Predicates of personal taste and normative meaning

The main aim of the paper is to reject the idea that predicates of personal taste express normative meanings. According to a recent theory proposed by Daniel Gutzmann, predicates of personal taste express both a truth-conditional content and a use-conditional content, the latter being normative. The purported normativity of predicates of personal taste is supposed to consist in that when producing utterances containing such predicates, their speakers suggest how other people ought to experience the objects of taste under discussion. The paper provides a bunch of evidence to show there are utterances containing predicates of personal taste that cannot be used normatively in this sense. In particular, it is claimed that one can make such an utterance and then add a related normative piece of information without doing anything redundant; one can also make such an utterance and then deny the related normative piece of information without generating any contradiction or infelicity. At the same time, it is admitted there are situations in which the speakers do communicate something normative by producing utterances containing predicates of personal taste. We thus need an explanation to cope with such cases. Although the paper does not offer one, it provides some reasons to the effect that a correct explanation has to be pragmatic rather than semantic.


Introduction
It is sometimes claimed that when making an utterance of a sentence containing a predicate of personal taste, 1 apart from presenting one's view regarding the flavor of a particular object, one says something normative regarding other people's views on the flavor. Suppose speaker A makes a serious and sincere autocentric utterance of 'X is delicious' (where 'X' designates an object of taste). 2 Filip Buekens argues that, besides having the assertive dimension that consists of expressing the proposition that X is delicious for A, the utterance also has the affective-expressive dimension that consists of manifesting an evaluative non-doxastic attitude towards X (see Buekens, 2011). Moreover, the affective-expressive dimension is said to be subject-transcendent (see ibid., p. 652) and one of its aspects can be portrayed in Kantian terms. 3 Famously, Kant claimed that judgments of taste are normative rather than merely descriptive; they are intended to be "converted into a rule for everyone" (Kant, 2007(Kant, [1790, p. 70) as "we believe ourselves to be speaking with a universal voice" (ibid, p. 47) when we make such judgments. Buekens has it that judgments of taste are supposed to influence how other people shall perceive the objects of taste under discussion. The affective-expressive dimension of autocentric taste utterances is thus normative at bottom.
Buekens's idea is that the affective-expressive dimension of autocentric taste utterances is not a matter of semantic content but pertains to the acts of making taste utterances. I take it that he is right in saying that some taste utterances are intended to set standards regarding other people's evaluative attitudes towards the objects of taste. 4 Now, granting the normativity of taste utterances, one might wonder whether it could not be the case that it is nevertheless more firmly connected with taste utterances than Buekens imagines. In particular, one might wish to consider the possibility that taste utterances are normative at the level of their lexical meaning or semantic content. Daniel Gutzmann recently proposed an interesting account along these lines according to which taste predicates express normative meanings as their non-truthconditional semantic contents (over and above their truth-conditional contents; see Gutzmann, 2016). The semantic rules for using taste predicates are claimed to suggest that by uttering taste sentences speakers not only express how they perceive the flavor of the objects of taste under discussion but also try to influence how other people shall affectively respond to them. There is thus a developed account inserting a normative ingredient into the semantics of taste predicates. This opens up space for us to examine the idea of the purported semantically driven normativity of taste utterances.
I aim to show that the meanings of taste predicates are not normative (in the above sense). That being said, I hasten to add that this is not to suggest that they cannot be used normatively in a suitable contextual setting. What I merely reject is that the normativity of taste utterances is semantically established.
The paper is structured as follows. I start with a brief description of how the normativity of predicates and utterances is understood in this paper (Sect. 2). Section 3 contains some illustrations of taste utterances that call for normative interpretation. They could be used to back up the idea that taste predicates express normative meanings. Gutzmann's theory to this effect is summarized in Sect. 4. Section 5 shows there are situations in which the normative interpretation of taste utterances is unavailable. Based on this, the normativity of taste utterances cannot be derived from the assumption that taste predicates express normative meanings. Section 6 contains some evidence that the normativity of taste utterances is to be explained in pragmatic rather than semantic terms. A brief summary in Sect. 7 concludes the paper.
Two warnings bear noting at the very outset. First, the purpose of this paper is negative. I aim to show that taste predicates do not express normative meanings, contrary to what Gutzmann's theory suggests. A positive account that would handle normative uses of taste predicates is not included. Second, I assume a contextualist semantic framework. This is because Gutzmann himself adopts a version of indexical contextualism for the truth-conditional content of taste utterances. Besides, I also find indexical contextualism rather appealing as an appropriate semantic framework for independent reasons I have no space to develop here. Since indexical contextualism is a mere working assumption in the present paper, I am not going to defend it against criticisms found in the literature.

Taste and normativity
Let us start with some cursory remarks on normative utterances. 5 In the present paper, the normativity of utterances is understood as connected with speakers' attempts to influence other people's non-doxastic (or evaluative) attitudes. 6 It seems to be a truism that by producing certain kinds of utterances, speakers express or imply (that they bear) positive or negative non-doxastic attitudes towards particular acts, objects, events, etc. If a speaker exclaims 'Yum!' upon trying something delicious she manifests her gustatory delight; we may describe her as express- 5 We need not take a stand on metaphysical issues surrounding normativity, such as whether there are normative facts and, if there are, what is their proper nature; see Finlay (2019) for a detailed discussion of both cognitivist and non-cognitivist explanations. 6 There are various conceptions of normativity; see Parfit (2011, pp. 267-269) for a brief overview. Some of them (like the reason-involving, the motivational, and the attitudinal approaches) could be helpful in the present context while others (like the rule-involving and the imperatival approaches) seem less fitting. I do not intend to imply that the notion of normativity outlined in the main text can be applied to all kinds of normative utterances.
ing a positive attitude of liking the thing in question. If a speaker utters 'I find the picture over there lifeless' she can be taken to suggest that she bears a negative attitude of disapproval towards the aesthetic qualities of the picture. In both cases, the speakers manifest their subjective attitudes. There is also a class of utterances that resemble the above examples by expressing subjective attitudes but differ from them in surpassing their subjectivity in a noteworthy way. Saying about some act that it is morally right does not amount merely to manifesting (that one bears) a positive attitude of approving the act but also to suggesting to other people the adoption of a similar attitude towards the act. Thus, apart from having the subjective dimension that is connected with expressing (non-doxastic) attitudes (of approval or disapproval, for example), utterances of this kind also have an intersubjective dimension that is normative in the sense relevant to this paper. In essence, the utterances of sentences like 'Performing Y is morally right' can be taken as a means to influence other people to adopt certain kinds of (non-doxastic) attitudes (towards performing Y, for example) rather than some other. 7 The above remarks are in line with the picture of normativity adopted by Gutzmann (2016). He builds on C. L. Stevenson's ideas according to which the major use of ethical statements "is not to indicate facts, but to create influence. Instead of merely describing people's interests, they change or intensify them. They recommend an interest in an object, rather than state that the interest already exists" (Stevenson, 1937, pp. 18-19; italics in the original). Given that "ethical terms are instruments used in the complicated interplay and readjustment of human interests" (ibid., p. 20; italics in the original), their meaning resides in the tendency "to produce … affective responses in people" (ibid., p. 23; italics in the original). The tendency to produce affective responses contains the tendency to influence (non-doxastic) attitudes (possibly among other things). An utterance of 'Performing Y is morally right' can be used with the intention to influence someone's moral attitudes towards doing Y and eventually prompt one to do Y in certain kinds of situations. It bears noting that according to Stevenson, the tendency to produce affective responses is a matter of meanings of ethical (and other expressive) terms; this is because the tendency is their persistent feature (see ibid.). 8 Let us turn to taste utterances and take 'X is delicious' as a paradigm taste sentence. 9 One might wish to claim that they also exhibit both dimensions. 10 Regarding 7 Apart from influencing other people's attitudes, the normativity of the above kind of utterances may have some other effects. For example, they can be used to influence other people's behavior, ways of conduct, preferences, inclinations, motivations, reasons, commitments, objectives, etc. I stick to the idea of normativity outlined in the main text because it is relevant for Gutzmann's proposal to be discussed below. The other manifestations of normativity, though equally important, are put aside in this paper. 8 Although Gutzmann is inspired by some of Stevenson's ideas, he does not adhere to his emotivist theory of meaning. Gutzmann's theory of meaning for expressive terms is much more sophisticated than a somewhat simplified emotivist approach. 9 I am going to discuss 'delicious' as a model taste predicate in my examples. Any other taste predicate could be used in its stead, though. If the reader preferred varied examples they are invited to make any modifications that would bring more diversity into the rather monotonous narration. Nothing important hinges on the choice of examples. 10 I assume that the normativity of taste utterances stems from the occurrence of taste predicates in them. If taste predicates are means to manifest attitudes it is possible to classify them as expressive terms; see Soria the subjective dimension, speakers use them to manifest their affective responses to the relevant objects of taste. A serious and sincere utterance of 'X is delicious' can be regarded as a manifestation of a speaker's positive attitude of liking X's flavor, while a serious and sincere utterance of 'X is distasteful' can be used to manifest the negative attitude of disliking X's flavor. Regarding the normative (intersubjective) dimension, speakers may produce taste utterances as attempts to influence other people to adopt certain kinds of affective responses to the objects of taste. One's utterance of 'X is delicious' is thus not just a way of manifesting (that one bears) a positive attitude of liking X's flavor but also a suggestion that one's audience is to adopt a similar attitude; analogously, one's utterance of 'X is distasteful' is used both to manifest her negative attitude of disliking X's flavor and to suggest that other people ought to perceive the flavor in the same way. 11 I take for granted that taste utterances exhibit the subjective dimension. Moreover, I assume without going into details that they do so due to the meanings of taste predicates contained in them. 12 It means that a serious and sincere assertion that something is delicious cannot be detached from the manifestation of the subjective attitude of liking or a similar non-doxastic attitude. This claim seems intuitively plausible because utterances like 'X is delicious, but I do not like its flavor' appear infelicitous-one cannot seriously and sincerely assert 'X is delicious' and immediately add that she does not like the flavor of the object without jeopardizing the consistency of her assertion.
Do taste utterances have the normative dimension? Does it hold that when one utters 'X is delicious,' one thereby intends to influence how other people shall shape their attitudes towards X's flavor? There surely are situations in which taste utterances are made with this intention; some examples of this kind are mentioned in Sect. 3. Moreover, there are situations in which taste utterances succeed in this mission. I doubt, however, that they license us to say that the normativity of the taste utterances is semantically established. They differ from ethical utterances in this respect because Ruiz & Stojanovic (2019) for some tests to distinguish between expressive and factual (uses of) terms. 11 Although both taste utterances and moral utterances are used to express non-doxastic attitudes, there seem to be some differences between them. Building on Blackburn (1998, pp. 9-11), John Eriksson observed that moral judgments exhibit an attitudinal complexity that taste judgments do not show (see Eriksson, 2016, p. 785). While taste utterances are used to manifest simple non-doxastic attitudes, moral utterances are means to manifest not just a positive or negative attitude towards something (such as approving or disapproving of a particular act) but also the disposition to call conflicting judgments into question. Moreover, the disposition to call conflicting judgments into question is claimed to be necessarily connected with moral judgments (see ibid., p. 789). See Pietroiusti (2021) for some persuasive arguments weakening the above contrast between taste judgments and moral judgments. In the present paper, I assume there is a normatively relevant contrast between taste utterances and moral utterances that is intuitively obvious, although it could be difficult to find an uncontroversial theoretical framework to spell it out. The intuitive difference between them will be utilized mainly in Sect. 5. 12 I proposed a theory of taste predicates that is compatible with this statement in Zouhar (2018). It might be argued that utterances of 'X is delicious' may not express positive non-doxastic attitudes (as an anonymous reviewer pointed out). I think a lot depends on how we understand the properties like being delicious and being tasty. If they are unpacked, for example, as interacting with someone's gustatory capacities in a way that produces agreeable gustatory experiences of an intrinsically desirable sort, as I have suggested in the above paper, any utterance of 'X is delicious' implies having a positive non-doxastic attitude on the agent's part. I leave this issue aside as orthogonal for the present purposes. normativity is not their persistent feature, to put it in Stevenson's terms. As we will see, there are taste utterances that are not intended to influence other people to adopt particular affective responses to the relevant objects of taste, meaning that serious and sincere assertions that something is delicious can be detached from the attempts to influence other people's non-doxastic attitudes towards the objects of taste. Thus, although it seems that a semantic explanation of the subjective dimension of taste utterances is legitimate, a semantic explanation of the normative (intersubjective) dimension of taste utterances must be rejected.

Normative taste utterances
Let us consider some situations in which taste utterances seem normative. One kind of situation in which they can be intended to influence how other people shall shape their taste-related non-doxastic attitudes involves assertions made by recognized experts. We often rely on experts' judgments in areas in which we take ourselves as less knowledgeable. The role of experts includes giving advice and directing other people's standpoints in the fields of their expertise. Wine critics recommend to us which wines deserve our estimate and thereby try to influence our attitudes towards the wines. 13 Imagine a situation in which a wine critic utters (1): (1) This 2015 Château Batailley Grand Cru Classé is delicious.
In uttering (1), the critic not only presents her subjective estimate of a particular wine but also attempts to shape the audience's attitudes towards it because she is entitled to expect that her estimate is relevant for influencing public positions. The critic thus conveys not only that the wine meets her taste standards but also that the audience shall find it delicious too.
It seems to me that the normativity of the above taste utterance is due to the social status of the wine critic rather than the meaning of 'delicious.' The audience's perception of the utterance as normative can be explained in terms of an inference from the premises that the speaker of (1) is an expert and that one should act as recommended by experts to the conclusion that the audience is to count the wine as delicious. If the utterance were made by someone else who did not share the same social status, it could not aspire to the same kind of effect because of the unavailability of the above inference.
Another kind of situation that could be interpreted in normative terms features taste utterances made in response to specific requests. Take the following exchange: (2) A: Which one of these wines would you recommend to me? B: This 2015 Château Batailley Grand Cru Classé is delicious.
13 Barry Smith describes wine critics as giving objective descriptions of the tastes of wines; see Smith (2007;2010).
There is a sense in which B's response to A's question is normative. A asks for advice, and B responds by recommending a particular wine. B indirectly answers A's question-by saying that the wine is delicious, B conversationally implicates what his recommendation is. The conversational implicature of B's utterance is the proposition that B recommends to A the 2015 Château Batailley Grand Cru Classé. According to a rather plausible explanation, this proposition can be appropriately calculated from what B said on the condition that the content of B's utterance is normative. If B's utterance is interpreted as saying that others, including A, are supposed to count the wine as delicious, and if it is assumed that B would recommend to A merely what A is expected to count as delicious, B's utterance would be an appropriate kind of response to A's query. The normativity of B's utterance is thus due to the peculiarities of the situation in question. Summing up, some situations contain normative taste utterances. They include those where a speaker is endowed with a proper kind of authority and those where taste utterances are made in communication in which their normativity is requested. Nevertheless, it would be a hasty generalization if all taste utterances were claimed to be normative and that, based on this, taste predicates expressed normative meanings. Below I sample several cases in which the normative interpretation of taste utterances is unavailable. Before doing so, however, I summarize Gutzmann's theory according to which taste predicates do have normative meanings. Gutzmann's theory is the primary target of the criticism to be given.

Gutzmann on taste predicates
Daniel Gutzmann proposed a theory according to which the meaning of taste predicates contains a normative ingredient (see Gutzmann, 2016). Based on the general framework of his multidimensional semantics for expressives (see Gutzmann, 2015), he recognized two meaning dimensions of taste predicates, namely the truth-conditional content and the use-conditional content. The motivation for discerning them is that some kinds of expressions contain truth-conditionally irrelevant ingredients but "impose conditions on the felicitous use[s] of the sentence[s] in which [they occur]" (Gutzmann, 2015, p. 7). This holds not just for various kinds of expressives (e.g., interjections like 'ouch' and 'oops,' expressive adjectives like 'damn' and 'bloody,' pejorative epithets like 'bastard' and 'idiot,' ethnic slurs like 'Boche' and 'Kraut') but also non-lexical ingredients such as intonation and syntactic operations (see ibid., pp. 26-38). Gutzmann has it that a complete rendering of such issues could be provided by a hybrid semantics that is in a position to specify the conditions under which the sentences containing such ingredients are true as well as the conditions under which they are felicitous (see ibid. pp. 7-8).
Before summarizing Gutzmann's proposal regarding taste predicates, let me place it on the map of the available semantic theories of this kind of expression. I assume the standard Kaplanian framework that distinguishes between the context of utterance and the circumstance of evaluation. Putting technicalities aside, the context of utterance determines the constituents of the semantic content an utterance expresses in the context while the circumstance of evaluation serves to determine the truth value of the semantic content so determined (see Kaplan, 1989, p. 494).
There are three main groups of theories according to which taste utterances express truth-conditional contents, namely absolutism, contextualism, and relativism, each of them comprising several approaches. 14 Let us assume that A is the speaker who makes an autocentric utterance of 'X is delicious' in the context of utterance, c u , and the utterance is assessed as true or false in the circumstance of evaluation, c e . Absolutism suggests that the utterance expresses, in c u , the proposition that X is delicious, and the proposition is either true or false depending on how things are in c e , where c e consists of a possible world (and time). 15 Contextualism takes it that the utterance expresses, in c u , the proposition that X is delicious for A, and the proposition is either true or false depending on how things are in c e , where c e consists of a possible world (and time). 16 Relativism claims that the utterance expresses, in c u , the proposition that X is delicious, and the proposition is either true or false depending on how things are in c e , where c e consists of a possible world (and time) and another parameter comprising A's standards of taste or a similar parameter. 17 There is a heated controversy between contextualists and relativists. Relativists blame contextualist theories for their inability to explain various phenomena, most notably (faultless) disagreements about matters of personal taste. For if speaker A autocentrically asserts 'X is delicious' and speaker B autocentrically asserts 'X is not delicious,' the contextualist theories predict that the former expresses the proposition that X is delicious for A while the latter expresses the proposition that X is not delicious for B, in which case the disagreement between them vanishes. Contextualists usually cope with this challenge by recognizing another level of content, either semantically or pragmatically obtained, apart from the above kind of propo-sition. 18 There are thus theories according to which, when uttering 'X is delicious,' the speaker expresses or communicates either a metalinguistic proposal about the standards of taste setting (see, e.g., Plunkett, & Sundell 2013;Sundell, 2011), or a presupposition of some sort or another (see, e.g., López de Sa, 2008;Parsons, 2013;Zakkou, 2019a;2019b), or non-doxastic (e.g., conative or evaluative) attitudes (see, e.g., Buekens, 2011;Huvenes, 2012;Marques, 2015;Marques & García-Carpintero, 2014;Zouhar, 2018). The disagreement phenomena are explained in terms of this additional level of what the speakers express or communicate.
Gutzmann's theory belongs to the contextualist camp. In particular, it is a specimen of an indexical contextualist approach. Assuming that ' [[e] ] t ' designates the truthconditional content of expression e, w ranges over possible worlds, and c ranges over contexts of utterance, we have (see Gutzmann, 2016, p. 38): Based on this, if speaker A utters 'X is delicious' in c, and A is the judge of c, the truth-conditional content of the utterance is identical to the proposition that X is delicious for A. The proposition is true in w provided X is delicious for A in w, and false otherwise. Given that the judge is a part of the proposition expressed, the content is subjective in a sense.
This proposal, as it stands, is not in a position to do justice to our intuitions regarding the possibility of (faultless) disagreements about matters of personal taste. Gutzmann thus follows suit and complements the truth-conditional dimension with another kind of content, namely the use-conditional dimension. However, his approach differs from the majority of contextualist treatments in two main respects. First, he pictures the expressions like 'delicious' as "lexically hybrid expressions that come with both meaning dimensions right from the lexicon" (ibid., p. 37; italics added; see also p. 38). In this respect, the taste predicates are modeled on expressions that are used to express simultaneously two types of meaning. For example, German 'Köter' and English 'cur' mean dog and also express a negative attitude (see ibid., p. 37). Second, unlike the other contextualist treatments, Gutzmann suggests that the additional meaning layer is normative in the case of taste predicates (see ibid., pp. 35 and 39). A taste utterance expresses not just that something is delicious (tasty, disgusting, etc.) for the judge of the context of utterance but also is an attempt to influence others to share the judge's attitude.
Assuming that ' [[e] ] u ' designates the use-conditional content of e, and w and c are as above, we have the following formal rendering (see ibid., p. 43): 18 Another strategy consists in denying that there do arise real disagreements about matters of personal taste. Since different speakers employ different standards of taste with respect to which they assess things as tasty or otherwise, the purported disagreements disappear when this is taken into account. See, e.g., Cappelen & Hawthorne (2009), Glanzberg (2007, Schaffer (2011), andStojanovic (2007). 19 Gutzmann discussed the truth-conditional content and the use-conditional content of 'tasty' rather than 'delicious.' Replacing 'delicious' for 'tasty' is innocuous for the present illustrative purposes.
The use-conditional content of A's utterance of 'X is delicious' in c is that X is delicious for A iff X shall count as delicious (in the context of utterance at hand). If A makes a serious and sincere assertion of 'X is delicious' to express the proposition that X is delicious for A, she thereby also makes a normative claim to the effect that X shall count as delicious in the context of utterance in which A finds herself. 20 If, on the other hand, A utters 'X is delicious' without asserting the proposition that X is delicious for A she does not make the respective normative claim either. 21 This happens, for example, when 'X is delicious' appears as a part of a conditional claim (see ibid., p. 42). A's utterance of 'If p, X is delicious' (where 'p' is a claim) is used to assert something else than the proposition that X is delicious for A; and, based on this, she does not express that X shall count as delicious in the context at hand either. In general, the use-conditional content of a taste sentence consists in that one is in a position to assert the (subjective) truth-conditional content only provided "one believes that the taste judgment shall objectively hold in the utterance context" (ibid., p. 43). Given the above biconditional, it is easy to see that the use-conditional content of taste predicates is firmly linked with their truth-conditional content, meaning that whenever one produces an utterance of 'X is delicious' to assert the truth-conditional content to the effect that X is delicious for the judge one also expresses the related use-conditional content. 22 Let us briefly look at the normativity of taste utterances. An assertion of 'X is delicious' is supposed to indicate that the context of utterance is such that X is taken as delicious in the context, i.e. that those involved in the context shall count X as delicious. The idea of counting something as delicious in the context of utterance is crucial for Gutzmann's understanding of the normativity of taste utterances. What does it amount to? 23 I assume that just as taking X as delicious by the speaker of an autocentric utterance of 'X is delicious' is tantamount to the speaker's bearing (and eventually manifesting) a favorable non-doxastic attitude (of liking, for example) towards X's flavor, so the audience, if claimed to count X as delicious, is expected to bear the same, or closely similar, non-doxastic attitude towards the same thing. Adopting such an attitude could be an appropriate affective response (to use Stevenson's term) of those involved in the context in which the utterance is produced. The 20 This is because the claim that X is delicious for A together with the claim that X is delicious for A iff X shall count as delicious give us, by the Modus Ponens, the claim that X shall count as delicious. Gutzmann decided not to state the use-conditional content as a simple claim that X shall count as delicious for the reasons summarized in (ibid., pp. 42-43). 21 Compare the following Gutzmann's remark (ibid., p. 43): "[T]he deontic force of a [predicate of personal taste] depends on the assertion of its truth-conditional component. When you do not assert that, say, tofu is tasty for you, you neither make a deontic claim that it shall count as tasty." 22 Gutzmann does not explicitly suggest what 'X is delicious' expresses when contained in A's assertion of 'If p, X is delicious.' Given some hints (see ibid., p. 42), he would perhaps admit that it still expresses the proposition that X is delicious for A but without asserting it. Since the expression of the related useconditional content of 'X is delicious' depends on asserting the truth-conditional content (a mere expression of the content does not suffice), the use-conditional content remains unexpressed. 23 An anonymous reviewer pointed out that Gutzmann's view on normativity is not fully stated. In particular, given that an utterance of 'X is delicious' suggests that X shall count as delicious in the context of utterance, Gutzmann is not explicit about what kind of normativity the 'shall count as delicious (in c)' locution presents. I agree with the reviewer's observation. normativity of the utterance is thus connected with an attempt to induce the audience to share a particular kind of attitude (cf. ibid., p. 40). It is left open, however, how the utterance can bring about such an effect. One might hypothesize, for example, that producing the utterance is just a way of suggesting that the audience is, or could be, motivated to adopt the attitude (and possibly act accordingly) or has some reason, or apparent reason, to do so; admittedly, some other stories invoking neither motivations nor reasons also are possible. Be that as it may, embracing a particular explanation of how the idea of counting something as delicious (in the context of utterance) connects with the idea of adopting a certain non-doxastic attitude on the audience's part is not necessary for the subsequent line of argumentation. For, the discussion is aimed at questioning the very view that utterances of 'X is delicious' (and other taste sentences) semantically express that X shall count as delicious in the context of utterance.
It seems that the theory is in a position to explain various phenomena. First of all, disagreements about matters of personal taste can be explained in terms of the use-conditional contents. Given their normative nature, the use-conditional contents are "up to dispute" (ibid., p. 35; see also p. 41) because speakers may have different views regarding what shall count as delicious in their (joint) context of utterance. If one speaker suggests that X shall count as delicious in their context and another speaker suggests that it shall not, they present incompatible views regarding the expected affective responses. The speakers thus place irreconcilable requirements on how people are supposed to perceive a particular object of taste, in which case they disagree with one another.
Second, since the use-conditional content of 'delicious' is normative, this approach should be in a position to explain the situations outlined in Sect. 3. When the wine critic utters (1) ('This 2015 Château Batailley Grand Cru Classé is delicious'), the theory portrays her as not only asserting that the wine tastes good to her but also as suggesting that others are supposed to count it as delicious too. This is what wine critics are expected to do, and Gutzmann's theory meets our expectations. And when B responds to A's request in the way she does in (2), it is again welcome that the theory portrays B's utterance as normative. A asks for advice on which wine to choose, and B asserts that the 2015 Château Batailley Grand Cru Classé is delicious. B's claim entails that A is expected to count the wine as delicious. This claim together with the overall setting of the situation can be used to derive B's recommendation that A ought to choose the wine in question.
Importantly, if there are situations where one can make assertions by producing taste utterances without expressing normative contents then the meanings of the uttered taste sentences have to be free from normative constituents too. And if the taste utterances express non-normative semantic contents, also the meanings of the taste predicates contained in the asserted sentences have to be non-normative. Based on this, the very possibility of the above kind of situation is decisive for demonstrating that taste predicates are not normative from the semantic point of view. This is not to say that there cannot be situations where taste utterances are used to communicate something normative; indeed, there can be-or even are-such situations. It just means that the normativity of the taste utterances in question is established on non-semantic grounds.
In this section, I gather three kinds of cases in which one may plausibly claim that the normative interpretation of taste utterances is not forthcoming.

Consent
The first case builds on the idea that people often discuss matters of personal taste and express their consent with, or dissent from, their fellows' taste utterances. Admittedly, if a taste sentence were normative, one should be in a position to express consent with, or dissent from, the normative content expressed by its utterances. We may test whether this is always the case. If it is possible to have a situation in which it is not, we would have evidence that the taste utterance did not express normative content and thus that the corresponding taste sentence was not normative at the semantic level either.
Let us start with an intuitively plausible instance. The following dialogue in which the deontic term '(morally) right' occurs seems acceptable: (3) A: Performing Y is (morally) right.
B: You are right that other people ought to perform Y.
Undoubtedly, the term '(morally) right' has a normative meaning. The structure of B's response is telling. B first says 'You are right' and then adds 'other people ought to perform Y.' The first part implies that B is going to approve of what A has said and the second part is thus expected to repeat the gist of A's meaning. In other words, the addition of 'other people ought to perform Y' after 'You are right' suggests that the former does not contain a novel piece of information that was not involved in A's utterance. Based on this, A's utterance is supposed to convey the normative claim that other people ought to perform Y. 24 24 It might be argued that saying that some act is (morally) right does not entail that people ought to perform it because one can intelligibly assert 'Performing Y is (morally) right, but people ought not to perform Y because it could have bad consequences.' (I am indebted to an anonymous reviewer for pointing this out.) As far as I can see, the consistency of this kind of assertion does not imply that '(morally) right' is not normative. It just implies that there are (practical or other) obstacles to doing the act. The intelligibility of the above assertion stems from the fact that although the danger of bad consequences is a reason for not doing Y in the situation under discussion, being (morally) right would amount to doing Y unless there were no bad consequences that would outweigh the benefits of doing Y. Based on this, B's response in (3) Turn to taste utterances. Given the previous considerations, the normativity of taste utterances implies that, by producing them, speakers intend to influence affective responses to the relevant objects of taste, i.e. motivate the audience to adopt certain kinds of non-doxastic attitudes towards the objects. The expression of consent with the normative constituents of their semantic contents thus consists of agreeing that people should be influenced in this respect. Take the following dialogue: If A's utterance were normative and a part of its content was about which kind of affective response others are motivated to take to the wine's flavor, B's line would have to be acceptable. Intuitively, however, it does not seem so. The utterance of 'You are right' has to be understood as relating to an already mentioned piece of information, and one may expect that the subsequent that-clause makes the piece of information explicit. It seems, however, that 'other people ought to like its flavor' brings a novel point that was not contained in A's claim. This is because prefixing 'You are right' to 'other people ought to like its flavor' is infelicitous, and thus A's utterance does not seem to contain a normative piece of information for B to agree with. 26 Comparing (4) with (3), 'delicious' does not strike us as being on a par with '(morally) right' when it comes to normativity. To strengthen this observation regarding (4), compare it with the following dialogue: B's utterance of 'You are right' is separated from her utterance of 'Other people ought to like its flavor.' The former utterance is naturally understood as expressing B's consent with A regarding the wine's flavor. The latter utterance makes good sense provided it is intended to push the dialogue further by making a novel point rather than could be understood as tacitly containing the 'other things being equal' phrase-B replies to A's assertion by suggesting that other people ought to perform Y, other things being equal. 25 B could express her consent with A in a number of different ways. Given Gutzmann's explication of the use-conditional content of taste predicates, she could respond by uttering 'You are right that it shall count as delicious.' It seems to me that neither this response nor any other in which the consent with a purported normative content of A's utterance is made explicit sounds good. 26 Let us admit that ascribing 'is delicious' to something can be regarded as a reason to like the thing in question. Based on this, B's response to A's utterance in (4) could run as follows: 'Yeah, there really is a reason for everyone to like it. It is just too bad that people do not follow reasons.' A response along these lines seems acceptable, as an anonymous reviewer pointed out. I agree. However, the plausibility of the modified dialogue does not make B's response in (4) acceptable. The dialogue in (4), as it stands, remains problematic, meaning that A's utterance is not normative. Second, even if we admit that being delicious is a reason to like the relevant thing, 'X is delicious' is not synonymous, or nearly synonymous, with 'There exists a reason for everyone to like X's flavor.' This is because 'X is delicious but I do not suggest that there is a reason for everyone to like X's flavor' is perfectly coherent. In the modified dialogue, B thus brings a novel point building on A's assertion rather than merely reiterating an already introduced idea. Based on this, even if normativity were explicated in terms of reasons, A's utterance would not be normative either.
merely specifying the target of B's consent; in other words, we do not perceive the utterance of 'Other people ought to like its flavor' as a repetition of what B claimed to agree with by her utterance of 'You are right.' This aspect could be made even more palpable if B made a pause between her utterances of 'You are right' and 'Other people ought to like its flavor.' Since the content of B's second claim in (5) that is explicitly normative seems to bring something new into the discussion, A's utterance cannot be plausibly understood as normative. And, analogously, since B's normative response in (4) seems problematic because it indicates that B merely repeats what A already has said, A's utterance is again better understood as non-normative. This contrasts with the predictions of Gutzmann's theory. It predicts that A's utterances in (4) and (5) convey normative use-conditional contents apart from the usual truth-conditional contents. Thus, the theory provides an inaccurate picture of what is going on in the dialogues.

Contrast
The second kind of case invokes linguistic constructions that contain 'although,' 'but,' or similar conjunctions expressing contrast between conjuncts. In general, 'p, although q' suggests that q contains a point that somehow contrasts with p. There are various kinds of contrast, but not all of them are suitable for 'p, although q.' In particular, if 'p, although q' is to be intelligible, q cannot be used to express content that would contradict the content of p. Given this observation, consider (6): (6) This 2015 Château Batailley Grand Cru Classé is delicious, although I do not mean to suggest that other people ought to like its flavor.
Notice that the second part of (6) is explicit about the speaker's intention not to issue any regulation regarding how other people ought to view the wine's flavor in a given context of utterance. Now suppose that the first conjunct expressed the normative content that the wine ought to be delicious in the context of utterance. If this were so, the addition of 'although I do not mean to suggest that other people ought to like its flavor' would be awkward. This is because the speaker would contradict what she tried to convey by uttering the first conjunct of (6). Yet, the utterance of (6) is perfectly intelligible. Based on this, the two conjuncts of (6) connected with 'although' do not contradict each other, suggesting that the first of them cannot be read normatively as an attempt to influence other people's non-doxastic attitudes towards the wine. 27 Compare (6) with (7) which concerns moral issues: (7) ?? Performing Y is (morally) right, although I do not mean to suggest that other people ought to perform Y. 27 Again, I do not mean to suggest that there cannot be a situation in which an utterance of (6) (or a similar claim) would turn out to be inconsistent. What is important is that there are at least some situations in which such utterances are not inconsistent. That is enough to deny that the first conjunct of (6) expresses normative content as a matter of its semantics.
Intuitively, (7) is awkward. A speaker of (7) says, on the one hand, that performing a certain deed is morally right but adds, on the other hand, that she proposes no regulation regarding our behavior. That is strange, and speakers who would seriously and sincerely utter (7) could hardly be said to use '(morally) right' in its ordinary meaning. Given this diagnosis, we can infer that 'although,' as used in (7), connects two conjuncts that do contradict each other. That is why (7), unlike (6), is infelicitous. 28 It bears noting that some uses of 'although' that appear in structurally similar taste utterances and moral utterances seem acceptable. Take the following pair of sentences: (8) This 2015 Château Batailley Grand Cru Classé is delicious, although not for you.
(9) Performing Y is (morally) right, although not for you.
There are acceptable readings of (8) and (9). According to one such reading, a speaker of (8) intends to say that the wine is delicious according to her taste standards but not according to her audience's taste standards, and a speaker of (9) intends to say that performing Y is morally right according to her moral standards but not according to her audience's standards. Thus, (8) is intelligible on the condition that its conjuncts contain implicit relativizations to various standards (and similarly for (9)). This kind of explanation is not available for (6) and (7). Summing up, Gutzmann's theory has to portray (6) as unacceptable. Since it predicts that the first conjunct of (6) conveys the normative piece of information that the wine in question shall count as delicious in a given context of utterance, the speaker of (6) would contradict herself by claiming that she did not mean to suggest that other people ought to like the flavor. However, the prediction is wrong.

Non-redundancy
The third kind of situation that illustrates the possibility of non-normative taste utterances of 'X is delicious' invokes the fact that the supplementations of explicitly normative claims to the utterances are not redundant. The basic idea has it that if an occurrence of a taste predicate resulted in an utterance's normativity, the addition of a certain kind of claim in which the normativity was made explicit should be felt like unnecessary repetition.
Consider the following sentence: (10) This 2015 Château Batailley Grand Cru Classé is delicious. Moreover, other people ought to like its flavor too.
If an utterance of the first sentence in (10) were normative, it would be intended to influence the affective response of a certain kind. Since the second sentence in (10) explicitly states how people are supposed to respond to the wine's flavor, we should take the sentence as redundant and feel the use of 'moreover' as somewhat inappropriate. However, neither the feeling of the inappropriateness of 'moreover' nor the redundancy of the second sentence obtains. The second sentence thus brings something new that has not been expressed by the first sentence. If the second sentence, which is normative, makes a novel point, the first sentence cannot be understood in the normative sense. We can strengthen this interpretation by the following pair of claims, where the second claim appears to be off-topic: (11) ?? This 2015 Château Batailley Grand Cru Classé is delicious. In other words, other people ought to like its flavor.
As 'moreover' suggests that something novel follows, the words 'in other words' imply that the same thing is going to be asserted once more using different wording.
The speaker would be perceived as making an awkward claim if she started by uttering the first sentence in (11) and subsequently uttered the second sentence. Based on this, the second sentence simply does not repeat the content of the first sentence. The first sentence thus expresses nothing normative.
(13) Performing Y is (morally) right. In other words, other people ought to perform Y.
The occurrence of 'moreover' in (12) indicates that it introduces a novel piece of information the previous part does not convey. It seems, however, that this is not so. That is why the addition made in the second sentence is redundant. To say that something is morally right amounts to prescribing a certain kind of behavior, which eventually amounts to saying that people are supposed to conduct themselves in line with this regulation. The second sentence in (12) is unnecessary because it merely repeats what the first sentence already contains. This fact is highlighted in (13) where it makes good sense to introduce the second sentence with 'in other words.' Let us sum up. Gutzmann's theory predicts that the first sentence in both (10) and (11) is used to make a normative claim. Based on this, the respective second sentences in (10) and (11) should be expected to bring nothing new. This expectation is, however, frustrated. Regarding (10), the opening phrase in the second sentence (i.e. 'moreover') suggests that a new piece of information will follow. Regarding (11), the opening phrase in the second sentence (i.e. 'in other words') suggests that the same piece of information will occur once more. The intuitive plausibility of (10), as well as the intuitive implausibility of (11), imply that 'This 2015 Château Batailley Grand Cru Classé is delicious' does not convey a normative piece of information, contrary to the predictions of Gutzmann's theory. Daniel Gutzmann raised a worry regarding the above tests with 'moreover' and 'in other words.' He suggested that the contrast between (10) and (12)-as well as (11) and (13)-does not show that taste utterances do not contain normative components. It merely shows that they do not contain them at the level of their truth-conditional content; yet, they can still be normative at the level of their use-conditional content, as proposed by Gutzmann's theory. This is emphasized by an analogy between taste utterances and utterances containing expressives. Gutzmann pointed out that (10) patterns with (14) and (11) with (15): (14) That damn guy was late again. Moreover, I really dislike him.
(15) ?? That damn guy was late again. In other words, I really dislike him.
The second utterance in (14) is acceptable and the corresponding utterance in (15) is not. Recall that 'moreover' is used to introduce a novel piece of information while 'in other words' usually introduces a piece of information that has already been provided. Based on this, the acceptability of (14) and the unacceptability of (15) could suggest that 'I really dislike him' presents something new that is not contained in 'That damn guy was late again.' Yet, 'I really dislike him' and 'damn' are closely related because both present a negative attitude. This asymmetry could be explained by the reference to different levels of content-the former expresses, at the truthconditional level, what the latter is used to manifest at the use-conditional level. Now, (10) mirrors (14) and (11) mirrors (15). By parity of reasoning, 'Other people ought to like its flavor' could be said to express, at the truth-conditional level, a piece of information that is contained in 'This 2015 Château Batailley Grand Cru Classé is delicious' at the use-conditional level. And since the former contains a normative component, the latter has to contain a normative component too, albeit as a part of its use-conditional content. Given this explanation, the acceptability of (13) and the unacceptability of (12) result from the fact that 'Other people ought to perform Y' captures a part of the truth-conditional content of 'Performing Y is (morally) right' also at the truth-conditional level.
There are some reasons to disagree with the idea that the utterances in (14) and (15) challenge my interpretation of the 'moreover' and 'in other words' tests. First, the tests are used to show that one's taste utterances (as a matter of their semantics) are not intended to influence other people's non-doxastic attitudes towards the objects of taste under discussion; semantically speaking, saying that something is delicious conveys nothing about whether other people shall find its flavor appealing. The utterances in (14) and (15) are supposed to undermine this view. They differ from those in (10) and (11) in an important respect, though; 'I really dislike him,' unlike 'Other people ought to like its flavor,' does not concern other people's non-doxastic attitudes. Thus, the former does not state the same kind of normative content as the latter. Consequently, the above utterances with expressives are rather irrelevant regarding the normative dimension of the utterances in (10) and (11). Moreover, notice that the utterances in (16) and (17), which are structurally more similar to (14) and (15) than (10) and (11), do not pattern with (14) and (15)  Based on this, the piece of information about the speaker's positive non-doxastic attitude towards the wine is contained in 'This 2015 Château Batailley Grand Cru Classé is delicious.' Admittedly, expressing such an attitude is a part of the sentence's semantic content. Nevertheless, this fact does not license us to say that its semantics also contains information about whether others ought to like the wine too, as the obvious difference between (10) and (16) on the one hand and (11) and (17) on the other demonstrates. Second, I am rather hesitant to admit that 'I really dislike him' in (14) and (15) captures, at the truth-conditional level, what 'damn' is supposed to express at the use-conditional level as a part of 'That damn guy was late again.' The intelligibility of 'moreover' in (14) and the unacceptability of 'in other words' in (15) could be explained by the fact that 'I really dislike him' is intended to convey something else, or something more, than a mere negative attitude manifested by 'damn.' In particular, the occurrence of 'damn' in 'That damn guy was late again' indicates the negative attitude towards the guy in connection with him being late. On the other hand, 'I really dislike him' presents the negative attitude broadly conceived-the speaker simply manifests her disliking the guy irrespective of the circumstances of the particular situation. We may strengthen this idea by the fact that (18) is acceptable: (18) That damn guy was late again. Yet, I really like him.
By uttering 'I really like him,' the speaker does not contradict what she conveyed by containing 'damn' in her utterance of 'That damn guy was late again' (although there undoubtedly is some kind of contrast in (18)). To conclude, it seems intuitively plausible that 'I really dislike him' brings something new that licenses the use of 'moreover' and 'in other words' in (14) and (15).
Third, and closely related, the utterances in (14) and (15) are inconclusive in another way too. Although the taste predicates in (10) and (11) seem to pattern with the expressives in (14) and (15), it is easy to find instances where this is not the case. Consider the following utterances with expressives produced by a speaker after trying a particular dish: (19) ?? Yuck! Moreover, I really dislike the dish.
(20) Yuck! In other words, I really dislike the dish.
If one utters 'Yuck!' to express one's disgust about a particular meal, it does not make good sense to utter 'I really dislike the dish' in the way contained in (19), but it makes good sense to do so as in (20). Admittedly, the 'Yuck!' part of (19) and (20) expresses a particular non-doxastic attitude at the use-conditional level, and the 'I really dislike the dish' part expresses the same attitude at the truth-conditional level. That is why the use of 'moreover' is infelicitous and the use of 'in other words' is felicitous-the attitude is not presented as new, irrespective of the fact that the two parts of (19) and (20) express it at different levels. Based on this, it could be the case that 'moreover' is acceptable in (14) and 'in other words' is unacceptable in (15) because 'I really dislike him' expresses something new (as also suggested in the previous paragraph) that has not been expressed by 'That damn guy was late again' and not because the former expresses as its truth-conditional content what 'damn' presents as its useconditional content.

Summary
Speakers are in a position to produce non-normative taste utterances. We noticed that one can make a taste utterance and then add a related normative piece of information without doing anything redundant. One can also make a taste utterance and then negate a related normative piece of information without producing any inconsistency. We further noticed that it does not make good sense to express consent with the alleged pieces of normative information in certain situations in which people discuss matters of taste.
Given the considerations in Sect. 3, however, it has to be recognized that there are taste utterances that are rather naturally interpreted in normative terms. That is why the results of the present section can hardly be interpreted as demonstrating that taste utterances cannot be normative. The proper lesson has to lie elsewhere. What the above considerations do demonstrate is that the normativity of taste utterances cannot be a matter of normative meanings of taste predicates. For if the meanings of taste predicates were normative, all taste utterances would have to be normative as well. Since not all taste utterances are normative, it could be promising to provide a pragmatic explanation of the normativity of normative taste utterances, as tacitly implicated by Buekens's approach mentioned at the very outset of the paper.

Pragmatic phenomenon
As I stated in the opening part of this paper, I do not offer a positive theory of the normativity of taste utterances. Instead, I am going to motivate the idea that it is merely a pragmatic phenomenon and thus that going pragmatic is the best strategy one could adopt to find a successful explanation. I outline two kinds of reasons for a pragmatically driven approach.

Explicit reference to a judge
Let us start with a seemingly unproblematic observation that taste utterances containing an explicit reference to a judge cannot be normative. Buekens motivated the idea as follows: "When I utter I find the roller coaster fun, you come to know something about me […]. Expressing my attitude towards the roller coaster, I speak, as Kant suggests, with a universal voice. Not literally or explicitly, of course: I achieve 'enough universality' by getting you to imagine yourself having a similar attitude or experience towards the roller coaster, something I cannot achieve by pointing out or asserting that I myself find the roller coaster fun. Uttering a sentence with that semantic content would make it impossible to exploit the subject-transcendent element involved in the affective-expressive dimension. Explicit reference to oneself or one's own experience in the semantic content of the utterance does not invite the intended audience to empathize with me." (Buekens, 2011, p. 653;italics in the original).
Given what Buekens states, a speaker, when uttering (21), cannot speak with a sufficiently universal voice: (21) I find this 2015 Château Batailley Grand Cru Classé delicious.
The speaker says something about her experience regarding the wine's flavor. Her utterance merely describes how she perceives it, without suggesting how others ought to experience it. The audience may register what the speaker said about her attitude towards the wine without feeling any pressure or invitation to empathize with the speaker. It seems we may further strengthen this impression by imagining the speaker explicitly adds that she does not aspire to make a normative claim: (22) I find this 2015 Château Batailley Grand Cru Classé delicious. I am not saying that also other people ought to find it delicious. What I say concerns merely my tastes.
The sentences contained in (22) form a non-contradictory whole. Based on this, the occurrence of 'delicious' in the first sentence (as well as its occurrence in the second sentence) is not normative. The consistency of (22) notwithstanding, it seems that Buekens is wrong after all. It is possible to produce utterances of (21) with such a degree of universality that could bring about normative effects. To make a comparison, consider a different kind of example that is not intended to bring normative effects: (23) I find you guilty of committing the murder.
If uttered by a judge at a court, the utterance of (23) is sufficiently universal (though in a non-normative manner) for everyone to accept, as a matter of fact, that the defendant is guilty of committing the murder. The judge produces a verdictive with subject-transcendent consequences rather than a mere description of her subjective view regarding the defendant and their involvement in the murder. When uttering (23), the judge has adequate powers to make a universally accepted judgment regarding the defendant's guilt and thus eventually make the accused person a convicted person in the eyes of society and public authorities. Admittedly, such an effect may not be achieved if (23) were uttered by a person who was not in a proper position to make this kind of verdictive. An utterance of (23) produced by an ordinary man from the street is a mere statement of their personal belief and is not in a position to make the accused person a convicted person. The comparison between an utterance of (23) produced by a judge and an utterance produced by a man from the street shows us that the subject-transcendent effects of such an utterance are a matter of the right kind of speech act produced by the right person in the right context rather than a matter of the meanings of the expressions contained in the uttered sentence.
Returning to (21), we can imagine that it also could be uttered as a verdictive or a similar kind of speech act-the utterance could attain a universal reading provided the speaker was endowed with proper authority. Imagine a situation in which (i) the speaker of (21) is a wine critic, (ii) the speaker's audience recognizes her as an expert in the relevant area, (iii) the audience acknowledges that she is entitled to make judgments to tell apart good wines and bad wines, (iv) the speaker believes that (ii) and (iii) are the case, and (v) the speaker has the corresponding communicative aims to make judgments regarding good wines or bad wines with the intention to shape her audience's attitudes towards the wines in question. The audience expects that the critic is going to make judgments to rank the available wines. Her utterance is not perceived as a mere description of her tastes but as a verdictive or a similar kind of speech act with a subject-transcendent effect. 29 Given the expert's verdict, the wine deserves high credit (the first prize in a contest, for example) not because it is appealing to a particular person's palate but because everyone with sufficiently refined tastes might find it appealing. Thus, (21) can be used to communicate something normative. Now, assuming contextualism, there is no significant semantic difference between (21) and (1) ('This 2015 Château Batailley Grand Cru Classé is delicious')-both sentences express the semantic contents that contain the judge of a context of utterance as their constituent. Based on this, if (21) can be used to produce normative utterances the same should hold for (1). The above paragraph could be easily modified to concern utterances of (1) instead of (21). 30 By uttering (1), the speaker can be in a position to issue a verdict (recommendation, suggestion, etc.) regarding the wine in question and produce the same kind of subject-transcendent effect provided she harbors the right kind of communicative intention and finds herself in the right kind of situation in which her audience recognizes her as an authority in the relevant area and acknowledges that she is entitled to make judgments of this sort. In such a case, her utterance of (1) is not a mere record of what she happens to believe; using Buekens's words, it is an invitation, addressed to her audience, to imagine having a similar experience with the wine.
Summing up, given the results of the discussion in Sect. 5, the normativity of normative utterances of (1) cannot be properly explained in semantic terms. We have seen that if a speaker is endowed with a proper social status and has appropriate communicative intentions, she can be in a position to produce normative taste utter-29 Such an utterance of (21) can be regarded as a hedged way of recommending the wine in question. (Thanks to an anonymous reviewer for pointing this out.) 30 In such a case, however, (1) could be regarded as a direct (rather than hedged) recommendation of the wine.
ances of (1). The possibility of making normative taste utterances is thus determined by the context in which the speaker utters a taste sentence and her communicative intentions.

Conversational implicature tests
Another kind of evidence justifying a pragmatic explanation of the normativity of taste utterances further builds on some of the results obtained in Sect. 5. We have seen that it was possible both to deny that a taste utterance expressed anything normative without producing inconsistency (Subsection 5.2) and to add a normative piece of information to what a taste utterance was supposed to express without producing redundancy (Subsection 5.3). These results could prompt one to hypothesize that if taste utterances are used to communicate normative contents the contents resemble (particularized) conversational implicatures or some other pragmatically obtained kinds of content.
Following Grice, (particularized) conversational implicatures are usually assumed to be cancellable and non-detachable. If a speaker utters p to communicate q in a given context, the latter is a conversational implicature of the utterance of p provided it is admissible for the speaker to add 'but not q' or 'I do not mean to imply that q' after saying p (see Grice, 1989, p. 44). 31 Thus, if the speaker is in a position to say 'p, but not q' in the context, and thereby successfully cancel q, without saying anything inconsistent, her utterance of p, if used to communicate q, does so on pragmatic (conversational) rather than semantic grounds. Similarly, if a speaker utters p to communicate q in a context, the latter is a conversational implicature of the utterance of p provided it is not possible to utter any other r, which says the same thing (i.e. expresses the same semantic content) as p while failing to communicate q (see ibid., p. 39). Thus, if the speaker is not in a position to detach saying q from what both p and r express in the context in question, her utterance of p, if used to communicate q, does so merely on pragmatic (conversational) grounds. Now it seems that the normative content of taste utterances is both cancellable and non-detachable. We have already seen that it is cancellable (without explicitly saying so; see Subsection 5.2). Consider a slightly modified version of one of the examples given in Subsection 5.2: (24) This 2015 Château Batailley Grand Cru Classé is delicious, but I do not mean to imply that other people ought to count it as delicious.
An utterance of (24) seems coherent; one can say about something that it is delicious without suggesting anything regarding how other people are expected to experience its flavor. The explicit cancellation of the normative content is unproblematic, suggesting that the normativity of normative utterances of (24) is pragmatically established. If someone finds (24) problematic because she or he has the idea that uttering 31 Grice distinguished between explicit cancellability and contextual cancellability (see Grice, 1989, p. 44). I put the latter aside but take for granted that if one can show that a communicated content is explicitly cancellable it is also contextually cancellable.
'delicious' instead of more cautious 'delicious for me' entails that the speaker does say something universal transcending her personal taste (see Buekens, 2011, pp. 643-644), a modified example could show that this is not the case. Imagine there is a continuation following the statement in (24) along these lines: (25) This 2015 Château Batailley Grand Cru Classé is delicious, but I do not mean to imply that other people ought to count it as delicious. I know they do not, but it is not in my power to do anything about it.
(25) sounds good. The second sentence provides a sort of justification for why the speaker does not intend to say about what others ought to find delicious. She simply sees no point in saying something like that because she is aware of her inability to influence other people's tastes. If (25) is acceptable, the normative content of (1) ('This 2015 Château Batailley Grand Cru Classé is delicious'), if there is any, is cancellable.
The normative content of a normative taste utterance, if there is any, cannot be detached from the utterance. If a speaker utters a taste sentence to communicate normative content, she would communicate it even if she used another sentence with the same semantic content instead of the former sentence. Imagine a situation in which the speaker produces a normative taste utterance by uttering (1) and communicates thus the content that other people ought to count the wine in question as delicious. As we have seen in Subsection 6.1, the same normative content could be communicated had the speaker uttered (21) ('I find this 2015 Château Batailley Grand Cru Classé delicious') instead of (1). Admittedly, any sentence saying the same thing the speaker could utter instead of (1) would communicate the same normative content provided the speaker's communicative intention is conveying something normative.
Normative contents of taste utterances thus pass ordinary tests for (particularized) conversational implicatures. Saying this, I do not wish to imply that normative contents of taste utterances are conversational implicatures. What I intend to say is just that they exhibit ordinary behavior of paradigmatic instances of pragmatically derived contents, meaning that it must be natural to treat normative contents as pragmatic rather than semantic. There is more to be done to show that normative contents are conversational implicatures rather than some other kind of pragmatically communicated content. Nevertheless, I take it that enough was said to motivate the idea that we should be looking for a pragmatic explanation of the normativity of taste utterances.

Conclusion
The primary aim of this paper was negative, namely to show that it is not the case that the meaning of taste predicates is normative, contrary to what Gutzmann suggested in the theory outlined in Sect. 4. The normativity of taste utterances was construed in terms of attempting to influence other people's affective responses (i.e. their nondoxastic attitudes) to the relevant objects of taste. The main evidence for the above negative thesis comes from taste utterances that express non-normative pieces of information. There are various situations in which speakers produce taste utterances without intending to influence other people's views regarding the flavor of the objects they talk about. Section 5 provides illustrations of some of them. That is not to say that no taste utterances are normative, though. Section 3 contained some instances we could naturally interpret in normative terms. We have thus mixed data regarding the normativity of taste utterances. Consequently, it seems natural to argue that the semantic content of taste utterances must be non-normative but that people sometimes make taste utterances to communicate something normative at the pragmatic level. Section 6 contained some evidence that going pragmatic in this respect could be a promising route.