Crossed disagreements

The topic of disagreement has been one of the focal points of the philosophical discussion over the last 15 years. Even though it started with so-called ‘faultless disagreement’, a wide variety of different phenomena has been identified and discussed within the fields of epistemology and the philosophy of language. Arguing over an issue, far from being a unitary straightforward phenomenon, is a kind of linguistic exchange that can adopt multiple forms. Some of these concern the object of appreciation. Discussing the length of the screen in front of me has something to do with discussing the aesthetic value of D's new record, but they are also activities that show significant differences. Within both realms, though, our exchanges can also adopt different shapes. Some times, on the one hand, those involved in a dispute assume a priori that there is an empirical finite method able to solve the disagreement, even if that method might be out of reach for them. Disagreements within every realm, empirical science, formal sciences, but also aesthetics or politics, can fall under this description. Some other times, on the other hand, the possibility of this method is not agreed upon, and the parties in the dispute might discover, along the discussion, that they do not share the standards, or perspectives, that support their respective positions. Also, depending on the information that agents have at their disposal during the discussion, we might talk of ‘peer-disagreements’, and discuss the best epistemic alternatives to face them. Objects of appreciation, attitudes toward disagreement, and the epistemic situation of those involved in the conversation might prompt different distinctions within the contemporary discussion on disagreement. The goal of this section is to succinctly cover these differences, only to the extent to which they affect the different attitudes that speakers involved in a disagreement might exhibit with respect to that particular exchange. Only when these attitudes are heterogeneous, in a sense to be specified, we will talk about crossed disagreements. The simplified taxonomy that we will push forward for the purposes of this paper needs to be introduced in the context of the ever-growing complexity of the current literature on the topic of disagreement.

When two or more individuals engaged in a linguistic dispute, assuming that there is an method that might be able to solve the disagreement, even if that method is out of reach for them, we say that they are involved in a fact-dependent, or, straightforwardly factual, disagreement.Footnote ¹ Suppose that Laura Galileo and John Aristotle, two young astronomy students, are arguing about the shape of the Sun. While John claims that the Sun is a perfect, clean sphere and Laura defends that it is full of sunspots, both agree that they could resolve their dispute by looking at the Sun through a telescope that they have at their disposal. Later on, they go to the telescope room and observe the Sun directly, confirming Laura’s suspicion – John then agrees that the Sun has several spots that show its shape to be imperfect. This kind of disagreement can be found not only in empirical sciences but also in, at least, some disputes about ethics. Imagine two friends discussing the morality of bullfighting. One of them claims that bullfighting is morally wrong because suffering is inflicted upon the animal, as it's shown through the animal's behaviour, while the other friend assures that bullfighting is not morally wrong because the behaviour of the bull does not determine whether the animal is really suffering or not. After all, it could be just a coincidence that the bull reacts in the way that it does. Both of them agree that the debate would be over if significant similarities could be established between the animal's nervous system and our own nervous system. An empirical method, in this particular case, would dissolve the moral debate, even though they might lack at the moment the appropriate resources to undertake such a deciding procedure. Speakers engaged in both conversations above are involved in straightforwardly factual disagreements.

However, there are disagreements where the method that would end the dispute is not agreed upon among the parties taking part in the debate. Sometimes we begin to discuss how we should support our perspective on the facts, or the kind of evidence that we should admit, or even what methods we should use to end the discussion.Footnote ² We are not discussing anymore the particular issue that prompted our conversation, but rather the standards that we endorse, those general rules or assumptions that support our respective positions. This kind of disagreements are deep disagreements:

Robert Fogelin argues that in such disagreements – what he calls a ‘disagreement over the underlying principles’ – the disputing parties can be impartial, free from prejudice, consistent and coherent, and yet still disagree. We are faced with such disagreements, then, when there is a confrontation of what Putnam called ‘framework propositions’, or what Wittgenstein understood as ‘rules’.Footnote ⁴ Deep disagreements are not necessarily straightforwardly factual, since the parties need not assume that there is a finite empirical or formal method to settle the matter. They can nonetheless be straightforwardly factual, when even though the discussion zeroes in on the standards accepted by both parties, they assume that a particular empirical discovery, a ‘smoking gun’, might settle the question.Footnote ⁵

Moreover, we can talk about disagreement among ‘epistemic peers’, this is, disagreements where the agents are reasonably symmetrical in certain relevant aspectsFootnote ⁶, they exhibit epistemic parity. Epistemic peers are said to be both cognitively and evidentially alike. Agents are cognitively alike over p if they have the same competence and/or virtues over the p domain. They exhibit ‘evidential equality’ if they have the same evidence on the p domain.Footnote ⁷ Suppose that you find yourself entangled in a dispute with someone whom you take to be an epistemic peer. What can we expect from this situation? Should we review our beliefs, or should we stand firm in our attitude? Two main positions in the epistemology of disagreement provide answers to this question: the conciliationist view and the steadfast view. According to the former, a doxastic review should be carried out in light of a peer disagreement, while according to the latter view, one should stick to one's own beliefs even in the face of such a disagreement. Conciliationists argue that one cannot remain as confident as one previously was when confronted with a peer disagreement.Footnote ⁸ The steadfast view, on the other hand, maintains that one should rather focus on improving one's own beliefs when faced with a peer disagreement, instead of automatically adjusting them.Footnote ⁹

In the literature on the epistemology of disagreement, the cases of disagreement between epistemic peers are usually presented as fact-dependent disagreements, this is, disagreements in which both parties accept that there is a smoking gun – a trace that cinches the case for a particular causal story – that, sooner or later, might settle the dispute. And it is in relation to these cases that the different doxastic attitudes one can adopt are to be evaluated. But this is not necessarily the case. The epistemological question concerning whether two or more agents have the same access to the relevant evidence, and the same cognitive abilities, is orthogonal to the semantic question regarding the kind of disagreement that they are taking part in. Two moral philosophers can discuss the principles behind a particular interpretation of trolley cases, for example – i.e., a deep disagreement where there is no empirical, finite method, accepted a priori, able to settle the dispute – and, nonetheless, be considered epistemic peers, if they have access to the same factual resources and they are cognitively alike. In other words, the question of what to do in the light of a disagreement does not need to be limited to fact-dependent, straightforwardly factual disagreements. Epistemic reasons, such as the ones that might lead us to declare that a given disagreement is a dispute between epistemic peers, are neither sufficient nor – obviously – necessary to establish the attitudes of the speakers toward that particular debate, how they conceive of the exchange. Epistemic peers can get involved in fact-dependent disagreements, but also in deep disagreements, and, as it will be shown below, in other varieties of disagreement. The aforementioned epistemic reasons are insufficient to establish the kind of disagreement that we have in front of us. Neither they can, crucially, determine whether the exchange is a homogeneous disagreement, one which discussants conceive in similar terms, or, as we shall see, they have conflicting views with respect to the kind of disagreement that they are involved in.

There is yet a different kind of disagreement that can be distinguished both from fact-dependent and deep disagreements. This new category covers disputes that focus on the meaning of certain expressions, and yet they cannot be deemed as merely verbal disputes. Forward-looking disagreements aim at establishing whether the extension of a particular concept should include a given individual or set of individuals.Footnote ¹⁰ Plunkett and Sundell also point to such phenomenon, albeit with a different terminology. They speak of ‘metalinguistic negotiations’, that is, disagreements that involve a negotiation over the proper use of a term.Footnote ¹¹ Imagine two students discussing whether chimpanzees and other animals should be considered non-human persons. One of them claims that chimpanzees, orangutans and great apes in general should be considered non-human persons, since they have proved to have high cognitive abilities. The interlocutor replies that they should not be considered non-human persons because there are notable differences between us, humans, and such animals. The first student might reply that she was not claiming that there were no such differences, even though we are genetically similar. What she meant was, instead, that they should be considered non-human persons, despite the characteristics that separate us. Thus, they continue to maintain the dispute, where the elements involved are both fact-dependent – they concern facts about different species, and evaluative – they point towards how the extension of the concept ‘non-human person’ should be determined.

Finally, there is a kind of nonfactual disagreement that is neither about our concepts, nor about the standards that we maintain in the dispute. These are non-straightforwardly factual disagreements, Footnote ¹² evaluative disagreements. In an evaluative disagreement, the parties might agree on every relevant fact, epistemically accessible or not, and yet disagree whether something is right, wrong, legal, justified, tasty, beautiful, etc. The goal of an evaluative claim is not to say something about our own standards, but about the particular object – individual, event, etc. – that we are assessing. A disagreement about the inherent musical quality of a thrash metal song does not necessarily have to turn into a disagreement about the musical standards that we endorse. Following Perry,Footnote ¹³ the dispute concerns the standards involved, but it is not about the standards. Our evaluations are only true or false with respect to some standards, the ones that we support, but they are not about these standards, as evaluative disagreements show. Sometimes a disagreement of this kind becomes a disagreement about the standards that speakers uphold, and thus becomes a deep disagreement, but sometimes it does not. When speakers involved in a disagreement discover that they have different standards, and yet their dispute does not automatically turn into a discussion about those standards, they are involved in an evaluative disagreement. Most of our conversations about what to do exhibit this feature. Many times we enter these conversations knowing perfectly well that our interlocutors support standards that are very different from our own. Yet this does not prevent these conversations from taking place, nor adds a tint of irrationality to our behaviour.

In this paper we turn a blind eye to the differences that affect the phenomenon of disagreement and depend on the domain of the object of appreciation, or the relative information available to the speakers – their epistemic status. We focus on the different ways in which a disagreement might be conceived. We will distinguish between three types of disagreements:

Imagine two theologians, A and B, discussing whether eating porcupine is allowed or not. They go over the details of the case, reviewing the possible difficulties of the hunt, the tenderness of the meat, and the transmission of illnesses possibly linked to eating such a particular species. At a given point during the discussion, they discover that their contradictory opinions towards the issue, whether eating porcupine meat is allowed, are based on different sacred texts – they discover that they are theologians professing different religions. At this point they realize that a crucial aspect of their conversation was misconstrued from the very beginning of the conversation. To A's mind, it is not an issue whether eating porcupine is allowed for B's religious community, and vice versa. It is only because they assume that they belonged to the same community, with the same sacred standards, that their dispute made any sense to start with. A and B were involved in a Type A disagreement, one that only made sense because they assumed that their standards were shared. Laura Galileo and John Aristotle are involved in the same kind of disagreement. Think about what would happen if at any point during the discussion Laura discovered that John's opinions about the nature of the Sun were based on a revelation provided by a two-headed dragon he encountered in an ayahuasca trip, beyond all scientific evidence. Most likely their discussion would stop on its tracks.

Sometimes we argue about what kind of standard we should maintain, and it is in the course of the conversation when we realize that the dispute has become about that matter. There are a lot of scenarios where this kind of disagreement can appear. De Cruz and De SmedtFootnote ¹⁴ relate the scientific controversy over the status of the Homo floresiensis, a hominid whose bones were found in Liang Bua, Indonesia. After their discovery, the paleoanthropological community disagreed about the interpretation of the fossils. As a result, two incompatible positions emerged:

1. - The insular dwarfism hypothesis (H _id): the fossils are from a hominid species different from the late Homo sapiens, which evolved from an isolated hominid ancestor in the island of Flores. They experienced insular dwarfism, hence the physioanatomic decreasing.

2. - The pathological modern human hypothesis (H _pmh): the fossils belong to our own species Homo sapiens. They suffered a congenital disorder that caused a reduction of the brain – microcephaly – and the body – congenital dwarfism –.

De Cruz and De SmedtFootnote ¹⁵ point out that the majority of the proponents of H _id come from a background in paleontology and archaeology, while the majority of the proponents of H _pmh have a physical anthropology background. Thus, the first group focus its research on the morphology of fossils, and formulates the hypotheses from its analysis of primitive and modern features, whereas the second group examines the fossils in the context of their clinical conditions and formulates hypotheses to explain the morphology of Homo floresiensis.

In this kind of disagreement, Type B, the epistemic standards that the parties support are not shared. Both groups might share the same level of competence and the same access to the relevant evidence, they just don't share the standards involved in the interpretation of the fossils. The way this evidence is interpreted depends on the epistemic principles that the paleoanthropologists adhere to, and so the different standards shape the way the two groups of researchers perceive and evaluate the same body of evidence. Unlike the disagreement between theologians A and B, when both groups of paleoanthropologists discover that their rivals favor different epistemic principles, their reaction is not to stop discussing the issue, chasing new evidence and publishing in order to convince the other side. When they theoretically engage in a discussion about the principles that motivate each position, their dispute has become a Type B disagreement, one in which the focus is on the standards.

Finally, imagine that two friends are arguing about the value of the latest single by the Spanish singer Rosalía. They have had very different musical upbringings, giving rise to completely different criteria for musical appreciation, and they know that. Yet, they have always appreciated the other one's point of view, and their conversations about music have been the source of some of their nicest, unexpected, musical discoveries. They fail to see eye to eye with respect to Rosalía's new single, though. For one of them is the best song of the year; the other one would avoid listening to it at any cost. They keep discussing, and exploring, the reasons for their liking. They are taking part in a Type C disagreement, one that makes sense under different standards –in this case, the parties know from the very beginning that they do not share standards for musical appreciation, and does not turn into a disagreement about the standards.

The phenomenon that we explore in this paper, that we have labelled ‘crossed disagreements’, is not directly related to the epistemology of disagreement, which is mainly focused on determining what is it that we should do once we realize that we are on a par epistemically with our interlocutor. And even though it can only be zoomed in with the aid of the work done so far in the semantics of disagreement, looking closely at crossed disagreements requires being able to distance oneself from some of the common assumptions in this realm. Perhaps the most salient one of these assumptions is the idea that disagreements are homogeneous. Most real-life conversational disagreeing scenarios, especially when they are part of public exchanges, cannot be confined to a single one of the categories listed above, types A to C. When we express an opinion on a given subject, particularly if this takes place in a public forum, we usually rehearse several different arguments to provide support for our position, and more likely than not, these normally belong to different categories. Replies do not usually track these different steps closely, and thus we end up with situations in which speakers clearly appear to be conceiving the disagreement under different lights. When the parties involved in a public dispute conceive the disagreement in significantly different terms, they are involved in a ‘crossed disagreement’.

This phenomenon has been recently spotted by Kappel:

Something really similar was amongst the examples of undermining propaganda that Stanley deals with in his (2015):

Crossed disagreements are not instances of talking at cross purposes. Crossed disagreements are genuine disagreements, and they should be treated as such. We need to learn how to spot them, and how to navigate them, in order to pass our points across more effectively, rather than dismissing them as cases where different parties confuse a mere exchange of private opinions with a genuine debate.

Before we move on, let's provide a couple of more elaborate toy examples along the lines of the cases introduced above:

Huntington's advice: Lisa and Thom debate the use of gender-inclusive language in public institutions. Lisa asserts that the changing sensibilities of the population should be taken into account, and that ultimately a proposal should be voted. Thom claims that only linguists – or other experts – can know the real impact of such changes, can distinguish genuine motivations etc., and therefore the legislator should only listen to their expert opinions.

Global warming: Lisa and Thom discuss publicly what policies to adopt with respect to global warming. Lisa thinks that it is clear that human intervention has boosted global warming, and that policies should reflect that. Thom thinks that it is not clear whether human intervention has boosted global warming, and therefore we cannot make it the focus of our environmental policies. Lisa insists that the evidence backing her position is overwhelming, while Thom asserts that experts are divided, no matter in which percentage, and that we have no way to adjudicate.

In Huntington's advice, on the one hand, Thom conceives the disagreement as a type A disagreement, while Lisa doesn't. The appeal to experts indicates, in this context, a refusal to discuss the standards that might support one claim or its negation. Thom, thus, presupposes that him and Lisa share, or should share, a particular set of standards. In case it is revealed that they do not, he would probably take this exchange to be over. In Global warming, on the other hand, Lisa conceives that disagreement as a type A disagreement, while Thom doesn't. Thom is seemingly open to discuss the standards, while Lisa is the one who, in this case, would have refused to continue the conversation had she discovered that their respective standards differ.Footnote ¹⁸

It is our contention that being able to detect and prevent crossed disagreements should be a desideratum of a theory trying to offer a systematic account of evaluative uses of language, since many crossed disagreements involve evaluations. The next section is devoted to highlighting the relevance of the phenomenon of crossed disagreements, and to specify the explanatory goals that can be derived from the study of this phenomenon for the analysis of the meaning of evaluative uses of language.

Why do crossed disagreements matter, and what's the issue about them

The effects that disagreements have on polarization and, overall, coordination in joint decision-making have been studied from different perspectives. Within the scope of the philosophy of language we can find those who argue that agreement on the moral or aesthetic standards from which we judge is an essential contribution to coordination.Footnote ¹⁹ We also find those who maintain that agreement is not the only source of coordination, but that disagreement itself can also boost coordination; exposing ourselves to the reasons of others contributes in a crucial way to our moral and aesthetic education, because this makes our opinions less extreme and therefore more open to accept courses of action that would result from deliberation.Footnote ²⁰ A similar split is found in the realm of political science between those who argue that disagreement produces more balanced opinions, and thus facilitates joint decision-makingFootnote ²¹ or at least does not increase polarization,Footnote ²² and those who believe that cognitive biases make agents exposed to discussions about a particular issue radicalize their positions about it.Footnote ²³

Polarization is the rupture of the normal distribution of a population's political beliefs and attitudes on a particular issue. DiMaggio et al.Footnote ²⁴ distinguish polarization as a state and polarization as a process:

Two main notions of polarization can be distinguished in the current literature on the issue: ideological and affective. Ideological polarization is an increase in the division of a population's opinions and beliefs – cognitive mental states. In particular, it has to do with the tendency towards the extremes of the ideological spectrum, and the degree to which they are divided into subgroups. Bramson et al.Footnote ²⁶ tracks up to nine different senses of ideological polarization: diffusion, dispersion, coverage, regionalization, community fracture, distinctiveness, group divergence, group consensus and equality of size. Affective polarization is the increase of hate and other negative emotions – non-cognitive mental states – that a group of people has towards the opposite group, and the increase of sympathy and other positive emotions – also non-cognitive attitudes – towards their own group.Footnote ²⁷

Contemporary democracies have multi-partisan institutions that provide checks and balances, granting that the decisions adopted are, in general, the result of a deliberative process between different parties. These institutions work significantly worse when public opinion is polarized. Polarization contributes to increase alienation from politics, and ultimately to the decline in party identification, political participation and electoral turnout.Footnote ²⁸ Furthermore, David R. Jones shows that higher party polarization in Congress is linked, at least in the US, to legislative gridlock. A political gridlock is a situation which affects the relative inability of the government to enact significant proposals on the policy agenda.Footnote ²⁹ As Cass Sunstein argues, polarization causes self-segregation into groups of like-minded people, which can in turn make it far more difficult to produce sensible solutions for the problems of society. And given this self-segregation, there is a pervasive risk that discussion among like-minded people will breed excessive confidence, extremism, contempt for others and, sometimes, even violence.Footnote ³⁰ According to Marc Hetherington and Thomas Rudolph, within a polarized society, ‘it would not be difficult for partisan elites to strategically manipulate or prime issue attention during the course of a political campaign […]. By guiding the public to focus its attention on a particular set of issues, elites can, in essence, strengthen or weaken the effects of political trust on citizens’ policy preferences’.Footnote ³¹ Similar negative effects of polarization have been recently studied by Levitsky and Ziblatt, Mounk, and others.Footnote ³²

One of the cognitive mechanisms linked to polarization is motivated reasoning.Footnote ³³ Motivated reasoning can be defined as the tendency that leads people to process information – both select and interpret it – in a way that is affected by their prior beliefs. By processing information in this way, people come to conclusions that fit their pre-existing beliefs, or the beliefs associated with their groups. The study of these cognitive mechanisms provides a suitable context to the empirical studies showing that polarization increases when the size and density of the argument pool that is aligned with the core beliefs of your activated political identity increases.Footnote ³⁴ The size refers to the number of different arguments that have to do with a given issue. Density refers to ‘the frequency of pertinent observations, thus constraining the number of learning trials on a topic afforded by the environment. Of two equivalent stimuli, the one object with the higher frequency or density of exposure is likely to be preferred just because there are more opportunities to learn about this object’.Footnote ³⁵ The more you are exposed to the ideas that support the core beliefs of your activated political identity, the more you become polarized.

Crossed disagreements, inasmuch as they are connected to the phenomenon of polarization, can be understood as an instance of undermining propaganda Footnote ³⁶. A public political debate is a disagreement whose purpose is to try and coordinate the people involved in it, those affected by the decisions under discussion. Having such debates contributes to the deliberative ideal linked to most contemporary democracies – the idea that political decisions must be the result of reasoned discussions where stakeholders are properly represented. However, when the parties involved conceive the disagreement in significantly different terms, they are not effectively tackling down the reasons provided by their interlocutors, and therefore they are not contributing to reaching a consensus, a joint course of action that every party accepts as the result of deliberation. Rather, crossed disagreements are used to expand the pool of arguments that support the views that we, and the other members of our political identity, already held, and to add density to that pool – by repeating the arguments that we already had. This is done, with the aid of crossed disagreements, in the safest possible way: while seemingly contributing to a well-respected democratic ideal. We are no longer serving this deliberative ideal, we are de facto undermining it.

Only a theory that is able to accommodate the difference between disagreements of types A, to C can be useful to detect and avoid crossed disagreements. In this respect, as we will show, expressivism has an edge over its competitors. Distinguishing between describing and evaluating seems to be necessary to set apart disagreements A, B, and C; and this, in turn, is necessary to be able to detect and prevent crossed disagreements. It is our contention that the difference between describing and evaluating can only be successfully accounted for if we assume that evaluative uses of language do not impart, or do not only impart, propositional content. This feature will be used to compare expressivism with three different varieties of contextualism: indexical contextualism, implicature pragmatics and presuppositional accounts.

Crossed disagreements: Expressivism and its competitors

Expressivism, in this context, is the label that might be applied to a family of theories. These theories share, essentially, the following features:

1. 1. The explanatory target of expressivist theories is evaluative language. More exactly, evaluative uses of language.

2. 2. Expressivist theories maintain that there is a distinction between saying how things are around us, and expressing an attitude towards those things –between describing and evaluating.

3. 3. Expressivist theories apply a pragmatist strategy – they explain the meaning of our evaluative uses of language by paying attention to whatever it is that we do when we perform these linguistic actions.

Most contemporary versions of expressivism take a dynamic form – specifying the semantic value of a sentence as uttered in a context c is making it explicit how context c is updated by this utterance, and distinguish between imparting, for that purpose, locational information, and orientational information.Footnote ³⁷ Locational information allows us to situate the actual world within a specific region of the context set, by eliminating possible worlds which are not compatible with the situation portrayed as occurrent in our utterance. Orientational information, on the other hand, does not allow us to update the location of the actual world. Rather, it is accommodated by making partitions in the context set, and by establishing hierarchies amongst these partitions.

When I say ‘Toronto just won the Finals’, in a suitable context, I propose that the common ground that my audience and I share is accommodated so as to eliminate every possible world where it was Golden State, rather than Toronto, that won the Finals. I'm providing information about the world, I'm describing my surroundings. If, on the other hand, I express an evaluation such as ‘It is great that Toronto won the Finals' I am still assuming that Toronto won the finals, but I'm not providing any extra information by claiming that that is great – I'm not describing how things are, I am not imparting any extra information about the world. This by no means implies that what I am doing while claiming that it's great that Toronto won the Finals is pointless. It is absolutely crucial to acquire this kind of evaluative information in order to coordinate our actions with others, and, more generally, to better understand human behaviour. Important as it is, though, again, I am not adding extra layers of content to ‘John stole the bike’ when I say ‘it's awfully wrong that John stole the bike’. I am simply letting my audience know what to expect reasonably from my future behaviour with respect to such actions, I am committing myself to a set of courses of action. Updating the context by adding this action-oriented commitments is the role that making evaluations play, according to the expressivist.

Conveying orientational information – expressing action-guiding (conative) attitudes – characterizes evaluative uses of language. Most gradable, multidimensional, subjective, adjectives are used by default to make evaluations,Footnote ³⁸ but these expressions can also be used to express something different. I can say ‘Elizabeth Warren would make a great President of the US’ to express my support for her candidacy – an evaluation, but I can also utter the very same sentence without aiming at expressing these action-guiding attitudes. If I'm simply taking a look at the common standards for evaluating presidential candidates, I can utter the same sentence without conveying any practical information at all. Being able to distinguish between evaluative and non evaluative uses of sentences that are uttered by default to make evaluations is one of the virtues of these varieties of contemporary expressivism, one that particularly pays off while discussing crossed disagreements, as we will see.

Expressivism's natural competitors in the contemporary market for the explanation of evaluative language are different varieties of contextualism. In what follows, we will briefly discuss three different sets of theories that require information from the context of utterance to specify the meaning of evaluative claims:

Within the first group of contextualist theories, ’Contextualism 1’, we find those that incorporate the information regarding the threshold and dimensions associated with gradable adjectives to what is said, to the propositions expressed by uttering sentences containing expressions used by default to make evaluations. According to the simplest instance of this first group, my seemingly n-adic evaluative predicates actually express n + 1adic evaluative predicates, ones that incorporate an argument place to include the standard of evaluation – as specified by the different dimensions and threshold – as it is recovered from the context of utterance. This standard, the norm that supports my claim, can preexist my evaluation, or be constituted by it,Footnote ³⁹ but it is part of what is said, and this is what turns out to be essential for this group. Classic contextualist positions in epistemology, such as De Rose'sFootnote ⁴⁰ fall under this category, but also those who were once called ‘relativists’.Footnote ⁴¹

This first group of contextualist theories runs into some trouble when trying to accommodate crossed disagreements. A successful theory of the meaning of evaluative language must, as we postulated above, be able to provide the tools to detect and prevent crossed disagreements. These theories fail to do so because they are unable to distinguish between evaluative and nonevaluative uses of sentences containing expressions that are used by default to make evaluations. As we saw above, there are certain expressions that typically feature in the sentences that we use while making evaluations – they involve taking into consideration dimensions, thresholds, but these very same expressions can be used to point out what follows from the adoption of certain norms, rather than committing ourselves to certain courses of action. Remember the example involving the difference between judging that Elizabeth Warren would be a great presidential candidate vs. considering that, according to the standards of the common folk in the US, Elizabeth Warren would be a great presidential candidate. When one of the parties takes a dispute to be about evaluations, while the other one is interpreting the claim in the center of the debate as non-evaluative, they are conceiving the disagreement in different terms, they are involved in a crossed disagreement. The first group of contextualist theories is unable to distinguish these uses, and therefore cannot provide any help with respect to the detection of a whole class of crossed disagreements.

While Contextualism 1 shall be able to provide the right predictions with respect to crossed disagreements where type A and types B or C are involved, it will be unable to detect crossed disagreements where type B and type C attitudes are respectively exhibited by each party. Including the standards adopted within what is said renders the same semantic values for evaluative claims – ‘Elizabeth Warren would be a great presidential candidate’, and for merely normative claims, those whose goal is to explore the consequences of adopting a given standard – ‘according to the standards of the common folk in the US, Elizabeth Warren would be a great presidential candidate’. So, even though they would be sufficient to explain cases as Huntingon advice and Global warming, they lack the theoretical resources to detect crossed disagreements involving types B and C.

Approaches under the general label ‘Contextualism 2’ explain what is characteristic about evaluative claims using the resources of Gricean pragmatics. They might adopt the contextualist 1's account of what is said, only to add, at the level of implicatures, a proposition that supposedly captures the special nature of evaluative claims. They are able, then, to distinguish between our saying, for example, that something is morally wrong, and our saying that something is morally wrong according to a certain moral code – which is not necessarily ours. They could, in theory, explain the difference between disagreements of the three types presented above, and thus, detect crossed disagreements involving any possible combination of these types. These approaches, no matter if they rely on conventional,Footnote ⁴² conversational particularized,Footnote ⁴³ or conversational generalized implicatures,Footnote ⁴⁴ face a different set of problems. Any kind of implicature-based approach, first, should be able to make the special effects of cancelling an implicature compatible with the peculiarities of retracting evaluative assertions. Taking back an evaluative assertion is compatible with still supporting the standards that were associated with the evaluation, and this seems to be at odds with the perceived effects of cancelling a conversational implicature – problems for the Principle of Cooperation, or a conventional implicature (Grice's ‘oddity’). Secondly, implicated content is propositional. Any kind of propositional content, as explained above, has a locational impact on the common ground. If the adoption of practical commitments is essentially associated with making evaluations, then an account of that makes evaluative uses of language peculiar must exceed the limits of propositional – locational – information. The mere definition of the notion of implicature blocks this theoretical move.

This problem, being unable to accommodate the non-propositional nature allegedly essential to evaluations, can also cause trouble for presuppositional accounts, Contextualism 3. Presuppositions, as they are normally understood, are propositions whose cancellation make what is said meaningless. They're also characterized by the particular projection behaviour that they exhibit.Footnote ⁴⁵ In order to be able to explain the difference between Type B and Type C disagreements, and consequently being able to detect crossed disagreements, a presuppositional account should be able to adopt a very particular form:

1. a. The presupposed content should be non-propositional – associative perhaps –.

2. b. Evaluative presuppositional failures are never catastrophic,Footnote ⁴⁶ while run of the mill presuppositional failures not always are.

3. c. Finally, the orientational information associated with evaluative claims exhibit a projection behaviour which is clearly different from the projection patterns of most presuppositions.Footnote ⁴⁷

Such a presuppositional theory would be able to successfully accommodate the difference between Type B and Type C disagreements. The consequences of such an extension of the notion of presupposition would have to be clarified for other realms where the notion is applied, and all of that just to provide Contextualism 3 with exactly the same explanatory power that expressivism has.