Cognitive Phenomenology: The Role of Introspection

In my first post I isolated Irreducibility as the main thesis in dispute about cognitive phenomenology:

Irreducibility: Some cognitive states put one in phenomenal states for which no wholly sensory states suffice.

In this post I am going to write about the role introspection should play in helping us decide whether Irreducibility is true.

There are two extreme views:

Extreme Optimism: Introspection alone can put us in a position to know whether Irreducibility is true.

Extreme Pessimism: Introspection cannot make any difference to our position to know whether Irreducibility is true.

To my knowledge, no one endorses either view. Fans of introspection generally supplement their introspective reports with further argumentation. Foes of introspection generally allow that our introspective reports at least form a data set that can be drawn on in theorizing about the mind. Still, I believe that Extreme Optimism and Extreme Pessimism are useful focal points: examining reasons to reject them does not serve any ad hominem purpose, but it does bring to light considerations relevant to developing a more nuanced view about the role introspection should play exploring cognitive phenomenology.

The picture suggested by Extreme Optimism is that just as you can tell by introspection whether you are having a headache, or whether it feels different from the itch on your ear, or whether it is sharp, you can tell by introspection whether Irreducibility is true. On its face this seems incredible. Suppose you can tell the following by introspection: that you are thinking about dinner, that your thought and your hunger are phenomenally different, and that your thought is intruding on the course of more lofty philosophical reflections. None of this suggests you can tell whether Irreducibility is true by introspection. Irreducibility is a (i) logically complex (ii) generalization about (iii) possible (iv) explanatory relations. The claims we typically know by introspection are (a) logically simple claims about the (b) actual (c) intrinsic properties of (d) particular mental sates. So there are these four differences—(i) instead of (a); (ii) instead of (d); (iii) instead of (b); and (iv) instead of (c)—that distinguish Irreducibility from the claims with respect to which introspection generally inspires confidence.

Consider also the fact that even after introspecting philosophers fail to agree about Irreducibility. Why is there no consensus? Bayne and Spener (“Introspective Humility.” Philosophical Issues 20 (1): 1.) helpfully distinguish four candidate explanations.

First, two philosophers might fail to agree because they simply have different experiences. Second, two philosophers might fail to agree because they use some terms differently, for example “phenomenal state.” Third, two philosophers might fail to agree because they have different background theories that bias their assessment of the introspective evidence. Fourth, two philosophers might fail to agree because introspection is insufficient to settle the question of whether Irreducibility is true. This is an example of what Bayne and Spener call operational constraints.

Here is a perceptual example illustrating the phenomenon. Suppose Al and Beth wonder what is inside an opaque box. They each have different conjectures. Suppose, further, that the only data they rely on in determining which conjecture is correct is what they can tell by looking at the box. It would be no surprise if Al and Beth fail to come to an agreement. Visual perception does not enable one to tell what is inside opaque containers. It just does not work like that. Maybe there are similar operational constraints on introspection. Introspection will tell you whether you are thinking about dinner, whether your thought and your hunger are phenomenally different, and whether your thought is intruding on the course of more lofty philosophical reflections. But it will not tell you whether Irreducibility is true. If this is so, then Extreme Optimism is false.

It is reasonable to suppose that there are at least some experiential differences, terminological variations, and divergent biases among philosophers. And it is reasonable to suppose that these do contribute somewhat toward the lack of consensus about Irreducibility. I doubt that they constitute the whole story though. Operational constraints are at least part—and probably a large part—of the story as well. Further, the applicability of this form of explanation to the failure of agreement about Irreducibility is a typical instance of a general phenomenon, sometimes exhibited by introspection with respect to other disputes about the mental and sometimes exhibited by other sources of knowledge with respect to various disputes about their own domains.

Consider the following disputed claims about mental states and observable properties:

Relationalism: the phenomenal character of a sensory state is constituted by the objects it makes one directly aware of.

Dispositionalism: each color is a disposition to cause sensory states with a certain phenomenal character in normal observers in normal conditions.

I might be able to tell by introspection whether it looks to me as if there is a computer in front of me, without being able to tell by introspection whether Relationalism is true. The first is about a mental state’s introspectable features. The second is about its underlying nature. I might be able to tell by perception whether a tomato is red, without being able to tell by perception whether Dispositionalism is true. The first is about the instantiation of a perceptible feature. The second is about its underlying nature.

Even after introspecting and perceiving philosophers fail to agree about Relationalism and Dispositionalism. Why? The central explanatory fact, it seems to me, is that introspection and perception alone do not settle these claims. Premises supported by introspection and perception can be used in philosophical arguments for or against Relationalism and Dispositionalism. But the extra philosophical argumentation is required. In general these sources of knowledge tell us about the “observable” aspects of their domains of application, but remain silent on the underlying natures of the items in those domains.

The natural continuation of the pattern is this: Extreme Optimism is false; introspection can support premises that might be appealed to in philosophical arguments for or against Irreducibility; but introspection alone cannot tell us whether Irreducibility is true or false.

This is a middle ground position. Another possibility, however, is to shift to the other extreme—Extreme Pessimism. In his well-known paper on the unreliability of introspection (“The Unreliability of Naive Introspection.” Philosophical Review 117 (2): 245–73.), Eric Schwitzgebel writes:

The introspection of current conscious experience, far from being secure, nearly infallible, is faulty, untrustworthy, and misleading—not just possibly mistaken, but massively and pervasively.

Schwitzgebel argues for this view by citing a number of disagreements about “current conscious experience”. One of those disagreements is the disagreement between proponents and opponents of cognitive phenomenology. If introspection is as unreliable as Schwitzgebel suggests, then perhaps Extreme Pessimism is warranted.

Let us make explicit how one might argue from disagreement about Irreducibility to Extreme Pessimism. There are two pieces of reasoning to be considered. The first is for the claim that introspection is unreliable with respect to Irreducibility. The second is for Extreme Pessimism. Here is how we might put the first:

(1) There are persistent disagreements among mentally similar, terminologically calibrated, unbiased, and equally introspectively reliable philosophers about Irreducibility

(2) If such proponents and opponents of Irreducibility persistently disagree, then it is because the proponents introspect that Irreducibility is true and the opponents introspect that Irreducibility is false

(3) Therefore, the capacity for introspection is unreliable with respect to Irreducibility.

Premise (1) in the argument rules out explanations of the disagreements in terms of experiential differences, terminological variations, or the influence of background theory. It also rules out explanations in terms of a difference in capacities for introspection. Let us assume premise (1) is true. The conclusion (3) plausibly follows from (1) and (2). So if there is any problem with the argument is must be with (2).

And indeed there is reason to suspect (2). As just noted, introspection alone might be silent about Irreducibility. Suppose we add operational constraints as a possible explanation of disagreements, or failures of agreement, about Irreducibility. Then we should revise (3) to:

(3*) Therefore, the capacity for introspection is unreliable with respect to or silent about Irreducibility.

Now one might think that there is still a continuation of the argument that results in Extreme Pessimism. This is the second piece of reasoning:

(4) A capacity that is unreliable with respect to or silent about Irreducibility cannot make any difference to our position to know whether Irreducibility is true.

(5) So introspection cannot make any difference to our position to know whether Irreducibility is true—i.e. Extreme Pessimism is true.

The idea is this. If introspection is unreliable with respect to Irreducibility, then we should not rely on it in determining whether Irreducibility is true. If introspection is silent about Irreducibility, then we cannot rely on it in determining whether Irreducibility is true. Either way introspection will not help us in determining whether Irreducibility is true.

Premise (4) is false. A capacity that is unreliable with respect to or silent about Irreducibility cannot settle on its own whether Irreducibility is true. But this is different from making any difference to our position to know whether Irreducibility is true. Even if it is unreliable with respect to or silent about Irreducibility, introspection can make a difference to our position to know whether Irreducibility is true by putting us in a position to know other facts about our mental states that can be appealed to in philosophical arguments for or against Irreducibility.

In order to rule out this possibility the argument from disagreement would have to be strengthened so that it is about all claims about mental states that could figure in arguments for or against Irreducibility. Premise (1), for example, would have to be revised to read: There are persistent disagreements among mentally similar, terminologically calibrated, unbiased, and equally introspectively reliable philosophers about all claims about mental states that could figure in arguments for or against Irreducibility. Replacing “Irreducibility” with “all claims about mental states that could figure in arguments for or against Irreducibility” throughout would yield an argument that establishes Extreme Pessimism.

But there is no reason to think such an argument is sound. Philosophers disagree about Irreducibility. But they tend to agree about claims such as the following:

  • I can immediately tell whether I am consciously thinking that 3 is prime.
  • There is a felt difference between hearing “Dogs dogs dog dog dogs” as a mere list and hearing it as a meaningful sentence.
  • There is a felt difference between merely entertaining the thought that if a < 1, then 2 – 2a > 0 and intuiting that if a < 1, then 2 – 2a > 0.
  • Achievements such as grasping what “Dogs dogs dog dog dogs” says and intuiting that if a < 1, then 2 – 2a > 0 contribute to making life interesting.
  • Unlike sensory experiences, thoughts occur in a discrete sequence and do not pass into one another by gradual transition.
  • We never experience, and we find it difficult to imagine, conscious cognitive states occurring absent any conscious sensory states.

There is no reason to think there are persistent disagreements among mentally similar, terminologically calibrated, unbiased, and equally introspectively reliable philosophers about these claims. And, as I’ll get into in the next post, these are the sorts of claims that figure in arguments for or against Irreducibility.