To appear in Journal of Linguistics:

Peter L. Patrick, Urban Jamaican Creole: Variation in the mesolect. Amsterdam: John Benjamins, 1999. Cloth, Pp. xx+329.

Reviewed by Salikoko S. Mufwene, University of Chicago.

ÝÝÝÝÝ Urban Jamaican Creole (UJC) is a stimulating and thought-provoking book, quite an informative and useful addition to the literature on speech continua, though there are respects in which, as I hope to show below, the authorís interpretations of the data remain disputable. Patrick addresses several interrelated questions about the creole mesolect which he summarizes in the conclusions of the book as follows: 1) what is ìthe nature of mesolectal grammar?î and 2) what is ìthe sociolinguistic structure of variation in the creole continuum?î (292). While the second question is answered to my satisfaction, the first is not, due in part to assumptions about the mesolect and its relation to both the basilect and the acrolect that I do not accept. I discuss the assumptions first, so that the reader can better understand my criticisms of a book that I otherwise think is competently put together.

ÝÝÝÝÝ Patrick presents the mesolect and its lectal continuum as a reality that can clearly be distinguished from both the basilect and the acrolect. This position is disputable, first because a basilect is just an analytical construct intended to depict the extreme systemic level of divergence from the acrolect that a linguist can infer of a creole. The standard, often misidentified with the acrolect (the speech of the educated and/or upper class), is the other extreme of the continuum within which most speakers in a creole community gravitate in one or the other direction. Natural speakers whose discourse evidences all basilectal features where they are expected are as rare as natural native standard English speakers who sound like a book. Much of acrolectal speech is colloquial and oftenÝ contains non-standard features. In the case of Jamaican society, even acrolectal speakers often display some of the features associated with Creole, such as the merger of 17th-century /Ê,Ý ,Ý / and /a/ into /a/. Thus positing a mesolectal grammar that is distinct from basilectal grammar is problematic, as much as I agree with Patrick that mesolectal grammar is structurally heterogeneous. In fact, grammars are generally not monolithic (Mufwene 1992). They exhibit principles that sometimes are not consistent, overlap, and thus compete with each other, as is evidenced by UJC. Utterances that they generate can be accounted for with two or more coexistent systems (Labov 1998), though it is sometimes difficult to determine unequivocally which system is at work. This is precisely where Patrick often errs, when he excludes almost a priori the possible role of the basilectal system in the production of some theÝ variants he discusses.

ÝÝÝÝÝ He is right is assuming that these featuresóbe they ìcreoleî or otherwiseó originate in English. However, one must also remember that ìcreole continuaî are partly a consequence of the fact that different dialects of their lexifiers came to coexist and presented conflicting models in the colonies.Ý The fact that the vast majority of the populations speak mesolectal varieties that diverge significantly from the acrolect reflects a contact history in which slaves far outnumbered the European colonists and the lower and working classes still far outnumber the upper class today.

ÝÝÝÝÝ The book is otherwise efficiently structured into eight chapters. In the first (1-22), Patrick situates the subject matter and formulates the central questions he addresses. He correctly states that mesolectal Jamaican speech is structurally heterogeneous, dissociates the continuum from ìdecreolization,î and argues that a multidimensional characterization of this linguistic situation is preferable to a unidimensional oneóa position he proves well in the chapters where he discusses variables other than ìphonolexicalî KYA (see below).

ÝÝÝÝÝ In Chapter 2 (23-64), Patrick presents Kingston as an urban setting, which developed differently from rural Jamaica, and he identifies the Veeton community in which he conducted his field research. He characterizes UJC as largely mesolectal, in contrast with rural, basilectal speech. Occasionally he identifies the latter varieties as ìconservative,î suggesting that rural speech is older, perhaps where contemporary Jamaican speech started overall. He actually observes that ìThe linguistic clocks in rural areas do not only run more slowlyóthey operate in a distinct social context and cannot be expected to slavishly follow urban developments a generation behindî (49).

ÝÝÝÝÝ His position is disputable. Chances are that since Kingston and the rural sugar cane plantations developed concurrently, all the lects of the Jamaican continuum (with probably the exception of the standard introduced through the scholastic medium) evolved at the same time, with the varieties closer to the basilect concentrated in the rural areas, where the slaves and their descendants have always been the overwhelming majority. Although rural speech is stigmatized, it is not obvious that people who always live in rural environment ever wish to speak like urbanites. Attitudes toward the latter are not always positive. Nevertheless, Patrick stratifies hisÝ consultants in a useful way that shows later in the book why education and social status, for example, cannot independently account for mesolectal variation.

ÝÝÝÝÝ Chapter 3 (65-82) explains the authorís field methods: a combination of interviews, English-to-Patwa translations, and standard English reading tests, complemented by a language attitude questionnaire and several informal observations during his interactions with the population in various Jamaican vernaculars, which he speaks fluently.

ÝÝÝÝÝ Chapter 4 (83-119) introduces the first variable for analysis: KYA, a convenient representation for words that contain a velar stop that is followed by a vowel that in American and British English mainstream varieties corresponds to /Ê, a:, ar/ or / /. The vowels /a:/ and /ar/ are conveniently represented as AR. In UJC they are both produced as /a:/ and thus distinguished from the other instances of /a/ by length. In lower mesolectal speech, only some of the words containing /a/ (those which should historically be non-back, assuming /a/ was a central vowel) show palatalization of the velar stop. Thus, cart, garden, and can but not cot, caught, got. The explanation is that this pattern represents continuity from 17th-century English (118). In varieties close to the acrolect, a phonological constraint prevents palatalization in words that have AR, as ìin the modern Midlands dialects, Northern Ireland, and Charlestonî (119).

ÝÝÝÝÝ Here we see the first evidence for systemic mixedness in UJC, with one group using a phonological principle in the production of the relevant words but another not. Patrick argues perhaps too fast against substrate influence, which must have favored the merger of /a:/ and /ar/ into /a:/ and the elimination of the phonological constraint still operating in the acrolect. On the other hand, he could have considered the following argument against substrate influence: In African languages, as in most others around the world, palatalization of velar stops usually occurs before high front vowels, not before lower ones. My point is simply that diverse influences are not necessarily mutually exclusive.

ÝÝÝÝÝ One practical problem readers not familiar with UJC may experience in reading this chapter and the following ones is that Patrick is slow in adducing the relevant examples in his arguments. They tend to be presented rather late. He also claims that the KYA variation represents change in progress, since it is mostly the older speakers who have no phonological constraint regulating the production of the relevant words. This position is disputable in the Jamaican context, since no evidence is provided of such an ongoing change in rural Patwa and the young may just reflect the fact that in Kingston the phonological constraint has been strong since UJCís inception. After all, most of the older speakers immigrated from the rural area.

ÝÝÝÝÝ Chapter 5 (121-166) is about the deletion of word-final /t, d/. It includes comparisons of constraints on this process with those that operate in American and British English varieties. For reasons of primarily analytical economy (159), Patrick concludes that this process operates in basically the same manner as in English dialects, except that ratios of ì(TD)-absenceî are higher in UJC (even before vowels) and regular verbs show the highest rate of all verbs, next to the weakening of negative nít (150).

ÝÝÝÝÝ Although there is independent evidence for hypothesizing consonant cluster reduction in UJC (e.g., lost [last] > [las] and last [la:st] > [la:s]), the high rate of absence in ìpast verbsî suggests that this continuum may be underlain by a non-monolithic system. In this case, lower mesolectal speakers may be using a basilectal principle that allows using the verb stem with PAST meaning when the discourse context makes this obvious. The minority of cases where /t, d/ are attested correspond to insertions under pronunciation specifications that overlap with those specified for ìdeletionî in other English varieties.

ÝÝÝÝÝ Patrick discards this alternative analysis, because it is not consistent with the traditional assumption that morphemes are inserted first and phonological rules apply to their outputs. First of all, this convention is not ipso facto an argument for psychological reality. Second, morphological specifications can also consist of amorphous abstract specifications, e.g., PAST, and abstract morphophonological rules can be posited that give phonemic shape to the morphological abstractions. Surely, early generative phonology in the 1970s provided evidence for preferring deletion to insertion rules for the sake of generality. However, aside from the fact that generality does not entail psychological reality, insertion would be just as general as deletion in this particular case, especially since the absence rate is almost the same for regular verbs as for semi-weak ones (e.g., send), viz., 56% and 59% respectively (157).

ÝÝÝÝÝ Chapter 6 (167-222) is about the ìpre-verbal past-markersî did and neva. Aside from noting accurately that their text frequency is very lowócontrary to what might be expected from theoretical analysesóPatrick concludes again that the pattern of their distribution is English. What he does not show is that in UJC, as in the basilect, neva also has another meaning, viz., PAST&NOT, which is more or less suppletive for no ben (ëdid/had notí) and different from the English meaning NOT EVER. Like basilectal ben, did is also part of a relative tense system and is not always a morphosyntactic alternant of ‚-ED. The structure of UJC may be more non-monolithic than Patrick admits. It is perhaps not by accident that the durative dida is so similar to bena. Did is also used almost exclusively by lower-mesolectal speakers (204-205), subject to discourse constraints similar to those of ben. These observations are not arguments against Patrickís position that the markers originated in English. My point is simply that English origin is not a sufficient reason for claiming that the grammatical functions of preverbal did and neva (as of ben) have remained the same as in the lexifier. The restructuring of English dialects into UJC involved some concurrent changes (though only minor in some cases) in the functions of the selected grammatical morphemes.

ÝÝÝÝÝ In Chapter 7 (223-266), Patrick helps us put the discussions in a larger picture, as he focuses on ìpast-marking by verb inflection.î Interestingly, even irregular verbs have a very low rate of past-marking, viz., 32%, as opposed to 44% for semi-weak verbs and 46% for regular verbs (231). Stative verbs are more often inflected than nonstatives (256), consistent with ‚ben-marking in the basilect. It becomes more difficult to resist the alternative analysis that most speakers know principles associated with Patwa and those associated with standard English and they alternate between both, but they do not command a uniform norm that first inserts a PAST morpheme and then allows them to delete it. Patrickís figures 4.7, 8.1, and 8.2 seem consistent with this interpretation of the linguistic performance of his consultants. The figures suggest clearly differentiated systems for some speakers but more or less blurred ones for others. The speakersí productions reflect the extent to which they command one or the other system better or their ability to gravitate between them. In this chapter, Patrick acknowledges that there are competing forms in the mesolect (251) and in fact competing grammars (264-265), but he fails to note that this conclusion is not quite like the single-norm position he defends in the previous chapters.

ÝÝÝÝÝ Chapter 8 (267-295) concludes the book, noting for instance that UJC is not a product of code-switching (292), that it is underlain by a mixed grammatical system (293), that the continuum model that emerges from his analysis is ìnon-discreteî (292), that there is an ìasymmetry of dual sets of norms for (synchronically-)related varieties in the creole continuumî (284), that ìmesolectal speakers do not have (or at any rate, use) a full basilectal grammar, but have not fully acquired an English oneî (293), that ìmost varieties show a relatively sharp break in their distribution across the population, but these breaks tend not to coincideî (283), and that a unidimensional account of the continuum is not valid (286). He admits that similar continua exist elsewhere in non creole-communities, but unfortunately he does not suggest ways in which the present study can enrich our understanding of such continua. I suspect that the working assumptions I disputed at the outset of this review have something to do with this shortcoming.

ÝÝÝÝÝ Nonetheless, the book remains very informative on UJC and leaves us with the challenge of how to best articulate non-monolithic systemsówhich are not two separate languages or dialects, but coexistent forms and overlapping principles that compete for use within the same system.

References

Labov, William. (1998). Co-existent systems in African-American vernacular English. In Mufwene, S., Rickford, J. R., Bailey, G., & Baugh, J. (eds.) African-American English: Structure, history, and use. London: Routledge. 110-153.

Mufwene, Salikoko S. (1992).Ý Why grammars are not monolithic. In Brentari, D, Larson, G., & McLeod, L. (eds.) The joy of grammar: A festschrift in honor of James D. McCawley. Amsterdam: John Benjamins. 225-50.