Partial unity of consciousness:

A prima facie defense

1 Introduction

Under the experimental conditions characteristic of the “split-brain” experiment, a split-brain subject’s conscious experience appears oddly dissociated, as if such a subject had two streams of consciousness. In most respects, however, split-brain subjects appear no different from “normal” subjects, whom we assume have only a single stream of consciousness. The tension between these impressions gives rise to a debate about the structure of consciousness in split-brain subjects.

That debate has for the most part been pitched between two possibilities: that a split-brain subject has a single stream of consciousness, associated with the brain (or with the subject) as a whole, or that she has two streams of consciousness, one associated with each hemisphere. Considerably less attention has been paid to the possibility that a split-brain subject has a single but an only partially unified stream of consciousness, a possibility that has been articulated most clearly by Lockwood, 1989 (see also Trevarthen, 1974; Moor, 1982).

The partial unity model of split-brain consciousness is interesting for reasons that go far beyond the split-brain consciousness debates themselves. Most saliently, the model raises questions about subjects of experience and phenomenal perspectives, about the relationship between phenomenal structure and the neural mechanisms and correlates of consciousness, and about the place for the content/vehicle distinction in psychology.

This paper examines two objections that have been raised to the partial unity model. These objections presumably account for how relatively little attention the model has received. Because I argue that neither of these objections impugns the partial unity model in particular, the paper constitutes a preliminary defense of the partial unity model, working to show that it is on par with its clearest contender, a version of the conscious duality model.

2 The split-brain phenomenon

The split-brain experimental paradigm typically involves carefully directing perceptual information to a single hemisphere at a time, to the extent possible. This is relatively simple to understand in the case of tactile perception. Suppose you blindfold a split-brain subject (or in some other way obscure his hands from sight) and put an object in his left hand, say, a pipe. Since patterned touch information transmits from each hand only to the contralateral (opposite side) hemisphere (Gazzaniga, 2000: 1299), tactile information about the pipe will be sent from the subject’s left hand to his right hemisphere (RH). In a “non-split” subject, the corpus callosum would somehow transfer this information to, or enable access by, the left hemisphere (LH) as well. In the split-brain subject, however, this tactile information more or less stays put in the initial hemisphere that received it. Meanwhile, in a large majority of the population, the right hemisphere is mute. A split-brain subject is therefore likely to say, via his LH, that he doesn’t know what he’s holding in his left hand. A few minutes later, however, while still blindfolded, the subject can select the object he was holding a minute ago from a box of objects, using that same hand, suggesting that the object was not only felt and recognized but that a representation of it was encoded and retrieved. The subject may even draw a picture of a pipe, again using the left hand, which is under dominant control of the right hemisphere (Levy 1969). Visual, auditory, olfactory, pain, posture, and temperature information may all be lateralized, to varying degrees, under some conditions

What makes such findings interesting for thinking about conscious unity is this. On the one hand, a split-brain subject can respond to stimuli presented to either hemisphere in ways that we think generally require consciousness. On the other hand, a subject can’t respond to stimuli in the integrated way that we think consciousness affords, when the different stimuli are lateralized to different hemispheres (or when a response is elicited not from the hemisphere to which the stimulus was presented, but from the other). For example, a very basic test for the “split-brain syndrome” is a simple “matching” task in which the subject is first required to demonstrate recognition both of an RH-presented stimulus and of an LH-presented stimulus—by pointing to a picture of the referents of the presented words, by drawing a picture, etc.—and is then simply asked whether the stimuli are the same or different. In the paradigmatic case, the subject can perform the former sort of task, but not the second, apparently simpler task. This is what first suggests—obviously not conclusively—that the hemispheres somehow have different streams of consciousness: after all, I could demonstrate what I was conscious of and you could demonstrate what you were conscious of, without either of us being able to demonstrate whether or not we’d been conscious of the same thing.

Such results notwithstanding, a number of philosophers have defended some kind of unity model (UM) of split-brain consciousness, according to which a split-brain subject (at least typically) has a single stream of consciousness. In the only version of the unity model invariably mentioned in the literature on the structure of split-brain consciousness, a split-brain subject has a single stream of consciousness whose contents derive exclusively from the left hemisphere. It’s actually not clear that anyone ever defended this version of the model; a couple of theorists (Eccles 1973, 1965; Popper and Eccles, 1977) are widely cited as having denied RH “consciousness”, but they may have been using the term to refer to what philosophers now call “self-consciousness” (see especially Eccles, 1981). The difficulty with that version of the UM is that a lot of RH-controlled behavior so strongly appears to be the result of conscious perception and control. (As Shallice said, of RH-controlled performance on the Raven Progressive Matrices Task (Zaidel, Zaidel, and Sperry, 1981), “If this level of performance could be obtained unconsciously, then it would be really difficult to argue that consciousness is not an epiphenomenon. Given that it is not, it is therefore very likely, if not unequivocally established, that the split-brain right hemisphere is aware.” (Shallice, 1997: 264)) Contemporary versions of the unity model (Marks 1981; Hurley 1998; Tye, 2003; Bayne, 2008), then, all assume that conscious contents derive from both hemispheres. I will make this same assumption here.

The major alternative to the unity model is the conscious duality model (CDM). According to the CDM, a split-brain subject has two streams of consciousness, each of whose contents derive from a different hemisphere. This model appealed particularly to neuropsychologists (e.g., Gazzaniga, 1970; Sperry, 1977; LeDoux, Wilson, and Gazzaniga, 1977; Milner, Taylor, and Jones-Gotman, 1990; Mark, 1996; Zaidel et al. 2003; Tononi 2004), and several philosophers have defended or assumed it as well (e.g., Dewitt 1975; Davis, 1997).[1]

Since both the CDM and contemporary versions of the UM allow that conscious contents derive from both hemispheres, what is at issue between them is whether or not RH and LH experiences are co-conscious with each other—that is, whether they belong to one and the same or to two distinct streams of consciousness. Unsurprisingly, there is disagreement about what co-consciousness (or “conscious unity”) is, and whether there is even any single relation between conscious phenomena that we mean to refer to when speak of someone’s consciousness as being unified (Hill, 1991; Bayne and Chalmers, 2003; Tye, 2003; Schechter, forthcomingb). It is however possible to articulate certain assumptions we make about a subject’s consciousness—assumptions concerning its unity—that appear to somehow be violated in the split-brain case. As Nagel says, we assume that, “for elements of experience… occurring simultaneously or in close temporal proximity, the mind which is their subject can also experience the simpler relations between them if it attends to the matter” (Nagel, 1971: 407). We might express this assumption by saying that we assume that all of the (simultaneously) conscious experiences of a subject are co-accessible. Marks, meanwhile, notes that we assume that two experiences “belong to the same unified consciousness only if they are known, by introspection, to be simultaneous” (1981: 13). That is, we assume that any two simultaneously conscious experiences of a subject are ones of which the subject is (or can be) co-aware.

Finally, we assume that there is some single thing that it is like to be a conscious subject at any given moment, something that comprises whatever multitude and variety of experiences she’s undergoing (Bayne, 2011). We assume, that is, that at any given moment, any two experiences of a subject are co-phenomenal. Although the paper is most centrally concerned with co-phenomenality, I will basically assume here that whenever one of these relations holds between two experiences, the other two do as well, and vice versa. (This assumption is actually quite controversial, but I don’t think its truth or falsity affects the central issues under consideration in this paper, so long as we view these relations as holding of experiences rather than contents; see Schechter, forthcominga.) For simplicity’s sake, I will also focus only on synchronic conscious unity—the structure of split-brain consciousness at any given moment in time—to the extent possible. Accordingly I will speak simply of the co-consciousness relation in what follows.[2]

Let us say that streams of consciousness are constituted by experiences and structured by the co-consciousness relation. According to the unity model of split-brain consciousness, a split-brain subject has a single stream of consciousness because right and left hemisphere experiences are all co-conscious. According to the conscious duality model, co-consciousness holds intrahemispherically but fails interhemispherically in the split-brain subject, so that the subject has two streams of consciousness, one “associated with” each hemisphere.

Despite their disagreements, the CDM and the UM share a very fundamental assumption: that co-consciousness—the relation between experiences, out of which streams of consciousness are constructed—is a transitive relation. In this one respect, these two models have more in common with each other than either of them does with the partial unity model (PUM). For the PUM drops the transitivity assumption, allowing that a single experience may be co-conscious with two others that are not co-conscious with each other. Streams of consciousness may still be constructed out of the co-consciousness relation, but it is not necessary that every experience within a stream be co-conscious with every other. In this model, then, conscious unity admits of degrees: only in a strongly unified stream of consciousness is co-consciousness transitive. According to both the UM and the CDM, in contrast, a split-brain subject has some whole number of strongly unified streams of consciousness, while according to the PUM, a split-brain subject has only a partially (or weakly) unified consciousness.

Because there are other notions of conscious unity besides those of that co-awareness, co-accessibility, and co-phenomenality, there are other possible partial unity models. The truth is that CONSCIOUS UNITY is (to borrow Block’s (1995) term) a “mongrel concept” (Schechter, forthcomingb); when we think of what it is to have a “unified” consciousness, we think of a whole host of relations that subjects bear to their conscious experiences and that these experiences bear to each other and to action. Talk of a “dual” consciousness may connote a breakdown of all these relations simultaneously. In reality, though, these relations may not stand or fall all together; in fact, upon reflection, it’s unlikely that they would. One intuitive sense of what it means to have a partially unified consciousness, is a consciousness in which some of these “unity” relations still hold, and others do not (Hill).

This is not what I mean by a “partially unified consciousness,” however. In one possible kind of partial unity model, some conscious unity relations, but not others, hold between experiences. In the kind of partial unity model under consideration here, however, conscious unity relations hold between some experiences, but not others. This point will be crucial to understanding the choice between the PUM and the CDM.[3]

The PUM of split-brain consciousness has several prima facie strengths. Most obviously, it appears to offer an appealingly intermediate position between two more extreme models of split-brain consciousness. The UM must apparently implausibly deny failures of interhemispheric co-consciousness; the CDM is apparently inconsistent with the considerable number of cases in which it is difficult or impossible to find evidence of interhemispheric dissociation of conscious contents.

Furthermore, the PUM that makes some kind of neurophysiological unity the basis for conscious unity. Against those who would claim that splitting the brain splits the mind, including the conscious mind, some philosophers argued that a putatively single stream of consciousness can be “disjunctively realized” (Marks, 1981; Tye, 2003). Lockwood’s defense of the PUM in contrast appeals explicitly to the fact that the “split” brain is not totally split, but remains physically intact beneath the cortical level: the cortically disconnected right and left hemisphere are therefore associated with distinct conscious experiences that are not (interhemispherically) co-conscious; nonetheless, these are all co-conscious with a third set of subcortically exchanged or communicated conscious contents. Many will be attracted to a model that makes the structure of consciousness isomorphic to the neurophysiological basis of consciousness in this way (Revonsuo, 2000).[4]

Another significant source of the PUM’s appeal is its empirical sensitivity, in a particular sense. Lockwood sought to motivate the PUM in part by considering the possibility of sectioning a subject’s corpus callosum one fiber at a time, resulting in increasing degrees of (experimentally testable) dissociation. Would there be some single (one wants to say “magical”) fiber which, once cut, marked the transition from the subject’s having a unified to a dual consciousness? Or would the structure of consciousness change just as gradually as the neural basis of her conscious experience? Lockwood implies that nothing but a pre-theoretic commitment to the transitivity of co-consciousness would support the first answer: “there remains something deeply unsatisfactory about a philosophical position that obliges one to impose this rigid dichotomy upon the experimental and clinical facts: either we have just one center, or stream, of consciousness, or else we have two (or more), entirely distinct from each other” (Lockwood, 1989: 86).

Lockwood’s thought experiment is not wholly fictitious: callosotomy is in fact now routinely performed in stages, with predictable degrees and sorts of dissocation evident following section at particular callosal locations (e.g. Sidtis et al., 1981). “Partially split” subjects really do seem somehow intermediate between “non-split” and (fully) “split-brain” subjects. Surely one appealing characterization of such subjects is that the structure of their consciousness is intermediate between (strongly) unified and (wholly) divided or dual.