意味論研究会2014年2月24日

とき：

2014年2月24日 2:00pm

ところ：

慶應義塾大学 三田キャンパス 南別館７階

発表１

講演者：

有井　巴 (東京大学／JSPS)

タイトル：

Setting the Standard in the Acquisition of Japanese Differential Comparatives

概要：

Comparison problems are notoriously difficult for kids, which require them to determine a differential between two sets, e.g., “A boy has three apples and a girl has two apples. How many more apples does the boy have than the girl?” Even young school children wrongly answer the number of the boy’s apples, three, instead of answering the differential between the two sets of apples, namely one. Children seem to interpret the differential phrase, how many as absolute. In this presentation, I will show that the absolute interpretation is not caused by children’s limited cognitive resources for quantifying a differential. Then, showing that Japanese children also assign similar non-adult interpretation to verbal comparatives (e.g. X-wa 2-ko fue-ta ‘X increased by two’) and locative comparatives (e.g. X-wa Y-yori 2cm ue-ni-iru ‘x is 2cm above y’), I will argue that when they interpret differential comparatives, children set the standard of comparison as the absolute zero by default.

発表２

講演者：

島田純理 (慶應義塾大学)

タイトル：

Entailments and Implicatures of Cumulative Sentences

概要：

A standard neo-Gricean theory says that the numeral “seven” in “John kissed seven girls” is taken to mean “exactly seven” due to scalar implicature. Krifka (1999) argues that essentially the same mechanism gives an “exactly” reading to the cumulative sentence in (1).

(1) Three boys kissed seven girls.

Krifka also claims that a pragmatic account may render the numerals in a cumulative sentence an “at least” or even “at most” interpretation in different contexts, so in (2), “three” and “seventy” are understood as “at most three” and “at least seventy”, respectively, because the sentence is meant to describe a bias in a statistical distribution.

(2) In Guatemala, three percent of the population owns seventy percent of the land.

To evaluate these accounts, I have investigated scalar entailment relations among cumulative sentences in two kinds of contexts and in different combinations of the count/mass types of the nouns involved. It turns out that (1) does not have the “exactly” reading as claimed and the observed interpretation of (2) is straightforwardly predicted without mention of a statistical bias. I will show that the revealed entailment relations predict interesting but controversial scalar implicatures for cumulative sentences.

発表３

講演者：

Orin Percus (Université Nantes)

タイトル：

Hidden indexicals

概要：

I will defend a Kaplanian picture on which semantic values are determined with respect to a context parameter, and I will present evidence that certain uses of descriptions trace back to a semantics implicating this parameter. I will suggest that, in these cases, silent indexical elements are sitting in argument positions projected by predicates located inside the description. I will discuss a number of instances where I think that adopting this position could yield insight.