World Englishes , Vol. 27, No. 1, pp. 9–25, 2008. 0883-2919 Contrast and convergence in Standard Jamaican English: the phonologicalarchitecture of the standard in an ideologically bidialectal community ALISON IRVINE ∗ ABSTRACT: Theacrolect,interritorieslikeJamaica,isdescribedintheliteratureinanumberofways–as“the local standard English” or as the theoretical upper end of the construct referred to as “the continuum”.Data was collected from a sample of Jamaican speakers whose work prospects require use of Standard Jamaican English. The analysis reveals that phonological variation in Jamaican English is structured insuch a way as to reflect the coexistence of Creole and English, in a speech community that values thespeaker who can use both varieties. In a linguistic context that is characterized by continuous variation, useof some variables is crucial in defining the boundaries between Creole and English, thus establishingthe variety the speaker is using. These I call “load-bearing phonological variables”. I show that it is notthe use of English variants per se that defines someone as speaking the acrolect; rather it is use of variantsof these load-bearing variables. INTRODUCTION This paper presents data from research into the phonological variation that occurs in spo-ken Standard Jamaican English (SJE). More precisely, the variation discussed here occursin a sample of speakers who have been identified as using SJE “properly” by virtue of their promotion to the front-line staff of the agency that employs them. These speakers aretherefore acrolectal, as they are representative of the “well-educated urban professional”(DeCamp, 1961: 82) whose language use exemplifies the top end of the theoretical lin-guistic continuum that has been used to typologize the Jamaican language situation. Thiscontinuum has been discussed in a number of studies with different approaches, 1 but allshare an essential core definition that describes continuous variation occurring betweentwo language varieties, Jamaican Creole (JC) and Jamaican English (JE).DeCamp’s early description is representative; he identifies a continuous spectrum of speechvarietiesrangingfromthespeechofthemostbackwardpeasantorlabourer(basilec-talJC)tothatofthewell-educatedurbanprofessional(acrolectalJE)(1961:82).Themono-lingual JC speaker – associated with little or no education, low income and traditionallyrural provenance – has low status in a social context that sees an inability to speak Englishalmost as moral failure (see Silverstein, 1996: 292–3). The following comment in one of Jamaica’s national newspapers is typical: Tosaythatwewillnowteachthevarioussubjectsinpatois[JC]becauseitistheonlylanguagetheyknowis not only completely ridiculous and laughable but also a sign of failure in our education system. As has become the norm in so many things in Jamaica today, we don’t pursue excellence per se, but rather lower ∗ Department of Language, Linguistics and Philosophy, University of the West Indies, Kingston 7, Jamaica. E-mail:treacle [email protected]
C 2008 The Author. Journal compilation C 2008 Blackwell Publishing Ltd., 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA02148, USA. 10 Alison Irvinethe bar of achievement to accommodate the inefficient, the slothful and the undisciplined. (‘The PatoisDebate’, Jamaica Observer , 18 March 2006) At the same time, the ability to use JC is considered a necessary element in the complexof sociocultural characteristics that define a Jamaican identity. The same writer continues: We all love patois and would never want to see it diminish or leave our shores, for it is indeed one of those things that bind us together as a people and make us quintessentially Jamaican. The Jamaican speech community therefore idealizes the speaker who can use both JCand JE, as the monoglot speaker is viewed as either socially inferior, if they only useJC, or “foreign” and pretentious, if they only use JE. Additionally, speakers believe thereto be two languages, best used in a socio-functional relationship that can accurately belabelled “diglossic” (Akers, 1981: 8; Winford, 1985; Devonish and Harry, 2005). And asBeckford-Wassink (1999: 72) notes, her study of language attitudes in Jamaica found that: Alltheprofessionaldomains ... (thosewheretheauditormightbeanon-familiarorsubordinate,aswhenansweringthetelephone,addressinganemployer,orteaching)weredeemedinappropriateforPatois[JC]usage. In this ideological and sociolinguistic context, data was collected from 51 educated Jamaicans whose work requires them to be SJE speakers and who are employed, in part,on the basis of how well they use SJE during their job interviews. The focus of this paper is the patterning of phonological variants in 10 variables used by this sample of speakersof the acrolect, SJE. LOCATING THE ACROLECT, STANDARD JAMAICAN ENGLISH In an earlier paper (Irvine, 2004: 42–6), I discussed some of the theoretical and method-ological weaknesses in the traditional approach to identifying the acrolect in speech com-munities like Jamaica. In brief, because the acrolect, the “local standard”, is putatively notso different from metropolitan Standard English (MSE), MSE has been used to typologizestructures found along the continuum. Judgements of forms as acrolectal or basilectal have been typically made with reference to MSE and not to “the local standard” that is referred to in passing. Jamaican speech has therefore often been analysed in terms of a variety thatis not actually spoken in the speech community, and the structures used by Jamaicans aredefined in terms of norms external to the community of speakers.Moreover,whathasemergedisakindofcircularityinthedefinitionofthepolarvarietieson the (Jamaican) continuum that, typically, takes this form: (a) Since the acrolect is notunlike MSE, phonologically as in [k h ] (in a word like cat ) or morphosyntactically as in did not take ; then (b) the basilect reflects structures that are maximally divergent fromthe acrolect – [kj]at or no ben ∼ neva tek ; therefore (c) acrolectal speech is maximallynon-basilectal and can therefore be expected to produce the forms [k h ]at or did not take .This circularity extends even to the identification of so-called basilectal speakers, so thatsamples of speech collected in the field are rejected when they are “too English” and therefore “noncreole” (Escure, 1997: 74, commenting on the general practice of creolists), because the linguist already knows what is supposed to be basilectal. C 2008 The Author. Journal compilation C 2008 Blackwell Publishing Ltd. Contrast and convergence in Standard Jamaican English 11 As a consequence, analyses of the Jamaican acrolect have generated some paradoxicalconclusions: • Forms have been typologized as basilectal in Jamaica even though they are attested to in the speech of all Jamaicans generally, irrespective of social context of use. • Forms are labelled as (upper) mesolectal even when found in the formal speech of educated Jamaicans or appear to be accepted as standard locally. • Conclusions are arrived at that suggest that no one in the speech community speaksStandard English, even as members of the speech community hold to an idea thatthere is such a thing and that there are such speakers. • The presence of non-MSE forms in the speech of educated Jamaicans is described as an emerging or new trend, even though such variation was identified some twocenturies ago.In this paper, I take an endonormative approach to locating the acrolect. The acrolectis the English normalized in Jamaica through public/formal use by speakers at the top of the sociolinguistic spectrum, which evolves within that local sociolinguistic context. Theevolution, of course, is more constrained by the prescriptions of written English and thegreater access to education. Additionally, the patterns that occur in the acrolect or SJEmust be informed by those that occur in the speech perceived to be Creole. This is whysome forms that occur in MSE are avoided by SJE speakers when they are also aspectsof Creole (see Irvine, 2004: 66). While some speakers in the Jamaican speech communityare vernacular speakers of a variety of JE, and many more are vernacular speakers of a variety of JC or both, SJE represents a situationally defined and defining variety thatis sanctioned and reinforced by institutions and agencies of the state for formal/ publicdiscourse. The acrolect, SJE, is therefore typically going to be phonologically distinctivefrom other varieties of Standard English.Crucially, there is evidence that Jamaicans themselves perceive accent and vocabularyalso to be the primary difference between Creole and English (Beckford-Wassink, 1999:66). Of 51 informants in her study, only 9 (18%) identified any aspect of morphosyntaxas distinguishing Creole and English, while 42 (82%) mentioned either accent (“how theysound their words”) or accent and vocabulary.The acrolect is therefore locally developed and particular to the specific speech com-munity in which speakers operate, as it is within the speech community that the shared indexicalityunderlyingcommunicativeactsofidentityandgroupnessevolves(Silverstein,1998: 407). This would necessarily seem to be most focussed around the phonology, espe-cially for speakers who pass through institutions that disseminate a near-universal literarystandard of English. DESIGN OF THE STUDY Data for this study was collected in 1994 at Jamaica Promotions (JAMPRO). 2 JAMPROhas a total staff complement of 188, in various offices in both urban and rural Jamaica.The principal JAMPRO building, from which all data was collected, had 153 members of staff at the time; 104 were interviewed and 82 agreed to have their interviews recorded.All interviews were done in an office on location loaned to me by the agency. Allinformants were given appointments, usually a day before the actual recording session,and told that I was conducting research on JAMPRO itself and the morale of staff. The 82, C 2008 The Author. Journal compilation C 2008 Blackwell Publishing Ltd. 12 Alison Irvine whose speech was later analysed, were recorded with a notebook in plain sight in whichI would, from time to time, make jottings. I was interested in eliciting formal speech,and was attempting to set up a situational and interactional context that would discourageinformality.Aninformantcan,ofcourse,incontextslikethis,exploitavailablelinguisticresourcestoconstruct for herself a particular persona or to construct an encounter as intimate, distant,friendly or otherwise (Milroy and Gordon, 2003: 206). The informant, like all speakers,is an active initiator of use of particular styles of speech (Rickford and McNair-Knox,1994), and might choose to depart from the SJE that would be considered, certainly amongeducated Jamaicans, unmarked in a formal, taped interaction with a stranger. As such, thisstudy rests on the assumption that for most, if not all, informants, the speech produced during the interviews reflects their use of this unmarked variety.Ofthe82recordedinformants,51belongedtoasubsectionofthestaffcalled“front-lineemployees”. This label is applied to those employees who actually interact with the public, both local and foreign, in the course of their duties. In certain departments at JAMPRO,some members of staff are required either to deal directly with the agency’s clients or torepresent the agency in overseas departments or trade missions.Selection of staff at JAMPRO is heavily dependent on performance during the jobinterview, particularly in the case of professional and (senior) management positions. 3 Front-line staff positions reflect an even more subjective set of selection criteria on the partof the agency, as these are employees who are the public face of JAMPRO. As explicitlystated by both senior management and the general staff, those in front-line positions areselected because they are perceived to be the type of person who can represent both theagency and the agency’s idea of a well-educated, competent, professional Jamaican with“a good command of the English language”. 4 The social factors that informants suggested would have some degree of associationwith membership in the ranks of front-line staff are: (a) a high level of education, (b)“middle-class” social status, (c) light skin colour, and (d) use of “proper” English. I arguehere that, since there seems to be some validity to and empirical support for the first threeoftheseperceptions, 5 thenthefourth,whichspeaksspecificallytothelinguisticcriteriafor front-line selection, should also be taken as important. Front-line staff are those employeeswhose English is believed to be suitable for a representative of JAMPRO, displaying theattributes of “diction” and English “at a standard”. THE PHONOLOGICAL DATA In the presentation of data below, I focus on the results for the 27 front-line informantswho have either a first or postgraduate degree from a university (labelled + Tertiary inthe tables below). Some front-line employees, such as the agency’s ancillary and supportstaff, have relatively low levels of formal education (labelled –Tertiary in the tables).This paper is specifically concerned with the formal speech of the “well-educated urban professional”. The results from the 28 non-front-line members of staff who do not havetertiary qualifications are included for contrast. 6 Along the Creole/English dimension we would expect the highly educated members of the front-line staff to use fewer Creole forms than those with relatively little education whowere not selected for interaction with the agency’s clients. C 2008 The Author. Journal compilation C 2008 Blackwell Publishing Ltd. Contrast and convergence in Standard Jamaican English 13 Word-initial glottal fricative /h/ /h/ is not phonemic in many varieties of Jamaican Creole (Akers, 1981: 32; Devonishand Seiler, 1991: 7; Meade, 2001: 40), though there is evidence that it is in WesternJamaicanvarieties(Wells,1973:12).HypercorrectuseofthefricativeiscommoninJamaica(Cassidy and Le Page, 1967: lxii), particularly in Kingston, where this JAMPRO samplewasrecorded.Inmytotalsampleofspeakers(82),11informantsproducedthishypercorrect pattern, as in F52 [hon i ] owning or M27 [ha υ t] out , and typically with single attestations.The more general pattern was h -dropping, a term that clearly assumes that speakers areusing an English phonological system, which does have /h/. Evidence for this is furnished in the pattern of use of h in the texts collected. Except for the 11 whose usage suggestsnon-contrastive [h], as in [ha υ t ∼ a υ t] out, most informants produced /h/ in the same wayas in internationally acceptable English (IAE), distinguishing lexical items like hand and and , without inserting [h] where it is not lexically specified in IAE. Table 1 presents thedata for use of /h/ in the sample of informants. Table 1. Word-initial productions of /h/: front-line and non-front-line staff compared Informants (n) h drop (%) [h] (%) Total + Tertiary front-line (27) 36 (5) 626 (95) 662 − Tertiary front-line (23) 47 (14) 335 (86) 382 Non-front-line (28) 101 (18) 447 (82) 548 For many Jamaicans, h -dropping is overtly stigmatized, though it is not at all clear thatthisisuniversallythecaseorthat h -lessnessisevenperceivedbythehearer.Itisastereotypeheld by others in the Caribbean of bad Jamaican speech (Wells, 1982: 569; Roberts, 1988:90). Educated front-line staff seldom h -drop. Indeed, 14 of the 27 informants never did.Two or three instances in each of the remaining 13 informants’ texts account for the 36 inTable 1. The low back stressed vowel / ɔ / In ‘not’ words the possible variants are [nat ∼ n ɔ t]. The “Creole” variant in this setis the low central vowel [a], as Jamaican Creole has one low vowel phoneme, /a/, with[at] produced for both hat and hot . Wells (1982) says that “block is acrolectally [bl k] (or sometimeswithanunroundedbackvowel[bl ɑ k),butbasilectallyhomophonouswithblack [blak] (576)”. 7 SimilaranalysesofthebasilectcanbefoundinAkers(1981:25)andDevonishandSeiler (1991: 5). Hypercorrect use of [ ɔ ] is one of the features that Patrick (1997: 48) associateswith “speaky-spoky”. 8 Forms like [s ɔ l ə r i ] salary (M47), [r i l ɔ ks] relax (F53) and [f ɔ :m ] farmer (F80) produced by some of my informants would suggest the kind of sensitivityto the feature discussed by Patrick and a belief that [ ɔ ] is the “correct” reflex of [a]. LikeWells, Beckford-Wassink (2001: 151) and Meade (2001: 42) describe acrolect (dominant)speech as clearly distinguishing /a/ and / ɔ / in ‘not’ words. The results for the JAMPROinformants are shown in Table 2.Again, educated front-line staff typically use the JE variant. In contrast, the non-front-line informants produce more than double the number of JC variants. The results for /h/ C 2008 The Author. Journal compilation C 2008 Blackwell Publishing Ltd.