There was a lot of discussion of the metaphysics of emergence at the Causal Powers and Sociology conference that we organized this past weekend. Almost everyone there was someone who’d thought about emergence a lot; many of the people there have published on it, several of the presentations addressed the issue directly; and most if not all of the discussion assumed a commitment to it in one form or another. But there was lots of disagreement, too, even amongst people who agree that there are such things as emergent wholes with emergent causal powers.
Some of the disagreement is semantic. People agree upon distinctions between different types of phenomena, but reserve the term ’emergence’ for different types. Some delimit the category such that emergent phenomena are few and far between; others are more permissive. Similarly, some use the term as a verb: entities and/or powers emerge out of other ones as a process of coming into being; others use it as an adjective: entities and/or powers are emergent, vis-a-vis some other entity and/or power(s). [As Jessica Wilson notes – via her attached paper in the comments, below – we can refer to these stances as diachronic and synchronic, when those terms serve to clarify — as many of us did in discussion during and after the conference.]
But there is also substantive disagreement, ranging from ontologically motivated disputes in favor or one or another of the semantic options, to disagreement about whether or not emergence under any description exists, to what emergence under any given description (if it exists) involves metaphysically.
The upshot is that there are all sorts of ways for people to talk past one another.
A useful place to begin, I think, is with the question of whether or not a whole is the same thing as a plurality of parts. In sociology, the classic formulation of this question involves a battalion. Is a battalion just another name for the total number of individual soldiers of which it is composed — such that, if there are 400 soldiers, the total number of entities in existence would be 400, rather than 401? Or does the battalion exist too, in addition to the 400 soldiers?
There are many ways to equivocate, in response to this question. In all cases it amounts (in my view) to wooly thinking. We need a yes or no answer to this first question. Not: “Well, battalions exist alright, but what they are is a plurality of soldiers.” Not: “Battalions aren’t anything in addition to a plurality of soldiers since what a battalion is is a plurality of soldiers duly arranged.” And for sure not: “Battalions are the same thing as a plurality of soldiers, but I can’t talk about the behavior of a battalion unless I help myself to the macro-level terminology, so I will, and it follows from this that I am not a reductionist.”
It’s fine if one answers “Yes” – i.e., “Yes, the battalion exists too, in addition to the 400 soldiers” – but also prefers to reserve the term ’emergence’ for something more than the fact that wholes do not reduce, ontologically, to pluralities of parts. To put it differently, one might think that emergence is something more than the negation of ontological individualism.
What would such a ‘something more’ be? One possibility would be to say that it’s not emergence unless the emergent whole is an entity that has no parts. Tim O’Connor and Jonathan Jacobs have defended this sort of view, though others do too — including Aquinas. Another possibility would be to say that an emergent whole is a relational phenomenon that retains its parts, but that the parts have to have been transformed in some way, such that they are qualitatively different from what they were on their own. Stephen Mumford and Rani Lill Anjum defend a view of this type, though again I don’t mean to ascribe the position to them uniquely. Stephen was at the conference, so it makes some sense to refer to him here.
Notice two things, with respect to these versions of emergence as something more than the fact (if it is one) that wholes are not equal to pluralities of parts — and that ontological individualism (i.e., metaphysical reduction in the sociological context) is therefore false.
First, neither version actually requires that one reject ontological individualism in the case of entities that don’t pass the recommended ‘something more’ test. A proponent of either version would be free to think that a battalion is not the 401st thing in an ontological inventory that includes 400 soldiers. A proponent of the ‘no-parts’ view who thought that battalions aren’t emergent entities could say that battalions reduce to micro-level phenomena because battalions have parts. Similarly, a proponent of the ‘transformed parts’ view who was a reductionist about ostensibly macro-level phenomena that flunk (their version of) the ‘something more’ test would say that the soldiers in a battalion have not been sufficiently changed for the battalion to count as a 401st thing. Of course, proponents of either version of the ’emergence requires something more than not reducing’ view would be equally able to think that battalions, even if they flunk the emergence test, do not reduce to a plurality of their parts.
Second, proponents of the ‘transformed parts’ view will have to be sure that it is the parts of the emergent entity that no longer behave as they did before, and not the whole emergent entity itself. For example, we can’t drink oxygen, and we can drink water. But if water does not reduce to a plurality of its parts, then it is not altered-oxygen that we drink, when we drink water, but water.
Even this much allows us to see more concretely some of the myriad of ways that people can wind up disagreeing and/or talking past each other. One might count all wholes as emergent, on the grounds that wholes don’t reduce. One might count only some wholes as deserving of the label ’emergent,’ but still think that no wholes reduce. One might think that only those phenomena that count as emergent according to a ‘something more’ (or, ‘something other’) test, don’t reduce. One might think that all purported sociological wholes reduce to individuals, but that the reduction stops there. One might think that it’s parts all the way down.
Sometimes the terms ‘strong emergence’ and ‘weak emergence’ are used to sort out these differences. I don’t find the terms helpful, so I’m avoiding them. Best, I think, is to get as close as possible to the alternate views of how the world is, without introducing potentially misleading names. Especially unhelpful, to my mind, is the use of the term ‘weak emergence’ to refer to what is already an equivocation (in my view), viz., the idea that a whole is a plurality of parts, “duly arranged” — where “duly arranged” does metaphysical work sufficient to distinguish such an arrangement from a plurality of parts as such (including, presumably, from heaps), but not sufficient to ground, or to be, an irreducible entity.
Now, I am a person who thinks that there is a difference between wholes whose parts retain their integrity and those whose parts are significantly transformed or whose material base is not best thought of as a proper part at all. I also think that there is a difference between living things and non-living things, and between conscious things and non-conscious things. But I don’t think that any wholes reduce to their parts, and I don’t think that it is possible to have it both ways with respect to “parts, duly arranged.” “Parts, duly arranged” are either a plurality of parts, or they are an irreducible whole; you don’t get the “duly arranged” for free.
So despite holding that there are important metaphysical differences between types of entities that don’t reduce, I still refer to all such entities as emergent. I do so because I think that the distinction upon which it is most important to have a roll call vote is that of reduction. Once we establish that no wholes reduce, we can sort the wholes into different metaphysical sub-categories.
A proponent of the ‘something more’ view – especially one who is a reductionist about everything except phenomena that pass the ‘something more’ test – can be expected to say that emergence so construed is too easy to come by. Everything and anything is going to turn out to be emergent.
I’d reply as follows. (Do, hereby, reply as follows.)
First, it would misleading to characterize the ‘something more’ view [when it is combined with reductionism about all of other (ostensible) wholes] as being a commitment to ‘strong’ rather than ‘weak’ emergence. In fact, rather than being strong on emergence, the position is much closer to a general reductionist program than is the view that no wholes reduce. The real objection, from this quarter, to the person who thinks, as I do, that no wholes reduce is that s/he has taken too hard of a line against reductionism.
Second, note that the logic of what I’ve just said becomes maximally apparent once we remove the option of fudging on “duly arranged.”
I say this because, third: if the equivocation is disallowed, then the proponent of the ‘something more’ position has to say either (a) that what they are counting as emergent entities are the only entities that don’t reduce, or (b) that non-reductive but non-emergent (as per their scheme) arrangements don’t reduce either.
And if they say the latter, then, fourth: the disagreement is not about there being too many non-reducible entities; it’s about not having enough names for the different types of non-reducible entities, and wanting to use the name ’emergent’ for only one of the types, rather than for the whole set of non-reducible entities.
Finally, notice that the otherwise reductionist proponent of the ‘something more’ criterion for wholes will not be allowed an emergent entity that is emergent vis-a-vis an organized entity one level “down,” unless that micro-level entity also passes the ‘something more’ test. It will be reduction all the way down to a “lower” level, if there is one, of part-less entity or entity whose parts have been sufficiently transformed.
Okay, so far so good. As I say, I like ’emergence’ for all wholes, with appropriate names for ones with different general compositional features (and I’m so far talking here about propertied entities, not properties themselves). But I don’t care too much about the names. We can call non-reducible entities ‘wholes’ if that’s better — and it might be. Or NRE’s. Or Biff.
I’ll go with ‘NRE’ for present purposes. Let me also assert that it’s in the nature of the case that if it’s an NRE, its properties are had by it, not by its parts or by a plurality thereof. If you want to tuck the properties of wholes into the properties of their parts, you’re simply denying that there are NREs. Whether or not it’s a plausible move – the tucking – is a separate issue.
Having said all of this really just as preliminary, in order to try to lessen the likelihood of miscommunication, I want to offer some quick & initial thoughts in relation to the disputes about emergence that were had at last weekend’s conference. There were two major points of disagreement.
First, people disagree about whether NREs are (a) caused by their parts or (b) constituted (but not exhausted by) their parts.
Second, there was disagreement about whether the properties of NREs [rather than the entities themselves (treat the distinction as you will)] are (a) caused by the properties of the parts or (b) had by the NRE, as part and parcel of its being what it is, rather than something else.
Some people answered (a) to both questions. Others (b) to both. And one person answered (b) to the first, (a) to the second. In all cases people agreed that the causal processes (i.e., displays of the powers of things) are diachronic. I was in the (b)-to-both group.
But what I thought even more than that the (b)’s seemed righter than the (a)’s, was that there were actually better ways to think about the whole thing, as it were (get it?), than the ways given by those choices so formulated.
However, because it is very, very late, and this post is now very, very long, I am going to end with: “To be continued in Part 2.” If you have read this far, I think that what I say in Part 2 will be clear, even if you don’t agree. That will (hopefully) have made Part 1 worth it.