[Linguistik] Invited talk on 14.11.18: Scalar implicature in comparatively modified numerals

[Dr. phil. Mingya Liu]

Dear all,

Next Wednesday, there will be a talk by Dr. Christoph Hesse (Leibniz-Center for General Linguistics, Berlin) between 12-14:00 in Room 50/E07 (Note: it will be preceded by a short student presentation of her MA project though). Everyone is cordially invited! If you would like to talk to him in additon, just let me know.
----------
Title: Scalar implicature in comparatively modified numerals
Abstract: It is widely assumed in the literature that numerals modified by the comparative quantifiers 'more than' and 'fewer than' do not generate scalar implicatures as the comparative quantifiers indicate that the speaker is ignorant of the exact number (Fox & Hackl, 2006; Breheny, 2008). However, speakers routinely employ experiential world knowledge to draw scalar implicatures in the presence of comparative modifiers. For instance, in the context of 'an exceptionally hot summer', speakers will use their knowledge of average summer temperatures and record temperatures to construct a range of potential values. Other contexts, such as 'the number of signatures on a petition', are less informative about potential values and speakers may lack sufficient information to form (prior) expectations.
We start from Cummins, Sauerland, & Solt (2012), who show that scalar implicatures are available from comparatively modified numerals. For example, the average number of people is much higher in (1b) than in (1a):
(1) a.  More than 110 people attended the public meeting.
b.  More than 100 people attended the public meeting.
Cummins et al. (2012) argue that this is because 100 is more round than 110. Based on our own experiments, we present a novel account which argues that the numerical property driving scalar implicature is magnitude.
 
In a series of three experiments, we followed up on Cummins et al. (2012) by (a) testing a greater range of numerals, (b) including absolute magnitude and order of magnitude as factors in addition to their factor of roundness. Results suggest that---contrary to Cummins et al. (2012)---, in impoverished, less informative contexts, scalar implicatures results from a Weber fraction on numerical vagueness, which might be linked to the approximate number sense (Dehaene, 2011).
--------------------
Best,
Mingya Liu