In response to “A Psychoanalytic Defense of Realism“, Chris over at the great science and psychology blog, Mixing Memory, asks:

OK, maybe I am missing something, but both your example from psychoanalysis and the nature of science seem to imply only one thing: that there is something out there that is the Other or the object of science. It doesn’t suggest that we can know anything about it. One could explain both with a Kantian position about the thing-in-itself’s unknowability, for example. So there does seem to be a leap not, as a previous commenter said, from epistemology/ontology to realism/anti-realism in epistemology, but from a realist ontology to a rejection of certain kinds of epistemology.

I thought this question might have occurred to others as well so I’m posting my response to Chris here, with a few additions, for anyone who might be interested in the ontology advocated by Bhaskar’s early transcendental realism.

Bhaskar’s realist claims are a lot more robust than a sort of nod to the Kantian in-itself. Bhaskar’s question, it will be recalled, is “what must the world be like for our science to be possible?” Where the transcendental idealist asks “what must our cognition to be like for such and such a type of judgment to be possible?”, Bhaskar instead inverts the question and makes it one about the world itself. He argues that minimally intransitive objects (objects independent of mind or society) must be causal mechanisms (what I’d call “difference engines”), that are structured and differentiated, and that act regardless of whether or not humans know about them or exist.

read on!

Additionally, he argues that these objects, difference engines, or causal mechanisms must be 1) capable of acting independent of experience and 2) independent of being actualized in an event in the world. Thus Bhaskar distinguishes three different levels that he believe are often confused in the philosophy of science:

1) The empirical or the level of what we sense

2) The actual or natural events that take place in the world.

3) The real or causal mechanisms, objects, or difference-engines.

Bhaskar’s thesis that causal mechanisms must be conceived as acting without producing an experience or a sense impression (the first level) or a an actual event (second level) might initially appear perplexing, so I’ll try to briefly explain why he is making this claim even though I’m dog tired (so please be kind!). One of Bhaskar’s major gripes with contemporary philosophy of science is that it tends to embody an implicit positivism and phenomenalism. This, for Bhaskar, is especially the case with how philosophy of science conceives causal claims. For this form of philosophy causal claims are not claims about objects themselves but about regularities between events in experience. These events are not to be confused with natural events but rather are phenomenal events or sense impressions. Under this interpretation a claim about causality is really a claim about two sensations constantly linked together in experience. Here we have a perfect example of an ontological issue (what causality is) being reduced to an epistemological issue (how we have access to causality).

For Bhaskar, this conception of causality renders scientific practice incoherent. The problem is two-fold: First, in ordinary experience the consequent of an event often doesn’t obtain in a regular fashion, i.e., the antecedent occurs without the consequent following. This should lead us to reject the causal claim as ungrounded, yet we do not. If a causal claim is a constant conjunction between sense-impressions and if such constant conjunctions almost never obtain in experience, then it is unclear as to why we would advocate the causal claim at all (here Bhaskar is really taking a shot at Popper’s model of falsification). Second, Bhaskar believes that philosophers of science don’t pay enough attention to the difference between closed and open systems. The reason that scientists create highly specified and controlled experimental settings is, according to Bhaskar, to get at the things-themselves. In open systems other causes or causal mechanisms intervene, preventing the antecedent from producing the consequent. The causal mechanism continues to act, according to Bhaskar, but it doesn’t produce a particular result because of these other factors.

What the closed system of an experimental setting allows is the release or “trigging” of these causal mechanisms in an environment where these other causal mechanisms are not intervening or interfering. In this way, the scientist occasionally discovers the causal power of a particular type of thing itself. In other words, what the scientist is interested in, according to Bhaskar, is not regularities of experience (generally these regularities don’t appear in experience because they are functioning in an open system), but rather causal powers of things themselves. When these powers are found, argues Bhaskar, we are warranted in making what he calls “transfactual claims” (he critiques induction and dismisses it as a model of science), where the “transfactual” is understood as the thesis that the causal mechanism (the thing itself) acts in open systems even though it doesn’t produce an experiencable event for a subject or observer and even when it doesn’t produce an actual natural event because of other intervening factors. Again, this is just a thumbnail but hopefully it gives a sense of just why he is making a far more robust ontological claim than Kant’s claims about the mere existence of things-in-themselves. For Bhaskar, knowledge claims are not claims about regularities in our experience or constant conjunctions of sensible events, but rather about things-themselves that continue to act regardless of whether they produce an experience for us. Moreover, for Bhaskar we are not merely coming to know a constant conjunction of sensations but the things themselves that exist independent of humans and that would act regardless of whether or not any humans existed to know them.