LingLunch : Dejan Matic

Dejan Matic (Münster)
Explicitness

Research on different amounts of information that can be linguistically conveyed has been particularly en vogue in the past decade, but the tradition of quantifying information is much older, going back to early attempts at information theory (Zipf 1949, Shannon 1948). The foundational observation of most approaches is that utterances and segments of utterances can vary immensely in terms of the amount of explicitly encoded information, while the basic conceptual device to account for this is the idea that the degree of explicitness is a product of the balance between effort and accuracy, such that more predictable and frequent information is as a rule less explicitly encoded, since its accurate transmission is ensured even with a minimal coding effort. This idea has been instrumental in formulating great many formal and pragmatic generalisations. Despite important differences, all approaches share the assumption that the measure of explicitness is a function of human cognition and therefore universally valid. I will present a tentative research programme which is based on a modification of this assumption. I will try to show that frequency and predictability can only be conceived of as interpreted against the background of expectations, which are in their turn rooted in conventions. For a certain degree of explicitness in a certain context type to count as predictable and thus less informative, it must be conventionalised in the given context. If degrees of explicitness are conventionalised, be it because of structural features of the language used, social norms, or for any other reason, then they must be at least partly language-specific or even vary across sociolects and registers. In other words, explicitness is not merely a function of cognition, but also of linguistic forms and norms of language use. I will illustrate this idea with a couple of select examples of variable explicitness across languages and registers and discuss possible methods of empirically investigating it.