Harman criticises the reductionism that he finds underlying Eddington’s account of the “two tables” paradox see my review here). However, Eddington’s own position already contains a far more effective critique of reductionism as based on false mental pictures inferred illegitimately from physics. This sort of illegitimate extrapolation of physical theory into paradoxical images is also the case for Harman’s account, with his false image of physics reducing everything to tiny components. Harman relies on a long abandonned paradigm of physical explanation as reducing everything to tiny irreducible atomic components.
The picture that Harman appeals to, that of a radical distinction between primary and secondary qualities, belongs first and foremost to a particular stage of physics, that of atomistic explanations, of which Eddington describes the process of dismantling. In which case, Harman cannot validly cite physics as backing up his partition, and so the distinction is no longer ontological or even epistemological, but contextual and pragmatic.
Harman sets up an ambiguous relation between what he calls “sensual” and “real” objects, where the objects of physics are consigned to the sensual realm, and so their qualities are re-categorised as secondary. The true primary qualities become in his ontological system the real qualities of real objects. This has the defect of obliterating the relation between real and secondary objects, which in the older view that he is criticising was a causal one: the real objects and relations were supposed to causally give rise to the sensual realm, and so there was no ontological rupture, merely an epistemological one, between the two.
The relation between real objects and sensual objects and relations is said by Harman to be one of “withdrawal”. This is an ad hoc addition to his system, as withdrawal is neither a sensual relation nor a real relation, and there is no place for it in the system. Withdrawal is a pure supplement, a protective device or epicycle, to plaster over the explanatory gap that arises once physical objects and causal processes are assigned to the sensual realm.
A second impossibility is what we might call “de-withdrawal”. How can a real object be conceived to manifest itself in any way? Withdrawal is not a scalar concept, it knows no degrees. Harman seems to recognise this problem, and appeals to a notion of emergence:
the elusive table number three, emerging from its components
This thesis has the unfortunate implication that the real table “emerges” from its unreal components.
Personally I am in favour of acknowledging the existence of many more than three tables. The problem comes when one table is promoted as the only real table, and the others are demoted to “unreal” status as “shams”. The table Harman proposes is a timid variant of the common sense table, separate and distinct in its autonomy. This separateness is not argued for, it just seems to respond to a “need”, in this case an intellectual need.
It must not be forgotten that Harman does not acknowledge different “levels” of reality, his ontology is not a continuum: there is just real and unreal (or “sensual”). The real object is “deeper” than the sensual. But it is not a deeper reality, it is the only reality, and the surface is not another level of reality, it is declared simply “unreal”.
It is a perfectly valid point to say that if OOO is so incoherent we should forget it and move on to something more positive, but I think that some useful and interesting lessons can still be extracted from OOO’s failure. This is why, despite encouragement from various sources, I never wanted to devote a whole book to OOO. However, a chapter in a book giving a vaster overview of the philosophical Zeitgeist seems to me to be quite acceptable.
My critique of Harman’s OOO highlights its dependence on the classical subject/object separation. This is a clear case of the conceptual regression his thought represents, despite grandiose claims to progress beyond Heidegger, Deleuze, Latour,etc. Harman’s thought is definitely pre-quantum theory, and pre-wave mechanics or electromagnetic field theory. That is to say that despite his attempts to immunise his theory from empirical critique, his claims do contain an empirical element that can be, and that already has been, falsified by developments in physics. Further, his rather idiosyncratic idea of time and his positing of a timeless unchanging realm of the real are decisive conceptual and empirical inadequacies.