Reflections on philosophy and culture


Thursday, May 3, 2012

Two Questions About Explanation and Metaphysics

Aristotle and Spinoza in a mild-mannered stand-off to the death

One of the themes I've been interested in is the role of explanation in metaphysics.  To what extent and in what way should our explanations of things be a guide to metaphysical reflection?  However, this question might be taken in more than one way.  Below, I want to try and distinguish two general issues that I think are evoked by this question -  the question of realism and the question of fundamentality.  What I'm primarily interested in is the question of fundamentality, which - as I'll try to make clear - bears directly on how we should treat metaphysical notions like the nature or essence of something.  To my knowledge, there isn't currently any philosophical work that addresses this second question head-on.  But since there is a great deal of work on the first, to forestall confusion it's important to simply get clear on the distinction between the two questions.

1. The Question of Realism

Much of the philosophical literature on explanation and metaphysics concerns explanation in the natural sciences.  And, much of it revolves around the rivalry between various forms of realism and anti-realism about scientific explanation.  The questions here concern whether various forms of scientific explanation are ever strictly speaking true, or are, at best, useful ways of organizing our empirical knowledge of the world.

Now, there are at least a couple of different forms that such debates might take.  We can understand the distinction I have in mind on the basis of a distinction between two aspects of the seeming truth conditions of explanations.

A canonical form in which a very wide range of explanations can be couched is 'P because Q,' where 'P' is the explanandum (what is being explained), and 'Q' is the explanans (what does the explaining).  One question we can ask about any such explanation concerns the truth of the propositions flagging the word "because."  More specifically, when debating the question of realism, what is usually at issue is whether the explanantia of a certain class of explanations are ever strictly speaking true - i.e. whether they describe real features and items in the world.

For an example, consider one classic debate in the philosophy of science. It might be thought that the point of explanation in the natural sciences is to explain observable events and observable features of things.  However, scientific explanations often appeal, in their explanantia, to unobservable entities.  Someone who is attracted to an austere form of empiricism might be skeptical whether we are ever entitled to assent to the truth of claims about unobservable entities.  On this basis, she might argue that claims about such entities are not strictly speaking true, but rather are merely useful tools for organizing our knowledge of observable ones.

Or, consider another classic example from the philosophy of science.  It might be thought that natural science is in the business of giving explanations in terms of exceptionless natural laws - e.g., that a canonical scientific explanation of event E will say that E occurred because some set of initial conditions C obtained, and some natural law tells us that something like E will occur, given C.  However, depending on her conception of what natural laws are like, someone might worry whether our statements of them are ever strictly speaking true.  Suppose, for example, that we think that natural laws have counterfactual force.  An austere empiricist might then complain that such claims can never be established solely on empirical grounds, so that appealing to such claims has no place in empirical scientific explanation.  This is one way of expressing worries that have their source in Hume's reflections on causation.

A second, distinct kind of question we might ask about the seeming truth conditions of explanations concerns, not the strict truth of their explanantia, but rather the reality of what we might call etiological relations.  An etiological relation in the sense I have in mind is the relation seemingly represented by the word "because" in any explanation.  As many have philosophers have noticed, the word "because" doesn't seem to represent a single relation, but rather a range of different relations.  One common way of reading Aristotle's account of the four causes, for example, is as a distinction between four different kinds of etiological relation.  On this reading, Aristotle is claiming that when we ask why a change occurs, there are really four distinct 'Why?' questions we might be asking, to each of which there corresponds one of the four causes - formal, efficient, material and final.

Given this, we might ask whether this or that kind of etiological relation ever obtains, and if so, what kind of relation it is.  In the philosophy of science, one prominent example concerns what Aristotle called final causation.  We give a final causal explanation when we say something of the form 'P in order to F,' where 'P' is some proposition, and 'F' is some verb phrase in infinitive form.1  For example, we might say that this sapling here is growing in order to develop into a mature oak tree, or that my heart is beating in order to circulate blood throughout my body.  However, although we often give these kinds of explanations in ordinary life, philosophers of biology have wondered whether they are ever strictly speaking true - or if they are true, whether final causal explanations are reducible to efficient causal ones (what philosophers sometimes simply call 'causal explanations').  Yet again, the ultimate worry here is often epistemic - i.e. an empiricist's worry whether we have clear empirical criteria for accepting final causal explanations.

To take another example: consider efficient causal explanations of the form 'E1 caused E2' - for example, 'Ball A's moving like this (i.e. rolling towards and striking Ball B) caused Ball B's moving like that (i.e., moving away from A when struck).'  The causal relation between E1 and E2 here is an etiological relation: the present explanation is a more precise way of saying that B moved like that because A moved like this.  But depending on her views about efficient causation, someone might again be skeptical that efficient causation ever really takes place.  For example, as Hume pointed out, it seems like a claim like 'E1 causes E2' has counterfactual force built into it: in some sense, E1's causing E2 involves its necessitating E2.  And again, someone who is skeptical that empirical evidence ever entitles us to accepting claims with counterfactual force might then argue that efficient causal relations never strictly speaking obtain - that instead, we merely mark something about our epistemic situation when we talk about them.

In summary, the two principal ways of pursuing the question of realism about scientific explanations are by debating about (a) the strict truth of their explanantia, and (b) the strict truth of their attributions of etiological relations.  Now, I don't mean to suggest that there is a crystal clear dividing line between these two versions of the question of realism.  For example, as is suggested in the two examples above of Humean worries about causation, it might seem unclear to us exactly where to place such worries.  Are they best thought of as worries about explanantia that appeal to counterfactual laws?  Or, are they best thought of as worries about the causing relation?  Both?  Neither?  How you come down on these questions will depend on your views about the canonical forms that scientific explanations take.

However, what matters for the present purposes is that both of these ways of pursuing the question of realism concern the seeming truth-conditions of scientific explanations.  The over-arching question in both cases is: does this or that feature of this or that sort of explanation describe a real phenomenon, or is it merely a kind of fa├žon de parler?

Now, a philosopher who is (as I was) introduced to explanation as a metaphysical theme using the foregoing sorts of issues in the philosophy of science might very well broaden her view.   She might look at how the realist/anti-realist debates play out when we look beyond the kinds of explanations characteristic of the natural sciences - i.e. to yet further forms of explanation.  However, there's a set of very broad metaphysical issues about explanation that seem, at least prima facie, to be orthogonal to the various realism/anti-realism debates we might have about explanations.  These are the issues addressed by what I'm calling the question of fundamentality.

2. The Question of Fundamentality

The structure of our explanations has held a privileged place in the history of metaphysics.  Metaphysicians have often taken themselves to be pursuing an account of the most fundamental features of what there is.  And this pursuit has often been guided by what we might call explanatorism.  What I'm calling the question of fundamentality can be understood as a question about explanatorism, so I'll begin here by sketching the view (or range of views) I have in mind.

The explanatorist view might be encapsulated in the slogan: what explains is of the essence.2  What do I mean by this slogan?  The issue here is not the reality of etiological relations, or of what we seem to be describing in the explanantia of our explanations.  Suppose we simply assume that a given kind of explanation has true instances that describe real features of the world.  An explanatorist about this kind of explanation will make a further claim - one that we might resist, even though we accept the relevant assumption.  Namely, she'll claim that when we give this sort of explanation for some phenomenon, what explains gives us the essence or being of that phenomenon.

Now, in the latter formulation, the expression "what explains" is ambiguous.  That is, given a certain explanation, "what explains" can refer to both (a) the explanation's explanans (b) the explanation as whole.3  This ambiguity is intentional: explanatorism comes in at least two different species, corresponding to these two different ways of understanding what explains.  To make clear the unity of the genus, I'll first discuss the two species.

In one sort of explanatorism, explanantia occupy the essence- or being-determining role.  On this view, the explanantia of our explanations (or some particular class thereof) tell us what is most fundamental about the subject matter being explained, and thus reveal its essence or being to us.

For an example of this sort of explanatorism, we can look at one common conception of essence: essence as real definition.4  On this conception, to give an account of something's essence is to give something like its definition - not a definition of any expression we might use to talk about it, or any concept we might use to think about it, but rather a definition of the thing itself.  But what is it that makes some feature of a phenomenon part of its definition?  There are a range of answers to this question, but one of them is that those features of a thing are part of its real definition that explain various things about it.  For example, we might, in a spirit of scientific naturalism, claim that water's essence is to have the structure 'H2O' because its having this structure explains all of the things water does and undergoes in varying circumstances.  Or, we might claim (as Kit Fine does) that metaphysical necessities must be explained by something more fundamental - the real definitions of objects and kinds.4  For example, we might think that although Socrates necessarily belongs to the set {Socrates}, this necessity doesn't have its source in Socrates's essence, but somewhere else (i.e. in the essence of the kind 'set').  In other words, although we have a genuinely necessary property of Socrates on our hands, it's not one he has because of anything in particular about him.  Rather, it's explained by something else: the real definition of the kind 'set.'

In another kind of explanatorism, to understand the essence or being of something, we look not so much at the explanantia of our explanations as these explanations as a whole.  The general approach here is to take certain kinds of explanation to reveal to us the essence of some phenomenon, or even what we might call its categorial being.  This approach to various phenomena is a core aspect of various recent attempts to revive aspects of Aristotle's metaphysics - specifically, attempts that, under the influence of Wittgenstein and Frege, understand many of Aristotle's metaphysical insights as really 'grammatical' or quasi-logical ones.  Because of how much work is being done nowadays in this vein, I take my example below from this revival.

This kind of explanatorism can be found in Elizabeth Anscombe's reflections on the nature of intentional action in Intention.  She begins by claiming that to be an intentional action is to be an event "to which a certain sense of the question 'Why?' is given application" - specifically, the sense "in which the answer, if positive, gives a reason for acting."  Intentional action is defined by its relation to a certain form of explanation, and an investigation of the conceptual and logical architecture of this form of explanation fills in our account of the nature of the phenomenon.  The basic move here is this: the form of a special kind of explanation for a phenomenon just is the nature of that phenomenon.  This is a particularly important form of explanatorism, since the model Anscombe constructs here became the standard model for analytic philosophy of action in the following half century, up until the present moment: in this tradition, it is usually presupposed that the metaphysics of action just is the metaphysics of reasons explanation, which in turn is grounded in an analysis of the logical or conceptual architecture of reasons explanation.

In both of its forms, what explanatorist pictures suppose is that our explanations of some phenomenon (or, more frequently, some elite class thereof) are the proper guide to understanding its essence or being - what it fundamentally is.  This philosophical maneuver is so deeply ingrained in the tradition that philosophers hardly ever remark that explanatorism involves a substantive metaphysical commitment about essence or being.  This might be due to a very deeply imbedded tendency of thought going back to the ancients: the equation of being with intelligibility.  After all, explanation is the paradigm example of a technique by which we render some phenomenon more intelligible.

In any case, what I'm calling the question of fundamentality is this: is the explanatorist supposition, in one form or another, a legitimate one?  Are we really entitled to proceeding with this supposition in tow?  This question is more likely to seem pressing, of course, if some pressure can be put on the explanatorist supposition - rather in the way that epistemic worries put pressure on our default realism about the explanantia of explanations we accept or the etiological relations they attribute.

I think such pressure can indeed be put on the explanatorist supposition.  However, for the purposes of the present post, I only want to make clear that the question of fundamentality is distinct from the question of realism.  In the latter, the truth-aptness of certain kinds of explanations is under debate - i.e., whether they describe real features of and items in the world.  In the former, something stronger is at issue: not simply whether certain kinds of explanations are ever true, but whether they in varying senses give us the essence or being of the phenomena being explained.  No doubt there are fruitful connections between these two questions, but they need to be distinguished carefully in order for these connections to be clear.


(1) I have in mind here, not the notion of final causation as it is found in Aristotle's full-blown metaphysical account of causation, but rather final causation in its basic form.


(2) The term "explanatorism" and the slogan I've used here to encapsulate it derive originally from a paper I co-authored with Sarah Coolidge and Joseph Almog "Life Without Essence: Man as a Force-of-Nature," in Philosophical Perspectives, 25, Metaphysics, 2011, pp. 43-77.  The present discussion is intended as a development of some of the ideas concerning explanation we talk about in that piece.


(3) Of course, there's yet a third way of using "what explains" - namely, to refer to the person giving the explanation.  But this third use is not relevant for the present purposes.


(4) The notion of real definition has its sources in ancient philosophy.  However, contemporary metaphysicians have attempted to revive the notion.  Cf., for example, Kit Fine's "Essence and Modality," which can be found on Fine's website here.


(5) Cf. Kit Fine op. cit.  The example I use to illustrate this approach - of Socrates and {Socrates} - is Fine's as well.

No comments:

Post a Comment