Contrasting causal connectives on the Speaker Involvement Scale

In our talk we would like to contrast Dutch and French causal connectives on the Speaker Involvement Scale (Degand & Pander Maat, to appear; Pander Maat & Degand, to appear). This scale is an alternative proposal to classify the use and meaning of connectives. Going beyond dichotomous and trichotomous classifications, we have proposed to represent (causal) coherence relations and connectives in a scalar way This scalar representation reflects the fact that (causal) connectives are not strictly domain specific, but that they nevertheless impose constraints on the contexts in which they can occur, with some contexts being more “natural” than others. In addition, a number of causal connectives seem to take an intermediate position between the traditional categories. According to us, this situation is an indication for the need of a scalar perspective on the spectrum reaching from nonvolitional causality in the content domain to epistemic and speech act causality. The scale we have developed is one of speaker involvement (SI), on which the inherent expressive power of connectives can be represented. Our hypothesis is that the different causal relations can be ordered along a scale from minimal to maximal speaker involvement. SI refers to the degree to which the present speaker is implicitly involved in the construal of the causal relation. More specifically, SI increases with the degree to which both the causal relation and the related segments vehicle assumptions and actions of the present speaker. Four characteristics of coherence relations may enhance the prominence of speaker assumptions in the relation, and thus enhance the level of Speaker Involvement of the relation: The involvement of a conscious protagonist, a lack of isomorphism between the relation and states of affairs in the real world, proximity of the relation to the present speaker and the time of speaking, and the implicit vs. explicit realisation of the protagonist. The different causal relations we distinguish are, in order of increasing SI: causal non-volitional and volitional content relations; causality-based and noncausality based epistemic relations, and causal speech-act relations. Point of departure of our contrastive analysis are connectives which are very close in meaning and seem to be easily substitutable by one another within one language, like French donc and dès lors (‘so/hence’), car and parce que (‘for/because’), or Dutch dus and daarom, or between languages like the supposed translational equivalents puisque and aangezien (‘since’). In addition, these connectives also show highly diverging frequencies in a newspaper corpus. The question we would like to tackle in our presentation is whether these divergences in frequency and subtle meaning differences are reflected in diverging SI distribution patterns. This presupposes to uncover the semantic profile of the given connectives as well as their interaction with the surrounding discourse. In our view, analyses of the SI potential inherent to connectives cannot do without systematic corpus analyses combining distributional data and semantic intuitions. In our presentation, we will develop how we proceed to determine the SI of a connective and how this SI level accounts for the substitution effects created in (1) (when addressed to a traffic agent who pulls you over because you ignored a ‘one way’ traffic sign), while (2) is perfectly acceptable.