Marmite and Metaphysics
Philosophers on Consciousness
1. Introduction: the open questions of consciousness
Have you ever tasted Marmite? I’ve not, so I don’t know what it tastes like. Query: do you think it’s possible to know what Marmite tastes like without having the experience of tasting it? You’d think that there’d be a firm consensus on this pressing issue, to the effect that to know the taste you’d actually have to taste it. But as it turns out some philosophers suppose that, given enough knowledge of human physiology, the experience wouldn’t be needed in order to know about that particular flavor. Your sensation of tasting Marmite, they argue, is an objective, physical fact, something in principle observable and knowable from the outside, not just subjectively available to you (see note 5). Moreover, some philosophers claim there’s nothing it’s fundamentally like to taste it. Flavors and other sensory qualities (“qualia” in philoso-speak), they say, are illusions.
The metaphysical status of tasting Marmite, among other experiences, is what’s at stake in Jack Symes’ entertaining and informative introduction to the study of consciousness, delivered in 12 lively interviews and essays by some of the leading thinkers in the field. At the end of each chapter we get further reflections (“Afterthoughts”) by Symes on the material just presented; questions to think about (some of which amount to an informal test of what you learned from the chapter, should you want to take it); and astute recommendations for further readings by and about the interviewee, graded by difficulty. So this is a great launching pad for anyone wanting to explore the several dimensions of the question of consciousness without getting too bogged down in technicalities.
And what could be more central to your life than your current conscious experience, your window on the world? We usually take it for granted as a direct presentation of reality, so what’s the big deal? But the first installment by Gregory Miller nicely motivates the controversies surrounding the nature and possible functions of consciousness covered in the ensuing dialogues. Symes’ deft, often humorous, touch, keeps the momentum going, and for those who need it there are frequent “info-boxes” explaining basic concepts and propositions, many of which are germane to more than just consciousness studies. What’s not to like about this book? Nothing.
Having thus recommended it, my purpose here is to convey the flavor, so to speak, of the views expressed, see where they coincide and diverge, and offer some critical reflections. Although Symes and his subjects don’t much discuss the major empirically grounded approaches to consciousness now in contention, such as Integrated Information Theory (IIT), Global Neural Workspace, or predictive processing approaches, they nevertheless demarcate some of the basic philosophical orientations to the problem. Necessarily these clash to some extent, which gives the book a bit of drama as opponents, in good philosophical fashion, sometimes accuse each other of missing the point, misrepresenting their position, or otherwise being dense or dogmatic. Fortunately, the discussants were not in a position to deliver sharp kicks to the shins, as one realist about qualia suggested might be in order (the “Thus I refute you!” response).
2. Consciousness realism
What no one does in this book is to doubt that consciousness exists, even if they might disagree on what consciousness fundamentally is and how it comes about. Can you assert that you’re not experiencing anything right now without contradicting yourself? Your conscious experience seems to be something that you can’t step outside of to doubt the existence of, which of course was Descartes starting point in his attempt to ground indubitable knowledge. Whatever consciousness is, it’s undeniably real; it’s something, not nothing. Psychologist Susan Blackmore’s favorite question about it, meant to shake you out of your complacency, is “Are you conscious now?”. If you’re in a position to answer, the answer will always be yes, even if you’re not sure about a moment ago. Realism about consciousness thus seems to be a self-evident empirical claim about what minimally exists, even if the nature of consciousness itself isn’t obvious, and the nature of the world outside one’s experience (assuming there is one) even less so. Conscious experience is square one in the epistemic game of figuring out what the world is like; arguably, we’re not in an observational epistemic (knowledge conferring) relation to it, rather we consist of it as conscious knowers. But the knowledge game of course has the nature of consciousness itself as an object – it’s presumably part of (or perhaps, some conjecture, the whole of) the world – so it looks like some sort of epistemic “strange loop” might be afoot. Stay tuned.
3. The essential features of consciousness
If the existence of consciousness is hard to doubt, what is it, precisely, as it presents itself in our lives? What are its essential features? Gregory Miller suggests we follow philosopher Thomas Nagel’s characterization: consciousness is, essentially, what it’s qualitatively like to have experiences – to undergo the feel of a sensation, or emotion, or perhaps even a thought or a hunch. This is what’s called in the trade phenomenal consciousness, phenomenality, or phenomenology, where the root word phenomenal points to the qualitative appearance of your experiences - their qualities. In her interview, Michelle Montague argues that it’s not just sensations like pain or seeing red that have qualitative character, but also things like thinking – the cognitive phenomenology of entertaining concepts – and the evaluative, emotional phenomenology of being happy or sad.
Experiences not only have qualities, they are subjective, that is, they are unshareably undergone by particular individuals. Experiences, on the face of it, seem inescapably private events, such that you can’t literally feel or see my pain or otherwise observe any of my experiences. This contrasts with the public, intersubjective observational availability of standard physical objects like my brain and body – they are objective, not subjective in the way experiences are. As philosopher Philip Goff puts it in his interview, “You can’t look inside someone’s head and see their feelings and experiences!” And philosopher Keith Frankish agrees: “…mental qualities seem to belong to a private world that is completely inaccessible to anyone else.” We can tell bees have brains by observing them, but we can’t observe their experiences, should they have any, so the question of whether bees are conscious awaits a settled account of consciousness. As a first cut, therefore, qualitativeness and subjectivity are what arguably pick out the essence of conscious experience, and are what we need to explain. It’s important to see that no metaphysical presuppositions (e.g., dualism, physicalism, idealism, panpsychism) are necessarily built into this initial characterization; it’s simply how consciousness presents itself as being. In particular, referring to the qualities present in experience (commonly called qualia by philosophers) is not necessarily to refer to anything non-physical or, for that matter, physical. We can, and should be, metaphysically agnostic when approaching consciousness.
Illusionism. Although most of those interviewed seem to accept the above characterization of consciousness more or less, two, Keith Frankish and Daniel Dennett, vigorously dispute it, saying that the phenomenal qualities that seem to make up sensory experience are illusions. What needs explanation, they say, is why we mistakenly think qualia exist. Consciousness, on their illusionist account – a brand of physicalism – is constituted by particular brain-based cognitive functions that are in principle fully objective and observable, so there’s nothing intrinsically subjective about conscious experience. As Frankish puts it:
We really do taste Marmite! What illusionists deny is that experiences involve awareness of non-physical, private mental qualities, presented like a show to some kind of inner observer. In other words, we deny that the Marmite taste is a private inner quality.
One wonders though: why would anyone pick out particular cognitive functions as being consciousness were they not accompanied by qualia, or the (purportedly false) appearance of qualia, or merely strong beliefs about qualia? It seems that phenomenality, even if illusory, is still doing the work of identifying the explanatory target, and it’s what makes explaining consciousness so challenging, what philosopher David Chalmers has called the “hard problem.” In any case, illusionism, a minority position at the moment, is itself faced with explaining why experiences so vividly seem to have private inner qualities, what Frankish calls the illusion problem.
In her interview (“The Grand Illusion”), Susan Blackmore seems to side with the illusionists, but it’s not clear if she thinks qualia don’t exist, period (what Dennett and Frankish claim) or if qualia, albeit real, aren’t what many folks take them to be – something non-physical. But one needn’t be an anti-physicalist to be a realist about qualia. Galen Strawson, a physicalist, rejects the illusionist proposal as manifest nonsense, the “silliest claim” philosophers have ever made, since he thinks that the appearance of something being qualitative just is an instance of a quality: “The seeming of an experience can’t exist unless there really is an experience. In other words, the illusion itself is already an instance of consciousness.” I happen to agree. If you admit, as does Frankish, that the taste of Marmite is real, and that it tastes like something (e.g., marvelous), then there is something it’s like to taste it; and that, my friend, is a quality (quale). Whether or not the taste quality (quale) of Marmite is physical, non-physical, or what is a further, open question, but its experiential reality shouldn’t be controversial. To be fair, illusionists vacillate between saying phenomenal qualities don’t exist (there is nothing that it’s like to feel pain) versus saying that they aren’t immaterial. Here’s Dennett, who has long denied the existence of phenomenality (see his book Consciousness Explained and his RoboMary paper), taking the latter position (underlining added):
They know what it’s like to feel pain and so do I. However, none of us knows everything about feeling pain. Most importantly, they don’t know that their conscious experience is immaterial. They can’t know that.
True. Since we don’t have a settled theory of consciousness, the physicality or lack thereof of conscious experience – phenomenality – is an open question. But it’s bootless to go around denying phenomenality exists, as illusionists such as Dennett sometimes do.
4. Does consciousness have a function?
It might seem obvious that phenomenal consciousness has a function, otherwise why would we be conscious? But it isn’t clear how the qualities of experience add causal power to the neurally instantiated behavior controlling mechanisms of the brain and body with which experience seems associated. If qualities are identical to such mechanisms, they don’t add anything, and if they aren’t identical, then you need an interactionist story about mental causation, which no one has yet come up with. That said, only Susan Blackmore embraces the conclusion that consciousness is epiphenomenal (lacks a function) since, she argues, a functionally equivalent creature without consciousness could do everything we do; therefore, evolution had no reason to select for it. That my functionally equivalent twin wouldn’t be conscious like me of course assumes that functional equivalence need not be accompanied by phenomenal equivalence, which I think is an open question to be decided, eventually, by a settled theory of consciousness. Neurophilosopher Patricia Churchland thinks functionality and behavior at our level likely entails phenomenality: “If Susan [Blackmore] can show me a person who has all of these structures functioning properly and can talk, dance, play music and paint a picture but isn’t conscious, I’d be very surprised!.” Perhaps Churchland is right, but again, we need a settled theory to know for sure. What’s not in question is that the cognitive functions accompanied by consciousness perform advantageous behavioral work that made them targets of natural selection, even if consciousness per se perhaps was not.
Philosopher David Chalmers thinks consciousness likely isn’t epiphenomenal, but in his interview only speculates that it might play some sort of role in higher cognitive processing, for instance by collapsing the wave function (I’ve not seen evidence for that): “If we can tell a story about how consciousness has a special causal role, then we can leverage evolution to help us explain why it is that we have consciousness.” A big if, depending on what consciousness turns out to be.
Biologist Massimo Pigliucci, a physicalist, is confident about it having a causal role: “Since consciousness requires a complex nervous system, and since complex nervous systems are metabolically expensive, consciousness probably evolved by natural selection in order to fulfil one or more functions.” But this equivocates between the neurally instantiated cognitive capacities that consciousness accompanies, on the one hand, and phenomenal qualities, on the other. The former were obviously selected for, but the latter need their own causal role to have been specifically selected for. If they are identical to the processes (an open question), they don’t contribute anything additional causally or functionally that evolution could have worked on, which is Blackmore’s point mentioned above.
Galen Strawson and Philip Goff, being panpsychists of different stripes, say in their chapters that phenomenology is a fundamental attribute of base reality, not essentially connected to or explained by brain activity. This suggests that on their view consciousness does not have a particular function or play a particular causal role that was selected for in evolution, e.g., in cognition or behavior control, since it was already on the scene at the micro-physical level.
5. Explaining consciousness: metaphysical options
Whether or not phenomenal consciousness adds causal power to its neural correlates, the big question is how to explain the fact that it exists at all. Strawson and Goff, as noted above, are panpsychists who posit that qualitative phenomenology is a basic attribute of, or the essential nature of, what physics can only describe in structural terms. Since physics is silent on intrinsic natures, why not suppose that something qualitative is what ultimately constitutes that which is structured? Strawson, a physicalist and a naturalist (as opposed to supernaturalist), holds that the ultimate nature of the phenomenal is physical: “Since we know that consciousness exists, the naturalist must take consciousness to be an entirely physical and natural phenomenon.” Since Strawson sees no possible explanation of consciousness as somehow emerging from non-conscious matter, he concludes that the widespread assumption that matter, “the basic stuff of the universe,” is fundamentally non-conscious must be mistaken. But how do we know for sure that no emergence-based explanation of phenomenality will be forthcoming? And why, as naturalists, should we assume the physicality of consciousness? Where is it written that whatever’s natural must be physical or that science is necessarily restricted to the study of physical phenomena? It seems to me that Strawson’s panpsychist claims, philosophical as they are (certainly not empirical), rest on shaky premises.
For Goff, a self-described anti-materialist, panpsychism is attractive because it sidesteps the hard problem of explaining why certain brain processes entail the existence of qualia, for which (like Strawson) he thinks no solution is possible, and because, if true, it answers the cosmic question of the nature of ultimate reality. But neither of these are good empirical grounds for endorsing panpsychism. There’s no observational evidence that phenomenality is basic or intrinsic to the world in its fundamental constituents, whereas there’s lots of observational evidence it supervenes (depends on) certain sorts of brain processes. So, if we’re looking to responsibly naturalize consciousness – bring it within the ambit of observational science, and not merely armchair speculation – then panpsychism isn’t a good bet, at least not until some evidence accrues for micro-phenomenality. That brain-based explanations of phenomenality are currently lacking is not a good reason to endorse panpsychism.
The most science-oriented discussants in this book (Churchland, Pigliucci, Dennett) are inclined to physicalism as the underlying metaphysics of consciousness; not surprising, since neuroscience and cognitive science have adduced considerable evidence for the hypothesis that consciousness depends on the brain, an uncontroversially physical system. Philosopher Frank Jackson (a former non-physicalist) joins them, counseling in his chapter that we should resist the conclusion that there are non-physical properties revealed by having experiences. Qualia, e.g., the sensation seeing red, feeling pain, or tasting Marmite, are, he says, simply informational states that tell us about physical properties of the world (respectively: light reflectance profiles of objects, bodily damage, the chemical composition of Marmite). That qualia are informative of (represent) the world seems right, but as Jackson acknowledges, they are not direct presentations of the physical states of affairs, but corresponding intermediaries. As tokens in our conscious representational economy, they have their own qualitative character that is simply arbitrary with respect to what in the external world, including the body, they represent. As Jackson would admit, if you’ve not tasted Marmite, you won’t know what it’s like to taste it. Whether the taste of Marmite, as distinct from its chemical composition, just is a physical property, and not, say, a representational property, is an open question. Perhaps a better understanding of the neural basis of qualia will show them to be identical to that basis, and thus straightforwardly physical, but I don’t think that Jackson has yet closed the explanatory gap in favor of physicalism, at least not in this necessarily cursory presentation.
Another sanguine physicalist is biologist Massimo Pigliucci, who, impatient with philosophy, advises us to stick with physical science as the explanatory default:
There is no hard problem [of consciousness] distinct from the (really hard) one of arriving at a neurobiological causal explanation of consciousness. Ultimately, Chalmers runs into the same problem as Jackson: he wants to do first philosophy, and first philosophy has firmly been replaced by science, so it’s a non-starter.
Like Pigliucci, I’m unsympathetic with scientifically uninformed hypotheses about consciousness, but the problem with seeking a causal explanation is that there’s no obvious candidate for something produced or caused by neural processing that’s qualitative and subjective. Any physicalist account of consciousness has to end up in an identity of phenomenality with certain neural or otherwise physical, objective events, not something those events produce or cause as a further effect. Or if there is a further effect that’s identical to phenomenality, that too would have to be a neural or physical event, according to physicalism. And it’s that identity claim that has to be proven, not assumed, when it comes to a theory of consciousness. Physicalists tendentiously tend to assume it, when in fact there are other possible mind-brain relations besides identity and/or causation, such as the relation between neural representational vehicles and their representational content, a matter of ongoing philo-scientific research.
A variety of idealism, more or less the metaphysical opposite of physicalism, is championed in the last essay by Jack Symes and Miri Albahari:
Panpsychist idealism says that there is no mind-independent material universe. It doesn’t exist. Everything physical is, in fact, mental… The world is grounded in consciousness and its qualities. Therefore, panpsychist idealism eliminates the problem of figuring out how conscious subjects can fit into a mind-independent world; the solution is that the world isn’t mind-independent, but mind-dependent.
Consciousness, on this view, is a pure, non-perspectival, content-free conscious field out of which our individual centers of conscious experience arise. In a thought experiment reminiscent of a session in a sensory deprivation tank, they imagine having all sensations and thoughts subtracted from your conscious experience, bit by bit, and claim that at the end all that’s left would be pure consciousness. Perhaps, but there’s good evidence that whatever consciousness remains, whether pure or still contentful, is still a function of the brain, possibly the ascending reticular arousal system (ARAS), as suggested by Thomas Metzinger in recent papers. And, as far as we know, sensory subtraction itself is a matter of taking various brain systems offline. Like other varieties of panpsychism, the idealist brand suffers from a lack of the empirical support that’s the hallmark of brain-based, materialist accounts in which the mental clearly depends – somehow – on the neural. That we don’t have the dependency relation figured out yet (the hard problem) is not, for the naturalist at any rate, a good reason to claim that ultimate reality – “the very fabric of the world” – is a field of pure consciousness.
6. Wrapping up: the hard problem persists
Symes says in his afterthoughts of the last chapter that perhaps we have to choose between consciousness and the external material world, echoing the pessimism many have about solving the hard problem: explaining how experience arises from non-conscious matter. Such pessimism clearly motivates some approaches to consciousness. Physicalist illusionists (Dennett, Frankish, perhaps Blackmore) hope to have eliminated phenomenality as the explanatory target, while idealists and other brands of panpsychists say it’s fundamental to reality; these views conveniently close the explanatory gap by eliminating opposite sides of it. Physicalists who seem to be realists about qualia (Churchland, Pigliucci, Jackson) have the much tougher job of showing how what’s qualitative and subjective is actually fully physical and objective. Those more on the fence about metaphysics, notably Chalmers and perhaps Blackmore, are non-committal about what the end game might be, thus more open to diverse explanatory options. Chalmers opines that “normal” science may not get us there; a cognitive revolution of some sort is therefore in order. Of course, we can’t know in advance what that might be.
Putting metaphysics aside, which I recommend we do in explaining consciousness, the hard problem for naturalists (as opposed to supernaturalists) is simply (ha!) to come up with a convincing, empirically-consistent-with-science account of how phenomenal consciousness fits into the world as a natural phenomenon. Contrary to Symes' assertion in an info box, naturalism isn’t necessarily equivalent to physicalism; our best empirically informed models of reality may end up including phenomena that are not obviously physical, but perhaps best understood as representational. What I found mostly missing in this volume (Jackson and Frankish come closest to mentioning it), is the thesis that consciousness is a basically representational phenomenon in which qualia are representational contents carried by neural representational vehicles. If representational contents, as the terms in which reality appears to us, are in general not found in the world they represent (we only find the vehicles), then this might help explain both the privacy and qualitative nature of conscious experience. It’s a natural but possibly representational – so perhaps not directly physical – phenomenon. But whichever way the explanatory cookie crumbles, and whichever metaphysics (if any) wins out, Symes’ book is an eminently enjoyable introduction to some of the explanatory options on hand, with a great selection of additional resources for further exploration. Although everyone in this book is probably wrong about something, you won’t go wrong in reading it.
- Tom Clark, 4/14/22
 On this point see Killing the Observer, and in this volume Susan Blackmore, Keith Frankish, and Daniel Dennett deny the existence of an inner self to which experience is presented.
 “Phenomenology” also refers to “the descriptive study of conscious experience, its structure, its flow, and its dynamics.” (Ramstead et al., 2022, emphasis added) See the first section of Ramstead et al. for more on this meaning of phenomenology.
 Miri Albahari and Jack Symes write about the categorical subjective privacy of consciousness in the last essay: “Conscious subjects like humans, penguins and kangaroos are deeply private entities; their experiences are sealed off from the rest of the world. You might see me wince with pain when I step on a sharp piece of glass, but you can’t feel the pain that I feel. Only I can feel that pain from the vantage point of my own unique perspective.”
 Symes sometimes leaves the impression that qualia are necessarily non-physical, as when he says in response to the illusionists: “If consciousness is just a trick, then we don’t need to believe in non-physical properties.”
 Note that if there are no private inner qualities to experience, there is nothing categorically subjective about it either. This means that you wouldn’t have to taste Marmite to know about its taste since it’s a fully physical phenomenon (e.g., certain neural processes), in principle available to observation. Here's Daniel Dennett in What RoboMary Knows saying (in the context of color sensation) that yes, it's possible to know what Marmite tastes like without having the experience: "I was saying that Mary [the color-deprived neuroscientist] had figured out, using her vast knowledge of color science, exactly what it would be like for her to see something red, something yellow, something blue in advance of having those experiences" (original emphasis). Note that Dennett is saying here that there is something it’s like that can be known without undergoing the experience, but it’s not private. But to be consistent with his illusionist denial of phenomenality, he’d have to say that there’s nothing qualitative about what it’s like to taste Marmite either.
 It isn’t clear to me why consciousness is so often thought to be obviously non-physical; a question for another day.
 Note that there’s no reason to suppose a being physically identical to Susan Blackmore at the microphysical level would lack consciousness, that is, be zombie Susan.
 Naturalists needn’t be physicalists about consciousness. Chalmers, for example, is a naturalist, but at the very least is agnostic about physicalism.
 For example, a mature science of representation might find that representational content, as distinct from its representational vehicles, is not perspicuously understood as a straightforwardly physical phenomenon.
 Other difficulties with panpsychism are raised in the last essay by Miri Albahari and Jack Symes, including the “combination problem”: how do bits of micro-phenomenality combine to produce your conscious experience, or mine?
 Jackson says: “…something’s looking yellow cannot tell us, in and of itself, about wavelengths and those processes in our eyes [involved in having the experience of yellow].”
 About Mary, the color-deprived neuroscientist, Jackson says “Of course, while in the black and white room, there is a sense in which she won’t know what it’s like to see something as red.” This contrasts with Dennett’s illusionist view, see note 5 above.
 Jackson writes: “Now we can see that when Mary is in her black and white room, she knows all of the properties there are to know. I’m sorry to disappoint, but Mary doesn’t learn about a new type of non-physical property when she escapes into the outside world. When we explain colour in terms of reflectance profiles, we see that the original thought experiment shouldn’t tempt us into rejecting physicalism." But he also says: "Of course, while in the black and white room, there is a sense in which she won’t know what it’s like to see something as red. What she can know, however, is that having this experience delivers transparently information about relationships between surfaces, but she won’t be able to recognize being in the state herself until she finds herself experiencing it. But that’s a point about recognitional capacities, not a point about the ignorance of non-physical properties."
So Jackson is denying that the experiences of seeing red or tasting Marmite involve non-physical properties while also asserting we can't know (recognize, remember) the "what it's like" of the experiences until we have them. So apparently, according to Jackson, there are physical properties - "what it's like" properties like the taste of Marmite - that are only known by instantiating them (unlike Dennett, see note 5). If so, then experiences on Jackson’s physicalism are constituted by a class of physical properties that are not intersubjectively accessible, unlike most physical properties.
 See for instance Representation in Cognitive Science, an open access 2018 prize-winning book by Nicholas Shea.
 “Minimal phenomenal experience: meditation, tonic alertness, and the phenomenology of ‘pure’ consciousness,” at https://www.philosophie.fb05.uni-mainz.de/files/2020/03/Metzinger_MPE1_PMS_2020.pdf and "The ARAS model theory: steps toward a minimal model of conscious experience as such," working paper at https://mindrxiv.org/5wyg7/.
 About this see Locating consciousness: why experience can’t be objectified, sections 6.3, 7, and 8.