1. The Seeming Problem

According to HO theories, what is to be explained is: being in a conscious state. If such theories are to provide more than a mere brute identification of conscious mental states with suitably represented mental states, they must show how that identification provides explanation. ^{Footnote 2} Here, it is important, as Josh Weisberg rightly stresses, to capture the correct target for explanation. ^{Footnote 3} HO theories aim to account for the specific notion of consciousness that is captured by the Transitivity Principle, which is, in Weisberg’s terms, an ‘extrinsic’ notion. It is not their aim to capture an ‘intrinsic’ concept of consciousness (one phrased in terms of qualia or something such), although HO theories may, of course, be extended to (potentially) cover such notions as well. I will argue that this seemingly neat distinction in the end requires some qualification—that the extrinsic characterization of the Transitivity Principle is not without traces of intrinsicality (see §3). However, what Weisberg’s distinction rightly does is to put aside issues concerning qualia and the like: HO theories do not directly aim to solve such issues.

Now, there is, of course, one obvious piece of explanatory value to HO theories: their identification of conscious states with states that are being suitably represented explains the platitude that motivates the Transitivity Principle, one formulation of which reads: “if a subject is in no way aware of herself as being in a mental state, that state is not a conscious state.” ^{Footnote 4} Now, this is a platitude linking intransitive and transitive consciousness. But what exactly does the transitive notion involved amount to? HO theorists should be concerned with this question, given that this transitive notion, ‘awareness-of,’ is their main explanans for the explanandum of intransitive consciousness (state consciousness). In particular, HO theorists should take note of platitudes relating to this transitive notion (just like they take note, and make thankful use of, the platitude recorded by the Transitivity Principle). Let me introduce one such platitude, one that links awareness-of with a certain notion of seeming:

As with all platitudes, it is important to be as clear as possible on what we take the Seeming Platitude to be saying, without thereby already engaging in potentially controversial theorizing. ^{Footnote 6} As a first approximation, the thought is that, if things do not seem to a subject to be a certain way, she is unaware of them being that way. (Here, neither the seeming nor the awareness-of are to be understood as involving intransitive state consciousness: we are focusing, to repeat, on the less demanding transitive notion of awareness-of.) In other words, a representational state amounts to awareness-of only if it makes its contents available to the subject, such that things seem to the subject to be the way those contents say they are. This involves, minimally, that the subject is awake, or at least not in a deep coma or immersed in a dreamless sleep (in other words, that the subject is ‘creature conscious,’ as it is often put). This connection between awareness-of and seeming, in fact, plays an important role for the HO theorist, for HO states are supposed to do precisely that: make their contents (i.e., the target mental state) available to the experiencing subject, such that it seems to the subject that she is in the target mental state. That appears to be the whole point of using the Transitivity Principle as a guide for an account of state consciousness.

Now, it appears that accounting for ‘awareness-of’ in terms of mere representation, as HO theorists routinely and confidently do, does not respect the Seeming Platitude. For what that platitude, in effect, states is that awareness-of is more than ‘mere’ representation. And this suffices to formulate a first rendering of the Seeming Problem:

Notice that the requirement of respecting the Seeming Platitude does not mention qualia or something such. It does not take us beyond the aim of HO theories: it merely highlights an aspect of the transitive notion ‘awareness-of’ that plays a pivotal role in the extrinsic characterization of consciousness based on the Transitivity Principle.

Strictly speaking, then, my conclusions in this article will be conditional; they can be resisted by challenging the additional requirement of explaining the Seeming Platitude, of explaining the connection between awareness-of and seeming I have just introduced. Doing so, however, would not exactly add to the plausibility of the HO approach, as I have indicated (see also §4 below). Notice, by the way, that I do not commit myself to a precise statement as to the relation between seeming and awareness-of. Perhaps they are just one and the same phenomenon; perhaps seeming is a feature of all and only states that provide awareness of their contents (or vice versa); perhaps there can even be seemings which do not amount to any awareness-of (although I don’t think that this last option makes sense, pre-theoretically). My argument only requires the claim that the awareness-of on which the HO approach rests implies seeming, and that seeming is somehow tied to ‘creature consciousness’—consciousness in the sense of being awake as opposed to being (dreamlessly) asleep (this will turn out to be of importance in §4 below). ^{Footnote 7}

On reflection, there seem to be three ways for the HO theorist to deal with the Seeming Problem. She can, first, claim that her notion of awareness-of does, in fact, provide materials to account for the Seeming Platitude, thereby simply denying the problem. Let us call this the ‘Ostrich Strategy.’ Secondly, she can accept the problem and attempt to tackle it by HO means—that is, by formulating an HO theory of awareness-of. Let us call this the ‘Repeat Strategy.’ And thirdly, she can accept the problem and attempt to tackle it by different (non-HO) means—that is, by formulating a theory of awareness-of that does not rely on postulating HO states. Let us call this the ‘Alternative Strategy.’

In the remainder of this introductory section, I will now argue that only the Alternative Strategy has any hope of being successful. Let’s start with the Ostrich Strategy: it is not difficult to see that it amounts to an objectionable trivialization of seeming. For consider the question of whether my first-order perceptual representation of a donkey suffices for it to seem to me as if there is a donkey here (i.e., suffices for its contents to be available to me in the way the Seeming Platitude suggests). If the HO theorist answers ‘Yes’ to this question, as the Ostrich Strategy recommends, it transpires that, on her view, there is seeming as soon as I represent anything. This trivializes seeming, and thereby awareness-of: there is no room for representations of whose contents one is unaware, on this understanding, for to represent something already implies that things seem to one as represented, and hence that one is aware thereof. Yet, there are bound to be many representations, for instance, in early visual processing, which in no way provide an awareness of their contents, as things never seem to the subject to be the way such representations say they are. (Recall that taking a representation to provide an awareness of its contents (transitive) is not the same as saying that that representation is a conscious state (intransitive)—the former is used by HO theorists to account for the latter.)

If, on the other hand, the HO theorist answers ‘No’ to our question (which seems more likely), the problem remains. For then, my being in a perceptual state representing the donkey does not suffice to provide me with a seeming of its contents (the donkey). Something more is required. Perhaps the most natural move for the HO theorist is to suggest an HO approach—that is the Repeat Strategy. However, upon reflection, it is quite obvious that this will not work. To see why, consider the following traditional regress worry for HO theorists that is usually quickly brushed aside in presentations of the HO strategy: if my first-order mental states are to be made conscious by second-order states suitably representing them, should not the same also hold for those second-order states, leading to a regress of states of ever higher orders? The standard answer is ‘No,’ because a second-order state doesn’t have to be conscious in order to confer consciousness on its target first-order state (though it might be, through a suitable third-order state). Now, the Repeat Strategy suggests that an HO state is required not just for establishing state consciousness but also for the target state to provide a genuine awareness of its contents, in line with the Seeming Platitude. But notice that this does amount to a viciously circular HO theory: the suggestion is, after all, to account for the awareness-of of the target state in terms of the HO state’s providing awareness of that state. ^{Footnote 8} That HO state would have to do so by providing awareness-of, and hence a seeming, of its own contents, namely, that I am in the target mental state. Here, the notion of providing awareness-of appears again, and hence the considerations so far repeat themselves: we need a further HO state representing the previous one in order to account for the latter’s constituting awareness of its contents, etc. Thus, the HO theorist cannot simply solve the Seeming Problem by sticking to her HO guns.

This leaves the Alternative Strategy, which suggests accounting for the seeming involved in awareness-of not in HO terms, but in alternative terms. Of course, the literature on HO theories shows no lack of suggestions as to what might do the relevant job. For various reasons having to do with the theory’s ability to account for intransitive consciousness, HO states are standardly supposed to have certain special features, such as taking the form ‘I am in mental state M,’ so the suggestion might be that this form, and the related features, can be simultaneously used to shed light on how the HO states having those features manage to display the relevant seeming. The crucial question thus becomes: what is it about an HO representation that allegedly makes it a case of awareness-of, a case of things seeming to its subject to be as its contents say they are? Before focusing on that question in more detail in §3, I will first elaborate on my presentation of the Seeming Problem by relating it to another issue: HO misrepresentation.

2. Higher-Order Misrepresentation

We have shifted our attention from the question of what makes a mental state conscious to the following question: what makes HO states such that, by virtue of hosting them, things seem to their subject as these HO states represent them to be? In other words: if awareness-of entails seeming, as the Seeming Platitude has it, and not all representations provide such seeming, then in virtue of what are the relevant HO states instances of awareness-of, instead of mere representations? Interestingly, a version of this question pops up in the discussion on how HO theorists may deal with another problem for their view: HO misrepresentation. It is instructive to see how it does.

What happens if I harbour an HO representation of myself as being in a perceptual state representing a donkey if, in fact, I am in no such perceptual state? There are two ways to go. One, illustrated in the domain of HOT theory by Jonah Wilberg’s ‘No Consciousness’ view, is to hold that there is no consciousness in that case, it only seems as though there is. ^{Footnote 9} The other way, illustrated in the domain of HOT theory by its creator Rosenthal himself, is to hold that “one’s being in a conscious state does not imply being in the state one is aware of being in.” ^{Footnote 10} Or, in the words of Jacob Berger, who defends the same view, “[w]hether the target of a HOT exists, and the degree to which it is accurately represented, are irrelevant to whether or not there is consciousness.” ^{Footnote 11}

The plausibility (or, indeed, intelligibility) of either of these strategies is not the issue here. What matters for present purposes is that both of them illustrate the Seeming Problem. Consider, first, the view Rosenthal and Berger defend, on which consciousness exists as long as there are suitable HOTs, independently from whether their target lower-order states exist. It seems to me as though I am in certain mental states, independently of whether these states exist. Being in a conscious state is here equated with seeming to be in such a state. Surprisingly, the situation is exactly similar on Wilberg’s account, except for the fact that he takes seeming to be independent from consciousness (on which more below, in §4): roughly, he equates being in a conscious state with veridically seeming to be in such a state. In both cases, how things seem to the subject is determined solely by her HO representations; the states that are supposed to be (intransitively) conscious by virtue of these HO representations do not play any role in this seeming. It turns out, then, that, as far as seeming is concerned, the HO theory is not a theory telling us what turns given mental states into states one is aware of oneself as being in, but rather a theory telling us which kinds of representations provide seeming—namely, precisely those seemingly direct, non-inferential representations whose contents take the form ‘I am in mental state M.’ ^{Footnote 12} On the Rosenthal/Berger view, seeming just is state consciousness; on Wilberg’s view, consciousness is something else—a claim he can only endorse at the expense of the plausibility and explanatory adequacy of his account. ^{Footnote 13}

Now, if HO states provide seeming merely in virtue of being representations, it follows that all representations provide seeming—which trivializes the relevant kind of seeming by making it impossible to harbour representational states that do not provide the relevant seeming. That is the Ostrich Strategy, and we have seen that it fails.

The HO theorist will have to reject this trivialization of the notion of seeming: her strategy requires the relevant HO states to provide a kind of seeming that other representations (such as those that play a role in early visual processing) lack. The question, then, becomes: what differentiates HO states that do provide seeming from representations that do not? Here, the HO theorist cannot stick to her HO guns: the HO state’s providing the sort of seeming in question cannot be a matter of that state’s itself being represented by another HO state. That is the Repeat Strategy, and it leads to a vicious regress, as we have seen.

My elaboration of the Seeming Problem makes it clear that seeming is (or is based on) a special feature of those representations that are supposed to constitute awareness of their contents (as opposed to being mere representations thereof)—special in the sense that it is transparent to the subject. ^{Footnote 14} It is a feature that the HO theorist needs her HO states to have in order for them to play the role she wants them to play—to provide awareness of their contents, which is what their targets being (intransitively) conscious mental states is supposed to consist in. The Seeming Problem, in effect, states that that feature can neither be a trivial one, such that all representations have it (the Ostrich Strategy), nor one that is amenable to HO analysis, because of the regress problem (the Repeat Strategy). The HO theorist must look elsewhere for a suitable theory of seeming (the Alternative Strategy). Let us, therefore, have a look at the features the HO theorist ascribes to her HO states in order to find out whether those features may indeed be jointly sufficient for the relevant kind of seeming.

3. The Alternative Strategy: Features of HO States

The Alternative Strategy for HO theorists to deal with the Seeming Problem is to make the following claim: what explains why suitably non-inferential HO states do provide for the relevant seeming, while other representational states do not, is a set of special features they stipulate such HO states to have. Usually, such states are claimed to be occurrent, non-inferential, assertoric and conceptual, and they are thought to involve a first-person reference that is linked to a mental state at the present moment. ^{Footnote 15} Let us see if these features can indeed do the job the Alternative Strategy requires them to perform.

When it comes to assessing such proposed features, it is important to keep in mind that they must make some kind of explanatory contribution. Otherwise, it would be just an exercise in getting the extension right, in which case one could just as well provide a list of all and only those states in the universe that we would label ‘seeming-providing’ and leave it at that. The theory would, then, boil down to a brute identity theory of the sort mentioned in §1—whereas the HO approach aspires to be explanatorily much more robust. ^{Footnote 16} In particular, there must be something about the relevant ingredients that explains why their combined presence entails the relevant seeming.

As an aside, note that one might, of course, resort to a different account of the relevant sort of seeming that highlights the extrinsicality it involves—after all, seemings are seemings for the subject, hence one could identify such seemings with representations that are suitably related to something that may function as ‘the subject.’ The result would be a functionalist theory of seeming akin to, e.g., Bernard Baars’s famous global workspace view of consciousness. ^{Footnote 17} Perhaps Robert Van Gulick’s ‘HOGS’ view can be read in this way. ^{Footnote 18} In that case, however, it becomes particularly clear that we are no longer dealing with a distinctively higher-order approach, for there, then, no longer seems to be reason to restrict the representations that are thus related to the global workspace to HO ones.

Moreover, our reflections on the HO misrepresentation problem in §2 above tell us that the features adduced must be such as to be available to the subject: it cannot be that the presence or absence of these features might simply escape the subject’s notice—the features are supposed to account for that sort of noticing, after all. (To repeat: for Rosenthal and Berger this ‘noticing’ coincides with consciousness itself, while for Wilberg it coincides with the seeming of consciousness.) This makes extrinsic features unlikely candidates: the presence or absence of such features will most likely not make any difference for the subject.

With these preparatory remarks in mind, let us now start with non-inferentiality. It is motivated by the common sense thought that, if I come to represent myself as being in a certain mental state through inference (or otherwise indirectly, e.g., through testimony), that will not count as conferring consciousness on the target state. Yet, non-inferentiality is an extrinsic notion, and hence is unlikely to make the difference that sets representational states providing the right kind of seeming apart from those that do not. Sure, if it seems to me that I am in mental state M, there is no need to infer that I am (save special reasons for thinking I am confused or something such). So non-inferentiality may be thought to help get the extension right. But it doesn’t explain much. For suppose that, through a clever neuroscientist or a cosmic fluke, there were to form a suitable HO representation in me of myself being in mental state M. That state is non-inferential, and in all other respects ‘seemingly immediate.’ Does it suffice for providing a seeming of its contents? Whether its target state exists is irrelevant, as we saw, so that cannot make the difference (and we can amend our thought experiment either way, of course). If ‘proper’ HO states do provide seeming, then the sort of HO state in our cooked-up example should too: there is no relevant difference. But now consider a case in which I arrive at a suitable HO state through inference; but as soon as the state has formed, a clever neuroscientist or cosmic fluke arranges for it to be the case that all traces of its being formed through inference are erased from my mind. In such a case, non-inferentiality is violated, but, again, the resulting HO state would be subjectively indistinguishable from a proper one. (This was why extrinsic features are unlikely candidates for the job of conferring ‘seeming.’) For seeming, then, non-inferentiality does not help.

It will, perhaps, be objected that the point is not that the HO state should be non-inferential but rather that it should seem non-inferential to the subject. In that case, my example of an inferential HO state for which all traces of inferentiality have been erased from the subject’s mind would suffice for seeming. However, this move relies on how things seem to the subject—and it is clear that the requirement of seeming non-inferentiality is of no help for explaining seeming.

Turning now to occurrence, we can be rather brief. A mental state’s being ‘occurrent’ can mean one of two things here. On the first reading, it is synonymous with ‘awareness-providing’ (which seems to be the most informative way of understanding the term, as ‘occurring to the subject’), in which case it, of course, cannot be utilized to illuminate awareness-of. On the second reading, departing from its opposite ‘dispositional,’ it means ‘manifested’—an occurrent state would, on this reading, be an actually existing representation, whereas a dispositional state would be the subject’s disposition to form that representation. It is clear that states that provide awareness-of and hence seeming will have to be occurrent in this second sense of being manifested; but for the rest nothing follows from a state’s being occurrent in this sense with respect to its awareness- or seeming-providing capacities—e.g., occurrent states in early perception of whose contents one is unaware are arguably ubiquitous. Still, occurrence in this second sense is, while non-inferentiality is not, a necessary ingredient for a proper account of seeming.

That the relevant HO states should be assertoric and conceptual makes sense with respect to seeming: if it seems to me that things are a certain way, then that seeming is bound to be assertoric and, presumably, conceptual. So, like occurrence (in the sense of being manifested), these are necessary ingredients for a proper account of seeming (though one might, of course, quarrel about conceptuality). They are, of course, not sufficient: for Freudian reasons, there certainly is room for occurrent (i.e., manifested), assertoric, conceptual (and even non-inferential) thoughts that by definition fail to provide one with awareness of their contents, and hence fail to display the right kind of seeming.

We are left with the mentioned restrictions on the contents of HO states: first-person reference and reference to a mental state, predicationally glued together with reference to the present moment. Setting aside the vexing (and unduly neglected) question as to how first-person reference functions at all, ^{Footnote 19} there seems to be nothing about these particular sorts of contents that would make a difference as to the alleged seeming-providing abilities of the relevant HO states—even when we hold fixed that they obey all the rest of the requirements. Obviously, one may again point to Freud-inspired counterexamples: ‘suppressed’ thoughts to the effect that one is in certain mental states (anger, say) fit the bill but are supposed to be exactly not seeming-providing. But, apart from such counterexamples, I fail to see how the proposal could work. The crucial point is that the relevant HO states should amount to a seeming that things are as they represent them to be. That is, the states should represent their contents in the right way. It may well be that only certain types of content are suitable for being represented ‘in the right way,’ but it seems obvious that, by delineating that suitable range of content, one has merely delineated what can be represented in the right way, not what ‘representing in the right way’ consists in. That is, these restrictions on the contents may again help in fixing the extension, but not in illuminating the emergence of seeming.

But perhaps these last reflections are too quick: could the HO theorist not simply hold that ‘representing in the right way’ just is self-representing? Then, HO states would provide seeming because they are self-representational (and fulfill the other criteria). There are two possible readings of this proposal. The first and most straightforward one holds that a self-representation just is a representation involving, as part of its content, a first-personal reference to its bearer. This seems to be what HO theorists usually have in mind (though not with an eye to accounting for seeming). However, this reading implies that every seeming is a seeming of oneself, and that is highly implausible. It simply need not always be the case that, if things seem to me to be some way, I am among those things. So, let us move on to the other reading of the proposal: instead of explaining ‘self-representation’ in terms of representation of oneself, the proposal now is to explain it as representation to oneself. Self-representation, on this view, is an altogether different sort of representation. This move, however, amounts to a departure from the HO approach tout court. After all, a theory of this special kind of representation plausibly amounts to a theory of intransitive consciousness, equating unconscious states with ‘ordinary’ representations, and conscious ones with self-representations. ^{Footnote 20} Of course, for such a strategy to be successful, a credible account of self-representation is required—in particular one that does not reduce to the above reading in terms of (ordinary) representation involving first-person reference. However, I will not pursue this thought any further here; this paper is concerned with the HO approach only.

I conclude that none of the usual strategies for capturing the correct sorts of HO states help to illuminate why those states, and not other, relevantly similar states, would amount to a seeming of their contents (and this conclusion is silent about the extent to which these same strategies may help establish the right notion of intransitive consciousness). Thus, the Seeming Problem for the HO approach stands: the HO approach relies on an understanding of awareness-of that does not respect the Seeming Platitude. In §1, we saw that the Ostrich Strategy and the Repeat Strategy do not solve the problem; in this section, I have argued that the Alternative Strategy is of no avail either.

Again, for a correct understanding of the survey of criteria I have just executed, it is important to keep in mind that I have been searching for an account of what makes the relevant HO states provide the right kind of seeming, which, according to the Seeming Platitude, is a necessary feature of a representation to amount to an ‘awareness of’ its contents at all (as I explained in §1). To my knowledge, that question has neither been posed nor elaborated on in this way; awareness-of has been largely treated as a given in HO theories. Still, there are some relevant considerations to be found in the literature; I will now turn to those, with an eye to illuminating the relation between seeming and consciousness.

4. Seeming and Consciousness

If the HO project is to really get off the ground, its conceptual underpinnings must be secure—in particular, the notion ‘awareness-of,’ also known as ‘transitive consciousness,’ must be clarified. I have raised a problem for that notion: the Seeming Problem. That problem relies on the idea that to be aware of things being some way involves it seeming to one that they are that way (the Seeming Platitude).

Rosenthal at one place problematizes that connection when he writes:

This is a remarkable statement. If awareness-of (‘transitive consciousness’) is consistent with being asleep, as Rosenthal asserts, it would follow that, while immersed in a deep, dreamless sleep, things seem to one to be various ways. Although one might use the term ‘seeming’ in that way, that is not the way it is used in the Seeming Platitude—after all, that platitude allows for a distinction between representations that do and those that do not enjoy such seeming, whereas Rosenthal’s notion of awareness-of does not. Thus read, Rosenthal is, in effect, pursuing the Ostrich Strategy, which leads to an objectionable trivialization of seeming.

Still, why can Rosenthal not simply bite that bullet and continue as he does? Well, let us see what it would mean for his HO view if this were indeed his last word on awareness-of and seeming. Suppose that through clever neuroscience or a cosmic fluke a state representing oneself as being in mental state M is created while one is immersed in a deep, dreamless sleep. On Rosenthal’s view, this is sufficient for state consciousness. It is hard to see why this should be impossible—yet, on Rosenthal’s view, it would amount to being in a conscious state while being unconscious, which is absurd. The absurdity arises precisely because Rosenthal on the one hand accepts a trivialization of the notion of awareness-of while relying on a more robust notion of awareness-of for his HO theory of state consciousness. In particular, he relies for his HO view on a notion of awareness-of that somehow presupposes (creature) consciousness. And, plausibly, the Seeming Platitude implies just such a connection between awareness-of and creature consciousness.

But perhaps Rosenthal can have his cake and eat it too by rejecting the Seeming Platitude altogether. That is, he could adopt a notion of awareness-of that can occur with and without seeming. Then, he could hold that awareness-of is just representation, and thus can indeed occur while being dreamlessly asleep, but add that, in the case of HOTs, there is, in fact, seeming, thus rebutting the sort of worry I just raised. After all, as we saw in §2, Rosenthal himself claims that HOTs suffice for it to seem to their subject that things are as they represent them to be (even in the absence of their target lower-order states). However, this does not help him escape from the Seeming Problem: he needs some account of this seeming, and if his account of awareness-of is silent on the matter, he will have to come up with an alternative. We have seen that an alternative along HO lines leads to a regress (the Repeat Strategy), while attempting to account for seeming in different terms (the Alternative Strategy) leads to abandoning the HO approach altogether.

Wilberg has a more elaborate discussion of seeming, in the context of his defence of the ‘No Consciousness’ view (see §2 above)—though he focuses on the relation between seeming and intransitive consciousness, not on the relation between seeming and transitive consciousness. Let us see whether his arguments have any force when rephrased so as to target the tie between transitive consciousness and seeming I have relied on in my argumentation. He argues that “consciousness is not solely a matter of appearance” by rebutting two arguments to the contrary: (1) that when one seems to be in a conscious state, one is in a conscious state, because these are subjectively indistinguishable; and (2) that one can only take subjects’ reports of their conscious states as evidence for those states if one takes those states to correspond to how things seem to the subject. ^{Footnote 22} Against (1), Wilberg notes, in effect, that the notion of subjective indistinguishability involved begs the question since it is treated as the measure for consciousness while capturing only seeming. Against (2), Wilberg alludes to the well-known fact that subjects can misreport their own conscious states. ^{Footnote 23}

Both considerations do not affect the seeming/awareness-of tie I have been working with, recorded by the Seeming Platitude. With regard to (2), I can be brief: people may have a hard time keeping a clear eye on how things seem to them over time; they may regularly confuse theorizing with observation, etc. That is not a problem for my claim: at the moment of reporting, things simply seem to them to be as these people say they do (at least if they are sincere and manage to hold things more or less fixed for as long as their reports last). And that may, of course, include incorrect seemings concerning earlier seemings. ^{Footnote 24} Moving now to (1), I have nowhere claimed that seeming is the same as being in a conscious state (though I certainly think that is true), precisely in order not to prejudge the issue. I have left the notion of state consciousness out in order to focus on what is doing the explanatory work in the HO accounts: transitive consciousness. If there indeed is an appearance/reality distinction when it comes to state consciousness, I thus urge that we focus on the appearance side of that distinction. When is it the case that things appear (seem) to one to be one way or the other? The HO theorist’s answer: just when there is transitive consciousness that things are as they appear (seem) to be. And that is where the Seeming Problem arises.

Since I have been formulating a challenge for ‘all higher-order theories,’ a brief reflection on William Lycan’s ‘Simple Argument’ challenge seems to be a fitting way to end this paper. Lycan’s argument runs as follows:

1. (1) A conscious state is a mental state whose subject is aware of being in it. [Definition]

2. (2) The ‘of’ in (1) is the ‘of’ of intentionality; what one is aware of is an intentional object of the awareness.

3. (3) Intentionality is representation; a state has a thing as its intentional object only if it represents that thing.

Therefore,

1. (4) Awareness of a mental state is a representation of that state. [2, 3]

And, therefore,

1. (5) A conscious state is a state that is itself represented by another of the subject’s mental states. [1, 3] QED ^{Footnote 25}

To be sure, what one is aware of is somehow represented; and if one accepts (1) as a definition, then there will be conscious states in the defined sense only if there are representations of those states (although the argument doesn’t settle that such representations must be additional states). So, to that extent, Lycan’s argument remains untouched by my argument. However, Lycan’s argument tells us nothing about awareness-of except that it is representational (4). And to repeat: definition (1) is motivated by pre-theoretical considerations; likewise, my linking awareness-of with seeming is motivated by pre-theoretical considerations (the Seeming Platitude). Hence insofar as Lycan’s argument is acceptable, so is my challenge to come up with a more substantive story as to what makes representations amount to seemings.