DANIEL R. DAVIS
There has been significant progress in the grammatical description of varieties of English in the past 25 years (Schneider 2003: 234). Specific grammatical descriptions play an important role in the recognition of different English languages, and demarcate a distinct stage in the history of the grammatical tradition. Nevertheless, the writing of these descriptions comes at a cost. They depend on assumptions drawn from various areas of linguistics and language study, and these assumptions limit the uses to which these descriptions can be put.
This chapter is inspired by the integrational linguistic approach set forth in Harris (1998), and draws upon the work of sociolinguists James Milroy, Lesley Milroy, and Deborah Cameron. The chapter is integrationist in its commitment to the assumption that current grammatical description, both in form and intent, owes a great deal to the general cultural background, the historical contexts, intellectual issues, and philosophical discourses of the English languages. Even the most basic grammatical terms are set within an intellectual tradition, and have political implications: There is no such thing as a value‐free description. This approach speaks to the experiences of those using, encountering, and analyzing world Englishes and varieties of English. Milroy and Milroy (1999) explore the importance that social networks and grammatical traditions have for social attitudes toward grammar, and Cameron (1995) draws out the political conditions and social implications of public discourse about grammar and related forms of what she terms “verbal hygiene.” These three sociolinguists have therefore called into question the supposed irrelevance of language prescription in linguistics.
Traditional prescriptive grammars of English reveal surprising openness to the question of varieties. Wallis (1972 [1653]: 108–113 [xxv–xxvii]) clearly intends to describe the language for the benefit of both foreign learners and native speakers and finds that accounts of English based too closely on Latin models are not suitable for this purpose (Michael 1970: 164–165; 495–496). This emphasis on description suggests the possibility of an empirical approach to the language of the community, and allows for the adaptation of terminology to reflect linguistic difference. Wallis nonetheless chooses to retain Latin terminology, a decision reflected in the terminology of descriptive grammar today. Kirkham (1833: 59, 63) questions the usefulness of his own prescriptive rules and allows that both singular and plural agreement work equally well with collective nouns, while arguing that incorrect agreement sounds “harsh.” Even a traditional and explicitly prescriptive grammar, in the right hands, is open to the problem of variation. Milroy and Milroy (1999: 30) define standardization as the suppression of optional variation and trace the development of this ideology in British and American culture from the seventeenth to twentieth centuries. Prescription is awarding prestige to one variant. Implicit in this is the descriptive act of recognizing that (given the analytical framework of the parts of speech) several variants exist.
In the late nineteenth and early twentieth century, historical linguistics and dialectology offered a basis for descriptive grammar tied to a social and historical conception of linguistic correctness. Even in the prescriptive and normative context of a school grammar, historical linguist H. C. Wyld (1925: 8–13, 205–206) defines grammar as the facts of a spoken language, places this in a community setting, and allows for variation across and within communities. His conception of the English language is explicitly pluricentric, and he recognizes the role of social and historical change in reconfiguring the standard and literary forms of the language (Wyld 1925: 220). His examples, however, are confined to British English dialects. Jespersen (1933: 16) mentions subdivisions of English, including geographical (Scottish, Irish, American) and social. Like Wyld, he defines descriptive grammar empirically (Jespersen 1933: 19‐20): “what is actually said and written by the speakers of the language investigated…lead[ing] to a scientific understanding of the rules followed instinctively by speakers and writers.”
This calls to mind Saussurean structuralism, in that language is situated in the community and recognized to be in a state of variation from one individual to another, and from one community to another. Nevertheless, this variable data is analyzed in order to derive an abstract set of rules (a language structure) followed by language users. It is ironic that traditional prescriptivism depending on descriptivist assumptions has been supplanted by a descriptivism assuming an underlying unity (J. Milroy 1999, for extensive analysis of the impact of standardization on linguistic description). This replacement has serious consequences for the representation of variation in descriptive grammar. As Harris says:
The situation in which an established descriptive format devised for one particular purpose is taken over and adapted to serve some new and quite different purpose is a situation fraught with potential errors and inconsistencies of all kinds.
(Harris 1981: 54)
Descriptive grammars draw heavily on the prescriptive tradition for their terminology and method. They are arranged on a traditional framework of the parts of speech, refined with the use of structuralist discovery method (Biber, Johansson, Leech, Conrad, & Finegan 1999: viii, 4 for confirmation of this, although the corpus‐based approach of this grammar admits context of use into the foundation of the grammar and gives it greater sociolinguistic value). The various frameworks of syntactic theory are not usually part of these descriptions, but rather use these descriptions as the basis for theory. Henry (2002: 267) discusses the way in which syntactic theory is for the most part predicated on assumptions that rule out variation.
Although Quirk, Greenbaum, Leech, and Svartvik (1985, hereafter termed “the Quirk grammar” unless specific page reference is given) discuss the possibility of grammatical variation in world Englishes in their introduction, the main body of this work adheres to the familiar pattern of presenting a core English with two equally prestigious varieties, each acceptable within its own regional monopoly. The index cites 150 sections or notes referring to American English constructions and 136 sections or notes referring to British English constructions. No other varieties appear in the index with constructions, except for “nonstandard’ with 26 constructions and “regional’ with 35 entries (not restricted to constructions). This is out of a total of 1,450 sections, and one might infer that approximately 9% to 10% of the sections of the grammar deal with variation between American and British English (Görlach 1991: 25), while only 2% to 3% of the sections of the grammar deal with other varieties, including nonstandard varieties.
It should be recalled that the Quirk grammar is not a direct reflection of the English language in its entirety but rather represents a notional “Standard English.” To take a convenient example, Quirk et al. (1985: 1247–9) discuss the nonpersonal relative pronouns which, that, and “zero” but make no mention of the nonstandard relative markers what or as. Therefore, although it is “descriptive,” the Quirk grammar cannot itself be used as evidence for the common core, as it excludes many nonstandard forms by definition. These forms may equally merit representation within a very different “core.” In addition, this section‐counting method does not indicate the comparative frequency of variable constructions, nor how these frequencies vary according to register. Finally, the market for the grammar is quite clearly those in search of an authoritative treatment, a description of what correct English is, that may be used prescriptively to say what is not correct English. Nonstandard forms and “regional” varieties are not in this picture, except in so far as forms encountered by learners must be explained (thus the account of relative pronouns, but also of the royal we and nonstandard us “Give us a job,” Quirk et al. 1985: 351).
Successors to the Quirk grammar conform to this pattern, for reasons of market, purpose, and methodology. Large‐scale descriptive grammars reflect the concerns of language learners, the publishing industry, and language specialists. Although both the Longman grammar and the Cambridge grammar identify their target audience as linguists (Biber et al. 1999: 45–46; Huddleston & Pullum 2002: xv), the size and expense of these works suggest that a large part of their market must be libraries in need of reference grammars. Like the Quirk grammar, they are consulted in order to find sanction for particular forms and usages. One cannot ignore the prescriptive power of a good description (Marenbon 1987, cited in Cameron 1995: 10).
The Longman grammar (Biber et al. 1999: 17–20, 25–26) is based on a 40‐million word corpus of British and American English and deals extensively with differences between American and British varieties, but also with differences between registers (conversation, fiction, newspaper language, and academic prose). Frequencies are given, making it possible to discern levels of normative agreement in different registers. An entire chapter is devoted to the grammar of conversation (1038–1125), with a small section devoted to nonstandard forms. The text asserts that most variation occurs in the area of morphosyntax, and that syntax is largely variation free, with the multiple negative and double comparative illustrated by “AmE” (1125). Nonstandard forms are mentioned in the text, as in the discussion of nonstandard relative markers what and as (Biber et al. 1999: 608). Reliance on the Longman Spoken and Written English Corpus (LSWE) corpus of British and American texts and conversations, which makes possible the frequency statements, also rules out discussion of World Englishes, although the authors direct readers to the International Corpus of English project (1133, n. 1). The Cambridge grammar (Huddleston & Pullum 2002) incorporates some reference to different varieties of English, although the emphasis is on syntactic structure derived from acceptability judgments. This has the effect of limiting serious consideration of varieties. For example, in the discussion of relative clauses there is no mention of nonstandard forms, and the goal is to describe the patterning of the standard relative pronouns and identifying the syntactic structures necessary to account for integrated and supplementary relative clauses (Huddleston & Pullum 2002: 1059–1061).
Kachru (1992: 304) identifies Quirk et al. (1972) as the moment of recognition for world Englishes. Quirk, Greenbaum, Leech, and Svartvik (1972) arrive at this recognition in the course of defining their object of study. The text uses the concept of a common core of English as a foundation for Standard English, which in turn is defined as the usage of the educated. On the basis of an analogy with taxonomy and intraspecies variation (the dog features or “dogness” of dogs embodied in different varieties of dog), the authors argue that “we need to see a common core or nucleus that we call ‘English’ being realized only in the different actual varieties of the language that we hear or read” (Quirk et al. 1972: 13). The empirical condition for the common core is that there are common grammatical features in all varieties of English:
The fact that in this figure the “common core” dominates all the varieties means that, however esoteric or remote a variety may be, it has running through it a set of grammatical and other characteristics that are present in all others. It is presumably this fact that justifies the application of the name “English” to all the varieties.
(Quirk et al. 1972: 14)
The logic of this passage, that there must be a common core shared by all varieties of English, and that this core consists in grammatical features, raises a number of difficulties (Kachru 1986: 83). Even if one accepts the premise that a common core is necessary in order to found a taxonomy of animals or languages, it is not clear that such a taxonomy is the purpose of a descriptive grammar of English. If it were, the grammar would have to contain information about the features of all varieties of English, and of other languages and their varieties historically related to varying degrees (and even after this work the linguistic features would merely suggest rather than confirm relationships). This information is to be found in an historical and comparative grammar.
In Quirk et al. (1985: 16), the core is no longer explained, but merely asserted, and it no longer “dominates,” “A COMMON CORE or nucleus is present in all the varieties so that, however esoteric a variety may be.” Still, the last sentence of the paragraph remains, “justifies the application of the name ‘English’ to all of the varieties” (Quirk et al. 1985: 16). The core has been called into existence in order to define the descriptive object of study, much as langue or the language system, also defined as a commonality, has been constructed as the object of study in Saussurean linguistics, “it is something which is in each individual, but which is none the less common to all” (Saussure 1983: [38]).
Quirk et al. (1985: 15) define Standard English from within this core, as the “supranational” usage of the educated, standing in opposition to the uneducated speech more closely aligned with the regional dialects. Like the core, Standard English is defined in opposition to variation, in terms of what is common to all educated speakers:
What we are calling national standards should be seen as distinct from the Standard English which we have been discussing and which we should think of as being “supra‐national,” embracing what is common to all. …[T]here are two national standards that are overwhelmingly predominant both in the number of distinctive usages and in the degree to which these distinctions are “institutionalized”: American English and British English. Grammatical differences are few and the most conspicuous are widely known to speakers of both national standards.
(Quirk et al. 1972: 17)
This quotation highlights problems with the core, and with the notion of Standard English (as used by Quirk et al.). Having first defined the core and the standard negatively, that is, as not containing any linguistic features not present in all varieties (or all varieties used by educated speakers in the case of the standard), it becomes necessary to identify two national standards used by educated speakers in their respective societies, precisely because these varieties do contain distinctive features which are institutionalized (that is, which are accorded the status of a standard). Dumas’s “one for all and all for one” has been replaced by Orwell’s “All animals are created equal, but some animals are more equal than others.” Quirk et al. (1972) are required by their argument to say that Standard English is different from the national standards. Otherwise, they must either insist that either American English or British English is not Standard English (leaving the other the winner on the world stage) or admit that English has split into at least two standards. The compromise position which they take is to hypothesize a Standard English of shared common linguistic features, which manifests itself in British and American contexts (most obviously publishing) as two separate national standards, each of which has additional features authorized as standard by the institutions of the respective society.
And this brings us to the moment of recognition of which Kachru tells us. Quirk et al. (1972) say:
At the opposite extreme are interference varieties that are so widespread in a community and of such long standing that they may be thought stable and adequate enough to be institutionalized and regarded as varieties of English in their own right rather than stages on the way to a more native‐like English. There is active debate on these issues in India, Pakistan and several African countries, where efficient and fairly stable varieties of English are prominent in educated use at the highest political and professional level.
(Quirk et al. 1972: 26)
Having defined community acceptance and support (institutionalization) as a way to explain the existence of nationalized British and American standard English alongside Standard English, there is no way to shut the door on any variety which can show distinctive features and institutionalization in the context of a nation (or perhaps a clearly demarcated community). Although the Quirk grammar does not follow through on this statement (opting instead for a common‐core model with ad hoc recognition of British and American differences), the stage is set for the variationist treatments of the early 1980s.
As we have seen, the problem with the Quirk grammar is that specific discussion of variant forms of grammar is limited to standardized varieties of British and American English. Space in such a large grammar is clearly at a premium, but one would be forgiven for assuming that these two varieties are the only ones sufficiently standardized to merit the attention of students and scholars. Variationist treatments of the early 1980s, including Bailey and Görlach (1982), Trudgill and Hannah (1982), and Platt, Weber, and Ho (1984), sought to correct this assumption by demonstrating the linguistic distinctiveness and social institutionalization of varieties of English.
Bailey and Görlach (1982) consists of a number of chapters by different contributors, each devoted to a particular variety of English. Each variety is treated in terms of its external history, linguistic features, and other sociolinguistically important or relevant aspects. The treatments of linguistic features are concise and tend to emphasize phonology and lexis (vocabulary). The purpose of the book is to trace the origin and institutionalization of target varieties, with an eye toward accounting for plurality by means of a social historical frame of reference.
Platt, Weber, and Ho (1984) organize their text around levels of analysis and grammatical constructions, which are then compared across varieties. This plan is somewhat more convenient for examining grammatical features found in more than one form of English, with separate chapters on variation in the noun phrase, variation in the verb phrase, semantic change grouped with derivational morphology, and syntax at the sentence level. The authors’ aim is to present both the unity and diversity of the new Englishes (nonnative varieties) in particular, and their treatment of grammatical features contributes to the establishment of unity.
Trudgill and Hannah (1982) do not attempt to treat, except incidentally, the social history and institutional contexts of different varieties of English. Instead they focus on linguistic features, using a terminology and organization reminiscent of the Quirk grammar. The book groups historically related and linguistically similar varieties into chapters, and within chapters proceeds according to levels of linguistic analysis. The effect is striking: each “standard” variety is awarded a section which lists the features distinguishing it from nearby varieties and from the most closely related “major” variety, either English English or American English. The text thus fulfills the project suggested by the Quirk grammar’s compromise: Englishes that have proven themselves to have educated speakers are given thorough delineation in terms of their features. This arrangement speaks to use by English language learners and teachers, who want to identify the varieties they encounter, adapt their usage toward a particular standard variety, and possibly make allowances for the diverse English language backgrounds of other speakers (Hundt 1998: 142). Recent editions have expanded the treatment of West Indian Englishes and creoles, African Englishes, Asian Englishes, and lesser‐known Englishes.
The problems of Trudgill and Hannah (1982) are far outweighed by its utility. Nevertheless, they illustrate the difficulties in applying a descriptive approach to varieties of English. First, the text divides the world into English English and American English sectors. This is not justifiable from variationist or historical linguistic perspectives, nor does it reflect, except in the crudest political terms, the complex histories of English around the world. Second, the dividing up of the English language into standardized varieties tends to fall into national stereotyping of varieties. Statements of linguistic features are generalized broadly throughout a national area, and the inclusion of a particular variety amounts to the recognition of those varieties that have made the grade either through having an educated population or an army and a navy. Within this view, Canada and Singapore make the grade (in the third edition 1994), but Martha’s Vineyard, or Ocracoke, or Hong Kong, do not. Trudgill (2002) has in his more recent work taken pains to correct this impression. Saint Helena thus receives increasing attention in the third and fourth editions, and the Miskito Coast is discussed in the fourth edition (Trudgill & Hannah 1994: 119; 2002: 118–119). African‐American English awkwardly holds the same status as dialects of American English until the fourth edition, when it is discussed in the context of postcreoles (Trudgill & Hannah 2002: 112). The point is not that the inclusion or exclusion of a particular variety is incorrect, but rather that the attempt to describe “standard” varieties must of necessity lead to a great deal of exclusion on nonlinguistic grounds. Third, despite frequent cross‐referencing, the compression of the book does not allow for recognizing the complexity of grammatical patterning, particularly overlaps in usage between certain British varieties and certain American varieties, and standard/non‐standard variation (for example, when British non‐standard usage resembles American standard, or vice versa). Changes made in successive editions show that the authors are aware of these shortcomings and have tried to ameliorate them. The problems derive from the Quirk grammar compromise and from the attempt to merge the function of a descriptive grammar with the representation of variation found in an historical and comparative grammar. Although new varieties are recognized and given license to exist (an improvement on the practice of the Quirk grammar), they are described in a way that can give rise to false impressions. An uncritical reader could well develop the following misconceptions: that world Englishes are derivative of British and American English, that they are arrayed as a sphere of particularized satellites diverging from the two dominant core varieties, and that national sovereignty alone authorizes linguistic variation.
This set of views can be traced in the reluctance of descriptions of world Englishes to admit the possibility of profound grammatical variation. It is almost as if to admit divergence from the norm would be a national disgrace and grounds for ejection from the commonwealth of the English language. The following set of comments from entries in McArthur (1992) give a sense of this:
Table 28.1 Describing the grammars of world Englishes.
McArthur ()1992
| African English | “The discussion of syntax tends to centre on deviation from standard English rather than a consideration of distinctively AfrE forms.” (21–22) |
| Australian English | “There are no syntactic features that distinguish standard AusE from standard BrE, or indeed any major non‐standard features not also found in Britain, but there are many distinctive words and phrases.” (92) |
| Canadian English | “Where CanE differs grammatically from BrE it tends to agree with AmE. However…Canadians are often more aware of both usages than Americans.” (181) |
| Indian English | “There is great variety in syntax, from native‐speaker fluency (the acrolect) to a weak command of many constructions (the basilect).” (506) |
| Anglo‐Irish | “Standard Anglo‐Irish is close to the standard BrE varieties. Non‐standard Anglo‐Irish syntax has six features also found outside Ireland.” (68) |
| New Zealand English | “Standard NZE is to all intents and purposes the same as standard BrE.” (696) |
| Pakistani English | “Distinctive grammatical features relate to uses of the verb, article, relative clause, preposition, and adjective and verb complementation, all shared with IndE. Features of the indigenous languages influence use of English and code‐mixing and code‐switching are common, including among the highly educated.” (742) |
Syntax as a topic seems to require linguists to assert that the variety they are describing has a standardized form which does not deviate from standardized forms of British or American English. Only Indian English and Pakistani English are described as allowing variation in grammar. One might assume that this table reports directly on the nature of the varieties in question, but again, the pressure of the same ideologies and approaches that inform the Quirk grammar cannot be ruled out of consideration.
How do the ideologies in question define descriptive grammar with respect to varieties? First, there is the pressure to be included in (literate) “Standard English”; this leads to the “commonwealth” statement that the grammar of the variety in question does not diverge from Standard (British or American) English. Second, there is nationalist pressure to identify a few characteristics that establish national identity. Third, when incontrovertibly profound grammatical variation is encountered, as in pidgins and creoles, the variety in question is reclassified as outside of “English.” As Quirk et al. (1985: 28) state, “It is a matter of debate, and to some extent politics, whether these should be regarded as falling within the orbit of the English language” (Compare Mühlhäusler 1996: 99–103, on the politics of labeling pidgin languages).
The grammatical description of world Englishes over the past 30 years has seen the convergence of techniques of data analysis from several fields. These include sociohistorical linguistics, the development of register‐specific (spoken vs. written) analysis of syntactic patterning, the detailed description of nonstandardized varieties, the application of sociolinguistic methodology to the grammatical variation in world Englishes, and the development of the International Corpus of English (ICE).
Romaine (1982) establishes a method for sociohistorical linguistics with reference to variable relative markers in Middle Scots, showing that the sociolinguistic study of syntactic change requires the use of corpora. Denison (1998) gives a thorough discussion of syntactic change during the present‐day period, using a descriptive terminology similar to the Quirk grammar. He makes impressive use of corpora and casts the widest possible net for nonstandard forms. His discussion of relative clauses includes mention of genitive that’s (“the house that’s roof was damaged”) and nonstandard as and but (“…not one of the children but was relieved to find that…”) (Denison 1998: 279–282). This chapter is an important resource for those who would require evidence for grammatical variation omitted from the present‐day syntactic and descriptive accounts.
Miller and Weinert (1998: 75–76, 397) demonstrate on the basis of cross‐linguistic data that the syntax of spoken and written language differ from one another significantly. They introduce the concept of magnasyntax to refer to the heavily documented morphology and syntax of the written English tradition. (377). This work interprets the difference between spoken and written, but also standardized and nonstandardized varieties, as a function of register and degree of analytical focus.
A range of recent studies employ various perspectives to undertake the serious systematic description of nonstandard English morphology and syntax, including Henry (1995) on the syntax of Belfast English, Wales (1996) on personal pronouns, and Anderwald (2002) on negation. These studies combine theoretical sophistication with a critical attention to detailed grammatical description. Cheshire and Stein (1997) theorize the differences between the syntax of standardized and nonstandardized varieties in terms of sociolinguistic function. Their contributors include valuable descriptive detail regarding the morphology and syntax of less‐standardized varieties (Seppäinen 1997, on the genitive of English relative pronouns, and Wright 1997, on second‐person plural pronouns).
The chapters in Cheshire (1991) apply sociolinguistic methodology to many instances of grammatical variation in world Englishes. Mesthrie (1991: 464–467) illustrates the descriptive inclusiveness of this method in his analysis of relative clauses in South African Indian English, including near‐relatives (“You get carpenters, they talk to you so sweet”), correlatives (“Which one haven’ got lid, I threw them away”), contact (zero‐subject) relatives, relatives with a resumptive pronoun, and other nonstandard relative pronouns including what. Mesthrie’s data indicate that younger middle‐class females seemed to be leading assimilation to the standard (Mesthrie 1991: 472).
Hundt (1998) assesses the degree of independence of New Zealand English language norms, using American, British, and New Zealand corpora to compare a list of morphological, syntactic, and lexicogrammatical variables. She tests numerous generalizations about the specific grammatical features of New Zealand English and adopts a pluricentric model to trace similarities with Australian, American, and British varieties. Her balanced conclusion recognizes the contingency of the notion of “a variety” while arguing that the data support a grammatical distinction between New Zealand English and these other varieties.
Gisborne (2002) undertakes a similarly open‐ended grammatical description while examining the contribution of relative clauses to the definition of Hong Kong English as a discrete system. He lists six types of relative clause: contact relatives, participial relatives with a relative marker, where‐relatives with a directional as well as locative sense, the omission of prepositions, resumptive pronouns, and the absence of restrictive/nonrestrictive contrast, and then considers the second type of relative in the context of the morphosyntactic feature system of Hong Kong English, using examples from ICE‐HK. Gisborne’s approach suggests that a sensitive application of an analytical framework to language data can result in a description which balances system and variation. He does not define Hong Kong English by the overgeneralization of one variable feature, nor does he ignore this variation in order to conform to the prestigious systematicity of another form of English. Cautious description of this sort will be extremely important in realizing the full potential of linguistic corpora in describing world Englishes.
A development in the descriptive grammar of world Englishes with far‐reaching implications is the use of corpus linguistics in connection with the International Corpus of English, introduced and explained in Greenbaum (1996). Meyer (2002: 46–53) presents the methodology of corpus linguistics, in particular discussing the emphasis that corpora place on native vs. nonnative speakers and the role of editors in shaping newspaper English in different varieties. He notes the problems that corpora have in reflecting sociolinguistic variation, especially dialect differences. Nelson, Wallis, and Aarts (2002) list recent research on British English using the ICE‐GB corpus. The grammatical model conforms to the Quirk grammar’s categories with some modification, and the text makes reference to the differences between this model and Government and Binding theory. Hundt (1998: 130) warns that the application of statistical tools on corpora is not useful for the discovery of grammatical differences between varieties. ICE will prove invaluable as a testing ground for grammatical and variational hypotheses, but the quality of these will still depend on the ingenuity of linguists. A good deal of ingenuity and careful description is to be found in the monumental study of the morphology and syntax of varieties of English in Kortmann, Burridge, Mesthrie, Schneider, and Upton (2004).
An outstanding example of the value of the description of world Englishes for linguistic theory in general can be found in Van Rooy (2008). Taking the view that “English has adjusted itself to new ecologies” (338), Van Rooy finds that earlier descriptions of aspect in Black South African English, couched in terms such as “the extension of the progressive” (340) exemplify the comparative fallacy as identified in second language acquisition studies. The comparative fallacy refers to those linguistic descriptions that negatively measure the distance between the variety described and some other, putatively more standardized variety. As an alternative, Van Rooy approaches the problem from the perspective of Emergent Grammar (Hopper 1998; Bybee 2006). In this approach, the language system does not exist in advance of the communicative event (a priori), but is seen to emerge through language users’ communicative practice in the context of situation. In this way Van Rooy not only accommodates his earlier observations of the progressive in Black South African English as having a persistitive (extensive duration) meaning (Van Rooy 2006), but he also offers a contextual analysis of specific texts, involving the interaction of idealization, frame of reference, and the relative importance of spatial versus temporal grounding. The approach is sociolinguistically and contextually sensitive as well as empirically and quantitatively grounded in language corpora. It enables Van Rooy to describe nuances of grammar within the context of situation, without resort to an external language system.
Language descriptions will continue to benefit from advances in the size, complexity, and refinement of linguistic corpora. However, these must be used with care, as they reflect language attitudes within national education systems, publishing industries, and media. Descriptive grammars will continue to be a flashpoint, as they are accorded prescriptive weight by their consumers. They can embody resistance to nationalist hegemony and traditionalist doctrine. On the one hand, they can be symbols of vibrant national literature, media, and intellectual life, and on the other, they can be a narrow nationalist stereotype, a betrayal of the richness and complexity of language heritage, language variation, and the negotiation and renegotiation of identities inherent in language. Language users and linguists would do well to allow this dialectic to inform their language practices.