1(8)
Contextualism in Ethics
Gunnar Björnsson
In more than one way, context matters in ethics. Most clearly, the moral status of an
action might depend on context: although it is typically wrong not to keep a promise,
some contexts make it permissible. More radically, proponents of moral particularism
(see PARTICULARISM) have argued that a reason for an action in one context is not
guaranteed to be even a defeasible reason in every context; whether it counts against an
act that it breaks a promise or inflicts pain might depend on the particulars of the
situation. In moral epistemology, Timmons (1999: Ch. 5) argues that whether a moral
judgment is epistemically responsible depends both on the basic moral outlook of the
moral judge and on whether the context of judgment is one of engaged moral thinking,
or one of distanced, skeptical reflection. In the former, the judge’s basic moral outlook
can serve to justify the judgment; not so in the latter (see EPISTEMOLOGY, MORAL).
Our focus here, however, will be on forms of metaethical and, more precisely,
semantic contextualism in moral discourse and moral thinking. According to these
forms of contextualism (henceforth “metaethical contextualism,” or just
“contextualism”), the meaning or truth‐conditions of a moral judgment or moral
assertion depend not only on the properties of the act it concerns, but also on features
of the context in which the judgment or assertion is made, such as the standards
endorsed by the moral judge or the parties of the conversation. If metaethical
contextualism is correct, it might be that when two people both judge that abortions
must be banned, one judge might be correct whereas the other is mistaken, because
they accept different fundamental norms. This would undermine the idea that there are
unique correct answers to moral questions.
Metaethical contextualism is supported from three directions. First, what is
expressed by terms such as “good” and “ought” seems to be context‐dependent when
used outside ethics, varying with speakers’ interests and concerns. One might therefore
expect similar context dependence when these terms are used to express moral
judgments, assuming a corresponding variety of moral interests and concerns. Second,
many have thought that deep moral disagreements show that the interests and concerns
behind moral judgments do vary in this way. Finally, contextualism promises to make
sense of what seems to be an intrinsic yet defeasible connection between moral
judgments and moral motivation, by tying the meaning or truth-conditions of moral
judgments closely to interests and concerns of moral judges. At the same time,
contextualism faces two broad kinds of challenge: to make sense of the seemingly
categorical or objective pretensions of moral claims, and to explain why the parties to
deep moral disagreement often behave as if they were disagreeing about substantive
issues rather than talking past each other. In the sections that follow, we look closer at
both sources of support and problems for contextualism.
Draft for the 2019 revision of the International Encyclopedia of Ethics (ed) Hugh LaFollette. Wiley.
CONTEXTUALISM IN ETHICS
2(8)
The Context Dependence of “Good” and “Ought”
It is well known that the content of what we say is determined not only by the words we
choose, but also by the context in which the words are used. Clearly, what claim is made
by an utterance of “I am here now” depends on the time and place of the utterance,
and on who is making it: “now,” “here,” and “I” pick out different times, places, and
individuals in different contexts. Similarly, the judgments expressed when we say of
someone that she is “tall” might variously measure her against the length of others her
age, or others her age and sex, or against that of, say, professional basketball players.
Interestingly, context‐sensitivity also seems to be a feature of expressions used to
make moral or normative claims, such as “obligation,” “right,” “wrong,” “must,”
“ought,” and “good.” We shall focus on “ought” and “good,” beginning with the latter.
Consider a sentence such as:
(1) “The weather is good.”
If it is uttered by tourists considering a day at the beach, it seems to say that the weather
answers to the requirements for a pleasant day at the beach; if uttered by farmers
considering crop yields, it would instead be saying that the weather answers to the
requirements for a high crop yield. Since these are different requirements, one utterance
of the sentence might be true while the other is false, even if the utterances concern the
weather at the same time and same place.
In light of these and other examples, it seems clear that “good” means, roughly,
“such as to satisfy requirement R,” where R varies across contexts (see Mackie 1977:
55–6; for similar suggestions, see Finlay 2014: Ch. 2; Thomson 2008: Chs. 1–4). The
relevant requirement, R, is often clear from the kind of thing called “good”: “a good
conversation” typically means one satisfying typical requirements on conversations, and
“a good knife” means one satisfying the requirements that we typically have when using
knives, i.e., cutting and handling well. At other times, R is clear from our specific
knowledge of the interests and concerns of the speaker: a scenographer talking about “a
good knife” might have in mind one fitting a certain scene well, whether or not it cuts
well. When the kinds of requirements in question are otherwise unclear, they are often
made more explicit: “good for the crops,” “good for a day at the beach,” “good for you
but not for me.”
“Ought” seems to display similar patterns of context dependence. In general, to say
that something ought to be the case or that someone ought to do it seems to express
that it is “favored” among some set of relevant alternatives. What varies from context
to context are the considerations that select the relevant alternatives and favor one of
them. For example, depending on context, an alternative might be favored because it
is probable given the evidence, or makes the achievement of some end most probable,
or conforms better to some ideal:
(2) “Since she left almost an hour ago, she ought to be here soon.”
(3) “[Given what we knew,] It ought to have dissolved; I wonder why it didn’t”
(Mackie 1977: 73).
3
(4) “To look more casual, Jill ought to wear blue jeans.”
(5) “To open the safe, Jill ought to use a Phillips screwdriver” (cf. Wedgwood 2007:
117).
As with “good,” context might make the italicized explications of the relevant
considerations unnecessary: watching Jill struggling to open the safe, we can utter (5)
without the explication, to the same effect.
Notice that “ought” judgments that relate to goals or ideals need not express the
speaker’s endorsement of the goal or ideal in question: (4) and (5) might be uttered as
pieces of conditional advice, or as matter‐of‐fact information about dress codes or
locksmithery. Similarly, “ought” judgments can be made relative to a restriction of
alternatives that we do not endorse, as in:
(6) “Since you are not going to stop shooting up heroin, you ought at least to use clean
needles” (cf. Jackson 1985: 181–2; Wedgwood 2007: 119).
Judgments can also be made both relative to evidence that we ourselves possess, and
relative to evidence possessed by an agent who we think ought to do something. For
example, looking down from above on someone making his way through a maze, we
might be correct in saying both of the following:
(7) “He has no way of knowing it, but he ought to turn left at this point.”
(8) “Given what he knows, he ought to turn right at this point” (Wedgwood 2007:
118; my italics).
Again, although explicit indications of the relevant states of evidence often help, the
right conversational context makes them redundant. (For different ways of
understanding the context dependence of “ought” judgments, see Mackie 1977: Ch. 3;
Wedgwood 2007: Ch. 5; Price 2008: Ch. 2; and Finlay 2014: Ch. 3. For criticism, see
Thomson 2008: Chs. 10–11.)
From Semantic Context Dependence to Metaethical Contextualism
The discussion in the previous section left open numerous questions about the exact
analyses of “ought” and “good,” and said nothing about how context determines the
content of claims involving these terms. What seems clear, however, is that what
utterances using those terms express does depend on context, and more precisely on
interests, goals, and ideals that are relevant in that context. Our question now is whether
this extends to uses of “good” and “ought” in ethical contexts, such that two people who
are considering whether an action is morally good, or whether it ought to be performed,
might be asking different questions. In other words, does the contextual variation
support metaethical contextualism?
We have already seen one possible example of this, illustrated by (7) and (8):
apparently some “ought” judgments mean to identify the alternative that is morally or
rationally ideal in relation to the knowledge or beliefs of the agent, whereas other
judgments also take into account information that the agent lacks (Jackson 1985; Finlay
CONTEXTUALISM IN ETHICS
4(8)
2014). If such information relativity is possible, perhaps differences between those who
tie the moral status of actions to their actual consequences and those who tie it to their
expected consequences stem from a focus on different moral questions. That there are
different such questions is further suggested by thoughts about what one ought to do in
cases of normative uncertainty:
(9) “I know that the correct theory is either utilitarianism (in which case I ought
to choose policy A) or prioritarianism (in which case I ought to choose policy
B). But since I have no idea which of these two theories is right, I ought to
choose policy C. For on either theory, policy C is a close second.” (Pittard
and Worsnip 2017: 168).
This line of reasoning seems to presuppose that the ought-claim of the conclusion is
compatible with the ought-claims of utilitarianism and prioritanism. The apparent
intelligibility of this reasoning, whether sound or not, thus suggests that we readily
recognize different ought-questions.
More contentious and radical than information relativity is the idea that the content
of moral judgments varies with the ideals operative in the context of judgment. Here,
proponents of metaethical relativism have argued that deep moral disagreement and
differences in moral outlook between cultures, or between, say, liberals and
conservatives or consequentialists and deontologists, make it implausible that the truth‐
conditions of their moral judgments coincide (see RELATIVISM, MORAL; DISAGREEMENT,
MORAL). If sentences involving “good” and “ought” generally express different claims
depending on the requirements, ends, and ideals that speakers take to be relevant to
their judgments, this would seem to provide a natural fit for the moral relativist,
accommodating both individual and cultural moral variation.
However, some who have stressed context dependence have simultaneously rejected
metaethical relativism. Even if one thinks that the terms used to express moral
judgments are context‐dependent, one might also think that, in moral contexts, our
judgments relate to the same fundamental requirements (see Mackie 1977; Wedgwood
2007). In the following section, we consider some of the concerns that people have raised
against metaethical contextualism.
Categoricality, Motivation, and Disagreement
One common objection to metaethical contextualism is that it fails to capture the
practical relevance or normativity (see NORMATIVITY) of moral judgments. If moral
“ought” judgments are always relative to some set of considerations, and if there are
numerous such sets that favor different alternatives, morality seems neither to offer
practical guidance nor demand any particular action (see, e.g., Montminy 2007). What
the contextualist can say, however, is that when we are asking ourselves what we ought
to do, then, insofar as our question has a determinate content, we are already relating
that question to one particular set of considerations. Moreover, insofar as these are
considerations that we ourselves endorse acting from, the judgment we arrive at will be
practically relevant for us.
5
Even so, one might think that metaethical contextualism fails to capture the sense in
which we intend our judgments to relate to the correct set of moral considerations, not
just the set that we happen to endorse. But it is unclear how this would be a problem
for contextualism. Insofar as one’s sense that considerations should be correct has
reasonably determinate content, the contextualist can understand it as constituting an
abstract or higher‐order way of specifying the considerations that one endorses. As long
as standards of correctness vary across contexts, the upshot would still be a form of
metaethical contextualism. Moreover, deep moral disagreement does seem to involve
variation not only in what parties count as relevant moral considerations, but also in
standards of correctness.
A related but importantly different objection focuses on the connection between
moral judgments and moral motivation. According to metaethical contextualism, our
motivation to act on a particular “ought” judgment depends on something external to
the judgment itself, in particular on whether we are motivated by the set of
considerations invoked by the judgment. This might seem highly plausible for some
explicitly relativized or conditional “ought” claims (such as the claim that, in order to
look more casual, I ought to wear blue jeans). But in normal cases that lack explicit
relativization, thinking that one ought to do something and being at least somewhat
motivated to do it do not seem to be two separate states: the motivation seems internal
to the judgment. Similarly, to think that an action is “good” often seems to already
involve being in favor of its performance. Many have taken this to suggest that moral
judgments are either beliefs with contents that are somehow intrinsically motivating or
desires or preferences of some sort rather than beliefs (see INTERNALISM,
MOTIVATIONAL; COGNITIVISM; NON-COGNITIVISM).
Though contextualism might seem incompatible with an intimate connection
between “ought” judgments and motivation, some have instead argued that
contextualism is particularly well placed to account for this connection. While denying
that the content and nature of “ought” and “good” judgments guarantee motivation,
Dreier (1990) and (in more detail) Finlay (2014: Ch. 5) suggest that, in identifying the
content of a given “ought” sentence, we normally assume that the set of considerations
that it relates to is one that motivates the speaker. (Similarly for the requirements that a
“good” sentence relates to.) The reason for this is that, in contexts where it is not
otherwise clear to what considerations an occurrence of “ought” relates, the audience
can reasonably assume that the considerations are ones that matter to the speaker:
otherwise, why make the claim? (This assumption is especially reasonable in contexts
where the speaker is engaged in practical deliberation.) Making an “ought” claim in
such contexts thus creates a presupposition that the speaker cares about the relevant
considerations, whatever they are. Since this presupposition precedes any assignment
of a specific content to the “ought” claim, it will naturally seem that the motivation is
inseparable from the judgment expressed.
If this pragmatic account is correct, it seems to capture the practical relevance of
moral judgments without postulating intrinsically motivating contents or taking on the
CONTEXTUALISM IN ETHICS
6(8)
difficulties of a non‐cognitivist analysis of moral judgments. Moreover, it leaves room
for cases suggesting that motivation is external to “ought” judgments: cynics, amoralists,
or apathetic people who realize that they ought not to do what they are doing, but who
are unmoved by this thought (see AMORALIST). Contextualism allows that such
characters make judgments in relation to the requirements and considerations that they
were once moved by, or that other people in their community are moved by.
Contextualist analyses of “ought” might also provide a natural fit for the idea that
what an agent ought morally to do must be an action that the agent could be (rationally)
motivated to do (see RATIONALISM IN ETHICS). In many contexts where we make
judgments about what someone ought to do, our judgments seem to have a rational
practical point, being intended to identify an action that the agent could (in principle)
be rationally convinced to do, or has internal reasons to do (see REASONS, INTERNAL
AND EXTERNAL; cf. Finlay 2006). If the truth‐conditions of “ought” judgments depend
on our concerns in making them, as examples of the context dependence of “ought”
suggest, the truth of judgments with this particular point could well depend on whether
the action in question is favored by an ideal or end that the agent could be rationally
compelled by. Furthermore, if agents’ varying subjective desires determine what ideals
or ends agents can be rationally compelled by, they would determine what agents ought
to do, in the sense expressed by these judgments. The upshot would be a version of
normative relativism, though one restricted to practical “ought” judgments with a
rational practical point.
Perhaps the most common and serious objection to metaethical contextualism is that
it misrepresents agreement and disagreement (see Lyons 1976 for a classical statement;
cf. Dreier 2009). For example, it seems natural to say not only that those on the other
side of seemingly intractable moral controversies about abortion, animal rights,
ownership rights, etc. ultimately relate to different moral considerations, but also that we
disagree with their judgments and, often, that they are wrong or mistaken. In saying this,
we seem to do two things. First, we attribute moral judgments to those on the other side,
understanding that these judgments are based on their fundamental moral outlook.
Second, we assess the correctness of their judgments based on our moral outlook. This
two‐part practice seems hard to reconcile with the contextualist assumption that the
truth‐conditions of moral judgments depend on the considerations that moral judges
relate to when making their judgments. Given this assumption, our assessments seem
blatantly insensitive to the truth‐conditions of the judgments assessed. Also, this
insensitivity would contrast starkly with how we assess judgments involving some other
context‐dependent expressions. If Alexander says of 7‐year‐old Beth that “she is tall,”
clearly comparing her with other children her age, we would not naturally say that
Alexander is wrong merely because Beth struck us as short compared to adults. If he is
wrong, it seems, it would be because he got his comparison wrong, not ours.
Metaethical contextualists seem to have two broad kinds of options in trying to
explain these phenomena.
The first sort of explanation would be that, in attributing disagreement and making
7
insensitive assessments, we make a mistake. Perhaps we make a semantic mistake, failing
to understand that the truth‐conditions of moral judgments depend on the
considerations that are relevant in the context of judgment. Given that such semantic
blindness seems constrained to some specific context‐dependent domains (including
moral discourse, but excluding attributions of tallness), this explanation would seem
plausible only if something about these domains would make this mistake particularly
likely (cf. Francén Olinder 2013). Or perhaps, in spite of the deep differences between
parties of intractable moral disagreements, we make a pragmatic mistake, failing to take
into consideration the possibility that the parties ultimately relate to different
considerations in making their judgments. The pragmatic suggestion gets some support
from empirical data indicating that insensitive assessments of correctness or
incorrectness become less common as the difference between moral outlooks becomes
increasingly radical and thus harder to ignore (Sarkissian et al. 2011). If this is right,
contextualists need not worry about seemingly insensitive assessments. But data also
suggests that we continue to attribute disagreements between parties with radically
different moral outlooks even while ceasing to insensitively assess correctness and
incorrectness (Khoo and Knobe 2016). Apparently, then, contextualists would still need
a nonstandard account of disagreement attributions.
Another sort of explanation sees attributions of disagreement and insensitive
assessments as perfectly adequate given the practical function of moral judgments. Moral
questions are largely shaped by concerns about how to behave, what attitudes to have,
and what behavior and attitudes to publicly support, and we normally expect people’s
moral judgments to correspond to attitudes guiding their behavior and emotional
reactions. Because of this, we have a practical interest in keeping track of when people
make moral judgments with conflicting attitudinal or behavioral consequences, as such
judgments constitute a kind of practical disagreement. We also have an interest in
assessing the correctness of moral judgments of others relative to the considerations that
we endorse as publicly upheld guides of conduct. Even given contextualism, this could
explain not only attributions of disagreement and why we make insensitive assessments,
but also why such assessments are limited when moral outlooks seem to rule out any
common guides of conduct. Moreover, it would account for the difference between
moral talk and thought and other context‐dependent domains – tallness judgments do
not normally have the same intimate connection to attitudes and action. (For
developments of this suggestion, see Finlay 2014 and Björnsson 2015.)
Whether any of these explanations succeeds is currently an open question, as is the
question of whether parties of moral disagreements do have fundamentally different
concerns (see DISAGREEMENT, MORAL).
See also: AMORALIST; COGNITIVISM; DISAGREEMENT, MORAL;
MORAL;
INTERNALISM,
MOTIVATIONAL;
NON‐COGNITIVISM;
PARTICULARISM; RATIONALISM IN ETHICS; REASONS, INTERNAL
RELATIVISM, MORAL
EPISTEMOLOGY,
NORMATIVITY;
AND EXTERNAL;
CONTEXTUALISM IN ETHICS
8(8)
REFERENCES
Björnsson, Gunnar 2015. “Disagreement, Correctness, and the Evidence for Metaethical Absolutism,”
in Russ Shafer‐Landau (ed.), Oxford Studies in Metaethics, vol 10. Oxford: Oxford University Press, pp.
160–87.
Dreier, James 1990. “Internalism and Speaker Relativism,” Ethics, vol. 101, pp. 6–26.
Dreier, James 2009. “Relativism (and Expressivism) and the Problem of Disagreement,” Philosophical
Perspectives, vol. 23, pp. 79–110.
Finlay, Stephen 2006. “The Reasons That Matter,” Australasian Journal of Philosophy, vol. 84, pp. 1–20.
Finlay, Stephen 2014. Confusion of Tongues: A Theory of Normative Language. New York: Oxford University
Press.
Francén Olinder, Ragnar 2013. “Moral Relativism, Error‐Theory, and Ascriptions of Mistakes,” Journal
of Philosophy, vol. 110, pp. 564–80.
Jackson, Frank 1985. “On the Semantics and Logic of Obligation,” Mind, vol. 94, pp. 177–95.
Khoo, Justin, and Joshua Knobe 2016. “Moral Disagreement and Moral Semantics,” Noûs, vol. 52, pp.
109–43.
Lyons, David 1976. “Ethical Relativism and the Problem of Incoherence,” Ethics, vol. 86, pp. 107–21.
Mackie, John L. 1977. Ethics: Inventing Right and Wrong. London: Penguin.
Montminy, Martin 2007. “Moral Contextualism and the Norms for Moral Conduct,” American
Philosophical Quarterly, vol. 44, pp. 1–13.
Pittard, John, and Alex Worsnip 2017. “Metanormative Contextualism and Normative Uncertainty,”
Mind, vol. 126 pp. 155–93.
Price, Anthony W. 2008. Contextuality in Practical Reason. Oxford: Oxford University Press.
Sarkissian, Hagop, John Park, David Tien, Jennifer Wright, and Joshua Knobe 2011. “Folk Moral
Relativism,” Mind and Language, vol. 26, pp. 482–505.
Thomson, Judith Jarvis 2008. Normativity. Chicago: Open Court.
Timmons, Mark 1999. Morality without Foundations: A Defense of Ethical Contextualism. New York: Oxford
University Press.
Wedgwood, Ralph 2007. The Nature of Normativity. Oxford: Oxford University Press.
FURTHER READINGS
Beebe, James R. 2010. “Moral Relativism in Context,” Noûs, vol. 44, pp. 691–724.
Björnsson, Gunnar, and Stephen Finlay 2010. “Metaethical Contextualism Defended,” Ethics, vol. 121,
pp. 7–36.
Björnsson, Gunnar 2017. “The significance of ethical disagreement for theories of ethical thought and
talk,” in Tristram McPherson and David Plunkett (eds.), Routledge Handbook of Metaethics, New York:
Routledge, pp. 275–91.
Brogaard, Berit 2008. “Moral Contextualism and Moral Relativism,” Philosophical Quarterly, vol. 58, pp.
385–409.
Dowell, Janice L. 2013. “Flexible Contextualism about Deontic Modals: A Puzzle about Information‐
Sensitivity,” Inquiry, vol. 56, pp. 149–78.
Enoch, David 2011. Taking Morality Seriously. Oxford: Oxford University Press.
Kratzer, Angelika 1977. “What ‘Must’ and ‘Can’ Must and Can Mean,” Linguistics and Philosophy, vol. 1,
pp. 337–55.
Plunkett, David, and Timothy Sundell 2013. “Disagreement and the Semantics of Normative and
Evaluative Terms,” Philosopher's Imprint, vol. 13, pp. 1–37.
Silk, Alex 2016. Discourse Contextualism: A Framework for Contextualist Semantics and Pragmatics. Oxford: Oxford
University Press.