Most of the discussion in this book has been concerned with non-standard varieties of Spanish, since it is there that most cases of variation and change are to be observed. However, account must also be taken of the standard varieties, which are perhaps best regarded as highly unusual forms of language, in the sense that they are rare (all humans use language, but only a minority use a standard language) and recent (they arose only in the last few thousand years of the multi-millennial history of human language). The reason why the process of standardization has an important place in any treatment of linguistic variation and change, in Spanish and other languages, is that an essential aspect of this process is reduction in variation within certain high-prestige varieties (see 7.2). However, paradoxically, we should not ignore the fact that for speakers of low-prestige varieties, the establishment of a standard may imply an increase in the range of variation available, since variants from the standard may enter the speech of non-elite groups and be added to the competition among pre-existing variants (see 7.3).

Although we have seen (1.1–2) that variation is inherent in language, the process of standardization may, in principle, reduce variation to zero in the variety which is subject to it. This elimination of variants, naturally, applies to instances of variation which are due to the normal effect of changes which are working their way through society (3.4), as well as to instances of variation which are more stable and long-established.

The large majority of Spanish-speakers today, perhaps 300 million of them, are to be found in the Americas, in a vast area which stretches from the southwestern United States to the far south of the southern continent, not to mention the large cohorts of Hispanophones to be found in major US cities such as New York. We find, as we expect, geographical variation across this area, although mutual intelligibility among varieties is rarely threatened, and certainly not among educated and urban speakers.

Until recently, detailed linguistic studies of American Spanish have most usually been focused on the language of individual countries, and the linguistic atlases so far published continue to be oriented in this way (see ALEC 1981–3, ALESuCh 1973, ALM 1990–4, Navarro Tomás 1974). Even excellent surveys of the language of the whole of Spanish America, such as John Lipski's (1994) or the collaborative work edited by Manuel Alvar (1996b), are, at least in part, internally organized on a country-by-country basis. The frequent implication (although disavowed in the best studies) is therefore that the features described have boundaries which are co-terminous with those of the country concerned. This is not so; in accordance with normal distribution, each feature observed in Spanish America occupies its own area, which rarely if ever coincides with the area of any other feature, let alone with political boundaries. That is to say that we are dealing here, as in northern Spain and many other parts of the world, with a dialect continuum which is intersected by the frontiers which separate one republic from another.

We have seen in Chapter 2 that all languages exist in a state of orderly heterogeneity, whether one is considering the spatial, the social or the diachronic aspects of variation, and in Chapter 3 we have considered the way in which such organized variation frequently determines the way in which language change proceeds. Many of these general issues have been illustrated with data taken from the Peninsular languages, but in the present chapter we come to a more systematic consideration of the distribution of linguistic features in the Spanish Peninsula. First we shall consider geographical variation, seeking an explanation for the main patterns of distribution of these features across the Peninsula. Then we shall turn to social aspects of variation, where reasons for particular patterns of heterogeneity will be hard to find, but where we shall study some of the many striking instances in which linguistic and social variation are correlated.

Geographical variation

The present geographical distribution of features in the Peninsula has been determined by two sets of circumstances, namely the existence of a northern dialect continuum, and the territorial expansion of northern varieties which accompanied the reconquest of Islamic Spain. The northern dialect continuum stretches across the northern third, approximately, of the Peninsula, and is part of the Romance dialect continuum which extends from northwestern Spain into France and thence into Belgium, Switzerland and Italy (see sections 1.1.1 and 4.1.2).

A common perception, among those who are not linguists, is that there is some difference in kind between a ‘language’ and a ‘dialect’. The question is often posed in the following form: ‘Is x a language or a dialect?’, where x is some such label as ‘Valencian’, or ‘Asturian’. And it is a question which the linguist, as linguist, cannot answer, first because of the insuperable difficulty of defining the concepts language and dialect (see 2.1 and 2.2), but secondly because any difference between these concepts resides not in the subject matter of linguistic description, but in the social appreciation accorded to particular codes of communication. The historical linguist will make it clear that every code to which the label ‘language’ is attached (e.g., ‘the Spanish language’, ‘the English language’, ‘the French language’, ‘the Latin language’) has its origins in what would usually be called a ‘dialect’, loosely defined in terms of geography (as the speech of a particular locality or area) and in terms of social class (as the speech of a particular social group, usually the dominant, educated, classes). Thus, the French language has its origins in the speech of upper-class Paris, specifically of the Court. If ‘dialects’ can gradually become ‘languages’, it follows that there cannot be any difference of kind between these concepts, but only differences of degree.

As this book unfolds, we will see that the verbs of the Athapaskan family exhibit two attributes, outlined in chapter 1. First, global uniformity exists – morpheme order is similar across the family. Second, so does local variability – some variability in morpheme order occurs both within a language and across the family. In this chapter I examine some possible hypotheses to account for both of these properties.

Two Hypotheses

One can imagine different ways of accounting for the fact that the Athapaskan language family as a whole exhibits both global uniformity and local variability. I outline two here. Under both hypotheses, global uniformity has its origins in the languages having a common source, but they differ as follows. Under one hypothesis, it is the common source and history of the languages that accounts for the global uniformity. It is an accident of history that certain properties remained stable across the language family while others were subject to change within individual languages. I call this the template hypothesis. Under a second hypothesis, global uniformity finds its origins in the languages having a common source, but results additionally from principles of universal grammar, which might be either diachronic or synchronic. I call this the universal or scope hypothesis.

In this chapter I examine why certain aspects of verb morphology are invariant while others have been susceptible to change, both across languages and within a particular language.

In this chapter I examine the ordering of material within the functional complex. In all of the Athapaskan languages except Hupa and Kato (see chapter 10), object pronominals systematically precede number subjects and aspectual material, be it subsituation aspect, situation aspect, or viewpoint aspect. Qualifiers are generally sandwiched between number subjects and aspect. Finally, first and second person subjects appear at the right edge, to the right of aspect. The basic schema is shown in (1).

1. (1) object – number subject – qualifiers – aspect – 1/2 subject

This schema is simplified, with much omitted. Recall that in chapter 10 I argued that the ordering of subjects and objects in general is not surprising. I am concerned in this chapter with ordering among positions; see chapters 10–12 for discussion of ordering within each position.

In this chapter I argue that the overall ordering in (1) is consistent with the scope hypothesis. Scope predicts some fixed orderings, such as objects appearing to the left of situation aspect, and also allows for some variable orderings.

The Ordering of Objects and Situation Aspect

Background

One striking fixed ordering across the languages involves objects and situation aspect – this is the only possible ordering across the family. Objects are consistently at the left edge of the functional complex (with the exception of Hupa and Kato; see chapter 10).

I have so far examined the lexical, pronominal, and aspectual complexes of the verb. A remarkable stability exists across the family in ordering within these complexes; differences between languages can be largely attributed to the noninteraction of systems and to the different importance assigned to contradictory universals. In this chapter I investigate ordering inside the qualifier complex. It is here that the greatest variation in morpheme ordering is found both across languages and within a particular language.

The Qualifiers: Content

Unlike the argument and aspect systems discussed in the previous two chapters, the qualifiers do not form a functionally homogeneous class. As discussed in chapter 9, qualifiers are of several types. Some are aspectual – inceptive, conative, egressive, inchoative, negative. Others are noun class markers that index properties of patient/theme arguments. A third set marks noun class qualities, but does not share all properties with noun class markers. Others mark classes of descriptive stative verbs in some languages. Another one interacts with the middle voice system in marking reflexive and self-benefactive. A final set is difficult to assign meanings to; they might be historical noun class markers that are now frozen in use with particular verbs. This last set is generally called thematic in the Athapaskan literature.

The Noun Class System

Many Athapaskan languages show some productive use of noun class markers, usually called gender in the Athapaskan literature, although the system is more productive in some languages than others.