• Aucun résultat trouvé

The reactive use of bound additives is an inherently interactive use on which the host of the bound additive is again a tensed F-marked verb. This use is exemplified in (196). In contrast to the polar use, the reactive use is characterised by the appearance of bound additives in agreeing reactions to an assertion made by another speaker. It is natural for the answer to be partly elided, as in (196a), and not fully spelled out, as in (196b). Note that on this use, the presence of the bound additive is not obligatory; a confirming reaction can just as well be expressed without it.

(196) Reactive use:–kin

‘Joni likes olives. – So he does.’

b. ?Joni

‘Joni likes olives. – So Joni does like olives.’

When the reaction targets a negative polarity assertion, as in (197),–kAAnis used instead of–kin.

As with–kin, both elided (a) and full (b) reactions are possible. While the presence of–kAAnis optional, the agreeing reaction is naturally formulated in a slightly different way in its absence, as shown in (197c).

‘Joni does not like olives. – No, he doesn’t.’

b. ?Joni

‘Joni does not like olives. – No, Joni doesn’t like olives.’

4.3. Reactive use

‘Joni does not like olives. – No, he doesn’t.’

The reactive use seems intimately connected to the confirming basic use presented in section 4.1.2; after all, both express agreement with or confirmation of an antecedent that was presented previously. The two uses are nevertheless not fully alike: while the confirming use is possible with unbound and bound additives, the reactive use is only possible with bound additives, as illustrated in (198).

(198) Reactive use is unavailable with unbound additives a. #Joni

Int. ‘Joni likes olives. – So he does.’

b. #Joni

Int. ‘Joni does not like olives. – No, he doesn’t.’

In addition to agreeing reactions that involve a bound additive and a tensed verb, it is also possible to express agreement by attaching–kinto an elative-marked ‘opinion holder’ phrase, as in (199a).

Note that the unbound additivemyösmay also be used in these structures, as shown in (199b). If an overt opinion holder phrase is present, the presence of some additive (bound or unbound) is obligatory, as shown in (199).10

10The negative polarity equivalents of (199) are shown in (i).

(i) Expressing agreement with unbound and bound additives and an opinion holder phrase a. Minu-sta

‘I think Joni is not worse. – Neither do I.’

b. Minu-sta

‘I think Joni is not worse. – Neither do I.’

(199) Expressing agreement with unbound and bound additives and an opinion holder phrase a. Minu-sta

I-ELA

Joni Joni.NOM

o-n

be-PRES.3SG

huono-mpi.

bad-COMP

– Niin so

minu-sta-*(kin) I-ELA-ADD

‘I think Joni is worse. – So do I.’

b. Minu-sta I-ELA

Joni Joni.NOM

o-n

be-PRES.3SG

huono-mpi.

bad-COMP

– Niin so

*(myös)

ADD

minu-sta I-ELA

‘I think Joni is worse. – So do I.’

As both unbound and bound additives are available when agreement is expressed through the use of an opinion holder phrase, as in (199), and these examples can quite naturally be analysed as instances of the basic use (where the opinion holder KP is the associate of the additive), it is worthwhile to consider whether the examples where reactive bound additives attach to tensed verbs could also be analysed in this way. In other words, one could assume that when the bound additive attaches to a finite verb, focus is nevertheless on a covert opinion holder, so that what are contrasted are speaker A’s statement and speaker B’s statement (the contents of which are the same). If this were the case, then it would have to be assumed that the host of the bound additive isnot necessarily its associate – in contrast to what we have been assuming so far, essentially following Karttunen and Karttunen (1976). Moreover, an analysis where the F-marked associate would somehow be deleted at PF would not fit in well with the general idea that it is Given material that usually undergoes ellipsis, not focused material.

Regardless, this discussion raises an interesting point: on the reactive use, speaker B agrees with something that speaker A has said, and the propositional contents of A and B’s statements are non-distinct. In other words, although the bound additive attaches to a tensed verb, the an-tecedent does not correspond to the polar opposite of the prejacent. If the prejacent and the antecedent can be non-distinct, it must be the case that the distinctness requirement that is usu-ally taken to hold of additives does not hold of allusesof bound additives. As we saw in section 3.1.3, it has been proposed before that the distinctness condition is a pragmatic effect that is not hardcoded into the semantics of additivity (Beaver, 2001). If this is the case, then it is possible that only bound additives have a reactive use because operator associated with unbound addi-tives does not accept non-distinct antecedents, while that associated with bound addiaddi-tives does.

Before concluding this section, I would like to note that the reactive use is not restricted to ac-tual conversations, but also appears in other contexts of dialogue, such as texts (where the writer is the speaker, and the reader is the hearer). In (200a), for example, the speaker is signalling their