¹ Ziff, Paul, Semantic Analysis (Ithaca N.Y.: Cornell University Press, 1960)Google Scholar, chapter 6.

² This last is the so-called ‘buck-passing’ theory. See Scanlon, Thomas, What We Owe To Each Other (Harvard University Press, 1998), 95–100Google Scholar.

³ Or so I should claim. See Ethics: Twelve Lectures on the Philosophy of Morality, (Harvard University Press and Penguin Press, 2006), chapters 11 and 12. See also 176–7.

⁴ London: Routledge & Kegan Paul, 1963.

⁵ See page 35 of the intellectual autobiography that he wrote for The Philosophy of Georg Henrik von Wright, edited by P. A. Schilpp and L. E. Hahn (Open Court, 1989).

⁶ See “Good and evil”, Analysis 17 (1956), 33–42.

⁷ Von Wright would have been suspicious of the assumption I am imagining the semanticist making of a simple and self-evident parallelism between a putative variety of kinds of goodness and a putative variety of senses of “good”.

The starting point for any mediation between von Wright and the semanticists is the evident need to supplement von Wright's rather restricted idea of ambiguity (cp. p. 14) and to make room for the thought that a good dictionary should afford just as many different readings of a given word as are needed (neither more nor fewer than are needed) to explicate its contribution at each and every one of its occurrences. So far I should side with the semanticists against von Wright. The measure of polysemy – the measure of “ambiguity” in this lexicographical sense (or of “homonymy” in the sense which, by transposing from the material to the formal mode, we derive from Aristotle's at Categories 1a1–3) – is the number of different headings/sub-headings a good dictionary needs to award to the word and to specific constructions involving the word in order for the dictionary to gloss the meanings of all sentences involving the word.

From this point onwards matters are less straightforward. The test for ambiguity just proposed would be easy to apply if it were (or could be) the business of lexicography to furnish a perfect analysis or a set of perfect analyses for each word in the dictionary. But there are good reasons why this was never to be expected (see my “Three moments in the theory of definition or analysis: its possibility, its aim or aims and its limit or terminus”, Proceedings of the Aristotelian Society, Vol. CVII (2008)). That is the reason why in stating the criterion of ambiguity I had recourse to the contestability-inducing verb “gloss”. Where glossing is concerned, do we want austere economy of means or do we also want generous illumination?

In practice – and now I am inclining to von Wright's cause – the anatomizing of the varieties of goodness has to proceed without looking for decisions concerning the number of entries an authoritative and theoretically perfectly parsimonious dictionary could allow itself. Decisions of that sort must come at the end, not at the outset. Only at the end can it be determined how many different things are meant (at the level of reference/Bedeutung) by the word “good” – and how many different varieties to recognize. Meanwhile let us anatomize varieties without trying to keep a final count.

⁸ At page 19, von Wright points to a difference that he marks by distinguishing “x is as a K good”, which implies that x is indeed a K, from “x is good as a K”, which only implies that (whether or not x is a K) x “can be used to advantage in the way Ks are normally used or performs well enough in the way characteristic of Ks”. A knife with a solid heavy handle can be good as a hammer even though, not being a hammer at all, it is not a good hammer. Not everything which is good for some purpose also belongs to some kind which is essentially associated with that purpose. Will not every theory need then to distinguish instrumental and useful goodness? In further explanation of that distinction, see Varieties 43–4:

⁹ One who restricts the univocity thesis in order to salvage some fragment of it will want Von Wright's assumptions to be spelled out here as they regard the relationship between “good” in “good at” and “good” in “good carpenter” – and then generalized. But at this point let us wrestle no more with an imaginary opponent and wait to hear from him.

This is the moment to make the obvious remark that in further studies of the varieties of goodness it will be necessary to attend much longer and harder to the syntax and semantics of all the varieties of grammatical construction built on the pattern: “is” + “good” + preposition + …

¹⁰ Compare also NE 1129a30 to which we shall return.

¹¹ On these sentences, see Segvic, Heda, “Aristotle on the Varieties of Goodness”, Apeiron 37 (2004), 151–176CrossRefGoogle Scholar. See also Shields, Christopher, Order in multiplicity: homonymy in the philosophy of Aristotle (Clarendon Press, Oxford 1999)Google Scholar. For discussions I have myself attempted of them, see “On sentence sense, word sense and difference of word sense”, in Steinberg, D. and Jacobovits, L., Semantics: an interdisciplinary reader (Cambridge University Press, 1971)Google Scholar. See also 249–253 of my “Replies”, in Essays for David Wiggins: Identity, Truth and Value, eds. Sabina Lovibond and Stephen Williams (Blackwell, 1996), 249. And see equally Stephen Williams's essay.

Aristotle's theory of focal meaning was clarified and made fully available to 20^th century philosophy by Owen, G. E. L.. See his Logic, Science and Dialectic (London: Duckworth, 1986)Google Scholar, especially Essay 10 (dating from 1957).

¹² Note that Aristotle doesn't mean by ‘somehow’ ‘just anyhow related’. For reasons of its own or of its speakers', the language settles upon some ways and refuses semantic recognition to other possibilities. (A medical professor who lectures on health and preventive medicine is not for that reason a healthy professor.) Something similar will apply with “good”.

This is the moment to say that, according to the focalist concerning “good”, undifferentiated good is to be understood not (as by the buck-passing theory of note 2) in terms of a second-level quantification over specific first-level properties, but as a first-level predicate that signifies by means somehow analogous to the possibility appealed to by Russell and Whitehead in Principia Mathematica at the point where they introduce the (essentially schematic) device of typical ambiguity. The analogy suggests that our understanding of undifferentiated good is always something incomplete or in process.

¹³ With this contrast come tangled issues. See Foot, Philippa, Natural Goodness (Oxford University Press, 2001)CrossRefGoogle Scholar, chapters 5 and 6. See also Varieties, 151: “It is of some importance for educationists of a certain bent of mind to remember that virtues are no ends in themselves but instruments in the service of the good of man, and important for educationists of a certain other bent of mind not to lose sight of the fact that it is only by being aware of harmful consequences of yielding to passion that man has a rational ground for aspiring after virtues. For, be it observed in this connexion, we do not commonly and naturally call the virtues “beneficial”. This is significant. The virtues are needed; absence of virtue is a bad thing for us. The goodness of the virtues is that they protect us from harm and not that they supply us with some good.”

¹⁴ Peter Hacker made this second suggestion to me.

¹⁵ On these, see Owen, G. E. L. “Aristotelian Pleasures” Proceedings of the Aristotelian Society 72, 1971–2, 135–52CrossRefGoogle Scholar.

¹⁶ It is not to be denied that some pleasures of this sort do make reference to the goodness of a being, but not all do. For a case where there is a connection of sorts, see David Hume, Enquiry into the Principles of Morals VI.1, Selby-Bigge, 243–4: “… the view [i.e. spectacle] of [happiness], whether in its causes or effects like sunshine or the prospect of well-cultivated plains …, communicates a secret joy and satisfaction” (italics added).

¹⁷ An objector may say: “This all sounds very nice, but what about “good henchman”? What about the good and faithful servants of an evil master. Remember that they too have their vocation, as well as certain skills or expertises. They too have a particular formation of character and will. They too work to no set hours. The elucidation of goodness in a vocation or calling simply does not need the sort of detour which your project has enforced upon you through the notion of the good of a being.” But I reply that the resemblance is very faint. The henchman's commitment is at best narrow and selective. One who is good in a vocation has a commitment to [human] beings at large.

¹⁸ Here I abandon altogether von Wright's approach to good/bad act and good/bad intention. For the difficulties of his approach, see Foot, Philippa's review of Varieties in The Philosophical Review 74 (1965), 240–244CrossRefGoogle Scholar. See also Lars Hertzberg “Georg Henrik von Wright on Goodness and Justice”, in “Georg Henrik von Wright: In Memoriam”, Acta Philosophica Fennica (2006).

¹⁹ In so far as Aristotle's own conception of a good (agathos or spoudaios) person (anthropos) fails to close the question – that issue does not concern us here – there are then at least two different ways for it to be open, the way it is open for us (see section 17) and the way it ought to seem open for Aristotle. In these connections see Glassen, P. ‘A Fallacy in Aristotle's Argument about the Good,’ Phil. Quarterly 7 (1957), 319–322CrossRefGoogle Scholar.

²⁰ I make a similar proposal for the linkage holding between the different senses of the adjective “just”. See Ethics: Twelve Lectures on the Philosophy of Morality, (Harvard University Press and Penguin Press, 2006), 285–290, with note 16.

²¹ See my Twelve Lectures, op.cit., chapters 3, 8, 9 and my Lindley Lecture, “Solidarity and the Root of the Ethical” (Lawrence, Kansas: University of Kansas Philosophy Department, 2008).

²² For von Wright's own view of the virtues see the citations in section 10 and note 13.

²³ For her explanation of the import and importance of this sentence, I am indebted to Heda Segvic, op. cit. 155.

²⁴ In section 22 we shall touch on one further difficulty having to do with the goodness of wisdom.

²⁵ Similarly, when we ask what is so good about something or someone we precisely don't specify in advance which variety we are enquiring about.

²⁶ To call the collar-bone kleis was an admirable piece of linguistic invention. But it fell short of investing the putative determinable kleides with more than etymological interest. Too much was lost in the transposition – a key as something lost or found, a key as leading inwards or onwards to something else that is to be kept safe… a key as something with which we lock, unlock and lock again at will. No sufficient semantic interest or point could attach to the serious putatively comprehensive attribution kleis.

²⁷ In their several ways, Ayer, Stevenson, Hare, Ziff and Scanlon all see this. Of course. How much more they see needs to be determined author by author.

²⁸ Nor, once you look long and hard enough, is this sort of good unintelligible to Aristotle (who does not assert that all agathon is prakton). See Metaphysics XII, chapter 10:

²⁹ Compare Williams, Bernard “Must a concern for the environment be centred on human beings?” Making Sense of Humanity (Cambridge University Press, 1995), 243CrossRefGoogle Scholar; David Wiggins, “Nature, respect for nature, and the human scale of values”, Proceedings of the Aristotelian Society, New Series 100 (1999–2000), section V. (The proper text of this address is given only in the Society's own bound Proceedings, published in 2000. The text in the Society's autumn 1999 instalment is not the published text.) On cognate matters, see also my “Sufficient reason: a principle in diverse guises, both ancient and modern,” Acta Philosophica Fennica, 61 (1996), 119–132.

³⁰ For comments and suggestions I owe a particular debt to Peter Hacker, and further debts to Roger Crisp (and his article “Goodness and Reason: Accentuating the Negative”, Mind 117 (2008)), Stephen Williams, Roger Scruton and Åsa Wickforss. This paper derives from an address to the Memorial colloquium for Georg Henrik von Wright, Ab, Finland (2006) and the S. V. Keeling lecture in Ancient philosophy at University College, London (2007).