The Finality and Instrumentality of Value in a Way

Abstract

Final value accrues to objects that are good for their own sakes, while instrumental value accrues to objects that are good for the sake of their effects. The following paper aims to show that this distinction cuts across some surprising areas of the evaluative domain. This means that there may be some unexpected types of value that can come in a final or instrumental form. The argument proceeds by looking at two prominent types of value, namely kind-value and personal value. The former accrues to objects that are good as the kinds of things that they are, while the latter accrues to objects that are good for someone. Substantive examples are offered in support of the idea that these types of value can come in final or instrumental form. The substantive examples are then given additional support by considering the structure and behavior of fitting attitudes.

Introduction

The following paper was presented at the Ninth European Congress of Analytic Philosophy and at the Higher Seminar in Practical Philosophy at Lund University. Gratitude is owed to the attendees of these events for their comments and criticisms. Special thanks to Toni Rønnow-Rasmussen, Wlodek Rabinowicz, Jakob Green-Werkmäster, Jens Johansson, Michael J. Zimmerman, Henrik Andersson, Ingvar Johansson, Berit Braun, Olof Leffler, Frits Gåvertsson, Fritz-Anton Fritzson, Gloria Mähringer, Mattias Gunnemyr, Eric Brandstedt, Sarah-Marie Babbitt, Jared Parmer, Jeroen Smid, and Francesca Bunkenborg.

The distinction between final and instrumental value is arguably one of the most prominent in the field of axiology. Final value is the type of value that accrues to ends, understood as objects that are good or bad for their own sakes.¹Instrumental value is the type of value that accrues to means, understood as objects that are good or bad for the sake of their effects. The following paper aims to show that there are unexpected types of value which are subject to the distinction in question. For the sake of brevity, the paper illustrates this point by focusing on two types of value, gathered here under the banner of value in a way. The first concept we will be considering is that of personal value, which denotes the type of value that accrues to objects that are good or bad for someone’s sake. The second concept is that of kind-value, which denotes the type of value that accrues to objects that are good or bad as the kinds of things that they are.²

The idea that an object can be good for its own sake for someone and that it can be good for the sake of its effects for someone has recently been defended by Toni Rønnow-Rasmussen (2011: 57–59). This idea is given additional support in this paper, as is its unexplored corollary in the case of kind-value: The suggestion is that an object can be good for its own sake as the kind of thing that it is and that it can be good for the sake of its effects as the kind of thing that it is. These ideas are important because if they are sound, it would be a mistake to treat value in a way as if it is in competition with final value for the centre stage of ethics. Some philosophers, such as Richard Kraut (2011), argue that final value does not matter as much as what is good for us.³ If it is true that value in a way can come in a final or instrumental form, then this result might recommend a more conciliatory view of the concepts in question.

The paper is divided into two main parts. Section 3 offers intuitive support for the ideas just mentioned in the form of substantive examples. The purpose of these is to show that value in a way can be of the final or instrumental variety. Section 4 is designed to strengthen the intuitive support, mostly by making observations about the structure and behaviour of fitting attitudes. The argument here proceeds from the observation that we can conceivably favour objects either as kinds of things or for someone’s sake, but in such a way that we also favour them for their own sakes or for the sake of their effects. Not only that, but the examples given in Section 3 could be taken to show that doing so is fitting. This places a burden of proof on potential sceptics, since they will have to insist that regardless of how it is conceivable or even fitting to favour objects, it is still not possible for them to have value in these ways. For the sake of clarity and precision, this argument is made with the aid of the fitting-attitudes analysis, which reduces value to fitting responses.

Some Illustrative Examples

Suppose first that we are considering the personal value of two distinct objects: One of the objects is hay fever medicine, which we judge to be good for people with allergies because it can alleviate their physical discomfort. The other object is pleasure, which we judge to be good for people because of the subjective qualities that the experience of pleasure is associated with. It seems intuitive to suggest that there is something fundamentally different about the personal value ascribed to the hay fever medicine and that ascribed to the pleasure in this case. One way of cashing out the intuition is to suggest that one of the objects is significantly better for us than the other. This suggestion would be right as far as it goes, but another plausible way of cutting up the cake is available: Insofar as hay fever medicine is judged to be good for people with allergies, it is typically good in an instrumental sense. When pleasure is good for the people who experience it, by contrast, it makes sense to say it is always good in a final manner.

The suggestion is not that hay fever medicine could only have instrumental value when it is good for people with allergies, nor that pleasure could only have final value for the people who experience it. It seems rather obvious that a mental state like pleasure could be useful and so potentially good for the sake of its effects as well. The suggestion is rather that insofar as hay fever medicine is good for people with allergies, one of the senses in which it is good for them is the instrumental one. Similarly, insofar as pleasure is good for the people who experience it, one of the senses in which it is good is in the final one. This would help explain the intuition that even if the objects under consideration are both bearers of personal value, their personal values would nonetheless have different normative flavours. The reason, to reiterate, is that one of the personal values, namely that of the hay fever medicine, is of the instrumental variety, while the other personal value, which accrues to the pleasure, is of the final variety.

Let us next consider the unexplored corollary to this suggestion in the case of kind-value: One of the objects under consideration, a toaster, fulfils the standards conventionally associated with toasters and so we judge it to be a good toaster. The other, a human being, conforms to whatever standards we associate with human beings and so we judge her to be good as a human being. An object conforming to the standards conventionally associated with toasters presumably must warm up bread relatively quickly, efficiently, and evenly. The issue of what it takes for a person to conform to the standards associated with human beings is a far more controversial one (cf. Rowland 2016b: 215). This is a significant problem for ethical theories that put a lot of emphasis on the importance of being good as a human. The present paper does not seek to defend theories of this sort, and it does not need to: The suggestion under consideration is that even if there is such a standard, the kind-value that accrues to the toaster and the kind-value that accrues to the human being differ fundamentally from each other.

The fundamental difference just alluded to cannot be understood simply by referring to the descriptive differences between toasters and human beings. There is something else that gives the value of being a good toaster a different normative flavour from the value of being a good human being. Several suggestions as to what this something might be suggest themselves, but one possibility is that the values in question fall on opposite sides of the distinction between final and instrumental value. The idea is that when an object is a good toaster, it is good in an instrumental manner. When someone is good as a human being, by contrast, she is good in a final manner. The suggestion that there is this difference is inspired by ideas concerning the dignity of persons that is often associated with the ethics of Immanuel Kant, although he was not concerned with kind-value or any of the key issues discussed here.

This is not to deny that human beings can also be good for the sake of their effects. Instrumental value is among the types of value that come rather cheap, especially for such useful beings as humans. However, in addition to whatever other values a human being has, it seems fair to suggest that she must be good for her own sake.⁴ One might point out here that objects can be bearers of distinct types of value at one and the same time. An object can be a good toaster while it also happens to be good for the sake of its effects, just as a person can be good as a human being while also being good for her own sake.⁵ This line of reasoning is certainly right as far as it goes, but the view defended here is not simply that distinct types of value can share the same bearer. Instead, the view is that an object can be good for the sake of its effects as a toaster, just as someone can be good for her own sake as a human being.

The examples just laid out are meant to illustrate that objects can have value in a way, but in the sense that they are also good for their own sakes or good for the sake of their effects. The examples appear to me perfectly coherent, although it is of course possible to explain them away and so make room for less reconciliatory views. In other words, if we want to uphold strict boundaries between value in a way and the distinction between final and instrumental value, there are always ways of doing so. Very few examples of this kind are entirely immune from deflationary treatment. The crucial question is whether such a treatment can be made to seem reasonable, or whether instead it ends up surrounded by an air of revisionism. One way of lending additional support to the examples just given is to carefully consider the structure and behaviour of fitting attitudes.⁶

Before we move on it is worth reiterating why any of this matters in the first place. Many people have intuitions about things being good for their own sakes. It seems to them that there are objects out there which should be treated as ends in themselves. Some of those people might also have intuitions about what is good in a way. It might seem to them like there are also objects out there which should be pursued for us, let us say. When we sit down and formulate ethical stories about how we ought to live, there might be a temptation to identify which of these values matters most. The possibility that objects can be good for us in a final way seems very important in that context. For it entails that instead of sacrificing the intuitions we have about one of these value types, we can instead reconcile them. Are there objects that seem to be good for us, but in a altogether different way from other objects that are good for us? One possible explanation is that the former objects are good for their own sakes for us.

The Argument from Fitting Attitudes

This section presents several applications of the fitting-attitudes (FA) analysis, which aims to reduce value to fitting responses.⁷ This focus makes sense given the clarity and precision of the FA analysis and the fact that it can elucidate the structure and behaviour of relevant attitudes in a way that will be helpful. As well as lending support to the examples mentioned in the previous section, the FA analysis will suggest how to account for the types of value they involve. The extent to which the views defended here really end up being dependent on the FA analysis will be discussed later. Until then an object should be taken to have value just in case it is the fitting object of a response of some sort. When applied to the distinction between final and instrumental value, the FA analysis is standardly taken to yield the following:

F: X has final goodness =_df it is fitting to favour X for its own sake.

I: X has instrumental goodness =_df it is fitting to favour X for the sake of its effects.

As is the custom in these contexts, favouring should be a placeholder for a positive response. This response could take the form of a non-doxastic attitude like admiring, respecting or loving. It could also take the form of an action, such as protecting, promoting or preserving. If we are curious about how things would look if we were to focus on badness, the analyses featured in this section can easily be modified. All we would have to do is to replace any mention of goodness with badness and any mention of favouring with disfavouring.⁸ As a next step, then, let us consider how the FA analysis might be applied to value in a way.⁹ One application proceeds from the intuition that value in a way cannot be understood without reference to a certain sort of choice situation:

KS: X is good as a K =_df it is fitting to favour X on condition that S (K).

PS: X is good for P =_df it is fitting to favour X on condition that S (P).

S (K) could perhaps be taken to specify the situation of our choosing or recommending a K. This makes it tempting to understand the favouring referred to by KS in similar terms as either the choice or recommendation of X. Making sense of personal value in this manner is admittedly not quite as straightforward. Perhaps S (P) could be taken to specify the situation of our choosing something which is conducive to the health or happiness of P. This would make it tempting to understand the favouring referred to by PS in similar terms as either the choice or recommendation of X.¹⁰ This means that an object is a good toaster just in case it is the fitting object of a choice or recommendation, given that we are choosing or recommending a toaster. An object is good for someone’s sake just in case it is the fitting object of a choice or recommendation, given that we are choosing something that is conducive to the health or happiness of the person in question.¹¹

Before we move on it is also important to note that the logical character of conditional favourings does not quite reflect the logic of material conditionals. To say that we favour an object on condition that something is the case is not to say that if something is the case, then we favour the object. Whenever we favour something on condition that something is the case we actually favour it. If the circumstances in question do not obtain, then all that happens is that the favouring misfires or becomes void in some way.¹² When the favouring consists in acts like choosing or recommending an object, this is not difficult to make sense of. For example, it seems reasonable to suggest that the circumstances obtaining will then constitute a felicity condition for the acts in question. Philosophers have a fairly good idea about how such conditions are to be understood and so when we look to acts like choosing or recommending an object, conditional favourings need not be as mysterious as appearances might suggest.

When a conditional favouring consists of attitudes, on the other hand, things are less straightforward, for these are not typically taken to be subject to conditions in this way. One obvious way of dealing with conditional attitudes would be to reduce them to non-conditional attitudes. For example, perhaps the favouring of X given that S (K) can be reduced to a preference for the state of affairs [X being chosen when S (K)] over the state of affairs [X not being chosen when S (K)].¹³ The problem here is that this pattern of analysis is also subject to the misdirection objection. It reduces the value of X to a fitting response which is directed toward a very complex state of affairs of which X is a mere part. The result would seem to be that it is the relative values of these complex states of affairs that we are capturing, rather than the value of X as such. We shall return to this issue in the next section, but let us first consider the following variations on KS and PS:

KFC: X is a finally good K =_df it is fitting to favour X for its own sake on condition that S (K).

KIC: X is an instrumentally good K =_df it is fitting to favour X for the sake of its effects on condition that S (K).

PFC: X is finally personally good for P =_df it is fitting to favour X for its own sake on condition that S (P).

PIC: X is instrumentally personally good for P =_df it is fitting to favour X for the sake of its effects on condition that S (P).

KFC and KIC are both meant to capture the general notion of being good as a kind of thing, but in a final or instrumental way. KFC captures the concept of final-kind value while KIC captures the concept of instrumental-kind value. PFC and PIC are meant to capture the general notion of being good for a person, but in a final and instrumental way. PFC captures the concept of final-personal value while PIC captures the concept of instrumental-personal value. Of course, if we do not share the intuition that value in a way cannot be understood without reference to a certain sort of choice situation, none of these analyses will seem very attractive. For this reason, it is perhaps worth considering what would follow if we took a somewhat simpler route. More specifically, we may want to consider the possibility that we can favour something as a kind of thing or for someone’s sake without ever adopting a conditional favouring. We thereby avoid the need to provide for an understanding of how attitudes could be subject to conditions in a way similar to acts such as choosing or recommending an object:

KS: X is good as a K =_df it is fitting to favour X as a K.

PS: X is good for P =_df it is fitting to favour X for the sake of P.

These analyses are simple in that they do not rely on any notion of conditional favouring. Instead, they treat the favourings called for by value in a way as a special type of attitude, placed on equal footing with any other type of attitude we might direct on to the world. This means that the notion of favouring an object as a kind of thing or for someone’s sake is taken for granted rather than reduced to something which might currently be more familiar in philosophy.¹⁴ Rønnow-Rasmussen (2007, 2011: 47–48) is among the philosophers to endorse a version of PS, while Skorupski (2010: 84–86) endorses some version of both KS and PS. The question is whether these simple analyses are amenable to further development which makes the subject to the distinction between final and instrumental value. The answer seems to be that they are. In fact, whenever an analysis of value is couched in terms of some set of discerning attitudes, it seems that the attitudes in question can always be made more discerning in this way:

KFS: X is a finally good K =_df it is fitting to favour X for its own sake as a K.

KIS: X is an instrumentally good K =_df it is fitting to favour X for the sake of its effects as a K.

PFS: X is finally good for P =_df it is fitting to favour X for its own sake for the sake of P.

PIS: X is instrumentally good for P =_df it is fitting to favour X for the sake of its effects for P.

Again, KFS and KIS are both meant to capture the general notion of being good as a kind of thing, but in final or instrumental way. KFS captures the concept of final-kind value while KIS captures the concept of instrumental-kind value. PFS and PIS are meant to capture the general notion of being good for a person, but in a final or instrumental way. PFS captures the concept of final-personal value while PIS captures the concept of instrumental-personal value.¹⁵ Rønnow-Rasmussen (2011: 57–59) endorses versions of FPS and PIS, but no philosopher seems to have explicitly endorsed KFS or KIS. Now, the present analytic endeavour could go on indefinitely, since countless variations on the basic ideas presented here seem to invite themselves. The analyses mentioned so far, however, should suffice to illustrate the crucial point: that there are attitudes that mirror the structure of the types of value purportedly exemplified in Section 3. In fact, it seems plausible that whenever we look at value in a way from the perspective of fitting attitudes, there is always a way to understand the relevant favourings in more discerning terms, so that they incorporate a final or instrumental character.

The objection might therefore be made that the views defended here rely too heavily on the idea that value is reducible to fitting responses. Against this idea, it is often observed that there can be reasons to favour objects that are not good, just like there can be reasons to disfavour objects that are not bad.¹⁶ The most popular example that is meant to illustrate the problem comes from Crisp (2000) and involves a case where we are forced to admire a worthless object on pain of dire consequences. The intuition is that in such a case we may very well have reasons to admire the object, without it being admirable. Examples that are closer to everyday life are not difficult to find: It seems reasonable to suppose that it is fitting for parents to love their own children more than the children of others, yet this does not mean that their children are more loveable (Orsi 2015: 144–145). Many sophisticated attempts have been made to solve the problem, but none of them have enjoyed widespread acceptance among philosophers.¹⁷

We cannot consider all the variations on the objection just mentioned, or the many sophisticated responses that they have inspired, so it is fortunate that we do not need to. In fact, the argument presented in the previous section does not rely too heavily on the idea that value is reducible to fitting responses. It has just been shown that there are attitudes that mirror the structure of the values exemplified in the Section 3. This indicates that it should be possible to favour objects either as kinds of things or for someone’s sake, but in such a way that we also favour them as such for their own sakes or for the sake of their effects. Moreover, the examples given in Section 3 also indicate that there are cases where it is even fitting to favour objects in this manner. These considerations seem to speak heavily in favour of the view that value in a way is also subject to the distinction between final and instrumental value. However, it is true that if the FA analysis is unsound, then those considerations may not conclusively establish the view that value in a way can come in the final and instrumental form.¹⁸

Axiological Attitude Problems

Looking beyond discussions about the connection between value and reasons reveals some very troubling objections that we would do well to consider. The most troubling objection concerns the attitudes that were invoked throughout the preceding section. For one thing, the question remains unanswered exactly what is involved in favouring objects in a way and how this differs from favouring objects for their own sakes or for the sake of their effects. Even if we did have an answer to this question, moreover, reasonable scepticism might persist regarding the more complex and exotic attitudes that were invoked in that section. In other words, even if it makes sense to speak in terms of favouring objects in a way, for their own sakes, or for the sake of their effects, it may be difficult to understand what it means, for example, to favour objects for their own sakes in a way.¹⁹ The fact that we lack a detailed story regarding these kinds of attitudes seems to raise serious concerns about the ideas defended in this paper.

Let us suppose that there is a person who favours an object for another person’s sake. This could be taken to mean that the first person favours the object and that the justification or causal explanation for this involves the other person in some way. To flesh out the example we can imagine that Jones favours going Smith’s birthday party for Smith’s sake. Jones believes that Smith would be happy if Jones attended Smith’s birthday party. The fact that Jones has this belief either gives him reason to favour going to Smith’s birthday party or causally explains his positive attitude about going to Smith’s birthday party. Although this understanding is probably not uncommon in certain contexts, it is not the understanding that is relevant here. One person can favour an object for another person’s sake without him being either justified or caused to do so by something involving the other person. Similar lines of reasoning are meant to apply to the other attitudes that have been mentioned here as well.

The lesson is that to claim that we are favouring an object in a way, for its own sake, or for the sake of its effects, is to make claims about intentional structure. To favour an object for a person is to favour the object with an eye toward that person (Rønnow-Rasmussen 2011: 56). The person will figure within the intentional content of the attitude, although not in such a way that it is the person that is favoured rather than the object. To favour an object as particular kind of thing is to favour the object with an eye toward the properties that ground its membership of this kind, but not in such a way that it is the properties that are favoured rather than the object. To favour an object for its own sake is to favour the object with an eye toward the object as such, whereas to favour an object for the sake of its effects is to favour the object with an eye toward its uses and consequences. Developing this further seemingly requires trudging through some of the very deep and murky waters in the philosophy of mind.

The metaphor of favouring an object with an eye toward something is borrowed from Rønnow-Rasmussen who was the first to use it in this context. When he challenges himself to elaborate on the metaphor he admits that “the prospects of specifying the characteristic features of this sort of attitude in a fully satisfactory way are bleak” (ibid: 56).²⁰ This seems right given the considerable risk of circularity inherent in finding such a specification. We might be tempted to say that when an object is favoured for someone, the object will be represented within the intentional content of the attitude as being good for this certain someone. Alternatively, when an object is favoured for someone, the properties of the object will be represented within the intentional content of the attitude as making the object good for this certain someone. This procedure can be repeated for the other attitudes as well. However, even philosophers without a general aversion to circularity are unlikely to find this to be fully satisfactory.

The other worry is that even if we understand the simpler attitudes invoked in the earlier section, there is no guarantee that the more complex and exotic ones will make any sense. Rønnow-Rasmussen (2011: 58-59) at one point considers the possibility that to favour an object for its own sake for a person’s sake can be analysed in the following way: The object is favoured for its own sake and the favouring of the object for its own sake is itself favoured for the person’s sake. This proposal is attractive because it entails that if we understand what it means to favour an object for its own sake and what it means to favour an object for a person’s sake, we will get an account of what it means to favour something for its own sake for a person’s sake for free. This is not such a complex or exotic attitude after all. Rønnow-Rasmussen (2011: 59-60) rejects the proposal because he insists on the possibility that when an object is favoured for its own sake and this favouring is favoured for the person’s sake, then it remains a meaningful question whether the latter is of the instrumental or the final form.

Perhaps the ideas presented in this paper can be defended from the worries just mentioned by appealing to our general ignorance about matters such as these. In other words, when it comes to offering philosophical accounts for the many attitudes that are relevant to the understanding of value, it seems fair to say that everybody hurts. Philosophers frequently speak in terms of our favouring objects in a way, for their own sakes, or for the sake of their effects. That we unfortunately lack a proper account for these attitudes have not stopped them from being theoretically useful, not only in the abstract realms of contemporary axiology, but in ethics more generally. Perhaps the same will be true of the more complex and exotic kinds of attitudes that have been developed in this paper, once they have been given their fair chance. Rønnow-Rasmussen has made the same point in his discussions:

The fact that they seem resistant to analysis might suggest that they are not genuine attitudes. Perhaps this is so. However, since analysis of these attitudes has been largely ignored by philosophers, to brush them aside at this stage would be premature. Of course, although they remain something of a puzzle, this does not mean that we cannot say anything about them (2011: 56).

Indeed, something informative has already been said: To claim that an object is favoured for its own sake and to claim that another object is favoured for us is to suggest that both objects are favoured, but in different ways. The intentional structures of the attitudes that are directed toward the objects in question are meant to differ from one another. When an object is favoured for its own sake for us, then the object is favoured with an attitude with a complex intentional structure that incorporates elements from both of the aforementioned attitudes. Exactly how all this is to be understood, and how we should cash out these claims about intentional structure, remains a problem. However, as things currently stand, it seems safe to suggest that this is a general problem in philosophy, rather than one which speaks specifically against the ideas defended in this paper.

Conclusions

The aim of this paper was to defend the view that the distinction between final and instrumental value cuts across the evaluative domain. To this end the paper focused on value in a way, understood in terms of the concept kind-value and personal value, respectively. The discussion was opened with intuitive support in the form of substantive examples. These examples were then strengthened by adding observations about the structure and behaviour of fitting attitudes. The argument here proceeded from the idea that we can favour things either as kinds of things or for someone’s sake, but in such a way that we also favour them for their own sakes or for the sake of their effects. Not only that, but the examples described in the early parts of the paper can be taken to show that doing so is sometimes perfectly fitting. It was argued that this places the burden of proof on potential sceptics, since they must insist that regardless of how we may actually or fittingly favour objects, objects still cannot have value in the way indicated.