Symposium on Joseph Gottlieb’s “Verbal Disputes in the Theory of Consciousness”

It’s my pleasure to introduce our next Ergo symposium, featuring Joseph Gottlieb’s “Verbal Disputes in the Theory of Consciousness“, including commentaries by Jonathan Farrell, Assaf Weksler, and Josh Weisberg. I’d like begin by thanking each of the participants for their great work. One of the basic rifts within consciousness research is between First-Order and Higher-Order theories of consciousness. This debate centers on the following disputed claim: TRANSITIVITY: A mental state is conscious only if one is in some way aware of that mental state. To illustrate, according to transitivity, if Michael’s perception of the tree in front of him is conscious rather than unconscious, then Michael is aware of his perception of the tree. Higher-Order theorists affirm this conditional, while First-Order theorists reject it. Although trenchant, Gottlieb argues that the debate over transitivity is ultimately verbal. The culprit, Gottlieb suggests, is the Nagelian Conception of consciousness: THE NAGELIAN CONCEPTION: For any mental state M of a subject S, M is conscious iff there is something it is like for S to be in M. As Gottlieb notes, The Nagelian Conception is taken as common ground in the First-Order/Higher-Order debate. The question is whether consciousness in the Nagelian sense requires higher order awareness. That is, the dispute is over the truth-value of: NAGELIAN TRANSITIVITY: A mental state is like something for its subject only if its subject is in some way aware of that mental state. Crucially, however, the central phrase of The Nagelian Conception (‘is like something for its subject’) is semantically ambiguous, and Gottlieb suggests that this ambiguity is at the heart of the controversy over transitivity. Specifically, Gottlieb suggests that First-Order and Higher-Order theorists disagree about the truth-value of transitivity because they tacitly disagree about the meaning of the expression ‘like something for its subject’. First-Order theorists interpret ‘There is something it is like for Michael to Φ x (e.g., to perceive a tree)’ to mean there is some way that Michael feels as a result of Michael’s Φ’ing x (‘The Affective View’). By contrast, Higher-Order theorists read the same expression as meaning there is some way such that it seems to Michael that Michael’s Φ’ing x is that way (‘The Operator View’). If so, we should regard the disputants in the First-Order/Higher-Order debate as actually speaking two languages: First-Order English—generated by adopting the Affective View of ‘what it is like’-sentences—and Higher-Order English—generated by adopting the Operator View of ‘what it is like’-sentences. Why should we regard the disputants in the First-Order/Higher-Order debate as speaking different languages? Gottlieb’s answer is that it provides the best rationalizing explanation for each side’s contribution to the debate. For example, Higher-Order theorists typically either explicitly endorse the Operator View or offer arguments that make sense only if we presuppose it, including (crucially) arguments from the truth of The Nagelian Conception to the truth of transitivity. And while First-Order theorists are less explicit than Higher-Order theorists about what they mean by ‘what it is like’-sentence, there is abundant evidence that First-Order theorist are not speaking Higher-Order English when using such sentences. Most fundamentally, if they were using Higher-Order English, they would be making an obvious a priori error in denying transitivity (p. 331-2). Basic charity therefore seems to require not interpreting First-Order theorists as assuming the Operator View and so as not speaking Higher-Order English. Further, once we interpret the First-Order theorist’s rejection of transitivity through the lens of the Affective View, the First-Order theorist turns out, even by the Higher-Order theorist’s own lights, to be asserting something indisputably true. Thus, we discover that, in affirming or denying transitivity, Higher-Order and First-Order theorists are giving consistent answers to different questions. Each party asserts something indisputably true in their own language. Having explained how we can charitably interpret the utterances of both First-Order and Higher-Order theorists as being true, Gottlieb concludes by examining another interpretive challenge: to explain how intelligent people have become absorbed in a purely verbal dispute without realizing it (p. 342). In response, Gottlieb hypothesizes that the First-Order/Higher-Order debate may be an instance of ‘metalinguistic negotiation’, wherein disputants negotiate the normatively appropriate way to use a term. On this interpretation, the First-Order/Higher-Order dispute is not a debate over the facts of consciousness itself, but over how we should use the word ‘consciousness’ in light of the significant normative strings that are attached to this word. A merit of this hypothesis, Gottlieb suggests, is that while the debate turns out to be purely verbal, it is nevertheless a debate well worth having, given the weight that a word can have in determining conduct. You can find the target article, commentaries, and Gottlieb’s response below. Target article Joseph Gottlieb – Verbal Disputes in the Theory of Consciousness Commentaries Jonathan Farrell – Commentary On Gottlieb’s ‘Verbal Disputes’ Assaf Weksler – Commentary on Gottlieb’s “Verbal Disputes in the Theory of Consciousness” Josh Weisberg – Commentary on J Gottlieb, “Verbal Disputes in the Theory of Consciousness” Author’s response Joseph Gottlieb – Replies to Commentators


  1. Aaron Henry

    Hi Joe,
    It was interesting to read your reasons for no longer being convinced of the paper’s main thesis. I think you’re probably right: more work is needed to identify an undisputed sentence between the two sides. In particular, I think you’re right that HO theorists need not claim that first-order states unaccompanied by higher order states feel some way to the subject (even if, as Rosenthal’s remarks seem to suggest, it’s open to them to say this). I’m curious about what broader conclusions you draw from this concession. It could still be true, for example, that the debate is purely verbal. After all, as you say in the paper, it would be “grossly uncharitable” to interpret FO theorists as using ‘what-it-is-like’ sentences in the sense of the Operator View, since that would involve an obvious a priori error. And it seems equally clear, for the reasons you give, that HO theorists *are* using those sentences that way, at least in certain contexts. That might be evidence that there is a terminological dispute here, even if we haven’t yet pinned down what FO theorists mean by ‘what-it is-like’ and so haven’t yet identified a pair of undisputed sentences. Or do you now agree with Josh that ‘what it’s like’-talk serves more as a rough way of introducing the subject matter under investigation. In that case, HO theorists provide one way (among others) to make that intuitive expression precise through the Operator View and, by implication, Transitivity. By the same token, though, on that understanding the Operator View and Transitivity cannot figure as premises in a non-question-begging argument for the account. They would be more like components of a hypothesis to be assessed on the basis of its explanatory virtues.

    • Joe Gottlieb

      Hi Aaron –
      Thanks for this. I agree that the debate still *could* be verbal. And I agree with you that there is something very curious with the way many HO theorist use WIL-talk in support of their theory. My problem here is that I am not sure what the new set of undisputed sentences would be, and so absent that, I’m inclined to operate on the assumption that the debate is in fact non-verbal. I do indeed often wonder—sort of how Josh suggests— whether the HO theorist could simply dispense with using WIL-talk in defense of her theory. That is, we could try to conduct the debate between FO and HO theorists without recourse to WIL-talk altogether. Much of the debate has already moved to empirical grounds, although there are some issues here too, as there is not a clean break between FO and HO theorists regarding which region of the brain subserves consciousness. (For example, not all HO theorists focus on the PFC.)

  2. Hi Joseph,
    I can’t find the time to think thoroughly about your interesting response, so here is a quick (perhaps confused) comment instead.

    You are now saying that, on the HO theory, U1 is true (or at least, there is no reason to deny U1). Something confuses me here. I’d appreciate clarification. It seems to me that U1 is committed to linking *feeling* with *high-order awareness*. It says something in the vicinity of: “if a state is responsible for a feeling then this state is the target of HO awareness”. Because, as you point out, obviously we do have unconscious feelings (e.g., anger), and Rosenthal grants this (even with respect to pain), it seems that Rosenthal will reject U1 (that is, he will deny the link between feeling and HO awareness).
    I’m not sure whether my point is clear. Here is a different way to look at it. You write that “for the HO theorist, when a first-order state has (what Kriegel calls) qualia or (what Rosenthal calls) mental qualities, a subject need not feel anyway in virtue of being in it.” This might be right, but this claim (call it Q) seems to me (on its face) to be orthogonal to U1 . Claim P says that (roughly) having qualia does not imply feeling. Claim U1 says that (roughly) feeling implies HO awareness. These are independent claims. So I’m not sure why you focus on P’s truth (for HO theorists), and then conclude that U1 is true (for HO theorists), or at least that there is no reason for them to reject U1. (again, there *is* a reason for them to reject U1 simply because of the existence of unconscious feelings, such as anger).
    I’m pretty sure I missed something:-)

    • Joe Gottlieb

      Hi Assaf –
      Thanks for these thoughts. The relevance of your Claim Q is best explained by going back to the analogy with Tye. We want to make sense of what Rosenthal might mean by an ‘unconscious pain’ or a ‘pain that we are not aware of.’ On one (I think common) usage of the term ‘pain’, pains by definition are feelings, and so by definition are states such that, when a subject S is in them, S feels some way. If *that* was our only construal, and so the construal employed by Rosenthal, then expressions like ‘unconscious pain’ or a ‘pain that we are not aware of’ would just amount to something like ‘a mental state in virtue of which we feel some way, yet are unaware of.’ And then Rosenthal would indeed agree, as I initially thought he would, that U1 is false. This sort of interpretation might be buttressed by Rosenthal’s contention that not only do such states have ‘mental qualities’ (or Kriegel’s analogous contention that they have ‘qualia’), but that the very same qualities that show up when we *are* aware of such states (i.e. when those states *are* conscious in Rosenthal’s sense). These qualities would perhaps be the ‘feely parts’: in virtue of the state having them, S feels some way. Yet Tye’s FO theory shows that this construal is not the only way to go. On Tye’s theory, a FO mental state M represents the very same property-complexes whether M or not is conscious. And since Tye identifies phenomenal character with those property-complexes, there is a sense in which an unconscious M ‘has’ phenomenal character, i.e. by representing it. But if Tye can say that S does not feel anyway in virtue of being in M until M is poised (and so is conscious), then why can’t Rosenthal say that S does not feel anyway in virtue of being in M until S is aware of M (and so is conscious)? Yet if Rosenthal *can* say this—and I don’t see why he can’t—then he can accept U1 as true. In other words, Tye’s theory gives us a model of how to think about what Rosenthal might mean by ‘pain that we are not aware of’ that does not require that he treat U1 as false. His mental qualities (like Tye’s phenomenal character) are ‘feely qualities’, but all this means is that they are the qualities that *would* characterize *how* S feels, *if* S felt any way at all. But S would only feel some way if her M was HOT-targeted.

      Hope this helps a bit.

      • Assaf Weksler

        Hi Joseph,
        I still feel that there are two separate issues here that are somehow conflated. One is whether Rosenthal grants in (e.g.) the extinction case that the subject feels something in virtue of being the state. I agree with you that following Tye’s case, Rosenthal could say that in the extinction case the subject does not feel anything in virtue of being in the state. So Rosenthal will hold that “there is something it it is like for the subject to be in the extinguished state” is false in FO-English. So this part of the dispute is indeed substantive and not verbal.
        HOWEVER, there appears to be *separate* issue, namely the issue of U1. As you write in the paper, U1 is ‘A mental state will be such that its subject feels some way in virtue of being in it only if its subject is in some way aware of it.’ It seems that despite what I said in the previous paragraph, Rosenthal will still hold that U1 is false. The reason is that U1 is a *conditional*. It says basically that feeling *implies* high-order awareness (roughly, if a state involves feeling, then it also involves high-order awareness). And we ALL agree that feeling does not require higher-order awareness. Your own example of anger shows this. All parties grants that we can have anger (a feeling) of which we are not aware (and surely a HOT proponent grants this). So it still seems to me that Rosenthal will hold that U1 is false.
        To emphasize, U1 is (or appears to be) silent on the issue of whether “feely qualities” must involve a feeling. (U1 literally says nothing about feely qualities, phenomenal character, and their relation to feeling. U1 only says that *if* there is feeling there is high-order awareness). U1 is therefore silent on the Tye issue (of phenomenal character without feeling). This is why I wrote in the previous comment that i don’t see why Tye’s issue is relevant to U1. I hope the point is clearer now.
        So overall, it seems to me that (a) the question whether the extingushed state involves a feeling is a substantive, non-verbal question, as the Tye case show. This means that Rosenthal will think that the claim “there is something it is like to be in the extinguished state” is false is FO-English. And so this issue is substantive, non-verbal. But (b) Rosenthal would still think that U1 is false. Since U1 is equivalent to TRANSITIVITY in FO-English, this means that Rosenthal would grant that TRANSITIVITY is false is FO-English. And so the debate *about transitivity* is verbal.
        What do you think?

        • Joe Gottlieb

          Hi Assaf –

          Thanks for this. You say “U1 is a *conditional*. It says basically that feeling *implies* high-order awareness (roughly, if a state involves feeling, then it also involves high-order awareness). And we ALL agree that feeling does not require higher-order awareness. Your own example of anger shows this. All parties grants that we can have anger (a feeling) of which we are not aware (and surely a HOT proponent grants this).” I think a lot here matters by what we mean by ‘feeling’ and ‘pain’ (I am setting aside anger). Suppose S breaks his leg. Presumably, S will be in pain. Call S’s pain state M. The Affective View, which forms the basis of U1, says (roughly) that M will be like something for S only if S feels some way in virtue of being in M. It is tempting to reason, as I originally did, that since pain is a feeling, if S is in M (a pain state), then S must feel some way in virtue of being in M. And then, when we consider cases such as being distracted intermittently and thus occasionally unaware of M (as I cite, following Rosenthal, on pg. 336), we are further tempted to say *of course* S can feel some way without being aware of the state (M) in virtue of which S feels some way. But I suppose what I am trying to say is that the HO theorist *need not* endorse the aforementioned conditional, viz. that if S is in M (a pain state), then S must feel some way in virtue of being in M. What they could say is that yes, M is a pain, but S won’t feel anyway in virtue of being in M unless S is aware of M. So when the HO theorist speaks of ‘pains that we are unaware of’ what they mean is something like ‘tissue damage that we are unaware of.’ In fact, Lycan (1996: 16) says something like this when he speaks of ‘unfelt pains’ and how this is contradictory only on some construals.

          Now hearing all this, you might wonder: why then invoke Tye’s theory for sake of comparison? My answer: to undercut the apparent evidence that Rosenthal *does* endorse that conditional, and thus evidence that he *must* say that U1 is false. The apparent evidence I have in mind is Rosenthal’s talk of M having ‘mental qualities’ and ‘thin phenomenality.’ For, prima facie, one might hear this and think that Rosenthal would say that S feels some way in virtue of being in M, even if S is not aware of M. (Why else would he say that M has ‘phenomenality’ absent S being aware of it?) But then if we think of these notions along the lines of Tye’s unconscious representation of property-complexes (‘phenomenal character’), this line becomes much less motivated.

          I am not sure if this helps, but it is the best I have for now.

          Citation: Lycan, W. (1996). Consciousness and Experience. MIT Press.

  3. Josh Weisberg

    Hi Joe,

    Wanted to take up the claim about “what it’s like” being technical again. My thinking is that yes, “what it’s like” talk of some kind has a ordinary usage, but once it gets into Nagel’s (and the community of philosopher’s) hands, it becomes technical. A term may begin its life in one locale and move to another. Further, it may have a more moderate kind of usage that fits with ordinary usage, and a more precise or contentious reading, one that requires technical apparatus to make clear. So it is, I think, with Nagel’s phrase. We get an ordinary use, meaning ‘experience’, but then Nagel contends in the 1974 paper that consciousness in his sense cannot be captured by any “recently developed” causal analysis (“because all are logically compatible with its absence…”). Now, is that part of the ordinary meaning? Lewis, for one, didn’t think so. Lots of others don’t either. So we jump from an ordinary sense to something bearing more philosophical weight. I think this happens all the time in philosophy, for what it’s worth. The fight then is on how to best fill-out and extend what had been an ordinary phrase to suit various purposes of philosophical inquiry. At this point, we have left the domain of ordinary discourse and are doing something technical, even if the terms used are homophonic to the old terms and even if the new uses carry some of the connections carried by the ordinary use. What happens next, I think, is that philosophers propose and justify readings of the term based on overall theoretic utility–something global, rather than semantically local. All this is fine, but it makes arguments about specific interpretations of the phrase look less exciting to me–what really matters is the overall theories behind the interpretations. Or so it seems to me.

    Also, a quick bit on the so-called “emphatic argument”. I don’t think HO folk make this argument. That is, we don’t argue from Nagel to transitivity and so to HO. Rather, we start with transitivity and argue for HO. Then Block says, “nice theory, but it’s not about p-consciousness–you know, what it’s like.” We respond, yes it is, so long as you read “what it’s like” in the right way, as involving something for the subject. (It was Nagel’s emphasis in the original, for what that’s worth.) So it’s kind of a rear-guard action to fend off Ned and co, not a positive claim we start out with. We are transitivity first, and then we pick up the rest as we go. The broader point is, if you think transitivity is where it’s at, you’ll find our reading of Nagel simpatico. If you think transparency is where it’s at, you’ll like some other reading. So it depends on which horse you rode in on–which theory you already buy into. Nagel interpretation is easy; figuring out which theory of consciousness (if any!) to have is not so easy. It may be positively hard, as Chalmers notes! (Notice that this isn’t far from your “verbal” claim.)

    Sorry–a bit rambling. Hope this makes some sense!

  4. Joe Gottlieb

    Hi Josh –

    Thanks for the reply. I think you make a fair point re: WIL-talk being technical. But on the emphatic argument, I suppose I am not sure what the difference is here. From what I can tell, you are saying that HO theorists start like this:

    1. Transitivity
    2. If Transitivity, then some form of HO theory is true.
    3. Thus, some form of HO theory is true.

    And then—and *only* then—is the Nagelian Conception invoked to ensure folks HO theorists are interested in p-consciousness. I suppose though the question is: how is the Nagelian conception being invoked? Presumably it’s in defense of one of these premises, i.e. in defense of (1) or (2). But surely it can’t be in defense of (2), since—as Lycan notes, which I am sure you know—(2) is basically trivial. [Here I am letting Hellie and Chalmers count as a HO theorists; what counts for being a HO theorist, broadly, is not that we have HO representation, but that we have some form of HO awareness. So acquaintance is cool. Premise (2) becomes a tad bit trickier if you assume HO representation.] Thus, it has to be in defense of (1). Yet that is just the emphatic argument.

    Perhaps though I am missing something.

  5. Joe Gottlieb

    Let me just add one thing for sake of clarification: I guess I don’t see the difference between using the Nagelian conception to argue for Transitivity versus using it to argue that *some* HO theory is true, since to implement Transitivity in *some* way *just is* to be a HO theorist. Calling one inference over the other ‘the Emphatic Argument’ doesn’t seem to matter.

    • Josh Weisberg

      Hi Joe–

      Thanks for that. I guess my thinking is that transitivity has its own justification from folk usage (Rosenthal), phenomenology (Kriegel), historical usage in philosophy (see Caston, Gennaro, others on Aristotle, Sartre, Locke), and theoretical utility (me, and maybe Lycan with his stipulative claim). So it has a decent claim to fix the data a theory of consciousness must explain. Then, HO is the best explanation of transitivity (this needn’t be trivial, if one adds naturalized representation, etc.). The Nagelian conception isn’t involved at all. It’s only when those who start with Nagel say we’ve “left out what it’s like” that we respond with the emphatic argument. So, does that mean, if we are correct in this last claim, that we’re using it to support premise 1 in your argument? I think no–we already are convinced of 1. But if you are not already convinced, then I suppose yes. But how convincing that will be depends on what theory you rode in with. So it won’t likely cut much ice. Or so it looks to me. In essence, we are arguing for our analysis from our theory. But I think that’s true with everyone in the debate, or so I’ve tried to argue. Still, it looks like Jonathan may not agree (see below)! I’ll have to get to his comment now. But I hope I’ve clarified a bit.

  6. Hi all,

    I disagree with Josh (and perhaps also Joe) that the ‘what it is like’ talk that we’re interested in here is technical. I think Josh is right that there is a technical or quasi-technical use of the phrase ‘what it is like’ that plausibly began with philosophers. This is the use of the phrases ‘what it is like’ and ‘something it is like’ (often written in scare quotes or hyphenated as‘what-it-is-like’ or ‘what-it-is-like-ness’) as noun phrases which just mean something like ‘experience’ or ‘phenomenal consciousness’.

    But that’s not the kind of ‘what it is like’ talk that Joe discusses in his paper, that Nagel uses in his, or that Stoljar is interested in. The ‘what it is like’ talk that Joe and Stoljar are interested, and which Nagel employs, is the appearance of these phrases in sentences like ‘I wonder what it is like to be a bat’ or ‘There is something it is like for Fred to see red’. In these sentences the phrases ‘what it is like’ and ‘something it is like’ aren’t noun phrases—if we try to read the phrases as noun phrases we end up with something ungrammatical.

    This use—the use in sentences like ‘I wonder what it is like to be a bat’—is non-technical (for the reasons I give in my 2016). And this is the use that is involved in what Joe calls the Nagelian conception of consciousness (NCC), and in the emphatic argument.

    Also, I don’t think that Nagel claims that it *follows from the meaning* of ‘what it is like’ talk that (then) recent approaches to the mental are incomplete. So the fact the he and (e.g.) Lewis disagree about whether Nagel is right doesn’t show that Nagel is using ‘what it is like’ talk in a non-ordinary way.

    • Josh Weisberg

      Hi Jonathan,

      Very well put! Thanks!

      I do wonder if the technical use, rather than the non-technical use, is at play in the debate between FO and HO theorists. Indeed, I titled my paper “Abusing the notion of ‘what-it’s-like-ness’” kind of for that reason—it seemed to me that Block was using a distinct technical notion, one which the HO theorist isn’t going to agree with. That also raises the possibility that the debate Joe is on about is not using the non-technical notion, despite his focus on Stoljar’s analyses.

      I also wonder (though you’ve gone far deeper into this stuff than I!) if Nagel starts the paper in ordinary talk, but shifts to a technical meaning at the crucial moment where he gives his “definition” of consciousness. Nagel says, “But fundamentally an organism has conscious mental states if and only if there is something that it is like to be that organism–something it is like for the organism.” I take it that your thinking that this does not, as it stands, rule out a causal analysis—some further argument is needed to do that. But for many philosophers, this is just what WIL talk picks out—something beyond the causal-functional. That’s now part of the meaning of the term, perhaps, for many people in current debates about consciousness. Let’s say that this is true—that some folk use Nagel’s definition to mean the more substantial claim. Does that mean this extended usage is technical? Or, in any event, it might be a contentious matter to be settled by something beyond the analysis of the non-technical usage. I think that this might be what lots of current folk mean by this sort of talk, even if they don’t employ the noun-phrase version.

      Another point. I’m not at all sure there’s a sharp line between technical and nontechnical usage. Non-technical terms get taken up by technical literatures and evolve new meanings. And technical terms move into common usage, like “denial” or “gravity” have. There may be no fact of the matter about which term is in which category. So, again, I wonder about the efficacy of the analyses in question if those rely on a sharp line, and on being on the non-technical side of things. And this says nothing of Putnam’s division of linguistic labor. (This maybe more to the point of your next post—more on that later!)

      Thanks, and again, very nicely put!

      • Hi Josh,

        Thanks for the reply! I think that the use by philosophers of phrases like ‘what-it-is-like-ness’ is technical in the sense that we don’t see non-philosophers using these phrases. But I think that when someone like Block uses it he’s just using it to mean something along the lines of ‘experience’ or ‘phenomenal consciousness’ or ‘the sort of feature of our mental lives that we point to when we say that there is something it is like to see an orange, or to stub your toe.’ So perhaps it would be better for me to say that I think that phrases like ‘what-it-is-like-ness’ are jargon, or a kind of short hand, rather than that they are theoretical terms. Describing these phrases as ‘technical’ does suggest that they (perhaps implicitly) embody some (perhaps controversial) theoretical commitments. But I don’t think that is true of ‘what-it-is-like-ness’ or uses of ‘what it is like’ or ‘something it is like’ in scare quotes. I think these phrases inherit their meaning from the (ordinary, non-technical) meaning of sentences like ‘What it is like to be a bat?’ or ‘There is something it is like to see a red rose.’

        You might be right that some philosophers take ‘what it is like’ talk to be about something which goes beyond the causal-functional. But adopting this stronger view is going to make things like Nagel’s paper, Jackson’s Knowledge argument, and zombie arguments (if they’re couched in ‘what it is like’ talk) all pretty uninteresting, and perhaps (depending on their precise form) even obviously question-begging. But I take it that these arguments _are_ interesting and are _not_ not obviously question-begging. So I don’t think that this is what Nagel is doing. And I don’t think that anyone _should_ do this as it liable to lead to confusion and talking past one another. (Maybe it’s okay to do this if you’re explicit about it, I suppose.)

        I agree that in some cases the the technical/non-technical border might not be a sharp one. And, as you say, this means that sometimes there won’t be a fact of the matter about whether a term is technical. But sometimes there will be. And even when there isn’t, I think it can be useful to try to clarify what the term might mean (regardless of its technical/non-technical status) for the reasons I gave in an earlier comment.

        • Josh Weisberg

          Hi Jonathan,

          Thanks for that, and sorry for the slow-motion response!

          On the question-begging bit: I worry that this is perhaps what’s going on (though maybe this isn’t as uninteresting as you think). Dennett, Lewis, later Jackson are all type-A materialists. They think zombies are inconceivable and that Mary does not learn any new propositional knowledge. One complaint about their views is that they “leave out what it’s like”. They might reply (and I think they do), that no, their views explain experience and that’s all that “what it’s like” talk picks out, so they haven’t left that out. Their opponents might say back, “no, that’s not what experience amounts to, so you’re not talking about what it’s like.” So, if you have some strong view of what experience is, you might think someone has left out what it’s like in (what such a person takes to be) the ordinary sense. Is this question begging? Maybe, but all that does is push the important question to “what is experience?” That is the question begged, and so we are left fighting about that. Once we have that answer, we’ll have some idea about what people mean by “what it’s like” (aka “experience”) in this debate. I don’t see how analyzing “what it’s like” talk in the absence of knowing the answer to the experience question helps. This is in a sense what I mean by WIL talk being “theory-laden.” See Richard Rorty’s chapter in Dennett and his Critics (ed. Dahlbom) for someone who claims that, yes, this debate is all question begging at the get-go and that makes it intractable (though not, perhaps, uninteresting!).

          In the same spirit, do you think it follows that if Lewis (or Dennett or postlapsarian Jackson) is right about consciousness that this means there’s really nothing it’s for one like to see red, in the ordinary sense of the term? I do think that that’s what some of their opponents think, and so I worry that those opponents take WIL talk to mean the stronger thing. And if it doesn’t really mean the stronger thing in its ordinary sense, then those opponents are using it in a technical sense.

          • Hi Josh,

            Thanks for your response. Here’s a probably-even-slower reply!

            I don’t take the ‘but you’re leaving out what-it-is-like-ness’ response to be begging the question. I think that, for the most part, both type-A materialists and proponents of the ‘leaving out’ response take themselves to—or at least start with the assumption that they—agree on the answer to the question ‘What feature of mentality are we talking about when we engage in ‘what it is like’ talk?’ (This needn’t require thinking that we can provide an explicit answer that is clear as we’d like.) I think that Chalmers’ taxonomy needs to assume that the ‘leaving something out’ response doesn’t beg the question otherwise his taxonomy will not be a taxonomy of accounts of _the same thing_. If the accounts are not trying to explain the same thing, then the taxonomy falls apart.

            I think that where type-A materialists and their opponents disagree is regarding the answer to a second question: ‘Does type-A materialism leave out that feature of mentality that we agree we’re talking about when we engage in ‘what it is like’ talk?’ Type-A materialists think it does not (either because that feature doesn’t exist so there’s nothing to leave out, or because that feature exists but can be accounted for by a type-A theory). Their opponents think it does leave something out.

            Whether the truth of type-A materialism entails that there is nothing it is like (in the ordinary sense) to see red seems to me to depend on exactly which type-A theory is correct. If the right theory is eliminativist (about what-it-is-like-ness in the ordinary sense) then there’s nothing it is like (in the ordinary sense) to see red. But if the right type-A theory is not an eliminativist one, there would still be something it is like (in the ordinary sense) to see red. Does that sound right?

            But if all of the above is wrong, and the ‘leaves something out’ objection _does_ beg the question, then I think the situation is as you describe: the argument shifts to answering ‘What is experience?’, and only once we’ve made progress on that question will we be able to make progress on explaining how ‘what it is like’ talk works. But then, insofar we’re primarily interested in certain features of mentality, rather than the language we use to talk about those features, the question of how ‘what it is like’ talk works doesn’t look particularly interesting. Would it be fair to say that this is the situation that you think we are in, and so this is why (perhaps only in part) you think that looking at ‘what it is like’ isn’t going to help us?

          • Josh Weisberg

            Great stuff, Jonathan—thanks.

            On Chalmers: I do think it’s a near thing with Chalmers. In his 2003 “Consciousness and its Place in Nature” he says that the “obvious problem with type-A is that it appears to deny the manifest” (109). What is manifest is that there’s more to consciousness than function. Later he says that it the debate between type-A and its opponents comes down to “the intuition that consciousness (in a non-functionally defined sense) exists…” (110). I think he intends WIL talk to pick our this non-functionally defined sense. If so, I think it does beg the question, if it’s coherent (reasonable?) for type-A to deny this sense. But you may be right that this is all a next step off of a more neutral reading by Chalmers. Still, for me the point may be that it’s not at all clear how to draw the lines here. (Also, full disclosure, Pete Mandik and I have a paper called “Type-Q Materialism” where, among other things, we challenge the stability of Chalmers’s taxonomy. We tactfully published it in an obscure festschrift so no one would know about it…)

            On eliminativism, I worry also about the difficulty of drawing the line between eliminativism and reductionism here. We seem to need to know what WIL really means before we can tell which theory does what. And that may not be do-able without putting more into WIL than is uncontroversially there in the ordinary usage. Thus, we’re back to the theory-laden issue—if I think WIL talk involves something “rich” I may be a type-A eliminativist; if I think it’s more “thin” I may be a reductionist. But is there really a fact of the matter about what the actual content of ordinary WIL talk is such that we might adjudicate this dispute?

            Your last bit nicely captures my worry, especially when the difficulty of drawing the question-begging/non-question begging line and the eliminativism/reduction line are taken into account.

  7. On a different topic: in his commentary, Josh expresses “a general skepticism about the employment of detailed linguistic analysis to elucidate these sorts of philosophical debates.” This is because “whatever we ultimately decide about the right analysis of language is beholden to more basic views we have about what’s true, what’s the best theory of the world, what science says, and so on.” The second claim strikes me as plausible, and I doubt that detailed linguistic analysis will end the disagreement between HO and FO theorists of consciousness. But the truth of this second claim doesn’t seem to me to offer much support for the general skepticism.

    Linguistic analysis doesn’t need to end the disagreement between FO and HO theorists in order to shed light on this disagreement. Such analysis might show that certain arguments which explicitly appeal to the meaning of ‘what it is like’ talk fail. It might show that participants in the disagreement are talking past each other. (I argue for both of these claims in my ‘Higher-order theories of consciousness and what-it-is-like-ness’ 2017, Joe argues for the latter in the paper this symposium is on). Or it might reveal that participants are not talking past each other. (One response to Joe’s paper could have been a chorus of ‘I’m assuming the Affective [or the Operator] View!’ from all sides.) Or it might encourage those on either side who appeal to the meaning of ‘what it is like’ talk in their arguments to be more explicit about exactly what they’re claiming, and why it is reasonable to claim so. Or it might make explicit interpretations of ‘what it is like’ that hadn’t previously been noted (maybe Assaf does this in his commentary with his IA-English) and so making unconscious slippage between this and other readings now less likely. Perhaps none of these things end the disagreement, but each counts as progress.

    • Josh Weisberg

      More slow-motion conversation:

      I’ve wondered why the HO theorist can’t simply accept the affect account. Indeed, as a HO theorist, I thought maybe that’s what I mean by WIL (or I don’t see why not). It’s just that we HO folk think that affect in the relevant sense either involves awareness, or we think the affect account must be amended to so it includes the claim that we’re aware of being so affected (or some such). Then we’d all be speaking the same language but we’d disagree about whether we need this sort of awareness or not. If we do need the awareness, the HO analysis is the one we ought to use; if not, then we can stick with the more minimal FO sense (without the built-in connection to awareness). Then the debate would be substantial. What’s more, the analysis bit is exposed as not doing much work beyond exposing where the deeper debate lies. And I agree that that’s not nothing!

      On the more general issue: despite your well-put plea, I remain skeptical of the power of detailed analysis to elucidate much. Maybe philosophers are talking past each other because they are using language differently, or maybe they’re talking past each other because they disagree about the facts. Or maybe these two are inextricably intertwined and so pulling out the one bit without making clear the other doesn’t show the whole story. Not that the language thing isn’t sometimes helpful, but it can’t bear much weight. The best we can do is employ “rough and ready” analyses, in Jackson’s term. On that, as I said in my initial response to Joe, WIL picks out experience and whatever is (seemingly) problematic about it. Any further step to fill out the meaning will be theory-laden.

      • Joe Gottlieb

        Josh –

        In a way, your suggestion about HO theorists employing the affective view connects to my reasons for abandoning the verbal disputes thesis. Roughly, on the affective view, a mental state M is conscious, so is like something for its subject S, iff S feels some way in virtue of being in M. Now I had thought that HO theorists said things which implied that S can feel some way in virtue of being M *without* being aware of M, and so *without* M being conscious. But I now see why that’s wrong; when HO theorists talk about being in a pain state they are unaware of, or states having ‘qualia’ or ‘thin phenomenality’ in the absence of our being aware of it, this need not mean that S feel some way in virtue of being in such states. Of course, this does not show that HO theorists *do* employ the affective view, but, as you suggest, maybe it does imply that they could.

      • Just a quick note about Josh’s suggested amendment of the Affective View. The amendment would produce a view that proponents of the Affective View (such as Stoljar) may well reject. Stoljar isn’t just interested in explaining the meaning of sentences of the form ‘There is something it is like for a subject, S, to be in mental state, M’. He wants to explain any sentence of the form ‘There is something it is like for S to F’. So the Affective View says that ‘There’s something it is like to have strychnine in your system,’ means something along the lines of ‘There is some way you feel as a result of having strychnine in your system’. But, I take it, neither of the ‘strychnine’ sentences entails that you are aware of having strychnine in your system. That’s not to say that looking into Josh’s amended view mightn’t be useful, but only that it involves a significant amendment to the view Stoljar discusses and endorses.

        On a different topic. Josh suggests (if I’ve understood him correctly) that perhaps the most that all sides can agree upon is that “WIL picks out experience and whatever is (seemingly) problematic about it.” I think there’s two ways to read this, and I wondered which reading Josh intended. It might be that we all agree that ‘what it is like’ talk picks out experience and a problematic feature of experience, and we all agree on what this problematic feature is. Or it might be that we agree that ‘what it is like’ talk picks out experience and a problematic feature of experience, but we disagree about which feature of experience ‘what it is like’ talk picks out as being problematic. Sometimes I wonder whether the second option best captures what’s going on with FO and HO understandings of ‘what it is like’ talk.

        • Josh Weisberg

          Right, Joe. Though there’s this other point I’ve been circling around that it’s not fully clear how we’d decide which is what’s being employed. But, for what it’s worth, I do think we HO folk often go with an “unconscious feel” way of putting things. But we don’t have to, as far as I can tell.


          I think one worry I have about Stoljar’s view is that it doesn’t really capture the Nagelian sense of WIL, at least from the HO perspective. We think there can be unconscious affect, and so we think there needs to be more to capture the sense Nagel identified. Of course, I think this intuition we have is theory laden, so I’m not claiming it as an independent bit of data. But I do worry that Stoljar’s desire to capture everything under one umbrella might have pushed him away from a characterization the HO folk can embrace. I take it, of course, he wants to avoid an emphatic reading as a short route to HO as well.

          About WIL talk and my “neutral-ish” characterization—yes, it’s the second way I intended. And then how one fills out the last bit will shape where one goes theoretically—if it’s awareness of, or transparency, or non-functional, etc. That bit I’ve been calling theory-laden, but the other bit might be ostensive and neutral enough to get us started fighting! Also, maybe the second bit brings in technical stuff, at least in some cases.

          • Hi Josh,

            You might be right that Stoljar doesn’t capture the Nagelian sense of ‘what it is like’ talk from the HO perspective. But that’s only a problem for him if there is such a thing to be captured. If the HO perspective on ‘what it is like’ talk is that the meaning of this talk fairly straightforwardly leads to a HO theory (which things like the emphatic argument suggest), and the Nagelian sense of this talk is the sense that Nagel takes it to have, then there is no such thing (or so I argue in my 2017): the HO understanding is not a Nagelian understanding. (But I don’t take this to beg the question against HO theories. That a HO theory doesn’t follow from the meaning of Nagelian ‘what it is like’ talk doesn’t entail that the correct theory of what Nagelian ‘what it is like’ is about is not a HO theory.)

            On the “neutral-ish” characterisation. Good! If there’s rough agreement that everyone is talking about experience, but a disagreement about which features of experience are problematic, then it may be possible to clarify what the two sides think the problematic feature is in a way which doesn’t require interpreting ‘what it is like’ talk. Here’s a boldly-stated, rash, and ill-thought attempt at doing so! I’d be interested to know if anyone thinks this is anywhere near the mark.

            The question HO theorists take themselves to be trying to answer is something like this: how can there be mental states which are such that, when you’re in them, they (i.e., the mental state, or perhaps your being in the mental state) seem/feel some way to you?

            The question FO theorists take themselves to be trying to answer is something like this: how can there be mental states which are such that, when you’re in them, things (often mind-external things like tomatoes and toes) seem/feel some way to you?

            These questions aren’t wholly unconnected, but I don’t think it would be surprising if answers to one don’t obviously shed much light on the other.

          • Josh Weisberg


            I guess I think that Nagel’s phrase is supposed to distinguish conscious from non-conscious states. Since it seems plausible that there can be nonconscious states that affect a subject, we need something more than affect to separate the conscious from the non-conscious states. So I don’t think Stoljar gets the Nagelian sense even if we don’t add the stuff about “awareness of.” Now, when we HO folk add that transitivity language, we do indeed beg the question. But everyone needs to put something in there to beef-up affect. So everyone is going to beg the question. And that’s ok! All that means is that we are arguing from our theoretical perspective and they are arguing from theirs. This stuff is theory-laden. But so is everything else. Does that mean we can’t argue about it, that things are “incommensurable’? No. We just have to try to make explicit all the theoretical baggage we’re bringing to the party. Then we tally up the theoretical virtues and see where we stand.

            On the other bit:

            I like your rash and ill-thought attempt! If you can’t engage in rash and ill-thought attempts on a blog discussion, when can you? ;>)

            I guess (in line with some of my other comments) that HO folk take transitivity to be wrapped up in what must be explained. FO folks take transparency to be central. But I do think there’s another element all reductive theorists of whatever stripe must deal with: why does it seem that consciousness can’t be functionally explained/isn’t physical/is special/etc. That is, I think there needs to be a concern about answering the meta-problem.

          • Hi Josh,

            Thanks for that, and apologies the extreme tardiness! Just a couple of comments on the last thing you said:

            I wasn’t sure I understood the first part of your comment. Stoljar wants to say that conscious states effect us in a particular way, not merely that they have some effect or other on us. So he’s saying a state is conscious (in the ‘what it is like’ sense) if there is some way the subject feels as a result (sometimes, more demandingly, in virtue) of the subject’s being in that state. So are you saying that it’s plausible that someone can feel some way in virtue of being in a state (without bringing in the “awareness of”) even though the state is an unconscious state?

            I take your point about the meta-problem, although I don’t think FO theorists need to take transparency to be central (or even true), although lots of them do. But if, as you say, HO theorists take transitivity to be involved in what must be explained, then the fact that FO theorists deny this is enough to distinguish between them.

          • Josh Weisberg

            Hi Jonathan,

            Yes, that is indeed what HO folk say–we think subjects can be in pain–a state marked by a distinctive feel (by a distinctive sensory quality) and that feel can affect them (it can make them act certain ways both psychologically and bodily) and nonetheless that very pain state can fail to be conscious. That is, there’s nothing it’s like for the subject to be in that state. And, for what it’s worth, we think our commonsense, ordinary way of looking at things licenses this claim (though it also has acceptance in empirical science as well). (Joe was on about this issue above.)

            I take it Stoljar wants to say either no, there can’t be real feels like that (and HO folk will debate him here) or he’ll say, ok, I meant “consciously affected”. This latter move might mess up the general claim (the economy doesn’t do this, etc.), but also it now looks uninformative and close to circular: WIL amounts to being consciously affected, which comes to “being conscious is being consciously affected”.

            Also, (with your recent paper in mind), this may be a spot where HO theorists cry “monadic”. There’s an assumption, maybe, that feel and consciousness go together–consciousness is a monadic property of states that feel some way to a subject. We reject that: consciousness and feel can come apart. Again, all this is from our theoretical perspective and no doubt some (strangely!) disagree. But the proper analysis of WIL seems to me beholden to these questions.

          • Hi Josh,

            Thanks for that. Just a quick recap of what I think’s been said in this thread: you suggested that Stoljar’s Affective View doesn’t capture the Nagelian sense of ‘what it is like’ talk even if we don’t include the claim that a state’s being conscious requires that the subject is aware of it. And the reason to think this is that nonconscious states can have an effect on subjects in those states. In your last comment you gloss this last claim as:
            (1) A state can have a feel which has an effect on its subject even though the state is not conscious.
            And you note that HO theorists accept (1).

            But I don’t see that (1) shows that the affective view doesn’t capture the Nagelian sense. To show this we need that:
            (2) A subject can feel some way in virtue of being in a state even though the state is not conscious.
            (E.g., a subject can feel pain even though their pain state is not conscious.)

            (1) and (2) are different. (1) is about a state having a feel, not (as (2) is) about a subject’s feeling some way. And (1) is about the subject’s being in the state having an effect on the subject (perhaps, as you suggest, making them act certain ways). But (2) about a very particular kind of effect (which looks different to the one you suggest)—the subject’s feeling some way—occurring in virtue of the subject’s being in the state.

            So I don’t see that accepting (1) shows that (2) is false. And I don’t think that in accepting (2), Stoljar has to deny (1) (although he might deny it for other reasons). Even if Stoljar denies (1) I think he can avoid the charge of circularity because he doesn’t have to say ‘by “the subject is affected by being in the state” I meant “the subject is consciously affected …”’. Instead he says ‘by “the subject is affected by being in the state” I mean “the subject feels some way in virtue of being in the state.”’

            I think this relates to your last point: accepting the affective view means that consciousness and a _subject’s_ feeling some way go together. But that might be compatible with denying that consciousness and a _state’s_ having a feel always go together. I’m still not sure whether it is plausible to deny the first of these links.

          • Josh Weisberg

            Jonathan (as usual, apologies for the slow reply!)–

            That sums it up very well, I think.

            And your claim (2) is right on point as well. Here, unsurprisingly, the HO theorist wants to accept 2. We hold that most of the characteristic effects of pain–say those given by a Lewis style functional analysis–can indeed occur nonconsciously (modulo the usual effect of being aware of oneself as being in pain). When that happens in a subject, they feel a specific way in virtue of being in that state, even though the state is not conscious. An unconscious pain can be caused by pinching shoes, can cause me to limp, can make it difficult for me to concentrate on a math problem, etc. So what is happening to me is certainly in virtue of my being in that state and, further, in virtue of the state being a pain (having the sensory quality distinctive of that sort of state). So this is plausibly unconscious pain and it appears to be affecting the subject. We are happy to call those affectations “feels”, though this is theory-laden. But if these are not feels, then we hold that we only feel pains when we’re aware of them. Then we’re happy with Stoljar’s analysis. But I take it that’s not how he’d read it. So the fight shifts to “feel.”

            (Also, I’m not sure the British economy feels anything in this sense, so we need some way to differentiate “feels” here to get Nagel to fit. )

            In any event, I think you got the dialectic just right!

  8. One more comment! What follows is a response to what Joe says in his reply to my commentary. Joe suggests that the points I make in my commentary about the precise statements of Nagelian Transitivity and the Operator View don’t matter so much because adopting HO-English still allows us to produce a sentence that is “undisputedly true”:
    U2* S’s being in M seems some way to S only if S is aware of herself as being in M
    I dispute the truth of U2*! Before explaining why I dispute its truth, it’s worth repeating how we get to U2*.

    If we combine _self_-relational (rather than state-relational) transitivity:
    SRT M is conscious only if S is aware of herself as being in M
    and the Nagelian conception of consciousness:
    TNC M is conscious iff there is something it is like for S to be in M
    we get:
    NT** there is something it is like for S to be in M only if S is aware of herself as being in M
    If we then apply the correct understanding of the operator view:
    O1 There is something it is like for S to be in M iff S’s being in M seems some way to S
    We get
    U2* S’s being in M seems some way to S only if S is aware of herself as being in M

    Why doubt U2*? I think that the following may be undisputedly true:
    U3 S’s being in M seems some way to S only if S is aware of S’s being in M
    U3 looks true to me because it is an instance of a more general claim, namely that
    X seems some way to S only if S is aware of X

    But U3 doesn’t entail U2* unless
    S’s being aware of S’s being in M
    S’s being aware of _herself as_ being in M.
    And this entailment doesn’t hold. S can be we aware of _S’s_ being in M without S being aware of _herself’s_ being in M. (If unbeknownst to me, I am the messy shopper, then I can be aware of _the messy shopper’s_ making a mess without my being aware of _myself’s_ making a mess). And S can be aware of herself’s being in M without S’s being aware of herself _as_ being in M (I can be aware of myself’s making a mess without being aware of myself _as_ making a mess).

    This won’t be a problem if we can get to U3 instead of U2*. We can do this if we change SRT to SRT*:
    SRT* M is conscious only if S is aware of S’s being in M
    But this will only work if SRT* is strong enough to capture the sense of _self_-relational transitivity that HO theorists are after. I’m not sure that it does, but I’d be interested to hear what others think about this.

    • Joe Gottlieb

      Hi Jonathan –

      Thanks for this great counter-reply. So I do wonder whether SRT* might be enough for at least some HO theorists. It requires some form of HO awareness, which is good enough to scare off FO theorists. Perhaps one broader but still relevant issue will be how a particular HO theorist thinks of the connection between p-consciousness and self-consciousness; if one thinks there is a tight link, then one will probably want self-relational transitivity, not SRT*.

      That said, Rosenthal (1997: 741-742) also seems to think that U3 does indeed entail U2*. Notice I am not speaking for myself here, and I do feel the force of your point. But for sake of completeness, here is what Rosenthal says. The subject S is supposed to be thinking about being in a mental-stake *token* (M), not simply thinking about a type of mental state. (I say ‘think’ here because the HO awareness is via a HOT.) However, Rosenthal adds, one cannot think about being in a mental-stake token unless one thinks that some individual is in that state. But since M is *ex hypothesis* a state of S’s, then S will be aware of herself as being in M. [This is the entirety of the argument as far as I can tell.]

      Now I take it that the question becomes: why does Rosenthal think token mental states are different in this way from other properties, like your Perry-inspired ‘making a mess’. I don’t know the answer to that.

      Citation: Rosenthal, D.M. (1997). A theory of consciousness. In Ned Block, Owen J. Flanagan & Guven Guzeldere (eds.), The Nature of Consciousness. MIT Press.

      • Hi Joe,

        Thanks for that. I agree that FO theorists will want to deny SRT*, so perhaps it’s one place where we can draw a line between FO theorists and maybe at least some HO theorists.

        On Rosenthal. Yes, that’s the question I’d ask. And maybe there is a principled reason why being in M is different to making a mess in a way that allows Rosenthal to answer that question. But it’s not obvious to me that that principled reason is.

  9. Joe Gottlieb

    I forgot that Kriegel filled in a bit of the details for Rosenthal’s argument, although he doesn’t endorse it per se. If you are curious, see his 2009 “Subjective Consciousness”, pg. 178.

Comments are closed.

Back to Top
%d bloggers like this: