We Can’t Have No Satisfaction

Many authors agree that there is a dimension of conflict expressed through discourse that eludes purely semantic approaches. How and why do conative attitudes conflict? The latter question is the object of this paper. Conflicts of attitudes are typically modelled on one of two models. The first imposes a Subjective Rationality constraint on conflicting attitudes, and the second depends on the impossibility of Joint Satisfaction. This paper assesses whether either of the two conditions can account for conflicting attitudes. First, it argues that Subjective Rationality cannot account for intersubjective conflicts. Second, it presents putative counterexamples to Joint Satisfaction. The counterexamples arise on the assumption that the attitudes are first personal. The paper then explores two alternatives: nihilism about attitudinal conflicts, and dropping the assumption that the relevant attitudes are first-personal states. Embracing nihilism would be devastating for expressivists and other non-cognitivists. But dropping the assumption on which the counterexamples to Satisfaction depend requires a new account of the conative attitudes expressed in value discourse. The paper concludes by pointing to an alternative.


Introduction
Many authors agree that conflict expressed through discourse eludes purely semantic approaches.The idea is crucial to metaethical expressivism.The explanatory task of expressivists has focused, to a large extent, on how sentences express certain kinds of conative attitudes.Yet theorists have not been entirely successful in answering the question of how and hy conative attitudes conflict.The latter question is the object of this paper.How should conflicting attitudes be explained?Conflicts of attitudes are typica ly mode led on one of two schemas.The first imposes a Subjecti e Rationality constraint on conflicting attitudes, and the second requires the impossibility of Joint Satisfa ion.
This paper assesses the two models of conflicting conative attitudes.The first section arues that Subjective Rationality cannot account for intersubjective conflicts.The second section presents putative counterexamples to Joint Satisfa ion.'The counterexamples arise on the assumption that the attitudes are first-personal sin ular.The third section explores two alternatives: nihilism about attitudinal conflicts, and drop ing the assumption that the relevant attitudes are first-personal states.Embracing nihilism would be devastating for expressivists and other non-cognitivists.But drop ing the assumption on which counterexamples to Satisfa ion depend requires a new account of the conative attitudes expressed in value discourse.The paper concludes, in the last section, by pointing to such an alternative.
Subjective rationality Stevenson's Facts and Values (1963) starts by clarifying the nature of ethical disagreement, drawing a distinction between doxastic and conative disagreements.Philosophers, mostly meta-ethicists, have assumed that his distinction tracks a real difference: Let us begin by noting that 'disagreement' has two broad senses: In the first sense it refers to what I shall call 'disagreement in belief.'This occurs when Mr. A believes p, when Mr. B believes not-p, or something incompatible with p, and when neither is content to let the belief of the other remain unchallenged.[…] In the second sense the word refers to what I shall call 'disagreement in attitude.'This occurs when Mr. A has a favourable attitude to something, when Mr. B has an unfavourable or less favourable attitude to it, and when neither is content to let the other's attitude remain unchanged (Stevenson, 1963, p. 1).
Stevenson then explains what he means by "disagreement in belief " and "disagreement in attitude": The difference between the two senses of 'disagreement' is essentially this: the first involves an opposition of beliefs, both of which cannot be true, and the second involves an opposition of attitudes, both of which cannot be satisfied (Stevenson, 1963, p. 2).
Expressivists hold that evaluative sentences express desire-like states.Hybrid forms of expressivism account for evaluative thought and discourse by combining the attitudinal conative component expressed in discourse and the truth-conditional semantic content that is encoded.Schroeder (2009) outlines the various alternatives on offer, as we l as the central questions a dressed by the different positions.According to him, the answers to four crucial questions constitute the conceptual map of hybrid expressivism.The questions are: Q1: Do different sentences containing the word 'wrong' express different desire-like states?Q2: Do different eakers express different desire-like states with the same sentence?
Q3: Does a given sentence have a different descriptive content for different eakers?Q4: Does the descriptive content of a sentence depend on the desire-like state it expresses?(Schroeder, 2009, p. 261) The fact that these questions demarcate its conceptual space shows that explaining conflicts of attitudes is crucial to hybrid expressivism.Any answer expressivists give presupposes that there are two forms of disagreement: disagreement in "attitude" and disagreement in "belief " .Sentences containing words like 'wrong' wi l (directly or indirectly) express desire-like states.
Schroeder is almost unique in the literature in voicing a wor y about expressivism (Schroeder, 2008).He points out that most expressivists merely assume that pairs of (different) conative attitudes are incompati le, or inconsistent, without explaining how they conflict.I share the concern: I think that none of these looks remotely satisfactory as an expressivist explanation of why 'murdering is wrong' and 'murdering is not wrong' are inconsistent.None answers the basic question of what makes disapproval and tolerance of murdering inconsistent with one another.Each posits that there are such mental states that are inconsistent with one another, but none explains why (Schroeder, 2008, p. 587).
Recently, contextualists and relativists have also tried to explain resilient impressions of disagreement in value discourse by pointing to the expression of conflicting conative attitudes.2Contextualists and relativists face the same challenge as expressivists: why are certain conative states incompati le?The cha lenge is to explain how and hy conative attitudes conflict.
How should conflicting attitudes be explained?Two conditions are implicit in Stevenson's quote above.The first condition is one of subjecti e rationality.This cor esponds to cases where one subject has a "favoura le attitude to something" , and the other has an "unfavoura le or less favoura le attitude to it" .The second condition is one of satisfa ion.The conditions can be expressed as: Subjecti e Rationality Subject A's φ attitude conflicts with subject B's ψ attitudes just in case it is not possi le fo an individual to rationa ly have a pair of attitudes φ and ψ.Satisfa ion Subject A's φ attitude conflicts with subject B's ψ attitudes just in case subject A's and subject B's attitudes cannot be jointly satisfied.
Several authors uphold Subjective Rationality as a general principle of doxastic and non-doxastic disagreements.For example, Kö bel (2004) and Egan (2012) clearly have a rationality condition in mind.In spite of the ap arent plausibility of a rationality condition for doxa ic disagreements, the condition is pro lematic in the non-doxastic case.
In the first place, it is not clear whether one is irrational just because one has a pair of different conative attitudes like desires, or even other emotions (e.g., love and hate, fear and hope).There are reasons to be sceptical.In the Treatise, Hume ar ued that: [I]t is only in two senses, that any affection can be called unreasonable.First, When a passion, such as hope or fear, grief or joy, despair or security, is founded on the supposition or the existence of objects, which really do not exist.Secondly, When in exerting any passion in action, we chuse means insufficient for the designed end, and deceive ourselves in our judgment of causes and effects.Where a passion is neither founded on false suppositions, nor chuses means insufficient for the end, the understanding can neither justify nor condemn it (Hume, 1978, II, iii, 3, p. 415).
For Hume, there are two ways for "an affection" to be unreasona le: (i) Its object does not exist, or (ii) The means chosen for the desired end are insufficient and the "affection" is not satisfied.
These two conditions point towards unreasona lenes , not ir ationality, and are grounded on the impossibility of satisfa ion.One may wish that one's deceased relative hadn't died.The wish may be unreasona le, because it is unsatisfia le, but one is not ir ational for desiring one's relatives to be sti l living.If Hume is right, the subjective reasona leness constraint for conative attitudes depends on the conditions of satisfa ion of those attitudes.A subject's attitudes are in conflict, in the Humean sense, if the satisfa ion of one attitude precludes the satisfa ion of the other attitude.More genera ly, it is unreasona le for a subject to have a pair of conative attitudes just in case they can't be jointly satisfied.

Reasona lenes Constraint
It is reasona le for A to have conative attitudes φ and ψ only if satisfa ion conditions C(φ) and C(ψ) are met.
In the second place, the fact that it is unreasona le for A to have certain attitudes φ and ψ, because A cannot satisfy both φ and ψ, te ls us nothing about a conflict between A and B when A has an attitude φ and B has an attitude ψ.After a l, it is possi le that A' s attitude φ and B' s attitude ψ are jointly satisfia le.For example, it may be unreasona le for A to want to end the weekend hiking in the Alps and sunbathing in Gran Canaria.Yet, there is no conflict between A and B when they desire to end their weekends doing different things (unless there is a presumption that they should be doing those things together).
Moreover, it is also possi le that two people are in conflict even when they have the same attitudes towards the same things.Donald Trump and Hi lary Clinton both wanted to be the 45 th President of the United States.In a relevant sense, they both desired the same thing. 3Yet only one of them could be the next president.It is hence possi le for two people A and B to have conflicting attitudes in spite of having the same desire to φ.
It fo lows that having different desires, or desiring different things, is neither sufficient nor necessary for having conflicts of desires, as the examples discussed testify.Further examples can be easily produced.David Lewis elegantly made the same point: If Jack Sprat and his wife both prefer fat meat, they desire alike.They are psychological duplicates, on this matter at least.But they do not agree in their desires, because no possible arrangement could satisfy them both.Whereas if Jack prefers the lean and his wife the fat, then they differ psychologically, they do not desire alike.But they do agree, because if he eats no fat and she eats no lean, that would satisfy them both.In general, they desire alike if and only if they desire de se to have exactly the same properties and they desire de dicto that the same propositions hold.They agree in desires if and only if exactly the same world would satisfy the desires of both; and a world that satisfies someone's desires is one wherein he has all the properties that he desires de se and wherein all the propositions hold that he desires de dicto.Agreement in desire makes for harmony; desiring alike may well make for strife (Lewis, 2000, p. 75, my emphasis).
Here, Lewis distin uishes two notions: desi ing ali e and agreeing in desires.The first ap lies to "psychological duplicates" , people with the same desires and dispositions.The second ap lies to the possibility of joint satisfa ion.Desiring alike may we l make for strife, i.e., having the same desires is no uarantee of agreement.
There are two reasons to reject the idea that Subjective Rationality gives the conditions for intersubjecti e attitudinal conflict.First, the subjecti e unreasona lenes of holding two attitudes φ and ψ depends on their joint unsatisfiability.Second, even if having some attitudes φ and ψ is subjectively unreasona le, that fact is neither sufficient nor necessary for intersubjecti e conflict between someone's attitude φ and someone else's attitude ψ.
The next section considers whether Satisfa ion offers more promising conditions for conflict, and presents an apparent counterexample.

Satisfaction
For Stevenson, a "disagreement in attitudes" involves an op osition of attitudes that cannot be jointly satisfied.If Satisfa ion is to capture the conditions under which conflicting attitudes occur, more has to be said about why certain pairs of attitudes give rise to conflicts.Now, the Satisfa ion principle does not ap ly to many attitudinal conflicts over ap raisals and evaluations where it presuma ly should ap ly.It fails because it is possi le that two different people p ima facie have an attitudinal conflict, but have jointly satisfia le attitudes.Sup ose that A and B have a dialo ue like (1), where they seem to express a conflict. (1) A: I like fatty meat.B: We l, I don't like fatty meat!4 The example is important for assessing the possibility of hybrid expressivism accounting for attitudinal conflicts, as we l as those of other recent theories, like some versions of relativism and of contextualism. 5 The dialo ue in (1) illustrates the kind of explanation of disputes that theorists seek.The sentences used semantica ly express truth-conditional ('descriptive') content.If we assume that A and B are sincere, what they say is true.Moreover, A and B express their re ective attitudes towards fatty meat-A likes it, B doesn't.A has a disposition to have certain pleasura le responses when eating fatty meat.'B has a disposition to have certain responses of dis ust when faces with the pro ect of eating fatty meat.
In (1), the non-doxastic attitudes conveyed can both be satisfied.A's disposition to enjoy fatty meat is satisfied if she indeed feels pleasure while eating fatty meat.B's disposition to be dis usted by it is satisfied insofar as B refrains from approaching fatty meat entirely.Since both dispositions towards fatty meat can be satisfied-A can eat what she desires and B can refrain from eating what she doesn't desire-there seem to be no grounds for those desires to be in conflict, apart from the fact that they are different.Notice that if the attitudes at stake were li ing and being indifferent to there would sti l be no conflict.A can eat what she likes and enjoys, and B has no preference either way. 6here is nothing extraordinary about this case.The same counterexamples arise whether the attitudes at stake are being fo or being against, a proving and disa proving, desi ing and not desi ing, etc.The important assumption here is: (A) The conative attitude expressed is first-personal sinula , e.g.: my desire for coffee.
This raises an ap arent pro lem for Satisfa ion.There are cases where people have first-personal attitudes that are jointly satisfia le but are in conflict.
We could ask why we cannot ap eal to the Subjective Rationality constraint.After a l, A and B are perceived to be in conflict in (1) and neither can accept the other's attitude without changing her mind.Unfortunately, this response ignores three issues.
First, it is not fortuitous that we mention rationality in connection with disagreement and conflict.The motivation for the Subjective Rationality constraint is not merely that it is not pos i le to have a certain pair of attitudes.The link between A and B disagreeing when A believes p and B believes not-p, and A' s irrationality in believing both p and not-p is not accidental.If p is a standard possi le world proposition, A and B cannot be both right because their beliefs are contradictory.For the same reason, A cannot have a coherent set of beliefs if she believes p and not-p.For standard propositions, the impossibility of joint truth grounds both the disagreement between A and B and A' s ir ationality.However, the first section warned us of the difficulties in extending this idea to conative states.
Second, the impossibility here is not one of rationality or lack of coherence.It is, at best, one of physical impossibility.In other words, we are assuming that having one disposition is physica ly incompati le with having the other disposition.We are assuming that if A is delighted to eat fatty meat, she cannot at the same time be dis usted or indifferent to it.And we may even be mistaken in sup osing that there is any physical incompatibility.Perhaps there is none.Perhaps there are things that delight A precisely because they dis ust her (some dark jokes, hor or films, Ha loween tricks, etc.).
Clearly, the presumed incompatibility here is not one of rational coherence.If A were delighted with fatty meat, and B indifferent towards it, B could not adopt A's attitude without 'changing her mind' .The presumed incompatibility does not result from the irrationality of having both attitudes, but from the physical impossibility of having an attitude towards something while being indifferent towards the same thing.
Third, not a l dialo ues like (1) express conflicts.When we give examples of conflicting attitudes like the above, we tend to fi l in the context of the dialo ue in such a way as to understand it as an expression of a conflict.But we can just as we l imagine that two friends are at a re aurant responding to the waiter's su gestions from the menu.
(2) Waiter: Today's two ecials are a traditional dish of the Cerdanya-fatty meat with mashed potatoes and green ca bage, or cod with gri led vegeta les.A: I like fatty meat.B: I don't like fatty meat.
A chooses one option, uttering 'I like fatty meat' , and B chooses the cod, uttering 'I don't like fatty meat' .In (2), A and B are sincere, and their utterances are litera ly true.The interlocutors conversationa ly implicate which dish they prefer.They can do so because it is relevant to the conversation to te l the waiter what they prefer.Notice further that it is sti l true that B could not acquire A's attitude towards fatty meat without changing her mind.And yet (2) does not a report conflict between A and B.
In what sense are A and B's different dispositions towards fatty meat in conflict in dialo ue (1)?As long as A and B concur in not forcing their choices on each other, both can have their preferences satisfied.An ap eal to different individual dispositions by itself does not explain hy even in this case they seem to disagree.If each of them expresses a personal preference, with no consequences for what the other wi l eat, where is the remaining conflict?Given that we have dismissed Subjective Rationality, and that Satisfa ion also seems una le to explain conflict when the attitudes at stake are first-personal, we are faced with a puzzle.The next section puts forward two alternative ways to han le the puzzle

Two responses
This section considers two responses to the puzzle of conflicting attitudes.The first adopts a nihilistic er or-theoretical view about attitudinal conflicts.The second drops the assumption (A) on which the putative counterexamples to Satisfa ion depend.

Nihilism
Sup ose that we take seriously the reasons against Rationality as a condition on attitudinal conflicts, and moreover that we accept that conflicts occur only when those attitudes cannot be jointly satisfied.This means that a great number of a parent conflicts of attitudes are not real.This is a disaster for various kinds of value theories.Any theory that relies on desire-like attitudes to explain "disagreement in attitude" would be undermined.At best, such a theory would be a le to accommodate doxastic disagreements-insofar as it predicts doxastic disagreements at a l.For instance, if what is evaluative and motivating in (3) is A's favoura le dispositions towards fatty meat, it would seem that those dispositions can be satisfied if and when A enjoys eating it.If B utters the sentence in (4) expressing her unfavoura le disposition towards fatty meat, it would seem that her dispositions would be satisfied as long as she does not eat fatty meat.
Although sentences (3) and ( 4) ap ear to contradict each other, the conative attitudes they express are not in conflict, since they are jointly satisfia le.
Expressivism about moral sentences holds that these sentences are conventional ways of expressing certain kinds of attitudes.7 Simple naïve expressivism makes sentences like 'fatty meat is bad' express attitudes akin to those expressed by 'Yuck, fatty meat!' and interprets 'fatty meat is bad' as an expression of a eaker's dis ust for fatty meat.A similar explanation is given for moral sentences, where a sentence like 'lying is bad' expresses attitudes like those expressed in 'Boo lying!'An alternative to simple expressivism is hybrid expressivism.It has recently been defended in different ways by various authors, like Barker (2000), Cop (2001), Finlay (2005), Ridge (2006), and also Boisvert (2008).Hybrid expressivism tries to preserve some form of cogniti ism in value discourse while at the same time holding that moral statements express conative attitudes.Different hybrid theories combine these two claims in some way: (I) An evaluative predicate, like ' good'/'bad' , expresses a property that contributes to the truth-conditions of the sentences where it occurs.(II) Speakers express a non-doxastic attitude towards that property (which accounts for the motivational force of moral discourse and thought).
A semantic contextualist can also advocate these two claims.For instance, Huvenes (2012) claims that two people disagree just in case they have incompati le or conflicting attitudes.He holds that disagreements arise from incompati le beliefs, but may also arise from conflicting desires o preferences.The kinds of attitudes that Huvenes considers are examples like li ing c illi and disli ing c illi, which resem le our li ing/disli ing fatty meat.
The ar uments of the first and second sections show that having different first-personal desires is not in itself a basis for conflict.If we cannot explain how and why conflicts of attitudes arise in these crucial cases, we cannot assume that a mere difference in first-personal attitudes amounts to conflict.Given the ar uments against Subjective Rationality and in favour of Satisfa ion, and given the assumption (A) (that the conative attitudes conveyed are first-personal sin ula ), it should fo low that we are genera ly mistaken in thinking that disagreements about value express conflicting attitudes.And if this is right, Stevenson's notion of a disagree ent in attitude over matters of value-an op osition of attitudes, where it is impossi le for both to be satisfied-does not cor espond to any real phenomena.

First-person plural
The conclusion reached in the previous section is that if the conative attitudes expressed in disputes over matters of value are first-personal sin ular, then we are mistaken in assuming that people disagree about values only when they have conflicting desire-like states.Should we reject assumption (A), that the relevant attitudes are first-personal sin ular?What would an alternative look like?On Lewis's (2000) account of value, "something of the ap ropriate category is a value if and only if e would be disposed, under ideal conditions, to value it" (Lewis, 2000, p. 68, my emphasis).
To value something, for Lewis, is to be in a certain sort of motivational mental state: to desire to desire it.This uarantees the internalist connection between value and motivation.Values are the things that e are disposed to desire to desire in certain ideal circumstances (of fu l imaginative acquaintance).On this theory, to find that fatty meat tasty is to be disposed in the right way towards fatty meat, i.e., it is to value ha ing pleasant ustatory expe iences hen eating fatty meat.And to find fatty meat is dis usting is to be disposed in the right way against fatty meat, i.e., to value not being in contact with fatty meat.Furthermore, fatty meat is dis usting just in case e are disposed in the right way against it.
On the Lewisian theory, the value property expressed involves the relevant group to which the eaker belongs.It is, if we want, a first-person plural, or de nobis,8 secondary property.The theory offers further advantages.It is cognitivist, since it accounts for the evaluative property expressed by the value predicate or word-and it can be true or false that fatty meat is tasty (or dis usting), and even that A (or B) can be mistaken about fatty meat's being tasty or not.At the same time, the theory is sufficiently subjectivist and dependent on people's desires to accommodate the perceived importance of conative attitudes in disputes of taste.So, on a dispositional account of value like Lewis' , a value is a dispositional property.The disposition at stake is first-peson plural.This should yield the desired result.A says that fatty meat is delicious.She ascribes to fatty meat the property of disposing us to having pleasure eating it, but B ascribes to fatty meat the property of disposing us to avoid it.If our desire to avoid fatty meat is satisfied, then we won't eat fatty meat.On the other hand, if our desire to have pleasure eating fatty meat is satisfied, then our desire to avoid it is not satisfied.The two value properties consist in desire-like states that cannot be jointly satisfied in the same world.Hence, not only is there disagreement, there is also an attitudinal conflict.
A Lewisian dispositional theory has the resources to explain conflicting attitudes-at least among people that are relevantly similar in certain re ects and constitute a group jointly refer ed to as "we" .It is not my goal here to discuss whether and how such a co lective identity is formed.My goal is to su gest that there is a value theory that essentia ly relies on desire-like states, and that offers better pro ects of accounting for "disagreements in attitude" .A theory like Lewis's seems we l suited to that role.It is better suited than purely expressivist theories, and presuma ly also better suited than some new hybrid theories.
As I have ar ued here, embracing nihilism would be devastating for expressivist and other non-cognitivist theories.By drop ing the assumption that the relevant conative attitudes expressed in value discourse are first-personal sin ular we can explain both disagreements in belief and disagreements in attitude, as Stevenson intended.