Friday, 6 November 2015

Steve Hyman's Loebel Lectures

At the end of the third Loebel lecture by Steve Hyman, the motivation for some of the more radical sounding claims from the second lecture became clearer. The more radical claims included the idea that because decision making processes are subserved by unconscious processes, our sense of agency for such decisions is misleading. (At the workshop I glossed his 'not veridical' as simply false. He flirted instead with incomplete. But I think that 'false' captured the intuition.)

One slide from the second lecture went as follows:

The eliminativist view (self, minds agency as illusion) misses important facts.

Evolution seems to have invested a great deal in our capacity-limited serial processor (consciousness) that serves as an important connector with parallel processing unconscious systems.

Impairment in sense of agency: a toxins stressor
Sense of being controlled: psychosis
Impairment in recognition of minds: autism spectrum

Yet from a neuroscience view, intuitions of eg. of agency, are not veridical. Rather they are proxies for unconscious processes.

Now this starts with a softening move in the comment about evolution. Further, in three cases, a lack of a sense of autonomy or sensitivity to others' agency is pathological. But there is that robust final comment that intuitions are not veridical but proxies for something else.

And this suggests a different reading of the earlier comment about evolution as having provided us with a useful fiction rather than a truthful awareness. Steve suggested I had him right on this.

One example he gave, again before today, was that of unconscious grasp of grammatical rules. He reamarked:  "When I am talking, I am not thinking 'now I need a verb'. These systems are not accessible by introspection."

But in my own talk earlier, I argued that the was no problem with high level intentional actions and context coming together to guide lower level basic actions and these being underpinned by motor capacities whose working is not available to consciousness. To take the industry standard example, crossing the road may lead to ironwork-avoiding action but also draw on unconscious adaptation of footfall to cope with slippery leaves. Or to use Steve's own example, in giving a talk about agency I may need to say something about crossing the road. But expressing / exercising that thought can simply draw on unconscious knowledge of grammar. I don't see any threat to agency here. 

(It would be revisionary of an understanding of agency that did require that no part of an action was not available to consciousness. But surely there is no reason to think that that is widely held?)

However, today's lecture concerned addiction and offered a kind of Pavlovian story of how dopamine plays a role in training the brain to attend to the scene at hand, as marking that something in the scene is better than expected. This role is then usurped by drugs that mimic that function. The thought as then that because the biochemical explanation explained how it could be that plant extract chemicals could overcome human free choice, this was the correct causal story of what happened. But it is typically not a story that is available to, or even consistent with the rationalising stories told by addicts.

So that was the real context for the rest of the worry that agency is a myth sold to us for evolutionary advantage but not for truth. But it leaves open a much more modest almost disjunctivist possibility. Agency goes wrong in the case of addiction, to be sure, and knowledge of the predicament also goes wrong. That's the bad disjunct. But in the good, there is no inference from the presence of something unconscious to agency's myth.