Improving methodology of quantifier comprehension experiments

Abstract

Recently, research devoted to computational modeling of uantifier comprehension has been extensively published in this ournal.McMillan, Clark,Moore,Devita, andGrossman (2005)using euroimaging methods examined the pattern of neuroanatomical ecruitment while subjects were judging the truth-value of stateents containing natural language quantifiers. The authors were onsidering two standard types of quantifiers: first-order (e.g., all”, “some”, “at least 3”) and higher-order quantifiers (e.g., “more han half”, “an even number of”). They presented the data showing hat all quantifiers recruit the right inferior parietal cortex, which s associated with numerosity, but only higher-order quantifiers ecruit the prefrontal cortex, which is associated with executive esources, like working memory. In the latest paper Troiani, Peelle, lark, and Grossman (2009) assessed quantifier comprehension in atients with corticobasal degeneration (CBD) and healthy subects. They compared numerical quantifiers, like “at least 3”, which equire magnitude processing, and logical quantifiers, like “some”, hich can be understood using a simple form of perceptual logic. heir findings are consistent with the claim that numerical quantier comprehension depends on a lateral parietal–dorsolateral refrontal network, but logical quantifier comprehension epends instead on a rostromedial prefrontal–posterior cingulate etwork. According to the authors of the mentioned studies, their results erify a particular computational model of natural language quanifier comprehension posited by linguists and logicians (see e.g., an Benthem, 1986). One of the authors of the present coment has challenged this statement by invoking differences – issed in (McMillan et al., 2005) – between logical (expressibil-