A direction effect on taste predicates

Abstract

Consider an utterance of ‘Fish sticks are tasty’ as made by a speaker who likes fish sticks. How will the speaker assess this claim when, at some later point in her life, she comes to dislike fish sticks? As true or as false? Will she retract her earlier statement or stand by it? More generally, will she use her present taste standard in assessing the claim or the standard she had at the time of the original utterance? The answer to this question is of vital importance for the recent discussion on the semantics and pragmatics of so-called “predicates of personal taste” (e.g. “tasty” and “fun”). The two major contenders for a semantics of predicates of personal taste are relativism and contextualism. Although these views make similar predictions in many cases, they crucially come apart when it comes to the indicated types of situations. Relativism predicts that the speaker’s assessment will depend on her later taste standard, that is, the taste standard governing the so-called “context of assessment” (e.g. Kölbel, 2003; Lasersohn, 2005; Stephenson, 2007; MacFarlane, 2014; Egan, 2014; Dinges, 2017). Meanwhile, contextualism predicts that the speaker’s assessment will depend on her original taste standard, that is, the taste standard governing the so-called “context of use” (e.g. Glanzberg, 2007; Stojanovic, 2007; López de Sa, 2008; Moltmann, 2010; Schaffer, 2011; Sundell, 2011; Huvenes, 2012; Plunkett and Sundell, 2013; Pearson, 2013; Marques, 2014; Zakkou, 2019). Which prediction is borne out by the data? Both relativists and contextualists typically take the data to favor their position. MacFarlane (2014: 141), for instance, holds that