Internalist Representationalists can easily appear to have a natural advantage over Distributed Enactivists when explicating visual hallucination. A hallucination, they would say, is to be understood as the obtaining of an internal representation of x in the absence of x, or at least in the absence of an appropriate causal impact of x on the experiencing subject. Enactivists on the other hand are all about our attunement to sensori-motor contingencies. In the absence of the object x, however, what sensori-motor contingencies do we have to do with?
There's a way in which a representationalist account of hallucination can start to seem like an inevitability. The inevitability comes, I think, from mistaking an internalist representationalist conception of visual perception for our actual pre-understanding of the phenomenon. It rears its head in the immediate 'Well, if there's no outer object to see, but it seems to me rather as if I'm seeing one, then there must be some other (inner) visibilium'. The move is clearly fallacious - after all the point of visual hallucination is that one isn't seeing anything. (This is why talk of 'hearing voices' is obviously misleading: the person with AVHs is precisely not actually hearing voices! Instead they are hallucinating - whatever that is!) In hallucination we have neither act nor object, but are inclined to posit both (the inner beholding of an inner object). We need to resist that inclination, since it obscures the explanandum.
Representationalists, in truth, only appear to have a theory of perceptual experience - and this is because they simply reduplicate the explananda. They deploy a notion - a 'visual representation' - which is in its root sense something which is itself a visibile. The having of pictures of x is not a good model for what it is to see x, since pictures themselves require to be seen. Our ability to engage with representations is derivative of, and not foundational for, our basic forms of world-engagement - such as visual perception. The very last thing we need, in explaining what it is to be able to see an external visual scene, is yet another (but merely internal) visual scene to contend with. What might look like a weakness of a non-representationalist account of hallucination is in fact precisely its strength, and we should precisely take comfort in the fact that it is going to try to explain perception and hallucination in terms which don't already presuppose the intentionality of visual experience. (So if, when thinking enactively, we find ourselves asking, e.g., 'but where is the hallucination?' we should invite ourselves to refrain from answering and instead invite ourselves to stop thinking of it as a thing witnessed by an interior observer - i.e. to step out of Dennett's Cartesian theatre.)
In truth Enactivism supplies a lovely theory of hallucination. Attunement to sensori-motor contingencies is a matter of being able to sort out which changes in sensory stimulation are due to environmental and which to own-bodily movement. As is well known, to learn to see we need to be able to move around the environment. And there are times when situations sensorially confound our grasp of the relevant contingencies. I step on a static elevator and my body lurches. I sit on a train in a station and cannot tell if it is my train or the one next door that is moving. There is an ambiguity between own movement and other movement. Paralyse my eye muscles and the world appears to move when it is my body that moves, or worse the coherence of the visual scene breaks down. What we meet with here are situations in which a history of entrainment leads to a tacit expectancy of null environmental stimulation, perhaps because (a la Frith's Helmholtzian model) the body is no longer adequately 'aware' of its own-initiated movement, which then, because we do have to do with some form of unacknowledged bodily movement, results in quasi-perceptual quasi-experience.
Hallucination, one might say, is a result of a failure to correct for the 'negative' of the visual sensory stimulation generated by planned or actual bodily movement. It results from a temporary failure in attunement to sensori-motor contingencies against a backdrop of prior world-relating perceptual experience. We don't need a theory of internal representations to get this off the ground. It isn't that a history of prior attunement to sensori-motor contingencies populates an internal stage with internal images. (See Myin and Degenaar's chapter on Enactive Vision in Shapiro (ed) Routledge Handbook of Embodied Cognition for further details on this.) Rather what we have is precisely a temporary failure of attunement.
Now the kind of thing we might expect a psychologist to say is that 'what is happening is that people are 'seeing' what they expect to see'. This however is nonsense. The whole point is that people are 'seeing' the opposite of what their bodies 'expect' to encounter. This, I suspect, is what accounts for the phenomenology of apparitions. We see ghosts when our bodies expect to encounter people. A ghost, one might whimsically suggest, is a form of perceptual anti-matter.
The theory also works for auditory hallucinations (i.e. 'voices'), and it connects nicely with the psychodynamics of schizophrenia. From a dynamic perspective schizophrenia involves a drastic splitting off from one's own aggressive and libidinal impulses. This 'id shadow' is then located externally to the subject. (We may also think of a breakdown of or splitting off from a pathological superego.) Remember, this is the whole point of the 'attunement to sensori-motor contingencies' enactivist theorisation of perception: the system must reach some dynamic equilibrium by chalking up sensory changes when the sense organs are still, or lack of sensory changes when the sense organs move, to either self or other. A failure of entrainment to aptly inform current attunement results in mis-allocation of motive or affect. Splitting off from one's own aggression results in a failure to anticipate it as arising from within. It is not that I expect to find it outside me; it is rather that I wrongly (and strongly, motivatedly) don't expect to find it within me: the external world therefore must take the hit. Dynamic equilibrium, which is nothing other than the bodily self maintaining its sensori-motor self-identity, is preserved - this is the dynamic aspect of the system functioning precisely as it should - but the cost of splitting+projection (which are not in reality two separate mechanisms, but rather just one) is hallucination.
Tactile hallucination works in a similar way. My body expects to touch but this expectation does not meet with its realisation. What therefore results is a quasi-experience (as) of being touched.
This theory in fact accounts I believe for the full range of autoscopies and coenaesthopathies, both as these occur in the schizophrenias but also as they occur in the sub-clinical experiences of schizotaxic individuals who so often chalk up their relevant inner and outer experiences to (inwardly) the living experience of 'subtle energies' / 'chakras' / 'meridians' and (outwardly) encounters with the 'spirit world'. The tragedy for some such individuals is when their difficulties in social-corporeal attunement motivates a substitute engagement with the 'spirit world' - an ongoing empty, deathly, and hollow form of unwitting narcissism in which they spend their life instead in relationship with the inverse of their bodies' sensory expectations.