¹ What I am calling ‘Hedonism’ is thus not what Thomson in Goodness and Advice (Princeton, 2001) calls ‘Hedonism about Goodness’ – the theory that an event is good if and only if it is some pleasure (p. 10) or rather includes net pleasure (pp. 9, 11). Thomson introduces ‘Hedonism about Goodness’ as a theory of the goodness of events, but it would seem to be about their intrinsic goodness instead. Hedonism (my kind) can call something ‘intrinsically’ good if it is good apart from whatever it represents beyond itself. This modifier intrinsically is kin to in the short run and ceteris paribus. Hence while every increment of intrinsic goodness is an increment of goodness, intrinsically good things need not be good. Note that on Hedonism the whole of what something good represents is good, not just intrinsically good.

² Thomson, Judith, ‘Goodness and Utilitarianism’, Proceedings and Addresses of the American Philosophical Association 67.4 (Jan. 1994)CrossRefGoogle Scholar.

³ Geach, Peter, ‘Good and Evil’, Analysis 17 (1956)Google Scholar.

⁴ Aristotle, Nicomachean Ethics 1096a23–29.

⁵ The defense of Hedonism may also defend similar theories that aim at different outcomes, such as preference-satisfaction. But pleasure seems simpler.

⁶ Similarly, a beer buzz may be intrinsically good – may, apart from any effects beyond itself, cause net pleasure – as opposed to no buzz, but not as opposed to a bigger buzz. The reader who thinks the straw challenge errs about goodness and the reply errs about causation might check to see whether her demurrals match.

⁷ Nicomachean Ethics 1096a23–29; compare Eudemian Ethics 1217b25–35 and Topics 107a3–12. Ackrill argues for his reading in ‘Aristotle on “Good” and the Categories’, Islamic Philosophy and the Classical Tradition, ed. S. M. Stern, A. Hourani and V. Brown (Oxford, 1972).

⁸ ‘Good and Evil’, p. 34.

⁹ Foot, Philippa, Natural Goodness (Oxford, 2001), p. 2 n. 4CrossRefGoogle Scholar.

¹⁰ Slote, Michael, ‘Object-Utilitarianism’, Pacific Philosophical Quarterly 66 (1985)CrossRefGoogle Scholar.

¹¹ Rawls, John, A Theory of Justice (Cambridge, Mass., 1971), p. 399 (changing ‘X’ to ‘K’)Google Scholar. It is unclear what range of significant phrases of the form ‘good K’ Rawls means his formula to cover.

¹² Rawls takes rationality to be thoughtful coherence of preferences, so that what is rational to want is what would, in one's best judgment (given one's ignorance), serve one's aims, or the aims one would have if rational (Theory, pp. 143, 397ff.). But while Rawls's formula is sensitive to the difficulty of knowing the effects of A's properties F and G, it is insensitive to the difficulty of knowing whether A has F and G.

¹³ ‘Other Ks’ is vague. ‘Most Ks’ would be wrong. Thomson points out in ‘Goodness and Utilitarianism’ that a good K need not be better, nor even a better K, than most Ks. ‘[A] big K is just a K that is bigger than most Ks, and it therefore cannot be the case that most Ks are big Ks. By contrast, a good K is not a K that is better than most Ks. It is not, but could have been the case that most, or even all, people are good people, and that most, or even all, hammers are good hammers’ (p. 12). (Of course, a big K is not exactly a K that is bigger than most. If there were just two sizes of fork, with more big than small ones, no big fork would be bigger than most. And since most fish are very small ones, some small fish are bigger than most fish. But these are details.)

Against Thomson we might object that the intuition she appeals to rates the goodness of Ks in other possible worlds according to the Ks in our world; and if size be rated similarly, then it could have been that all fleas are big ones. But Thomson can reply that even in our world there are good nails that are not better nails than most. Hence ‘good’ is not like ‘big’, if a big K is a K that is a bigger one than most.

Thomson's observation reflects the fact that commonly a normal or standard K is a good K. That fact poses a problem for the enterprise of defining a good K against other Ks as alternatives. For if to be a good K is to be (in some respect) good as opposed to the normal or standard or average K, then the normal or standard or average K must not be a good K. Perhaps the solution is that each K gets credit for some notional cost of replacing it with another. Any nail is a good one if having it as opposed to others is good (in relevant ways), given that one already has it and not the others. For a nail is worse than nothing if replacing it is worthwhile. Or perhaps the solution is that we are to divide Ks (actual and possible) into good and bad ones, or into good, bad and indifferent ones, at whatever points on the scale draw the most useful and salient distinctions.

¹⁴ We can distinguish properties from circumstances even for events, as in ‘good cry’.

¹⁵ But there are bad earthquakes. If we write ‘bad’ for ‘good’ throughout the definiens of the Crude Account, including the formulation of criterion (b) for standard scenarios, then we get an account that seems to capture ‘bad earthquake’ well enough.

¹⁶ Dispute about whether a deft nasty painting is a good painting can reflect disagreement about consequences, as well as a certain flexibility in the role of the noun in ‘good painting’: the skill displayed in a painting can count toward its being a ‘good painting’, as discussed below.

¹⁷ ‘Good and Evil’, p. 41.

¹⁸ Foot, Philippa, ‘Utilitarianism and the Virtues’, Mind 94 (1985)Google Scholar.

¹⁹ What knowledge counts as being reasonably available? Some common standpoint-relative terms such as ‘indicates’ seem to be very flexible in allowing different contexts to specify different degrees of objectivity.

²⁰ Standpoint-relative Hedonism allows for simple goodness if standpoints can be defined into objects of evaluation – X as from our standpoint. And of course it allows us to hypothesize superhuman standpoints, and standpoints that are unspecific as to cognitive resources.

²¹ For comparison, consider two possible limitations of Rawls's theory that a ‘good K’ is such as to be ‘rational to want, given what Ks are used for, or expected to do’. The first is that it may not explain which institutions or actions are good ones. The second is that it may be wrong about which batches of crack are good ones. To see the latter point, note that by ‘given what Ks are used for . . .’, i.e. given that Ks are used for Z, Rawls might mean given that (a) Z is what is rational to want from Ks, or (b) Z is what is incorrigibly wanted from Ks, or (c) Z is what is actually wanted from Ks, or (d) Ks are in fact used for Z. But each version is problematic. For in fact the rational crack to want in the standard case is aversive or vanishingly weak, and so not good crack. Hence if K is crack and Z is a nice high, (a) and arguably (b) are false of the standard case, and (c) and (d) pick out the wrong crack. To help Rawls handle crack one might amend his account, replacing rationality by some even thinner thoughtfulness. Or one might try to give a unified account that overcomes both limitations mentioned above, as follows. (1) Find a way to conceive reasonableness (involving respect and concern for others) and rationality as full and thin versions of one thing (as Rawls does not: Political Liberalism (New York, 1996), pp. 51–2). And (2) propose that to be good is to be reasonable to want, and that the standard standpoints for ‘murder weapon’ and ‘crack’ involve attenuated capacities for reasonableness and even for rationality, respectively.

²² Sen, Amartya, ‘Rights and Agency’, Philosophy and Public Affairs 11 (1982)Google Scholar, considers the possible non-epistemic agent-relativity of intrinsic or ultimate goodness. By contrast, our three relative Hedonisms make intrinsic goodness standpoint-relative in the sense that we may reasonably differ about how much pleasure a certain festival or future includes.

²³ Perhaps, though, we would apply this version's concept of goodness to some other things easily enough. What would one be predictably causing if one caused an earthquake? It might be natural to leave out of account details about the hypothetical causer (whether Smith understands the consequences of earthquakes, and whether Jones could cause an earthquake only with a nuclear bomb).

²⁴ Thomson, ‘On Some Ways’, p. 101.

²⁵ A rational desirability theory of goodness can accommodate indicative goodness by holding that it is incoherent to want something without wanting to hear it coming.

²⁶ The evidential approach may have an advantage in handling coordination problems; for my deciding in a certain way is some evidence for me that others do too, and so it promises that they do. For debate on this and other points, see Paradoxes of Rationality and Cooperation: Prisoner's Dilemma and Newcomb's Problem, ed. R. Campbell and L. Sowden (Vancouver, 1985).

²⁷ ‘Goodness and Utilitarianism’, pp. 8–12. See also Thomson, ‘On Some Ways in Which a Thing Can Be Good’, Social Philosophy and Policy 9 (1992); ‘Moral Objectivity’, Moral Relativism and Moral Objectivity, ed. G. Harman and J. Thomson (Oxford, 1996); ‘The Right and the Good’, Journal of Philosophy 94 (1997); Goodness and Advice (cited earlier); and ‘The Legacy of Principia’ and ‘Reply to Sinnott-Armstrong’, both in The Southern Journal of Philosophy 41 (2003). Thomson holds that good is ‘incomplete’. Her other examples of incompleteness – interested (in what?), knowledgeable (on what?), efficient (at what?), equal (in what?) – suggest that she would not rule out defining goodness by a single account with unbound variables (‘Goodness and Utilitarianism,’ pp. 11–12). See Christian Piller's discussion in ‘Ways of Being Good’, Acta Analytica 16 (2001).

²⁸ ‘Goodness and Utilitarianism’, pp. 8–12. For Thomson's views on the possible relations between the two kinds of ways of being good, see ‘Goodness and Utilitarianism’, p. 11; ‘Moral Objectivity’, pp. 144–52; ‘The Right and the Good’, passim; and Goodness and Advice, pp. 43–82.

²⁹ Even when we seem to find it clear that Alfred is a better person than Bertha, we may fall into doubt by thinking of Bertha's hard circumstances. Such doubt may reflect the fuzziness of the line between properties, as counting towards one's being a good person, and less permanent or integral features, as not counting.

³⁰ ‘Goodness and Utilitarianism’, p. 9.

³¹ For Thomson's own analyses see ‘Moral Objectivity’, pp. 133–40.

³² This doctrine need not imply that it is good for Smith to drug her into permanent mindless euphoria. For if doing so makes her less a person and so less Smith, then it may make the greater pleasure less Smith's and so decrease Smith's net pleasure.

³³ Thomson thinks the two phrases are not parallel: ‘Moral Objectivity’, pp. 140–44.

³⁴ Other topics include Smith's good, her quality of life, her well-being, etc.

³⁵ Hedonism is not very plausible if it implies that even the best of us have been evaluating their options by unreasonable methods. Hence, for Hedonism to be plausible, calculative deliberation must be self-defeating in this sense: that switching over from the sorts of deliberating good non-Hedonists do, towards highly calculative inquiry into the hedonic implications of one's options, must normally recognizably yield choices that are worse, given that goodness is what Hedonism says it is. That is, the best available practical method of judging the goodness of our options, if goodness is what Hedonism says it is, must privilege other kinds of thinking, including reliance on sensibility and testimony. I hope on another occasion to show that this is the case.

³⁶ For example, see Ross's, W. D. use of ‘good for’ in The Right and the Good (Oxford, 1930), pp. 34–5Google Scholar, quoted and discussed by Thomson, ‘Goodness and Utilitarianism’, p. 20 n. 10.

³⁷ Against something like this claim about (2), Philippa Foot writes, ‘[W]hen I was told by a certain philosopher who wanted to explain “good” in terms of choices, that the good roots of trees were roots of the kind “we should choose if we were trees,” that finally confirmed my suspicion of the kind of philosophy that was his’ (Natural Goodness, pp. 25–6; cf. p. 1 n. 1). But if, as seems likely, people who are asked ‘What sorts of roots would plants choose if they could?’ do describe the right sorts of roots, then the account is not relevantly absurd.

³⁸ ‘Goodness and Utilitarianism’, p. 12.

³⁹ Many thanks to Alfonso Gomez-Lobo, Ellen Kline, Christine Korsgaard, Adam Kovach, T. M. Scanlon and David Schmitz for helpful comments.