top of page


Taste discourse may be affected by individual differences, such as one's native language, gustatory sophistication, and this project explores whether this is so. 


This research cluster explores intuitions about taste. Wyatt, Ulatowski, Hwan Ryu (Miami), and Dan Zeman (Warsaw) have designed a questionnaire that explores intuitions about taste predicates. And the study seeks to uncover whether native English speakers and native speakers of Korean and Romanian harbour different intuitions about taste. 

The topic of this study is to discover how people, philosophers and non-philosophers alike, respond to questions that test their intuitions about taste discourse. Since the major analyses of taste discourse entail empirical predictions about how the person on the street uses discourse of this sort, this study will ask participants to respond to a series of prompts intended to uncover their views about taste.

The research questions include: What are people’s views about the obligatoriness of retraction when one’s gustatory tastes change in relevant ways?  How do people evaluate the truth and falsity of their prior taste-related assertions when their gustatory tastes change in relevant ways?  What are people’s views about what one says when one makes a taste-related assertion?  Are people’s views about these topics stable across different linguistic communities, and are their views influenced by their gustatory sophistication, gender, socio-economic status, or political affiliation?

The objective of this investigation is to determine whether individual differences such as native language, gustatory sophistication, gender, socio-economic status, or political affiliation matter for one’s intuitions about taste discourse. Past studies have not shown that such individual differences exist, and indeed, this is the first cross-linguistic study on these topics.  However, past studies have shown that English speakers’ views about taste discourse are influenced by the order of polarities within vignettes. This project seeks to replicate and to extend the previous studies.

Relevant Background

Contemporary debates about taste discourse have generated a massive literature over the past 20 years which includes contributions from philosophers, linguists, and social scientists (for overviews, see Kölbel (2015 a,b), López de Sa (2011), Wyatt (2021), and Zeman (2020)). These debates have focused on central issues within all the represented fields, including the nature and value of disagreement, how different kinds of disagreements ought to be resolved, the distinction between ‘subjective’ and ‘objective’ linguistic expressions, the meaning of gradable adjectives, the scope of indexicality in natural language, the nature of context-sensitivity in natural language, and the nature of linguistic meaning.  


These debates have been particularly exciting for two reasons. The first is that they are interdisciplinary and thus more productive than many other debates within philosophy that are conducted solely by philosophers. The second, which is particularly germane to the study proposed here, is that they have generated a range of empirical predictions about how taste discourse is used in ordinary language. The trouble is that for the most part, these predictions have not been tested. A growing number of researchers, including Alexander Dinges and Julia Zakkou (2020), Elsi Kaiser (Kaiser forthcoming, Kaiser and Lee 2017 a,b, Kaiser and Rudin (2020)), and Markus Kneer (2021), have recently begun to empirically evaluate these predictions. This study will focus on three empirical phenomena pertaining to taste discourse, and in doing so will aim to replicate the findings of Dinges and Zakkou (2020) and Kneer (2021) following improvements to their experimental designs which are intended to add new dimensions to empirical work on taste discourse.


The first phenomenon is retraction.  In his highly influential defence of a relativist semantics for taste discourse, John MacFarlane (2014) argues that the key empirical diagnostic for testing such a semantics against its competitors pertains to retraction. Specifically, MacFarlane asks us to think about a taste shift case, in which a person A enjoys the flavour of a food f (e.g. fish sticks) at time t1 but then ceases to enjoy f’s flavour at time t2. We are to suppose that at t1, A asserts the sentence ‘F is tasty.’ MacFarlane hypothesises that competent speakers of English will hold that in such a taste shift case, A is obligated to retract at t2 the assertion that they made at  t1. Moreover, he argues that only his relativist semantics for taste discourse can accommodate this hypothesis and thus that this semantics is preferable to competing semantics.  


In a recent empirical study, Markus Kneer (2021) found that his participants judged that retraction is actually not obligatory in taste shift cases. Kneer’s findings thus disconfirm MacFarlane’s hypothesis and look to undermine MacFarlane’s case for his relativist semantics. Our aim is to determine whether Kneer’s findings can be replicated in light of certain improvements to his experimental design. 


First, the participants in Kneer’s study were all native English speakers, whereas our participants will be native speakers of English, Romanian, or Korean. We anticipate that the data we collect from English speakers will replicate Kneer’s study, but we will also seek to expand upon Kneer’s study by testing the hypothesis that Romanian and Korean speakers will have different intuitions about retraction than English speakers. In this respect, the scope of our study is considerably broader than that of Kneer’s study. Second, we will include vignettes with inverted polarity orders to determine whether the ‘direction effect’ that was observed by Dinges and Zakkou (2020) in connection with truth-value assessments (described below) is also manifested in participants’ intuitions about retraction. Third, whereas Kneer’s prompts were framed solely in terms of whether retraction is ‘obligatory,’ our prompts will test participants’ intuitions using a range of normative terms, including ‘obligatory,’ ‘required,’ ‘ought to,’ ‘need to,’ and ‘should.’ This will enable us to determine whether participants’ intuitions about retraction vary, depending upon how the prompts are framed.


The second phenomenon that we will investigate is truth-value assessments. In investigating this phenomenon, we will again have taste shift cases in mind. We will aim to identify our participants' views on how a speaker such as A above should evaluate the truth or falsity of their assertion at t1, given that the assertion conflicts with their tastes at t2. Each of the five major analyses of taste discourse makes different predictions about how ordinary speakers will think about this issue, so our findings will enable us to evaluate these analyses.


In a recent empirical study, Dinges and Zakkou (2020) obtained two results in connection with truth-value assessments. The first is the ‘direction effect,’ wherein the polarity order within a taste shift vignette (e.g. A likes and then dislikes f’s flavour vs. A dislikes and then likes f’s flavour) affects participants’ intuitions about truth and falsity. The second is the ‘even split,’ which is the result that the number of Dinges and Zakkou’s participants who favoured their ‘truth’ option was not significantly different from the number of their participants who favoured their ‘falsity’ option. Our aim here is to determine whether Dinges and Zakkou’s findings can be replicated in light of certain improvements to their experimental design.  

First, the participants in Dinges and Zakkou’s study were all native English speakers, whereas our participants will be native speakers of English, Romanian, or Korean.  In this respect, the scope of our study is considerably broader than that of Dinges and Zakkou’s study.  Second, Dinges and Zakkou’s prompt was ‘For each of the following responses, please tell us how likely you would be to give this response to [character]’s remark in the given context.’ This prompt is ambiguous, as it can be read in terms of either correctness or naturalness. It is participants’ intuitions concerning the correctness of [character]’s remark that are relevant  within this debate, so our prompt will be phrased specifically in terms of correctness. Third, Dinges and Zakkou’s study included only a ‘truth’ option and a ‘falsity option.’ Anticipating that some participants may be inclined to hold that A should classify their assertion at t1 as neither true nor false, we have added a ‘neither true nor false’ option. In assessing our findings, we will thus aim to determine whether Dinges and Zakkou’s ‘even split’ result was an artefact of their experimental design.


The third phenomenon that our study will cover is object-language speech reports.  Kölbel (2008, pp. 13-14) has pointed out that reports of this kind constitute an underinvestigated empirical diagnostic for testing the major analyses of taste discourse. Kölbel offers some preliminary hypotheses, but neither he nor any other researchers working on taste discourse, have investigated object-language speech reports empirically.  In our investigation of these reports, we will aim to determine how participants think the contents of certain taste-related utterances, e.g. an assertion of “It’s delicious” upon trying a food that one likes, ought to be described.  Our findings will enable us to evaluate all of the major analyses of taste discourse on this score.  Our methodology is novel within this area, insofar as we will be gathering intuitions from native speakers of English, Romanian, and Korean.  



Dinges, A. & Zakkou, J. (2020). A direction effect on taste predicates. Philosophers’ Imprint, 20(27), 1-22.

Kaiser, E. (forthcoming). Effects of sensory modality on the interpretation of subjective adjectives: comparing sight, smell, and taste. Proceedings of the 2018 Berkeley Linguistics Society Meeting.

Kaiser, E. & H. Lee (2017a). Experience matters: a psycholinguistic investigation of predicates of personal taste. Proceedings of SALT 27, 323-39.

Kaiser, E. & H. Lee (2017b). Predicates of personal taste and multidimensional adjectives: an experimental investigation. Proceedings of the 35th West Coast Conference on Formal Linguistics. 224-31.

Kaiser, E. & D. Rudin (2020). When faultless disagreement is not so faultless: what widely-held opinions can tell us about subjective adjectives. Proceedings of the Linguistic Society of America 5(1), 698-707.

Kneer, M. (2021). Predicates of personal taste: empirical data. Synthese,

Kölbel, M. (2008). Motivations for relativism. In M. García-Carpintero & M. Kölbel (Eds.), Relative Truth (pp. 1-37), Oxford University Press. 

Kölbel, M. (2015a). Relativism 1: representational content. Philosophy Compass, 10(1), 38-51.

Kölbel, M. (2015b). Relativism 2: semantic content. Philosophy Compass, 10(1), 52-67.

López de Sa, D. (2011). The many relativisms: index, context, and beyond. In S. Hales (Ed.), A Companion to Relativism (pp. 102-117). Blackwell. 

MacFarlane, J. (2014). Assessment Sensitivity: Relative Truth and Its Applications. Oxford University Press.

Wyatt, J. (2021). The nature of disagreement: matters of taste and environs. Synthese,

Zeman, D. (2020). Faultless disagreement. In M. Kusch (Ed.), The Routledge Handbook of the Philosophy of Relativism (pp. 486-495). Routledge.

bottom of page