How Valuable Is It?

Let us begin by showing how, rather obviously, a non-gradable interpretation is not compatible with the Normative Component View. Non-gradable interpretations can be traced back to an early formulation of the FA analysis that was championed by Franz Brentano. Brentano claimed that an object, x, is valuable if and only if love towards x is ”correct” (richtig).¹⁰ Among more contemporary commentators we find that ”correct” has been substituted with ”sufficient reason” or with ”conclusive reason,” or ”ought”¹¹ All of these different views have, for our purposes, one salient feature in common, i.e., they are all formulated in terms of a non-gradable normative component. Due to the non-gradable normative component it should be obvious that it is not compatible with the Normative Component View because the Normative Component View requires a gradable normative component to get off the ground. If the non-gradable interpretation is correct, then Normative Component View is implausible. On the other hand, if one finds the Normative Component View to be intuitively appealing, then this could be taken to count against the non-gradable interpretation.

It seems as if the Pro Tanto Reading of the FA analysis is straightforwardly compatible with the Normative Component View, albeit not in any way committed to it. In order for the Normative Component View to make sense, the normative component must be able to come in degrees, and pro tanto reasons is a bona fide example of something normative that comes in degrees, namely degrees of strength. On this view, if we have strong pro tanto reasons to favor x, then x has a high value. If we have strong pro tanto reasons to favor x and only weak pro tanto reasons to favor y, then x is more valuable than y. How much more valuable x is than y is accounted for by the difference in strength between the pro tanto reasons to favor x and the pro tanto reasons to favor y.

This simple picture of the Pro Tanto Reading, however, needs to become a bit more sophisticated. An object can be valuable in certain respects but also all-things-considered valuable.¹² For instance, global warming might be good in respect to wine production in northern Europe, but bad overall. If the improved prospect for wine production in northern Europe is a pro tanto reason to favor global warming it seems as the initial formulation of the Pro Tanto Reading entails that global warming is good. Likewise, if the prospect of millions of people dying due to global warming is a pro tanto reason to disfavor global warming then global warming is bad. Global warming therefore turns out to be both good and bad at the same time. Furthermore, since it is both good and bad, global warming turns out to be better than itself. The claim that global warming is both good and bad and better than itself might not seem that problematic; as the analysis is formulated above it doesn’t state whether it is an analysis of all-things-considered goodness or a more fine-grained notion of goodness. Consequently it can be argued that the above reasoning does not entail that global warming is all-things-considered bad and all-things-considered good but rather it entails that global warming is bad in some respects and good in other respects. We are hesitant to this suggestion since one would expect the more interesting FA analysis to be an analysis of “value period” and not only “value in a respect” and this suggestion seems to commit us to the claim that the Pro Tanto Reading can only give us an analysis of “good in a respect.” It can also be shown that this line of reasoning comes with other problems.

First of all, even if we admit that things can be good in a respect this is not sufficient to avoid the objection. That is, for some valuable respects the same sort of objection that was presented above can be iterated. Let us illustrate with an example. Global warming is good with respect to wine production in northern Europe and bad with respect to melting ice caps. However, note that there very well could be a reason against bringing about global warming that relates to wine production in northern Europe. In order to determinately know that there are no such reasons the adherents of the Pro Tanto Reading must claim that if there is a pro tanto reason to favor (or disfavor) an object with regards to a certain respect there can be no other pro tanto reasons to disfavor (or favor) that object with regards to the same respect, nor can there be any other pro tanto reasons to favor (or disfavor) that object that is more or less strong than the initial pro tanto reason. This claim should be hard to accept since it seems central for the concept of “pro tanto reasons” that when there is such a reason in favor of ψing there could be a contrasting reason against ψing. ¹³

The natural solution to the dilemma is the idea that what determines the value of an object are not the individual pro tanto reasons, but rather the balance of reasons. It is not fitting to favor global warming just because there are some pro tanto reasons to favor it, rather whether it is fitting or not depends on the strength of the pro tanto reasons taken together. This solution, however, is tantamount to abandoning the Pro Tanto Reading and instead opting for some sort of Overall Reading. This view roughly says that there may be pro tanto reasons to favor x and pro tanto reasons against favoring x and x is only positively valuable if the overall balance of reasons favors favoring x. To illustrate, this would mean that global warming is all-things-considered valuable if and only if the overall balance of reasons favors favoring global warming, and global warming is valuable with respect to wine production in northern Europe if and only if the overall balance of reasons favors favoring global warming with respect to wine production in northern Europe. How valuable it is is analyzed by the balance of reasons.

As previously argued, an FA analysis in terms of non-gradable concepts such as ”sufficient reason” and ”conclusive reason” is a non-starter. It is perhaps possible to reconstrue these views in a manner similar to the above, making room for gradable concepts. However, such a view would thus end up being an Overall Reading and not a strict non-gradable reading.¹⁴

From this we submit that the Normative Component View is more reasonable given the Overall Reading. That is, the Normative Component View is not a plausible view for those who endorse a non-gradable reading of the FA analysis.¹⁵ While we have no great qualms against the Overall Reading as a reading of the FA analysis, it is a disadvantage for the Normative Component View that it is committed to a specific reading of the FA analysis. Especially if one takes into account that, to our knowledge, very few, if any, have in text endorsed an FA analysis in terms of the overall balance of reasons. ¹⁶

Before we compare the Normative Component View with its rival the Attitudinal Component View we first need to properly introduce the Attitudinal Component View.

There is of course a lot to be said about how to understand attitudes, what they are, and how to understand the intensity of an attitude. One way would be gauge the phenomenological experience of what it is “felt like.”²¹ Another way would be to understand it in functional or dispositional terms. Hate, for instance, not only seems to have a certain feel, but is also associated with other characteristics such as a tendency to attack, confront, or be filled with a blinding rage. The more nuanced details in how to understand the intensity of an attitude need not be our main concern; for our present purposes it suffices that we have some sort of intuitive grasp of having different attitudes to a higher or lesser degree of intensity.²²

If we, for reasons of presentation, formulate the intensity of attitudes in terms of the phenomenological experience of what it is like-ness of having an episode of that attitude we can grade the intensity of attitudes on a one-dimensional scale. This has the advantage that the intensity of different attitudes can be placed on the very same scale in terms of their intensity, ranging from lower to higher intensity.²³

It is essential for some attitudes to fall within a certain range on the intensity scale. It seems, again, central for hatred to be quite intense, even low-intense hatred is still quite intense. If it lacked sufficient intensity, we would not experience hatred but rather dislike or something of that kind.

Some attitudes may have a large range, perhaps even covering the entirety of intensity scale. This might be the case for an attitude such as favoring.²⁴ It must also be made clear that pro-attitudes may overlap con-attitudes in terms of their intensity range, love and hatred may for example have the same lower- and upper bounds intensity wise. The intensity does not take into account the valence of the attitude. The Attitudinal Component View claims that it is the intensity, in this wide sense, of the attitude that captures the amount of value an object has. In other words, by knowing that it is fitting to favor x we know that x is valuable, and by knowing with what degree of intensity it is fitting to favor x we know how valuable x is.

It should also be noted that we have mentioned a complex version of the Attitudinal Component View. A much simpler version holds that value ought to be analyzed in a single attitude, e.g., desire. If this view is correct then there is no need to compare the intensity of different attitudes and the attitudinal view would be a rather straightforward analysis.

It could be argued that certain objects may be very valuable, yet it is impossible to have the corresponding attitude with such great degree of intensity. The Attitudinal Component View cannot account for such a case. Note that the problem is not that it may be psychologically impossible to have such an intense feeling of admiration – the FA analysis is silent on what is psychologically possible – rather it is the fact that it seems to be conceptually impossible given the characterization above. The possible lack of symmetry between the intensity of the attitude and the amount of value could thus give rise to a conceptual hurdle for the Attitudinal Component View. However, in order for this objection to get off the ground we need plausible examples that establishes that it is possible that there are objects with a certain degree of value for which the corresponding attitude cannot be felt with the correct intensity. Possible examples might be God, humanity, or the good will, all of which are examples of something that might be claimed to be of infinite value. While we are suspicious towards the notion of infinite value in the first place, we still want to push back on the notion that if something has an infinite value then this should entail that we should, let’s say, admire it with an infinite intensity. We fail to see why an infinite value would need to correspond to an infinite intensity of attitudes, rather, than perhaps the maximal intensity. Furthermore, if one is of the view that there are infinitely valuable objects then perhaps this could be captured by the idea that it is always fitting to, e.g., love God more than what one does. In this way, we are able to capture an object being infinitely valuable by it being the case that it is always fitting to have adopt a more intense attitude than what one currently has.²⁵

Another potential objection is that we make too strong assumptions in characterizing The Attitudinal Component. Let us clarify. If the FA analysis is spelled out in the following way:

Generic FA analysis: x is valuable if and only if it is fitting to favor x.

From this it follows that if it is fitting to favor x then x is valuable. Furthermore, if we make the reasonable assumption that anything that has a value, has a certain amount of that value, it follows that x has a certain amount of value.²⁶

This means, according to the Attitudinal Component View, that if it is fitting to favor x it is fitting to do so with a certain intensity. Reasons for attitudes are thus, strictly speaking, reasons to have a certain attitude to a specific degree of intensity. It might seem odd that it cannot be the case that it is fitting to favor something without it being fitting to favor it with a certain degree of intensity. We do not, however, take this to be an objection to the Attitudinal Component View; it might just as well be seen as a remarkable result that follows if we accept the most reasonable interpretations of the FA analysis. The fact that this feature has not been discussed by others, may be due to the fact that previous work on the FA analysis has focused on the question of how to understand what it is, or means, for x to have value, and in order to answer this question there is no need to specify any degree of intensity of attitudes; it is enough to say that an object is valuable if and only if it is fitting to favor it.²⁷

This connects to another potential problem with the Attitudinal Component View. The proposed view seems to be incompatible with a wider understanding of the FA analysis that includes not only pro-attitudes but also pro-responses construed more broadly. Examples of such responses include actions, such as e.g., preserving an object. Since actions do not come in degrees, it seems odd to claim that it is intensity of the pro-response that corresponds to the amount of value the object has. One cannot preserve, promote, or protect something more or less intensely. On the other hand, it seems intuitive to say that certain objects are fitting to preserve. We see two possible ways to accommodate this intuition while remaining a proponent of the FA analysis and the Attitudinal Component View. The first possibility is to interpret ”being fitting to preserve” in a less literal way. For example, one could take it to mean: ”being fitting to have a disposition to preserve.”²⁸ Dispositions can be more or less strong. The second possibility is to assume some type of bridge-principle that links reasons for certain attitudes to reasons for certain actions. An example of such a principle is John Skorupski’s Bridge principle: “Whatever facts give x reason to feel φ give x a reason to do the φ-prompted action, in virtue of being a reason to feel φ.”²⁹ This principle gives us the means to explain why certain objects are fitting to preserve without thereby saying that they are valuable in virtue of it being fitting to preserve: It is fitting to feel a pro-attitude towards an object, and in virtue of this it is fitting to do the action prompted by said attitude. It being valuable is, strictly speaking, understood purely in terms of fitting attitudes, but there is still a close connection to the fittingness of other pro-responses.

There is perhaps a temptation to opt for a hybrid view. On the hybrid view, how valuable an object is, is determined both by the strength of the normative component and the intensity of the attitudinal component. We advise against going this route. A hybrid view seems to retain all of the problems of the original views without retaining their simplicity. In addition, the hybrid view has its own set of problems. For instance, if both the normative component and the attitudinal component play a role in determining how valuable an object is it is still clouded in mystery exactly what role that is. The hybrid view cries out for a function that tells how the two components together account for the exact value of an object. This function could take many different forms but it is hard to see how one would go about to argue for one formulation rather than the other. Should the strengths be added together, multiplied, or is it a more complex function? Finding leverage for one function rather than another seems like an impossible task.³⁰

Let us once again consider Moore’s objection and compare the Normative Component View and the Attitudinal Component View. We took Moore to highlight two things. First, the structural claim: If “valuable” means “worthy of love” “more valuable” should mean “more worthy of love.” Secondly, an intuition that it is the normative component that ought to do the explanatory work of accounting for how valuable something is. These two issues can easily be addressed. Let us begin with the first issue. The structural claim seems to be based on the linguistic intuition that the comparative form of a predicate F is ”Fer than” or ”more F than” consequently the comparative form of ”worthy of love” ought to be ”more worthy of love.” Indeed, as has been pointed out by Bykvist this principle of compositionality is quite reasonable.³¹ Therefore, the Attitudinal Component View would potentially be in a lot of trouble if one assumes that the FA analysis is a semantic theory about what it means for something to be good, i.e., the view that “good” is synonymous with “worthy of love.” Luckily, however, we are not advancing the FA analysis as a semantic theory, but rather an ontological theory about what it is for an object to be good, i.e. the metaphysical explanation of an object being good is that it is worthy of love.³² Because of this, Moore’s argument that the Attitudinal Component View would violate some semantic principle of compositionality is beside the point.

Moore’s second, and more worrying argument, is that if something has more value, then surely this must be reflected in the normative component of the analysis; the normative seems after all to be central for values. In other words it may seem confused to claim that an object is more valuable and thereby has a stronger normative pull, and yet the analysans fails to capture this. We find this line of response to be harder to refute. We can note, however, that we are in good company when we put forth an analysis that downplays this alleged connection between the evaluative and the normative. Consider for example standard FA analyses of different kinds of value:

Final Value: An object has final value if and only if it is fitting to favor an object for its own sake.³³

Instrumental Value: An object has instrumental value if and only if it is fitting to favor it for the sake of its effects.³⁴

Personal Value: An object is good for P if and only if it is fitting to favor the object for P’s sake.³⁵

Attributive Value: An object is good as an F if and only if it is fitting to favor the object in so far as one wants an F.³⁶

All of these are similar in that they employ the concept of fittingness. What is being modified in these different analyses is not the normative component but the attitudinal component. Consider, for example, the analysis of instrumental value. It can be questioned whether this is a value at all, it is only a means to something that is valuable, yet the normative component is the same as the normative component in the analysans of final value. Surely Moore’s objection should apply here as well: that an object has final value rather than instrumental must be reflected in the normative component of the analysis, by potentially introducing final reasons to favor and instrumental reasons to favor.

From this we submit that if Moore´s objection is to be taken seriously, it should not only be construed as an objection to the Attitudinal Component View, but rather as an objection against most attempts to extend the generic FA analysis to also apply to different kinds of value. We take it, however, that most proponents of the FA analysis do not find Moore’s objection too worrying and indeed accept analyses of different kinds of value such as those above. It does seem unfair to claim that the normative component must do all the heavy lifting in the analysis, the analysans does after all have two parts: a normative component and an attitudinal component.³⁷ Indeed looking at contemporary attempts within in the FA analysis tradition of accounting for various different kinds of values in terms of modifying the attitudinal component rather than the normative component shifts the burden of proof. It establishes some sort of precedent and presumption in favor of the Attitudinal Component View rather than the Normative Component View. If the Normative Component View is correct in this instance, then be this would reason to doubt the prior analyses of e.g., final and instrumental value. However, the outlooks of understanding final value in terms of e.g., final reasons rather than favoring objects for their own sake look quite bleak.

We should not, however, dismiss the intuition that the strength of the normative component plays some role in accounting for the value of the object. Indeed, it is quite persuasive. We share this intuition, and instead of dismissing it we want to propose a different explanation for what role the normative component plays when it comes to the value of object. We suggest that the strength of the normative component accounts for the resilience of the object’s value. The resilience of an object’s value is a dimension of value that has been overlooked in the current debate. This is unfortunate since conceptualizing it may be helpful when trying to get a fuller grasp of the domain of value.³⁸ This notion merits further investigations but we can give a brief characterization of what we take resilience to be. The value of x can be more resilient than y’s in the sense that x value can withstand a greater shift in the balance of reasons and still remain valuable. To illustrate, assume that an object is valuable if and only if there are conclusive reason to favor it. Whether a reason is conclusive depends on the balance of reasons. If the balance of reasons were to shift so that we no longer have conclusive reason to favor the object then the object would no longer be valuable.

The more the balance of reasons can shift while it still being the case that we have conclusive reason to favor x the more resilient the value of x is. Thus, if there is a conclusive reason to favor x to a certain degree of intensity, and a conclusive reason to favor y to the very same degree of intensity, x and y are equally as valuable. Their respective value is determined by the attitudinal component. However, in our scenario let us assume that the pro tanto reasons to favor x vastly outweighs the pro tanto reasons to not favor x, i.e., the balance of reasons is heavily in favor of favoring x. Furthermore, the pro tanto reasons to favor y only minimally outweigh the pro tanto reasons against favoring y. In our pictured scenario, x’s value is more resilient. The value of x is more resilient than y’s value because it would take a greater shift in the balance of reasons to make it the case that we do not have conclusive reasons to favor x than in the corresponding case of y.

With the concept of resilience we can account for the intuition that the normative component plays a role in determining the value of an object. It does not determine the degree of value, but rather another dimension of the object’s value – its resilience.

We have claimed that too little has been said in the literature about how the Fitting Attitudes analysis of value accounts for how valuable an object is. This should be interpreted to mean that little has been said about degrees of value, however, there are suggestions in the literature on how a proponent of the FA analysis should understand comparative value. More specifically there is a suggestion of how to analyze “better than” that seems to be widely endorsed in the FA analysis tradition:³⁹

FA Better than: An object x is better than y iff it is fitting to prefer x over y.

This analysis tells us how good x and y relative to each other without referring to the strength of the normative component or the degree of intensity of the attitudinal component. That is, we know that x is better than y without knowing whether it is more fitting to favor x than it is fitting to favor y, or whether it is fitting to favor x with a higher degree of intensity than y. Prima facie we seem to have an analysis that is neutral with regards to the Normative Component View and the Attitudinal Component View. However, we are not given any information about x’s and y’s monadic value. It might be that both are bad, but that x is less bad than y. In other words, we are given an account of value comparisons without expressing anything about the compared objects’ monadic values. This is of course a possibility but it seems odd to have an analysis that can account for how valuable the object is in relation to other objects without accounting for how valuable the object is by itself, i.e., the analysis is silent on whether the object is very good, somewhat good, or even bad. This line of reasoning is familiar from debates on whether “better than” is more basic than “good.”⁴⁰ Note, however, that the aforementioned objection is even more problematic in this context since the aim of this text is to provide an analysis of how good something is. Providing an answer to how valuable x is (i.e., accounting for its monadic value) in terms of FA “better than” is only a serious contender if it is possible to understand “good” in terms of “better than.” The prospects of this has recently, however, be shown to be quite slim by Johan Gustafsson.⁴¹

Let us clarify what it means for the comparative analysis to be neutral with regards to the two views under consideration. First, it can be concluded that this view is compatible with both the Normative Component View and the Attitudinal Component View. According to the Normative Component View if x is better than y then it must be more fitting to favor x than it is fitting to favor y. This may very well be compatible with the fact that it is fitting to prefer x over y. There is no contradiction in these claims. According to the Attitudinal Component View, if x is better than y then it is fitting to favor x with a stronger intensity than y. This as well seems to be compatible with it being fitting to prefer x over y.

Even though there are no contradictions between these views, it is not obvious how this comparative formulation of the FA analysis relates to the two views discussed here. Can this account be derived from the strength of the normative component and/or the intensity of the attitudinal component that is directed to the objects in isolation as is suggested in the above accounts? Or is it a completely separate account?

Let us first consider what the possibilities are for the Attitudinal Component View. Consider the claim that how valuable the object is should be understood in terms of the degree of intensity of the attitude it is fitting to have towards the object. It may be a conceptual necessity that if it is fitting to favor x with a certain intensity and it is fitting to favor y less intense, then it is fitting to prefer x over y. After all, if it is fitting to favor x with a certain intensity and it is fitting to favor y less intense, then clearly it is fitting to favor x more strongly than y. And to favor something more strongly seems to entail preferring it. Or at least the more moderate claim seems true: If one prefers x over y in terms of admirations, then one admires x more than one admires y.

Of course, “favoring” is here taken as a placeholder for any pro-attitude. The reasoning in the previous paragraph, however, assumes that there is a corresponding comparative attitude to all monadic pro-attitudes. But one can question whether all comparative value have a comparative attitude. On the other hand, it seems to be the case that all attitudes with intensity can have a comparative aspect. As long as the attitude is more intense it seems to follow that one can express a comparison using this attitude. Consequently, the Attitudinal Component View seems to provide us with an answer to why “better than” can be analyzed in terms of fitting preferences.

We can provide an outline of an argument that attempts to show the connection between the Normative Component View and the above analysis of “better than.” In order to argue that the FA analysis of value comparatives is derivable from the Normative Component View, one seems forced to argue that it is the very fact that it is more fitting to favor x than it is fitting to favor y that makes it fitting to prefer x to y. In other words, one is committed to the following principle:

P2 seems to follow from a reasonable view on the logic of preferences. If one prefers x over y then ones desire to x is stronger than ones desire to y. Which entails that if one prefers x over y in terms of admirability then one admires x more intensely than y. From this we can assume that if it is fitting to prefer x over y in terms of admirability, then it is fitting to admire x more intensely than y.

Consider the following example to see why P1 and P2 are not compatible. Imagine that it is very fitting to admire Ann with a low intensity and merely fitting to admire Beth with a high intensity. From P1 it follows that it will be fitting to prefer Ann to Beth in terms of admirability. However, according to P2 this entails that it is fitting to admire Ann more intensely than Beth. This seems implausible considering that it is very fitting to admire Ann with a low intensity and merely fitting to admire Beth with a high intensity.⁴²

To conclude; FA analysis of comparatives does not seem to be as straightforwardly derivable from the Normative Component View on its own. This could of course be taken as a reason to reject the Normative Component View. The fact that the most accepted version of how one ought to analyze value comparatives is not straightforwardly derivable from the Normative Component View should count for something.⁴³ Furthermore, both the Attitudinal Component View and the Normative Component View would turn out to entail that if x is better than y in terms of admiration then it is fitting to admire x more intensely than y. The difference is that the Attitudinal Component View arrives at this conclusion more naturally than the Normative Component View, where it is perhaps at most viewed as an unfortunate consequence. The simplicity of the Attitudinal Component view seems to speak somewhat in favor of it over the Normative Component View. The FA analysis of “better than” consequently seems to be compatible with the Normative Component View and the Attitudinal Component View, but only derivable from the latter.