top of page

The interface of scalar implicatures and effect:

A comparative study of Dutch and Japanese

 

Marina Terkourafi, Wataru Uegaki,Jens Branum,Roxanne Casiez

Leiden University

Imagine that you are offered cookies out of a plate with the words “You can have some cookies.” How many should you take? One? Two? Can you take them all? The answer can depend on the emotive valence of the conversation you are having.

 

In preliminary research using a crowd-sourced sample of US English respondents we found that participants tended to interpret “some” as ‘potentially all’ when the speaker was expressing solidarity and approval (e.g., a mother rewarding the kids for playing quietly on their own) and as meaning ‘less than all’ when the speaker was seen as distancing themselves from the listener (e.g., a mother irritated with the kids who have been pestering her for cookies). These results support the conclusion that emotional valence (positive vs. negative) plays an important role in utterance interpretation (cf. Kousta et al. 2009).

 

Existing research has largely assumed that scalar interpretations are triggered in a uniform way across languages and cultures. However, the interplay of the inferential patterns and emotion discovered in our preliminary study suggests that language-specific factors (lexical semantic properties of the expressions investigated) may play a significant role in triggering different inferential mechanisms under different context conditions.

 

In this paper, we test the crosslinguistic validity of this hypothesis by reporting on a comparative study of scalar expressions in Dutch and in Japanese. Dutch and Japanese are two languages with different sets of scalar expressions and opposite interactional orientations (individualism vs. collectivism), which are expected to interface differently with (positive vs. negative) emotional valence. Moreover, since Japanese offers a distinct set of scalar expressions (e.g., two-way distinction in the disjunction (‘or’) particles: ka and ya) from Dutch and English, it presents an interesting test case for a general theory of the relationship between scalar inference and emotion. Our results contributes to our understanding of how emotive aspects of the situation can regulate the informational content of utterances as well as provide new data and insights into scalar implicatures from a pair of languages previously unexplored in this respect.

bottom of page