The varieties of English in the Americas, like everywhere else, correlate with the parameters of region, social class, and style, and in most cases it is impossible to draw clear-cut, qualitative distinctions. Typically, select features tend to occur more frequently in certain varieties than in others; hardly ever are there any uncontroversial shibboleths to be observed (for instance, even the prototypically Southern pronoun y’all has been shown to be spreading outside of the South; Tillery, Wikle and Bailey 2000). Nevertheless, it is possible to state some broad tendencies which as such are of interest.

Broadly speaking, phonology tends to vary regionally while grammar varies socially in the first place. Pronunciation differences delimitate dialect regions of North American English most clearly and consistently, and the contributors to the pronunciation papers point out local, regional and supraregional phonological or phonetic features. Of course, accents go by social class as well, but the standard assumption for American English is that even educated speakers, from certain regions at least (most notably New England and the South), at times use regional pronunciation characteristics and thus speak “with an accent”; hence, despite the persistent belief in a homogeneous “General American” accent or notions like “network English” there is in fact no single American norm of pronunciation that corresponds to RP in England, being a non-regional class dialect. (Kretzschmar, in this volume, defines a “Standard American English” as an accent deliberately held free of features associated with particular regions.) In contrast, the phonologies of Caribbean varieties of English are underresearched – the strong focus of the discipline upon creole genesis, reflected in the grammar of creoles, has made this a Cinderella of creole studies (Plag 2003 deliberately sets out to remedy this situation). Clearly there are both supra-regional features and tendencies and regional or local forms of pronunciation, but no systematic survey of such similarities or differences is available to date.

Unlike phonology, in North American English grammatical variation is primarily socially determined. This is perhaps less true for nonstandard morphology (like irregular nonstandard verb forms or noun plurals), where dialectological research has identified some regional correlations (Atwood 1953), and a small number of minor syntactic patterns may be pinned down to specific regions; but basically using nonstandard grammar betrays a speaker’s social class background, not his or her regional whereabouts. Many of these patterns (like multiple negation, left dislocation, or intonation-marked but uninverted questions) are not even distinctly American but constitute elements of informal English, presumably British-derived, in many countries around the globe. Quantitative distinctions from one dialect to another exist in America (i.e. some features occur more frequently in certain regions or contexts than others), but basically it is the particular configuration, the specific sub-set of such forms and patterns available in a given region or community, that identifies and distinguishes individual varieties of North American English.

This particular aspect, the uniqueness of the mixture of forms at a given location rather than a diagnostic role of any individual variant, can be stated for the Caribbean situation as well, although the creole continua found there provide for quite different, and certainly no less complex, linguistic ecologies. As is well known, creole grammars are characterized first and foremost by the use of pre-verbal markers for categories of tense, mood and aspect, in addition to several other “characteristically creole” features (e.g. specific copula uses, the functional conflation of pronoun forms, or serial verb constructions), while, conversely, they display very little inflectional morphology on verbs, nouns, or other word classes. Some of these forms characterize certain sub-regions (most importantly, a few forms appear to mark off the eastern as against the western Caribbean), but the most important parameter of variation here is the class and style stratification that is captured by the notion of a creole (or “post-creole”) continuum, the systematic variation between acrolectal (or near-standard), mesolectal and basilectal (“deep creole”) choices. Bickerton (1975), following deCamp (1971), described this variation as “implicational scales”, with both lects (distinct “grammars”) and their features arranged in such a tabular format that the presence of certain forms in certain lects predicts the presence of all other “more basilectal” forms in all other “more basilectal” lects. On the other hand, several aspects of this model have been challenged in recent years, including its monodimensionality and its diachronic implications (the assumption that creoles started out as basilects and have “decreolized”, i.e. exchanged basilectal creole forms by corresponding acrolectal English forms, in the course of time). In fact, the scholarly concentration upon the putatively pure, basilectal creole has led to the paradoxical situation that basilects are at the center of creole studies even if no one has ever documented a pure basilectal creole, while mesolects, the forms that are really in use, have only recently begun to be the objects of scrupulous investigation (Patrick 1999).

Words, finally, vary readily and mostly by region, with the range of their spread extending from the strictly local through the regional to the quasi-national domain. Variation in the lexicon is considerably more resistent to systematic investigation – which is why the contributions project cover regional vocabulary only incidentally or not at all. Regional lexicography identifies the ranges and conditions of the uses of individual words (Kurath 1949; Carver 1987), and in the present context the main dictionaries to be consulted are the Dictionary of American Regional English for North America (Cassidy et al. 1985-) and the Dictionary of Caribbean Usage (Allsopp 1996) for the Caribbean.