Does Armstrong need states of affairs?

Abstract

In 1997, David Armstrong argued that the world is a world of states of affairs. In his latest book, Truth and Truthmakers, he remains strongly committed to the existence of states of affairs, despite now advocating an ontology in which they are not needed, ‘as an ontological extra’. States of affairs remain needed, Armstrong says, ‘to act as truthmakers for predicative truths’. In this paper, I attempt to shed light on what Armstrong might mean by this claim. While there is a straightforward sense in which states of affairs are not needed in Armstrong's amended ontology, I suggest that Armstrong might be charitably interpreted in a manner that justifies his claim. However, in clarifying the manner in which states of affairs remain needed in Armstrong's ontology, it becomes unclear whether they are needed in any ‘deep’ sense, or rather are merely parochial to his ontology. I examine Armstrong's rejection of Resemblance Nominalism on the grounds that it does not provide adequate ‘minimal’ truthmakers. I then argue that he has significant additional work to do in explaining this concept before his rejection of Resemblance Nominalism can be justified, and thus before the need for states of affairs can be asserted generally, rather than just within particular ontologies, such as Armstrong's amended one.

Notes

¹‘The universal is a one that runs through its particulars just as the particular is a one that runs through its universals’ [2005: 243, my emphasis], as well as the analogy of the instantiation (partial identity) relationship to a table in which columns are particulars, rows are universals, and the intersections states of affairs [2004: 48].

²This inference implicitly appeals to the thesis that Armstrong, following Stephen Read, labels ‘expressibility’. It is the thesis that ‘For all being, there is a proposition (perhaps one never formulated by any mind at any time) that truly renders the existence and nature of this being’ [2004: 9]. If the expressibility thesis were not true, then a (or F) might exist without there being any propositions that truly correspond to it. Armstrong also affirms ‘maximalism’, the thesis that every truth has a truthmaker [2004: 5, 7 – 8].

³Armstrong is discussing the Martin-Heil view of property-tropes as non-transferable, and thus puts his argument in terms of tropes rather than universals. But he immediately follows the argument by claiming that ‘the situation is the same if we turn to the theory I now incline to: a's being F being necessary because a and universal F intersect, and are thus partially identical’ [2004: 49].

⁴States of affairs will be simpler parts of thick particulars and saturated universals so long as the particulars have multiple properties and the universals are multiply instantiated. It is possible that there might be a thick particular, a, that instantiates only one universal, F, and that is the only particular that instantiates F. In this case, the state of affairs, a's being F would be identical to both a and to F. An anonymous referee commented on this possibility by saying, ‘I am then unclear as to what, if any, distinction there is between that thick particular [or saturated universal] and the state of affairs of a's being F. This matters because one's grasp of the main issue begins to slip’. (bracketed text added) Indeed it does; the reader must judge the extent to which that is due to this author, or due to Armstrong.

⁵States of affairs might be posited as the basic ‘facts of resemblance’ in a Resemblance Nominalist's ontology. Gonzalo Rodriguez-Pereyra explicates Resemblance Nominalism in this way 2002: 85 – 7]. These states of affairs would not fit Armstrong's understanding of states of affairs as a particular having a property; Armstrong's states of affairs would be constituted by sets of several resemblance facts. For instance, where a, b, and c, are all the white particulars, aRb is a basic resemblance fact, while a's being white is {aRa & aRb & aRc}. Interestingly, Rodriquez-Pereyra argues against Armstrong's understanding of Resemblance Nominalism in terms of particularized natures on the grounds that it betrays the spirit of the view by making the resemblance between particulars derivative upon the natures of particulars [2002: 87 – 9].

⁶When Armstrong introduces the entailment principle, he notes that he would like to put limitations on this entailment in order to avoid any truthmaker for a contingent truth also being a truthmaker for every necessary truth. After discussing a problem for Frank Jackson's suggestion that the values that may be substituted for p and q might be limited to contingent truths, Armstrong suggests that this general approach might work if the values are limited to purely contingent truths. Without going into that notion here, I merely wish to point out that if Armstrong endorsed this understanding of the entailment principle, then one could not use it as I have here, since any proposition truly expressing resemblances between particularized natures and any predicative truths will be necessary given Resemblance Nominalism. Armstrong does not endorse the ‘amended Jackson thesis’ however, for he appeals to the entailment principle in providing truthmakers for truths of mere possibility, which are of course necessary [2004: 83 – 5]. It remains unclear exactly how the entailment principle is to be limited (as Armstrong admits), but surely the relationship between truths of resemblance and truths of predication given Armstrong's understanding of Resemblance Nominalism is just as ‘tight’ as that between a contingent truth, p, and the truth of <it is possible that not-p>.

⁷In introducing the notion of a minimal truthmaker, Armstrong points out that the world is both the least discerning truthmaker and the most promiscuous truthmaker, making all truths true. He does not suggest that the most discerning truthmakers should also be the least promiscuous, but if one could easily judge a truthmaker's promiscuity, I think he would be sympathetic to the comparison.

⁸In order to avoid the problem of the same particular(s) having unrelated but coextensive properties, Rodriguez-Pereyra is forced to appeal to a Lewisian realism about possible worlds [2002: 99 – 100]. Attaining one kind of ontological parsimony (only particulars exist) at the cost of another kind of ontological largesse (adopting Lewis's concrete possibilia) may seem too high a cost to pay for Resemblance Nominalism, but the theoretical point that one may attain sufficiently discerning minimal truthmakers without appealing to states of affairs stands. Note that Rodriguez-Pereyra includes states of affairs, or facts, within his ontology, but that these are not the same kind of entity as that with which Armstrong is concerned (see footnote 5 above.)

⁹I would like to thank Carl Gillett, Ronald Loeffler, Michael Rea, and the anonymous referees of this journal for their comments on previous drafts of this paper.