September 9th, 2009

Some further comments on “object-oriented philosophy”, based on a personal communication with Nick Srnicek:

Gabriel Catren has come up with a clever argument which Nick Srnicek kindly pointed out to me; it’s summarized pretty well in his article “Can Classical Description of Reality Be Considered Complete?” in which he essentially argues that quantum mechanics is not an incomplete description of reality; for mathematical, aesthetic, and symmetry reasons, one can argue that it is classical mechanics, which specifies both the precise position and momentum (i.e., velocity) of a particle, which is overly specific. While I have objections to the idea of objects, I think Catren’s argument is a clever way to capture what is probably the intuition of most physicists today, and has certainly always been my intuition: that there are no “hidden variables” which allow a particle to take up a definitive position and momentum at any given moment; that the Uncertainty Principle represents a fundamental quality of nature, not a limit to our ability to know underlying reality. In essence, he’s arguing against Einstein more than anything else. If I have any dispute with his argument it’s perhaps his desire to use the word “objective” over and over — I suspect he has a sort of attachment to the idea that physics is discovering “objective” properties of the universe which are driven by mathematical necessity or elegance. I certainly think it’s reasonable to posit that the universe has stable objective properties, but no matter how elegant the math we can’t know for certain that the patterns we have uncovered are that way out of necessity, or that future discoveries won’t replace our current models with radically different ones with completely different terms; though, naturally, those new models wouldn’t completely invalidate the old ones, as typically all paradigms capture some patterns in the world which are beyond the purely subjective with some degree of accuracy, or we wouldn’t adopt them at all.

My objection to “object-oriented philosophy”, however, isn’t an attempt to suggest that reality has no objective structure (even if we can’t know what that structure is for certain). The patterns we’ve uncovered so far do seem not only moderately stable but we may even be able to speculate as to their logical necessity, though not without positing postulates of some kind which themselves can’t be known to be correct for certain. In other words, I don’t doubt there is some non-subjective aspect to the structure of the universe which is reflected to some degree in our models of the world, but this mostly has to do with structures at a very basic, underlying level; a totally different level of organization from the objects that people ordinarily talk about, yet these objects are being considered for ontological status by object-oriented philosophy. What Catren is talking about, and what fundamental physics is about, is at a very different level of organization. Furthermore I believe the measurement problem in QM, which hasn’t been resolved, is another very important factor here (Catren’s argument doesn’t resolve that problem; he’s talking about the uncertainty principle, which is related to the measurement problem but by no means the entirety of it.)

Another problem with object-oriented philosophy is that I do not believe it works even if you assume a classical (i.e., non-quantum) world. There are thinkers like Gregory Bateson and Brian Cantwell Smith who start with the presumption of a classical universe, and then reason about how cognitive processes would work in such a universe. There’s no bizarre uncertainty principle, no measurement problem, just particles or a mass-energy flux of some kind, and they ask: what would creatures who are aware be like in such a universe? When you look at what they discover, you end up seeing all sort of odd things pop up: in particular you see that self-reference (loops of information or differences travelling around causally-connected loops) becomes central, and you see that while thinking about self-referential causal loops you end up seeing a lot of the subject-dependence and subject-object coupling claimed by many thinkers who Meillassoux and others are calling “correlationist”. So, even if you are not a radical anti-realist, there’s a lot to be learned by looking at “correlationism” which cannot be easily understood in simple terms as an artifact of translation by an object A sending differences to a receiver object B. Decomposing information and subject-object relationships into a point-to-point translation I don’t think works, in other words, even when you’re simply talking about purely classical systems. Both Bateson and Smith end up saying that the observer and observed end up coupled in a way where you cannot separate them cleanly, and Smith ends up saying that the “natural” perspective for cognitive systems operating in a classical world is one which has striking resemblance to the phenomenological picture, centered on the subjective — even though, as I say, he begins with a provisional assumption of a classical world of a matter-energy flux. None of this involves privileging the human version of subjectivity, nor does it imply a radical anti-realism; but it’s remarkable how many “correlationist” insights seem to naturally arise even when you start from a presumption of a classical world. (Naturally, Smith recognizes that assuming the world is comprised of a classical mass-energy flux is known to be incorrect, but it’s interesting to see what happens when you DO assume that.)

Object-oriented philosophy is placing itself within the context of a larger movement called “Speculative Realism”, which apparently was inspired by Meillassoux. I’ve written earlier about my problems with his thinking, especially the “ancestrality” argument which I think is fairly bogus, but there are a lot of different currents in speculative realism, and while I find fault with Meillassoux as well as with the notion of objects, I find myself in general agreement with the idea that one can talk about, or at least speculate about the structure of, a subject-independent ground of Being. That is to say, the thesis that the universe has some sort of patterned regularity to it which we can investigate, reason about, speculate about, and model, to me seems like a postulate which one can adopt without much risk. Yet, even while taking up this view, the insights of the “correlationists” remain not only valid, but profound; hardly “trite” or trivial as some are claiming.

I think many of the “correlationist” insights can be related to very tricky and interesting issues with cybernetic systems that have cognitive properties (i.e., are computational systems in some sense), as well as interesting problems in theory of computation. Correlationism, in other words, if not taken to an anti-realist extreme, I believe can be made compatible with what people are now calling speculative realism; just as quantum mechanics and classical mechanics have areas of strong overlap. I agree with SR that you can go farther than pure “correlationism”; in fact I’ve taken this approach myself and encountered some resistance to some of my speculations (in which, like Bateson and Smith, I start with an assumption of a physical naturalism, which causes some of my more anti-realist friends to look at me with some suspicion); but the fact that you can say more doesn’t mean that what correlationists say is trivial or even inconsistent with some realist perspective of some kind (as long as you’re scrupulously avoding any sort of correspondence theory).

A further comment on Catren’s argument; as I said above, it addresses the uncertainty principle but it doesn’t resolve the measurement problem itself, which remains unsolved and still very mysterious. An interesting result I read about a number of years ago suggests that if you take the universal wave function (i.e., everything in the universe for all time) and you think about it in terms of the relative state interpretation, it suggests that the universal wave function, taken as a whole, has zero information in it, i.e., it is just in some sense empty, uniform (I’ll have to try to find that reference at some point). If you think of it in this way, then “objects” as we experience them are dependent upon cognition in the sense that cognition (as a sort of information loop) is correlated with objects via rules which may themselves be somewhat objective, but the objects are nevertheless strongly dependent upon the observational process. I.e., if relative state is correct, then this paper implies that every “thing” we see is “carved out” in some sense from a formless block of potentiality which has mathematical properties but no “things” in them a priori — the things exist because of the information loops/awareness. Of course many worlds is just one of many interpretations, but it is quite popular among physicists and it happens to be my favorite. This sort of interpretation ends up sounding correlationist — yet it doesn’t privilege the human subject at all, and it doesn’t in any way suggest that reality has no structure whatsoever. It may have a very exact structure, yet still not have “things” in it totally independent of observation.

However: even if something like this “featureless universal wave function” perspective is in some sense “correct” — and there are actually ways to run experiments about this, to distinguish it from, say, Bohmian mechanics or the transactional interpretation, contrary to popular belief — this doesn’t mean that if objects only exist in relation to feedback loops that one cannot talk about independent objects in some approximate sense. Because even if it requires some sort of perceptual feedback loop to “carve out” the object from a featureless universal wave function, because the universe has an objective patterned-ness to it, this means that one cannot just “change the rules” of how these objects behave by, for example, merely changing one’s ideas about them. Obviously, for example, rocks can be modelled as independent objects and they will follow to a large degree many of the rules of classical mechanics. I.e., once you “see” a rock, then it is going to behave like a rock, regardless of what you believe about it, so there remains something objective and mathematical about the rock, at least insofar as the rock is dependent upon some underlying structures which are themselves part of the Being of the universe.

However I believe it’s also important to realize that the rock, and the geological processes operating on the rock, and the system that you’re a part of including the rock and geology and ecology and so forth, all can be seen, in some sense, to be part of a single system that isn’t separated out into subject and object as clearly defined parts. That is to say, whatever the underlying basis of the rock, the sense in which we define a rock as a rock may not be mappable in a one to one, clearly-defined fashion, to that underlying reality; so one has to include the whole information processing loop, the entire ground of Being, as a encompassing system, to really understand what we mean when we talk about a rock. That’s the crux of my concern about OOP — I agree with SR that one can talk about underlying reality in non-human-subjective terms, that there is some sort of ground of Being that has some patterns we can “compress” using Dennett’s terminology; but that we can fruitfully decompose it into *objects* which we can think of as completely independent and separable (except, perhaps at the subatomic level) I find highly questionable.

permalink | comments(rss)

comment trackback