Correlations

In the last sections we have seen various ways in which the morphologies of languages can be classified. Typologists are interested in more than classification, however. They are also interested in seeing whether there are any predictable correlations between particular morphological characteristics or between morphological characteristics and other (syntactic or phonological) aspects of grammar. In other words, they seek to find whether there are any patterns to the kinds of morphology one finds in a language, and if there are, why those patterns might exist.

For example, as Whaley (1997: 131) points out, isolating languages usually have rigidly fixed word orders. This correlation makes perfect sense: if a language has no morphological way of marking the function of noun phrases in a sentence, those functions must be signaled by the position of a noun phrase in a sentence.

Linguists such as Joseph Greenberg (1963) and Joan Bybee (1985) have looked at many languages and observed that there are several other correlations to be found. For example, Greenberg noted the two patterns in (50):

If a language has separate morphemes for number and case, and if both are either prefixes or suffixes, the number morpheme almost always occurs closer to the base than the case morpheme.

Observations such as these are called implicational universals. Based on observations of lots of languages, they are not true in every language, but they are true in a statistically significant number of languages.

Bybee (1985) also observed a statistically significant trend in the ordering of inflectional affixes in the languages of the world. What she noticed is that in languages with a number of different inflectional affixes on verbs, those affixes tended to come in a particular order. For example, if a language exhibits both tense and person/number affixes, the tense affix usually comes closer to the verb stem than the person/number affix. And if there are aspectual affixes, these tend to precede tense affixes.

We must still ask, however, why these particular correlations should exist. Bybee, for example, claims that the order of inflectional morphemes on verbs has something to do with their relevance to the verb itself. We might think of this in terms of the concepts of inherent and contextual inflection that we discussed. For example, tense and aspect are inherent categories for verbs, but person and number are contextual: they signal agreement with one or more of a verb’s arguments (its subject or object, for example). So inherent inflection comes closer to the verb stem than contextual inflection. The second of Greenberg’s implicational universals might be explained in the same way. Number is an inherent inflection on nouns, but case is contextual, and inherent marking comes closer to the noun stem than contextual marking. Whether this is generally the case, however, is something that will require looking at many more languages.