Abstract
This paper presents a puzzle about moral responsibility. The problem is based upon the indeterminacy of relevant reference classes as applied to action. After discussing and rejecting a very tempting response I propose moral contextualism instead, that is, the idea that the truth value of judgments of the form “S is morally responsible for x” depends on and varies with the context of the attributor who makes that judgment. Even if this reply should not do all the expected work it is a first step.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Another question: don’t we attribute responsibility even in cases of (moral) bad luck, that is, also in cases like (1)? Perhaps, but this kind of responsibility is different from the one discussed here. Thanks to a referee for pressing me both on this point and the one above.
So, the following reverse principle is false:
(Reverse RESP) If there is a sufficiently high (conditional) probability that a consequence will happen, given an action, then the agent is responsible for the consequence of their action.
On why (Reverse RESP) is not true, see also the following remarks.
The intention is, of course, crucial for the description of the act: Saying that Mafioso M was just doing his job underdescribes M’s intentional killing of another person.
Perhaps it is knowledge which is required or justified belief or something else; similar things will hold then, mutatis mutandis. For more on this, see below.
Here is an additional, graded principle (leaving aside, again, further precisifications like, e.g., ceteris paribus clauses):
(RESP*) The higher the (conditional) probability that a certain consequence will happen, given a certain action, the greater the responsibility of an agent for the consequence of their action (should it happen).
A referee presented me with the following objection. Suppose A has a gun with one bullet in its six chambers, knows this but doesn’t know which chamber is “loaded”, aims at a person, pulls the trigger and shoots and kills that person. Suppose further that B has a gun with five bullets in it, knows this but doesn’t know which chamber is empty, aims at someone, pulls the trigger and shoots and kills that person. According to (RESP*), B is more responsible for murder than A. But aren’t they both responsible, and to the same degree? No. To see more clearly why not consider the following (parallel) cases. Mary knows that there is a 1-in-a-Million chance that her porch will collapse if she has her neighbors over to sit on it with her; she invites them over, they come over, sit with her on the porch and, alas, the porch collapses. Chuck knows that there is a 50% chance that his porch will collapse if he invites his neighbors over to sit on it with him; he invites them over, they come over, sit with him on the porch and—no big surprise—the porch collapses. It seems simply false to say that both Mary and Chuck are responsible for what happened on their porch and that they are responsible to the same degree. Since the shooting cases above are structurally identical, the same things hold in those cases. (RESP*) makes the correct prediction. For the sake of simplicity, I will stick with the first, non-graded version of the principle. Nothing essential hinges on this.
Thanks to Tony Milligan who pressed me on this point.
For purely expository reasons, I did not include this identity in the initial description of (1) and (2) above. However, nothing substantial depends on this “surprise” strategy of mine. The problem does not go away even given that the driver is aware of the fact that (1) = (2). He (like us) still has to choose between thinking about the event in terms of Z-street or in terms of the X-neighborhood. This should become even clearer below. Hence, it is also not necessary to “update” the probabilities for (1) or (2) upon mentioning that (1) = (2).
For a short mention of the problem see my 2008.
To avoid misunderstandings: This example is about probabilities to live to a certain age and as such not confined to subjective expectations to get to that age.
The reference class problem arises for all the different interpretations of "probability". See Hájek (2007).
The question what belongs to the act and what to the circumstances need not worry us here. No matter how we "slice the cake" and no matter whether our problem is to identify the relevant description of the circumstances or the relevant description of the act, we have to deal with the same basic problem of the relevant reference class.
See also Greco (2006) for a nice series of cases (involving drunk drivers) which illustrates a closely related problem concerning moral luck.
A contextualist response like the one proposed below does not count as a “solution” in the strict sense which would enable us to indicate unique relevant reference classes in a straightforward way, that is, without further relativization or contextualization. It rather constitutes a “sceptical solution” which accepts the initial problem and does not dissolve it. It is in this sense only that the proposal below is meant as a “solution”.
Reichenbach restricts the conditions to those for which we do have statistical information concerning their relation to E. However, if we want to know what really is the relevant reference no matter the limitations of our information, then we have to choose the formulation in the text above.
Talk about "the probability of single cases" is ambiguous here. It can either mean the probability that a certain combination of properties which is only instantiated once is F. Or it can mean the probability that a certain individual a is F. Fortunately, this difference does not matter here.
Thanks to a referee here.
I am using the word “truth” in a minimal sense here, that is, in a sense that does not imply any commitment to moral realism. Contextualism about “knowledge” holds that the truth conditions of sentences of the form “S knows that p” vary with the context of the speaker of an utterance of a sentence of that form. For epistemic contextualism about terms like “knowledge” (cf., e.g., Cohen 1987; DeRose 1999; Lewis 1996).
Thanks to a referee for pressing me on this. For more on error theories, see below.
For subject-sensitive invariantism about epistemic terms like “knowledge” (see Hawthorne 2004; Stanley 2005). According to this view, knowledge does not only depend on epistemic aspects of the subject’s situation (evidence, belief, etc.) but also on non-epistemic factors (like what is at stake for the subject). Another form of invariantism not to be discussed here is truth-relativism: the same content can be true for one assessor and not true for another assessor (see MacFarlane 2005).
This view is sometimes also called “contextualism”—because of the relevance of further, traditionally neglected factors in the situation of the subject. In the early days of the discussion of epistemic contextualism, people distinguished between “subject-contextualism” and “attributor-contextualism” (see for this DeRose 1992). Only the later view is nowadays called “contextualism”. We should stick with the parallel restriction in the moral case and use “contextualism” in the restricted sense explained above.
Dreier (1990), Norcross (2005a), Brogaard (2003, 2008), Greco (2008), Wedgwood (2006b), Montminy (2007) and Jenkins and Nolan (2010) propose contextualism for core moral terms (see also Unger 1995 and critically: Weatherson 2008, Sect. 4 and Schroeder 2009, 284–287). There is also growing debate on the context-sensitivity of normative ought-claims: (see e.g., Björnsson and Finlay 2010, esp. the first two sections and Wedgwood 2006a, 151–152). Hawthorne (2001) and Rieber (2006) have tried to apply attributor contextualism to the concept of a free action while Norcross (2005b) defends a contextualist analysis of "harm". Sinnott-Armstrong (2006) defends contrastivism with respect to justified moral belief but does not hold a contextualist version of contrastivism.
One could propose a kind of “relationalism” according to which there is a further parameter implicit in responsibility judgments, namely a parameter for reference classes. This would still qualify as a kind of contextualism.
For a discussion of legal implications of the reference class problem, see Colyvan et al. (2001).
According to some theories of epistemic contextualism, it is the conversational context and what is being mentioned in a conversation which determines what is salient to the participants (and thus also the contextually relevant epistemic standards; see Lewis 1996). This is less plausible in moral cases like our example. Salience is rather determined by other, psychological, factors.
Thanks to a referee who pressed me on the last two points. One more issue. As soon as the attributor becomes aware of the indeterminacy of the relevant reference class and of the relativity of their own view to their own interests and perspectives, something similar to Moore-paradoxality and incoherence threatens: “A is suffering from bad luck—but that’s just me and my perspective on things!” sounds self-defeating. It seems that the attributor can only suspend judgment under these conditions. See, however, the above remarks on error theories.
References
Baumann, P. (2008). Problems for Sinnott-Armstrong’s moral contrastivism. The Philosophical Quarterly, 58, 463–470.
Beebe, J. R. (2004). The generality problem, statistical relevance and the tri-level hypothesis. Noûs, 38, 177–195.
Björnsson, G., & Finlay, S. (2010). Metaethical contextualism defended. Manuscript.
Brogaard, B. (2003). Epistemological contextualism and the problem of moral luck. Pacific Philosophical Quarterly, 84, 351–370.
Brogaard, B. (2008). Moral contextualism and moral relativism. The Philosophical Quarterly, 58, 385–409.
Cohen, S. (1987). Knowledge, context, and social standards. Synthese, 73, 3–26.
Colyvan, M., Regan, H. M., & Ferson, S. (2001). Is it a crime to belong to a reference class? The Journal of Political Philosophy, 9, 168–181.
DeRose, K. (1992). Contextualism and knowledge attributions. Philosophy and Phenomenological Research, 52, 913–929.
DeRose, K. (1999). Contextualism: An explanation and defense. In J. Greco & E. Sosa (Eds.), The Blackwell guide to epistemology (pp. 187–205). Oxford: Blackwell.
DeRose, K. (2009). The case for contextualism. Knowledge, skepticism, and context (Vol. I). Oxford: Clarendon.
Dreier, J. (1990). Internalism and speaker relativism. Ethics, 101, 6–26.
Fetzer, J. H. (1977). Reichenbach, reference classes, and single case probabilities. Synthese, 34, 185–217.
Gillies, D. (2000). Philosophical theories of probability. London: Routledge.
Greco, J. (2006). Virtue, luck and the Pyrrhonian problematic. Philosophical Studies, 130, 9–34.
Greco, J. (2008). What’s wrong with contextualism. The Philosophical Quarterly, 58, 416–436.
Hájek, A. (2007). The reference class problem is your problem too. Synthese, 156, 563–585.
Hawthorne, J. (2001). Freedom in context. Philosophical Studies, 104, 63–79.
Hawthorne, J. (2004). Knowledge and lotteries. Oxford: Clarendon.
Heller, M. (1995). The simple solution to the problem of generality. Noûs, 29, 501–515.
Heller, M. (1999). The proper role for contextualism in an anti-luck epistemology. Philosophical Perspectives, 13, 115–130.
Hempel, C. G. (1965). Aspects of scientific explanation and other essays in the philosophy of science. New York/London: Free Press/Collier.
Jenkins, C. S., & Nolan, D. (2010). Maximising, satisficing and context. Noûs, 44, 451–468.
Lewis, D. (1996). Elusive knowledge. Australasian Journal of Philosophy, 74, 549–567.
Lewis, D. (1998). The paradoxes of time travel. In P. van Inwagen & D. W. Zimmerman (Eds.), Metaphysics: The big questions (pp. 159–169). Oxford: Blackwell.
MacFarlane, J. (2005). The assessment sensitivity of knowledge attributions. In T. Szabó Gendler & J. Hawthorne (Eds.), Oxford studies in epistemology (Vol. 1, pp. 197–233). Oxford: Oxford University Press.
Mackie, J. (1977). Ethics. Inventing right and wrong. New York: Penguin.
Montminy, M. (2007). Moral contextualism and the norms for moral conduct. American Philosophical Quarterly, 44, 1–13.
Nagel, T. (1979). Moral luck. In T. Nagel (Ed.), Mortal questions (pp. 24–38). Cambridge: Cambridge University Press.
Norcross, A. (2005a). Contextualism for consequentialists. Acta Analytica, 20-2(35), 80–90.
Norcross, A. (2005b). Harming in context. Philosophical Studies, 123, 149–173.
Reichenbach, H. (1949). The theory of probability. An inquiry into the logical and mathematical foundations of the calculus of probability. Berkeley & Los Angeles: University of California Press.
Rieber, S. (2006). Free will and contextualism. Philosophical Studies, 129, 223–252.
Salmon, W. C. (1966). The foundations of scientific inference. Pittsburgh, PA: University of Pittsburgh Press.
Schiffer, S. (1996). Contextualist solutions to scepticism. Proceedings of the Aristotelian Society, 96, 317–333.
Schroeder, M. (2009). Hybrid expressivism: Virtues and vices. Ethics, 119, 257–309.
Sinnott-Armstrong, W. (1996). Moral skepticism and justification. In W. Sinnott-Armstrong & M. Timmons (Eds.), Moral knowledge? New readings in moral epistemology (pp. 3–48). Oxford: Oxford University Press.
Sinnott-Armstrong, W. (2006). Moral skepticisms. Oxford: Oxford University Press.
Stanley, J. (2005). Knowledge and practical interests. Oxford: Clarendon.
Thomas, A. (2006a). Value and context. The nature of moral and political knowledge. Oxford: Oxford University Press.
Thomas, A. (2006b). Inferential contextualism and moral cognitivism. Manuscript.
Timmons, M. (1993). Moral justification in context. Monist, 76, 360–378.
Timmons, M. (1996). Outline of a contextualist moral epistemology. In W. Sinnott-Armstrong & M. Timmons (Eds.), Moral knowledge? New readings in moral epistemology (pp. 293–325). Oxford: Oxford University Press.
Timmons, M. (1999). Morality without foundations: A defense of ethical contextualism. Oxford: Oxford University Press.
Unger, P. (1995). Contextual analysis in ethics. Philosophy and Phenomenological Research, 55, 1–26.
Weatherson, B. (2008). Attitudes and relativism. Philosophical Perspectives, 22, 527–544.
Wedgwood, R. (2006a). The meaning of ‘ought’. In R. Shafer-Landau (Ed.), Oxford studies in metaethics (Vol. 1, pp. 127–160). Oxford: Oxford University Press.
Wedgwood, R. (2006b). The context-sensitivity of ‘ought’, and the normativity of rationality. Manuscript.
Williams, B. (1981). Moral luck. In B. Williams (Ed.), Moral luck. Philosophical papers 1973–1980 (pp. 20–39). Cambridge: Cambridge University Press.
Acknowledgments
I am grateful to conversations with and comments by a number of people (and I hope I’m not forgetting anyone): Michael Clark, Christoph Fehige, John Greco, Tony Milligan, Bob Plant, Walter Sinnott-Armstrong, Barry Smith, anonymous referees as well as audiences at universities at Aberdeen, Essen, Glasgow and Stirling.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Baumann, P. A Puzzle About Responsibility. Erkenn 74, 207–224 (2011). https://doi.org/10.1007/s10670-010-9259-6
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10670-010-9259-6