Abstract
This paper explores a problem involved in defining the act utilitarian notion of a “right action” within the framework of stit semantics. The problem is there seem to be two different, and conflicting, ways of defining this idea, both intuitively attractive. Previous work has addressed this problem by developing separate theories to account for our conflicting intuitions. This paper shows that the problem can also be addressed through a single theory that allows actions to assessed from different perspectives.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
In an effort to find language that is both gender neutral and unobtrusive, I assume here that the agents are impersonal acting devices, such as robots, which it is appropriate to refer to using the pronoun “it”.
- 2.
- 3.
- 4.
A discussion of this terminology can be found, for example, in Sections 2.1 and 13.1 of Luce and Raiffa (1957). Of course, the legitimacy of the distinction between uncertainty and risk is itself an issue: following Ramsey (1931) and Savage (1954), many writers in the Bayesian tradition assume that an agent’s assessment of the possible outcomes in a given situation can always be represented through a probability measure , so that uncertainty always reduces to risk. However, there is an important tradition of resistance to the assimilation of uncertainty and risk in a single numerical measure. A classic paper in this tradition is Ellsberg (1961); for more recent work on decision theory in situations that mix elements of risk and uncertainty, see the papers contained in Parts II and IV of Gärdenfors and Sahlin (1988).
- 5.
This pattern of reasoning is first explicitly characterized as the “sure-thing principle” in Savage (1954), but the principle appears already in some of Savage’s earlier work, such as (Savage 1951, p. 58), where he writes concerning situations of uncertainty that “there is one unquestionably appropriate criterion for preferring some act to some others: If for every possible state, the expected income of one act is never less and is in some cases greater than the corresponding income of another, then the former act is preferable to the latter.”
- 6.
Regan does not actually require that these choices must be simultaneous (though simultaneity is part of Gibbard’s earlier description), but he does require the choices to be independent, and we guarantee independence through simultaneity.
- 7.
Gibbard adopts a similar viewpoint in his original discussion of this example, evaluating each agent’s selection only under an assumption about the action selected by the other (Gibbard 1965, p. 215). And Sobel defends Gibbard’s strategy as follows: “It is perhaps natural to feel that Gibbard’s first case is objectionable just because it includes assumptions concerning what agents will and would do. But this can be no objection since it is obvious that such assumptions are essential to the application of AU; without such assumptions the dictates of AU could not be determined …” (Sobel 1968, p. 152).
- 8.
These definitions may seems to be needlessly general, but please bear with me; the generality will help us later on.
- 9.
- 10.
This problem was originally presented in a trio of papers: Goldman (1976), Sobel (1976), and Thomason (1981). Further discussion can be found, for example, in Bergström (1977), Carlson (1995), Feldman (1986), Goldman (1978), Greenspan (1978), Humberstone (1983), Jackson (1985), Jackson (1988), Jackson and Pargetter (1986), McKinsey (1979), and Zimmerman (1990).
- 11.
- 12.
- 13.
Although this point is “visually obvious,” it actually relies on the technical constraint of “no choice between undivided histories,” not discussed in this chapter, according to which histories that are still undivided at a given moment cannot be separated at that moment by the Choice partition.
- 14.
In the group case, this fact actually follows from the previous definitions of State Γ m as the set of states confronting Γ at m and \(Choice_{\varGamma}^{m}/X\) as the actions available to Γ that are consistent with X. However, it is set out separately here in order to conform to our treatment of the individual case, where the corresponding notion must be introduced through a definition.
References
Belnap N (2001) Double time references: speech act reports as modalities in an indeterministic setting. In: Wolter F, Wansing H, de Rijke M, Zakharyaschev M (eds) Advances in modal logic, vol 3, CSLI Publications, Stanford pp 1–21
Belnap N (2004) Future contingents and the sea battle tomorrow, manuscript, Philosophy Department, University of Pittsburgh
Belnap N, Perloff M, Xu M (2001) Facing the future: Agents and choices in our indeterminist world. Oxford University Press, Oxford
Bergström L (1966) The alternatives and consequences of actions, Stockholm Studies in Philosophy, vol 4. Almqvist and Wiksell, Stockholm
Bergström L (1977) Utilitarianism and future mistakes. Theoria 43:84–102
Carlson E (1995) Consequentialism reconsidered, theory and decision library, series A: Philosophy and methodology of the social sciences, vol 20. Kluwer Academic Publishers, Dordrecht
Ellsberg D (1961) Risk, ambiguity, and the Savage axioms. Quarterly Journal of Economics 75:643–669
Feldman F (1986) Doing the best we can: An essay in informal deontic logic. D. Reidel Publishing Company, Dordrecht
Gärdenfors P, Sahlin NE (eds) (1988) Decision, probability, and utility: Selected readings. Cambridge University Press, Cambridge
Gibbard A (1965) Rule-utilitarianism: merely an illusory alternative? Australasian Journal of Philosophy 43:211–220
Goldman H (1976) Dated rightness and moral imperfection. The Philosophical Review 85:449–487
Goldman H (1978) Doing the best one can. In: Goldman AI, Kim J (eds) Values and morals, D. Reidel Publishing Compamy, Dordrecht, pp 185–214
Greenspan P (1978) Oughts and determinism: a response to Goldman. Philosophical Review pp 77–83
Horty J (2001) Agency and deontic logic. Oxford University Press, Oxford
Humberstone IL (1983) The background of circumstances. Pacific Philosophical Quarterly 64:19–34
Jackson F (1985) On the semantics and logic of obligation. Mind 94:177–195
Jackson F (1987) Group morality. In: Pettit P, Sylvan R, Norman J (eds) Metaphysics and morality: Essays in honour of J. J.C. Smart, Basil Blackwell Inc., Oxford pp 91–110
Jackson F (1988) Understanding the logic of obligation. In: Proceedings of the Aristotelian Society, Supplementary Volume 62, Harrison and Sons
Jackson F, Pargetter R (1986) Oughts, options, and actualism. Philosophical Review 99:233–255
Luce RD, Raiffa H (1957) Games and decisions. John Wiley and Sons, New York
MacFarlane J (2003) Future contingents and relative truth. Philosophical Quarterly 53:321–336
McKinsey M (1979) Levels of obligation. Philosophical Studies 35:385–395
Parfit D (1984) Reasons and persons. Oxford University Press, Oxford
Prior A (1956) The consequences of actions. In: Proceedings of the Aristotelian Society, Supplementary Volume 30, Harrison and Sons
Prior A (1967) Past, present, and future. Clarendon Press, Oxford
Ramsey F (1931) Truth and probability. In: Braithwaite RB (ed) The foundations of mathematics and other logical essays, Routledge and Kegan Paul, Oxford pp 156–191, originally published in 1926
Regan D (1980) Utilitarianism and co-operation. Clarendon Press, Oxford
Savage L (1951) The theory of statistical decision. Journal of the American Statistics Association 46:55–67
Savage L (1954) The foundations of statistics. John Wiley and Sons, New York second revised edition published by Dover Publications, 1972
Sobel JH (1968) Rule-utilitarianism. Australasian Journal of Philosophy 46:146–165
Sobel JH (1976) Utilitarianism and past and future mistakes. Nous 10:195–219
Thomason R (1970) Indeterminist time and truth-value gaps. Theoria 36:264–281
Thomason R (1981) Deontic logic and the role of freedom in moral deliberation. In: Hilpinen R (ed) New Studies in Deontic Logic, D. Reidel Publishing Company, Dordrecht pp 177–186
Zimmerman M (1990) Where did I go wrong? Philosophical Studies 59:55–77
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix: Act Utilitarianism for Groups
Appendix: Act Utilitarianism for Groups
This appendix shows how perspectival act utilitarianism can be extended from individual actions to group actions , and how the relation between the right actions available to groups and individuals can then be seen to depend on the standpoint from which these actions are evaluated.
The extension of perspectival act utilitarianism to group actions is straightforward, involving nothing more than a generalization of several of our previous notions. We have already seen, in the text, how the set Choice Γ m of actions available to the group Γ at the moment m can be defined, with each group action identified as a pattern of actions available to the individuals from that group. The states confronting the group Γ at m can then be defined as the patterns of actions available at m to all agents except those from that particular group :
And where X is some proposition, weak and strong dominance relations under the condition X can be defined among the actions available to a group in a way exactly parallel to the definition for individual actions.
Conditional dominance ordering on group actions : Let Γ be a group of agents and m a moment, and let K and K′ be members of Choice Γ m, and X a proposition. Then \(K \preceq_X K'\) (K′ weakly dominates K under the condition X) if and only if \(K \cap X \cap S \leq K'\cap X \cap S\) for each state \(S \in State_{\varGamma}^{m}\); and \(K \prec_X K'\) (K′ strongly dominates K under the condition X) if and only if \(K \preceq_X K'\) and it is not the case that \(K' \preceq_X K\).
The set of actions available to the group Γ under the condition X can be defined as those among the available actions that are consistent with this condition :
And the optimal actions available to the group under this condition can then be defined as the actions available under this condition that are not dominated under this condition by any other such actions:
Finally, the set of states confronting Γ at m that are consistent with the proposition X can be represented just as before:
Footnote 14 And likewise the proposition that one of these states holds:
Given these materials, we can now introduce a form a perspectival act utilitarianism for groups, according to which an action available to the group Γ at a moment m is right from the standpoint of the moment m′ just in case that action is optimal under the conditions in which the group Γ finds itself at m, where these conditions are judged from the standpoint of m′:
Perspectival act utilitarianism for groups: Let Γ be a group of agents and m and m′ moments such that either \(m < m'\) or \(m' < m\) or \(m = m'\), and suppose \(K \in Choice_{\varGamma}^{m}\). Then the action K is right at m from the standpoint of m′ if and only if \(K \in Optimal_{\varGamma}^{m}/State_{\varGamma}^{m}(H_{m'})\), and wrong otherwise.
As with individual actions, this perspectival account supports the orthodox intuitions concerning group actions when the moment m′ of evaluation is later than the moment m of action, while the dominance intuitions are supported when m′ is earlier than or identical with m.
Now that the perspectival account has been extended from individuals to groups, let us turn briefly to two of the most central questions concerning the relation between individual and group act utilitarianism. First, if each individual belonging to a group performs a right action, does that entail that the group itself performs a right action? And second, if a group performs a right action, does that entail that the individuals belonging to the group do so?
The answer to the first question is No. This fact is well-known and can be illustrated with the Whiff and Poof example from Fig. 11.6, which was originally formulated to make exactly this point. Still, it is useful to consider the question separately from the dominance and orthodox perspectives, since the contours of this negative answer differ.
Suppose, first, that we evaluate the actions available at the moment m in Fig. 11.6 from the standpoint of m itself, adopting the dominance perspective. Then it is easy to verify that each action available to either agent is classified as right from the standpoint of m. So suppose that Whiff pushes his button, performing the action K 1, while Poof refrains, performing K 4 – each agent therefore performing an action that is classified as right. Then the group \(\varGamma = \{\alpha, \beta \}\) containing both Whiff and Poof performs the action \(K_1 \cap K_4\), which is clearly non-optimal, leading to a utility of 0 while 10 is possible, and so classified as wrong from the standpoint of m. Indeed the group action \(K_1 \cap K_4\) is not even in equilibrium: each agent would be better off with a different choice, given the action chosen by the other. Individual satisfaction of dominance act utilitarianism, then, not only fails to guarantee group satisfaction , but has the even more depressing consequence that the pattern of actions chosen, each right from an individual perspective, may not be an equilibrium pattern .
Next, suppose Whiff and Poof both refrain from pushing their buttons, performing the individual actions K 2 and K 4. The outcome of this pair of actions is the history h 4. So let us evaluate these actions from the standpoint of some later moment along this history, thus adopting the orthodox perspective. It is easy to see that both of these actions are then classified as right from the standpoint of this later moment, and also that the pair of actions is in equilibrium : each agent is performing a best available action, given the actions performed by the other.
This example illustrates the general rule: whenever each individual member of a group of agents performs an action that is right from the standpoint of a later moment – and so right from the orthodox perspective – the pattern of actions performed by the entire group is in equilibrium . However, this does not mean that the group action is itself right. In this case, the group action \(K_2 \cap K_4\) is non-optimal, and so wrong, since it yields a utility of 6 while the available group action \(K_1 \cap K_3\) yields a utility of 10. If each member of a group performs an action that is right from the standpoint of a later moment, then, the overall pattern of actions will be in equilibrium , but it still may not be a right action for that group to perform, since there may be better equilibrium patterns.
Now to the second question: if a group action is right , does it follow from this that the actions of the individuals belonging to that group are also right? The standard answer to this question is Yes. Regan, for example, writes that “for any group of agents in any situation, any pattern of behaviour by that group of agents in that situation which produces the best consequences possible is a pattern in which the members of the group all satisfy AU” (Regan 1980, p. 54). And Jackson, that “if the right group action is actually performed, then that group action’s constituent individual actions must be right” (Jackson 1988, p. 264). In the case of this question, however, the dominance and orthodox perspectives yield different answers.
Both Regan and Jackson adopt the orthodox perspective, evaluating actions from the standpoint of a later moment, and from that perspective what they say is right. In our current language, it can be put like this: if a group action performed at m is right from the standpoint of a later moment m′, then the actions performed by the individual members of that group are also right from the standpoint of m′.
However, the implication fails if we consider the matter from the dominance perspective, evaluating actions from the standpoint of a moment at or before the moment of their performance: where m′ is identical with or earlier than m, it might well be possible that a group action performed at m is right from the standpoint of m′, while the individual action of some member of that group is wrong from the standpoint of m′. This possibility is illustrated in Fig. 11.8. Here, it is easy to see that the action \(K_2 \cap K_3\) performed at the moment m by the group \(\varGamma = \{\alpha,\beta\}\) is right from the standpoint of the moment m itself, since this group action leads to an outcome of utility 1, the highest available, and is therefore optimal. But the component action K 2 by the agent α is wrong from the standpoint of m, since it is dominated by K 1. Of course, from the standpoint of some future moment along the history h 3, we can see the action K 2 by α was performed under circumstances in which β performed the action K 3, so that an outcome of utility 1 was achieved; from this later standpoint, the action K 2 is therefore right. But at the moment m itself, while it is still unclear which action β will perform, the choice of K 2 allows for an outcome of utility 0, and is therefore dominated by K 1, which guarantees an outcome of utility of 1.
Rights and permissions
Copyright information
© 2011 Springer Science+Business Media B.V.
About this chapter
Cite this chapter
Horty, J.F. (2011). Perspectival Act Utilitarianism. In: Girard, P., Roy, O., Marion, M. (eds) Dynamic Formal Epistemology. Synthese Library, vol 351. Springer, Dordrecht. https://doi.org/10.1007/978-94-007-0074-1_11
Download citation
DOI: https://doi.org/10.1007/978-94-007-0074-1_11
Published:
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-007-0073-4
Online ISBN: 978-94-007-0074-1
eBook Packages: Humanities, Social Sciences and LawPhilosophy and Religion (R0)