Introduction to #value_theory

We use value talk all the time. Most of the time, we don’t need to mention the word ‘value’ or similar terms like ‘worth’ or ‘merit’. More commonly we use evaluative expressions like ‘good’, ‘bad’, ‘better’, ‘the best’, ‘great’, ‘fine’, ‘excellent’, ‘poor’, ‘terrible’ and so on. To utter such evaluative words of course is not always to evaluate something positively or negatively.

The exclamation ‘Good heavens!’ normally doesn’t say anything about the value of the heavens. Also, we can use ‘good!’ or ‘that’s good’ simply to express or report our satisfaction with something, regardless of whether we think it worthy of our satisfaction. At other times still, evaluative words are used to register other people’s evaluations, without joining in ourselves: if I know nothing about row-crop tractors, except the ranking that an expert commission gave of models produced in 2013, I might call one 2013 model better than another simply to report the commission’s ranking.

given its primary evaluative association, such a judgement might be misleading, since you might well infer that I do believe one model to be better than another (and thus, perhaps, that I am advising you to get that), whereas all I mean is really that the commission decided so.

It is also important to keep in mind that evaluation need not be articulated via value terms, or indeed be verbally articulated at all. Often we can tell a person’s values and commitments from her behaviour and emotional reactions much more than from what she says (to others or even to herself). We can understand a given community’s unspoken and unwritten values or norms, or at least some of them, by interpreting their social practices, rituals, internal conflicts, and so on.

However, in this book I will explore conceptual questions about value, rather than about evaluation as a social and psychological phenomenon. . Therefore evaluations which are verbally expressed or at least expressible in value terms are the starting point.

First, I draw a distinction between different evaluations, in order to focus on the value concepts that will be discussed. Then I define and explain the project of value theory, as I pursue it here.

In section 1.4 I lay out the central assumption or guiding thread: value is normative. two views intended to articulate this notion:

1. the #fitting_attitude_account
2. the #buck-passing_account

1.2 Which evaluations?

Evaluation comes in so many linguistic forms and spans so many different categories. But there is a conceptual distinction that can help to isolate the sort of evaluations that will keep us occupied in what follows.

1. #thin_evaluation #think_concept: Some evaluations are #pure, or purer than others, i.e. they carry minimal or no descriptive or informative content about the thing evaluated. Following Bernard Williams (1985), these can be called thin evaluations. When a certain concept is used in such an evaluation, it is a thin evaluative concept.

2. #thick_evaluation #thck_evaluation: which do carry more than minimal information about Value and Normativity.

not easy to find this theoretical distinctions simply linguistically, but few remarks may help to grasp the point:

• thin evaluation can be seen in unqualified uses of words such as ‘good’ and ‘bad’ and their comparatives . These can be applied to concrete objects or states of affairs: ‘this movie is bad’, ‘following her advice was good’. In philosophy, thin general evaluative statements are often made about abstract objects: ‘pleasure is good’, ‘knowledge is generally better than ignorance’. But one can also use abstract nouns like ‘value’ (or ‘good’ as a noun): ‘friendship is/has the greatest value’,

• Now, these terms can be qualified by adverbs or adverbial phrases like ‘intrinsically’, ‘in itself’, ‘instrumentally’, or by the respective adjectives. Depending on the adverb, the resulting evaluation will be more or less thick , that is, will carry more or less descriptive or informative content about the thing evaluated. Sometimes it won’t carry any additional descriptive content, but simply stress the importance of a value: ‘helping others in need has incalculable value’. In other cases it will carry some minimal content: to say that knowledge is instrumentally good hints at the effects or consequences of knowledge, albeit not saying what these are, except that they are good.

• **in other cases, qualifications can carry quite a lot of informative content : ‘This photograph has personal value (for me)’, ‘Wedding rings have sentimental value (for many people)’, ‘This song has no artistic value’, These qualifications assign the thing evaluated to some category or other of value, and so imply some substantial information about them, or about their relations to other things or people. Such information need not be very detailed, of course. To speak of personal or sentimental value implies something about the historical or affective connection of an object to people.

**- To speak of artistic or scientific value turns one’s attention to the sort of features which justify the inclusion of an object in either category. In these cases, we are making thicker evaluations.

evaluative words can be used predicatively, as in the sentences used so far, which have the general form:

• x is good.

But very commonly they are used attributively, i.e. they accompany a noun: ‘Sarah is a great dancer’. In general:

• x is a good F Often, these attributive evaluations are quite thick

it is not at all obvious what information is carried by the statement that WWII was the worst war ever, because it is not at all obvious what a war has to be like, as a war, to be worse than another. All we definitely know, of course, is that WWII was a war, and so that there was fighting and the like (for another grey-area example: ‘Valerie is a very good child’). Attributive evaluations then are not guaranteed, just in virtue of their form, to be particularly thick ones.

There are two sorts of concepts which are often said to belong by their own nature to the thicker end of the thin-thick continuum:

1. #value-maker_concepts
2. #-able_concepts

Example of (1):concepts like courageous, honest, cowardly, corrupt, elegant, tacky, melodious, insightful and so on. Employing these concepts, at least normally, means both evaluating and describing the thing or person one way or another. For a woman to be courageous, she has to meet certain descriptive conditions, like being able to overcome fear in certain kinds of situations.1 While it is controversial how to best understand these terms, this much can be said: they bear an intimate relation to some relevant thin evaluative concept.

• A courageous woman has something good about her, i.e. merits a positive evaluation of some sort.

not only do they serve to evaluate, but, unlike thin concepts, and unlike other thick evaluations, they normally also present a reason for the evaluation: by saying the she is courageous, we are also saying that she is good (if she is) because she is courageous. Being courageous makes her good.

However, as before, some of these terms need not be very thick: for instance, ‘virtuous’ (as in ‘possessing virtue’, rather than ‘chaste’) does convey the idea that the person is morally good because of certain personal qualities, but nothing more than that, partly because what makes a person virtuous is a much more contested issue than what makes her courageous or cruel.

Example of (2) : #-able_concepts such as valuable, desirable, admirable, enviable, contemptible, etc. These concepts wear on their sleeves, so to speak, the idea that the thing so evaluated merits or is worth a certain attitude or response: valuing, desiring, admiring, etc. Obviously the ‘-able’ suffix need not appear: the concepts fearsome, trustworthy, amusing, shameful have exactly this same structure. (**they are response worthy)

The thickness of these concepts lies not in any particular descriptive condition the thing so evaluated has to meet, but rather in the fact that they describe a specific response as merited. By contrast, ‘good’, ‘bad’ and the like by themselves tell us nothing about merited responses – or at least, nothing as specific. The idea of merited, fitting, or appropriate response, as I explain in a moment, plays a central role in what follows.

Evaluations on the thinner end of the spectrum provide the background for most of the subject matter of this book

value as such: how is the value of something related to what we ought or have reason to do (in a broad sense)? How does value ‘depend’ on non-evaluative features? How do we work out the value of complex states of affairs? Must there be a common measure to compare different values?

thicker evaluation will be of more direct concern: Chapter 3, where I consider (and reject) arguments to the effect that thin predicative evaluation (‘x is good, period’) is based on a mistake, and Chapter 4, where I take some steps towards understanding what surely is a thicker type of evaluation, namely that involving what is ‘good for me’

1.3 The idea of value theory

The name ‘value theory’ doesn’t refer univocally. There are at least three different sorts of philosophical inquiries moral philosophy, political philosophy, aesthetics, and possibly epistemology, regarded as a normative discipline. In this sense, it contrasts with all other ‘theoretical’ branches of philosophy, like metaphysics or philosophy of language.

A second, much more restricted use, stands for all substantive views about what is fundamentally good and bad, and the debate among them. For instance, hedonism is a value theory, holding that pleasure is the one fundamental positive value. As such, it is opposed to other theories, e.g. perfectionism, holding that the excellent development of certain abilities is the fundamental positive value, or to some pluralist theory, holding that there is more than one fundamental value, for example pleasure, knowledge, and moral virtue.

despite the names, normative theory and value theory are both normative inquiries, in the broad sense that to adopt a certain normative or value theory is to come to believe that one ought or at least has reason to behave (act, desire, feel, judge, etc.) in certain ways for certain fundamental reasons. (For example if hedonism is true, then in principle one should take pleasure, one’s own or others’, as the fundamental goal.) While certain meta-ethical views may ultimately have normative implications for conduct, the guiding aim of meta-ethics is not evaluative, prescriptive, or directive, but rather explanatory, reconstructive, or descriptive.

I will not engage in axiology (value theory in the second sense), since I will not argue for or presuppose any particular view about what is fundamentally valuable. Only for the sake of argument will I assume in some places that a certain axiological view (e.g. hedonism) or a certain evaluative intuition (e.g. about the value of punishment) is correct.

Value theory, in this third sense, purports to explore a host of structural questions about value, in the hope of understanding at least some salient properties that things have in virtue of being good, bad, and so on

When we speak loosely of ‘value(s)’ as a countable term, we may refer to three different things:

**(i) the object or state of affairs that has value,

**(ii) the features that make an object or state of affairs valuable,

**(iii) the fact that an object or state of affairs has value (is good, bad, etc.).

An axiological theory (a ‘value theory’ in the second sense above) is typically a theory about the features that ultimately make anything valuable, that is, about (ii). As a result of the ‘value-making job’, we can then speak of values in senses (i) and (iii): there will be objects and states of affairs that are made valuable, or are bearers or carriers of value, and there will be facts or truths about such things being good or bad. In other words, the idea or fact of value-making is logically prior to the existence of valuable objects and of evaluative truths. But since here I am not doing value theory in the sense of providing an axiology, when I loosely speak about e.g. the value of friendship or of a wedding ring, I mean to refer to (iii): the fact or truth that friendship or the wedding ring is valuable, rather than friendship as a value-making feature, or friendship as a bearer of value (although it will of course be true that friendship is a bearer of value).

1.4 Value and normativity

Evaluative concepts in a very broad sense include not only good, bad, etc., but also ought, right, wrong, duty, obligation, practical or epistemic rationality, and good reasons for actions, attitudes, or beliefs. in this book : the sense at issue when we judge things to be good, bad, and the like, with or without qualifications.

‘Normative’, however, should not be thought to refer to anything as norm- or rule-based as a quasi-legal normative system. In this broad sense, commonly used in philosophy, normative truths can bear the form of a rule as well as that of particular statements about what you should do here and now. Nor are actions the only normatively relevant objects: intentions, desires, emotions, beliefs are all just as relevant, insofar as it makes sense to hold that there is or there could be good reasons for and against them. T. M. Scanlon labels all these responses as ‘judgement-sensitive attitudes’, in the sense that ‘an ideally rational person would come to have [them] whenever that person judged there to be sufficient reasons for them and [they] would, in an ideally rational person, ‘extinguish’ when that person judged them not to be supported by reasons of the appropriate kind’ (1998: 20).5 Distinct as they may be, still the evaluative and the normative bear an intimate relation: value is normative. This can be said to be one of value’s fundamental properties.

**truths about value, at least, regularly entail normative truths of some sort about actions or attitudes. Here are some examples: that a certain hotel is overall better than another seems to entail that you should prefer it to the other; that a certain conduct was morally good seems to entail that there is some reason to admire it; that a movie is terrible seems to entail reasons to avoid going to see it, and so on. What is relevant is that the connection between value and reasons is entirely general: it applies to values or value claims and reasons as such. It is not because of what better hotels are like that they should be preferred, but simply because of what it is for something to be better than something else. Of course, the specific attitude or action that is appropriate to take will depend on the valuable object at stake: for instance, a great painting should be appreciated in a certain aesthetic way, and a poor one aesthetically censured, while morally outstanding conduct is to be praised, and evil conduct blamed or condemned. But what remains constant across these diverse cases is that a positive value is to be responded with a positive attitude, and a negative value with a negative attitude of some sort.

The most straightforward way to articulate the idea that value is normative is to say that what it is for something to be good is nothing else than for there to be reasons, in principle for anyone, to respond favourably to it, and likewise for what it is to be bad, better, etc. Goodness, or being good, simply consists in the existence of such reasons.

Evaluative truths are really normative truths in a shortened form. In the philosophical literature, there are two existing accounts of this form: #fitting_attitude_account

• FA: x is good = it is fitting to respond favourably to (‘favour’) x. starting with the German philosopher Franz Brentano.6 The crucial term is of course ‘fitting’. This is meant to be a fully normative term, expressing the idea of a normative or ideal match between the object x and the favourable response. But other terms have been used in this connection: appropriate, suitable, correct, worthy, deserved, merited, required, ought (x is good = x ought to be favoured). While these terms have different connotations, they are all supposed to be general normative terms. In particular, they should not be assumed to have a moral meaning.

#buck-passing_account (called by Scanlon)

• Buck-passing Account 1: x is good = x has the property of having other properties that provide reasons to favour x.

This differs from FA in two respects:

1. it uses the normative concept of a reason to favour.
2. makes clear a certain structural claim about value: x’s being good (and other value properties on the thinner end of the spectrum7 ) is not itself a reason to favour x, but rather is the fact that x has other features, i.e. other than goodness, which provide reasons to favour x.

2 means that: To talk of x’s goodness thus is not already to indicate one reason to favour x, but rather to announce that there are reasons to favour x. In this sense goodness ‘passes the normative buck’, i.e. the ability to provide reasons for attitudes, down to the features which make x good, rather than keeping the normative buck, i.e. providing a reason itself

x has certain properties (e.g. is pleasant) → (there are reasons to favour x = x is good)

• The arrow → signifies that being pleasant provides or grounds a reason to favour x, and correspondingly grounds x’s goodness.

[T]he reason why it is proper to admire anything must be constituted by the qualities which make the object of admiration good … The ground [for an attitude] lies not in some other ethical concept, goodness, but in the concrete, factual characteristics of what we pronounce good. Certain characteristics are such that the fitting response to what possesses them is a pro attitude, and that is all there is to it

FA articulates the normativity of goodness and other thin concepts (and obviously of what I called ‘-able’ evaluative concepts). But is it applicable to what I called thick ‘value-maker concepts’? For example, courage makes a person good or admirable, but the property of being courageous may not be the property of having other properties which give reason to admire the courageous person. Courage doesn’t have to pass the buck to courage-making features: it is itself a perfectly good reason to admire somebody. The normativity of these thick evaluative concepts seems to lie in their immediate reason-providing (or good-making) ability, not in their being analysable in FA terms.

Both FA and the buck-passing account then make two central claims:

1 Reduction Claim: value properties such as goodness are reduced to normative relations of fittingness of, or of there being a reason for, attitudes;

2 Normative Redundancy Claim: value properties such as goodness do not themselves provide reasons for attitudes (make attitudes fitting) over and above the good-making features which already provide reasons for attitudes (make attitudes fitting).

The two claims are related.

• If the Reduction Claim is true, then for x to be good is for x to have properties that provide reasons to favour x. Consequently, if x’s goodness were a reason to favour x, such reason would consist in the fact that x has properties that provide reasons to favour x.... but:

‘Why should I favour x?’ The answer ‘because x has properties that provide reasons to favour x’ is very unsatisfying. What we want to know is what reasons there are for favouring x, and all we are told is ‘the fact that there are reasons’. In a reason-giving exchange, at least, mentioning goodness is little or no use. Second, the complex fact about reasons in which goodness consists obviously depends on other properties already providing reasons to favour x: for instance, that x is pleasant. To hold that goodness provides a reason of its own over and above these other features, or on the same standing as them, is to ignore this fundamental dependence.

Imagine again a reason-giving exchange: ‘Why should I favour x?’ ‘Because x is pleasant, and x is good.’ X’s goodness would not add any normative weight to the case for favouring x. The case for favouring x has already been made in order for x to be good. In this sense, if goodness is reduced to a fact about reasons (or about fitting attitudes), goodness would be redundant as a provider of reasons.

On the FA/buck-passing picture, then, value is normative because it just is a normative relation, and not because it is a provider of good reasons or something that itself makes an attitude fitting.

---

Is there any significant difference between the concepts of fittingness and reasons? Should one choose FA or the buck-passing account? Arguments can be made on either side.

in favor of FA

**first

-to have a reason to do something, one must be able to do it (reason, like ought, implies can), whereas this doesn’t hold for fittingness, since it expresses a more ideal sort of normativity.

we can clearly imagine things being good or bad in a world where there are no agents capable of responding as they have reason to: e.g. the plight of dinosaurs’ extinction was arguably something bad, if non-human pain is bad at all, but nobody could do anything about it, and therefore nobody had reason to do anything about it. This doesn’t change the fact that the plight of dinosaurs merited, even back then, a negative response.

However, it is still true that, in such situations, if there were agents capable of responding, they would have reason to do something about it. So one might simply qualify Scanlon’s account accordingly: Buck-passing 2: x is good = x has the property of having other properties that provide reasons to favour x, or would provide such reasons to suitably situated agents

**second

Another possible reason in favour of fittingness comes from what is known as the wrong kind of reasons problem (WKR).

There can be all sorts of reasons for positive attitudes towards things, people, and situations, which are not good. A commonly used example asks us to imagine that a demon will torture us or the whole humankind if we fail to admire him. It seems that, based on this threat, we have abundant reasons to admire him. If the buckpassing account is true, then it follows that the demon is good. But clearly the demon is evil rather than good, in part precisely because of the threat. His threat is a wrong kind of reason to admire him, because it doesn’t make him good or admirable. Such a reason to admire him does nothing to show the demon’s positive value. So, one reaction to this example might be that, whatever contingent reasons there might be to admire the demon, that doesn’t change the fact that it is unfitting to admire him: a being like the demon simply doesn’t deserve admiration. Fittingness truths might seem to have the sort of fixedness that goes well with truths about value. But buck-passers can again modify their formulation, so as to pick out the right sort of reasons as those that stem from fittingness:

**Buck-passing 3: x is good = x has the property of having other properties that provide reasons of fittingness to favour x, or would provide such reasons to suitably situated agents.

Response to FA : On the other hand, advocates of a buck-passing account might complain that the notion of fittingness really belongs to the evaluative rather than the normative realm. This might be for two reasons:

First, the idea of an ideal match applies also outside of the domain of human responses: a key fits a keyhole, a chord fits a certain melody, a certain trait makes a species fitter for a certain environment, etc. These statements express evaluations of a functional, aesthetical, or biological kind, without directly addressing what anyone should do. Second, there is a similarity between reasons and fittingness that is worth bringing out: certain features of x provide reasons for an attitude towards x or an action, just like an attitude towards x or an action fits certain features of x. In other words, both notions relate objects and their properties with responses. But reasons seem to essentially relate to the agent or the subject of those responses in a way that fittingness need not. Reasons are always reasons for someone to do something, or they are not reasons at all, while a certain attitude might be called fitting prior to being fitting for someone to take. Of course, when appropriate, we can always add some ‘for A’ qualification, but so can we add like qualifications to claims about value: e.g. it is (would be) good of you to act generously. But intuitively we don’t think that goodness requires mention of an agent, even when it is the goodness of an action. So, since normativity has to do with claims addressed to agents, and reasons, but not fittingness, essentially apply to agents, reasons may seem better equipped to express the normativity of value.

However, these remarks are insufficient to make a case in favour of reasons over fittingness. First, the fact that the word ‘fitting’ and related words have other, non-normative uses shouldn’t give us pause, because the same applies to the word ‘reason’. Moreover, the same concept can be expressed by other words like ‘deserved’ or ‘merited’, whose meanings are unmistakeably normative.

Regarding the second remark, there can indeed be an intuitive difference between reason (essentially ‘agential’, belonging to agents) and fittingness. But, for one, it shouldn’t be overstated. The ‘agent’ argument in a reason-predicate often has to be left unfulfilled – whose are the counterfactual reasons to prevent the plight of dinosaurs’ extinction? Do we learn anything normatively significant by assigning those reasons to potential agents? Secondly, the fact, if it is such, that fittingness, like goodness, is not essentially agential might cut both ways. If reasons are relations to agents, and value is not, then a definition of value in terms of reasons looks less promising than one in terms of fittingness, because it would metaphysically require ‘more’ from value than we might have thought: the presence, albeit possibly only counterfactual, of agents. Perhaps we can accept that there is a level of normativity where the specification of an agent or a subject is an inherent possibility (after all, fitting responses will be someone’s responses), but no agent need feature as a separate term of the normative relationship.

I will mostly talk in terms of the fitting attitude account of value, or more precisely, I will speak as if FA provides the true account of the normativity of value. The question whether it is the right account will be directly addressed in the final chapter, where I will consider some objections against FA and possible alternatives to it

Assumptions?

doesn’t FA assume that (1) value judgement is cognitive rather than non-cognitive – a belief about an evaluative matter of fact, rather than a desire or emotion; (2) there 18 VALUE THEORY are evaluative truths; (3) value properties are non-natural properties, since fittingness and the like are non-natural just as goodness and the like? Let’s consider these seeming assumptions in turn.

1- FA does not assume cognitivism, for two reasons. First, FA can be put forward as a general conceptual thesis about the content of value judgements: when we judge that something is good, we also judge (or are committed to judging) that it is fitting to favour it. Scanlon himself sometimes speaks in these terms, and it is notable that Allan Gibbard, a prominent non-cognitivist, accepts accounting for notions such as goodness or wrongness in terms of the rationality (or their being warranted) of certain feelings and attitudes, like approval, admiration, remorse, indignation, etc. (Gibbard 1990). It is then a further question whether these judgements of rationality (or fittingness) are to be understood in non-cognitive terms, e.g. as expressing a non-cognitive state of acceptance of a system of norms that permit certain actions and attitudes (as in Gibbard’s analysis), or as expressing beliefs about matters of fact concerning fitting attitudes (as on a cognitivist approach).

The second reason why FA does not assume cognitivism is that, even if FA were an account of value properties, and thus in turn implicitly suggested a cognitive account of evaluative judgements as beliefs purporting to represent those properties, still non-cognitivism has over the years developed into what Simon Blackburn calls quasi-realism (Blackburn 1984). Quasi-realism is the project of reconstructing (or ‘earning the right to’) such notions as normative truth, properties, belief, and knowledge, working with the scant materials of a non-cognitivist approach and a generally empiricist worldview. Accepting a fitting attitude account of value properties doesn’t seem to make such project any harder than accepting a different account of value. If value is normative, a quasi-realist analysis will deploy whatever resources it already has to reconstruct such a basic feature of value

2- John Mackie’s error theory holds that affirmative propositions about value, when understood as implying objective and categorical demands or reasons, are all false. To speak of objective value, or alternatively of attitudes that are fitting regardless of what anyone’s actual attitudes are like, is to commit oneself to a metaphysically and epistemologically queer view (Mackie 1977). However, FA doesn’t assume the truth of any evaluative proposition. It only means to articulate what evaluative properties would have to be like, if there were any, and so, if any proposition about them were true. 3. Another remark worth making in connection with Mackie relates to the third question about the natural/non-natural distinction. Mackie takes common sense to be committed to the existence of value or fittingness conceived as entities constituting non-natural facts: facts which lie outside the world described by the natural sciences, including psychology, and which therefore require some non-empirical (and, for Mackie, queer) means of accessing and knowing about them. But in fact FA does not presuppose any of all this. FA says nothing about whether value, and therefore the fittingness of attitudes (or reasons for attitudes, as in the buck-passing account) can be understood in naturalistic terms, for instance as depending somehow on an agent’s actual or potential attitudes. The massive mistake imputed by Mackie to common-sense evaluative thinking lies not in FA, but in a certain meta-ethical (ontological and epistemological) understanding of the concepts and properties mentioned in the account.

FA can be taken on board by naturalists, error theorists, and non-cognitivists, not as the ultimate meta-ethical truth about value, but as a general truism about value which one’s theory will aim to explain or otherwise come to terms with. To the extent that FA, or something close enough, is a guiding assumption of this book, there is no reason for all such philosophers to decline their interest in the value-theoretic questions explored here.

1.7 Value theory: The questions

If x is valuable implies that it is fitting to favour x, then we can ask a number of questions:

IT IS FITTING TO FAVOUR X

FOR THE SAKE OF WHAT, OR WHOM? Final/Non-final Value (ch.2) Personal/Impersonal Value (chs 4, 7)

WHY? Intrinsic/Extrinsic Value (ch.2) Organic Unities (ch.5

AS WHAT? Absolute/Relative Value (ch.3)

HOW AND HOW MUCH? Respect, Promote, Prefer … (ch.7)

BY WHOM? Agent-neutral/Agentrelative Value (ch.4)

MORE/LESS THAN Y? Value Relations, Comparability (ch.6)