Приглашенный доклад Бенжамена Спектора

28 января на семинаре лаборатории с докладом "On the interpretation of embedded questions. Do weakly exhaustive readings exist?" выступил Бенжамен Спектор, научный сотрудник Института Жана Нико (Institut Jean Nicod) в Париже.

В докладе рассматривались интерпретации вложенных вопросительных клауз, вводимых предикатами разной степени фактивности.

Аннотация прошедшего доклада:

According to most recent work (including experimental work), embedded questions under 'know' trigger an ambiguity between a strongly exhaustive reading and a weakly exhaustive reading:

(1) Mary knows which students did the homework
a. Strongly exhaustive: If S is the plurality of all the students who won the game, Mary knows that S did the homework and that no student beyond S  
b. Weakly exhaustive: If S is the plurality of all the students who won the game, Mary knows that S did the homework and that no student beyond S 

In my talk I will argue, on the one hand, that the weakly exhaustive reading under 'know' and a number of other factive responsive verbs is an illusion, which results from the interaction between the strongly exhaustive reading and domain restriction. On the other hand, I will argue that emotive factive responsive predicates such as `surprise' do really support a weakly exhaustive reading.

My argument will be two-fold: 
a) Degree questions do not support weakly exhaustive readings under 'know' (and many other verbs), but do under surprise-type verbs.
b) One can construct examples where there is a perceived weakly exhaustive reading under `know' and yet the standard mechanism for the weakly exhaustive reading would not in fact predict this reading, while the domain-restriction view would.

I will then, more speculatively, discuss a) what constraints on domain restriction we can formulate so as to not generate unattested readings b) whether the approach I advocate can capture the so-called `no-false-belief constraint' on (apparent) weakly exhaustive readings.