Nathan Nobis
nobs@mail.rochester.edu
Philosophy Department
University of Rochester, NY 14618 USA
http://mail.rochester.edu/~nobs/
Vagueness, Borderline Cases
and Moral Realism: Where's the Incompatibility?!?
In Philosophical
Writings (UK), No. 14, Summer 2000, pp. 29-39.
Abstract: Shafer-Landau
argues that occasional indeterminacy in the metaphysics of morals--i.e. that
moral predicates are vague and, thus, that some moral judgments are neither
true nor false (i.e. indeterminate in truth value)--is compatible with 'moral
realism,' the view that there are objective moral truths. I argue not that his
argument is unsound, but that the view he tries to undermine--that
indeterminacy and moral realism are incompatible--is a position held
by no philosopher informed on the issues surrounding vagueness and moral
realism (nor is it one that would be held upon sustained reflection by
those informed on these debates). Thus, since there is little antecedent reason
to presume that indeterminacy poses a problem for a realistic conception of
morality, Shafer-Landau's argument for their compatibility shows us little.
Keywords: Ethics, Moral, Vagueness, Indeterminacy, Borderline Cases, Realism, Anti-Realism, Bivalence
Shafer-Landau argues in "Vagueness, Borderline Cases and Moral Realism" that occasional indeterminacy in the metaphysics of morals is compatible with there being objective moral truths. For those that accept the existence of some linguistic indeterminacy and believe there is nothing particularly unique in kind about moral predicates and properties, i.e. that moral properties and predicates work in about the same ways as most non-moral predicates and properties do (and that an argument would be required to show that moral predicates and properties are sui generis), Shafer-Landau 's article contains nothing particularly disagreeable, but also nothing particular new and insightful for moral realists who accept or wish to develop a "moderate" account of vagueness (i.e. non-nihilistic and non-epistemic). From this perspective, the question is this: If moral predicates and properties are not very different in kind from non-moral predicates and properties, and it is sometimes indeterminate whether a non-moral predicate applies and whether a non-moral property obtains, then why suppose the situation is any different in the moral case?
I will argue that a good response to this question is that there's not much reason to suppose the lack of parity between moral and non-moral predicates and properties. Furthermore, I will note that Shafer-Landau gives no argument to deny this, i.e. no positive argument for the incompatibility of moral realism and vagueness. Thus it seems that Shafer-Landau 's primary aim is to undercut a not very-well-thought-out position on moral realism and vagueness (one that probably would be articulated only by a moral realist not familiar with the literature on vagueness), which is not a particularly difficult task.
Those that reject occasional indeterminacy in favor of nihilism (global indeterminacy or lack of truth values) or the epistemic view (global determinacy in truth values but irremediable ignorance of many of these truths) will find Shafer-Landau 's arguments against these views question-begging and his own moderate account of vagueness undeveloped and a little too comfortable relative to naive common sense. Those that are fond of a different unmentioned position on vagueness, especially one that is incompatible with Shafer-Landau 's, will be disappointed that these are the only three that he considers. If the truth about vagueness lies elsewhere than these three options, it remains open to show whether vagueness is compatible with the existence of objective moral truths.
I. The Sources of Moral Indeterminacy
Let's call the following the "indeterminacy
thesis" [(IT)]:
(IT) Moral predicates (hereafter
"M") are sometimes vague: it is sometimes indeterminate
(in a not merely epistemic sense), i.e. neither true nor false whether M
does or does not apply to a thing; there are cases where something is borderline
M.
The evidence for the moral indeterminacy thesis (IT), apart from the avowedly non-epistemic interpretation of indeterminacy, is that sorites-type arguments can be constructed containing predications of moral virtue and vice, e.g. ". . is generous", ". . is malicious", etc., predicates such as ". . is wrong to kill", ". . is a 'person' and is thus morally significant," ". . is morally responsible for x," ". . is a just society," etc. Also, there are comparative moral evaluations such as "x is kinder (or braver) than y." These examples show that vagueness, indeterminacy, and borderline cases are common in moral philosophy. (1)
Shafer-Landau 's account of how vagueness and indeterminacy are generated in the moral realm is basically the following (pp. 85-91): first, all moral predicates are multi-dimensional predicates.(2) So, for each moral predicate M, M's application depends on the application of its constitutive predicates, say, A, B, and C. Predicates A, B, and C are either unidimensional predicate or are themselves composed of constitutive, multi-dimensional predicates which are constituted by further sometimes vague predicates. Either way, the constitutive predicates of M are themselves sometimes vague. Since M's application depends on the application of sometimes vague A, B, and C, M is sometimes vague, so sentences of the form "x is M" are sometimes indeterminate.
But this is only half the account. The second factor accounting for vagueness is that for M to apply, not only must predicates A, B, and C apply, they must apply in a correct ranking. For example, let's suppose that for "generous" to apply to a person, that person must A, give a sufficient amount of goods away, B, give with a particular attitude, and, C, give to worthwhile causes, not frivolous ones. For a person to be generous, A is the most important factor in the ranking, but even if a person "maxed-out" in A, but greatly lacked in B and C, she might not be generous. Also, for someone that had so much B and C, even though she lacked much A, "generous" might truly apply to her. We could tinker with ascriptions of A, B, and C until we created a character for whom it is indeterminate whether she is generous or not. The point is that any moral predicate's application will depend on the rankings of the constitutive criteria and that these rankings can sometimes generate indeterminate results. So vague predicates can be sometimes be ranked in indeterminate rankings. Comparative moral evaluations--whether "x is M'er than y"--are all the more messy in that there is another factor for which indeterminacy can arise.
This is Shafer-Landau 's account of
how vagueness, indeterminacy and borderline cases arise in the moral realm. In
essence it mirrors a common account of how vagueness, indeterminacy and
borderline cases arise in the non-moral realm. Vagueness in ethics seems to be
generated in the about same way that vagueness about mundane, non-moral topics
is generated.
II. The Alleged Incompatibility of Moral
Indeterminacy and Moral Realism
Shafer-Landau believes that the
indeterminacy thesis (IT) is true and that it implies that "some morally
difficult questions admit of no right [i.e. true] answer" (p. 83). Since
(IT) states that "it is sometimes neither true nor false whether M
does or does not apply to a thing," Shafer-Landau is correct here. He reports,
however, that moral indeterminacy "is taken by most metaphysicians to
imply some sort of [moral] antirealism," (p. 83) and that "there is a
deep intuition that [moral] realism and indeterminacy are as complementary as
oil and water" (p. 84). He wishes to undercut this "deep
intuition" and show that the moral indeterminacy thesis (IT) is compatible
with what he calls "moral realism." By "realism" in
general, Shafer-Landau means ". . a theory about the source of truth
conditions for statements in a given domain," (p. 84) and, more
specifically, the view that
"the alethic status of propositions is
determined independently of the intentional attitudes taken by actual or ideal
persons towards such propositions or states they represent" (p. 84).
Shafer-Landau describes "moral realism"(3) as the view that:
(MOB) "moral judgments are
made true by things other than the intentional stances taken towards such
judgments or towards the facts they convey" (p. 84).
The following is a discussion of terminology which is necessary due to Shafer-Landau's non-standard labeling and juxtapositioning of metaethical positions.
Here are two brief, standard statements of moral realism.
More recent statements of moral realism are very similar to these:
". . moral realism claims that
there are moral facts and true moral propositions whose existence and nature
are independent of our beliefs about right and wrong. Moral realism's
metaphysical claim suggests the semantic claim that moral judgments and terms typically
[my emphasis] refer to moral facts and properties and the epistemological claim
that we have some at least approximate moral knowledge" (Brink, p. 24).
"Moral realists hold that morality is
objective: moral facts are discovered not legislated or created. The truth of
our moral beliefs is independent of our evidence for them and of our feelings
of approval or disapproval; they have objective truth conditions. On this view,
the wrongness of inflicting pain on humans for the fun of it is objective. We
disapprove of it because it is wrong, it is not wrong because we disapprove of
it. . . In addition moral realists hold that . . [s]ome moral beliefs are
instances of objective moral knowledge" (Tolhurst, p. 44).
Thus, most statements of moral realism entail semantic, metaphysical, and epistemic theses. Shafer-Landau 's "moral realism," which focuses on the metaphysical component of moral realism, is often called "moral objectivism," [I've called it (MOB)] the view that moral judgments have objective truth conditions: moral truth and falsity is not a direct, immediate, or simple function of a person's or persons' beliefs, attitudes, feelings, etc., but rather something else. (4)
The positive view of what this "something else" is a matter of normative ethical theory, but what Shafer-Landau wants to probably wants to avoid are views commonly expressed in a claim like, "If I (or we) approve of X then it's true that X is good!" This latter view about the source of truth conditions is sometimes called "moral subjectivism" or "relativism," as the truth of moral judgments is a function of a subject's or subjects' beliefs, attitudes, etc. and thus is relative from subject to subject (person to person) or community of subjects to community of subjects. Shafer-Landau 's calling the view, which I am calling moral objectivism, "moral realism" is confusing and misleading because it suggests that the person that denies (MOB) should be called an "anti-realist," but this clearly need not be so: someone could deny (MOB) and be a subjectivist or a relativist. Anti-realism, in its usual formulations, is a more extreme position than subjectivism or relativism because it implies, at least, that no moral judgments are true, which the subjectivist or relativist denies.(5) Because it will eventually become helpful to do so (when the distinctions become important), I will discuss Shafer-Landau 's "realism" under the label "objectivism."
It is important to note that SL is not defending his view that objectivism and indeterminacy are compatible from any stated arguments that they are not compatible. As far as I can tell, no one has argued, at least in print, for their incompatibility. SL only reports "intuitions" that this is so and that "most metaphysicians" agree (or would agree, perhaps if they reflected on the question), although he mentions no particular metaphysicians or ethicists that have addressed the question. This is interesting and puzzling because there is nothing prima facie strictly incompatible between (IT), roughly, the view that vagueness, indeterminacy, and borderline cases sometimes obtain in the moral realm, and (MOB), the view that the 'truth makers' for moral judgments are objective. To make the incompatibility apparent, we must clarify and disambiguate Shafer-Landau 's statement of moral objectivity.
Shafer-Landau's formulation of
(MOB) is ambiguous in at least two ways: should (MOB) be prefaced with a
"some" or "all" quantifier'? Two plausible ways to
understand (MOB) are:
(MOB') Some moral judgments are true and some are false, and their truth or falsity is determined by something objective.
or
(MOB*) All moral judgments are either true or false
(i.e. determinate), their truth or falsity is determined by something
objective, and some are true.
The indeterminacy thesis (IT) only says that some, not all, moral judgments are indeterminate, neither true nor false. (IT) presumably implies that judgments that are not indeterminate are determinate, either true or false, and for these, in itself (IT) says nothing concerning the nature of their truth conditions. The truth of (IT) does not preclude there being objective truth conditions for judgments that are true. Thus, (IT) is clearly compatible with (MOB').(6)
(MOB*) however implies the falsity of (IT) in that (MOB*) says that all moral judgments are true or false and (IT) says some are not. And, on a concrete level, (MOB*) implies that for every moral question there is a true answer. On the face of it, this sounds like something we would expect someone that calls herself a moral realist or objectivist, or someone "enthusiastic" about morality, might want to affirm.(7) She might reason to this conclusion along the following lines: take any moral judgment that is definitely true, for example, that unloading a dump truck full of babies with a pitchfork for fun is wrong. (MOB*) says its truth value will be determined by something objective, something apart from anyone's beliefs, attitudes, and feelings. If this is so, then there could not be any cases where that 'something other than beliefs, attitudes, and feelings, etc.' fails to be there and thus fails to make the non-definite cases, and therefore all the cases, either true or false. If this is so, then the indeterminacy thesis (IT) is false. Reasoning along these lines probably provides the initial, intuitive support for (MOB*) being the correct statement of moral objectivity, not (MOB').
So, the question again is whether
(IT) Moral predicates (M)
are sometimes vague: it is sometimes indeterminate (in a not
merely epistemic sense), i.e. neither true nor false whether M does or
does not apply to a thing; there are cases where something is borderline M.
is incompatible with
(MOB*) All moral judgments are either true or false
(i.e. determinate), their truth or falsity is determined by something
objective, and some are true.
Both have some intuitive appeal, but they can't both be right and a rational person can't accept them both. So, we can either reject (IT), reject (MOB*), or reject them both.
Let's first consider rejecting (IT). (IT) expresses the minimal conception of vagueness: sorites arguments show that there are borderline M's, cases where it's indeterminate whether x is an M, and thus that M is vague. What's controversial about (IT) is that it implies that nihilistic and epistemic interpretations of indeterminacy are false. Shafer-Landau does argue, however, for this assumption. He adopts the 'standard line' that there are no vague properties and that vagueness is a linguistic phenomena (p. 83). From here he considers three (and only three) interpretations of vagueness.
The first interpretation Shafer-Landau
considers is the nihilists', who he calls the 'radicals' (p. 83). Nihilists
accept a principle something like "for any predicate, if it is vague,
there are no properties that it designates." This, of course, implies that
there are no bald or tall men, well-behaved dogs, and freckled kids (p. 83). It
also implies that there are no hairs, grains of sand, people, or particular
material objects (and that all our beliefs about these things are false). In
the moral realm, nihilism generates this argument:
1. M is a vague predicate.
2. If M is a vague predicate, there are no properties that M designates.
3. If there are no properties that M designates, then no judgments of whether x is M can be true.
4. Therefore, no judgments of whether x is M can be true.
5. If it is true that no judgments of whether x is M can be true, then moral realism, of any form, is false.
6. Therefore, moral realism, of any
form, is false.
So, if nihilism is true, there are no truly and objectively unjustifiable killings, infidelities, or acts of generosity. Nihilism is incompatible with (MOB*), but Shafer-Landau rejects nihilism because of its radical implications, both moral and non-moral.
The second interpretation Shafer-Landau considers is the epistemicists', who he calls the "know-nothings" (p. 84). The "know nothings" say that there's always a fact of the matter in borderline cases--properties' extensions are perfectly determinate so that every person either is or is not charming, smart, a good dancer, rational, and so forth. This is extreme, but he shows that view is actually even more extreme since, for example, because the epistemicist says a single hair that makes the difference between bald and not-bald and since hairs are infinitely divisible, the epistemist is actually committed to the view that 1/10million+ (or whatever the greatest number a hair is divisible into) of a hair makes the difference between being bald and not bald. He thinks this is absurd and that even in the less absurd cases, the epistemic hasn't given a convincing argument for, first, why there are so many unknowable truths, and second, why we are so ignorant (p. 93). For these reasons, he rejects the epistemic view in general and, in particular to account for vagueness in ethics. If the "know nothings" get it wrong with baldness, then surely they get it wrong with any moral terms.
Shafer-Landau wants to keep the ordinary, commonsense, clear-cut truths (and our knowledge of them) that the nihilist denies and he wants to deny the epistemists' conclusion that there is always a true judgment in borderline cases and their predicament of having to explain away our ignorance. He wants a view that will support his saying that
"We
are occasionally stumped about whether someone is generous or laudable because
sometimes there is no fact of the matter awaiting discovery. In the absence of
such facts, the statements or sentences describing the relevant states of
affairs may be neither true or false" (p. 84).
(IT) fits the bill well and if it is the only other
alternative to epistemicism and nihilism, then (IT) is the one to go with. But
a question is, is (IT) the only choice left? Shafer-Landau only gives us three
options, and doesn't do much to develop his own account.(8)
There are more developed, nuanced
positions on vagueness that are not as easy to deny as nihilism and
epistemicism. In refuting these views, he does not do much to provide positive
support for his view, other than stating that it is preferable to two highly
contentious and, perhaps, absurd views. What would be more interesting would be
to see how vagueness in ethics fares in relation to a better, more substantive
account of vagueness. It would be interesting to discern whether moral
indeterminacy (IT) is incompatible with a better account of vagueness.
III. Should a Moral Realist be Committed to
Determinacy?
But, having said that, it seems that the indeterminacy thesis, with its non-epistemic interpretation of indeterminacy, is true. Also, since it is somewhat broad, it is compatible with a number of different accounts of vagueness. So, (IT) is a rather minimal thesis, and one that I think should be accepted, although probably not only on the basis of Shafer-Landau's rather sparse and non-comprehensive arguments.
We saw that the indeterminacy thesis (IT) is compatible with (MOB'), the view that some moral judgments are objectively true. The problem came when we saw that (IT) is incompatible with (MOB*), the view that all moral judgments are objectively true or false. We saw one informal argument supporting (MOB*) and noticed that Shafer-Landau does not articulate an argument for (MOB*), he just mentions intuitions that (MOB*) is the moral realist's view. So, we should wonder exactly why a moral realist or objectivist should accept (MOB*) or believe that (MOB*) is a statement of moral objectivism, especially a statement of the essence of the moral objectivist's point of view.
Shafer-Landau seems to get to (MOB*) by following Dummett's characterization of realism in general. Dummett writes that realism is that claim that ". . statements in a given class relate to some reality that exists independently of our knowledge of it, in such a way that reality renders each statement in the class determinately true or false, again independently of whether we know, or are even able to discover, its truth value" (p. 55, emphasis mine). Thus, Dummett's view is that realism implies bivalence and determinacy. While Dummett is free to use the term "realist" however he likes, why suppose that Dummett's characterization of realism is correct? Dummett's account of realism and what realism allegedly does and does not imply has been subject to severe scrutiny.(9)
Why suppose that a realist, especially a moral realist, accepts or should accept Dummett's view? Surely there are other general accounts of realism that do not imply universal bivalence and determinacy. None of the characterizations of moral realism above (Brink and Tolhurst) made any remark to the effect that realists are committed to determinacy; I have found no others that do so either. Sayre McCord, in an essay on moral realism, writes that ". . realism . . seems perfectly compatible with acknowledging that some of our predicates have indeterminate extensions" (p. 3). Sinnot-Armstrong notes that moral realism has been defined ". . as the claim that moral judgments obey the law of excluded middle" (190). He notes that Wiggins, 'Truth, Invention, and the Meaning of Life' requires this. He says that 'Such claims derive from Dummett, but I do not see why even extreme moral realists cannot accept that some moral judgments are neither true nor false in some cases' (243). Judith Thomson also agrees that there can be indeterminate in their truth value (154).
As far as I can tell, no actual moral realists take (MOB*) to be their position. If (MOB*) should be their position, Shafer-Landau should argue for this, and then show that their view is incompatible with moral indeterminacy. If moral realists should accept universal determinacy in the moral realm, presumably they should also accept determinacy in the non-moral realm as well, as, presumably, what makes it true that a person with no hairs is bald is as certainly as objective as the truth makers for any moral claim. But, obviously, we don't want to be committed to universal determinacy for baldness, so we shouldn't be want to be committed to determinacy in ethics either. The case between baldness and moral terms seem psychologically different in that nothing of much importance hangs on correct judgments about baldness, while moral judgments are often very important. This seems to play a psychological trick on some of us: because there are great stakes in getting some moral judgments correct, and objectively correct, it should be very important to get them all correct and that their be a truth about all of them. But why should this be so? I don't see anything in the views themselves that would imply this, and I don't see why it is morally important that there always be a true answer to every moral question. It's just not clear what difference occasional moral indeterminacy would make for the practice of making important moral judgments. If it does make a difference, Shafer-Landau should shown how this is so.
It seems that the genuine moral
realist is committed to (MOB'), not something as extravagant as (MOB*). Since
this is so, there is no incompatibility between moral realism and vagueness. Shafer-Landau
may have shown this, but since there is no reason to believe realism and
vagueness were incompatible in the first place, his victory is small.
1. Shafer-Landau gives this example to show that there can be borderline cases of generosity where it is indeterminate whether the person is generous. This shows that 'generous' is a vague predicate: 'A person's generosity, for example, is a matter of degree. If the frequency of her contributions diminish, and their amount dwindles, and the grace with which she gives continually fades, then at some point she clearly is no longer generous. Somewhere in the transition, however, we might be irremediably puzzled about whether to count her as a generous person' (p. 85).
2. For a discussion of multi-dimensional predicates see Keefe and Smith's introduction (p. 5).
3. His use of the term 'moral realism' is somewhat idiosyncratic since most statements of moral realism have epistemic implications beyond their semantic or metaphysical implications (i.e. there are some true moral judgments and their truth is objective), such as that there is some moral knowledge and/or justified moral beliefs. Given the epistemic worries about vagueness, it would have been better had Shafer-Landau not used the term 'moral realism' due to its traditionally epistemic implications.
4. Shafer-Landau's notion of "realism" or "objectivism" is not as extreme as Michael Wreen who takes ". . an object, property
(-instantation), or relation (-instantation) to be objective if and only if its existence is, or can be, absolutely independent of all mental states; and subjective otherwise," (p. 461). This strong notion of objectivity seems to imply that if there are any objectively true moral judgments, they could obtain even if there were no sentient beings. Many who think that morality is "objective," in a more common sense, are not committed to such as extreme view. But if this view is correct, most substantive ethical theories are trying to account for "subjective" properties. But, if these all deal with "subjective" properties, we need to distinguish between the crude subjectivisms (e.g. simple relativism) and the not-so-crude ones (e.g. consequentialism, Kantianisms, contractarianism, etc.). Wreen's distinction is so extreme and revisionist it's unhelpful. It creates the need for new, further distinctions.
5. For an anti-realist view, see Mackie.
6. (IT) is compatible with at least some (possible) theories that posit subjective (or non-objective) 'truth makers' for moral judgments.
7. In 'Is There A Single True Morality?' and other essays, Gilbert Harman, an anti-realist, argues that the answer to his question is "no," but that the implications for this view are not as unsettling as many might suspect. A realist or objectivist or non-relativist typically argues that there is a single true morality that makes, for example, what Hitler did really bad and that some of us believe this because it is true (it is not true because some of us believe it). The realist here in my example below seems to be to want to generalize from the clear-cut cases of moral evaluation to the conclusion that there must be a definite moral evaluation in all cases.
8. Shafer-Landau argues that 'If the arguments of the radicals and the know-nothings can be undermined, then we can retain our commonsense ontology and reject the prospect of absolutely precise properties and predications,' (p. 89).
9. See, for example, Rosen's critical review of Dummett's The Seas of Language.
· Brink, David. "Externalist Moral Realism," Southern Journal of Philosophy 24 (1986) Supplement, pp. 23-42.
· Dummett, Michael. "Realism," Synthese, 52 (1982), pp. 55-112.
· Harman, Gilbert. "Is There A Single True Morality?" pp. 27-48, in Copp, David and Zimmerman, Michael (eds.), Morality, Reason and Truth, (New Jersey: Roman and Allanheld, 1985).
· Harman, Gilbert and Thomson, J.J. Moral Realism and Moral Objectivity (Blackwell, 1996).
· Keefe, Rosanna and Smith, Peter, eds. Vagueness: A Reader (Cambridge: MIT Press, 1997).
· Mackie, J.L. Ethics: Inventing Right and Wrong (New York: Penguin Books, 1977).
· Rosen, Gideon. "The Shoals of Language," Mind 104 (1995), pp. 599-609.
· Sayre-McCord, Geoffrey, "The Many Moral Realisms," Southern Journal of Philosophy 24 (1986) Supplement, pp. 1-22.
· Shafer-Landau, Russ. "Vagueness, Borderline Cases and Moral Realism,"American Philosophical Quarterly 32 (1995), pp. 83-96.
· Sinnott-Armstrong, Walter. Moral Dilemmas (Oxford: Basil Blackwell, 1988).
· Tolhurst, William, "Supervenience, Externalism and Moral Knowledge," Southern Journal of Philosophy 24 (1986) Supplement, pp. 23-42.
· Wreen, Michael. "Vagueness, Values, and The World/Word Wedge," Australasian Journal of Philosophy, 63 (1985), pp. 451-464.