In what was going to be Part 2 but will now be Part 3, I was (and still am) going to say a few things about how I think of emergence, in keeping with what I said in Part 1.
But before I do that I want to register some problems that I see with maintaining, as Stephen Mumford and Rani Lill Anjum do, that emergent phenomena (a) have parts; which parts, (b) in virtue of being parts of an emergent whole, are different from what they (the parts) are like on their own, i.e., when they are not parts of such a whole (recall: the parts having been sufficiently transformed is Stephen & Rani’s test for the whole being emergent); and which parts (c) either in their original form or in their changed form (it seems as though it would have to be in their original form, but it’s not clear), cause the emergent whole of which they are the parts. I’m not opposed to (a) – some emergent wholes do have parts, in my view – and I’m fine with (b) too (though not as a criterion of emergence), since in some cases the parts are indeed transformed. It’s (c) that I reject. I don’t think that the relationship between an emergent whole and its parts (if it has parts) is causal.
Stephen and Rani are neither alone in thinking that emergent entities are caused by the lower-level phenomena to which those who deny emergence think purportedly emergent entities reduce, nor the first to think it. But they think it and also hold an explicitly anti-Humean, anti-passivist, productive account of what causation is. This is true of Doug Porpora too, for that matter, but Stephen presented (a)-(c) as a package at the conference that Phil Gorski & I organized earlier this month, so for that reason I will direct myself to him and Rani now.
Here are some points that seem problematic to me, in no especially important order.
1. Is it the parts pre- or post-transformation that cause the emergent entity? [Actually, Stephen and Rani prefer to talk of powers being causes, not powerful things. So to be precise it’s probably the powers of its parts that, as they have it, cause the emergent whole (and its powers), not its (powerful) parts that do so. I’ll use “micro-level phenomena” whenever possible (or, stylistically viable), which can refer either to powers of parts or to powerful parts, as one prefers.]
At first glance, it looks as though it’s going to have to be the pre-transformation micro-level phenomena that are the cause of the emergent whole, since the parts being different than they would otherwise be (i.e., if they weren’t parts of an emergent whole) presupposes that the emergent whole (of which they are the parts) already exists.
But if we say that the causing of the emergent whole is done by its pre-transformation parts (or the powers thereof), we will then need to know what the relationship is between the resultant emergent whole and its (now sufficiently) transformed parts — since (by stipulation) the latter parts (or the powers thereof) aren’t the ones that are its cause. This way seems to leave us with too many groups of parts, only the first of which (or the powers thereof) is the cause. Moreover, if the transformed parts (or the powers thereof) aren’t the ones that cause the emergent whole – if, e.g., they instead constitute it in some way – then the position starts to sound less like an alternative to the constitution view, since the instant the emergent whole exists at all, its parts are indeed the transformed ones (as per Stephen and Rani’s account of what an emergent entity is).
2. If we say instead that the emergent whole is caused by (the powers of) its transformed parts, we will be able to just set aside the pre-transformed parts (or powers thereof). They have nothing to do with the emergent entity in question. So now at least we don’t have an extra batch of parts to deal with.
But there are complications this way too. First is the issue of the transformed parts presupposing the existence of the emergent whole. Second, in explaining why it’s the already-transformed parts (or the powers thereof) that cause the emergent whole, rather than the parts as they were before they were parts of the whole, it looks as though it will be hard to avoid saying that the reason why the pre-transformation parts (or powers thereof) aren’t relevant is that, not being transformed, they amount to a plurality of parts, not an emergent whole. But as soon as we start talking about what things amount to, it’s harder to make out why it wouldn’t be that what we’re talking about is constitution, rather than causation.
This will also be the end-result, I think, of saying that the reason why there is no paradox in the case of transformed parts causing a whole that they presuppose is that causes can bring about effects instantly. Here, of course, we need the full effect, viz., the whole, to be brought about instantly, such that the transformed parts can instantly be the transformed ones that we need to cause the whole in the first place. But once you say this it starts to sound very much as though what we’ve got is constitution, expressed as a wildly implausible causal scenario.
3. What if we say that emergent wholes are caused first by pre-transformation micro-level phenomena, and then, after that, by the now-existent wholes’ now-transformed parts (or the powers thereof)? This seems messy too. If nothing else, it is going to be a question how two different sets of powers [those of the (soon-to-be) parts pre-transformation and those of the parts post-transformation] have caused the same emergent entity.
4. Since the emergent entity doesn’t ever stop having (now-transformed) parts, it looks as though emergent wholes are going to be being caused throughout their existence. At no point will the cause have issued in the effect, i.e., the existence of the emergent whole. This doesn’t seem right either. Especially since we need the whole to be caused instantly, in order to underwrite the transformation of the parts.
5. If emergent entities are caused by their parts, transformed or not, it would seem that all real entities (i.e., all entities that don’t reduce, metaphysically, to other entities) that have parts will have to be caused by their parts. That alone needn’t be a problem. The proponent of this view will just say that all real entities with parts are, by definition, emergent. But it does raise the question of what the principled difference is between causation and constitution, on this powers-based view of emergence as a causal relation.
6. And it does matter that it’s a powers-based view of emergence as a causal relation. If one were to hold, for example, that causation is counterfactual dependence, then to claim that the emergent whole is “caused” by its parts would be to say nothing more than that it wouldn’t exist if its parts didn’t. The realist about productive causal powers (who means something entirely different by “caused by” than counterfactual dependence) is perfectly entitled to deny that emergent entities are caused by their parts (transformed or otherwise) but affirm that, absent their parts (or their base, be this parts or not), they would not exist.
7. My hunch is that the powers-based argument over whether or not emergent phenomena are caused by their parts (rather than being constituted though not exhausted by their parts, if they have them) — my hunch is that disagreement about this ultimately intersects with the issue of whether or not one retains substances (or objects) in one’s ontology, if only since an ontology that bottoms out in displays of powers is one in which issues of constitution may not figure at all.
8. For what it’s worth, Roy Bhaskar argued against the view that societies, as one type of emergent whole anyway, are caused by individuals. He didn’t think that societies are composed of (but not exhausted by) individuals, either, but even so, critical realists who think that societies are emergent wholes that are caused by their (transformed) parts, where the parts are individuals, should note that in saying so they are rejecting Roy’s charge that such a position amounts to voluntarism, and that individuals must be seen as re-producing society, rather than producing it. It’s fine to disagree with Roy Bhaskar – I argued against his concept of alethic truth as early as 1997; I mean only to point out the connection between the present issue and Roy’s defense of the TMSA (transformational model of social activity) and the so-called “ontological gap” between individuals and society (though the gap issue is not identical to the “not caused by” point).
9. Again, I am working up to a post in which I offer some positive thoughts on these matters. But, as before, I hope that these preliminary comments will be of interest in the meantime.