1 A Hedgehog’s Unity of Value

Joseph Raz
Ronald Dworkin was nothing if not an inventive and innovative theorist. Though,
like all of us deeply embedded in his time and the ideas of his time, he was
carving his views out of his own imaginative resources, to an ever-growing
degree free from the need to grapple, in his own contributions, with the
conventional paradigms set by others, and at the same time, in his critical
commentaries on events and ideas, dissecting the presuppositions, ideas, and
writings and exposing the fallacies of opponents.

As is to be expected, some ideas, or perhaps it is better to call them intellectual

tendencies, marked, often dominated, the movement of thought in much of his
writings. One dominant trend is the striving towards unity. And unity is my topic
today, or more specifically the unity of value in Dworkin’s Justice for
Hedgehogs. I will reflect on some of the many things he writes in dealing with
that theme. My main aim is to clarify his view about the unity of value. In doing
that, I will meander in different directions, trying out some interpretations before
turning to others. In other words, I will try to interpret his views in ways that will
turn out not to fit them— in part to show that the interpretations do not fit, and in
part to see how closely his views resemble them even so. I hope that by the end
of this journey, we will better understand his view about the unity of value.

One general caveat before we start: While drawing distinctions between values,
virtues, reasons, rights, duties, etc., where appropriate, Dworkin also uses
‘value’ in a more indiscriminate, all-encompassing way. Its scope is similar to
what other writers regard as the domain of the evaluative or normative, or their
combination. The unity of value is about value in that broad sense, and I will use
‘value’ in that sense in this paper, namely, to refer to reasons, norms, virtues,
etc., as well as to values in the narrower sense.

1. The Unity of Value: an Introduction

The theme is introduced in the opening sentence: “This book defends a large
and old philosophical thesis: the unity of value.”1 The first couple of pages make
clear that the one big thing that the hedgehog knows is both how to live well and
that value is one. As will emerge below, one lives well if one lives responsibly,
namely discharging successfully the responsibility project (a task in which no
one can be completely successful). The responsibility project leads one towards
the unity of value, and in living well that unity is manifested in one’s life. 2

What does the unity of value mean? You may think that it means that there is
but one value, and all the different values we may have in mind are but different
names for it (on the paradigm of the view that there is one deity, and that
different religions, and sometimes the same religion, have different names for
it). But that is not Dworkin’s thought. Or, you may think that it means that there
is but one value, and the different values we have in mind are but different
aspects of it. That may be closer to Dworkin’s thought. But that formulation is
itself obscure: What makes justice and liberty different aspects of the same
value rather than two different values?
History is rife with examples of such views. Perhaps, as utilitarians have it, there
is one value, say pleasure, and the different aspects of it are different causes of
its instantiations, say poetry and push pin. Or perhaps they are different
contexts in which it manifests itself. Perhaps the single value is desire
satisfaction (or some subclass of it) and the aspects are its manifestations
within family life or in one’s professional life (as when one’s desire to have
supper in the company of one’s family, or to complete one’s assigned task on
time, is satisfied). Or, if the only value is being virtuous, perhaps there is only
one virtue, say wisdom, though it can be manifested in different contexts, as
when one is courageous, which is being wise when facing danger, or generous,
which is being wise regarding the needs of others, etc. But there is no reason to
think that Dworkin understands the unity of value in any of these ways.

Some passages may suggest that Dworkin simply means that there is no
conflict between different values. But that could be at most part of his
meaning;3 it cannot be all he means. After all, very disparate values may not
conflict. For example, some jokes are funny, and that makes them good, at
least to a degree or in one way. And sleep can be restful, and that makes that
kind of sleep good, at least to a degree or in one way. It may be impossible for
the value of restful sleep to conflict with the value of funny jokes 4 (at least I do
not know what it could be (p.5) for them to conflict), yet it does not appear that
the two are aspects of one value. Or rather, the fact that they do not conflict
does not establish that they are such aspects.
Similarly, that one value cannot be instantiated unless another is (for example,
assuming that life and generosity are both values, one cannot be generous
unless one is alive), does not in itself establish that they are aspects of one
value. And nor does the fact that one value is a constituent part of another
establish that they are but aspects of one value. For example, possibly a
country is not democratic unless its residents are both free and literate, and
possibly these are constituent components of democracy (and I assume that
they are all valuable). That does not establish that democracy is an aspect of
freedom or of literacy, nor that there is nothing more to the values of freedom or
literacy than their contribution to democracy. Therefore, so far as this
consideration is concerned, the three are distinct values. 5 Finally, that all values
are values does not show that they are aspects of one value. It merely shows
that they share something—a common property or properties.

2. The Unity of Value: Consistent with Value

Pluralism
I mention these points to distinguish Dworkin’s thesis from other familiar theses
that are sometimes presented under the same name. Although he does not
offer a definitive formulation of his thesis, he says much from which his meaning
can be, at least partly, inferred. But before turning to that, it may be worth
examining briefly the opposing, or what may be taken to be the opposing,
thesis, namely that there are many distinct values.

Needless to say there are different versions of, different views about the nature
of value pluralism, not all of them incompatible. I will merely point to some
features of pluralism about values that seem to me right. They presuppose a
certain understanding of value.

For example, I will be assuming, along with Dworkin, 6 that evaluative properties,
namely, features of an activity, object, event, or whatever, that make it valuable
in some respect can in principle be understood, meaning that given favorable
circumstances (which are metaphysically possible), beings with the capacities
that humans commonly have can comprehend what is good about things that
are good and why. Again, with Dworkin, I assume that explanations of values
are not reductive. They employ other value concepts. Therefore, values do not
come in isolation. The value of anything will relate to some other values, which
may be constituents or consequences of it, or related to it in some other way
that makes it helpful to refer to them in an explanation of the value we are
explaining.

(p.6) Furthermore, our views about the value of things shape our attitudes to
ourselves, and the world around us. One aspect of that is that the value of
actions, of those actions that are options for us, is a reason for performing them.
That connects value to the quality of our life, for at the very least, a major factor
that determines its quality is that our life goes well when we engage in activities
and have attitudes that we have reason to have. But note that the connection is
asymmetric: Our life is good because we engage in activities that are good and
that connect us to valuable aspects of the world. It is not the case that these
activities are good because they contribute to the goodness of our life.
So far—some observations about features of value in general. Now to value
pluralism:

The question is: how are values individuated? That is, what makes one value
property distinct, and different from another value property, so that the
instantiation of one manifests a different value from that of the other? Think of
an example: humour is good and so is camaraderie. But they are different
goods, or values, as is manifested by the fact that the explanations of their
value will be different. We can, for example, expect that the explanations will
refer to different human capacities and dispositions excellence in which these
values bring out. Such facts will be part of the explanation of their value.
Possibly that will be all that they share, a reference to human capacities and
dispositions, and ways, different ways, in which their manifestations can be
valuable. The way these capacities or activities excel will be different. We may,
metaphorically, say that their point is different.
There is no way of avoiding metaphors when discussing the difference between
values. When explaining the difference between various derivative values,
those properties that are valuable because of their relations to other values
(e.g., instrumental values), we proceed by pointing to their dependence on
different values. But non-derivative values, precisely because they do not derive
from any others, deprive us of a non-metaphorical way of explaining their
differences except by engaging in detailed explanations of each value, or value
property, and observing their differences.

This understanding of the plurality of value allows for a great inflation in the
number of values. For example, there are various kinds of humor: There is
satire and irony and broad humor, and sarcasm and so on. In explaining the
value of each, we would identify them as species of humor, but will also explain
their differences, which make each one good in a different way. One marker of
that difference is that they are not interchangeable: what makes a satirical
observation suitable to the situation may not make sarcasm suitable. Of course,
what makes a remark about Jane suitable on one occasion may not make the
same remark about Liz suitable for the same occasion. But with satire and
sarcasm, the fact that they are not interchangeable is due to the different
excellences they display, whose display is typically appropriate in different
contexts.

(p.7) It is not merely that there are, on this view, many values. There is no end
to the possibility of fissure, the possibility that any good kind may develop
distinctively good subkinds,7 though, of course, some values are quite remote
from one another, and we normally think of those when referring to value
pluralism.
Some people would greet this value inflation with grave suspicion. They may
even take it to be an objection to this understanding of value pluralism. I draw a
different conclusion from it. It deflates the importance of the difference between
one value and another, but without losing sight of their distinct character. In a
way, these reflections bring value pluralism, the value pluralism I am discussing,
closer to Dworkin’s thought. I believe that Dworkin’s discussion is not hostile to
this view of value pluralism. His idea is not that there is only one value. He is
not concerned with criteria for the individuation of values, and does not rely on
such criteria to establish the unity of value. His thesis is not that there is only
one value but that there are certain relations among values that establish what
he calls their unity. When illustrating the unity of value in his first chapter,
Dworkin sees it in the fact that the different values mesh together, that they are
integrated, or that “the various concepts and departments of value are
interconnected and mutually supportive.”8

3. Exploring the Unity Thesis

This last articulation of the thesis (he describes that statement as “the more
general thesis of this book”) may strike one as rather distinct from what we
would normally understand by the unity of value (e.g., we expect different
scientific theories to be mutually supportive and interconnected without being
but aspects of one theory). But so long as we are not misled into taking Dworkin
to be contending for something he is not in fact arguing for, no harm is done. If
one insists we can take him to be arguing for a particular version of value
pluralism, one in which the different values are interconnected and mutually
supportive.

There is an understanding of such connections that would incline one to speak

naturally about the unity of value. Suppose that, while there are various distinct
values, one of them is supreme in that (1) all the others contribute to its
realization, perhaps by being constituent elements of it, or by their realization
being a precondition of its realization, and (2) if it is realized to the highest
possible degree, then there is nothing that can improve the way things are. The
second condition makes it the supreme value. In combination, all other values,
though important in themselves, are fully realized in their contribution to its
realization, for no improvement in them matters once it is realized to the highest
degree possible. But again, that is not Dworkin’s thought, and he says nothing
that suggests it.

(p.8) His discussion keeps returning to the idea of an “interconnected and

interdependent system of principles and ideas.” 9 The unity of value seems to
consist in that. But it is unlikely that we would find the key to unity in the kind of
connections or dependencies that Dworkin has in mind when discussing
specific values. He does not advance a view that a specific relationship or a
number of such relationships obtains between values generally and constitutes
their unity. On the contrary, he is open to the existence of many different types
of connections and dependencies, including ones not yet envisaged by anyone:
“what can count as an argument for a moral conviction is a substantive matter:
we must wait to see what connections among different departments of value
seem pertinent and appealing.”10 Rather the unity consists, as this quote
illustrates, in the fact that the connections constitute reasons for evaluative
beliefs, and in the case for the inescapability of an interconnected system of
beliefs of that kind.11 The argument goes to the conditions of truth about value:
a value judgment “can be true only if there is an adequate case … that supports
it. … [T]‌hat case must contain further value judgements. … . None of those
further value judgements can be … true [unless] a further case can be made
supporting each of them, and that further case will ramify into a host of other
judgements … [that] need yet further cases to show them true.”12
Dworkin is not discussing here epistemic reasons for believing that value
judgments are true. We can reach warranted beliefs about, say, the time of the
next train to Brighton, via various routes, each one providing sufficient warrant
for a belief about the time, and we can have various different epistemic reasons
providing alternative routes to a warranted evaluative belief that, let’s say,
Eggers’s latest novel has important lessons to teach us about the direction of
contemporary culture. We may be justified in believing that that is so by the
testimony of discerning friends, or by our knowledge of Eggers’s work and of his
interests, and of the themes explored in the novel (that we did not read), etc.
But the “cases” for value beliefs that Dworkin is writing about in the quoted
passages are not epistemic cases. They are—in his view—the evaluative
analog of truth- conditions, or of truth makers, in relation to non-normative
beliefs. Their existence is what makes value beliefs true. Perhaps we could say
that we are looking for the (p.9) grounds of the truth of value beliefs, and of
evaluative propositions more generally, using ‘ground’ in the meaning in which it
is used in recent writings.13
The question Dworkin addresses is not when do reasons to believe some value
judgment warrant belief in it, but what grounds the truth of a true evaluative
belief. His answer begins with the observation, “The truth of any true moral
judgement consists in the truth of an indefinite number of other moral
judgements and its truth provides part of what constitutes the truth of any of
those others.”14 So he is not looking for the grounds of the truth of value
judgments, for the relation between the grounds and what they are grounds of is
asymmetric, and so generally (though with some exceptions) is the relation
between truth makers and what they make true. As so often, Dworkin’s thought
defies current philosophical categories, and is of course none the worse for that.
A short way of referring to what he is after is useful, and I will use the
expression ‘a constitutive case’. His thesis is about the constitutive case for the
truth of value propositions and beliefs.

The constitutive case, he explains, is reflexive. It includes the belief for which it
is the case. But there is no question begging or vicious circularity here. A value
belief is not sufficient for its own truth, it is merely a small part of an indefinite
number of propositions that taken together would be the constitutive case for its
truth, if it is true. The quotation above is about moral beliefs. But the constitutive
case for the truth of true moral beliefs is not limited to other moral beliefs.
Dworkin explains:

“Morality is only one department of value … Is there any limit to the range of
convictions to which we might appeal in making a case that some action is
morally right or morally wrong? Or that someone is virtuous or vicious, or that
something is beautiful or ugly, or that some life is successful or unsuccessful?
Could a case for the unfairness of affirmative action include an aesthetic
judgement as well as a moral one? Could a case for the right way to live include
claims about the natural evolution of the universe or about the biological
heritage of animals in human beings? I see no conceptual or a priori reason
why not. What can count as an argument for a moral conviction is a substantive
matter: we must wait to see what connections among different departments of
value seem pertinent and appealing.”15
I already remarked on the fact that Dworkin’s case for the unity of value does
not include and does not rest on a view of the type of interconnections between
values. The question we examine now is how far the interconnections go. Given
what values are, each of them and each value proposition or value belief rests
on a constitutive case, and the values included in these cases themselves rest
on further constitutive cases. These cases, the quotation above explained, may
contain (p.10) any other principle or proposition. There is no general argument
that excludes any kind of principle, or conviction, evaluative or non-evaluative,
from being part of the constitutive case for the truth of any value belief. Dworkin
here, and elsewhere in the book, gestures towards the view that all the values
are interconnected in a chain of constitutive cases, which possibly include all
other true propositions as well.
But anyone who expects Dworkin to provide an argument to that effect may be
disappointed. In spite of the repeated reference to the unity of value residing in
the “interconnected and interdependent system of principles and ideas,” when it
comes to providing an argument it seems, as it does reading the quotation
above, that after all Dworkin does not know whether values are such a system,
whether they are united in that way, unless of course the ignorance is dispelled
elsewhere in the book. But it is not. There is nowhere in it a case for taking the
totality of true propositions to be the constitutive case for the truth of any single
value proposition, nor for the constitutive cases being connected in a chain of
justifications that embraces all values, let alone all other propositions.

But perhaps there is, or perhaps Dworkin thinks that there is, such a case in the
passage we are discussing. That is, does not the mere possibility that any
proposition is part of the constitutive case for any value proposition make it part
of that constitutive case? Of course, Dworkin does not claim that it is possible
that the constitutive case for any true value proposition is the totality of all true
propositions, nor does he say that the case for any proposition is chain-linked to
the case for any other proposition. All he says is that he sees no reason why
that is not so. Possibly he suspends judgment on the issue. But perhaps this is
just an understatement. Or, perhaps he believes that not seeing an objection to
a possibility shows that it is a possibility. Let that be as it may. The substantive
question is worth pondering: If it is possible that a proposition is part of the
constitutive case for some true proposition, does it follow that it is part of that
constitutive case? So put the answer is clearly “No”. It is possible, I take it, that
some propositions that are in fact false are true (For some p, p is possibly true
and p is false). It seems to follow that if all true propositions are possibly part of
the constitutive case for all true value propositions, then so are some false
propositions. But no false proposition is, as Dworkin understands matters, part
of the constitutive case for any true value proposition. This may be wrong.
Whatever Dworkin had in mind, possibly false propositions can be part of the
constitutive case for true propositions, for example, if they are part of a reductio
ad absurdum argument for them. Further thought is required.

If, however, we accept that no false proposition can be part of the constitutive
case for any true proposition, then we can rephrase the argument under
consideration to avoid the objection above. Perhaps it is the case that for any
proposition, if it is true then if possibly it is part of the constitutive case for some
proposition, it is part of that constitutive case.16

(p.11) The obvious cause for doubting that view is that in denying that the
constitutive case for a true proposition discriminates between relevant and
irrelevant true propositions, it renders the category unhelpful. But that criticism
could be rejected if the set that constitutes the constitutive case includes no
redundancy at all, that is, if none of its members can be excluded without
undermining the case for the proposition; in other words, without rendering it
untrue.
If that is the case for inclusion in a constitutive case, then propositions that
merely possibly contribute to establishing its truth cannot be part it. Only those
that actually contribute can be. One may doubt that test. It excludes the
possibility of over-determination. But that is best dealt with by taking
propositions whose truth is over-determined as having two or more constitutive
cases, with no redundancy within any single one of them.

Of course, without all true propositions belonging to the constitutive case of

each value proposition, we have no reason to think that this account of the truth
of value propositions guarantees the unity of value. Perhaps it does. But to see
that, we need a better understanding of what makes a proposition belong with
the constitutive case for another proposition. So far we avoided that question,
for if all of them belong there, possibly the question why they belong there need
not be faced, at least not when considering only the unity of value. Now the
question is inescapable. And it is difficult to answer because of an important
omission in Dworkin’s account. I noted earlier the need to distinguish between
epistemic reasons to believe in a value proposition and the reasons that
constitute the case for its truth. Many, perhaps all, reasons that belong to the
constitutive case for the truth of a value proposition are also, or can be
depending on circumstances, reasons to believe that the proposition is true.
That would be the case at least whenever it is possible to know the constitutive
reasons, and know that they are constitutive reasons, independently of knowing
the truth of the proposition whose truth they establish.

However, many epistemic reasons are not constitutive reasons. Some are easy
to tell apart, for example, testimony. But with others the distinction is less
straightforward. I already mentioned the difficulty I encounter as to whether the
propositions that figure in a reductio argument form a constitutive or an
epistemic case for believing its conclusion. Here is just one other example.
Often, and to some minds inescapably, we explain the value of something, as
well as value properties themselves, by analogy. The excellence of some poetry
is similar in some respects to the excellence of some music. The value of
patriotism is similar in some ways to dedication to one’s family, etc. Such
explanations help us to see that whichever side of the analogy we were less
clear about is valuable and how it is valuable. They provide reasons, not
necessarily conclusive reasons, to believe in propositions about those values.
But are they constitutive of the case for the truth of these propositions? Or, are
they merely ways of enabling us to “see” that those propositions are true on
grounds that do not include the analogy, which is after all merely a gesture to a
similarity, one that can also be misleading if taken on its own, etc. etc.?

(p.12) In my introductory observations about value in general, I endorsed, as

does Dworkin, the intelligibility of value, meaning the possibility of making
people with ordinary capacities and experience understand what is valuable
about possessing value properties, i.e., why they are value properties. Good
analogical reasoning is sufficient to secure intelligibility. It is far from clear,
however, whether this shows that it is part of the constitutive case for values
and evaluative beliefs. The distinction between epistemic and grounding or
constituting reasons for a value is important, and without it we lack a proper
understanding of what makes a reason part of the constitutive case for a value,
or what makes it a ground for the value.
The mere dependence of values on constitutive cases does not secure the unity
of value. It is consistent with the constituent cases belonging to disparate,
possibly even mutually exclusive domains of value. A substantive argument is
required to show that this is not so, that each value depends on all others, and
all values are mutually supportive. Dworkin does not offer such an argument,
and his failure to provide a distinction between epistemic and constitutive
grounds for value beliefs adds to the difficulty in finding in the book any steps
towards such an argument.

4. Conflict and Incommensurability

Referring to conflict of values Dworkin remarks: “If I am to sustain my main
claims in this book, about the unity of value, I must deny the conflict.” 17 The
relatively relaxed way that, as I suggested, Dworkin’s view about the unity of
value is to be understood does not compel that conclusion. Unity, as he sees it,
is in interconnections and mutual support, and these are matters of degree. A
degree of conflict is compatible with a degree of unity, as (in a different context)
is well known to all members of families. Moreover, even Dworkin does not
deny all practical conflicts. He denies conflict (though not the appearance of
conflict) between different values. Practical conflicts can exist even if there is no
conflict of values. The familiar example of a lifeguard on a beach illustrates that.
Two people will drown if he does not save them. He can save one but not both.
Which one ought he to save? The only values relevant, I will assume, are that of
saving or preserving life, and the value of doing his duty as a guard, which he
owes both to his employer and to the two people, given that his presence there
was publicized, and was the reason bathing there was safe and permissible.
Both considerations apply in equal measure to both bathers. He ought to save
each of them and he cannot save both. This is a not untypical case of practical
conflict, but it does not involve a conflict between different values. 18 However,
not being a conflict (p.13) between values these conflicts do not weaken the
thesis about the unity of value in the way that conflicts between values do. It is
that form of weakening of unity that he resists.
Dworkin is aware that some unity theses are consistent with conflict. He writes:

“My claim is not just that we can bring our discrete moral judgements into some
kind of reflective equilibrium—we could do that even if we conceded that our
values conflict … . I want to defend the more ambitious claim that there are no
genuine conflicts in value”19
Why do values not conflict? Because, as Dworkin sees it, it is never the case
that the realization of one of them to a greater degree restricts the degree to
which any other value is realized. Dworkin does not deny that some values can
be realized to a greater or lesser degree. He does deny that the limited
realization of one value can secure the realization of a second value to a
greater degree than would have been possible had the realization of the first not
been so limited. This, as Dworkin knows, does not strike many people as
obvious. Is it not the case that the geography of a region may be awe inspiring,
but if it changes in certain ways while losing that character, it may become
idyllic as it is not now, or that it may be beautifully colorful, with saturated colors,
but if it gets drier it would be less beautiful but more pleasant to live in? All four
qualities I mentioned in the two examples are non-instrumental value qualities,
or can be so understood. Yet in each example, one value is realized at the
expense of another. Are they not examples of cases in which one value is
realized at the expense of the other? Perhaps not, perhaps appearances
deceive. But first, two differences between these examples and the cases
Dworkin discusses.

First, he is often concerned with moral values: honesty and the avoidance of
cruelty, and such like. True, but his central thesis, about the unity of value,
applies (p.14) to all values. I would be the first to remind us that it does not
stand or fall with the absence of conflict. The unity may be greater or smaller,
and the presence of some conflicts, or of conflicts in some departments of
value, as Dworkin refers to them, is consistent with some kinds or some degree
of unity. Dworkin, however, sees a degree of unity, among values in general,
that excludes the possibility of conflict.
Second, my examples deal with situations in which the facts that impede or
restrict the realization of value may not result from human activity. They
certainly need not result from human activity in the pursuit of the values
mentioned in the examples. The conflicts that Dworkin is interested in are those
that face human agents who have to choose among options that appear to
realize one value at the expense of another. True, but the values in my
examples can be affected by human actions. Humans have been known to
interfere with the landscape to enhance its value in one way even while
detracting from the degree to which it manifests another value. It is possible that
when people confront such choices, one option is supported by a better reason
than all the others. But that in itself does not, as Dworkin reminds us, show that
the choice does not manifest a conflict of values.

So why does he deny that the realization of one value may be at the expense of
another? Let us look first at cases in which agents are confronted with what
appears to be a conflict of values and in which there is a conclusive reason to
choose one of the available options over any of the others. Why does Dworkin
think that the existence of a conclusive reason results from there being no
conflict rather than pointing to the right way of reacting to a conflict? Because, I
think the answer is, he cannot think what else it could be. He does not so much
offer an argument in support of his view as ask for an explanation of how any
alternative makes sense. Here is what he says:

“A colleague asks you to comment on a draft … and you find it bad. You will be
cruel if you are frank but dishonest if you are not. … the way to think further is to
further refine our conceptions of the two values. We ask whether it is really cruel
to tell an author the truth. Or, whether it is really dishonest to tell him what it is in
his interests to hear and no one’s interest to suppress. However we describe
the process of thought through which we decide what to do, these are the
questions that, in substance, we face. We reinterpret our concept to resolve our
dilemma: the direction of our thought is toward unity, not fragmentation. … .
What other story might one tell? Consider this one: “Moral conflict is real. … not
an illusion produced by incomplete moral interpretation; it is a matter of plain
fact.” But what in the world could that supposed plain fact consist in?”20
And he continues to remind us that there are reasons for moral truths, etc. And
how could conflict be ultimate if there are such reasons? This is a question that
requires an answer. Dworkin is right both that the process he describes
here (p.15) can make one realize that what one took to be a conflict of values is
not one, and that the thesis that values can conflict needs an explanation: How
is it that they conflict and why? The need for an explanation applies to all
practical conflicts, not only to those that involve a conflict among conflicting
values but those that do not, the ones that Dworkin implicitly allows. Of course,
quite a number of answers to these questions have been offered. Even my
sketchy observations earlier about the nature of value point to a family of such
explanations: different values have different points, bring out or enable different
excellences, and so on. Given that that is how they are individuated, there are
no grounds to doubt that the conditions for their realization may be
incompatible, thus yielding a conflict. True to his intention expressed at the
beginning of the book, Dworkin does not consider any of the explanations of the
possibility of conflict.21
Any account that allows for conflicts among values confronts not only the
question about the possibility of conflicts, but also the question: What ought one
to do when facing a conflict? I share Dworkin’s feeling that this task is hard to
discharge, even though I am not as pessimistic as he is. I think that often, more
often than is sometimes realized, when reasons conflict, no option is backed by
a conclusive reason; rather several of them are backed by incommensurable
reasons. Dworkin discusses incommensurability at some length.22 He insists on
one important lesson: Do not assume that two values are incommensurable or
that conflict between them is indeterminate just because you do not know any
better. Do not take incommensurability or indeterminacy to be a default, a view
to endorse in the absence of sufficient reasons for either alternative.

There are a good number of different phenomena often described as value or

reason indeterminacy or incommensurability. And his advice is sound regarding
all of them. I want to make one comment about the one kind of
incommensurability that interested me most, namely, when on a particular
occasion, an agent confronts several options backed by incommensurate
reasons. As Dworkin points out, if this is so, then there are explanations that
make it intelligible why it is so, and they operate on two levels. First, they
explain how it is that reasons of these kinds can be incommensurate with one
another. Second, they explain why the reasons that apply to this particular case
are incommensurate. But of course, there are cases in which we have reason to
believe that something is the case even though we have no explanation of why
or how it is that it is the case. So we may have adequate reasons to believe that
the value-based reasons for action in a particular case are incommensurate
even when we do not yet have an explanation of why they are incommensurate.
In some cases, the nature of the problem and the values that bear on it, coupled
with failure to find any grounds for holding that the reasons supporting any
option are conclusive even after due investigation, (p.16) may warrant the
conclusion that the reasons are incommensurate. Like many beliefs backed by
evidence, the conclusion may turn out to be mistaken. We are fallible even
when we form beliefs on rational grounds. Our fallibility does not establish that
the beliefs are unwarranted.
5. Constructivist Unity?
It is possible that the preceding discussion seriously misunderstands Dworkin’s
view.23 One ground for such a doubt is the absence from my discussion so far of
any mention of Dworkin’s ideas about personal responsibility. I have
approached Dworkin’s views about value as expressing an objective stand
based on what I will call object-dependent truths (ODT): Truths about value are
independent of any single person’s view about what values there are; the
constitutive case for them consists of values or propositions about values. You
may say that it is a form of realism about values, though Dworkin would regard
that term as confusing his view with the belief that there are bare facts about
values (meaning that there are values that cannot be made intelligible and
beliefs in which are not capable of being vindicated by a constitutive case).
According to the ODT interpretation of his view, individuals can appreciate what
is valuable and why, by engaging in interpretive reasoning. If they do so well
they will come to realize truths that are independent of the beliefs about these
truths that each one of them holds. Much that Dworkin writes suggests the ODT
interpretation of his views. But there are other indications that may point in a
different direction. Consider the following early passage:
“Interpreters have critical responsibilities, and the best interpretation of a law or
a poem or an epoch is the interpretation that best realises those responsibilities
on that occasion.”24
Could it be that Dworkin has a constructivist (to use current jargon)
understanding of truth about values? Does he mean that the correct or true
interpretation will be discovered by an interpreter who lives up to his
responsibilities on that occasion? Or, does he mean that the correct or true
interpretation is made correct by being the one that an interpreter who acts
responsibly will come to endorse? On the first reading, the one I have been
following in the previous discussion, the correctness of a belief about values is
independent of the way it is arrived at, leaving open the possibility that even a
flawless interpretive reasoning (p.17) that leads to endorsing it may fail and lead
to a mistake. This is how we normally think of truths about, for example, the
physical world. But possibly Dworkin has the second view: A correct interpretive
reasoning makes a belief true. The truth of any belief consists in the fact that
the person who has it reached it, or could have reached it through correct
reasoning. If he has it because of such reasoning he cannot be wrong. Which is
Dworkin’s view? If it is the second, if he is a constructivist about values, then the
unity of value is not so much a fact about values as a feature of correct
reasoning about values.
In an important passage Dworkin writes:

“We unreflectively interpret each [of our abstract concepts] in the light of the
others. That is, interpretation knits values together. We are morally responsible
[i.e. act in a morally responsible way, succeed in being responsible] to the
degree that our various concrete interpretations achieve an overall integrity so
that each supports the others in a network of value that we embrace
authentically. To the extent that we fail in that interpretive project—and it seems
impossible to wholly succeed—we are not acting fully out of conviction, and so
we are not fully responsible.”25
This pregnant passage is amenable to the second way of understanding
Dworkin, and it ties the unity of value not to how values are, independently of
what we may discover about them by interpretive reasoning, but to the
foundation of interpretation. Interpretation is, among other things, a process of
knitting values together. If your conclusion does not show them to be knitted
together, you failed to interpret as you should have. It seems that we find unity
because we unite values, not because they are united. And of course, we could
not unite values unless we made them through our interpretation. This is by no
means the only way to understand this passage. But it appears to be supported
by its end, which is remarkable in itself.

When our understanding (or aka interpretation) of the various moral values (or
aka concepts) does not present them as a network of mutually supportive
values (1) we did not succeed in being responsible, and (2) we do not
authentically believe, do not believe with conviction what we think we believe
(we are not acting fully out of conviction).26 How can that be? Suppose that the
unity of value (p.18) is an objective feature of value, independent of the view of
the person searching for the truth about value. In that case, by failing to realize
that they are united, one is ignorant of some truth. One may even have, as a
result, some mistaken beliefs. But neither ignorance nor mistakes normally
mean that the views one has, even the mistaken views, are not authentically
embraced or not held with conviction. If, however, the values and their unity are
a product of your interpretation of them, provided it is properly done and follows
the correct principles of interpretation, which include the goal of establishing
unity among the values, then failure to come up with unity is a mark of
not reallyinterpreting, and if you did not really interpret, given that beliefs about
values are your interpretation of the value concepts, it can perhaps be said that
your beliefs are not really, not authentically, your beliefs.
This is strongly supported by much else that Dworkin writes. “Our moral
responsibility,” he explains, “requires us to try to make our reflective convictions
into as dense and effective a filter as we can.” The image of a filter indicates the
role of our beliefs in containing, reshaping, or blocking, motivations and
opinions we have due to the accidents of our history.

“This requires that we seek a thorough coherence of value among our

convictions. It also requires that we seek authenticity in the convictions that
cohere: we must find convictions that grip us strongly enough to play the role of
filters when we are pressed by competing motives that also flow from our
personal histories. … We interpret each of … [our] convictions, so far as we can,
in the light of the others and also in the light of what feels natural to us as a
suitable way to live our lives. … Much of the rest of this book is an illustration of
how we might pursue that responsibility project.”27
It does look as if the unity of value is a constitutive requirement of the
responsibility project. We do not discharge our responsibility properly if we do
not find unity in value, or rather if we do not knit value into a unity, which
possibly we can do only if what values there are result from our doing so. This
passage introduces a further element through its explanation of the authenticity
requirement. A responsible interpretation is one that meets two conditions: (1)
unity that makes the values intelligible, and (2) being natural for each one of us,
given our diverse histories. The second condition has a dual aspect. One is
conventional: Our value beliefs, based on our understanding (or aka
interpretation) of values, restrain us from acting in some ways that our natural or
historically acquired inclinations incline us to behave. The other aspect of that
condition is more surprising: Our histories, and the inclination they produce in
us, also shape our interpretation. They shape our interpretation of values to
ensure that they have a strong grip on us, and that enables them to act as
filters, curtailing and shaping our motivations, etc. That grip is what ‘authenticity’
refers to in Dworkin’s discussion. Perhaps the authenticity condition is needed
to make beliefs in values have the power to make (p.19) us conform to the
values that bind us.28 Absent that kind of authenticity, we can understand why
Dworkin might think that our belief in the values we believe in is not complete.
But does it also show, as that and other passages suggest, that the values we
imperfectly believe in are not the true values, and our (imperfect) belief in them
is not a true belief, not being the result of a correct interpretation?
That extra step invites the interpretation that not only do we generate values by
correctly reasoning in the interpretive way about values (or value concepts), but
that the result of this construction of values, being influenced in each of us by
our personal histories, will generate different or differently unified values for
each of us. On that understanding, Dworkin’s theory of value is one of
perspectival constructivism. “Real” values are the products of successful
interpretive reasoning guided (possibly along with other considerations) by the
aims of unity and authenticity.

At times, it seems clear that Dworkin rejects the constructivist view of value.
“Morally responsible people may not achieve truth, but they seek it.” 29 But is this
a refutation of the constructivist interpretation? The constructivist interpretation
relates to successfully responsible people, but often Dworkin uses ‘responsible’
to refer to nothing more than the people who seek to be completely responsible.
He warns us that the term is used for various closely related ideas. It could be
that this quotation merely indicates that people who seek the truth, seek to be
fully responsible, may fail to construct values by failing to be fully responsible.

6. The Role of Interpretation: Unity Through

Division
In much of the paper, I explored the possibility that Dworkin anchors his thesis
about the unity of value in the fact that values are intelligible and that each is
grounded in a constitutive case for it, a case that relates it to other values in a
more or less seamless continuous web. In the previous section, however, I
explored the thought that he anchors it elsewhere, namely, in the responsibility
project that is identified as the project of integrating values, i.e., establishing
their unity, in an authentic way (in the special sense that that term has in the
book).

This second way of establishing unity raises, of course, the question of why we
are bound by the responsibility project. Dworkin says a fair amount about its
value, and though this may strike one as circular, given that on the constructivist
reading the responsibility project is the foundation of all value, this
circularity (p.20) may not be damaging. What is doubtful is whether what he
says about the value of responsibility is sufficient to establish it as the
foundation of values, including the duties that are stringently binding on all,
including those who have no interest in the responsibility project, or who doubt
the cogency of its conception or value. But these doubts do not matter, for
Dworkin would reject the allegation that his view is constructivist in the way I
explained.
He would also reject the ODT interpretation of his view. He would reject the
distinction between the two ways of understanding his view about value and its
unity. In Dworkin’s view they are one because the fact that being responsible
consists in part in finding unity in value connects with the case for unity that
derives from the dependence of any value on a constitutive case. 30 The
connection is revealed in, and is driven by, his view of interpretation. Dworkin
explains:

“Interpretation is pervasively holistic. An interpretation weaves together hosts of

values and assumptions of very different kinds, drawn from very different kinds
of judgement or experience, and the network of values that figure in an
interpretive case accepts no hierarchy of dominance and subordination. The
network faces the challenge of conviction as a whole …”31
As always, we ask: Is the interpretation in question an epistemic activity,
namely, one aimed at discovering what is there, what is the truth, independently
of it? Or is it an innovative interpretation that constitutes its object through the
activity of interpretation, when correctly done? Only if it is innovative would
there be a direct route to attributing characteristics of interpretative activities to
the truths that they yield and to the domain those truths are about. Even if
interpretation is the only epistemic access to that domain, its features cannot be
attributed to that domain just because they are features of interpretations and
without additional premises.

But there is an additional premise, one that I share, and in one form or another
many writers accept it and its ramifications, at least in part. It is that value truths
are intelligible, that people can understand them. That is a crucial and
challenging premise. Challenging—for it is as difficult to explain its meaning as
it is to establish its truth. However, that is another topic. For our purpose two
(possible) implications of the premise are relevant. First, if value truths are
intelligible then necessarily the only (non-derivative) way to establish what they
are is through interpretive reasoning. Second, it follows that necessary features
of sound Dworkinian interpretive reasoning, or more particularly features that
sound interpretive reasoning necessarily assigns to its domain, to truths about
value, really belong to that domain. Sound interpretive reasoning as Dworkin
understands it, is necessarily, according to him, the only non-derivative way to
establish truths about values, and to gain understanding about values. Hence
features that are necessarily attributed to values by such interpretations are
features (p.21) of values: In being necessary features of Dworkinian interpretive
reasoning, they are, inescapably, features that any promising attempt to
understand values attributes to them. Therefore, if—as we assume—we can
gain understanding of what we interpret, essential features of interpretation are
also features of the domain we thus understand.32
This argument enables us to sidestep the question whether Dworkinian
interpretation is epistemic or innovative. Even if it is merely interpretive it can—
given the two premises—reveal essential features of values. But perhaps that
should warn us that Dworkin might not have accepted this argument as a
correct representation of his view. It means that there is no more to the unity of
value than what can be established by the epistemic view of Dworkinian
interpretation, and that means that there is no more to it than what can be
established by ODT.

Dworkin may have taken interpretations as inherently more than merely

epistemic.33 In the end, I do not think that that matters. The features that
constitute the ODT approach are, according to Dworkin, essential features of
interpretive reasoning, as he understands it. So both when relying on
Dworkinian interpretation to provide an account of the unity of value, and when
approaching the task independently, the case for unity rests on the ability of the
ODT approach to establish what it is.

Furthermore, at the end of the day, Dworkin sees the case for engaging in
interpretive reasoning, as he understands that process, that is the case for
understanding values through Dworkinian interpretation, as resting on the
responsibility project. It is what responsibility requires of us. The case for the
responsibility project is that it is valuable, and its value must in the last resort be
vindicated by the ODT approach. The ODT approach, on the other hand, does
not need the responsibility project to be cogent. It stands and falls by the
argument that the dependence of value truths on constitutive cases for them
establishes the unity of value.

If the previous observations about Dworkin’s way of understanding the unity of

value are along the right lines, they show how the different aspects of his theory
are integrated. They direct us through various channels to ODT as the ultimate
foundation of the doctrine of unity. I see one problem in the case for the unity of
the different approaches: Although the authenticity condition can be read
into (p.22) some of Dworkin’s observations about interpretation (including the
passage cited above) in ways that mesh well with its foundation in the
responsibility approach, it has no place in the ODT approach. However, it is
possible to develop ODT to include a perspectival component that will bring it
into line with the responsibility approach.
7. Conclusion
Where do these ruminations lead us? As I see things, Dworkin gave us a
research project. Given that truths about values are grounded in constitutive
cases themselves consisting (in part) of truths about values, each one of which
depends on a constitutive case, and so on and so forth, we should research (1)
whether, and if so to what degree or in what ways, do the links thus existing
between truths about values connect all truths about all values, or only some of
them; and (2) how tight are the connections between values so established (do
they allow for conflict? Indeterminacies? Etc.)? Dworkin himself does not
provide reasoned answers to these general questions, though on some of them
his own beliefs about the results of such inquiries are clear.

References

Bibliography references:

Correia and Schnieder (2012). Metaphysical Grounding, Correia, F. and

Schnieder B. (eds.), Cambridge: Cambridge University Press.

Dworkin, R. (1986). Law’s Empire. Fontana Press.

Dworkin, R. (2011). Justice for Hedgehogs. Cambridge: Harvard University

Press.

Fine, K. (2012) “Guide to ground.” In Correia and Schnieder (2012).

Raz, J. (2003). The Practice of Value. Oxford: Oxford University Press.

Raz, J. (2011). From Normativity to Responsibility. Oxford: Oxford University

Press.

Rosen, G. (2010). “Metaphysical Dependence: Grounding and Reduction.”

In Modality: Metaphysics, Logic, and Epistemology, Bob Hale and Aviv
Hoffmann (eds.). Oxford: Oxford University Press.
Notes:
(*) I am grateful to the many participants in the Analytical Legal
Philosophy Conference, held in Oxford, 2014, who made many helpful
and instructive comments.
(1) Dworkin (2011: 1).
(2) I am grateful to Robin Kar for encouraging me to make clear the
connection between the unity of value and living well.
(3) Even that is not clear given that Dworkin allows for the possibility of
conflict (Dworkin 2011: 120). If he thought that that possibility is
sometimes realized then he allowed that the unity of value could co-
exist with conflict of values (as I believe that it can).
(4) Though they may be derivatively related to some conflicts (e.g.,
between buying a book of funny jokes and securing restful sleep).
(5) Things are different if the constituents of one value are not
independently valuable. Then they can be thought of as mere aspects
of the value to which they contribute.
(6) See Dworkin (2011: 113-115).
(7) I am presupposing here that the existence of values, at least of
some of them, is historically contingent. SeeRaz (2003).
(8) Dworkin (2011: 10).
(9) Dworkin (2011: 116).
(10) Dworkin (2011: 117).
(11) Nicos Stavropoulos suggested to me that Dworkin took the unity of
value to consist in the existence of some specific kind of connections
between values, connections whose nature and existence are an open
question. According to Stavropoulos, Dworkin is merely asserting that
possibly such a unifying connection or connections exist, and that
there is some reason to think that they do. Their existence should be
explored, and if they are found, then we will know that the unity thesis
is true. I believe that that view misconstrues Dworkin’s view. He argues
that the unity follows from the very nature of values in a way I explain
in the text, and therefore can be known even though we do not know
which connections between values instantiate it.
(12) Dworkin (2011: 116).
(13) E.g. Rosen (2010), Fine in Correia and Shnieder (2012) and other
essays in that volume.
(14) Dworkin (2011: 117).
(15) Dworkin (2011: 117).
(16) I am grateful to Ori Simchen for suggesting this modification of the
argument.
(17) Dworkin (2011: 118).
(18) See Raz (2011: ch. 9). That is not the only kind of practical conflict
that leaves Dworkin unperturbed. He also allows for conflicts between
desiderata (“Desiderata almost always conflict. … A community wants
the highest level of security, the best educational system … . But its
budget is tight.” Dworkin (2011: 118).) and between desiderata and
values (“Values often conflict with desiderata. … Some steps we might
take to improve safety from terrorists, which we certainly desire, would
compromise liberty and honour.” Dworkin (2011: 118).) ‘Desiderata’ is
used by Dworkin not in its common meaning (desirables) but to refer to
what is desired (“Desiderata are what we want but do no wrong not to
have.” Dworkin (2011: 118).). We want what we want for what we take
to be reasons. But even so, we often want what we should not want,
what is worthless and pointless to have or to want. Needless to say,
there may be conflict between what we want and what is of value. But
Dworkin’s examples, or most of them, are not of this kind. They are of
wanting what is worth wanting, indeed in some cases wanting what we
ought to want (even on his view, given that he probably believes that a
community should strive to protect itself from terrorists). In as much as
desiderata are backed by reasons, by values that the wanting or its
satisfaction realize, it would seem that the case of such conflict is the
same as that of conflict between values. If desiderata can conflict, so
can values. However, if persuaded that I am right about these conflicts
Dworkin would have denied that desiderata can conflict either.
(19) Dworkin (2011: 119) Scott Hershovitz suggested that as Dworkin
writes that the view that values conflict is “conceivable and perhaps
someone might make it seem plausible” Dworkin (2011: 120), he does
leave the possibility of conflict open. In a sense that is true, it is a
“conceivable” possibility. But it is not a matter on which his theory is
silent or agnostic. As the quotation above shows, the theory he is
advancing denies that that conceivable possibility is ever realized.
(20) Dworkin (2011: 119).
(21) I am not considering the separate question of whether although
values conflict in application they cannot conflict in themselves. An
intriguing and difficult question, even to understand, it is not one that
engages Dworkin in this book.
(22) Dworkin (2011: especially 90-96).
(23) I am grateful to Sari Kisilevsky for alerting me to another issue: Is
my inquiry inconsistent with Dworkin’s rejection of a separate domain
of metaethics? I think that I said and implied nothing that offends
against Dworkin’s views in this regard. But the relevant issue is
whether the questions raised here merit an answer. If they do, then if
they are inconsistent with his view on metaethics, so much the worse
for that view.
(24) Dworkin (2011: 7).
(25) Dworkin (2011: 101).
(26) Matthew Kramer suggested to me that Dworkin is not saying that,
in that case, our beliefs lack conviction, or authenticity. Rather, if our
view of value lacks unity, we cannot act on it, and our actions lack
conviction and authenticity as they do not match our beliefs. This is an
ingenious reading of the text. It attributes to Dworkin the view that
evaluative beliefs that are not sufficiently united are not only
incomplete, representing only part of the truth, but that they are
defective in a more fundamental way, perhaps in ways analogous to
the ways in which irrational beliefs cannot be rationally understood.
There is no reason I know of to attribute to Dworkin that view. The one
possible textual support for Kramer’s view is the reference to people
not then being fully responsible, but that could be both because their
beliefs are not fully theirs as well as because their actions do not
match their beliefs. The text is about the interpretations (i.e.,
interpretive beliefs) not being authentically held and the actions not
being done out of beliefs that we hold with conviction.
(27) Dworkin (2011: 108).
(28) Crude Kantian views may stop with the first aspect of the second
condition: Value beliefs stop us from acting on some inclinations, but
allow us to act on others, so most of our morally permissible actions
are not guided by reasons for action, but simply caused by our
inclinations. Dworkin is no Kantian. His thought does not rely on any
such distinction between value beliefs and inclinations. Rather, as
explained in this passage, it takes the interpretation of value to be
shaped by our inclinations, making our value beliefs authentic as well
as enabling them to filter out other inclinations.
(29) Dworkin (2011: 113).
(30) Dworkin (2011: 112).
(31) Dworkin (2011: 154).
(32) That only features necessarily attributed to its domain by
interpretation of values can be attributed to values makes the view
consistent with the possibility that even correct interpretive reasoning
could yield a mistaken conclusion about values, for the attribution of
value to empirical situations usually depends also on ordinary inductive
reasoning.
(33) Though Dworkin always allowed that ‘interpretation’ designates
different kinds of activities. In Dworkin (1986: 52-53) he distinguished
scientific interpretation—which is really interpretation only in a
metaphorical sense—conversational interpretation and creative, which
is constructive, interpretation. In Justice for Hedgehogs, the variety of
types of interpretations and their dependence on the specific purposes
they serve is taken very seriously. Hence, the question about the
character of interpretive reasoning about value, and the possibility that
there is more than one kind of such interpretive reasoning is consistent
with Dworkin’s approach to the matter.