2.1 Introduction
To the extent that we justify actions and attitudes in terms of some value they produce or express, we do not normally need to reach for fundamental, basic, or ultimate values. But certainly there are plenty of contexts where our attention focuses on a deeper dimension of value. For instance, many of us would take a critical attitude towards a businessman intent on accumulating money exclusively for the sake of it, quite regardless of how money impacts on other aspects of his life. The feeling is that money, however important it may be, is the sort of thing to be valued not for its own sake, but for the sake of something else; at any rate, it cannot be the only thing to be valued for its own sake.
Conversely, someone who appears to treat their own children exclusively as potential workforce for the family would be the target of blame: whether or not people can have some #instrumental_value, few of us believe that one’s own children can be regarded exclusively as potential workforce. In these and similar examples, our criticism is based on an intuitive distinction between what is valuable for its own sake (one’s own children) and what is instrumentally valuable (money).
The aim of this chapter is to make distinctions between value concepts which are often not distinguished, either within philosophy or in ordinary value talk: final value, exclusively final value, unconditional value, intrinsic value, necessary/essential value. The idea is to alert the reader to the importance of keeping these concepts separate.
I provide some defence for the idea that something can be valuable for its own sake (i.e. finally valuable, the sense in which one’s children are valuable and money arguably is not), and yet doesn’t have to be unconditionally valuable, nor must it owe its value to its intrinsic properties. This idea is worth exploring, since it has been assumed by most philosophers that what is valuable for its own sake can only be valuable in itself, that is, in isolation from other things, and not conditionally on other values.
#Final_value, according to this tradition, must be the stopping place of our evaluations, or is not really ‘final’. I will argue that this tradition, which has a distinguished history and has recently been revived, considerably restricts the ways in which it is appropriate to value things, and depends on a questionable assumption on what it means to be ‘final’.
I also explain why we should distinguish merely instrumental extrinsic value from other forms of extrinsic value.
2.2 Final and unconditional value: Some philosophical examples
How are we sure that Aristotle, Kant, Bentham, and Mill are talking about the same thing? First, despite different terminologies, eudaimonia, pleasure, good will, or happiness are all contrasted with things that are ‘good for the sake of something else’, that is to say, things that are good because they stand in a certain (typically but not necessarily) instrumental relation to the former items. For example, from a hedonist point of view, health is good insofar as it produces, facilitates, or simply contains experiences of pleasure, or prevents painful experiences. Devoid of such relationships to pain and pleasure, health would be neither good nor bad. Second, they all mean to pick out what can be described as final values.
Something is #finally_valuable when it is valuable for its own sake, that is, as suggested in Chapter 1, when we have reason or it is fitting to favour or value it for its own sake. The idea is fully normative: something that is finally valuable deserves, as such, a certain kind of treatment.
One is the concept of something that is not only finally valuable, but #exclusively_finally_valuable**: all the value it has or could ever have is final value. This is what after all for Aristotle sets eudaimonia apart from things like reason and virtue, which are finally valuable but not exclusively finally valuable, since they are valuable also for contributing to other things, such as happiness. and Kant agrees with this.
And the same goes for hedonists: of course a particular experience of pleasure could give us further future pleasures, and to that extent be instrumentally as well as finally good, but there is no non-pleasure-related state which might make an experience of pleasure additionally valuable in this way. **So, alongside contrasting final to non-final values, these philosophers meant to pick out exclusively vs. non-exclusively final values.
It is important to note, however, that the search for exclusively final value is not inevitable. Those who hold a pluralist axiology, whereby at least two things are finally valuable – e.g. pleasure and virtue – would have no problem admitting that virtue is both finally valuable, and also valuable when and because it produces pleasure, and vice versa for pleasure. One such example is W. D. Ross’s axiology, which includes innocent pleasure, knowledge, and virtue as final values (2002 [1930]).
Exclusively final value thus is a notion that only monist axiologies require: if there is no other final value than V, then there is nothing else for the sake of which V could be valuable. Of course, if you are a pluralist, you might still believe that each of your chosen two (or more) final values is exclusively finally good: e.g. that the value of virtue cannot be enhanced by pleasure or vice versa. But the monist needs to believe in a value that is exclusively final.
A second notion implicit in the quotations above is that those final values are also unconditional values, in the sense that their value does not presuppose the value of anything else. Again, if you are monist, e.g. you think that pleasure is the only final value around, then it follows that the value of pleasure cannot depend on the value of anything else. For an example of a conditional value, Kant famously mentions happiness: someone’s happiness is valuable only if they are worthy of it (due to their good will). In this sense, Kant denies that happiness as such is valuable for its own sake. It is a problematic question whether Kant thereby counts as a monist, since in some interpretations it appears that now two things are finally valuable: good will (unconditionally), and deserved happiness (conditionally on the value of good will)
For a different example, consider a valuable relationship such as friendship. Friends are usually disposed to help each other out, and they take pleasure or derive happiness from their relationship. Plausibly, if their happiness and altruistic dispositions had no value, their friendship would have no value either. In this sense, the value of friendship is not unconditional. But this does not mean that we are to value friendship simply for the sake of these other things (happiness, altruistic action)
Similarly, the value of altruistic action itself can plausibly be said to presuppose the goodness of happiness and the badness of suffering: if suffering were not bad, there would be no merit in alleviating others’ suffering. But this, again, does not mean that altruism is to be admired and encouraged simply for the sake of minimizing suffering. In sum, goods like friendship, altruism, or happiness (in Kant’s view) can be finally valuable and yet (unlike Aristotle’s eudaimonia or Kant’s good will) not unconditionally so.
But this is a dependence on factual, rather than evaluative, conditions: enabling conditions of this sort do not refer to the value of being mentally sophisticated or the value of there being people around. In this sense, an unconditional final value can and normally does depend on factual conditions:
nobody can have a good will if they are not sufficiently mentally sophisticated creatures. Being a mentally sophisticated creature is a necessary condition for someone to have a good will, and therefore for the value of the good will to be instantiated. Moreover, plausibly nobody can be a mentally sophisticated creature in isolation from others (say, their parents): there being other people around is another necessary condition for the value of good will. But this is a dependence on factual, rather than evaluative, conditions.
Any experience we may have depends on us being alive: dead people have no experiences. Therefore also the value of our experiences (e.g. our pleasures) depends on us being alive. But it does not follow that experiences are valuable for the sake of being alive. Rather, being alive appears valuable, if at all, for the sake of the experiences it makes possible. The value of being alive (as an enabling condition) is therefore non-final.
2.3 Intrinsic value and final value
The phrases ‘for its own sake’ and ‘in itself’ naturally suggest a further consideration: what makes these things finally valuable must somehow be found ‘in’ the valuable thing, rather than in something totally or partially ‘outside’ it. This is why such values have historically been called intrinsic, and therefore contrasted with extrinsic ones. G. E. Moore so defines intrinsic value: ‘To say that a kind of value is ‘intrinsic’ means merely that the question whether a thing possesses it, and in what degree it possesses it, depends solely on the intrinsic nature of the thing in question’ (1993 [1922]: 286). And in order to determine what has intrinsic value, he devised the so-called isolation test: ‘it is necessary to consider what things are such that, if they existed by themselves, in absolute isolation, we should yet judge their existence to be good’ (1993 [1903]: 236). Now, certain things just cannot exist in absolute isolation: e.g. a state of pleasure necessarily has a subject and (at least for some sorts of pleasures) an object (what it is pleasure at). So the test seems unserviceable as it stands: we cannot coherently imagine a world containing states of pleasures but no subject and no object of those pleasures. However, in the light of Moore’s definition of intrinsic value, we can propose an amended isolation test: in order to determine the intrinsic value of x, we have to focus on x’s intrinsic nature alone, and on that basis see if we judge x to be good – presumably, this is something we can do with pleasures and other states.
But is all final value intrinsic value? So did Moore assume. And until not long ago, such a question might have sounded otiose. If something is to be valued for its own sake, then it seemed obvious that what makes it worth valuing in such a way must be found in the intrinsic nature of the object. And if what makes something worth valuing is to be found partly ‘outside’ the object itself, then it seems that the value of the object cannot be final.
Equating final with intrinsic is not necessary
final/non-final refers to the correct or appropriate way of valuing something, while intrinsic/extrinsic refers to the metaphysical location of the good-making properties of something. So a Moorean view which equates final and intrinsic value need not be obviously true.
In particular, many believe that final values can be extrinsic: objects which are valuable for their own sake partly thanks to their relations to other objects.
For a more concrete example: a particular fur coat might be regarded as valuable for its own sake, as an outstanding piece of handicraft, yet so only assuming an appropriate evaluative background. If fur coats were not in general instrumentally valuable for protection against the cold they provide, this particular coat could not have any value, let alone any final value. (. In other words, intrinsic value need not be metaphysically mind-independent value 1983 as Korsgaard mentions)
Another case is where final value might be enhanced by an object’s relational properties: a fine work of art, beautiful and thus already valuable on account of its intrinsic features, might have its value increased by its being a unique or rare piece (a kind of relational property, since it implies the absence of other things, or many other things, like it). And such uniqueness or rarity need not make the object additionally valuable simply in instrumental terms, say because it increases its economic value, or because it provides particular pleasure to viewers or owners.
Yet another case is where the final value of an object does not just counterfactually depend on the value of something else, but results from or is exhaustively grounded on some relation to other objects. Uniqueness or rarity can give an otherwise unremarkable item (say, a stamp) a new final value. A certain type of car might have the ability to race at unusually high speed: this relational, indeed causal, feature may conceivably make the car valuable (e.g. worth maintaining) for its own sake. Were it not for this feature, the car would simply have the instrumental, non-final value that most other cars have. So in these cases, final value is extrinsic in two ways: first, it is grounded in relational properties of the object, and second, such relational properties make the object finally valuable only on the condition of something else than the object itself being valuable.
If these examples are persuasive, then final value can be extrinsic. It also follows that we cannot always apply Moore’s isolation test to discover what has final value. Remember that we had to imagine the object in isolation from any other thing. Then we would miss the final value of Napoleon’s hat, since we would ignore who the hat belonged to, and likewise the final value of a rare stamp, since we would ignore how many other stamps like it are around. We would miss the additional value of a unique artwork, since we wouldn’t know that there are no other exemplars of that. We would miss the final value of the fur coat, since we would have to ignore the point of producing and wearing fur coats. And we would possibly mistake the final value of any given friendship, since we wouldn’t know if it is actually good for the friends – we would ignore an external condition for its final value.
2.4 The reduction to facts
Given the plethora of examples available, it seems that the burden of proof lies on those who take final value to be always intrinsic value, i.e. value which depends only on an object’s intrinsic features. How could one hope to show that?
One strategy starts from the question: what are the real bearers of value, i.e. what sorts of entities really have value? Most of the examples assume that individual objects (coats, hats, stamps …) are the sort of entities that can have final value. The strategy invites us to reconsider this assumption.
It seems that we could, in principle, always understand or translate the putative extrinsic final value of an individual object in terms of the final value of a fact, or a state of affairs, which includes the object together with the relevant relational properties. For example, the final value of Napoleon’s hat can be seen as really the final value of the fact that (F1) there is a hat which belonged to Napoleon. Alternatively, it could be seen as the final value of the fact that (F2) this hat belonged to Napoleon. Either F1 or F2 would be both worth valuing for its own sake, and in virtue of their intrinsic or internal properties: that the hat belonged to Napoleon is indeed what these facts are all about, be it F1 or F2. So this translation strategy, or move from objects to facts, would give the result that a Moorean needs: a picture where final value is always intrinsic. Similar moves can be made, in principle, with other putative cases of extrinsic final value
It is interesting to note that the F2-style translation, which avoids the existential commitment, is perhaps preferable to F1 for two reasons. First, it is not obvious that to value something for its own sake always means to value its existence. This might work for artefacts which are worth preserving (and thus keeping into existence), but what about wishing a painless death to our sick dog as a way of valuing him for his own sake? Second, precisely this point shows that in general we value the existence of x only if we value x in the first place. F1-style translation instead sees valuing x’s existence as a prior condition on valuing x, so it seems to get things the wrong way around.
Why can’t we be happy with final values that are extrinsic? One immediate reason might be simplicity. Extrinsic final values give us a less unified picture of how value comes into existence. And, as we have seen, we wouldn’t be able to use Moore’s isolation test, which might have given us a handy procedure to look for the things to care about for their own sake. A more theoretical reason, suggested by Michael Zimmerman (2005: 194), is that if there are long stories to be told about, e.g. why Napoleon’s hat is a valuable thing, i.e. stories that must refer to values other than the value of the hat itself, then what we have is at most a derivative value. But final values, in principle, should rather play the role of ‘endpoint values’: once we reach them, no helpful explanation of them can be given, except saying that they are good ‘as such’ – i.e. ‘for their own sake’. That’s why extrinsic values of the sort indicated above cannot be final values for Zimmerman: explanation (or rather justification) does not stop at them. Only values that are intrinsic can play such an endpoint role, because there’s no looking beyond them to explain why they are valuable. And it seems that only facts, and not individual objects, can be guaranteed to play this double role of carrying final, i.e. non-derivative, and intrinsic value.
We obviously do not only evaluate objects or persons, but also facts or situations: we say that it is good that this and that happened, or that it would be better if that did not happen, and so on. And it also matches with the idea that many of our fitting responses to value are what are generally called propositional attitudes, such as desire or preference, e.g. desiring that the war end.
If, as claimed in Chapter 1, values must be the sort of things towards which fitting responses can be directed, then it is easy to see why values at least could consist in proposition-like entities, such as facts or states of affairs.
However, the reduction manoeuvre is not easily accomplished. I will mention two challenges.
First, since the strategy has to be carried out for the whole spectrum of final values, its proponents are committed to the claim that only propositional attitudes are fitting responses to final value. If final values are all facts such as F2, then it seems that attitudes, such as respect, preservation, or certain forms of love, which are thing- or person-oriented, rather than fact-oriented, cannot be appropriate responses to final values. We respect and love individual people (or animals), and this is a different attitude than loving that such people have this or that characteristic. We want to preserve Napoleon’s hat, rather than the fact that it belonged to Napoleon – indeed, preserving the latter seems to make almost no sense. According to the reduction manoeuvre, thing- or person-oriented attitudes might still be in some sense fitting to their objects, but they won’t match with final value: final value will rather belong to the sorts of states which we can value by a relevant propositional attitude. So the fittingness of preserving Napoleon’s hat will reduce, it seems, to the fittingness of valuing or cherishing F2, the fact that the hat belonged to Napoleon. While this latter example may give some plausibility to the strategy (after all preserving the hat is a way of valuing F2), it is not equally clear that, e.g. respect and love for individual persons can always be seen as fitting because of the fittingness of a kind of respect and love towards some fact about these persons. In a sentence, the reduction of finally valuable things to finally valuable facts commits one to a reduction of fitting responses (to final value) to fact-oriented attitudes, and the latter reduction might be hard to sustain, given the variety in the kind of responses that we normally identify as ‘valuing for its own sake’. Perhaps then bearers of final value are just as various as our responses: ‘We value many different kinds of things, including at least the following: objects and their properties (such as beauty), persons, skills and talents, states of character, actions, accomplishments, activities and pursuits, relationships, and ideals’ (Scanlon 1998: 95)
Second, the theoretical justification suggested by Zimmerman for regarding all final value as intrinsic is weak. There is no initial reason to think that final values, understood as things worth valuing for their own sake, also need to play the role of non-derivative or endpoint values, in the sense that no explanation of their value can be given, except pointing to their own intrinsic nature. Conceptually, we are again dealing with different notions: final value, which refers to the appropriate way of responding to something, and non-derivative value, a structural-metaphysical notion. Consider again friendship. The value of friendship, and thus the respect a particular friendship deserves from third parties, can be partly explained by reference to other values: the mutual desire for happiness that friends typically have is good; the happiness thus achieved is also good; and, at least on some views, friends must make each other happy by and large in morally permissible ways. The value of friendship then is not an ‘endpoint’ value, because appreciating the value of any given friendship is not independent of appreciating the value of other things (happiness, morality). But valuing, e.g. respecting, a particular friendship on the condition that it doesn’t involve immoral conduct, obviously does not mean valuing it for the sake of something else, namely moral permissibility. One can recognize a value as conditional on other values, and in this sense ‘derivative’, and still fittingly care about it for its own sake. What does seem to be true is that, if a final value is derivative or conditional on other values, then this reflects on the appropriateness of responses. If I care about Napoleon’s hat for its own sake, but show otherwise utter indifference towards Napoleon’s sword, or indeed to objects belonging to comparable historical figures (say, Sitting Bull’s war bonnet), then my valuing is not of the fitting kind, precisely because it is insensitive to the structure of the final value towards which it is directed (assuming, of course, that the hat is only valuable for its historical connection). But such interrelation among fitting attitudes does not subtract anything from the final value of the object – it is not as if its value gets ‘thinned out’ in the network of attitudes required towards similar objects.
essential values, i.e. values which depend on the essential properties of the valuable object. It is an essential property of F2 that it concerns Napoleon’s hat and its having belonged to Napoleon: a different hat or a different owner would make up a different fact. In general, a fact like ‘that x is P’ has the essential property of concerning x, or ascribing P to x. If a fact is valuable because of such intrinsic property, then it is also essentially valuable. In this sense, final values turn out to be incorruptible, that is, they remain constant for as long as their bearer remains what it is. It is a good question, to say the least, whether it is a defensible consequence of such a view that only unconditional and essential values are really worth valuing for their own sake, whereas values that are had conditionally or contingently do not deserve that same kind of response.
2.5 Intrinsic and conditional value
Having defended the possibility of final extrinsic values, we can now also clarify the relations between intrinsic/extrinsic and unconditional/conditional value.
Unconditionally good things do not depend on the value of anything else. Therefore whatever it is that makes them good must be looked for among their intrinsic properties: in their intrinsic nature. As said, even unconditional values might depend on factual enabling conditions: the unconditional value of pleasure depends on the subject being alive. Now, such enabling conditions can be intrinsic or extrinsic properties of the valuable object. i.e there being other people around might be a factual extrinsic condition for the value of a good friend.
But this doesn’t mean that therefore the good will has an unconditional extrinsic value. This is because Moore’s definition of intrinsic value and the amended isolation test for it presuppose that whatever necessary conditions for x to exist. and be intelligible as such, be they intrinsic or extrinsic properties of x, have already been counted in. So, for instance, it would make no sense to apply Moore’s test to the good will, and conclude that the good will has no intrinsic value because one instance of the good will in a world without other agents would strike us as valueless. Rather, when we apply the test, we should imagine a world where necessary conditions for a good will to exist and be intelligible as such are taken for granted (such as the existence of other people besides the agent). Factual dependence on external conditions doesn’t make a value extrinsic.
What about conditional value? Here a certain object x is dependent for its value on the value ‘of something else’. Now, this ‘something else’ can be intrinsically or extrinsically related to x. Earlier I considered the value of a friendship, and suggested that it may depend not only on whether it produces happiness for both friends, but on whether such happiness is indeed good for them or more generally valuable. Happiness, and its value, are in this way extrinsically related to any given friendship: when we apply Moore’s test on friendship, its producing valuable happiness or not is one of those facts we should abstract from, because it is not a necessary condition for any friendship to exist – friendship can produce misery as well as happiness. So the value of friendship appears to be conditional and extrinsic (but not for these reasons non-final, as argued above).
On the other hand, if the ‘something else’ is intrinsically related to x, then x has conditional but intrinsic value. A good example is offered by Thomas Hurka’s theory of virtues, in which virtues are attitudes of this form: ‘loving (desiring, etc.) the good for its own sake’ and ‘hating (avoiding, etc.) the bad for its own sake’ (Hurka 2001). Virtues are therefore second-order final values. Their structure makes their value at once conditional, because it presupposes the first-order value of the objects of love and hate, but also intrinsic, since it is an intrinsic feature of a virtue that it involves an attitude towards the good or the bad.
The crucial point from this discussion is that intrinsic value can be conditional value. This is significant, since it is easy to confuse the two notions: it is natural to wonder, how can something’s value be intrinsic, belong to ‘the thing itself’, if it depends on the value of something else? But just like we shouldn’t assume final value to be necessarily intrinsic, so should we not assume that intrinsic value is always unconditional.
2.6 Elimination of extrinsic value?
There is a lingering concern with extrinsic values. If x owes its value to the value of something else y, then the worry is that x really has no value at all. Some philosophers are tempted to draw this conclusion when considering extrinsic value of the instrumental sort. Thus Ross on the value of acts: ‘Whatever value [an act] has independently of its motive is instrumental value, i.e. not goodness at all, but the property of producing something that is good’ (2002 [1930]: 133, my emphasis). There is a sense in which the act does not contribute any value to the world. In computing how much value the world contains, we are not going to add instrumental value on top of whatever value the act has caused. A world where the same valuable states of affairs occur through other means would contain the same amount of value (other things being equal). Hence the act is good-causing, but not literally good.
But could we draw eliminativist conclusions for all extrinsic value? For instance, some philosophers speak of ‘signatory value’: an X-ray is signatorily good if it indicates something else that is good (e.g. that the tumour has gone). (X-rays of course are in general also instrumentally good, as aids to medical knowledge.) Or there can be ‘contributory value’: a certain motif in a painting is good in this sense if it contributes to the (aesthetic) value of the painting as a whole. It is good ‘as a part’. It would be tedious to recount all forms of extrinsic value, because it would require drawing a list of all relevant relations: causing, being a sign of, being a part of a whole, being historically connected to (as in Napoleon’s hat), etc. But the question now is: if we grant that one sort of relation (instrumental) to value means that an object (an act, say) really is valueless, why not generalize and conclude that the concept of extrinsic value is, at best, a handy way of talking about value, but does not capture a genuine evaluative reality? And if extrinsic value doesn’t capture a genuine evaluative reality, then a fortiori all final value must be intrinsic.
not all extrinsic value can be eliminated like that. The argument for elimination assumes that what has instrumental value does not contribute any value over and above the value of its causal consequences. However, this does not seem to apply to the cases of extrinsic final values considered above. First, it is true that Napoleon’s hat has no value apart from its historical connection. But Napoleon’s hat contributes value precisely in being an exemplar of a supposedly valuable category of things – objects (maybe of a certain kind) that belonged to important historical figures. Destroy the hat, and you have directly reduced the amount of value contributed by this category. On the other hand, once you imagine away acts and in general things that only have instrumental value – while keeping constant the amount of valuable states of affairs they would otherwise produce – you will not have diminished in the slightest the amount of value in that world.
Second, regarding values such as friendship which are conditional on external factors, the elimination argument would work if it were true that the fact of friendship would contribute no value, over and above the value contributed by other facts which are conditions for friendship’s value (e.g. mutual happiness produced within a friendship). But it is hard to see how to show this much. While it might be held that a given friendship loses in positive value if it makes one or both friends worse off, what loses in value is the complex of expectations, mutual feelings, shared history, etc., which define friendship for what it is, and which – prior to philosophical arguments to the contrary – constitute its final value and its specific valuable contribution ‘to the amount of value in the world’. Of course some philosophers (e.g. hedonists) would be ready to make the substantive claim that friendship is only instrumentally valuable, e.g. insofar as it promotes the general happiness. But this move would not be acceptable, since here we were looking at possible reasons why extrinsic values are eliminable (and a fortiori non-final) qua extrinsic. It seems that no general argument for elimination can be found.
In this chapter I have distinguished several pairs of value concepts:
- final vs. non-final value: what is fitting to favour for its own sake vs. what is fitting to favour for the sake of something else;
- exclusively final vs. non-exclusively final value: what is fitting to favour only for its own sake vs. what is fitting to favour for its own sake and for the sake of something else;
- unconditional vs. conditional value: what is fitting to favour (for its own sake) independently of whether it is fitting to favour something else vs. what is fitting to favour (for its own sake or not) not independently of whether it is fitting to favour something else;
- intrinsic vs. extrinsic value: what is fitting to favour because of its intrinsic properties vs. what is fitting to favour (for its own sake or not) partly because of its extrinsic properties;
- essential/necessary vs. contingent value: what is fitting to favour in all possible worlds where it occurs vs. what is fitting to favour in some but not all possible worlds where it occurs.
In particular, I have suggested that final value need not be unconditional value, and criticized a strategy to equate intrinsic and final value based on the idea that only facts or states of affairs are bearers of final value. Since the resulting view is that there can be final extrinsic values, then I had to defuse a worry that the very category of extrinsic value could be dispensed with.