«Exemplification as Explanation Anna-Sofia Maurin Lund University Abstract: In this paper I critically investigate an unorthodox attempt to ...»
Published in Axiomathes online first (2011). [doi: 10.1007/s10516-011-9174-8].
If you want to quote this paper but do not have access to the published version, feel free to e-mail
me at: email@example.com and I will send you the final product.
Exemplification as Explanation
Abstract: In this paper I critically investigate an unorthodox attempt to metaphysically explain in virtue of what
there are states of affairs. This is a suggestion according to which states of affairs exist thanks to, rather than, as is the common view, in spite of, the infinite regress their metaphysical explanation seems to engender. I argue that, no matter in which form it is defended, or in which theoretical framework it is set, this suggestion cannot provide us with the explanation we crave.
You are sitting at your desk thinking about exemplification. Distracted, you reach for the cup of coffee to your right. You touch it, flinch, and coffee spills all over the work you have laid out in front of you. Annoyed you ask yourself: Why is the cup hot? Then you remember that the reason the cup is hot is that only minutes before you touched it your colleague was kind enough to top it up for you.
Mystery solved. You calm down and take another look at the mess in front of you. But instead of cleaning it up you get distracted once more. Why, you repeat – now wearing your philosopher’s hat – is the cup hot? But now the answer which satisfied you as a response to the first, seemingly identical, question will no longer do. It will not do because what you are asking for is not an account of what caused the cup’s temperature; what you desire, rather, is an account of what reality must contain – how it must be ontologically structured – in order for it to be the case that the cup is hot. What you want is a metaphysical explanation. What you need, as we shall see shortly, is to think some more about exemplification.
Metaphysical explanation – a rough characterisation It is easier to say what a metaphysical explanation is not than it is to say what it is. Negatively speaking, therefore, we can safely say that a metaphysical explanation is not a causal explanation.
But if it is not causal, what is it? A distinctively metaphysical explanation, we might respond, is a constitutive explanation.1 But exactly what this characterisation should be taken to entail is, again, not obvious. It does, however, seem at least not unsafe to claim that, when we say that A metaphysically explains B if B in virtue of A, what we mean is not just that A somehow necessitates B or that A makes B exist/obtain (although we do mean that as well). What we are saying is also, minimally, that A is this necessitating existential condition for B, and hence makes B exist/obtain, in virtue of being what B (mereologically or non-mereologically) consists of or, at least, by being the base on which B somehow supervenes or from which B emerges. Metaphysical explanations, thus understood, are not subjective; i.e., their value is not contingent on the sense of insight or revelation they give to their recipients (although a sense of insight may be, in fact very often is, a positive sideeffect of a successful metaphysical explanation). For this reason, a perfectly legitimate metaphysical explanation may be of the primitivist form – B in virtue of B – where the insight gained is at most very little. In general, however, a metaphysical explanation in terms of something more rather than of I borrow the distinction between causal and constitutive explanations, though not the use made of it, from W. Salmon (1984).
something equally fundamental is supposed to have a higher explanatory value so that, ceteris paribus, we should prefer an explanation of the form B in virtue of A (where A is more fundamental than B) over one of the form B in virtue of B. Now, it is normally said that we can metaphysically explain not just existence but also truth. However, clearly, to explain the truth of some proposition cannot be the same as listing its constituents. Instead, the idea that truth as well as existence can be metaphysically explained trades on the very reasonable idea that propositions are somehow made true by what exists or, at least, that their truth in some sense depends on or supervenes on what exists. A metaphysical explanation of the truth of a proposition looks, in the first instance, very much like a simple instance of the T-schema, e.g.: The cup is hottrue in virtue of The cup is hotexists.2 A metaphysical explanation, but not every instance of a T-schema, is however such that, that which supposedly explains the truth of the proposition in question is both distinct from and independent of the proposition whose truth it explains. Full metaphysical explanation, moreover, requires that not just the truth of the proposition, but also the existence of the state of affairs (which presumably explains the truth of the proposition), receives a metaphysical explanation.
A full explication of our notion of a metaphysical explanation is beyond the scope of this paper.3 For the moment, therefore, we will have to make do with the rough, but hopefully relatively uncontroversial, characterisation just given. As will become evident, however, to be able to evaluate the different answers to the question – why is the cup hot? – proposed in the literature, this first characterisation will have to be complicated. A complication which, as we shall see, involves distinguishing between different kinds of metaphysical explanatory tasks.
Why is the cup hot? – The Orthodox View4 Back to the cup and to thinking about exemplification. The explanatory task we have now set out for ourselves often is (though it does not have to be) put in terms of our seeking to metaphysically explain the truth of some proposition, in this case: The cup is hot. A metaphysical explanation of this kind, we have just noted, starts out looking very much like a simple T-sentence: The cup is hottrue in virtue of The cup is hotexists. Or, in general;
Explanation1: Fatrue in virtue of Faexists
Now, as we have also seen, this is not a full metaphysical explanation. For, Fa can only explain the truth of Fa if Fa can, in turn, be metaphysically explained. What grounds the existence of Fa? Fa, in virtue of what? Fa could, of course, be given a primitivist explanation – Fa in virtue of Fa – but this is not the preferred view among proponents of what I will call “the orthodox view”. For, according to proponents of this view there are strong independent reasons for thinking that Fa is a contingently existing, ontologically complex, entity constituted by a (a particular) and F-ness (a universal), which means that its metaphysical explanation should (at least) include reference to those constituents.
Explanation2: Faexists in virtue of F-nessexists and aexists
The problem with explanation2 is that it cannot, appearances to the contrary, explain what it is supposed to. The reason why not is that as the existence of Fa is only contingent (the cup does not have to be hot), then given the existence of a and F-ness, the existence of Fa is not guaranteed.
Therefore, Fa does not exist in virtue of, or has its existence grounded in, the existence of just a and Throughout this paper, anything put between ‘’ and ‘’ should be understood as a proposition.
For some recent, more thorough (and sometimes conflicting), views on the topic of metaphysical explanations, see e.g.
Betti (2010); Correia (2008); deRosset (forthcoming); Schaffer (2010); Schneider (2010); Wieland & Weber (2010).
Proponents of the Orthodox view include, as we shall see, all philosophers who agree that metaphysically explaining the existence of states of affairs by adding a universal exemplification-relation to a and F-ness leads to a vicious infinite regress.
This means that there are a lot of proponents of this view. One prominent example is D. M. Armstrong (1978; 1997; 2004).
F-ness. More is needed. On the Orthodox view, what is needed is some way to ground the existence of Fa in not just the existence of the prima facie constituents of Fa, but in those same constituents somehow united. Enter exemplification. Fa exists, we can now say, in virtue of the existence of not just a and F-ness, but of exemplification as well, for to say that Fa exists is to say that apart from a and F-ness there is this exemplification-relation which holds a and F-ness together. Explanation2, therefore, should be discarded and replaced by the following, supposedly more accurate
Explanation3: Faexists in virtue of aexists, F-nessexists, and Exemplificationexists But, again, clearly, this explanation cannot be the last word. For the same reason that made us rule out an explanation of the existence of Fa that pointed to (just) a and F-ness, we cannot rest content with an explanation that simply adds one more constituent to this list, for as before, what is now on the list, it seems, could exist and yet a not exemplify F-ness, which means that those things could well exist yet Fa not exist. For metaphysical explanation3 to succeed more is, once again, needed.
What we need now is not just some way of binding a to F-ness (which was what we were looking for when offering explanation3) but also some way of binding Exemplification to a and F-ness.
Explanation3 should therefore be discarded and replaced:
nd Explanation4: Faexists in virtue of aexists, F-nessexists, Exemplificationexists, and (2 order) Exemplificationexists But now we can see that this explanation, just as explanation3, creates rather than satisfies an explanatory need. In virtue of what does a, F-ness, and Exemplification exemplify Exemplification? It is clear that if we continue along the same lines – and answer that a, F-ness, and (1st order) Exemplification exemplify (2nd order) Exemplification in virtue of a, F-ness, (1st order) Exemplification, and (2nd order) Exemplification exemplifying (3rd order) Exemplification – not much is gained. We end
up in an infinite regress that has the following appearance:5
That this regress is vicious is, again, the orthodox view. Explanation1 can only succeed if it can, in turn be explained, but whichever explanation we suggest, there is always something missing. Explanation is infinitely deferred, and hence never given. The infinite regress functions as a reductio. Given the regress, F-ness and a cannot explain the existence of Fa, which means that Fa cannot, after all, explain the truth of Fa, which means that we still do not know in virtue of what the cup is hot.
Does this failure to metaphysically explain the existence of Fa, and hence the truth of Fa, mean that there is no way in which Fa (and hence Fa) can receive a metaphysical explanation? Not necessarily. By finding out what doesn’t work, proponents of the orthodox view tend to argue, we are well on route towards finding out what does, and the number of suggested ways out are many and varied.6 This is however not a paper on how, in spite of the infinite regress This regress is perhaps most famously formulated by F. H. Bradley in his (1908 ). Cf. also: Maurin (2010).
More or less extreme ways out include e.g. accepting monism and regarding all distinction as appearance (Bradley’s own solution) and (less extremely) tinkering with how we regard the nature of exemplification by either taking exemplification to be “non-relational” (cf. Armstrong (1978)) or by taking it to be such that it necessarily relates specific relata (cf.
Bergmann (1967); Maurin (2002; 2010; 2011)), etc. Another escape, one that we will have reason to get back to in this text, generated above, we can nevertheless metaphysically explain either the existence of Fa or the truth of Fa. Instead, this is a paper on how, according to some philosophers, those same things can be metaphysically explained thanks to the regress any attempt to provide such an explanation will generate. This is the unorthodox view.
Why is the cup hot? – The Unorthodox View How can an infinity of disjointed entities, entities whose existence is perfectly compatible with the non-existence of Fa, nevertheless metaphysically explain Fa? The answer, of course, is that it cannot.
But then, proponents of the unorthodox view argue, the infinity to which our attempt to metaphysically explain the existence of Fa commits us, is not this infinity of disjointed entities in Fa.
To see this, they argue, we need to consider, once again, our admittedly regressive explanation.
Suppose, therefore, and as before, that what we in the end want to explain is the truth
of Fa, and that our explanation should have the following appearance:
Explanation1: Fatrue in virtue of Faexists
Suppose next, and again in line with the orthodox view, that Fa, which you agree is a contingently existing ontologically complex state of affairs constituted by the union of a (a particular) and F-ness (a universal), must in turn receive a metaphysical explanation. Now part company with the proponents
of the orthodox view and suggest the following explanation of the existence of Fa:
Explanation2: Faexists in virtue of the state of affairs that a exemplifies F-nessexists Again, clearly, this explanation cannot be the last word. For, in virtue of what does the state of affairs
that a exemplifies F-ness exist? The answer, says the proponent of the unorthodox view, is this:
But now we can see that this explanation, just as explanation2, needs to be metaphysically explained.
In virtue of what does the state of affairs that a and F-ness exemplify (1st order) Exemplification, exist? Again, we end up in infinite regress. So, how is this different from before? It is different, proponents of the unorthodox view argue, because this time, by not expressing the regress in a misleading way, we can see that it does not have the structure envisaged by the proponent of the
orthodox view. Instead it has the following structure: