109

. Neural coalitions But there is a problem in the reasoning of the last section. Whenever asubject reports such phenomenology, that can only be via the activation of the frontal neural basis of global access. And how do we know whether those frontal activations are required for– indeed, are part of– theneural basis of the phenomenology? Metaphysical correlationists say that they are; epistemic correlationists say we can’t know. This section draws together strands that have been presented to argue that both kinds of correlationism are wrong because we have empirical reason to suppose that activation of working memory circuits are not part of the neural basis of phenomenology (not part of either the core or total neural basis). A preliminary point: Pollen (in press) summarizes evidence that prefrontal lobotomies on both sides and other BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Block: Consciousness, accessibility, and the mesh frontal lesions do not appear to decrease basic perceptual content such as luminance or color. Frontal damage impairs access but it doesn’t dim the bulb (Heath et al. 1949). Still, it could be said that some degree of frontal activation, even if minimal, is part of the background required for phenomenal consciousness, and, epistemic correlationists would allege, once there is so much frontal damage that the subject cannot report anything at all, there is no saying whether the person has any phenomenal consciousness at all. In the rest of this section, I give my argument against this view, the one that the second half of the article has been leading upto: If wesupposethatthe neural basis ofthe phenomenology does not include the workspace activations, we can appreciate a neural mechanism by which phenomenology can overflow cognitive accessibility. There is convincing evidence that the neural processes underlying perceptual experience can be thought of in terms of neural network models (see Koch 2004, Ch. 2, pp. 19, 20). In visual perception, coalitions of activation arise from sensory stimulation and compete for dominance in the back of the head, one factor being feedback from the front of the head that differentially advantages some coalitions in the back. Dominant coalitions in the back of the head trigger coalitions in the front of the head that themselves compete for dominance, the result being linked front and back winning coalitions. Support for this sort of model comes from, among other sources, computerized network models that have confirmed predictive consequences (see Dehaene & Nacchache 2001; Dehaene et al. 1998; 2006). Furthermore, some recent experiments (Sergent & Dehaene 2004) provide another line of evidence for this conclusion that is particularly relevant to the themes of this article. This line of evidence depends on a phenomenon known as the “attentional blink.” The subject is asked to focus on a fixation point on a screen and told that there will be a rapid sequence of stimuli. Most of the stimuli are “distractors,” which in the case of the Sergent and Dehaene version are black nonsense letter strings. The subject is asked to look for two “targets”: a white string of letters, either XOOX or OXXO, and a black name of a number, for example, “five.” One or both targets may be present or absent in any given trial. At the end of the series of stimuli, the subjects have to indicate what targets they saw. In the standard attentional blink, subjects simply indicate the identity of the target or targets. The standard finding is that if the subject saw the first target (e.g., “XOOX”), and if the timing of the second target is right, the second target (e.g., “five”) is unlikely to be reported at certain delays (so long as it is followed by distractors that overwrite the phenomenal persisting representation, as in Figure 5). In this setup, a delay of about 300 msecs makes for maximum likelihood for the second target to be “blinked.” Sergent and Dehaene used a slight modification of this procedure in which subjects were asked to manipulate a joystick to indicate just how visible the number name was. One end of the continuum was labeled maximum visibility and the other was total invisibility. (See Fig. 5.) The interesting result was that subjects tended to indicate that target 2 was either maximally visible or maximally invisible: intermediate values were rarely chosen. This fact Figure 5. TheAttentional Blink. A sequence of visual stimuli in which the first target is a white string of letters, either XOOX or OXXO and the second target is the name of a number. At the end of the series the subject is asked to indicate how visible target 2 was and whether target 1 was present, and if so, in which form. suggests a competition among coalitions of neurons in the front of the head with winners and losers and little in between. Usually, the coalition representing the number word either wins, in which case the subject reports maximum visibility, or it loses, in which case subjects report no second target and there is no cognitive access to it at all. I have guessed (Block 2005) that there can be coalitions in the back of the head which lose by a small amount and thus do not trigger winning coalitions in the front, but which are nonetheless almost as strong as the back of the head coalitions that do trigger global broadcasting in the front. The subject sees many things, but only some of those things are attended to the extent that they trigger global broadcasting. A recent study (Kouider et al. 2007) suggests that indeed there are strong representations in the back of the head that do not benefit from attention and so do not trigger frontal activations. (See also Tse et al. 2005 for convergent results.) Kouider et al. contrasted a subliminal and supraliminal presentation of a stimulus, a lower-case word. In the subliminal case, the stimulus was preceded and succeeded by masks, which have the effect of decreasing the visibility of the stimulus (and, not incidentally, decreasing recurrent neural activation– see Supe`r et al. 2001b). In the supraliminal case, the masks closest to the stimulus were omitted. The supraliminal but not the subliminal stimulus could be identified by the subjects when given a forced choice. In the key manipulation, the subject was told to look for an upper-case word, ignoring everything else. In those conditions of distraction, subjects claimed that they were aware of the lower-case stimuli in the supraliminal case but that they could hardly identify them because they were busy performing the distracting task on the upper-case stimulus (which came later). The difference between the supraliminal and subliminal stimuli in conditions of distraction was almost entirely in the back of the head (in occipito-temporal areas). Supraliminal stimuli activated visual areas in the back of the head strongly (subliminal stimuli did not) but did not activate frontal coalitions.11 The strong activations in the back of the head did, however, modulate frontal activity. Kouider et al. (2007) and Dehaene et al. (2006) acknowledge that there are highly activated losing 497 BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Block: Consciousness, accessibility, and the mesh coalitions in the back of the head. They argue that such losing coalitions are the neural basis of “preconscious” states– because they cannot be reported. But the claim that they are not conscious on the sole ground of unreportability simply assumes metaphysical correlationism. A better way of proceeding would be to ask whether a phenomenal state might be present even when it loses out in the competition to trigger a winning frontal coalition. Here is the argument that the second half of this article has been building up to: If we assume that the strong but still losing coalitions inthebackof thehead are the neural basis of phenomenal states (so long as they involve recurrent activity), then we have a neural mechanism which explains why phenomenology has a higher capacity than the global workspace. If, on the contrary, we assume that the neural basis of phenomenology includes workspace activation, then we do not have such a mechanism. That gives us reason to make the former assumption. If we make the former assumption– that workspace activation is not part of the neural basis of phenomenology– we have a mesh between the psychological result that phenomenology overflows cognitive accessibility and the neurological result that perceptual representations that do not benefit from attention can nonetheless be almost as strong (and probably recurrent) as perceptual representations that do benefit from attention. The psychological argument from overflow showed that the machinery of phenomenology is at least to some extent different from that of cognitive accessibility, since something not in cognitive accessibility has to account for the greater capacity of phenomenology. What the meshargumentaddsisthat the machinery of phenomenology does not include the machinery of cognitive accessibility. Of course, my conclusion that the neural machinery of cognitive access is not partially constitutive of phenomenology leaves room for causal influence in both directions. And it may be that top-down causal influence is almost always involved in making the phenomenal activations strong enough. But that is compatible with the possibility of the relevant amplification happening another way, for example, by recurrent loops confined to the back of the head or even by stimulation by electrodes in the brain, and that is enough to show that top-down amplification is not constitutively necessary. Myfirst conclusion then is that the overlap of the neural machinery of cognitive access and the neural machinery of phenomenology can be empirically investigated. Second, there is evidence that the latter does not include the former. These points are sufficient to refute the correlationism of the sort advocated by Dehaene and his colleagues and to answer the question posed at the beginning of this article. Further, this theoretical picture leads to predictions. One prediction is that in the Sperling, Landman et al., and Sligte et al. experiments the representations of the unaccessed items will prove to involve recurrent loops. Another upshot is that if the activations of the fusiform face area mentioned earlier in the patient G.K. turn out to be recurrent activations, we would have evidence for phenomenal experience that the subject not only does not know about, but in these circumstances cannot know about. The fact that the fusiform face activations produced 498 in G.K. by the faces he says he doesn’t see are almost as strong as the activations corresponding to faces he does see, suggests that top-down amplification is not necessary to achieve strong activations. The mesh argument suggests that workspace activation is not a constitutive part of phenomenology. And given that actual workspace activation is not a constitutive part of phenomenology, it is hard to see how anyone could argue that potential workspace activation is a constitutive part. Further, as noted a few paragraphs back, it is doubtful that potential workspace activation is even causally necessary to phenomenology. 15. Conclusion If we want to find out about the phenomenological status of representations inside a Fodorian module, we should f ind the neural basis of phenomenology in clear cases and apply it to neural realizers inside Fodorian modules. But that requires already having decided whether the machinery of access should be included in the neural kinds in clear cases, so it seems that the inquiry leads in a circle. This target article has been about breaking out of that circle. The abstract idea of the solution is that all the questions have to be answered simultaneously, tentative answers to some informing answers to others. The key empirical move in this article has been to give meshing answers to psychological and neural considerations about overflow. ACKNOWLEDGMENTS Earlier versions of this paper were presented at the following venues: the 2005 NYUColloquium in Mind and Language; Georgia State University; MIT; Harvard Medical School; Dan Pollen’s course at Harvard College; Dartmouth College; NYU Medical Center; Aarhus University, Denmark; Stockholm University, Sweden; Go¨teborg University, Sweden; Umea ˚ University, Sweden; Indiana University, Brown University; the Pacific APA 2006; University of Western Ontario; Cal State, LA; UCLA; the Association for the Scientific Study of Consciousness; the Institut Jean Nicod; the University of Toronto; and at Katalin Balog’s class at Yale. I am grateful to Thomas Nagel for his role as chief inquisitor when this material was defended at the NYU Mind and Language Colloquium and for some of his suggestions; and I am grateful to Katalin Balog, Alex Byrne, Tyler Burge, Susan Carey, Max Coltheart, Leonard Katz, Sean Kelly, Victor Lamme, Mohan Matthen, Alva Noe ¨, Dan Pollen, David Rosenthal and Susanna Siegel for comments on a previous draft. NOTES 1. See Siegel (2006a; 2006b) for discussion of what kind of thing the content of a phenomenal state is. 2. Phenomenal persistence and persistence of accessible information should be distinguished from what Koch (2004, Ch. 9) calls gist perception. We have specialized detectors for certain kinds of scenes, and in learning to read, we develop similar detectors for words. These detections can take place in 100 msecs and seem to require no attention. (See Potter 1993; Rousselet et al. 2002.) 3. The inattentional blindness view can be found in Rensink (1997), Simons (1997), Noe¨ (2004), and O’Regan & Noe¨ (2001). Views more closely related to the inattentional inaccessiblity view can be found in articles by philosophers– (e.g., Block 2001; Cohen 2002; Dretske 2004)– and by psychologists– (e.g., Simons & Rensink 2005a; 2005b; Wolfe 1999). 4. The “cerebral celebrity” view of consciousness is not the view in Dennett’s Consciousness Explained (Dennett 1991), BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Commentary/Block: Consciousness, accessibility, and the mesh but was introduced a few years after that, I think first in Dennett (1993). I argued for a distinct notion of “access-consciousness” in Block (1990; 1992). 5. Iusedthis terminBlock(2001), but I discovered years later that Nigel Thomas published pretty much the same idea first, deriving it from Marvin Minsky’s “Immanence Illusion.” Minsky’s (1986, sect. 15.5) Immanence Illusion is this: “Whenever you can answer a question without a noticeable delay, it seems as though that answer were already active in your mind.” At least in what I have read, Minsky does not focus on the idea that potential phenomenology is supposed to be confused with actual phenomenology. Thomas does focus on phenomenology, arguing for a view similar to that of O’Regan and Noe¨ mentioned earlier: The seeming immediate presence of the visual world to consciousness does not arise because we have built a detailed internal representation of it, rather it is (like the ever shining fridge light) a product of the “immanence illusion” (Minsky 1986). For the most part, the visual perceptual instruments ask and answer their questions so quickly and effortlessly that it seems as though all the answers are already, and contemporaneously, in our minds. (Thomas 1999, p. 219) 6. Noe¨ (2002; 2004) suggests an even more pervasive form of such an illusion– that all experience is a matter of potentiality, but precisely because it is so pervasive, he does not regard the view as one that postulates an illusion. (See Cohen 2002.) 7. The first classical “visual” cortical area is V1; later classic “visual” areas include V2, V3, V4, and V5. The latter, V5, has two names, “MT” and “V5” because it was identified and named by two groups. I put “visual” in scare quotes because there is some debate as to whether some of the classic “visual” areas are best thought of as multimodal and spatial rather than as visual per se. The motion area I am talking about in the text is actually a complex including MT/V5 and surrounding areas, and is often referred to as hMTþ. (See Kriegeskorte et al. 2003). 8. TMS delivers an electromagnetic jolt to brain areas when placed appropriately on the scalp. The effect is to disrupt organized signals but also to create a signal in a quiescent area. Thus TMScanbothdisrupt moving afterimages and create phosphenes. A comparison is to hitting a radio: the static caused might interrupt good reception going on but also cause a noise when there is no reception. (I am indebted here to Nancy Kanwisher and Vincent Walsh, personal communication.) 9. To experiences phosphenes for yourself, close your eyes and exert pressure on your eye from the side with your finger. Or if you prefer not to put your eyeball at risk, look at the following website for an artist’s rendition: http://www.reflectingskin.net/phosphenes.html 10. TMS stimulation directed to V1 may also stimulate V2 (Pollen 2003). Perhaps V2 or other lower visual areas can substitute for V1 as the lower site in a recurrent loop. Blindsight patients who have had blindsight for many years can acquire some kinds of vision in their blind fields despite lacking V1 for those areas. One subject describes his experience as like a black thing moving on a black background (Zeki & ffytche 1998). Afterimages in the blind field have been reported (Weiskrantz et al. 2002). Stoerig (2001) notes that blindsight patients are subject to visual hallucinations in their blind fields even immediately after the surgery removing parts of V1; however, this may be due to a high level of excitation that spreads to other higher cortical areas that have their own feedback loops to other areas of V1 or to other areas of early vision such as V2. (See also Pollen 1999.) 11. In a different paradigm (de Fockert et al. 2001), working memory load can increase the processing of distractors. Open Peer Commentary Access to phenomenality: A necessary condition of phenomenality? doi: 10.1017/S0140525X07002798 Katalin Balog Department of Philosophy, Yale University, New Haven, CT 06520-8306. katalin.balog@yale.edu http://pantheon.yale.edu/~kb237 Abstract: Block argues that relevant data in psychology and neuroscience show that access consciousness is not constitutively necessary for phenomenality. However, a phenomenal state can be access conscious in two radically different ways. Its content can be access conscious, or its phenomenality can be access conscious. I argue that while Block’s thesis is right when it is formulated in terms of the first notion of access consciousness, there is an alternative hypothesis about the relationship between phenomenality and access in terms of the second notion that is not touched by Block’s argument. Ned Block (in Block 1990; 1995b; 2002; and the present target article) has made a conceptual distinction between the phenomenality of a mental state (a.k.a. its phenomenal character or the quale it instantiates) and access consciousness of the same state. There is a view– Block calls it “epistemic correlationism”according to which the metaphysical relationship between these two is not scientifically tractable. While cognitive accessibility is intrinsic to our knowledge of phenomenology, it might not be constitutive of the phenomenal facts themselves. According to the epistemic correlationist, there is no possible empirical evidence that could tell us one way or another. This view is Block’s main target. His thesis is that the issue of the relationship between phenomenal and access consciousness is an empirical one; and that moreover, the issue is approachable by the same empirical methods we employ in science in general. Block’s aim is to show that by looking at the relevant data, and employing the method ofinference to the best explanation, we can mount an argument for the specific thesis that access consciousness is not constitutively necessary for phenomenality. If this is so, he has given reasons to reject “metaphysical correlationism” as wellthat is, the view that the cognitive access relations that underlie reportability are constitutive of phenomenology. Block’s thesis needs further clarification. A phenomenal state can be access conscious in two radically different ways. Its content (or part of its content) can be access conscious, or its phenomenal character can be access conscious. Say, I am having a visual experience of a red circle in an orange background. In this case, both the content and the phenomenal character of this experience can be access conscious. I can be aware that I am seeing a red circle in an orange background, and I can also be aware that my experience has such and such a phenomenal character. However, in the experiments that Block discusses in his article, the two kinds of access come apart. As a result, although Block’s thesis holds if understood as involving the first notion of access, there is an alternative thesis involving the second sense of access that is untouched by Block’s arguments. After clarifying Block’s thesis, I will briefly sketch this alternative hypothesis. Consider the following kind of experiment, which provides crucial support for Block’s thesis. Following Sperling’s (1960) famous experiments, Landman et al. (2003) showed subjects eight rectangles in different orientations for half a second. The resulting experience e has a– presumably non-conceptual (pictorial or iconic)– representational content r, and, according to the introspective reports of subjects, a phenomenal character p.1 Given the model of access consciousness assumed in Block’s paper as broadcasting of conceptual representations 499 BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Commentary/Block: Consciousness, accessibility, and the mesh in a global workspace (Baars 1988; 1997), Block takes e to be access conscious if and only if conceptual representations of e’s content are present in the global workspace. In other words, e is access conscious if and only if there are conceptual representations in the global workspace that extract the content of e (e.g., “There were rectangles of the following orientations…”). That typically we are not access conscious, in the sense described above, of all aspects of a phenomenal experience’s content is convincingly shown by the Landman et al. (2003) experiments. After the brief exposure, subjects are only able to report on the precise orientation of up to four of these rectangles. These experiments show, to my mind conclusively, that access consciousness of this sort– that is, the existence of conceptual representations in global workspace that extract all the relevant content of e– is not constitutively necessary for the phenomenality of the experience. This finding is further supported by the neurophysiological data Block cites, which show the neural implementation of sensory representations and the neural implementation of global access to be physically separate and independent from each other. However, these experiments– which comprise the bulk of Block’s supporting evidence– do not show that no access is constitutively necessary for phenomenality. Notice that the aforementioned interpretation of these experiments crucially relies on the subjects’ introspective report of the phenomenality of their entire visual experience, including those aspects of the experience whose content is not access conscious. Introspective access to the phenomenality of the entire experience was part of the evidence in the Sperling and the Landman et al. experiments for why access to the conceptualized content of the experience is not necessary for phenomenality. But these data leave room open for the hypothesis that access to the phenomenality of the experience is constitutively necessary for that phenomenality. How exactly should we think about access to the phenomenality of the experience if it is not access to its conceptualized content? Notice that the representations in the global workspace that are not constitutively necessary for phenomenality are separate from the representations whose phenomenality is in question. Phenomenal experience quite plausibly involves non-conceptual representation; representations that enter the global workspace, on the other hand, are conceptual representations. There are different representations involved. What about access to the phenomenality of the experience itself? It seems plausible that the relationship between phenomenality and the representation of it that is in the global workspace is more intimate. Here is an idea: Perhaps phenomenality requires that a conceptual representation of the phenomenal character of the experience, more precisely, a judgment to the effect that the relevant phenomenal experience occurs, itself is in the global workspace. Plausibly, this would not involve any old conceptual representation of the phenomenality of the experience, but a phenomenal representation involving phenomenal concepts. There is a plausible account of phenomenal concepts, the constitutional account (see, e.g., Papineau 2002), according to which phenomenal concepts– in their canonical, first person, present tense applications relevant to these experiments– are partly constituted by the experience they refer to. That is, the first-person, present-tense judgment that e has phenomenal character p is partly constituted by e itself. Notice that here the experience whose phenomenality is at issue and the state in the global workspace that constitutes access to it are not separate and independent. The conceptual representation in the global workspace involves e itself and this adds to the plausibility of the idea that this kind of access is intrinsic to phenomenality. Unlike the thesis Block is criticizing (let’s call it the Accessc thesis), this thesis (let’s call it the Accessp thesis) seems to be a viable hypothesis. None of the data discussed by Block rule it out, or even make it implausible. But if the Accessp thesis is true, then some interesting consequences follow– for example, 500 that despite suggestions to the contrary by Block, activations in the “fusiform face area” of “visuo-spatial extinction” patients, or any other early visual state that is not accessp conscious, could not be phenomenal. ACKNOWLEDGMENT I would like to thank Georges Rey for useful suggestions. NOTE 1. I want to sidestep the issue of representationalism about qualia here. All I assume is that e has r and p; I won’t discuss the relationship between r and p. Psychology supports independence of phenomenal consciousness doi: 10.1017/S0140525X07002804 Tyler Burge Department of Philosophy, University of California, Los Angeles, Los Angeles, CA 90024. burge@ucla.edu Abstract: Inference-to-best-explanation from psychological evidence supports the view that phenomenal consciousness in perceptual exposures occurs before limited aspects of that consciousness are retained in working memory. Independently of specific neurological theory, psychological considerations indicate that machinery producing phenomenal consciousness is independent of machinery producing working memory, hence independent of access to higher cognitive capacities. Ned Block argues, “the machinery of cognitive access is not included in the machinery of phenomenology” (target article, sect. 9, para. 13). His argument is plausible, but I think psychological considerations support his conclusion, independently of neurological conjecture. The view that all consciousness must be available to higher cognitive faculties is motivated by worry that without “reportability,” consciousness cannot be studied scientifically. Either the view tries to rule apriori on empirical matters– how could it be apriori that animals that lack propositional attitudes cannot feel pain?– or it envisions too narrow a range of possible empirical evidence. Block has widened the range. Here, I think he underplays psychological considerations. In the Sperling (1960), Landman et al. (2003), and Sligte et al. (2008) experiments, subjects, using iconic memory, take themselves to have seen a relatively specific number of items, experimentally investigated to have been in the 8 to 32 range. The items are displayed long enough for normal perceptual processing to be completed. In any given trial, using working memory, subjects can make use of information on only four instances of specific types of items, say, specific alphanumeric characters. They can do this for specific types of any 4of8to 32, if cued after presentation. I believe these experiments support two conclusions: (a) In any given trial, there are phenomenally conscious perceptions of specific types of items not accessed by working memory; and (b) causal machinery produces specific types of phenomenal consciousness on given occasions, but on those occasions does not register those types in working memory. Blockaccepts, butdoesnothighlight, conclusions (a) and (b). He argues for a stronger conclusion: (c) The machinery of working memory does not overlap the machinery of phenomenal consciousness. Hereaches this third conclusion in three steps. First, he holds that if one accepts the Sperling-type experiments at face value, the minimum concessions required of someone who believes working memory(and throughit, “reportability”) is constitutive to phenomenal consciousness are: (d) “the ‘capacity’ of … the visual BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Commentary/Block: Consciousness, accessibility, and the mesh phenomenal memory system, is greater than that of the working memory buffer that governs reporting” (sect. 9, para. 11), and (e) “the machinery of phenomenology is at least somewhat different from the machinery of cognitive accessibility” (sect. 9, para. 13). (That is, (a)-(b) entails (d)-(e), but not vice-versa.) Second, he argues for accepting Sperling-type experiments at face value. Third, he invokes neurological conjecture to support (c). One might accept (a)-(b) and (d)-(e), but insist that working memory and “reportability” are constitutive to phenomenal consciousness. One might hold that although some specific phenomenally conscious items do not appear in working memory, all phenomenal consciousness depends constitutively on some items’ being accessible to working memory. Block marshals neurological considerations for (c) against such a position. In his second stage, Block opposes Dehaene’s attempts to avoid taking Sperling-type experiments at face value. Block effectively criticizes postulating what he calls a refrigerator-light illusion, and points out that it is question-begging to invoke “change-blindness” to support the position that the subjects in Sperling-type experiments are under an illusion that they had phenomenal experiences of items that do not appear in working memory. The two cases are disanalogous in a way that Block does not note. On Dehaene’s view that Sperling-type subjects are phenomenally conscious only of items actually in working memory, the subjects cannot have had, before the cue that selects those items retained in working memory, a phenomenologically conscious perception of any of the specific 8–32 items that they seem to have experienced. On that view, subjects’ sense of having consciously perceived even specific retained items before they appear in working memory is illusion. No one postulates analogous total illusion in “change-blindness” cases. Even proponents of the (I think mistaken) view that items that change unnoticed are not consciously seen do not claim that nothing is consciously seen. I believe that Sperling-type experiments support (c), not just (a), (b), (d), (e). I argue by dilemma. If retained and unretained items are held not to be conscious before any items are retained in working memory, what is the evidence that memory of their having been conscious is total illusion? Exposure is long enough for perceptual processing to be complete. Why should phenomenology, even of specific retained items, have been missing? We have independent evidence about working memory. It does have constructive functions: making consciousness more vivid, rehearsing to facilitate retention and reproduction of imagery (Andrade 2001; Pearson 2001). But its primary function is to preserve perception already formed. Holding that its preservations convey systematic illusion is ad hoc. The fact that subjects seem to remember having seen all items, and could be cued to retain any item specifically, supports believing that even specific unretained items are phenomenally conscious. Now suppose that all, or at least the retained, specific items are held to be conscious before being preserved in working memory. What is the evidence that mere accessibility to working memory is constitutive to their being occurrently conscious beforehand? Such a view labors under heavy empirical burden. Consciousness is an occurrent, not a dispositional, condition. We have no good idea how meredispositional accessibility to working memory could be causally necessary to occurrence of consciousness before working memory operates. Why should the door’s being open matter to the occurrence of something that does not use the door until after it already occurs? Such a view would require very special evidence and explanation. In the absence of specific empirical support, the idea is not a serious contender. The best explanation of current evidence is that conscious perception of the specifics of items later retained, indeed of all 8–32 items, occurs independently of working memory. The machinery of phenomenal consciousness appears to be independent of the machinery of working memory. Conclusion (c) is supported independently of Block’s neurological conjecture. Further evidence for (c) may lie in the formation speed of at least generic phenomenally conscious aspects of visual perception. Some super-ordinate object categorization occurs in less than 150 msec– before a signal even reaches working memory (VanRullen & Thorpe 2001; Rousselet et al. 2004a; 2004b). Such considerations are tentative. But it is important not to be so fixed on neurological matters that one underrates the force of psychological considerations in supporting psychological conclusions. Do we see more than we can access? doi: 10.1017/S0140525X07002816 Alex Byrnea, David R. Hilbertb, and Susanna Siegelc aDepartment of Linguistics and Philosophy, Massachusetts Institute of Technology, Cambridge, MA 02319; bDepartment of Philosophy, University of Illinois at Chicago, MC 267, Chicago, IL 60607; cDepartment of Philosophy, Harvard University, Emerson Hall, Cambridge, MA 02138. abyrne@mit.edu hilbert@uic.edu http://web.mit.edu/abyrne/www www.uic.edu/~hilbert ssiegel@fas.harvard.edu http://www.people.fas.harvard.edu/~ssiegel Abstract: One of Block’s conclusions, motivated by partial-report superiority experiments, is that there is phenomenally conscious information that is not cognitively accessible. We argue that this conclusion is not supported by the data. Block’s overall argument appeals to the lemma that “in a certain sense phenomenal consciousness overflows cognitive accessibility” (target article, Abstract), which Block takes to be supported by Landman et al. (2003) and Sligte et al. (2008). (For reasons of space we will ignore the latter.) Block summarizes his discussion of these two papers as follows: ThemainupshotoftheLandmanetal.andtheSligteetal. experiments (at least on the surface– debunking explanations will be considered later) is along the same lines as that of the Sperling experiment: The subject has persisting experiences as of more specific shapes than can be brought under the concepts required to report or compare those specific shapes with others. (sect. 9, para. 10) Thus, in the first condition of the Landman et al. experiment, Block holds that the subjects have persisting experiences as of [a circle of] eight rectangles, with the horizontal/vertical orientation of each rectangle specified. And if that is so, then, as Block says, the subject’s experiences are not completely accessible, because the subjects can report the orientation of only four (or so) rectangles. Although most of Block’s discussion is couched in terms of “phenomenal consciousness” and the like, for present purposes we can talk instead (as Block himself sometimes does) of what the subjects see. Put this way, Block’s claim is that the subjects continue to see each rectangle as oriented horizontally or vertically after the stimulus has been replaced with a gray screen. In the terminology of Coltheart (1980), this is an example of visible persistence. Coltheart distinguishes visible persistence from informational persistence. The latter is defined not in terms of seeing, or phenomenal consciousness, but in terms of the persistence of rich visual information about a stimulus after it has been replaced. Sperling-type experiments show that stimulus information is held in a high-capacity but transient memory, and thus that there is informational persistence. One might hold that there is informational persistence simply because there is visible persistence; that is, stimulus information continues to be available because the subject continues to see the stimulus. Coltheart argues, however, that the phenomena are not connected so intimately. One consideration is that informational persistence lasts longer than a few hundred ms, the duration of visible persistence. (As Block notes, the duration of informational persistence found by Landman et al. is about 1,500 msecs.)1 501 BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Commentary/Block: Consciousness, accessibility, and the mesh With this distinction in hand, consider Landman et al.’s conclusion: The present data agree with the presence of two parallel types of short term memory. …Almost all items enter the first type of memory. It is like iconic memory, becauseit has a high capacity and it is maskable. … The second type of memory is one that resists interference by new stimuli. When new items enter the visual system, they replace the old items, except the ones that have entered the second type of representation. … The cue-advantage arises because the subjects selectively transfer the cued item from iconic memory to the more durable working memory. …(Landman et al. 2003, p.162) Landman et al. are, then, concerned with informational persistence, not visible persistence. Their paper contains no data concerning visible persistence. Since informational persistence is consistent with no visible persistence at all, Block’s appeal to Landman et al. must be somewhat indirect. And indeed it is. Block’s argument for visible persistence is based on subjects’ reports: “[subjects say they are] continuing to maintain a visual representation of the whole array” (sect. 9, para. 6). We have three points about this. First, Block needs only the weaker claim that the subjects in the Landman et al. experiment saw each rectangle as oriented horizontally or vertically, not the stronger claim that the subjects remain in this state after the stimulus has been replaced. The weaker claim implies Block’s conclusion about inaccessibility for the same reason that the stronger one does. We do not dispute that information about the orientation of each rectangle persists and is not as a whole accessible; we do dispute Block’s claim that this inaccessible information characterizes what the subjects see. Our second point is that it is unclear that subjects’ reports unequivocally support Block. Block needs subjects to agree that they saw each rectangle as oriented horizontally or vertically (even if they can’t report which orientation each rectangle has). More precisely: for each rectangle x, either they saw x as horizontal, or they saw x as vertical. If the subjects merely say that they saw eight rectangles, some horizontal and some vertical, or that “they can see all or almost all the 8 to 12 items in the presented arrays” (sect. 9, para. 11), this is insufficient. According to Landman et al., selected stimulus information is transferred from the transient iconic memory to the more durable working memory. Working memory therefore contains less information about the stimulus than iconic memory. If that is all that working memory contains, and if working memory governs subjects’ reports about what they see (as Block supposes), then subjects would simply say that they saw a circle of rectangles and saw some of them as oriented horizontally/vertically. They would not, then, agree that they saw details, some of which they can’t report. So our third point is this: Block must deny that the contents of working memory are simply a subset of the contents of iconic memory, which is to go beyond the results of Landman et al. If Block is right and subjects report (correctly) that they saw each rectangle as oriented horizontally or vertically, then the contents of working memory should include, not just certain information about the stimulus transferred from iconic memory, but also the meta-information that some information was not transferred. We are not saying that this proposal about the contents of working memory is wrong, but only that the Landman experiment does not address it. NOTE 1. The question of the exact relationship between visible and informational persistence remains open. Loftus and Irwin (1998) argue that the many measures of visible and informational persistence pick out the same underlying process. Nevertheless, the distinction is still useful and our discussion does not rely on the assumption that it marks a real difference. 502 Experience and agency: Slipping the mesh doi: 10.1017/S0140525X07002828 Andy Clark and Julian Kiverstein School of Philosophy, Psychology and Language Sciences, George Square, Edinburgh EH8 9JX, Scotland, United Kingdom. andy.clark@ed.ac.uk s9903600@sms.ed.ac.uk http://www.philosophy.ed.ac.uk/staff/clark.html Abstract: Can we really make sense of the idea (implied by Block’s treatment) that there can be isolated islets of experience that are not even potentially available as fodder for a creature’s conscious choices and decisions? The links between experience and the availability of information to guide conscious choice and inform reasoned action may be deeper than the considerations concerning (mere) reportability suggest. In this elegant and tightly argued target article, Ned Block seeks to persuade us that phenomenal consciousness routinely “overf lows” cognitive accessibility. By this he means that we can (and do) have experiences even in cases where we lack the kind of access that would yield some form of report that such and such an experience had occurred. The case Block makes for such an apparently hard-to-support judgment rests on a “mesh” between psychological results and work in neuroscience. The psychological data seem to show that subjects can see much more than working memory enables them to report. Thus, in the Landman et al. (2003) experiments, for instance, subjects show a capacity to identify the orientation of only four rectangles from a group of eight. Yet they typically report having seen the specific orientation of all eight rectangles. Working memory here seems toset a limit on the numberofitemsavailable for conceptualization and hence report. Workinneuroscience then suggests that unattended representations, forming parts of strong-but-still-losing clusters of activation in the back of the head, can be almost as strong as the clusters that win, are attended, and hence get to trigger the kinds of frontal activity involved in general broadcasting (broadcasting to the “global workspace”). But whereas Dehaene et al. (2006) treat the contents of such close-seconds as preconscious, because even in principle (given their de facto isolation from winning frontal coalitions) they are unreportable, Block urges us to treat them as phenomenally conscious, arguing that “the claim that they are not conscious on the sole ground of unreportability simply assumes metaphysical correlationism” (sect. 14, para. 9; italics in original). That is to say, it simply assumes what Block seeks to question– that is, that the kind of functional poise that grounds actual or potential report is part of what constitutes phenomenology. Contrary to this way of thinking, Block argues that by treating the just-losing coalitions as supporting phenomenally conscious (but in principle unreportable) experiences, we explain the psychological results in a way that meshes with the neuroscience. The argument from mesh (which is a form of inference to the best explanation) thus takes as its starting point the assertion that the only grounds we have for treating the just-losing back-of-thehead coalitions as non-conscious is the unreportability of the putative experiences. But this strikes us as false, or at least premature. For underlying the appeal to reportability is, we suspect, a deeper and perhaps more compelling access-oriented concern. It is the concern that any putative conscious experience should be the experience of an agent. The thought here is that we cannot make sense of the image of free-floating experiences, of little isolated islets of experience that are not even potentially available as fodder for a creatures rational choices and considered actions. Evans (1982) rather famously rejects the very idea of such informationally isolated islands of experience. According to Evans, an informational state may underpin a conscious experience only if it (the informational state) is in some sense input to a reasoning subject. To count as a conscious experience an informational state must: BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Commentary/Block: Consciousness, accessibility, and the mesh [serve] as the input to a thinking, concept-applying and reasoning system: so that the subject’s thoughts, plans, and deliberations are also systematically dependent on the informational properties of the input. When there is such a link we can say that the person, rather than some part of his or her brain, receives and processes the information. (Evans 1982, p.158) The real point here is (or should be) independent of Evan’s appeal to conceptualization. What matters, rather, is that the information must be available to the agent qua “reasoning subject,” where this may be unpacked in many different ways, not all of them requiring full-blown concept-use on the part of the agent (see, e.g., Bermu´dez & Macpherson 1998; Hurley 1997). Evans’ insight is that the notions of conscious experience and reasoned agency (here very broadly construed) are deeply intertwined: that there are non-negotiable links between what is given in conscious awareness and the enabled sweep of deliberate actions and choices available to a reasoning subject. Such a story opens up a different way of interpreting the Sperling (1960) and the Landmanetal.(2003)results. In these cases (we suggest) subjects report phenomenally registering all the items because information concerning each item was, at that moment, available to be deployed in the service of deliberate, reasoned, goal-directed action. Such momentary potentiality is not undermined by the (interesting and important) fact that the selection of a few items to actually play that role then precludes the selection of the rest. Contrariwise, Block argues that a subject like G.K. can be having an experience of a face and yet it be impossible for him to know anything of this experience. Block takes G.K.’s phenomenal experience to be constituted by recurrent processing in the fusiform face area. We believe that G.K. can be consciously experiencing a face only if this experience is at least momentarily poised for use in reasoning, planning, and the deliberate selection of types of action. Recurrent processing in the fusiform area will no doubt prove to be among the conditions necessary for realizing a state that plays this causal role. The contents of conscious phenomenal experience, if all this is on track, must be at least potentially available for use in the planning and selection of deliberate, stored-knowledge–exploiting, and goal-reflecting and goal-responsive, actions. Block’s justlosing coalitions fail to trigger winning frontal coalitions and hence fail to be in a position to contribute their contents in this manner to the full sweep of the agent’s deliberate acts and choices. It is this fact (rather than the more superficial indicator of unreportability) that should motivate our treating the contents of the just-losing coalitions as non-conscious. If this is correct, then the staging post for the argument from mesh is called into question. Until the considerations concerning links between experience and rational agency are more fully addressed, it remains unclear whether the kind of “fit” to which Block appeals can really favor his conclusion over our own. ACKNOWLEDGMENTS Thanks to MatthewNudds, Tillmann Vierkant, and the participants in the Edinburgh Philosophy, Psychology, and Informatics Reading Group for stimulating discussion of the target paper. This commentary was prepared thanks to support from the AHRC, under the ESF Eurocores CNCC scheme, for the CONTACT (Consciousness in Interaction) project, AH/E511139/1. Why babies are more conscious than we are doi: 10.1017/S0140525X0700283X Alison Gopnik Department of Psychology, University of California at Berkeley, Berkeley, CA 94720. gopnik@berkeley.edu ihd.berkeley.edu/gopnik.htm Abstract: Block argues for a method and a substantive thesis– that consciousness overflows accessibility. The method can help answer the question of what it is like to be a baby. Substantively, infant consciousness may be accessible in some ways but not others. But development itself can also add important methodological tools and substantive insights to the study of consciousness. Infants and young children cannot report their phenomenology. This has led some philosophers to argue that babies’ consciousness must be limited. Even if we think that babies are conscious it might seem impossible to recover the particular character of their experience. A version of Block’s abductive method can at least partially solve this problem. We can consider a wide range of functional and neural correlates of conscious experience in adults and then look to see similarities and differences in babies. In adults attention is highly correlated with vivid consciousness. This consciousness has a particular subjective quality– “the spotlight”– with a defined– “brightly lit”– focus and surrounding darkness. Both phenomenologically and functionally, attention to one event seems to inhibit consciousness of other events– as in inattentional blindness. Thereisadistinction betweenexogenousandendogenousattention. Exogenous attention is driven by information-rich external events. These events may be intrinsically salient. But exogenous attention may also be driven by subtle unexpectedness. Exogenous attention is marked by characteristic event-related potential (ERP) signatures, eye movements, decelerating heart rate, and parietal activation. Exogenous attention and vivid consciousness characteristically fade as information is obtained, a process of habituation. Endogenous attention is the sort of top-down “paying attention” that is motivated by specific goals rather than by intrinsic interest. It has been the focus of the adult literature for methodological reasons, and endogenous and exogenous attention are often not distinguished in discussions of consciousness. Frontal activation seems to be particularly important for endogenous attention. These functional features of attention correlate well with neurological patterns. In adult animals, endogenous attention leads to the release of cholinergic transmitters to some parts of the brain and inhibitory transmitters to other parts. Attention increases both theefficiency of a particular part of the brain and its plasticity; and it inhibits activation and plasticity in other brain areas. So for adults there is an elegant if undoubtedly oversimplified story about how consciousness works. In the canonical case, the goal-directed control systems in frontal cortex indicate that a particular kind of event is important. The perceptual system, guided byattention,zoomsinonjustthateventandthebrainextractsinformation about the event and modifies itself, that is, learns accordingly. Significantly, though, this whole process is highly focused; other parts of the brain may actually be shut down in the process. Vivid spotlight consciousness is the phenomenological result. Even very young infants have extensive exogenous attention capacities. When they are presented with even highly subtle and conceptually unexpected novel events, they immediately focus their gaze on these events, and show similar heart rate deceleration and ERP signatures to those of adults. Indeed, this is the basis for the habituation technique that is our principal source of information about infant’s minds. However, infants develop endogenous attention much later, and it is still developing during the preschool years. Moreover, and probably correlated with this fact, infants and young children appear to have less focused attention than older children– for example, they show better incidental memory. Theneurologysuggestsasimilarpicture.Theparietalandsensory systems involved in exogenous attention are on line at an early age. The top-down frontal regions and connections that control endogenous attention only mature later. Young animals’ brains are far more plastic than adult brains and this plasticity is much less focused and attention-dependent. Cholinergic transmitters are in place early, while inhibitory transmitters emerge only later. So again, an undoubtedly oversimplified but suggestive picture emerges. When infants and young children process information 503 BEHAVIORAL AND BRAIN SCIENCES (2007) 30:5/6Commentary/Block: Consciousness, accessibility, and the mesh there is much less top-down control and inhibition. Infants are sensitive to any information-rich stimuli, not just those stimuli that are relevant to their goals and plans. And because they have much less experience, more stimuli will be informationrich for them than for us. With less top-down inhibition their processing and plasticity will be more distributed and less focused than those of adults. These pictures also make sense from an evolutionary point of view. Childhooditself,thatparadoxicalperiod ofhelplessimmaturity, is moreprotractedinspeciesthatrelyheavily on learning– and mostprotractedinhumans.Childhoodreflectsanevolutionarydivision of labor. Children are very good at learning about everything and are not very good at doing anything– adults use what they have learned in childhood to design effective actions. AsBlocksuggests,wecanusethismeshoffunctional,neurological, and evolutionary facts to make a hypothesis about phenomenology. Babies are moreconsciousthanweare.Theyare ceaselessly and broadly engaged in the kind of information-processing and learning that adults direct only at limited, relevant events. And babies are less subject to the processes that actively cause unconsciousness in adults– inhibition and habituation. When adults are placed in a situation that is functionally similar to babies, such as traveling in a strange country or meditating in certain ways, we experience a similarly vivid but distributed phenomenologyconsciousness becomes a lantern instead of a spotlight. For babies, consciousness may also overflow accessibility. But babies make the ambiguity of “accessibility” vivid. Obviously, information in infant consciousness is not accessible for reporting– it may also not be accessible for goal-directed planning nor be the subject of the endogenous attention that accompanies such planning. But it may indeed be accessible for purposes of induction, prediction, generalization, and intuitive theory-formation– the principal psychological activities of babies. Does that constitute overflow or not? This leads to a broader methodological point, and to the greatest advantage of developmental data. Human adults are our chief source of information about consciousness. But in adults, many very different functional and neurological processes are highly correlated. Attention is focused, inhibitory, and top-down, and it leads to plasticity and learning, and it is reportable, and it is in the service of goals, and it is connected to a sense of self, and so on, and so forth. In development, this apparently unified picture breaks apart in unexpected and informative ways: Babies may have access for induction but not planning, have exogenous but not endogenous attention, and have plasticity and facilitation without inhibition. Rather than a single story of capital-c Consciousness, there may be many very varied relations between phenomenology and function.