[1] Philosophical Papers, Volume II (Oxford: Oxford University Press, 1986), p. ix.

[2] A recent example is Bas van Fraassen's critique of contemporary non-Humean accounts of modality in Laws and Symmetry (Oxford: Oxford University Press, 1990).

[3] On this see G. D. Romanos, Quine and Analytic Philosophy (Cambridge: MIT Press, 1983), Ch. 3.

[4] Philosophical Investigations, 3rd. English edition (Oxford: Basil Blackwell, 1968), p. 224; see also sections 23-4, for example.

[5] Quine, Theories and Things (Cambridge, Mass: Harvard University Press, 1981), p. 98

[6] Theories and Things, pp. 97-8

[7] Word & Object, (Cambridge, Mass: MIT Press, 1960), section 27.

[8] There has been much discussion as to whether there are autonomous levels of explanation in science. Those who argue that there are such levels commonly insist that nevertheless, physics has ontological priority. Functionalists about the mental will often profess to ontological physicalism, for example, though claiming that psychology involves autonomous modes of description of (certain special arrangements of) physical entities or properties. For present purposes I shall assume that this ontological criterion is capable of bearing the weight, and hence that such a view does count as a form of reductionism in the sense of Figure 1 (in contrast say to Cartesian dualism). If this assumption were to prove untenable, then I think the defenders of multiple levels of description, like those I am calling lapsed Humeans, would be hard pressed to distinguish their view from a discourse pluralism about the scientific hierarchy.

[9] Bear in mind that the relevant notion of reduction is an ontological one; see the previous footnote.

[10] Ethics: Inventing Right and Wrong (Harmondsworth: Penguin Books, 1977).

[11] See in particular his Spreading the Word (Oxford: Oxford University Press, 1984).

[12] Why not say "... as if this realm were more inclusive than it actually is," so that the alternative to non-factualism becomes additive monism? Here I anticipate a little--we shall see that the availability of a substantial distinction between factual and non-factual uses of language turns out to be a requirement for the additive monist, as much as for the non-factualist.

[13] See David Armstrong, What is a Law of Nature (Cambridge: Cambridge University Press, 1983), for example.[ ]It is my contention that these views would be better explicitly recast in terms of discourse pluralism, but clearly they are not intended as such.

[14] See "Moral Realism" in Casey, ed., Morality and Moral Reasoning (London: Methuen, 1971), 101-24; Spreading the Word (Oxford: Oxford University Press, 1984); and "Supervenience Revisited" in Hacking, ed., Exercises in Analysis: Essays by Students of Casimir Lewy (Cambridge: Cambridge University Press, 1985), 47-67.

[15] Strictly, a monist might consistently hold that all discourse is factual (a position that David Lewis adopts in "General Semantics," for example; see his Philosophical Papers, Volume I (Oxford: Oxford University Press, 1983), 198-232). What matters is not the factual/non-factual distinction as such, but the characterisation of the factual itself.

[16] These also exhibit striking superficial similarities, looking to the untrained eye like so much raw meat. Who could have guessed that the kidney was really a filter, the brain a thinker, etc.? The pluralist is suggesting that the functions of language might exhibit a similar diversity and modularity.

[17] Physicalism might after all emerge as a degenerate form of discourse pluralism, namely as the view that in fact there is only one autonomous and legitimate discourse, although there could have been many.

[18] In Facts and the Function of Truth (Oxford: Basil Blackwell, 1988).

[19] The odd thing is that Quine himself never seems to notice that his scepticism has this consequence. As we noted at the end of Section II, he takes his minimalist intuitions about truth to support the view that there is a single category of descriptive discourse, contra for example those who would take an instrumentalist view of talk of abstract objects. But he fails to notice that this is the sort of "victory" that we achieve by sweeping the pieces from the board. The default position is a kind of pluralism--not the non-factualist's kind, but the Wittgensteinian doctrine that superficial unity of our talk of truth masks wide differences in the functional roles of discourses or "forms of life."

[20] As we have already noted, Quine's minimalist views about truth fall into this category.

[21] "Consistency and Realism," reprinted in Problems of the Self (Cambridge: Cambridge University Press, 1973), 187-206, at p. 202.

[22] In the original emotivist case, this difficulty has recently been highlighted by the inconclusiveness of several attempts to exclude the possibility that desires or evaluative attitudes might themselves be represented as a special class of beliefs. See in particular Michael Smith, "The Humean Theory of Motivation," Mind, XCVI, 1987, 36-61; Philip Pettit, "Humeans, Anti-Humeans and Motivation," Mind, XCVII, 1988, 530-33; David Lewis, "Desire as Belief," Mind, XCVII, 1988, 323-32; John Collins, "Belief, Desire and Revision," Mind, XCVII, 1988, 333-42; Michael Smith, "On Humeans, Anti-Humeans and Motivation: a Reply to Pettit," Mind, XCVII, 1988, 589-95; Huw Price, "Defending Desire-as-Belief," Mind, XCVIII, 1989, 119-127; and Philip Pettit and Huw Price, "Bare Functional Desire," Analysis, XLIX, 1989, 162-9.

[23] See "Realism - the Contemporary Debate: Whither now?," forthcoming in J. Haldane and C. Wright, eds., Reality, Representation and Projection (Oxford: Oxford University Press, 1992).

[24] See my "Sense, Assertion, Dummett and Denial," Mind, XCII, 1983, 174-88, and "Why 'Not'?," Mind, XCIX, 1990, 221-38.

[25] "Realism - the Contemporary Debate: Whither now?," typescript p.11.

[26] This point was made vivid to me in discussions with John Campbell. Discourse pluralism seems compatible with what Campbell calls the Simple View about entities or properties of a given kind. See, e.g., his "A Simple View of Colour," forthcoming in J. Haldane and C. Wright, eds., Reality, Representation and Projection (Oxford: Oxford University Press, 1992).

[27] This misinterpretation of Quine seems to me to parallel a misunderstanding concerning the role of best explanation of science. It is part of the practice of science to accept the best current explanations of observed phenomena. This is hardly more than a truism: to say that an explanation is the best we have is ipso facto to indicate that we give it more credence than anything else on offer. There is a tendency to make a second order principle of inference out of this: to say that because they provide best explanations, the theories concerned are likely to be true, and therefore worthy of a further degree of credence. Another similar mistake is that of thinking that the principle that the best choice is the one that maximises one's expected utility assigns a further value to the choice in question, over and above its expected utility.

[28] I argued in Section III that the monist's unifying principle cannot be entirely ontological in any case, but must involve a semantic component. The present point is that Quinean quietism undercuts the ontological strategy at first base, as it were.

[29] We pluralists think that many of those who try to deflate the lapsed Humean in the first way also fall prey to this misconception.

[30] For Dennett's own recent response to this concern, see his "Real Patterns," this JOURNAL, LXXXVIII, 1991, 27-51.

[31] Recent work attempting to generalise the notion of a secondary quality might be seen as falling under this heading. See for example Crispin Wright, "Realism - the Contemporary Debate: Whither now?," and Mark Johnston, "Objectivity Refigured," both forthcoming in J. Haldane and C. Wright, eds., Reality, Representation and Projection (Oxford: Oxford University Press, 1992); Johnston's "Dispositional Theories of Value," Proceedings of the Aristotelian Society, Supp. Vol. LXIII, 1989, 139-74; and Philip Pettit, "Realism and Response-Dependence," Mind, C, 1991, 587-626. For a critical view of this program, offering an alternative more in keeping with discourse pluralism, see my "Two Paths to Pragmatism," Working Papers in Philosophy, Research School of Social Sciences, ANU.

[32] I am grateful to audiences in Canberra, Brisbane, Sydney and Wollongong, and particularly to John Burgess, John Campbell, John O'Leary-Hawthorne, Michaelis Michael, Philip Pettit, Michael Schepanski and Mark Walker, for many comments on earlier versions.