I gave up on reductive self-representationalism for quite general reasons, reasons affecting all representationalisms. As such, one might be tempted to suggest adopting some non-reductive form of S theory. For example, if one adopts the phenomenal intentionality[30] view, one might hold that whatever phenomenal representation is, consciousness represents itself in that way. It seems like this view might be just another way of describing the same phenomenological facts belabored in the previous section. If that is so, the phrases “phenomenal intentionality” and “acquaintance” are going to be basically synonymous, and the advocate of the former terminology can just translate. If we build nothing into the notion of representation other than the idea that something (an object, property, episode of consciousness, or whatever) is phenomenally manifest (to someone), then the views are indistinguishable at the phenomenological level and, maybe, the ontological level as well.
If this is not what is intended, however, then it is probably because the phenomenal intentionality theorist wants to mark an important distinction between intentionality (representation) and acquaintance. Perhaps they would prefer not to be committed to acquaintance if possible, and there are several reasons they might want to avoid such a commitment. But I will argue that in a certain sense, to be plausible at all, all forms of representationalism, reductive and non-reductive (including a phenomenal intentionality-based representationalism), ought to embrace a type of acquaintance relation.
Consider, for a moment, fictionalist representationalism about sensory qualities (projectionism about colors, for example). One could embrace a view according to which the sensory qualities are phenomenally manifest, though they in fact are never really instantiated by anything. In such a case, one would not want to think of sensory phenomenal consciousness as a matter of bearing a real acquaintance relation to such qualities or quality instances. Instead one might prefer an adverbial construal of the situation that avoids any commitment to anything literally having (or perhaps even to there being) the properties phenomenally represented. On this view, one denies that there is a relation that supports existential quantification over these immediate objects (whatever they are), and one cannot conclude from the fact that one is phenomenally conscious of a red patch that there exists a red patch of which one is conscious.
Of course, this failure of existential quantification won’t apply in the case of one sort of object, namely the conscious episode itself. But it will not be because it is an object of phenomenal intentionality that one can validly, existentially generalize from it; generally that fails, just as in other intentional (and intensional) contexts. Rather, it will be because it is the subject or bearer of phenomenal intentionality that one can validly generalize from it. In other words, we take episodes of consciousness to be individuals that have this pseudo-relational property. That is why we can quantify over them, and not because of anything that they pseudo-bear that pseudo-relation to. Such “objects,” after all, can be nonexistent. Thomas Reid’s “ambulo ergo sum” would be appropriate here, not the Cartesian Cogito conceived in a phenomenologically performative way.
This situation is rather paradoxical. If the only mode of awareness of our own consciousness (even supposing ubiquitous self-manifestation) is via phenomenal intentionality so construed, then our evidence for the very existence of our own consciousness is really no better than our evidence for the existence of phenomenal colors. Just as we might be persuaded that there really are no phenomenal colors, perhaps we could become persuaded that there is no such thing as phenomenal consciousness either. I regard this as absurd. It is like saying that perhaps we only think we think, or that perhaps it only appears to us that things appear to us. Consequently, consciousness must bear some evidentially relevant relation to itself and to its own being, other than the phenomenal intentionality pseudo-relation it pseudo-bears to phenomenal colors.
Thinking of consciousness as “being-appeared-to-existingly” does not help here, since that applies to phenomenal colors and all other perceived pseudo-objects and pseudo-qualities as well. Any theorist committed to self-manifestation should not try to construe this as just a case of phenomenal intentionality as just described. From our self-consciousness we can conclude that we do exist, and this is not just because we know by inference or in some other way that we are the bearer of a property, as in Reid’s Ambulo. We must be acquainted with our own existence—in the sense that every episode of consciousness, however individuated, is acquainted with its own existence. This applies to the subject-pole. What about the object-pole?
In the context of the theory of perceptual consciousness, I think it is a mistake to maintain that any view according to which one can always legitimately quantify over the “immediate objects of conscious awareness” is committed either to some form of direct realism (or perhaps a disjunctivist version thereof) or to old-fashioned sense-datum theory. Any plausible form of representationalism—fictionalist or realist, externalist or internalist, reductive or non-reductive, is, I’ll argue, committed to such quantification, though this must be understood in a particular way. I am not, of course, saying that if we seem to consciously visually perceive a pink rat then we can infer that there exists a pink rat that we see. There is, however, something other than just the conscious state itself (qua whole) that we can legitimately, existentially quantify over.
Our conscious perception of differentiation (in unity) entails, even on a representationalist view, that there exists something of which we are aware, namely, at the least, differentiation (or contrast) itself. For example, suppose I hallucinate purple and pinkish smoke clouds arising from stereo speakers as “Fairies Wear Boots” comes on. Evidently I cannot conclude that those purple and pinkish clouds exist. Still, I maintain, we can conclude that there exists some differentiation or contrast of which we are aware. By hypothesis, we cannot say that the difference is that between the pink smoke cloud and the purple one, since they do not exist. Differences between non-existent objects cannot be appealed to in order to make sense of real differences.[31] But we are aware of some real and phenomenally manifest differentiation here. If we say no to that, we’d have to assume that reflection is simply inaccurate when it comes to such hallucinations; that we seem to have a differentiated experience when in fact there is no phenomenal difference at all. But if that itself is a phenomenal state, say a conscious reflection on an ongoing hallucination, we have the same problem all over again.
If the difference we are aware of is not and is not to be accounted for by a difference in the objects (since they do not exist), it must be a matter of the difference in the representations. Hence, albeit in an indirect manner and, as it were, under the guise of a difference in the pink and purple clouds, we must be aware of some differentiation inherent in the representational states themselves.[32] If we reject disjunctivisms, then we ought to maintain that in every case of differentiated phenomenal awareness we are, in fact, acquainted with (and not merely representing) the differences inherent in our episodes of phenomenal consciousness. This is, at any rate, what I think is the most plausible account, even if the considerations just given don’t absolutely clinch it. Again, it is not that there cannot be some sort of representationalist response.[33] It is, rather, that I regard the line I take to involve fewer epicycles.
We cannot make good sense of the appearance of a phenomenal difference without direct awareness of differentiation. But, by hypothesis, in the case of hallucination it cannot be that we are aware of a real difference in the objects of representation. Moreover, it cannot be a difference in something that is hidden from conscious awareness—some difference in the externalist conditions determining the content of the representational states, for example—that we are aware of. The most plausible candidate, then, is that we are directly aware of (acquainted with) differentiation or modifications in consciousness itself (and hence the Transparency Intuition (see page here) is, strictly speaking, false; we are indeed aware of features of consciousness itself even in so-called “first-order” awareness). This applies to both reductive and non-reductive forms of representationalism. If this line of thought is correct, representationalist theories really presuppose some sort of non-representationalist, acquaintance theory.
Implicit in the above discussion is something like this definition of acquaintance:
Acquaintance =Df (1) the relation (R) the subject (s) of consciousness (i.e., the episode or stream itself) bears to the differentiated phenomenal manifold (D<x1, x2….xn>), such that (2) if sR[D< x1, x2….xn>], then we may infer truly that (∃x)(sRx).
Of course, clause (2) can be taken as redundant, given the usual understanding of real relations and that the R of clause (1) is so taken. But in this context it is important to emphasize the point. The first clause is just an inner-ostensive phenomenological characterization that assumes that the relational appearances are indeed the reality; the second is a logico-ontological characterization. Importantly, we can “quantify in” here: If, in any concrete particular case, we stand in that relation to some phenomenally differentiated field, then we can truly infer that there exists something differentiated we stand in that relation to. However, it is in general not the represented (or intentional) objects that we are thus acquainted with. It is, rather, the common factor of all episodes of phenomenal consciousness, be they hallucinations, dreams, or the “perceptions” of brains in vats. This, again, is often precisely what is denied when one says that a state is one of representation as opposed to acquaintance. If it is true that I represent A, I cannot infer from this that there is some X such that I represent X. Adverbialisms and other forms of representationalism were, recall, developed precisely around this insight in order to overcome the problems of sense-datum and other relational theories of perception. Is the theory I am suggesting here a form of old-fashioned sense-datum theory?
Unfortunately I cannot give a short answer to that question and can’t give all of the long answer here. This will have to suffice: (1) We can regard sensory qualia (or hyle) as being complex, relational properties of consciousness (and its concrete embodiment in brain processes); in fact, they could be something like irresolvable structural properties that appear simple precisely because they mark a limit of our sensory resolution. (2) In order to flesh this out, we must reject the Revelation Thesis—the thesis that acquaintance yields up all of the properties of sensory qualia. In particular, we can (and should) reject the idea that acquaintance tells us all of the categorial properties of sensory qualities. There is no good reason to believe that it does. Hence, they could fail to seem relational and yet still be relational. This is a solution to the “Grain Problem”—a problem arising from the fact that brain properties are “complex” and relational while sensory qualities (phenomenal colors, tones) do not seem to be. If we infer from the appearances then we cannot consistently hold that they are identical to brain properties. But we have no good reason for making that inference.[34] (3) It is not hard to understand why the sensory qualities would be integrated into a spatialized and “intentionally animated” grid that can serve as a “user interface” for us to deal with the external world, yielding a “transparent” manifold in Metzinger’s sense, a manifold we are built to systematically and automatically “see right through”—causing us to suffer from a sort of delusion of direct realism (see Williford et al. 2012; Williford 2013; Metzinger 2004, p. 163, and Revonsuo 2006). Finally, (4) appeals to the “Transparency Intuition” (in Tye’s sense of “transparency”) thus carry no serious weight. All the phenomenological data in question are accounted for by 1–3, and there are good independent lines of reasoning for each of these (that we do not have time to go into here).
I’ve argued that the notion of acquaintance, when interpreted in the rather minimal, phenomenological, and logico-ontological way proposed, is the proper notion for characterizing the relationship between consciousness and the differentiated but unified multimodal experiential manifold. Moreover, on the view proposed here, consciousness bears this same relation, generically understood, to itself.
If the episode of consciousness bears the relation to itself, then evidently there is something to which it bears that relation. But, nontrivially, we could not have the sort of direct evidence of its existence that we do have if consciousness were not self-acquainted—and acquainted with its own existence. And if the episode of consciousness bears the relation to the differentiated manifold that constitutes the surface that serves as its contact with a differentiated reality beyond it—i.e., if it bears it to a differentiated portion of itself—then there is something differentiated of which it is non-representationally aware. One is directly aware of the difference or differentiation even if one only, strictly speaking, represents what the things so differentiated happen to be or interprets them as being such and such (mental, physical, surfaces of objects, internal sense data, quotidian objects, etc.). In other words, I can see that red is not blue even if I do not know what colors are exactly, or if they are in physical space or only in a virtual space in my brain. One does not merely represent this difference or differentiation. One is acquainted with oneself and with the differentiation one contains. Of course, one is also acquainted with the apparently intrinsic properties that mark these internal differences, but again, this need not mean that the properties are in fact non-relational and simple. In fine, we are self-acquainted and acquainted with a differentiated manifold and thus, at some level, with real differences in the mind, the world, or world-mind boundary. [35] The acquaintance relation consciousness bears to itself is, generically speaking, identical to the relation it bears to sensory qualia (or hyle)—which are taken here as ultimately just transient modifications in the unfolding embodiment of consciousness. It is important to understand that this does not imply that there is a special type of sensory quality (a “me-ish” quale) peculiar to consciousness. It is as diaphanous as G.E. Moore said. Remember that the acquaintance in an instance of acquaintance with phenomenal red is identical with the acquaintance in an instance of acquaintance with phenomenal C#, even though phenomenal red and C# are utterly heterogeneous.
One might reasonably ask for a more substantive definition or account of acquaintance. The definition given relies on phenomenology and logic and is otherwise quite empty. But this is as it should be, in my view. Any further account of the nature of acquaintance, of what the acquaintance relation is, will be the result of empirical inquiry and a well-supported a posteriori identification.