The New Explanatory Objection Against the Fitting Attitude Account of Value

Abstract

The explanatory objection against the fitting attitude account of value states that if the properties of attitudes explain fittingness facts, but do not always explain value facts, then value facts cannot be identical with or reduced to fittingness facts. One reply to this objection is to claim that the constitutive properties of attitudes also explain value facts, for they are enablers for the value possessed by an object (the “enabling maneuver”). In this paper we argue that the enabling maneuver exposes FA to a new explanatory objection, to the extent that the explanatory role played by the constitutive properties of attitudes in value facts is assumed to be different from the explanatory role they play in fittingness facts.

1 Introduction

One version of the fitting attitude account of value (hereafter referred to as “FA”) maintains that value can be either identified with or reduced to the fittingness of attitudes. Good objects are the fitting targets of positive attitudes and bad objects are the fitting targets of negative attitudes.¹ The debate about FA has almost entirely focused on counterexamples to the biconditional claim entailed by FA, for example: Something is (finally) good if and only if it is fitting to favor it (for its own sake). One line of attack, by now well-known as the wrong kind of reasons problem, presents cases where it may be fitting to favor something (for its own sake), but the object is not (finally) good.² These are cases where the reasons why it may be fitting to favor the object (for its own sake) intuitively have nothing to do with its (final) value. Another line of attack, sometimes referred to as the solitary goods problem, presents cases where an object is good or bad, but it is not fitting for anyone to favor it or disfavor it, and therefore it is not the case that it is fitting to favor or disfavor it.³

However, we have recently proposed an entirely new problem for FA: even if the relevant biconditionals were true, value facts (like x being good) should not be identified with or reduced to the corresponding fittingness facts (like it being fitting to favor x), because the normative explanation of value facts and the normative explanation of fittingness facts are not identical.⁴ We (2021: 1209) present what we call “the explanatory objection” as follows:

• P1 If property P is identical with or reducible to the property Q, then any fact that explains why Q is instantiated also explains why P is instantiated.

• P2 Not all facts that explain why the property of being a fitting target of attitudes is instantiated also explain why the property of having value is instantiated.

• C Therefore, the property of having value is neither identical with nor reducible to the property of being a fitting target of attitudes.

P1 could be supported by an application of the principle of identity of indiscernibles to properties. The principle would then state that P cannot be identical with or reduced to Q unless both properties share the exact same explanatory base. However, as we point out, P1 only needs the weaker principle that P cannot be identical with or reduced to Q in the event that there is some fact that explains Q but that does not also explain P.⁵ While it is easy to see why P1 is intuitively plausible, especially if “explains” is meant in an objective (rather than interest-oriented) sense, P2 requires more careful consideration. We argue that P2 is the result of two claims (ibid.: 1210):

• 1. Fittingness needs attitudes: The fittingness of attitudes is always explained, inter alia, by both properties of the fitting attitudes A and properties of their fitting targets O.

• 2. Value does not need attitudes: Value is explained, inter alia, by the properties that make objects valuable, and while these need to include the properties of the object O, they need not include properties of the attitude A.

The thought behind the two claims is the following. Suppose that we ask why a certain attitude, like admiration, is fitting to a certain object, like a beautiful painting. It seems clear that a plausible answer to this question must invoke both the properties of the object (say, the aesthetic qualities of the painting) and the constitutive properties of the attitude (in this case admiration). After all, fittingness appears to be a two-term relation, and so it would be surprising if the properties of one term did not have any explanatory role to play for the relation. Suppose next that we ask why a certain object, like an act of sadistic torture, is intrinsically bad. It seems that a plausible answer to this question could simply invoke the intrinsic and bad-making properties of the object. For example, we might judge that the act of sadistic torture is bad for its own sake precisely because it is an instance of selfish and deliberate cruelty. At the very least, what seems clear is that the constitutive properties of negative attitudes towards the object are explanatorily irrelevant for the badness of such an object—or at least it would seem so, prior to accepting a view like FA. If all of this is correct, then the normative explanation of fittingness appears to be different from the normative explanation of value.⁶ And if P1 is true, then FA is false.

One initial complication with this argument is that, if the explanatory objection to FA is delivered while accepting the relevant biconditionals (e.g.: Something is (finally) good if and only if it is fitting to favor it (for its own sake)), then it becomes something of a mystery why the biconditional is true. How can two facts (necessarily) co-obtain if the explanations for each fact diverge in the way prospected by the explanatory objection? However, this is a false problem. We do not deny that the properties that make an object good are the same properties that make it a fitting object of favoring, as it were, from the “object-side” of the fittingness relation. The normative explanation of fittingness facts includes the properties that explain why an object is good, it is just that it requires something in addition to this, namely certain properties of the fitting attitude. Therefore, it follows that, given certain properties of favoring, anything that is good will also be a fitting object of favoring, and anything that is a fitting object of favoring will be good.⁷

In this paper we focus on a second complication, which has been discussed by Rabinowicz and Rønnow-Rasmussen (2021) in their reply to our original paper. Claim 2 above (Value does not need attitudes) includes an inter alia clause:

• Value is explained, inter alia, by the properties that make objects valuable, and while these need to include the properties of the object O, they need not include properties of the attitude A.

The clause seems designed to leave room for factors other than the good-making properties to play some explanatory role with respect to value facts. While this might be a reasonable qualification, it also opens the door for an obvious response to the explanatory objection to FA: properties of attitudes could be some of the alia that contribute to explaining value facts, and therefore, contrary to P2, all facts that explain why the property of being a fitting target of attitudes is instantiated also explain why the property of having value is instantiated.⁸ To this extent, FA would be safe from the explanatory objection.

In what follows we first outline Rabinowicz and Rønnow-Rasmussen’s reply (Section 2). We then argue that a new explanatory objection challenges FA (Section 3). Then we consider whether interpreting FA as a claim about grounding avoids the new explanatory objection (and argue it does not) (Section 4). Finally, we discuss whether explanatory objections (old or new) spell doom for FA (Section 5).

2 Properties of Attitudes as Enablers for Value

Rabinowicz and Rønnow-Rasmussen argue that while properties of attitudes do not make things good or bad in any ordinary sense, they enable other properties to become value-makers, and are thus part of the complete explanation of value. We hereafter refer to this as “the enabling maneuver”. Here is what they say:

[T]he properties of an attitude enable the properties of the object to be value-makers simply because they determine the nature of the value in question. Clearly, admirability is a value whose nature in part is determined by what the attitude of admiration consists in. Likewise, the nature of desirability in part is determined by the constitutive properties of desire, and so on. The particular relation of value-making, which obtains between the object’s properties and the value it has, must depend on the nature of the value in question. Consequently, the properties of the attitude, which determine the nature of the corresponding value, have an enabling role with respect to the value-making relation itself. As such, they need to be mentioned in the complete explanation of the object’s value. (2021: 2477)

In this way, attitudinal properties play an explanatory role both for the instantiation of the property of being a fitting target of attitudes, and for the corresponding value property. Therefore, so far as attitudinal properties go, all that explains fittingness facts explains value facts as well.

We in fact anticipate the enabling maneuver, and initially claim that a view on which value is necessarily enabled by certain attitudinal properties would violate a requirement of substantive neutrality (Orsi & Garcia, 2021: 1218). Invoking enablers usually involves substantive evaluations; e.g. to claim that the moral permissibility of a joke enables that joke to be funny (or more precisely, it enables some of the joke’s other properties to make the joke funny) is to hold a substantive value judgment about what is funny. But Rabinowicz and Rønnow-Rasmussen believe that this worry does not apply to their enablers.⁹ The properties of attitudes would not act as substantive enablers. The best way to illustrate the difference is to take a case where there are different kinds of enablers at play with respect to the same value property. For example, on certain views, the fact that a painting is morally unproblematic enables other features of the painting (say, its aesthetic qualities) to make it (aesthetically) admirable—if the painting were morally problematic, it would not be admirable, because those aesthetic qualities would be barred from playing an admirability-making role. Being morally unproblematic is a substantive enabler for the admirability of the painting: it is a substantive question whether a morally problematic painting is admirable or not. One way to spell this out is that two parties can disagree about this question without necessarily possessing a mistaken concept of admirability.

On the other hand, Rabinowicz and Rønnow-Rasmussen think of the “constitutive properties” or “essential properties” (2021: 2477) of admiration and other attitudes as enablers of a different kind: they “enable the properties of the object to be value-makers simply because they determine the nature of the value in question” (ibid.: 5). The idea is that, unlike the property of being morally unproblematic, the constitutive properties of admiration are not substantive enablers for the admirability of the painting, but (to stick with the authors’ terminology) what we might call “value-determining enablers”. This would not seem to be a substantive axiological question: If admiration did not have the constitutive properties it has (whatever they are), would the painting be admirable? The thought must then be that the answer to this question is “no”, and one that is reached on purely conceptual grounds.¹⁰ After all, without an attitude with the constitutive properties of admiration, the relevant value of the painting (its admirability) would fail to get “determined”, and therefore whatever aesthetic qualities the painting may possess, they would not make it admirable. If this is correct, then the enabling maneuver does not incur a loss of substantive neutrality for FA.

3 A New Explanatory Objection Against FA (and Why Two Other Options Won’t Work)

It is unclear whether most or any FA advocates would concur with Rabinowicz and Rønnow-Rasmussen’s enabling maneuver, as nothing like it has ever been proposed before in these discussions. Having said this, in this section we outline a new problem for FA, should its proponents employ the enabling maneuver in response to the explanatory objection.

As just pointed out, for Rabinowicz and Rønnow-Rasmussen the nature of the relevant fitting attitudes plays a non-substantive, but rather “value-determining” explanatory role with respect to value facts. But this move immediately sets up a new explanatory problem for FA, if (as Rabinowicz and Rønnow-Rasmussen concede) the nature of the relevant fitting attitudes does instead play a substantive role with respect to fittingness facts. Here is a paragraph worth quoting in full (bold is for our own emphasis):

Does FA-analysis then go well together with the idea that facts about attitudes are a kind of enablers with regard to value? We think so. Both the properties of an attitude and the properties of its object are included in the resultance base of the fittingness relation between the attitude and the object.¹¹The value of the object is not this relation, though. It is not a relation at all; it is a property of the object. On FA-analysis, it is a relational property: the object has it in virtue of it being a target of a fitting attitude of a certain kind. The relation on which this property is grounded does not of course connect the object to any concrete token of an attitude in question (indeed, no such token might be in existence); it connects it to an abstract attitude type. What, then, is the resultance base of value? Unlike the resultance base of the fittingness relation, it only includes properties of the valuable object (the ones that make it a fitting target of a pro-attitude), while the properties of the pro-attitude that would fit that object instead play the role of enablers. (2021: 2479)

As can be seen from the highlighted passages, Rabinowicz and Rønnow-Rasmussen think of the complete explanation of fittingness facts and value facts as including the same elements playing, however, different roles: in fittingness facts, the properties of the fitting attitudes are part of the resultance base (they make the attitude fitting, together with the properties of the object); in value facts, they are not part of the resultance base (they enable facts in the resultance base to make something good or bad). See illustration 1 for a graphic rendition of this view.

Illustration 1. Rabinowicz and Rønnow-Rasmussen’s view. Shorter light arrow for “is included in the resultance base for”, longer bold arrow for “enables … to be included in the resultance base for”.

A new explanatory objection against FA can now be stated:

• P1* If property P is identical with or reducible to the property Q, then any fact that explains why Q is instantiated also explains why P is instantiated and plays the same explanatory role for P as it does for Q.

• P2* Not all facts that explain why the property of being a fitting target of attitudes is instantiated also explain, with the same explanatory role, why the property of having value is instantiated.

• C Therefore, the property of having value is neither identical with nor reducible to the property of being a fitting target of attitudes.

P2* is what follows from Rabinowicz and Rønnow-Rasmussen’s view. As for P1*, it seems as intuitive as P1 in the original argument: identity or reduction requires not just identity of explanatory factors, but identity of those factors’ explanatory roles too. Here is an analogy. Suppose one said that facts about the optical system of a normal observer directly contribute to making it the case that certain objects appear red to the observer—they are in this sense “appearance-makers”—but then proceeded to claim that the same facts about the optical system of a normal observer only enable those objects to have the property of being red—they are therefore not “red-makers”. This difference in the explanatory role played by facts about the optical system would seem to speak against identifying redness with (or reducing it to) the property of appearing red to normal observers. The same thing seems to be going on in the case of the enabling maneuver. In general, the identification or reduction of one property to another property does not just require that, somehow, all facts that explain the one property manage to sneak into the explanation of the second property, but also that their explanatory roles be identical for the two properties.¹²

Interestingly, Rabinowicz and Rønnow-Rasmussen do not seem to be bothered by this new explanatory asymmetry resulting from their enabling maneuver, while clearly acknowledging it. We will explore why in the next section, but before moving on it is worth briefly discussing two theoretical options that might have occurred to the reader, both of which would avoid the new explanatory objection against FA.

The first option is to think of the constitutive properties of attitudes as part of the resultance base for both fittingness and value: the nature of admiration, for example, as contributing both to making it fitting to admire Mona Lisa and to making the painting admirable. Or consider the fact that an act of sadistic torture is bad, which might be explained by reference to its being a selfish act of deliberate cruelty. The option under consideration invites us to say that in addition to the act being selfish and deliberately cruel, the act is further made bad by facts about the properties of the negative attitudes that it is a fitting target of, like moral resentment and anger. This option would establish explanatory symmetry. See illustration 2 below.

Illustration 2. The first option.

Rabinowicz and Rønnow-Rasmussen do not take this route, and this seems correct to us. The problem with this option is that it would commit FA to leaving room only for a very specific kind of substantive value theory, where properties of attitudes are always contributing to value-making. While it might be an open question whether the role of enablers is always substantive, or whether (as Rabinowicz and Rønnow-Rasmussen argue) there might also be non-substantive enablers, it is always a substantive move to assign to certain properties the role of value-makers, or to view them as part of the resultance base for value facts. In this case, it would be rather hopeless to argue that FA does not violate the requirement of substantive neutrality.

The second option to establish explanatory symmetry between fittingness facts and value facts is to think of the constitutive properties of attitudes as enablers both for fittingness facts and for value facts, as opposed to being part of the resultance base for fittingness but only enablers for value. On this picture, fittingness facts and value facts would be identically explained by properties of the object (qua part of the resultance base for both facts) and by properties of the fitting attitude (qua enablers for both facts). For example, this would mean that the constitutive properties of admiration would be enablers with respect to both the admirability of the Mona Lisa and the fact that the painting is the fitting target of admiration. Likewise, the constitutive properties of moral resentment and anger would be enablers with respect to both the badness of a deliberate act of cruelty and the fact that the act is a fitting target of these negative attitudes. There would therefore be no difference in the explanatory stories behind value facts and behind fittingness facts after all. See illustration 3 below.

Illustration 3. The second option.

We suspect that this is the option many fitting attitude advocates would take, if pressed on this point. Early proponents like A. C. Ewing seemed to suggest as much: “the ground [for a certain attitude to be fitting] lies … in the concrete, factual characteristics of what we pronounce good. Certain characteristics are such that the fitting response to what possesses them is a pro attitude, and that is all there is to it” (Ewing, 1948: 172). The last sentence (“and that is all there is to it”) appears even to suggest that only the properties of the object are relevant to the explanation of fittingness facts.¹³ The option we are considering makes the further claim that the characteristics of the attitude, too, play a role, albeit as enablers for the fittingness fact. This option might as well be foreshadowed in claims like this one: “the total nonevaluative facts are sufficient for goodness—for it to be fitting to value the thing” (McHugh and Way 2016: 594), assuming that the “total nonevaluative facts” include facts about attitudes and not only facts about the object.

However, this reconciliatory move is not convincing, since it entails a rather odd way of looking at the metaphysics of fittingness facts. Remember that the role of enablers is to enable other properties to play their role as part of the resultance base. In the case of fittingness, this would mean that the constitutive properties of admiration enable the aesthetic properties of the painting (e.g. its chromatic harmony, etc.) to make admiration fitting towards the painting. Similarly, the constitutive properties of negative attitudes such as moral anger and resentment would have to enable other properties, such as the deliberate cruelty of an act, to make the act a fitting target of the negative attitudes. This would lead to an explanatory imbalance of sorts between the two terms of the fittingness relation: to give rise to fittingness relations, the object-side (involving the properties of the Mona Lisa and the deliberate act of cruelty, respectively) of the relations would stand in need of enabling by the attitude-side (involving different kinds of constitutive attitudinal properties). This is odd because in fittingness relations, it seems that the terms or arguments of the relation must contribute equally and in the same capacity to the “making” of the relation. One side does not have explanatory priority over the other. Thus, if any enabling happens (and it often does when it comes to normative phenomena), it is operated by factors lying outside the immediate terms of the fittingness relation itself.

Here is a factual analogy to make these points clearer: properly cutting my toenails enables my feet to fit into my shoes. But neither the shape of my foot nor the shape of the shoe enjoys any explanatory superiority over one another in determining the fit. They rather “need each other”, and they both may need relevant external factors (like properly cutting toenails) for the fit to happen. In the case of the Mona Lisa, the properties of the painting and the properties of admiration play an equal role in explaining why the painting is the fitting target of admiration. If anything enables these factors to determine that the Mona Lisa is the fitting target of admiration, then it is probably something that is not already a term or argument of the fittingness relation. Consider the feature of being morally unproblematic as an enabler: a painting being morally unproblematic would seem to enable both the aesthetic qualities of the painting and the properties of admiration to work together in making it fitting to admire the painting. Similarly, in the case of the deliberate act of cruelty, the qualities of the act and the properties of moral resentment and anger seem to play an equal role in explaining why the act is the fitting target of such negative attitudes. If anything enables these factors to determine that the act is a fitting target of moral resentment and anger, then it is once again something that lies outside the terms or arguments of the fittingness relation itself. Indeed, it is difficult to think of anything that might play the role of an enabler in this case. What makes an act of deliberate cruelty a fitting target of moral anger and resentment are the properties of the act and the properties of the negative attitudes. Their contribution to the explanation of the fittingness fact appears to be exactly symmetrical. Neither set of properties seem to need enabling to do their explanatory work, either by one another or by anything outside of the fittingness relation.

In sum, it seems FA cannot escape the new explanatory objection by recalibrating the explanation of fittingness facts and value facts in either of the two ways just described.

4 FA Meets Grounding

As pointed out, Rabinowicz and Rønnow-Rasmussen appear unfazed by the explanatory asymmetry between fittingness and value which is entailed by their enabling maneuver. The reason why may lie in their interpretation of FA, which is contained in the same passage reported above (bold is, again, for our emphasis):

Does FA-analysis then go well together with the idea that facts about attitudes are a kind of enablers with regard to value? We think so. Both the properties of an attitude and the properties of its object are included in the resultance base of the fittingness relation between the attitude and the object. The value of the object is not this relation, though. It is not a relation at all; it is a property of the object. On FA-analysis, it is a relational property: the object has it in virtue of it being a target of a fitting attitude of a certain kind. The relation on which this property is grounded does not of course connect the object to any concrete token of an attitude in question (indeed, no such token might be in existence); it connects it to an abstract attitude type. What, then, is the resultance base of value? Unlike the resultance base of the fittingness relation, it only includes properties of the valuable object (the ones that make it a fitting target of a pro-attitude), while the properties of the pro-attitude that would fit that object instead play the role of enablers. (2021: 2479)

This interpretation does not quite match the characterization of FA that we originally set out from. In our article, FA was understood as the metaphysical claim that value is identical with or reducible to the fittingness of attitudes, but when Rabinowicz and Rønnow-Rasmussen clarify what it might mean for attitudinal properties to be enablers, they seem to adopt a different understanding of FA: not as an identity claim, but as a claim of grounding.

More precisely, for Rabinowicz and Rønnow-Rasmussen FA appears to be the claim that the value property an object has—x being admirable, say—is grounded on, or is a property which x has in virtue of, the relation of x being the fitting target of an attitude of a certain kind, here, admiration. Being admirable is not identical to this relation (“the value of the object is not this relation”). Being admirable is a property belonging to the admirable object (say, to the painting), though one that the object possesses purely in virtue of standing in a relation of fittingness with admiration (or in virtue of such relation obtaining)—hence being admirable is a relational, rather than non-relational, property of the object. Relational properties of an object are therefore to be metaphysically distinguished from the corresponding relations the object stands in. To mark the difference, we could use symbolism such as this (“F” for “is a fitting target of”):

• Relational property: x [F admiration]¹⁴

• Relation: x F admiration

Note that the distinction between the relational property and the corresponding relation that grounds it had better be metaphysical and not purely notional: because otherwise FA would again amount to an identity claim between value properties and fittingness relations. Moreover, it needs to be pointed out that Rabinowicz and Rønnow-Rasmussen surely have in mind what is usually called “metaphysical”, as opposed to “normative”, grounding here (Bliss & Trogdon 2016): if x being admirable were normatively grounded on admiration being fitting towards x, then fittingness facts would implausibly add to the stock of value-makers on top of the object’s natural properties.

Having now settled the contours of Rabinowicz and Rønnow-Rasmussen’s interpretation of FA, the central question can be asked: can FA, understood now as a claim of grounding rather than identity, tolerate the explanatory asymmetry brought out in the new explanatory objection? In other words, can the explanatory story behind what does the grounding (here, relations of fittingness) differ from the explanatory story behind what is grounded (here, value properties)? Rabinowicz and Rønnow-Rasmussen are committed to answering “yes”, or their grounding version of FA would fall prey to the new explanatory objection just as the identity version of FA does. In particular, they are committed to the claim that if a property (or relation) P grounds a property Q, and X is part of the resultance base for P, then it need not follow that X is part of the resultance base for Q. This is because, on their view, fittingness relations ground value properties, but while facts about attitudes are part of the resultance base for fittingness facts, they are not part of the resultance base for value facts—they are enablers for value facts (see again illustration 1, now placing a right-to-left grounding sign instead of the identity sign). But why doesn’t this asymmetry in explanatory role undermine grounding value properties on fittingness relations?

Rabinowicz and Rønnow-Rasmussen do not provide an answer. Now this is not the place for a review of the literature on grounding. However, we will note that putative cases of metaphysical grounding do not tolerate such asymmetries—or, at least, not without further elaboration. Let’s go back to the analogy with color presented above. Suppose that a theory of color claims that the property of being red is not identical to, but is instead grounded on, the property (or relation) of appearing red to normal observers. Suppose further that the relevant explanatory asymmetry holds: certain facts about the optical system of a normal observer directly contribute to making it the case that certain objects appear red to the observer—they are in this sense “appearance-makers”—but the very same facts about the optical system of a normal observer only enable those objects to have the property of being red—they are not “red-makers”. How can color appearances then ground colors, if one element in the supervenience base of facts about color appearances (here, facts about the optical system) plays a different explanatory role in the supervenience base of facts about colors? It seems that any claim of a strong dependence (let alone identity) between color appearances and colors is at least prima facie undermined, if the supervenience base for color appearances is heterogeneous with respect to the supervenience base for color itself in the way indicated above. The same, again, would hold for any claim of strong dependence (let alone identity) between fittingness and value.

Note: metaphysical grounding may, perhaps, tolerate that some element in the supervenience base of the ground be missing from the supervenience base of what is grounded. Not all that explains the ground need be relevant to explaining what is grounded. But in the case at hand things are different, because for FA facts about fittingness relations and facts about value properties are logically equivalent. Therefore the supervenience base for each must be made up of the same elements. Rabinowicz and Rønnow-Rasmussen certainly would agree—that is why they seek to find some room for the constitutive properties of attitudes within the supervenience base for value properties. But assigning two different roles or positions to constitutive properties of attitudes is what, according to us, sounds alarm bells. And this is a particularly urgent point for Rabinowicz and Rønnow-Rasmussen to address, since, as noted above, the metaphysical distance between the ground and what is grounded is, by their admission, very narrow: it is the difference between a relation (x being a fitting target of admiration: x F admiration) and the corresponding relational property belonging to one term of that relation (x being admirable, i.e. a fitting target of admiration: x [F admiration]). Even granting that this difference is real, and not only notional, it is to say the least unexpected that two such closely related metaphysical items, standing in a relation of grounding, and being on top of that connected by logical equivalence, could tolerate any difference in their respective explanatory stories.¹⁵

The upshot is that Rabinowicz and Rønnow-Rasmussen must do more to show that, unlike the identity version of FA, a grounding version of FA is not subject to explanatory concerns. Until this account is fully spelled out, the idea that unlike identity, grounding can tolerate explanatory asymmetries between value and fittingness such as the one pointed out (and admitted by the authors) risks coming across as an ad hoc claim.¹⁶ Moreover, until this task is fulfilled, one cannot proceed to extending a grounding-style FA analysis from response-mentioning value properties (like admirability) to value properties such as being good for its own sake, as Rabinowicz and Rønnow-Rasmussen propose to do at the end of their article (2021: 2480–81).

5 Conclusion: Different Versions, Different Costs and Benefits

In this paper we have laid out no fewer than six versions of FA:

• a) Identity (between fittingness facts and value facts) without explanatory symmetry: Attitudinal properties as fittingness-makers but only enablers for value.

• b) Identity with explanatory symmetry: (1) Attitudinal properties as both value-makers and fittingness-makers; (2) Attitudinal properties as enablers both for value facts and for fittingness facts.

• c) Metaphysical grounding (of value facts on fittingness facts) without explanatory symmetry: Attitudinal properties as fittingness-makers but only enablers for value.

• d) Metaphysical grounding with explanatory symmetry: (1) Attitudinal properties as both value-makers and fittingness-makers; (2) Attitudinal properties as enablers both for value facts and for fittingness facts.¹⁷

As argued in Sections 3 and 4, versions (a) and (c) are challenged by the new explanatory objection (or at least Rabinowicz and Rønnow-Rasmussen have not said enough about why (c) would not be challenged). As argued in Section 3, version (b)1—and therefore by extension (d)1—seem implausible (because they would violate a requirement of substantive neutrality), although they are unchallenged by the new explanatory objection. Versions (b)2 and (d)2 are also unchallenged by the new explanatory objection, but in Section 3 we argued that (b)2—and therefore by extension (d)2—present us with an odd way of looking at fittingness facts.

It is a difficult call to decide which version of FA is the most plausible. But overall the explanatory objections (old and new) seem sufficiently worrisome, and therefore FA theorists had better defend a version of FA (whether as an identity or grounding claim) which avoids explanatory concerns, that is, either (b) or (d). Such versions are not costless. However, the cost of versions b(2) or d(2)—the resulting unintuitive picture of fittingness facts—seems to us inferior than the cost of versions b(1) or d(1)—FA violating a neutrality requirement. After all, fittingness is a partly theoretical construct—theorizing on which is still at a relatively early stage—nor are intuitions about what does the “fitting-making” and what does only the enabling as clear-cut as one might hope. By contrast, substantive neutrality is a well-established requirement in the debate over the analysis of value.

In conclusion, we believe that the costs (and benefits) of a version of FA such as (b)2 or d(2) will have to be weighed against the costs and benefits of rivals to FA: the value-first view, whereby fittingness facts are identical with (or grounded in) value facts; and the no-priority view, whereby fittingness facts and value facts (necessarily) co-obtain, but are not identical, nor does one ground the other (see Rowland 2019 for this taxonomy, albeit against the background of the buck-passing rather than fitting attitude account of value). However, the value-first view appears vulnerable to the same explanatory worries as FA: how can value facts be identical to, or ground, fittingness facts, if attitudinal properties are needed for the explanation of fittingness facts but not for that of value facts? It seems that value-first views will have to resort to recalibrating moves similar to (b) and (d) above, with the added complication that assigning an explanatory role to attitudinal properties with respect to value appears here even more ad hoc than for FA. After all, on FA, it is fittingness facts that “lead the way”, and fittingness facts do require a role for attitudinal properties. But on value-first, it is value facts that “lead the way”, and value facts did not seem, at least prima facie, to require a role for attitudinal properties.

As far as explanatory objections go, therefore, the main contenders seem to be on the one hand FA, in its b(2) or d(2) versions, and on the other hand a no-priority view, which does not require a perfect symmetry between the explanation of value facts and the explanation of fittingness facts, provided (as we pointed out in Section 1) that a plausible explanation for why fittingness facts and value facts co-obtain is available (short of identity, reduction, or grounding in either direction). We leave it to further work to weigh up the overall costs and benefits of these two rival views. ¹⁸.