Some Questions on Heterophenomenology

Here are two things Dennett says about heterophenomenology:

(1) Scientists should interpret a subject’s first-person reports as expressions of the subject’s beliefs (about their consciousness experience)
(2) Scientists should treat people as incorrigible about what it’s like to be them.

Since (2) seems to contradict Dennett’s often repeated claim that people can be wrong about their experience, let me give you a quote:

“You are not authoritative about what is happening in you, but only about what seems to be happening in you, and we are giving you total, dictatorial authority over the account of how it seems to you, about what it is like to be you” (Dennett 1991, 96).  (In other places, Dennett uses the term “incorrigible”.)

I’d like to know what (1) and (2) amount to, whether they are logically independent, and why we should believe them.

Claim (2) has been extensively dissected and criticized by Eric Schwitzgebel.  Eric published a helpful paper on this topic, which was followed by a reply by Dennett, which was followed by further discussion on Eric’s excellent blog.

After much searching, Eric settles on the following interpretation of (2) (which has been endorsed by Dennett via email to Eric):

“The key idea is that there are two sorts of “seemings” in introspective reports about experience, which Dennett doesn’t clearly distinguish in his work.  The first sense corresponds to our judgments about our experience, and the second to what’s in [the] stream of experience behind those judgments.  Over the first sort of “seeming” we have almost unchallengeable authority; over the second sort of seeming we have no special authority at all.” (from here.)

So according to Eric, when Dennett says subjects are authoritative about “what it’s like to be them,” he means they are authoritative about their judgments.  Eric also says that he “can agree” with Dennett on this point.  But I’m still confused, so I can’t figure out whether I should agree.

First, where does the “almost” before “unchallengeable” come from, and is it anywhere to be found in Dennett’s writings?  As Dennett’s quote shows, Dennett wrote “total, dictatorial authority” (and elsewhere of “incorrigibility”).  And what exactly does it mean that something is “almost unchallengeable”?  “Almost P” normally implies NOT P.  If it’s almost unchallengeable, it must be challengeable after all.  I would agree with that, but then I wouldn’t know what all the fuss is about.  Because of this, I will drop the “almost” from Eric’s interpretation, pending further clarification on this point.

Second, how should we understand Eric’s talk of “judgments,” such that people have unchallengeable authority over them?  One option would be to think of the judgments as identical to the first-person reports themselves.  In this sense, subjects surely have unchallengeable authority:  they get to issue their own reports.  But this is a trivial kind of unchallengeable authority – Eric himself considers it as a possible interpretation of (2) in his article, calls it “paltry”, and argues that Dennett must mean more than that.  The other option I can think of is that the judgments are identical to the subject’s beliefs about her experience.  Under this reading, (2) reduces to (1) plus unchallengeable authority.

But this assimilation of (2) to (1) plus authority creates more problems than it solves.

First problem:  what reason do we have to accept that every time a subject issues a first-person report about her experience, she first has to generate a belief about the experience?  In many cases (e.g., button pressings during signal detection experiments; notice that Dennett explicitly assimilates such button pressings to first-person reports), it seems more plausible that the report is not mediated by any belief.  This problem can be somewhat alleviated by loosening the notion of belief to something like any state that causes first-person reports to say what they say.  This trivializes the unchallengeable authority claim again – now (2) boils down to the claim that first-person reports are caused by some state inside subjects that causes them to say what they say.  True, but uninteresting.  Another way to alleviate this problem is to appeal to Dennett’s intentional stance:  a belief is simply a theoretical fiction posited by the interpreter to measure the cognitive state of the subject.  While Dennett explicitly advocates this treatment of (1) (Dennett 2003, 20), this interpretation trivializes again the claim of unchallengeable authority – now (2) boils down to the claim that interpreters can attribute subjects beliefs with the same propositional contents as their first-person reports (but this is not to say that the subjects’ minds contain any explicit representations of such contents).  In addition, I don’t think Dennett’s intentional stance is an adequate reconstruction of the way either ordinary people or scientists use the term “belief” – I think people are more realistic about beliefs than Dennett’s intentional stance allows.

Given that I find the above deflationary interpretations of “beliefs”/”judgments” unsatisfactory, I will assume that “beliefs”/”judgments” are explicit propositional representations in the minds of subjects – states separate from, and intermediate between, both the experiences subjects are reporting about and the verbalization process.  Also, let’s assume we have compelling evidence (which as far as I can tell, we lack) to posit that subjects always have beliefs as intermediaries between experiences and the verbalization process. 

Second problem:  now, I really don’t see any reason to grant unchallengeable authority to subjects in expressing their beliefs.  Why couldn’t subjects make mistakes in verbalizing their beliefs?


  1. Hi Gualtiero,

    Re: “First, where does the “almost” before “unchallengeable” come from, and is it anywhere to be found in Dennett’s writings?”

    One place to look is his paper “A case for rorts”. There’s a discussion there of a robot so complicated via successive generations of kluges that, while it’s in-theory corrigible, for all practical purposes is best regarded as incorrigible. Also, I’d say the “almost” comes from D’s Quineanism. He’s thus no fan of analytic certainties and necessary truths.

    Re: “what reason do we have to accept that every time a subject issues a first-person report about her experience, she first has to generate a belief about the experience?”

    On this topic Dennett borrows pretty heavily from Rosenthal’s analysis of expressing and reporting. See, for example, the sections in Consciousness Explained where D sets up his zombies and zimboes arguments. The gist of the relevant Rosenthal is that something just wouldn’t be a report unless it was the expression of a belief.

  2. Thanks, Gualtiero, for this interesting discussion of my critique and interpretation of Dennett! One thing I want to emphasize from the start — and perhaps did not do so well enough in my 2007 blog post — is that this interpretation of Dennett is filtered through me, and although he assented by email to my write-up of it, which I summarized on my blog, I can’t be sure how thoughtfully he did so. He might retract on further thought, or at least modify the details.

    I’m not entirely sure what “first person reports” are in your sense. I’ll grant that there probably are some that are not introspective judgments about conscious experience. I mean my comments only to apply to introspective judgments about conscious experience.

    What kinds of “seems” claims are authoritative, on my view? Let’s start with a non-introspective example. If I say that “the Democrats will lose” then what I say seriously risks falsehood. If the Democrats do not lose, what I said was false. If, on the other hand, I say “It seems to me that the Democrats will lose” (which is not normally I think an introspective statement for reasons explained by Evans), that might still be true even if the Democrats do not lose. Under what conditions is it false? Well, on one way of interpreting it, as long as it’s sincerely and “normally” produced (I realize these qualifiers open up some fishy wiggle room) – for example, as long as it’s not produced by epileptic seizure or sleep-talking or at gunpoint or on the theater stage or without understanding the terms – then it expresses a judgment. Now there will be some weird exceptions, but it seems to me that occurrent judgment and verbal expression are pretty tightly linked. (The connection between these two and dispositional belief is a much looser matter, though.)

    Now let’s go introspective: If I sincerely say “It seems that 30 degrees are visual arc are all clear simultaneously”, there’s an authority there: In all likelihood that *is* my judgment. It would be an unusual set of circumstances that justify your responding “No, it doesn’t seem to you that there is!” when the “seems” here is interpreted in same way as in my claim about Democrats. However, there is also a phenomenological reading of “seems”, not as a kind of epistemic qualifier or signal of hesitation, on which such a statement is more or less equivalent to “My visual experience is that 30 degrees are all clear simultaneously”. On that reading, the statement could quite easily be false.

    “Seems” is confusing in this way, so I recommend avoiding it in discussions of phenomenology. But seeing this ambiguity in “seems” and the sense in which “seems” statements are almost unchallengable when read one way, and quite easily challengeable (often factually wrong) when read another, can I think help us sort through some of the tensions in Dennett – though I don’t think we can *entirely* remove the tensions in Dennett by this maneuver.

  3. gualtiero

    Pete, Thanks for these helpful comments. Now I wonder, how do your two points fit together? The first points denies analytic truths, while the second seems to assert an analytic connection between reporting and believing. Can anyone help me understand this?

  4. gualtiero

    Eric, thanks for your reply. So, are you (and Dennett) simply suggesting that there is an exact parallel between first-person reports about consciousness and any other report, such that subjects are “authoritative” about how things seem to them? In this case, isn’t it misleading to bring this up in the context of heterophenomenology, and suggest that this authority has something to do with consciousness?

    Many philosophers have argued that people have a special authority about the contents of their conscious mind. Now if I understand correctly, Dennett is saying that while people do NOT have a special authority about the contents of their conscious mind, they have autority about how things (either in their concious mind or anywhere else) seem to them to be. But why isn’t this trivial and irrelevant to consciousness studies?

    Here is another way to put the point. Suppose you issue the report, “P”. Presumably your report is corrigible and carries no special authority, but I can always reformulate it as the incorrigible/authoritative report, “It seems to me that P”. Obviously I can do this for any report, regardless of subject matter. Is this all that Dennett is saying? And what does this have to do with studying consciousness?

  5. When Quineans who hate analyticity offer things that look like analyses, what they are probably really doing instead are offering non-analytic Quinean explications. For more on explication as a successor-activity to analysis, see the stuff toward Word&Object on the Ordered Pair as Philosophical Paradigm

  6. Gualtiero —

    I agree with you that the authority of epistemic uses of “seems” has nothing to do with consciousness. Consequently, it’s somewhat misleading of Dennett to speak as though it does. But in interpreting Dennett I think we’re faced with a choice between “somewhat misleading” and “flatly self-contractory”; obviously the first is the more charitable.

    I think we can perhaps even get more charitable by building a bridge partway to consciousness: Epistemic uses of seems reveal, authoritatively, your “notional world” (as least occurrently) and in a way I don’t fully understand or, to the extent I do understand, agree with, that will be tied up somehow with selfhood which is tied up somehow with consciousness. (An orthodox Dennettian or Dennett himself could probably draw this out better.)

    Anyhow, that’s my take. Let me emphasize again that Dennett might find all this to be hogwash if we pushed him further on it!

Comments are closed.