Tuesday, March 29, 2005

Wanted: New Theory of Causality

Unacknowledged problems with the concept of causation plague philosophical debates. Take, for instance, the issue of the causal closure of the physical world. This seems to require that first person experience must be either epiphenomenal (no causal impact) or, if we insist it can cause, we are left with interactionist dualism (which fails because it has no plausible way to connect the physical and non-physical worlds).

This is the heart of the free/will determinism debate. To say a human has free will, we must reject causal closure or find some way to redefine free will to make it seem compatible with determinism.

The real problem is that we do not have an adequate concept of causality which fits the true workings of the natural world (of which humans are a part!)

In many of these debates, it seems pretty clear that people on both sides have in their minds a causality based on something like classical mechanics. While there is a philosophical literature on causation which highlights the controversies surrounding the concept, I rarely see it invoked in philosophy of mind or free will discussions. The notable recent exception I’ve discussed here is Gregg Rosenberg’s book, A Place For Consciousness, which presents a detailed new theory of causality and connects it to the body/mind problem. I'll be going through the book soon for a third time and will continue to assess its causal model.

Here are 3 reasons to think we need a new treatment of causality:

1. Quantum mechanics has overthrown the classical picture, but when philosophers bring it into the discussion, they oversimplify by assuming it is a simple probabilistic version of the classical.

2. Complex non-linear systems resist reduction to classical causation, but no one has offered a new explanation, except to say higher level features somehow emerge.

3. Remember also the difficulty we have of finding an explanation of the asymmetrical "flow" of time, in which causality takes place. This concept is not part of the laws of mechanics. While time asymmetry arises in the 2nd law of thermodynamics, this statistical law dangles unconnected to the other laws of physics.

Let’s start with the case of QM, where the causal process is indeed richer than the classical one. We have a unitary process of evolution described by the wave function, and then we have a second process: measurement. There is obviously more going on here than in the classical picture of billiard ball ‘A’ effecting billiard ball “B”. One or both systems involved in a measurement need to have an additional (natural) property in order to have a quantum causal event. This “ability to measure” or “ability to observe” or “ability to receive information” property is an integral part of the picture.

A new concept of causality should make this “ability to measure” property explicit and work out the role it plays in construction of natural systems, including complex macroscopic ones like us.

With regard to complex systems, they appear to manifest a binding or coordinating dimension to causation which supplements classical micro-causation of the parts of the system. A new theory should include an explanation of this phenomenon.

With regard to time, all the evidence shows that the asymmetrical flow of time and causality is an consequence of the first person perspective, rather than an objective feature of nature. A new theory of causation should show how the causal interactions of an individual system constitute the flow of time.

The best case is that these 3 requirements come together in a new theory. The theory would explain how a natural system implements quantum events, coordinates them, and generates a system-relative flow of time.

Wednesday, March 16, 2005

Perception Grounded in Action

This new paper by Alva Noë describes his action-oriented approach to perception. Interestingly, he discusses the “constancies” of perception quite a bit and disagrees with Sean Kelly’s account (discussed here) on some points. Whereas Kelly stresses two modes of perceptual attitude (the engaged and the detached), Noë views both ways of perceiving an object (with and without constancy) to be concurrent – a “two-dimensional” experience. I tend to prefer Kelly’s take on this. But much more important than these differences is what I take to be the two philosophers’ broad agreement on the main way to approach the problem of perception: that is, perception is grounded in bodily engagement with the world.

According to Noë, one’s relation to perceived elements of the environment is essentially sensori-motor. Even in the case of vision (which somewhat unfortunately dominates discussion of perception), our perception of objects is shaped by how we could move and probe the object in its environmental context. This is how we perceive the missing part of the object which is partly obscured behind a picket fence, and how we view the penny to be round despite its being turned at an angle which presents an elliptical image to the retina.

This approach is the best way to address the long-standing philosophical “problem of perception” discussed in this SEP article recently posted. Historically, different approaches to the problem suffered from explicitly or implicitly viewing perceptions as representations or other entities in the mind. Instead we should view perception as actually based on direct interaction with the world. How does this address the phenomena of illusion and hallucination? If perception is direct contact with the external world, how are these possible? Well, through other causes impacting our body/brain, it becomes as if we have the perception. But if we cut out the middle-man of sense-data or representations, the existence of illusions or hallucinations shouldn’t lead to radical skepticism about the reliability of normal perception.

Thursday, March 03, 2005

Another Hobbit Post on The Loom

Can't get enough hobbit blogging. Here is the link. The Loom has had a great run of other interesting posts lately as well.

(Note to visitor: I'm sorry I haven't yet introduced a hack which would remove the tag referring you to "continue reading" on short posts where there isn't any more to read. I hope to fix this at some point, but right now I can either make it appear on all posts or none.)

Wednesday, March 02, 2005

The Limits of Introspection

In the talk I mentioned in my recent post, Sean Kelly discussed the ways one might account for the “constancies” in perception. This refers to the phenomenon whereby the color, size, and/or shape of an object is perceived to be the “same” from different perspectives (distance from object, angle, lighting), even as the details of the image presented to our retinas shift. An example used in the talk was a penny turned at an angle: while presenting an elliptical image, we still perceive it as round.

The key to the discussion was that to correctly characterize our perception in these cases one must distinguish between two modes of perceptual attitude which Kelly called the engaged and the detached. The engaged mode is our mode the majority of the time, when we are immersed in our activities. This mode, shaped by evolution and early development, is one where perceptions are guided (essentially automatically) in ways which serve our bodily goals (“motor-intentionality”). The detached mode describes the introspective investigation of our perceptions we undertake when we “step back” and reflect on a phenomenon.

The engaged mode is the one in which the “constancies” properly reside. In contrast, an account of perception which explicitly or implicitly makes use of the detached mode leads to misleading accounts of the phenomenon.

I’ve said in the past that investigations of consciousness which assume the primacy of what I’ve called the reflective or introspective mode of self-consciousness can be flawed. This introspective capability (which is of recent historical vintage and almost surely unique to humans) is part of our highest cognitive skill set, and yet, when applied to our own conscious experience, can obscure as well as illuminate the details of that experience.

This is discussed in detail in this SEP entry, “Phenomenological Approaches to Self-Consciousness”. The authors outline a distinction between reflective and pre-reflective self-consciousness with which I concur.

One place these limits come into play is in Benjamin Libet’s work on the temporal dimension of conscious experience. The experiment I’m thinking of is the one where subjects are asked to flick their wrists at a time of their choosing. Libet timed the wrist action and compared it to both a measured “readiness potential” detected in the motor cortex of the brain as well as the subjects’ own reports on when they experienced the desire to perform the action. Famously, the readiness potential preceded the reported time of experienced volition by an average of 0.5 seconds. Now, I think much of the commentary on this experiment suffered from a simplistic separation of the subjects’ internal states into the unconscious (the readiness potential) and the conscious (the report). By reporting on the experience, the subject had to bring the experience into reflective self-consciousness. This probably introduced a delay beyond what would be the case in the mode of pre-reflective self-consciousness (which couldn’t directly be measured in this setup). Nothing I’ve said means the experimental findings are not interesting and meaningful, but just that it may tell us more about the limits of the reflective mode of self-consciousness than it does about the divide between the conscious and the unconscious.

One final note I find interesting: the phenomenological account of self-consciousness in the SEP article I noted above dovetails nicely with the account described in this previous post, which was derived within the context of analytic philosophy.