980:25). The speech of eight-middle aged people and eight young people was recorded in each of the three areas (Ballymacarrett, Clonard, Hammer), as well as a random sample of forty-three households throughout the city, producing a corpus of over one hundred participants. Additional constraints on the fieldwork methodology were necessary because of what Lesley Milroy describes as the “generally disturbed situation” in the city at the time of the research. As well as forming a connection to the communities in order to establish trust, the fieldworker had to be a woman and she had to enter the communities alone.

Women were much less likely to be attacked than men, and since male strangers were at the time viewed with considerable suspicion in many parts of Belfast, they were likely to be in some danger if they visited one place over a protracted period. (L. Milroy 1980:44).

Social Network Theory

One of the questions the Milroys attempted to address in their research was why BUV persisted against these standardising pressures. The solution they presented to this problem was that speakers of BUV were experiencing pressures from inside their own community which operated in the opposite direction from the ideas of prestige and social improvement that were being imposed from above. Standardising pressures from the education system and the media were being opposed by counter-prestige that favoured and enforced vernacular norms. They called this idea ‘social network theory’.

Pressure to maintain the vernacular is likely to be strongest in small communities where almost everyone knows everyone else (a ‘dense’ network), and knows each person in the network in several capacities (a ‘multiplex’ network). For example, person A knows person B not only as a friend, but works with him, is related to him by marriage and lives on the same street. Under these social pressures, speakers will prefer the solidarity expressed by use of the vernacular, to the status they might gain by using a more standardised form.

City life is particularly conducive to the formation of strong networks, as new residents experience an urge to gravitate towards and form ties with those similar to them. The close-knit network is a form of protection (A. Giddens, The Constitution of Society, 1984 quoted in J. Milroy and L. Milroy 1992: 7).We might expect this urge to be even stronger in a city such as Belfast, in which two antagonistic communities live at close quarters, each feeling threatened by the other. In fact, the anthropologist Thomas Horjup argues that some degree of insecurity is necessary in order for strong networks to be formed,

The solidarity ethic would collapse and network ties become weaker if economic and political conditions allowed workers to feel secure (T Horjup, quoted in J. Milroy and L. Milroy, 1992: 20).

Dense and multiplex networks are most likely to form in working-class areas with little turnover in population, where the inhabitants work together and have ties of kinship L. Milroy and S. Margrain 48). These factors were taken into account as part of the Milroys’ research, and each informant was assigned a ‘network strength score’ from 0-5 which measured their integration into local networks. A score of zero represented a person with few strong ties to the local community, whereas a score of 5 represented someone whose networks within the community were dense and multiplex (L. Milroy 1980: 139). It is easier for “normative consensus” to be imposed on speakers who are members of dense, multiplex networks, so these networks tend to maintain the vernacular and impede linguistic change, whereas weak networks encourage change and the abandonment or alteration of vernacular language features (L. Milroy and S. Margrain 48, J. Milroy 1991: 76). In a dense, multiplex community, using the vernacular is construed as a symbol of community loyalty, whereas using standard forms is considered self-aggrandizing, artificial, and a way of disavowing association with the community and its values.

Conclusions of the Milroys’ research: phonological variation in Belfast

Now that I have outlined the methodology and theory behind the research, I will discuss several phonological features which the Milroys identified as subject to variation, and question what social trends can be seen as responsible for these changes in phonology.

The effects of gender roles and community loyalty on BUV phonology

Ellen Douglas-Cowie proposes that although both sexes may have the same language resources in their linguistic repertoire, women engage in more marked style-shifting and avoid using the vernacular in formal situations. Male speakers are more likely to cleave to vernacular forms. From this, Douglas-Cowie suggests that we can categorise women as more innovative in adopting changes in language, whereas men are more conservative in their language habits (543). Of course this does not apply to every individual man and woman, but it is visible as a general trend when the speech of large numbers of people is analysed. We would expect from this line of thought, that linguistic change in Belfast would be spearheaded by women.

One long-standing feature of Belfast dialect where the differences between male and female linguistic behaviour can be observed is the distinctive BUV rounded vowel. Vowel rounding means that the standard Northern Irish English [ü], becomes

[ʌ], causing words such as ‘pull’ to be pronounced to rhyme with ‘dull’, rather than ‘pool’, and words such as ‘shook’ to be pronounced to rhyme with ‘luck’. David Patterson mentions words with this variable in his pamphlet; The provincialisms of Belfast pointed out and corrected (1860). Although it is a feature that originally entered Belfast speech from rural dialect, this pronunciation was actually more frequent among young men than older speakers at the time of the Milroys’ study. James Milroy suggests that they “feel it is a prominent marker of Belfast dialect, and they use it as a symbol of community loyalty” (J. Milroy 1981: 30). Perhaps the need to express allegiance to an inner-city community was strengthened by the severe disruption caused by the Troubles (which started in 1969) to these communities.

The Milroys gathered data for four groups: men 40-55, women 40-55, men 18-25 and women 18-25. In the older age group, women use the ‘dull’ pronunciation more than men, in both Interview Style (IS) and Spontaneous Style (SS), however in the 18-25 age group men use the dialectal variant much more than women. This gender difference is particularly strong in SS, with only 20% of utterances by women aged 18-25 using the ‘dull’ pronunciation, compared to 61% of those by men in this age group (J. Milroy 1981: 95). So, what was seen as a stigmatised dialectal variant has become a marker of community identity, but only for men; why might this be?

James Milroy indicates that both variants may be present in the linguistic repertoire of the informant, and that the speech of the inner-city shows a high incidence of “phonolexical alternation” between possibilities (J, Milroy 1991: 78). This can be demonstrated by referring back to the figures for the PULL variable; in every age and sex grouping, the vernacular pronunciation is used significantly more in SS than in IS. This suggests that informants are capable of realising both variants, and which one they select is the result of a choice based on the formality of the situation, who they are speaking to, who else is present and the topic of conversation. The alternative choices can also be used to send social signals, with the vernacular variant encoding “messages of social nearness” or community with the addressee, whereas the standard variant encodes “social distance” (J Milroy 1991: 79). Women, however, are much more likely to use a variant closer to the standard. Every speaker will ‘correct’ his or her speech towards what is perceived to be ‘standard’, but women will do it to a much greater extent than men. There is no certain explanation for why this occurs, but James Milroy offers a suggestion:

This is a general truth – not confined to Belfast or any single place. It may be interpreted to mean that women are more linguistically aware or that they are better at languages generally: certainly, they do make a greater effort to use the pronunciation they judge appropriate to the circumstances in which they are speaking (J. Milroy 1981: 38).

Another variant which is used much more heavily by men than my women is the deletion of ð, e.g. dropping the ‘th’ sound from words such as ‘mother’ and ‘brother’, so that they become ‘mo’er’ and ‘bro’er’. In the inner-city areas used in the study, virtually everyone uses this variant to a greater or lesser extent, but as with the PULL variable, men use the vernacular variant more often than women in all age groups and styles, and the younger men use it more than the older men. If we looked solely at the difference between usage by young men and older men, we might conclude that use of the vernacular form was spreading, and that today, almost thirty years after the study was carried out, dropping of ð would be higher still if we were to test it.

However, Milroy points out that by looking at the female scores, we can tell that ‘th’-dropping is actually stable overall in the community: younger women use it almost as much as older women. He suggests that the increased use of the vernacular variant among young men, is a consequence of this group’s social attitudes. They admire the ‘roughness’ of the vernacular and wish to use it, but as they grow older and “begin to define their role in society differently” (e.g. as they get married, become householders and are fully employed) they shift towards using more standard forms; in this case, pronouncing ‘th’ (J. Milroy 1981: 36-7). The vernacular pronunciation will always be used more by young men, but it is not spreading to other groups. This feature is very strongly sex-marked: when relevant words were spoken, young men dropped the ‘th’ 80% of the time in SS and 89% in IS, whereas young women only dropped it 47% of the time in SS and 30% in IS.

/A/ backing, is a another phonological feature which is strongly affected by gender roles and community ties. In this feature, the pronunciation of /a/ is moved backwards in the mouth, so for example, ‘hand’, man’ and lad become ‘haund’, ‘maun’ and ‘laud’. At the time of the study, /a/ backing was most prevalent among East Belfast males, however it was spreading to West Belfast. The discovery that men use the vernacular variant more is not surprising. Ballymacarrett is an area that was heavily influenced by the shipyards. In this community unemployment was low, gender roles were clearly differentiated (hence the corresponding lack of use of backed /a/ by women) and most families had been in the area for generations. Most of the men in the area were employed in shipbuilding, and according to the social network theory these conditions are exactly those which tend to create strong, multiplex networks (the men lived in the same area and worked together). The stronger and more multiplex the network, the more pressure to maintain vernacular forms. Backing of /a/ was being used to encode meanings of informality, social proximity and male identity. One question worth addressing in any future studies of Ballymacarrett is whether the decline of the shipyards as an employer in the area (from over 30,000 employees in the 1970s in to 130 in 2003, BBC News) has led to a weakening of networks and an erosion of vernacular language features.

The Spread of Vernacular Features across the City.

The figures provided by John Harris (221) show that among older men, ‘th’ dropping is much more prevalent in Ballymacarrett. Here 89% dropping occurs in SS, as opposed to 56% in the Hammer and 62% in the Clonard. On the basis of these figures, we might conclude that it is a feature of East Belfast dialect, however among young men it is equally prevalent in all three areas. This suggests that the feature started in the East of the city and has spread from East to West. The situation becomes more interesting when we compare female scores across the three areas. Although scores for young women are lower than those for young men in all cases, the scores for young women in the Clonard are significantly higher than those for young women in Ballymacarrett (SS: B 15%, H 57%, C 70%). This confuses the theory of ð deletion as a clearly East Belfast feature. What we can say is that it is a feature that in East Belfast is heavily used by men, and it has spread to young West Belfast men and women, with disproportionately high figures among young women in the Clonard. Might this be because young women were the group who introduced the feature to West Belfast, and why might they be particularly likely to be linguistic innovators? I will return to this question in more detail after a discussion of whether language features vary based on ethnicity.

Phonology and Ethnicity – can they be linked?

One of the features that is often identified as Catholic is palatalisation, i.e. pronunciation of a ‘y’ sound after initial ‘k’ and ‘g’ sounds, i.e. ‘car’ is pronounced as ‘kyar’, and gas as ‘gyas’. However, Milroy argues that the origin of the speech community is more important in determining whether speakers palatalise or not than the ethnicity of the speaker. In Northern Ireland, some areas have a majority of Catholics and some have majority of Protestants. As I mentioned earlier in this article, the Ulster-Scots, from which the people in East Belfast mainly came, are heavily Protestant, so Ulster-Scots features will lead a listener who is aware of these regional divides to hypothesise that the speaker is Protestant. Likewise, West Belfast was mainly populated from the Lagan Valley and west and south Ulster, where there are more Catholics and palatalisation is more common. Thus, palatalisation might lead a listener to guess that the speaker is Catholic. Linguistic clues can be used to guess at ethnicity, but this is not to say that ethnicity produces phonology, simply that knowing the phonology of different regions, and the ethnic make-up of those regions, can lead the listener to an educated guess. As James Milroy puts it,

There is a fairly high probability that you will be right if you identify a South Armagh speaker as a Catholic and a North Down speaker as a Protestant, but this is obviously only a probability and not a certainty (J Milroy 1981: 42).

A minority Catholic speaker from an Ulster-Scots area will still use Ulster-Scots phonology, although this is associated with Protestantism.

To return to Belfast, on the basis of the immigration patterns that formed the areas we would predict that East Belfast pronunciations would be identified as Ulster-Scots-influenced and therefore Protestant, whereas West Belfast pronunciations would be identified as influenced by west and south Ulster, and therefore more likely to be Catholic. However, it is necessary to remember that there are communities of Protestants living in West Belfast, the Hammer being one of them. Do these Protestants palatalise or not? Are they more like the Catholics who live in close proximity to them in West Belfast, or like the Protestants across the river?

The answer is that they are somewhere between. Harris notes that palatalisation is a recessive feature that is “now almost entirely restricted to older male speakers” (214). Accordingly, the older men in the Clonard are the ones that use it the most (62% of times when reading a word-list), while Ballymacarrett men in this age group do not use it at all. Men from the Hammer use it much less than their neighbours in the Clonard (14%). There are no figures available to measure whether men in the Hammer have been palatalising less over time, but it is likely that the presence of the “Peace Line” dividing them from the Clonard might have caused divergence between the communities since the time of the research. However, the Hammer men are palatalising, whereas men in Ballymacarrett are not doing it at all, which would seem to suggest that it is a West Belfast feature rather than solely a ‘Catholic’ one. Also, neither the Milroys nor Harris provide any figures for palatalisation in IS. It is possible that the feature, which is stigmatised as rural, might appear more in unguarded, spontaneous speech.

How ‘weak’ ties facilitate language change

I mentioned that the backed /a/ feature originated in Ballymacarrett. Only one group in West Belfast were using backed /a/ more than their counterparts in Ballymacarrett: the young women in the Clonard (J. Milroy 1992: 186). This group of women appeared earlier as possible innovators in connection with the variable ð. Women in this area are actually using the vernacular variant more than young men in the same area. Why might this be? How could young Catholic women be picking up a stereotypically Protestant, male pronunciation when the “Peace Line” divides their communities?

James Milroy’s answer to this is to refer to Mark Granovetter’s theory of how weak ties operate. Granovetter argued that although strong ties maintain vernacular norms within a community, weak ties are a “crucial bridge” between communities. Weak ties connect us to people different from ourselves, outside our immediate community, and are a conduit through which linguistic change may be spread (Granovetter 106, 108). Social conditions in the Clonard and Hammer (which also shows a high incidence of female /a/ backing, although not as high as the Clonard) were quite different at the time of the study from those in Ballymacarrett. There was no cohesive industry, and male unemployment rates were around 35%. Lesley Milroy claims that “women were much more inclined than men to look for work outside the locality”, often earning the family wage. As a result, gender roles were much less clearly marked than those in Ballymacarrett (L. Milroy 1982: 162). The women often found jobs outside the community at city-centre shops, were they were in weak-tie contact with “large numbers of people from all over the city, both Catholic and Protestant” (J. Milroy 1992: 187). East Belfast phonologies picked up from the many weak contacts they might have in their employment could then be spread back into their own community.

Conclusions

These studies demonstrate that in the Belfast of 1977, phonological variants were clearly being used as social markers. To choose a vernacular variant over a standard one was an expression of solidarity with a particular local identity. Which variants speakers used depended on their sex, their age, their location within the city, the origins of their speech community and the strength of the networks in their area. There was stronger pressure in East Belfast to adhere to vernacular norms, particularly among young men, but the greater mobility of young women within the city meant that they were the speakers who carried linguistic change over barriers between the communities. As much as this research can tell us about BUV phonology, it is almost thirty years after the research was carried out. It is reasonable to assume that it is now outdated, and new research is needed to investigate the effects thirty years of segregation have had on these communities and their speech. This research might investigate whether gender differences are still clearly marked in BUV, and whether the diffusion of features from one side of the city to the other has continued at the same rate or been slowed down by segregation.

Works Cited

Bardon, Jonathan. Belfast: An Illustrated History. Dundonald, Belfast: Blackstaff Press, 1982.

BBC News. “Cranes to remain on city skyline”. Thurs. 9 Oct 2003. <http://news.bbc.co.uk/1/hi/northern_ireland/3176184.stm> , accessed 19 Oct 2006.

Boal, Frederick W. “Territoriality on the Shankill-Falls Divide, Belfast: The Perspective from 1976”. An Invitation to Geography (2^nd ed.). Ed. David A. Lanegran and Risa Palm. New York: McGraw-Hill, 1978. 58-77.

Douglas-Cowie, Ellen. "The sociolinguistic situation in Northern Ireland”. Language in the British Isles. Ed. Peter Trudgill. Cambridge: Cambridge University Press, 1984. 533-545.

Foulkes, Paul and Gerard Docherty. Urban Voices: Accent Studies in the British Isles. London: Arnold, 1999.

Granovetter, Mark. “The Strength of Weak Ties: A Network Theory Revisited”. Social Structure and Network Analysis. Ed. Peter V. Marsden and Nan Lin. Beverly Hills/London/New Dehli: Sage, 1982.

Harris, John. Phonological Variation and Change. Studies in Hiberno-English, Cambridge Studies in Linguistics, Supplementary Volume. Cambridge: Cambridge University Press, 1985.

Kirk, John M. “The Northern Ireland Transcribed Corpus of Speech”. New Directions in English Language Corpora: Methodology, Results, Software Developments. Ed. Gerhard Leitner. Berlin, Germany: Mouton de Gruyer, 1992. 65-73.

McCafferty, Kevin. Ethnicity and Language Change: English in (London)Derry, Northern Ireland. Impact: Studies in Language and Society. Amsterdam and Philadephia: John Benjamins Publishing, 2001.

Milroy, James. Regional Accents of English: Belfast. Dundonald, Belfast: Blackstaff Press, 1981.

----“The Interpretation of Social Constraints on Variation in Belfast English”. English around the world: sociolinguistic perspectives. Ed. Jenny Cheshire. Cambridge: Cambridge University Press, 1991. 75-85.

----Linguistic variation and change. On the historical sociolinguistics of English. Book. Oxford: Blackwell, 1992.

Milroy, James and Lesley Milroy. “Social Network and Social Class: Toward an Integrated Sociolinguistic Model”. Language in Society, vol.21, no.1 (1992). 1-26.

Milroy, Lesley. Language and Social Networks. Language in Society 2. Oxford: Blackwell, 1980.

Milroy, Lesley. “The Effect of Two Interacting Extra Linguistic Variables on Patterns of Variation in Urban Vernacular Speech”. Variation Omnibus. Ed. David Sankoff and Henrietta Cedergren. Edmonton, Alberta: Linguistic Research, Inc., 1982. 161-170.

Milroy, Lesley and Sue Margrain. “Vernacular Language Loyalty and Social Network”. Language in Society, vol. 9, no. 1 (1980). 43-70.

Northern Ireland Statistics and Research Agency. Northern Ireland Census 2001. < <http://www.nicensus2001.gov.uk>, accessed 28 Nov 2006.

Patterson, David. The provincialisms of Belfast pointed out and corrected. Belfast: Mayne, 1860.

Wolfram, Walt and Natalie Schilling-Estes. American English: Dialects and Variation, Language in Society 24. Oxford: Blackwell, 1998.