The Weekly Newsletter of MIT Linguistics

Benjamin Spector at MIT

Benjamin Spector (Centre national de la recherche scientifique) will be visiting the department this week. In addition to his Colloquium talk on Friday, he will be offering a mini-course on the relationship between logical entailment and contextual knowledge. Details below:

Speaker: Benjamin Spector (CNRS)
Title: Understanding the interactions between contextual knowledge and logical entailment: scalar implicatures and presuppositions
Time: Wednesday, October 11th, 12:45-2:15pm and Thursday, October 12th, 3:30-5:00 pm
Place: 32-D461 (Wed), 4-237 (Thurs)

I plan to revisit a number of problems/puzzles in semantics/pragmatics which all have to with the relationship between logical entailment and contextual knowledge. These puzzles include phenomena usually understood in terms of

– Maximize Presupposition
– Interactions between scalar implicatures and presuppositions
– Blind scalar implicatures

(i) Maximize Presupposition

Why can’t you say “A president of the US came”? Standard answer: because you’d better say “The president of the US came”. What’s responsible for that is Maximize Presupposition.

(ii) Interactions between scalar implicatures and presuppositions

Relevant reading: Spector & Sudo, Presupposed ignorance and exhaustification: how scalar implicatures and presuppositions interact

(1) Mary is un/aware that some of the students smoke

(1) suggests that in fact not all of the students smoke. The most straightforward theories of scalar implicatures fail to predict this.

(iii) Blind implicatures (Magri oddness cases)

(2) Every professor gave the same grade to all of their student.
#Mary, who is one of the professors, gave an A to some of her students.

The puzzle here is that in the context of the first sentence, the second one is equivalent to ‘Mary gave an A to all of her students’. Yet it is infelicitous, feels nearly contradictory, possibly because ‘some’ triggers a ‘not all’ inference. However, this inference is not expected if the notion of informativity used to compute scalar implicatures takes into account background knowledge: in this context, the sentence with ‘some’ is as informative as the sentence with ‘all’, so there should not be any particular pressure to use ‘all’ rather than ‘some’.

Relevant reading: Giorgio Magri, Another argument for embedded scalar implicatures based on oddness in downward entailing environments.

My goal will be a) to present some of the recent literature about these topics, b) to discuss problems with current theories, c) time-permitting, whether it is possible to unify these different cases, based on on-going work by a student of mine, Amir Anvari.