Tuesday, February 09, 2010

Cognitive Shielding

Here's a concept I'm playing with and may soon have occasion to deploy in my work on introspection: cognitive shielding.

Normally when I reach judgments, the processes driving the judgment are wild. I don't attempt to control the influences on my judgment. I just let the judgment flow from whatever processes might drive and affect it. I look out the window and think about whether it will rain. I'm not sure what exactly causes me to conclude that it will. Presumably the appearance of the clouds is a major factor, but maybe I'm also influenced by my knowledge of what month it is and how common rain is this time of year. Maybe I'm influenced, too, by wind and by temperature, reflecting sensitivity to contingencies between those and oncoming rain -- contingencies I may have no conscious knowledge of. Maybe I'm influenced by knowledge of yesterday's weather, of this morning's weather report, and who knows what else. I don't attempt to control any of this, and the judgment comes.

Sometimes, I intentionally launch processes with the aim of having those processes influence my judgment. So, for example, I might think to myself: "In the northern hemisphere, storms spin in such a way that the wind of the leading edge tends to come from the south. So I really should consider the direction of the wind in reaching my judgment about the likelihood of rain." [How true this generalization actually is, I don't know.] I notice that the wind is indeed from the south and this increases my confidence that it will soon rain. The decision to consider a particular factor launched a process that would not otherwise have occurred, with an influence on the conclusion.

And finally, sometimes I try to shield my judgments from certain influences. Maybe I know that I'm overly pessimistic and am biased toward anticipating rain whenever I'm planning a picnic. I am in fact planning a picnic, and I don't want the resulting pessimism to affect my judgment, so I attempt to put the picnic out of mind or compensate somehow for the bias it would otherwise introduce. Or -- a familiar example for professors -- in grading student essays I might be legitimately concerned that my like or dislike for the student as an individual might bias my grading. I might attempt to compensate for this by not looking at the names on the essays, and then no cognitive shielding is necessary. But sometimes I do know who has written the essay I am grading. I might then try to shield my judgment about the essay's quality from that potentially biasing influence. Wild judgment might unfairly favor the student if I like her, so I try to reach a judgment uninfluenced by my opinion about her as a person.

Two issues:

(1.) It's not always clear whether some series of thoughts is wild or launched. Similarly for shielding. Possibly there is a large gray area here. But if the distinction between spontaneously considering certain factors and intentionally considering (or setting aside) certain factors makes sense -- and I think it does -- then I think these distinctions can fly, despite the gray area.

(2.) It seems that launching will normally be successful. Shielding on the other hand, may be difficult to execute successfully. One might try not to be influenced by certain things and yet nonetheless be influenced. But this is no objection to this taxonomy as long as it's clear that we can try to shield our judgments from certain influences.

Thoughts? Reactions? Does this make sense? Is there someone in the literature who has already laid this out better than I?


Gary Williams said...

Hey Eric,

I am sure that I am by now sounding like a broken record, but I think Julian Jaynes had a lot of insightful things to say on these type of issues.

What you call a "launch process" he called a "struction" (as in "instruction"). Discussing the 1901 Marbe experiment, which established that the process of judging is not introspectively accessible, he says:

"What seemed to be happening was that thinking was automatic and not really conscious...This was a remarkable result. Another way of saying it is that one does one's thinking before one knows what one is to think about. The important part of the matter is the instruction, which allows the whole business to go off automatically. This I shall shorten to the term struction."

He also says, in speaking or writing, "I am not now consciously entering my lexical storehouse and consciously selecting items to string on these syntactical structures. Instead, I have what can best be described as intentions of certain meanings, what I call structions, and then linguistic habit patterns which take over without further input from my consciousness."

I think this more or less overlaps with your distinction between wild and launched cognitive processes. Thus, in order to answer issue (1), I would say that, except in extreme cases,it is self-evident when we launch an intention because part and parcel of that intention is that it happens within a narrative structure such that "I" am making the intention, "I" know that "I" am making the intention, and that, moreover, this becomes stored within my autobiographical memory system as something I intended. To launch an intention you must know that you have launched it and be able to articulate why you have launched it, and so on.

As for (2), I think your assessment is very nice. I would add that in the case of a shielding failure, there is always the possibility of retroactive ignore-ance, in which case we could probably "overlook" or simply forget anything that slipped through.

Michael Metzler said...


I think it likely that any novel narrative structure is interpretively imposed after the thought of V and the launch of V - although perhaps previous to the linguistic judgment related to V. I have in mind Daniel Wegner's The Illusion of Conscious Will (2002) recommended by Dennett (2003). Wegner likes discussing 100 year old experiments too.

Gary Williams said...


If I understand your point correctly, I would say that there are probably two components to the narrative structure: the underlying functional substrate and the process which is "run" on that substrate. It very well might be true that the "run" occurs rather late in the total process and in such a way that it is retroactively imposed or something.

However, while I would agree with Wegner that the vast majority of goal-directed "action" is automatic rather than conscious, I am wary of any attempt to say that ALL intentions are illusionary or "interpretively imposed". Eric's thought experiment demonstrates that the "conscious" will does have some causal force insofar as it can provide an initial instruction, which sets off a causal chain.Sure, you can probably cash that instruction out in automatic terms as well, but as far as I am concerned, if it happens within the narrative I-structure, it is a legitimate form of top-down control and denying it's causal efficacy is simply bad phenomenology.

But perhaps I misunderstood your point.

makvan said...

our arguments eventually rely on our inuitions, so if we do our best to be fair while our intuitions differ, we can not get to a same result. the result of a reaserch shows that peoples with different nationalities may have different intuitions(I think I read it in"externalism about mental content" in SEP). doesn't it?

Eric Schwitzgebel said...

Thanks for the comments, folks! I'm just back from out of town and can't quite think straight yet. Coherent thoughts soon.

Michael Metzler said...


Thanks for the reply. Your first comment seems to make use of a computer metaphor of sorts; although this has been a popular approach traditionally, I am inclined to think that the architecture determines the nature of cognition more than this metaphor suggests. Although, there certainly seems to be a higher level aspect to 'narrative structure' in contrast to lower level cognitive mechanisms partially explained in neurobiological terms. If this is where you are headed, I would be in agreement.

As for the second point, I am skeptical. I am not sure that Eric has revealed causal force. Consider for example his statement: "So I really should consider the direction of the wind in reaching my judgment about the likelihood of rain." First, it is not clear to me that this is a prototypical thought (to but it mildly). Second, even if there was a prototypical thought, it need not play a role of intentional force; the thought that "I should consider X", for example, often fails to cause the consideration of X, and the consideration of X often fails to produce actions pertaining to X. Further, it is not clear to me that the thought that "I should consider X" amounts to a "decision", or any necessary part of a decision process. So the way seems cleared for empirical evidence to suggest that the causal connection between the linguistic interpreted thought "I should consider X", the consideration of X, and the later narrative judgment, is not only automatic, but importantly indirect. No?

Eric Schwitzgebel said...

Gary and Michael: I'm inclined to agree with Gary against Michael. I'm not fully on board with Wegner. The existence of illusions about the will does not show that the experience of conscious willing is always illusory. I also remember thinking that Terry Horgan had a pretty good criticism of Wegner at one point, but I now no longer remember its content and I'm too lazy to look it up right now. Nahmias also I believe points out that even if knowledge of the consciousness of willing is subsequent to consciousness of willing itself, and sometimes dissociated from it, that doesn't show that there is no conscious willing.

It seems a pretty skeptical view that holds -- if this is what you hold, Michael -- that we can have no intentional control over the process of our conscious thinking.

Eric Schwitzgebel said...

Makvan: Yes!

Andrey Fedorov said...

It seems a pretty skeptical view which holds [...] that we can have no intentional control over the process of our conscious thinking.

To the contrary, I think this follows directly from your own distinction. In particular, the existence of judgements made outside of our control (as you say, "in the wild") necessitates that all thinking is determined subconsciously.

To see why, consider the chain of "higher order" judgements of (a) whether or not it is worth to consider your biases in a given situation, and (b) which biases to consider. In all of the examples you give, it seems that the second-order judgement was "wild" - instead of grading the paper, you just happened to stop to consider your biases. But now please pause again and consider - why did you stop grading and start reflecting? Because whatever it is that made you reflect is surely going to determine the direction that reflection will take [1]. Now that you've found the reason, realize, of course, that there are biases at the heart of *that* choice and *those* judgements, as well.

So while we can find judgement to recursive on ad nauseam, we will run out of articulable biases pretty soon, and out of conceptually unique biases even sooner. This line of reasoning is like a small child having just discovered the recursive nature of asking "why? why? why?". If we are to ever finish, you have to stop and say "just because", make a judgement "in the wild", and surrender the basis of your reasoning to intuition and whatever wild things that come with it.

Hence, not only are you incapable of complete intentional control of your conscious thinking, but the very detail that recurses out of the realm of conceivability is precisely the determining factor of the direction your reasoning will take.

Therefore, you have no intentional control of your conscious thinking (because in order to, you'd first need intentional control of your intentions, and of those intentions as well, etc. :)).

My formal training is in math, not philosophy, so please forgive any inappropriate terminology.

1. Maybe you are protecting your "fair professor" self-image. Or maybe you're afraid the other students resent their peer for perceived bias. Or maybe you suspect a vocal student will complain to the department about a bad grade. Or maybe the sight of her handwriting, a trace of her perfume, or a smudge of her lipstick has sent your mammalian brain into a frenzy, and your neocortex justifies thinking of her only in the context of "legitimate concern". Or maybe something else? Take a moment, imagine alternatives, judge them, and decide.

Michael Metzler said...


Thanks for the thoughts. Yet, I had intended to accomplish the following (willfully): 1) differentiate Wegner's view on indirect causality from Gary's comment on automaticity; 2) note the fact - and an important one it still seems to me - that your stated example gave little or no evidence of intentional causality; and 3) conclude modestly that "the way seems cleared for empirical evidence to suggest that the causal connection between the linguistic interpreted thought 'I should consider X', the consideration of X, and the later narrative judgment, is not only automatic, but importantly indirect."

Concern of a "pretty skeptical view" - which, I suspect you would agree, cannot help escape some sort of moral connotation - is not a response I would have expected, however much I might be indebted to this criticism. "Intentional control over" seems to provide unclear reference to the subtlety of the issues, and filtered through Dennett's view on consciousness, 'illusion' is apparently not a necessary way of framing Wegner's point. It seems important to emphasize that the issue is the concept of 'intentional control' generally, and not whether the process of our conscious thinking is somehow uncharacteristically out of our grip. Lastly, I call attention to the difference between a) claiming that we have no intentional control of our thinking and b) claiming that "the way seems cleared for empirical evidence to suggest that" our ownership and control of actions is less direct than our folk conception. Unleash the wonders of science, I say, against all my intuitions!

Thanks for the reference to Horgan. As for Nahmias, I am not sure how the consciousness of willing could be separated from the knowledge of consciousness of willing, unless you are speaking of alleged propositional knowledge, in which I case I would agree in so far as I presumed such a thing existed. . .