The diachrony of horizontal constructional relations
Language change as a “threat” to horizontal constructional relations
In the previous sections, a number of horizontally organised constructional networks have been discussed, to wit, vowel features, verbal inflection, position of the finite verb, case frames in experience predicates, and integration of subordinate clauses. Evidently, each of these networks can get tousled when one of the differential values increases its scope. Suppose all vowels become [+front], then the difference between [+ front] and [- front] is no longer able to discriminate words. Or suppose that analogical leveling in verbal inflection paradigms blots out the person desinences. In that case, the inflectional paradigm ceases to exist.
This is not so far-fetched: the history of the verbal inflection in Proto-Germanic, Dutch, English and Afrikaans is largely one of far-reaching analogical leveling.
In the syntactic case studies, introduced in section 2, language change similarly can bring down the network of horizontal relations. In principle, Dutch could extend its V2 to subordinate clauses, thereby obliterating the semiotically relevant distinction between V2 and Vn. Again, this is not far-fetched, as such a change occurred in English. In the following sections 3.1 and 3.2, a closer look is taken at the history of the other two syntactic examples, as these have in fact come under pressure by diachronic drifts elsewhere in the language system. The case frames in Dutch experience predicates came under pressure in Middle Dutch, and the integration level of subordinate clauses has recently come under pressure.
In both cases, however, it can be shown that the semantic differences that are formally expressed by the horizontally related nodes, survive. The reason is that these semantic differences are degenerately expressed. Crucially, the degenerate strategies are not instances of what is traditionally called “renewal” (Hopper and Traugott 2003:122-124): it is not the case that language users come up with new ways to express semantic distinctions which they used to express differently. Rather, grammatical strategies that already existed in the language and already played a role in the functional domain at issue are seized upon to “rescue” the system.