The Many Uses of Predicates of Taste and the Challenge from Disagreement

Abstract In the debate between contextualism and relativism about predicates of taste, the challenge from disagreement (the objection that contextualism cannot account for disagreement in ordinary exchanges involving such predicates) has played a central role. This paper investigates one way of answering the challenge consisting on appeal to certain, less focused on, uses of predicates of taste. It argues that the said thread is unsatisfactory, in that it downplays certain exchanges that constitute the core disagreement data. Additionally, several arguments to the effect that the exchanges in question don’t amount to disagreement are considered and rejected.


Introduction
Disagreement has played an important role in the current debate about the semantics of predicates of taste -expressions like 'tasty ', 'disgusting', 'fun', 'boring', 'astonishing', 'cool' etc. Specifically, the intuition of disagreement has been used by certain parties to the debate to argue against their opponents by showing that they cannot account for this intuition and thus that they are incomplete. Relativists, in particular, have used this objection against the major competitor of their view, contextualism.
Needless to say, this objection has not remained without answers. A number of strategies to deal with the challenge from disagreement posed by the relativist have recently surfaced in the literature -such as relegating the disagreement to a pragmatic level, or understanding disagreement in a different way than the relativist does. However, important as those strategies are, in this paper I won't engage with any of them. Rather, I will Dan Zeman investigate a less solid, yet nevertheless recognizable trend present in the work of several contextualists. The trend consists in appealing to various, mostly neglected uses of predicates of taste to explain disagreement, coupled with considerations against taking a certain type of exchange the relativist bases her challenge on as a case of disagreement. My aim is to show both that the dialectical effect of appealing to those neglected uses is close to null and that the case made against taking the relevant exchanges as cases of disagreement is not convincing. My aim is not to show that contextualism cannot account for disagreement in the relevant exchanges when other strategies are appealed to, or when the trend scrutinize here is combined with such strategies. The focus is entirely on the ability of the said trend in itself to provide a full answer to the challenge.
The paper is structured as follows. In the first section I introduce contextualism and relativism and present the challenge from disagreement, by means of a scenario that has the role of pumping intuitions. In section 2 I present the various uses of predicates of taste the contextualists engaged with have pointed to when tackling the issue of disagreement. In section 3 I assess the dialectical gain contextualists hope to reap from those uses and illustrate the type of exchange that the relativist should base her challenge on. I then show that various considerations extracted from contextualist literature to the effect that the exchange presented cannot be taken as a case of disagreement are far from being compelling. The outcome of the paper is that the contextualist trend investigated, by itself, goes a very short way towards answering the challenge from disagreement.

Contextualism, relativism, and disagreement:
setting the stage I start with a scenario containing an exchange involving predicates of taste that I take to be a case of genuine disagreement, and introduce contextualism and relativism by showing how each treats the utterances involved in the exchange. Thus, Anne, Bob and Claire are three graduate students from the same department who are living together. Claire likes to make brownies, so most days there is a plate with brownies in the kitchen. On this particular evening, Claire is not home but Anne and Bob are, working on their term essays. At a certain point, Anne and Bob meet in the kitchen. In front of them is the usual plate, this time with a single brownie left on it. Anne and Bob look at each other and tacitly decide to split the brownie. Being done with their workload for the day, they linger in the For relativism about predicates of taste, on the other hand, the required standard is not part of the content of utterances of sentences containing such predicates, but of the circumstances of evaluation 5 with respect to which such utterances are evaluated. As with contextualism, further variation within relativism pertains to which context it is that provides the required standard 6 , as well as to the question of whose standard it is that gives the value of the standard of taste parameter of the circumstances. The same options that were available for contextualism are available for relativism too; and, as before, I will assume a flexible version of relativism, according to which the standard of taste that enters into the circumstance depends on the details of the context in which the utterances to be evaluated appear. And assuming that in the exchange above the relevant standard is that of the speaker, DIALOGUE will be rendered as

DIALOGUE (R)
Anne: The brownie is tasty. Bob: The brownie is not tasty. Now, it is a common view that disagreement is related to semantic content in such a way that, for example, two people disagree if they assert contents that contradict each other. 7 Since this is the notion of disagreement the relativist bases her challenge on, and since it will come up often in what follows, let us flag it for convenience: Conversational Disagreement (CD): Two interlocutors disagree in a conversation if they assert contradictory contents in that conversation. 8 If such a notion of disagreement is accepted 9 , by comparing the renderings given to DIALOGUE by relativism and contextualism it is easy to see that the former can, while the latter cannot, yield disagreement. As can be seen in DIALOGUE (C), the contents of Anne and Bob's utterances don't contradict each other. And as can be seen in DIALOGUE (R), the contents of Anne and Bob's utterances stand precisely in that relation. Thus, what I call 'the challenge from disagreement' basically amounts to the contextualist's duty of explaining disagreement in exchanges like DIALOGUE in a coherent, non-ad-hoc way -on pain of losing the upper hand to the relativist. 10

Putting more flesh on the bones: the various uses of predicates of taste
As briefly mentioned at the beginning of the paper, several broad contextualist (or contextualism-compatible) strategies to answer the challenge from disagreement have recently surfaced in the literature. One is to reject (CD) as the right characterization of disagreements arising in exchanges like DIALOGUE. Instead, such disagreements are explained in terms that don't involve contradictory contents, perhaps even contents that are not propositional. To this effect, various authors have appealed to what is known from expressivist literature as 'disagreement in attitude' (Huvenes 2012(Huvenes , 2014Marques 2014;López de Sa 2015) or practical disagreement (Stojanovic 2011(Stojanovic , 2012Marques and García-Carpintero 2014). Another strategy proposed by contextualists is to keep (CD) as the right notion of disagreement when it comes to exchanges like DIALOGUE, but to relegate disagreement to a different level of content altogether, one that pertains to pragmatics rather than to semantics. Thus, various authors suggest that the perceived disagreement arises at the level of presuppositions (López de Sa 2008;Parsons 2013) or at the level of implicatures (Sundell 2011;Schaffer 2011;Huvenes 2012), with other pragmatic strategies being possible too (see Sundell (2011) for a comprehensive list). Now, as also mentioned at the outset, in what follows I will concentrate on certain considerations that, while strictly speaking not amounting to a full-fledged strategy in answering the challenge from disagreement, can be found in the work of a significant number of contextualists tackling the issue of disagreement. The common thread of those considerations is to point to certain uses of predicates of taste that haven't been on the relativist's radar in launching the challenge from disagreement and which, contextualists claim, can yield disagreement even if (CD) is accepted and the disagreement is deemed genuinely semantic. The type of consideration just described stems from a legitimate dissatisfaction with scenarios such as DIALOGUE in that they are too underspecified to support intuitions of disagreement; as Schaffer (2011: 211) puts it, 'the case for relativism relies on a misrepresentative sample of underdeveloped cases'. The contextualists I engage with aim to exploit such dissatisfaction to their advantage by showing that once the scenarios are fleshed out by using the predicates of taste appearing in exchanges like DIALOGUE in certain ways, disagreement is not problematic anymore. In this section I present four uses that contextualists have mentioned vis-à-vis the issue 83 Brought to you by | Facultad de Filologia y Authenticated Download Date | 11/29/16 3:36 PM of disagreement, while in the next section I assess their dialectical import for the view.

Exocentric uses
One of the uses of predicates of taste that contextualists (e.g., Cappelen and Hawthorne 2009) have mentioned in connection with the challenge from disagreement is known as their 'exocentric' use (the term has been coined by Lasersohn (2005)). A speaker uses a predicate of taste exocentrically if she is taking on someone else's point of view -if she puts herself in someone else's shoes, as the saying goes. This use contrasts with the more widespread one, the 'egocentric' or 'autocentric' use, in which the speaker takes on her own point of view. Employing such a use, perfectly natural exchanges can be constructed in which one interlocutor uses a predicate of taste egocentrically, while the other exocentrically -from the point of view of the first interlocutor. To illustrate, let's go back to our initial scenario and imagine that at some point the issue of whether Anne finds the brownie tasty arises. A disagreement over that issue ensues, with Anne using 'tasty' egocentrically, while Bob using it exocentrically:

DIALOGUE-EXO
Anne (speaking egocentrically, from her own point of view): This brownie is tasty. Bob (speaking exocentrically, from Anne's point of view): No, it's not tasty. Remember the one you had one yesterday? You found it way too sweet, lacking in chocolate and dry.
The exchange in DIALOGUE-EXO is most naturally interpreted as Bob intending to correct Anne about herself finding the brownie tasty. As with DIALOGUE, the intuition of disagreement is present here too, but this time the contextualist is able to account for disagreement. The flexible contextualist will give the first two sentences of the exchange above the following rendering:

DIALOGUE-EXO (C)
Anne: The brownie is tasty for Anne. Bob: The brownie is not tasty for Anne.
Not only is disagreement accounted for in this case, it is also accounted for in a way that preserves (CD): the contents of Anne and Bob's utterances are contradictory. 11

Collective uses
Another use of predicates of taste that contextualists (e.g., Glanzberg 2007;Cappelen and Hawthorne 2009;Huvenes 2012) have mentioned in connection with the challenge from disagreement is what could be called a 'group' or 'collective' use. A speaker uses a predicate of taste collectively if she is taking on the point of view of a group/collective. Employing such a use, perfectly natural dialogues can be constructed in which two interlocutors each use a predicate of taste collectively, with the group relevant for the interpretation of both utterances being the same -for example, a group that both interlocutors belong to (a use that could be dubbed 'egocentriccollective'). To illustrate, let's go back again to our initial scenario and imagine that at some point the issue of whether the students at Anne and Bob's department find the brownies made by Claire tasty arises. A disagreement over that issue ensues, with both Anne and Bob using 'tasty' egocentric-collectively:

DIALOGUE-COL
Anne (speaking egocentric-collectively, from the point of view of the students at their department): This brownie is tasty. Claire took the other brownies to the department and every student liked it -I just spoke with her on the phone. Bob (speaking egocentric-collectively, from the point of view of the same group): No, it's not tasty. You forget that I don't like it at all. Neither does Claire, by the way.
The exchange in DIALOGUE-COL is most naturally interpreted as Bob attempting to correct Anne about the group that both belong to finding the brownie tasty. As with DIALOGUE and DIALOGUE-EXO, the intuition of disagreement is present here too, and this time the contextualist is again able to account for disagreement. The flexible contextualist will give the first two sentences in the dialogue above the following rendering ('S' represents the group of students at Anne and Bob's department):

DIALOGUE-COL (C)
Anne: The brownie is tasty for S. Bob: The brownie is not tasty for S.
Not only is disagreement accounted for, it is also accounted for in a way that preserves (CD): the contents of Anne and Bob's utterances are contradictory. 12

Generic uses
Another use of predicates of taste that contextualists (e.g., Stojanovic 2007;Glanzberg 2007;Cappelen and Hawthorne 2009;Moltmann 2010;Snyder 2013) have drawn attention to is known as their 'generic' use. A speaker uses a predicate of taste generically if she has in mind what is generally, or usually, considered to be the case among the members of a (contextually determined) group. 13 Employing such a use, perfectly natural exchanges can be constructed in which the two interlocutors use a predicate of taste generically. To illustrate, let's go once again back to our initial scenario and imagine that at some point the issue of whether people in general 14 find the brownies made by Claire tasty arises. A disagreement over that issue ensues, with both Anne and Bob using 'tasty' generically:

DIALOGUE-GEN
Anne (speaking generically): This brownie is tasty. Most people I know like it. How could they not? It's great! Bob (speaking generically): No, it's not tasty. The majority of people have never even heard of brownies. And from the ones who have, the majority doesn't like them.
The exchange in DIALOGUE-GEN is most naturally interpreted as Bob attempting to correct Anne about people in general or usually finding the brownie tasty. As with the other exchanges presented in previous subsections, the intuition of disagreement is present here too, but, again, the contextualist is able to account for disagreement. There is more than one way to give a contextualist semantics that captures the generic use of predicates of taste 15 , but the following (flexible) contextualist rendering of the relevant sentences of DIALOGUE-GEN seems to capture the spirit, if not the letter, of all those proposals:

DIALOGUE-GEN (C)
Anne: The brownie is tasty for people in general. Bob: The brownie is not tasty for people in general.
Not only is disagreement accounted for in this case, it is also accounted for in a way that preserves (CD): the contents of Anne and Bob's utterances are contradictory.

Metalinguistic uses
The last use of predicates of taste that contextualists (e.g., García-Carpintero 2008;Sundell 2011;Plunket andSundell 2013, 2014) have focused on in connection with disagreement that I investigate is their 'metalinguistic' use. This use is illustrated by the following example found in Barker (2002), which served as a model for most contextualists taking up the issue: Normally, (1) will be used in order to add to the common ground new information concerning Feynman's height: (1) Feynman is tall.
But (1) has another mode of use. Imagine that we are at a party. Perhaps Feynman stands before us a short distance away, drinking punch and thinking about dancing; in any case, the exact degree to which Feynman is tall is common knowledge. You ask me what counts as tall in my country. "Well," I say, "around here, ..." and I continue by uttering (1). This is not a descriptive use in the usual sense. I have not provided any new information about the world, or at least no new information about Feynman's height. In fact, assuming that tall means roughly 'having a maximal degree of height greater than a certain contextually supplied standard', I haven't even provided you with any new information about the truth conditions of the word tall. All I have done is given you guidance concerning what the prevailing relevant standard for tallness happens to be in our community; in particular, that standard must be no greater than Feynman's maximal degree of height. The context update effect of accepting (1) would be to eliminate from further consideration some candidates for the standard of tallness. My purpose in uttering (1) under such circumstances would be nothing more than to communicate something about how to use a certain word appropriately -it would be a metalinguistic use. (Barker 2002: 2-3) Thus, a speaker uses a predicate of taste metalinguistically when the purpose of the conversation is to offer information not about what is tasty or not, but about what 'tasty' does or should mean in a given context. 16 Employing such a use, perfectly natural exchanges can be constructed in which two interlocutors use a predicate of taste metalinguistically -perhaps in a situation like the one described by Barker in the background. Thus, imagine that Lydia, a foreign student who just arrived at their department, visits Anne and Bob, and wanting to find out about the local culinary habits and the locals' way to use 'tasty' to better blend in, inquires into what the locals consider tasty. A disagreement over the meaning of the word ensues, with both Anne and Bob using 'tasty' metalinguistically: The exchange in DIALOGUE-META is most naturally interpreted as Bob attempting to correct Anne about what the meaning of 'tasty' is in the context in which the exchange takes place. 17 As with the other exchanges presented previously, the intuition of disagreement is present here too, but, again, the contextualist is able to account for disagreement. I'm not sure how exactly to express the content of a sentence that includes a predicate of taste used metalinguistically, but the following rendering of the relevant sentences of DIALOGUE-META seems to capture the spirit of the proposal:

DIALOGUE-META (C)
Anne: The meaning of 'tasty' is such that brownies fall under it. Bob: The meaning of 'tasty' is such that brownies don't fall under it. 18 Not only is disagreement accounted for in this case, it is also accounted for in a way that preserves (CD): the contents of Anne and Bob's utterances are contradictory. 19

The relevant use of predicates of taste, and the alleged impossibility of relevant disagreement
So far I have merely presented the uses of predicates of taste contextualists have mentioned in connection to the challenge from disagreement. For each of those uses, it is shown that, when sketchy exchanges like DIALOGUE are fleshed out in the right way, disagreement can be easily accounted for, even if understood along the lines of (CD). But what exactly is the dialectical gain of pointing to such uses for the contextualist? The claim cannot simply be that the envisaged ways of fleshing out DIALOGUE are possible and that the relativists have ignored them in launching their challenge, because that would be a move with very little dialectical value. Although it is true that most scenarios devised by relativists are relatively sketchy and that they have focused on a rather limited number of cases (a noteworthy exception is Lasersohn (2005) the contextualist that such ways of fleshing out DIALOGUE are possible. Given that the version of relativism assumed here is a flexible one, and thus that the relevant standards of taste can vary according to the details of the context, there is no problem for the relativist in accounting for the disagreement in DIALOGUE-EXO, DIALOGUE-COL and DIALOGUE-GEN. As for DIALOGUE-META, there is no impediment for the relativist to accept the existence of metalinguistic disagreement, and thus accounting for cases involving predicate of taste in which such disagreement occurs is equally unproblematic.
So the contextualist claim that accompanies the showcase of various uses of predicates of taste surveyed above must be stronger: namely, that disagreement in exchanges like DIALOGUE is possible only if the predicates of taste involved are used in the ways presented. Such a claim would solve the challenge from disagreement because now disagreement is relegated to cases in which the contextualist can explain it -even if understood along the lines of (CD). But although this claim would give the contextualist the dialectical strength needed, there is at least a prima facie problem with it: namely, the existence of (what I take to be) a very natural, entirely straightforward case of disagreement in which both interlocutors use predicates of taste egocentrically. As I mentioned above, a speaker uses a predicate of taste egocentrically when she takes on her own point of view. To illustrate such a case, let's go back for the last time to our initial scenario and imagine that the issue at stake is whether the brownie made by Claire is tasty for each of them. A disagreement over that issue ensues, with both Anne and Bob using 'tasty' egocentrically: The exchange in DIALOGUE-EGO is most naturally interpreted as Bob attempting to correct Anne about the brownie being tasty. As with the other exchanges presented in section 2, the intuition of disagreement is present here too, but this time the contextualist cannot yield disagreement, for her rendering of the relevant sentences of DAILOGUE-EGO is the following:

DIALOGUE-EGO (C)
Anne: The brownie is tasty for Anne. Bob: The brownie is not tasty for Bob.

Dan Zeman
The contents of Anne and Bob's utterances don't contradict each other, so disagreement along the lines of (CD) is not secured. 20,21 So, in order for the contextualist claim to have a bite, the exchange in DIALOGUE-EGO has to be ruled out as a case of disagreement. That contextualists intend to do so is obvious from the fact that scenarios fleshed out in similar ways to DIALOGUE-EGO are systematically missing from the lists of admissible disagreements various contextualists provide (e.g., Stojanovic 2007;Glanzberg 2007: 15;Cappelen and Hawthorne 2009: 110-111;Snyder 2013: 297), to give just some examples). However, such an absence is not always fully motivated by contextualist -at least not explicitly.
One motive for the contextualist exclusion of the exchange in DIA-LOGUE-EGO from the category of disagreements could be that the intuition of disagreement in such a case is missing. It is hard (perhaps impossible) to argue with intuitions, but let me nevertheless note that basing the refusal to recognize the exchange in DIALOGUE-EGO as a case of disagreement on lack of intuitions is a very ad-hoc move. It is important to stress what exactly this move amounts to: namely, to claiming that the intuition of disagreement is present in all the other ways to flesh out DIALOGUE except DIALOGUE-EGO. Given that this is the central case the relativist relies on in raising the challenge from disagreement, it is difficult not to see the move as methodologically dubious. It should be also noted that, even allowing this divergence in the intuition of disagreement across cases, that the result is clearly not progress-conducive: what ensues is at best a stalemate -contextualism will be right for those with contextualist intuitions, while other views (for example, relativism) will be right for those with different intuitions. This, in my opinion, could hardly be seen as progress on the matter.
Another, more principled move for the contextualist would be to appeal to some additional, theory-external ancillary hypothesis such as semantic blindness (the idea that speakers of a language are generally and systematically ignorant -'blind' -when it comes to some aspects of the workings of their own language) to explain away the recalcitrant data. According to such a hypothesis, both the intuition of disagreement and the interlocutors' feeling that they disagree are due to the fact that we are blind to how predicates of taste work and as a result we perceive disagreement where there is none. Now, this is a move that the contextualist might very well be entitled to make. However, two things merit stressing in this connection: first, not all contextualists that should acknowledge the need to appeal to semantic blindness do so; second, among those who do, very few actually give detailed explanations of why speakers exhibit such ignorance (something that

The Many Uses of Predicates of Taste and the Challenge from Disagreement
I gather most theorists would agree is desirable). 22 In addition, the following methodological point can be made: what we are after is assessing a view's ability to account for a certain phenomenon; by appealing to additional ancillary hypotheses we are not testing the view's ability in itself, but that of the conjunction of the view's tenets with the theory-external hypothesis. It is important to keep this in mind in comparing the view with its rivals. So, while appeal to semantic blindness, when suitably motivated, might not be a bad thing in itself, it certainly constitutes an additional theoretical burden that will have to be weighted when a thorough assessment of the view is made. 23 Luckily, not all contextualists have been silent with respect to DIA-LOGUE-EGO. Although considerations about this case are rather sketchy, several (possible) arguments can be reconstructed from remarks found in the literature. I don't pretend to be exhaustive here; perhaps more arguments (or beginnings of arguments) can be found. However, the following seem to be the most prominent. 24

The argument from similarity
One way to support the claim that exchanges like DIALOGUE-EGO are not cases of disagreement is to provide similar exchanges in which the intuition of disagreement is lacking. Thus, after making it very clear that they don't take exchanges like DIALOGUE-EGO to give rise to disagreement, Cappelen and Hawthorne draw attention to the following case: Consider a case where one speaker says 'That is F' and another says 'That is not F', where F is a predicate of personal taste. In cases where F is clearly being used [egocentrically] in both cases, there is no intuition of disagreement. (...) [T]he 'Lost Disagreement' problem for contextualism has been considerably exaggerated. (2009: 126) The two authors attempt to offer some intuitive support to their conclusion by putting forward two types of case. The first type involves 'fun', and the conclusion is supported by exchanges like the following: a child who is very happy to be going to a summer camp utters 'The summer will be fun', while his father, who will have to work the whole summer to pay the child's camp utters 'The summer will not be fun'. The second type involves 'filling', and the conclusion is supported by an exchange between a human (or a child) who utters 'This steak is filling' and a talking lion (or an adult) who utters 'This steak is not filling', while each eat the same quantity of food. In neither of these cases do we have the intuition of disagreement, and thus, they think, the conclusion is supported.

91
Brought to you by | Facultad de Filologia y Authenticated Download Date | 11/29/16 3:36 PM I find neither of the two types of example well chosen to support the conclusion that no disagreement arises in exchanges like DIALOGUE-EGO. To start with the second, 'filling' seems to be quite different from other predicates of taste like 'tasty'. Whether something is filling or not depends on a (more or less) determinate, objective amount of food that one has to eat and is connected to the (species-and age-specific) capacity of one's stomach. Surely the amount of food needed for something to be filling is different in the case of a human (or a child) and that of a lion (or an adult). So it is no wonder that we don't judge the human and lion as disagreeing.
In the case of other predicates of taste (like 'tasty') no such determinate, objective amount exists. To acknowledge this, one need not deny that 'filling' is a predicate of taste (although that is certainly an option); one has only to be aware of the fact that they function differently and that what is needed for the interpretation of one is not needed for the interpretation of the other (see also Lasersohn (2011, footnote 4)).
In relation to the first example, one way to explain the lack of disagreement could be to claim that the child and the father, despite using 'the summer' to designate the same period of time, do in fact refer to a series of completely different events: for the child, it is the various activities he will partake in during the summer camp; for the father, tedious hours of work. Surely, whether a certain period of time is judged as fun or not depends on the events that have taken place during that period of time. Since the activities the two will partake in are very different, there's no wonder why we don't judge them as disagreeing. As Lasersohn (2011: 437) notes, sometimes we predicate a property only about certain aspects of an object, not about the object as a whole; similarly, 'we may also apply predicates to parties or summers based on their applicability to relevant parts'. Which part is relevant, Lasersohn argues further, is a matter of context. Thus, the feeling that there is no disagreement in the child/father case comes from contextual effects that have nothing to do with standards of taste. 25 The consequences of such an explanation might be considered too hard to swallow; however, note that even if Cappelen and Hawthorne are right with respect to this particular example, this only shows that certain exchanges about whether a certain period of time was fun or not don't give raise to disagreement and is silent with respect to exchanges about whether a certain object (like our brownie) is tasty or not. Judging a period of time to be fun seems significantly different from judging an object to be tasty, and Cappelen and Hawthorne don't provide a similar example involving 'tasty'. It is fair then to say that their cases don't support the conclusion that the exchange in DIALOGUE-EGO is not a case of disagreement.

The argument from awareness of difference in standards
Addressing the issue of disagreement, Stojanovic (2007) discusses a case similar to DIALOGUE-EGO. Her objection is the following: Now, even though, in matters of taste, people sometimes reach some kind of agreement by realizing that they like different things and that their taste matters to the truth of their statements, at other times they persist disagreeing (...) What do speakers who disagree on whether something is delicious tout court actually disagree about, if the truth of what they say depends indeed on a particular judge and his or her taste? If such speakers are aware that what they say cannot be evaluated for truth unless a judge, or a point of view, has been supplied, and if they supply different judges, they must realize that their seemingly contradictory utterances may be simultaneously true, simply because of being evaluated at different judges. But (...) there would be no matter for disagreement in such a case. (Stojanovic 2007: 694, my  The argument put forward by Stojanovic and Moltmann seems thus to be a reductio: assuming that the exchange in DIALOGUE-EGO is a case of disagreement, and taking the interlocutors to be aware that they hold different standards of taste, then there won't be anything for them to disagree about. Since there is no matter of disagreement, it follows that there is no disagreement at all. Several things can be said in reply. First, note that the fact that sometimes people continue to disagree even if they are fully aware that their interlocutors hold different standards (I take this to be a fact, since I have myself been part of such disagreements, and perhaps the reader too 26 ) has to be explained. If the contextualist thinks that there is no disagreement in an exchange like DIALOGUE-EGO while the interlocutors continue to behave as if they disagree, then the only way to explain why they continue to behave in such a way is by making the interlocutors semantically blind. As I mentioned above, this move comes with a cost for the contextualist, while for the relativist explaining the fact that the interlocutors persist in disagreeing comes very easy: they do it because they disagree in the first place. Second, suppose that the contextualist is right and (what they take to be) the impression of disagreement disappears when interlocutors become aware that they hold different standards of taste. Does this necessarily show that there was no disagreement in the first place? The contextualist takes the interlocutors' realization that they don't hold the same standards as the moment in which the impression of disagreement vanishes; but such a realization could also coincide with the moment in which the 93 Brought to you by | Facultad de Filologia y Authenticated Download Date | 11/29/16 3:36 PM exchange transforms from one in which disagreement is present into one in which it is not. Thus, the interlocutors' realization is compatible with them disagreeing in the first place. Furthermore, even if we agree that there is no disagreement in exchanges in which the interlocutors are aware that they hold different standards, what does this show about exchanges in which interlocutors are not aware that they do so? Here the point is perhaps that such a situation is impossible, and that interlocutors register that they diverge in standards of taste pretty quickly. But how quick is quickly? As several authors have noted (see, e.g., Lasersohn (2005), but also contextualists like Francén (2010)), disagreement need not be persistent: short-lived disagreement is disagreement alright. Exclusive focus on exchanges (or stages of such exchanges) in which the interlocutors are aware (or become aware) that they hold different standards in lieu of exchanges (or stages of such exchanges) in which they are not (do not) strikes me as methodologically biased. 27 As for the matter of disagreement -Stojanovic's main worry abovethere is a way to claim that the interlocutors disagree over whether the object to which the predicate of taste is applied has the property denoted by the predicate or not. Thus, in DIALOGUE-EGO, Anne and Bob's disagreement would be over whether the brownie is tasty or not. As we have seen in section 2, predicates of taste can be used in a variety of ways. Lasersohn (2009) associates such uses with taking a stance, but in addition to the stances corresponding to the uses above, he also talks about an 'acentric stance': a neutral stance in which speakers don't take anyone's point of view, not even their own. It is reasonable to claim that, when people abstract away from the exchanges they happen to find themselves in and judge that something is tasty or not according to this or that standard, they adopt such a neutral stance. However, this doesn't impede speakers from taking an egocentric stance when involved in an exchange like DIALOGUE-EGO. When speakers take such a stance, the matter for disagreement is simply whether the brownie is tasty or not. Since this is the stance that is relevant in DIALOGUE-EGO, Stojanovic's worry is put to rest.

The argument from retreating to explicit relativizations
Some further remarks by Stojanovic (2007) could be interpreted as the basis of another argument against cataloguing the exchange in DIALOGUE-EGO as a case of disagreement. Thus, after mentioning the possibility that interlocutors in exchanges like the ones focused on in this paper retreat by using explicit relativizations like 'This brownie is tasty for me', Stojanovic submits that 'acknowledging that one has been talking about one's own 94 Brought to you by | Facultad de Filologia y Authenticated Download Date | 11/29/16 3:36 PM taste strongly suggests that there was no genuine disagreement in the first place' and that the interlocutors 'meant to be talking of themselves and of their own tastes, and, therefore, did not really mean to contradict each other ' (2007: 693).
I think this argument can be interpreted in two ways. According to the first interpretation, retreating to explicit relativizations has to actually happen in a dialogue. According to the second, retreating can remain a mere possibility for the speakers. 28 In its first interpretation, the argument is weak: even if retreating in the way Stojanovic suggests is possible, or even widespread, interlocutors can keep disagreeing without retreating (as Stojanovic herself acknowledges) -depending on their spare time, whim or argumentative skills. Ignoring such cases in favour of the ones in which an interlocutor actually retreats is methodologically faulty. Further, retreating to explicit relativizations is not necessarily a sign that disagreement was missing in the first place. There are many reasons why interlocutors might retreat: among them, the mere desire to stop the exchange, the signaling of a concession without giving up entirely, the realization that their position is problematic and that more thought needs to be put into it etc. None of these situations is incompatible with there being a disagreement before the retreat was made. (In fact, some of the reasons mentioned above only make sense if we assume that there was a disagreement -for example, conceding.) So, the most that retreating by using relativizations of the kind mentioned shows is that the interlocutors don't disagree anymore, not that they didn't disagree in the first place. The second interpretation of the argument fares no better: despite having the advantage of not overlooking certain data, the considerations above apply mutatis mutandis. Invoking retreating to explicit relativizations -whether actual or possible -as a separate argument against treating exchanges like those in DIALOGUE-EGO as cases of disagreement simply doesn't work.

Conclusion
In the previous section I surveyed several considerations found in the literature that could be used as arguments (or beginnings of arguments) for the refusal to consider the exchange in DIALOGUE-EGO as a case of disagreement. As we have just seen, these arguments are not very convincing. Given that appeal to sheer intuitions is not a methodologically kosher way to go here, the conclusion is that no good reasons have been given to exclude the exchange in DIALOGUE-EGO as a case of disagreement. Such an exchange can thus stand as the central case on which the relativist bases her challenge. Given that the contextualist strategy investigated in section 2 is silent with respect to DIALGUE-EGO, it follows that it is not a satisfactory answer to the challenge from disagreement. Now, another strategy is available to the contextualist that appeals to the uses of predicates of taste showcased above. The strategy consists in accepting that the exchange in DIALOGUE-EGO is a case of disagreement, but claiming that, while disagreement in the other exchanges (DIALOLGUE-EXO, DIALOGUE-COL, DIALOGUE-GEN and DIALOGUE-META) is both semantic and should be explained by appeal to (CD), the disagreement in DIALOGUE-EGO is either not semantic or should be explained by appeal to other notions of disagreement -for example, those mentioned at the beginning of section 2. Such a mixed contextualist view would have the advantage of not having to reject an exchange that intuitively looks like a case of disagreement, of not making ad-hoc moves and of not having to appeal to additional theoretical claims. However, such a mixed strategy would be viable only if it provides a principled reason for treating the disagreement in DIALOGUE-EGO and the disagreement in the other exchanges differently. For, at least on the face of it, all the exchanges are extremely similar -in fact, the sentences that appear in them are identical. And since a good methodological principle is to treat similar cases similarly, the contextualist owes us an answer. Of course, the contextualist can claim that disagreement in none of the cases above is semantic or that it should be accounted for by appeal to (CD). As I made clear, investigating such strategies is beyond the reach of this paper. As far as the strategy focused on here is concerned, accounting for disagreement in exchanges like DIALOGUE-EGO remains a challenge. 29 N O T E S 1 Intuitions are a tricky business, and it is far from clear to what extent they should be trusted; however, I'm assuming here that intuitions such the one present in DIALOGUE should be trusted to some degree. It should be also noted that the contextualists I will be dealing with in what follows, although perhaps rejecting the intuition of disagreement in certain cases, don't reject it in exchanges like DIALOGUE across the board.
2 Options: by complying with the character of predicates of taste, conceived as indexicals (perhaps following Rothschild and Segal's (2009) treatment of color terms); by saturating a variable for standards of taste found in their syntactic configuration (e.g., Stojanovic 2007; Schaffer 2011); by saturating more than one variable, when construed as akin to gradable adjectives (Glanzberg 2007); by pragmatic processes of the kind postulated by truth-conditional pragmatists (e.g., Recanati 2004).