In her study of the West-Flemish particle dè (realized as /dQ /, /dE /, /dA /, /tQ / and /tE /), Liliane Haegeman (1993) discusses sentences like (1) and (2).

Haegeman claims that dè only occurs with questions to which the answer is directly given in the situation of speaking. The hearer therefore does not have to reply to the question itself but is requested to provide the questioner (she for convenience) with information that will help her to mentally grasp what is becoming evident. For example, the questioner can unmistakably see for herself that the hearer is back, in (1), or that it is a sheep that the hearer is taking along, in (2)—she does not need a confirmation for that. Rather, by adding dè, the questioner makes clear that this situation clashes with her expectations and so urges the hearer to give an explanation.

In this paper, I would like to argue that this instruction for the hearer is part of the interpretation of dè, not of its meaning. In other words, it is on the basis of language-independent (i.e. extra-linguistic, "pragmatic") principles that the hearer in (1) and (2) knows that he does not simply have to answer the question itself—which is not always the case. In the following situations, for example, the questioner probably does not have a clue as to the answer to the "bare question" (i.e. the question without dè):

In (3) and (4), the hearer has no right to assume that the questioner already knows the answer to the bare question, and so, he does have to answer that question.

If I argue that the hearer knows this on the basis of independent cognitive principles, this is not to say that the particle dè cannot be described in semantic terms. On the contrary, I would like to propose a fairly accurate description of its meaning, following the example of Wierzbicka’s (1986) natural language paraphrase approach. A corpus of 100 authentic utterances (as opposed to Haegeman’s self-invented examples) reveals that it is impossible to posit a single invariant meaning covering all its uses. However, Wierzbicka’s approach naturally accounts for prototype phenomena.

Liliane Haegeman (1993). The interpretation of the particle da in West Flemish. Lingua 90, 111-128.