3.1 Occam's razor, it might rightly be countered, is a mere methodological precept, not a law of nature. Yet Nature seems to be remarkably lazy. She has very few principles at all [perhaps even only a single principle? : the ubiquitous, scale-invariant conservation of a condition analogous to zero???]. Such simplicity of principle apparently suffices to generate the extraordinarily rich structures and phenomenal complexity we're stuck with today. A non-interactionist dualism, on the other hand, is [epistemically] feasible but still messy and ad hoc. It defeats the spirit of Wheeler's surmise that "...we will grasp the central idea of it all as so simple, so beautiful, so compelling that we will all say to each other, 'Oh! How can it have been otherwise? How could we all have been so blind for so long?'"

3.2 Panpsychism, it should be stressed, isn't a recipe for feel-good philosophy and the re-enchantment of the world. In fact generating the universal feel-good factor will demand a remarkably mechanistic strategy in all but name. Moreover we've no evidence whether the different values of brute psychons themselves are just abject 'tickles', fragments of the godhead, or something else altogether - though in some branches of the universal wave-function ersatz demi-Gods must occur via wholly naturalistic processes. Unhappily for the purposes of re-enchantment, too, the promised non-redundant causal efficacy for mental states [conceived 'narrowly'] doesn't buy, for instance, free-will or freedom from mathematico-physical law. One may recall [particularly if one has mislaid the reference; can anyone help?] the experiment where a merely locally-anaesthetised subject abruptly raised his hand in response to electrode stimulation of his psycho-motor cortex. "Wow, you did that", he said to the surgeon, or something to that effect. Then the experimenter stimulated an adjacent pre-motor area - and the subject suddenly decided to raise his arm. Interesting. Apparently the phenomenology of willed action no more indicates a radical new form of agent-causality than any other sort of tickle. Real freedom of choice would involve someone, somewhere, doing something violating the continuous, linear, unitary and deterministic evolution of the universal wave function. Not a trick I've managed myself.

3.3 Moreover, although panpsychism scotches some worries about the possibility of zombies, it presents others in turn. Given a minimal panpsychist ontology as coded by the QM formalism, then zombies are logically impossible. Phenomenological minds [and the sort of substrate-irrelevant computational-functional virtual minds which phenomenological-minds sometimes simulate] both naturally and logically supervene on the properties of the stuff whose behaviour is described by the laws of physics. Yet at first blush the low-level facts don't rule out silicon etc quasi-zombies. Quasi-zombies lack true experiential manifolds. They are merely functional assemblages of quasi-punctate what-it's-likeness. Perhaps we'll be able to build them one day; though generating zombies via silicon VR rather than building physical robots might be easier. In any event, these notional quasi-zombies can be systematically interpreted as expatiating on their sentience and its mysteries. Their verbal behaviour is consistent with their agonising about The Problem Of Other Minds - but they lack the unitary phenomenal experience to match.

3.4 What's going on here? Should we worry about the possible existence of such exotic beasts? Perhaps meaning-externalists should; but ultimately, IMO, no. This is because carbon chauvinism and functionalism, an admittedly unlikely mésalliance, are actually quite natural bedfellows. For a panpsychist analogue of what we mathematically model - and naively visualise - as a carbon atom is micro-functionally unique in its valence properties. Comparatively warm, silicon etc partial functional analogues to organic psycho-wetware can't muster the presumptive quantum mechanically-described coherence to support even fleetingly unitary room-temperature phenomenal minds. So 'carbon micro-functionalism' might be a better title for this position. It's neither as naïve nor as chauvinistic as popular prejudice imagines; not that popular organic prejudice runs very deep around here.

3.5 But let's assume, counterfactually, that carbon isn't uniquely unique. Grant instead that full-blown silicon quasi-zombies are practically feasible. What are they on about when they speak; or rather when they utter the sorts of vocables which we do when talking about our beliefs and desires, our pains and pleasures, our existential angst and indigestion?

3.6 Strictly, I doubt if they're on about anything; or rather the paradox arises only when two radically different senses of 'content' are conflated. Now one-paragraph theories of meaning tend to be over-ambitious. They are preferable to completely unargued one-liners; so, as usual, what the heck, this is cyberspace, here goes. There are two different kinds of content worth distinguishing here, one which is real, narrow and intrinsic; the other simulated, broad and extrinsic. The former, a.k.a. intrinsic intentionality, inherits [albeit non-additively; QM] its subtle variant textural properties from the minimal mind-dust of which it's composed. The cognitive phenomenology making up this sort of content may nonetheless for many [e.g. syntactic, linguiform] purposes be functionally irrelevant - both to the role which intrinsically intentional thought plays in the informational economy of each virtual world in which it occurs and also to its role in any broader teleo-functionally-described eco-system of which it is a part. Natural selection has allowed fundamental self-intimating, self-referential what-it's likeness - phenomenal belief episodes - adaptively to simulate what we dub 'broad' content. An intrinsic, phenomenological 'aboutness' is simulating another ineffable form of 'aboutness' which we just don't, and perhaps can't, understand. Further complications ensue because there are other ways to simulate broad content. It can be simulated by functionally analogous but quasi-zombified robots. Well-programmed or properly trained-up silicon systems apparently do the job just fine. Self-reference, on the other hand, is intrinsically intentional; this thought isn't 'about' anything external to itself. The broad/narrow distinction collapses.

3.7 We thus arrive at another reason why zombies aren't possible. It consists in what is ostensibly a uniquely human ability - but is arguably symptomatic of something elemental and omnipresent in Nature itself. Zombies are by hypothesis behaviourally and neurofunctionally indistinguishable from non-zombies. Nevertheless, they are incapable of authentically self-referential thoughts, such as this one. They are quite capable of behaving in ways systematically interpretable as protesting that of course they can undergo self-referential thoughts. Some parts of their functional aggregate of constituents can monitor other parts; and in a limited sense, this may be loosely analogous to self-reference. But it's actually something else altogether. Indexicality and self-reference might seem a late development in evolution: specialised, linguistic and surely of no great cosmic import. Yet though scientific culture - abetted by neo-behaviourist philosophy and cognitive science - is geared to promoting introspective illiteracy, one finds that linguistic self-reference and indexical terms have introspectibly discernible phenomenological roots. I'm going to argue that their pre-linguistic counterparts are the stuff of the world.

3.8 Even on the orthodox "light-bulb" model of sentience - in which a kind of low-wattage consciousness abruptly switches on from pure insentience when a given level of neurofunctional complexity is attained - then something akin to a minimal self-intimating, self-referential what-it's-likeness is needed on pain of multiplying miracles beyond necessity. This isn't intuitively obvious: "ownerless" feelings, sensations or thought-episodes might seem a problem for panpsychism, but not for approaches which confine the incidence of qualia to full-blown persons or proto-persons. But unless the phenomenology of personhood and each initial infantile feeling of, say, pain switch on at exactly the same time, then rudimentary flecks of what-it's-likeness must be generated before they can be organised into anything phenomenally and structurally interesting. A primordial pain, for instance, is only ownerless in the sense that it is not instantiated in a person. It is still painful by its nature and to itself - whether or not it significantly interpenetrates a wider neural system.

3.9 Yet surely the success of science, and escape from the false prison of solipsism, both depend on the failure of this kind of subjectivism? When one thinks, verbally or otherwise, about chairs, tables or photons, one needn't be thinking about this chair here right now. Language allows us to abstract from the specificities of time, place and person to construct the scientific 'view from nowhere'. Indexicality and self-reference are puzzling anomalies; but surely not the building blocks of a new world order.

3.10 Once again, this response underestimates the power of the stuff of the world to simulate something radically different from what a pre-reflective consideration of its intrinsic attributes would suggest. On an orthodox physicalist stance, for instance, inert matter can self-organise to simulate something described by alien laws and principles (natural selection and the whole neo-Darwinian synthesis), namely life. Silicon atoms, too, can be arranged to yield virtual worlds which simulate properties and laws radically different from their virtual parental host; hence computers. Equally, I'm arguing, fields of pure subjectivity - courtesy of substrate-neutral 'universal Darwinism' - can adaptively simulate something else entirely thanks to the pressure of natural selection. In dreams, for example, one may trustingly presuppose that one is inhabiting an impersonal objective world whose features can be captured and referred to in third-person perspective. Yet the pseudo-public language one uses to do so is covertly indexical. This is because one's entire dreamworld is an autobiographical experiential manifold.

3.11 Surely it's different by day? When one speaks about tables, chairs and intermediate vector bosons, one isn't constrained to talking in some private language of thought about this table or chair or intermediate vector boson - or about myself.

3.12 Well, let's start on a boringly uncontentious note again. When in the grip of one's nightly psychoses, one assumes that one is noisily talking away in a public language about a public world too; but this is a misconception. For one's body-image is applying, or lends itself to description in terms of its applying, pseudo-public criteria to make pseudo-public utterances about a pseudo-public but intrinsically subjective world. Dreaming is nonetheless contextually different from being awake in two notable ways. First, the virtual world simulations of one's dreams don't causally co-vary (etc) with the mind-independent environment. Second, an effective muscular atony stops one's extra-cranial body from acting out the dramas internal to the virtual worlds which it's hosting. An organism that lacks this highly adaptive [but experimentally reversible] decoupling of its mind/brain and musculature really does act out its dreams. Human beings occasionally suffer from this rare and dangerous condition. Even though they may behave and speak as if they are awake, their motions are mere by-products of dream-dramas starring a somato-sensory-homunculus. The vocables which their host body emits while asleep are just by-products of an obliquely indexical mentalese internal to its dreams.

3.13 In the illustrative fable Alone Amongst the Zombies (HI; 2.15) I sketch how natural selection could orchestrate virtual dreamworlds into sometimes amazingly well-integrated organic virtual realities. Peripheral impulses deriving from a host organism's surface transducers serve to select dreamworld states; they don't add to the menu of potential dreamworld options. Environmentally-honed dreamworlds serve as adaptive models. They dynamically simulate the extra-mental environment as (un)faithfully as our own world-making mechanisms. Over millions of years, the DNA-driven coupling of the body-image/somato-sensory cortex and the typical host musculature allows these progressively scientised dreamers to simulate a third-personal impersonal 'view-from-nowhere'. They don't enjoy any broad semantic privileges over silicon etc systems. For their symbols are 'grounded' purely internally within the alienated subjectivity of each extra-homuncular dreamworld. Lacking any muscular atony, however, the bodies of host genetic vehicles obliviously go on to construct an advanced technological civilisation as they act out internal dreamworld dramas. The indexical and self-referential properties of all their dreamworld pseudo-public idiolects gets ever more heavily disguised. Some shy folk even get embarrassed about using the first-person pronoun too freely in their very own dreamworlds; for one does not wish to seem egotistical even to oneself.

3.14 Access to an authentic third-personal perspective would be very useful. Doubtless psychic remote-viewing, too, would be highly adaptive. It might yield up an Aladdin's cave of real semantic goodies as well. Alas such folkish stories of perception are best grouped with other psi-phenomena as spectacularly unproven. If something isn't publicly on offer, however, people tend to use home-grown substitutes in its place. Pseudo-publicity is better than no publicity at all; and what can't be had can be simulated instead. At any rate, a more elaborate adaptation of the dreamworld fable could be used to illustrate how a subjectivist ontology described by laws type-identical to ours could yield a world empirically indistinguishable from our own. Spelling out the differences between them and us in a non-question-begging manner is harder.

3.15 But I'm special. Even if others might suppose me imprisoned in a false mind-set of concepts, I know my meanings roam the Cosmos, non-locally alighting at will; and I feel like stamping my foot to prove it - virtually or otherwise. All I need to underwrite my conviction is a naturalised theory of meaning; but this has been unaccountably delayed.

3.16 If I'm right, then zombies of a sort are thus much more common than one might suppose. When dreaming, at least, one encounters them all the time. At a minimum, then, a big part of our lives is spent as dupes of mindless phantoms spawned by our own imaginations. When we are awake, of course, then the animated, sensuous, flesh-and-blood figures popping in-and-out of existence around one's egocentric body-image don't look like zombies - yet then neither do their dream-dwelling counterparts. More pertinently, philosophers' zombies don't act like [Hollywood or stuporous Haitian] zombies either: the idea that one could "catch out" a mere humanoid is excluded by the terms in which the usual thought-experiments are couched.

3.17 Recurrent, Night-of-the-Living-Dead dream-psychoses, are thus part of the Human Predicament. But by what manner of means does one temporarily escape their madness in one's early-morning mystical (er...I mean, hey presto, 'naturalistic!' phew!) "awakening" experience? It's not at all clear that one does...

3.18 On an inferential realist analysis of perception, what's essentially changed in waking as distinct from dreaming consciousness is not that a radically different suite of psycho-cortical sub-nets is getting fired up to celebrate the new day. We don't have one set of modules used for vector transformations in dreams and another to mediate self-transcendence - a.k.a. "perception". If we embodied such amazing dual-architecture mind/brains, then one set of mechanisms to avoid thermally-induced decoherence would be needed to support our panoramic dream-experiential manifolds. A second (and frankly incredible) set of decoherence-avoiding mechanisms would have to be invoked to explain our supposedly direct awake presentation with a unitary extra-mental macroscopic world. The mind boggles.

3.19 A far better sketch of a story can be told, though it brings scant comfort to zombiephobes. It's that when one wakes, the occipito-temporal homunculi which form the zombies of one's dreams no longer spontaneously self-generate. They now get activated only after one's genetic host vehicle receives peripheral stimulation from its environment. The zombie-generating mind-dust is still all there. The zombies lurk as dispositions in one's psychoneural weight-space. Their building blocks are in place all ready to propel them into action - loveable, hateable or boring, bless 'em - given half the chance; though since in neural-net-speak they embody neither transparent nor projectible representations, the 'lurking' metaphor shouldn't be pushed too far.

3.20 Ultimately, however, zombie-infestation doesn't depend on peripheral selection for its occurrence, either as a necessary or sufficient condition: zombies are natural both to the ecology of dreaming minds and to "awakened" ones. Their antics, however, tend causally to co-vary with goings-on in other host vehicles and virtual worlds elsewhere only when one is awake. In practice, the types of zombie-pattern which co-vary with one's biological mother, for instance, were originally differentially triggered in the presence of one's natural biological female parent. It was her presence, interacting with one's genetic host vehicle, which enabled the normal trained-up unfolding of one's idiolect of the language of thought in its characteristic pseudo-public guise. Without occipito-temporally generated mother-avatars to train (set the psycho-neural weights etc) one's locally distributed language network, then the normal epigenesis of Wittgensteinian anti-private language arguments couldn't in practice occur. Neither philosophical zombies nor silicon robots are endowed with this conscious, mentalese idiolect as distinct from a grungy neuralese imitation. Nor do they enjoy a conscious virtual world in which their subjective body-image is able to use it: "it's all dark inside" [though perhaps in one sense this turn of phrase is misleading. Darkness is a mode of visual experience; whereas zombies would suffer from a generalised and congenital version of Anton's syndrome. Victims of this syndrome have lost all their visual and visual-associative cortex. They are not merely blind. They don't experience darkness. They don't know what blindness is, nor do they recognise what their residual lexicon of visual terms would normally signify. This is because visual concepts no longer have any meaning to them.]

3.21 Thus positing brute sentience as the stuff of the world doesn't automatically exorcise the spectre of zombies on the cheap. For whether electrode-stimulated or not, dreaming or awake, silicon- or organically- triggered, we each occupy [infelicitously] virtual realities all the time. A fair scattering of the Multiverse consists in them. In common with florid schizophrenics, we (almost) all hear voices which we locate "out there" in an index-linked autobiographical world. Thanks to natural selection, and (sometimes) happily for most of us, waking voices really do causally co-vary (etc) with the dramas in virtual worlds of other genetic vehicles nearby. Yet it does get a bit lonely in here at times.