A few days ago, Graham Harman gave a hint of his (soon-to-be-published) response to my (soon-to-be-published) Whiteheadian critique of him:
Shaviro is wrong to say that I am inconsistent in allowing this sort of infinite regress while being horrified it in the case of Latourâ€™s mediators, where Joliot mediates between politics and neutrons, but further mediators will be needed between Joliot and neutrons, and so on infinitely. Using the phrase â€œinfinite regressâ€ for both is an equivocation, as I will explain in my response to Shaviro in The Speculative Turn. The first of the two is merely strange, but contradicts no obvious facts. Latourâ€™s infinite mediators, however, contradict themselves: if there is a problem with politics touching neutrons directly, then there will be just as big a problem with Joliot touching either of these, and so on to infinity, with the result that no contact between any two things will ever occur. But this is absurd, while the infinite regress of objects is merely strange, not illogical.
Now, I don’t feel I understand Latour well enough to comment on the problem Harman raises as to the supposed infinite regress of mediators between mediators in Latour’s account of networks. Harman is claiming, in effect, that Latour falls victim to a Zeno’s Paradox kind of situaiton. I cannot judge whether this is right or not.
However, this wasn’t quite the point I was making when I compared Harman’s acceptance of the infinite regress of objects or substances with his rejection of the infinite regress of relations. Harman’s complaint about Whitehead is that, in Whitehead’s brand of relationalism , an entity is
nothing more than its perception of other entities. These entities, in turn, are made up of still other perceptions. The hot potato is passed on down the line, and we never reach any reality that would be able to anchor the various perceptions of it. (Guerrilla Metaphysics, p. 82)
This sort of infinite regress has nothing to do with the Zeno’s Paradox-like regress of which Harman accuses Latour. It is rather a regress of the same “merely strange, not illogical” sort that Harman himself so cheerfully accepts when it comes to substances. Therefore, I still don’t accept Harman’s distinction between the bad infinite regress of relations, and the good infinite regress of substances.
Harman’s objection to Whitehead’s relationalism comes down to his sense that, given this regress of relations, an entity is nothing but its prehension (or, what Harman calls “perception”) of other entities. However, this is wrong. Whitehead explicitly rejects the “nothing but” that is unspoken, but seems to be taken for granted, in Harman’s account. The “nothing but” is what Whitehead calls the “sensationalist principle” of Locke, Hume, and Kant: the assumption that perception is purely passive and receptive, the “bare entertainment” of data. This assumption is wrong, for several reasons. First, because prehension or perception has much more to do with affectivity, or with “obscure feelings,” than it does with the “clear and distinct” ideas of “presentational immediacy” (which is the only form of perception recognized by Hume, and, responding to him, by Kant); and second, because prehension is active as well as passive (or, better perhaps, because it resists being described in terms of an active/passive dichotomy). When what Whitehead calls an “actual entity” constructs itself by prehending or perceiving other entities, it is in fact engaged in elaborate processes of revaluing the prehended data: processes of choosing, adding, subtracting, relating, juxtaposing, tweaking, and recombining: ultimately of deciding and selecting. Decision and selection are left out of Harman’s account of relationalism.
I think that the big problem with Harman’s attack on relationalism is that he doesn’t distinguish between different notions or senses of relation. His complaint that relationalism destroys the integrity of individual entites, and does not allow for novelty to emerge, might well be correct if applied to Hegelian, dialectical relationalism, in which the “labor of the negative” allows everything to be swallowed up within a self-reflecting totality, as well as if applied to Saussurean or “structuralist” relationism, where there are no positive terms, but only negative determinations of the differential elements within a closed order by one another. But this is not the way that relations work in William James, or Whitehead, or (to the extent that I understand him) in Latour. Relations proliferate wildly in these thinkers; there is no overarching totality or system in accordance with which they are deployed. This difference is somewhat akin, I think, to Delanda’s distinction between the “relations of interiority” that we find in Hegel and Saussure and the “exteriority of relations” that Delanda himself, following Deleuze, favors (I have written about Delanda’s logic here). Where Harman gives us a vision of independent substances wrapped in vacuums and only communicating vicariously, James and Whitehead and Latour and Delanda give us a world in which things are continually jostling up against one another, touching and affecting one another (but this does not mean that they are totally interpenetrating one another, or “determining” one another through some sort of ironclad causality).
It may well be that an ungrounded infinite regress is not such a bad thing (as Harman says, for instance, here). There are, however, other ways to nuance the question of infinite regress. Kvond suggests as much here, raising the point that what stops the regress from being infinite might be of another nature than the entities among which the regress takes place. (This could be seen in a number of ways; I am inclined to think of it in terms of Schelling’s notion of a ground, as opposed to Hegel’s totalizing closure). But I need to think about this some more, so I will postpone further discussion until another time. For now, I will just conclude by saying that the great thing about Harman’s philosophy is the way it makes us so aware of the multiplicity and diversity of things:
Atoms and molecules are actants, as are children, raindrops, bullet trains, politicians, and numerals. All entities are on exactly the same ontological footing. (Prince of Networks, p. 14)
But I agree with William James that relations, both “conjunctive” and “disjunctive,” are every bit as real as the terms they place into relation:
The relations that connect experiences must themselves be experienced relations, and any kind of relation experienced must be accounted as ‘real’ as anything else in the system” (Essays in Radical Empiricism, p. 42).
A “system,” in this sense, is an open one, rather than a closed Hegelian or Structuralist totality; these “relations” are ones of exteriority rather than internal determination; and the “experience” that encounters them is not exclusively that of human beings, but applies to all the elements (terms and relations) involved. I am looking for a “speculative realism” that does justice to the multifariousness of relations, as well as to the multifariousness of things or substances.