Thursday, September 2, 2010

EC or not EC?

Deric Bownds here suggests that this research in which emotional negativity was reduced for participants who placed a written recollection of a regretted past decision or unsatisfied strong desire inside an envelope provides a case of embodied cognition.

What do you say if your are Andy Clark?
What do you say if you are Rob Wilson?

My guesses: No. Yes.  Andy, I think, (as I've posted before) thinks that one has to have informational resources in order to have extended cognition, but Rob, I think, does not have this requirement.  (Why I think this about Rob will only come out in my forthcoming posts on his "Meaning making the and the mind of the externalist".  Please stay tuned.


  1. Voting 'no' on the grounds that it isn't. :)

  2. Andrew,

    I'm sort of surprised by your answer, given the following you had in an earlier post:

    Step 2 is to identify what is performing that work. If any of the necessary work is being done by something other than the brain, then it's surely game on as far as embodied cognition is concerned.

    If you have some time, maybe add a clarification. (I'm hoping to get a bit of work done on the long weekend ...)

  3. By 'work' I mean 'necessary work'; this is why a task analysis at the right level is so important. In this example, the reduction of emotional negativity just reflects a side effect of the task they had people do, not any necessary component required to actually achieve that task, and it's therefore not embodied anything.

    Let me know if this makes sense: I've been banging this idea around for years and I always have great difficulty getting cognitive-type people to see what I mean; I've never been entirely sure of what the barrier is and I'm keen to try and figure it out :)

  4. Ok. So, suppose that the task is to reduce emotional negativity. In this case, is putting something in the envelope part of the agent's extended cognitive processing? Does it really matter if using the envelope is the only way to reduce emotional negativity, hence that using the envelope is (at least in some sense)necessary work?

  5. The measure was of emotional negativity; you could have gotten this number with or without the envelope bit (as I think they show in a control experiment). The act of putting the past regret in the envelope did indeed moderate that measure, but it wasn't in any sense required for the measure to occur.

    Even if the task was 'reduce emotional negativity', the envelope still isn't necessary! It just so happens that using it lowered emotional negativity - that's not the same as saying it's literally impossible to lower emotional negativity without an envelope.* That fact means that envelopes (or things invoking enclosure, more generally) then remain separate from the cognitive system of interest, (the one that reduces emotional negativity) and there's no way it is a non-neural, extended element of that system.

    A good task analysis is an exhaustive list of only the necessary and sufficient conditions for something to happen. An example**: I think 'moving so as to be at the right place at the right time so as to be able to intercept a fly ball' is a better task analysis for the outfielder problem than, say, 'predict where the ball is going to land and move to there'. Why? Because the former makes no assumption about implementation while the latter has already moved to 'prediction' as the mechanism. The former is good because it sticks to what has to be true: to catch a fly ball you do indeed have to move from where you are to where you need to be. If any of that is not achieved, you fail at the task. That latter is bad because it is not necessarily the case that prediction is involved; if prediction doesn't happen you can still catch a fly ball (as evidenced by all the studies on the perception-action strategies used).

    *A side note - writing sentences that weird also makes me feel we're not talking about a task that makes any sense.

    **Another side note: I always go back to perception-action examples because a) they're what I know and b) because they are much more robust examples. These weird, abstract 'purely cognitive' tasks are a complicated mess that frankly may yet just go away when psychology stops assuming things about mechanisms. But that's another story.

  6. It seems like "task analysis" is a technical term for you. It's not just a bit of the vernacular.

    Now, it seems to me that "moving so as to be at the right place at the right time" and "predict where the ball is going" do indeed differ with respect to what they presuppose is going on. But, to me they just look to be different tasks. The first task does not by definition require thought, where the second task at least appears to. It might turn out that humans do not catch fly balls by prediction, but that would be just to show that they do not perform that task.

    I have not encountered this conception of a task analysis before. So, I'll have to mull it over.

  7. It might turn out that humans do not catch fly balls by prediction, but that would be just to show that they do not perform that task.
    Precisely. My (methodological) point is that you should at least start by describing the task the first way because it leaves the door open for data to show that the second way is the specific implementation, if indeed it is. Psychology is, however, firmly in the grip of the assumption of computational representation and they therefore start too far upstream. This is a real problem, I believe.

    I need to work on a blog post about this though (so your comments are helping me a lot to plan that out :) - it's a very important issue that gets very little attention, and frankly it's at the heart of the Gibsonian critique of things so it's worth attending to (for me, anyway :)