Fabian Dorsch "Hallucination and Phenomenal Error*"

Two important characteristics of hallucinatory experiences are that they do not involve a perceptual relation to some perceived aspect of the world, and that they at least sometimes may be subjectively indistinguishable from veridical perceptions. The first feature concerns the metaphysical structure of hallucinations, while the second concerns their phenomenology and, as we will see, also their rational role.

In this paper, I would like to get clearer about certain aspects of the relationship between these elements of hallucinations. More specifically, my aim is threefold:

(a) to show that, partly because of their subjective indistinguishability from perceptions, hallucinations involve some kind of phenomenal error;

(b) to argue that this error is perhaps best accounted for in terms of intentionality, understood as part of the phenomenology of the mental episodes concerned;

(c) to illustrate that the resulting intentionalist view is in principle compatible with, but also puts some considerable pressure on metaphysical and phenomenal versions of disjunctivism about perceptual experiences - pressure which a phenomenologically based and non-disjunctive account of the nature perceptual experiences can much more easily avoid.

1. Subjective indistinguishability and rational role

Before I will begin to pursue each of my three aims in turn, it will be necessary to clarify what subjective indistinguishability amounts to, and how it is related to the rational role of the episodes concerned.

1.1. Subjective indistinguishability is primarily a phenomenological matter. Two episodes are subjectively indistinguishable just in case they present themselves to us in phenomenal consciousness in such a way that we cannot tell them apart from our subjective perspective (i.e., 'from the inside'). But subjective indistinguishability is perhaps compatible with a difference in phenomenal character, that is, a difference in what it is like to have, or be in, the respective episodes. The two may go apart, say, because we may be able to subjectively notice qualitative differences between episodes, but not phenomenal differences in particularity (e.g., as they may arise when we perceive two qualitatively identical, but numerically distinct twins). However, if two episodes are subjectively distinguishable, there will also be a difference in what it is like to have, or be in, them.

1.2. But subjective indistinguishability is also closely linked to the rational (or, more generally, functional) role of mental episodes. In particular, we treat subjectively indistinguishable episodes in the same way - say by (not) relying on them as providers of reasons - when rationally forming beliefs or intentions. My suggestion will be that the best explanation of this link is that how episodes present themselves in phenomenal consciousness typically indicates or reflects their rational role.

2. Accepting phenomenal error

The argument to the effect that the phenomenology of hallucinatory experiences involves some kind of error proceeds in three steps.

2.1. The first is to argue that part of the subjective difference between veridical perceptions and sensory imaginings can be adequately captured in terms of the former, but not the latter, presenting themselves in phenomenal consciousness as genuine relations to objects present in the actual and mind-independent environment of the subject. This involves that perceptions, but not imaginings, present their objects as mind-independent and actually there before us. And this subjective difference reflects the fact that perceptions, but not imaginings, have the role to be endorsed in belief and action concerning the external world.

2.2. Next, I will try to show that there are two plausible ways of understanding what it may mean for hallucinations to be subjectively indistinguishable from experiences which present themselves in phenomenal consciousness as relations to the external world.

Assuming that perceptions and hallucinations do not possess distinct phenomenal characters, it should mean that perception-like hallucinations, too, present themselves in phenomenal consciousness as relations to the external world.

But assuming that perceptions and hallucinations possess distinct phenomenal characters, it will more likely mean that perception-like hallucinations present themselves in phenomenal consciousness as having the phenomenal character (or phenomenal nature) that perceptions present themselves as having (and which, in the perceptual case, can be truly described in terms of them presenting themselves as relations to external objects).

In both cases, the fact that we rationally rely also on perception-like hallucinations can be explained by reference to how they present themselves in phenomenal consciousness: either the explanation is exactly the same as in the case of veridical perceptions; or it exploits the fact that how hallucinations present themselves has to be elucidated in terms of how perceptions present themselves.

2.3. The third and final step of the argument is to note that, following either of the two interpretations, there is a phenomenal error in hallucinating due to a mismatch between the phenomenology of hallucinations and some of their metaphysical features. Either hallucinations present themselves as relations to actually present objects, but in fact are not such relations. Or they present themselves as having a certain phenomenal character (or phenomenal nature), which they in fact do not possess. In both cases, there is a gap between how hallucinations present themselves in phenomenal consciousness and how they in fact are. And my contention is that this mismatch constitutes a genuine error in the light of the fact that how mental episodes present themselves is meant to, and typically does, indicate their rational role.

3. Explaining phenomenal error

The third part of the paper will prepare the case for the view that the phenomenal error involved in hallucinations can and should be accounted for in intentionalist terms.

I intend to support this conclusion partly by putting forward considerations to the effect that the error cannot be traced back to some kind of introspective or reflective error. The main problem for such an approach is to capture the phenomenality of the erroneous awareness. Hallucinations are phenomenally conscious, at least in the sense that they present themselves in phenomenal consciousness and may be subjectively indistinguishable from conscious perceptions in the sense specified above. What would accordingly be needed is some higher-order account of how they present themselves, which maintains that this is a matter of how some higher-order state represents them to be. To avoid a regress, the higher-order state should be taken to be nonconscious - an implicit or tacit representation, say, or a dispositional state. But it remains mysterious how such a state can give rise to the open and explicit subjective similarities and differences between the various kinds of sensory episode at issue.

As a better alternative, I would like to propose an account of phenomenal error in intentionalist terms. Three claims are central to this view: (i) one aspect revealed by how both perceptions and subjectively indistinguishable hallucinations present themselves in phenomenal consciousness is that they are appearances of something; (ii) more specifically, they are token-reflexive appearances, that is, appearances not only of aspects of the external world, but also of themselves and their relation to those worldly aspects; and (iii) what it means for a perceptual experience to be a token-reflexive appearance can be satisfactorily spelled out in intentionalist terms.

As I understand the term, a central part of what it means for a perceptual experience (and presumably other mental episodes) to be intentional is that they are subject to the (weakly) normative constraint to occur - or at least to be relied on in the rational formation of beliefs and intentions - only if certain conditions on the world are satisfied.

Furthermore - and following the phenomenologist rather than more recent representationalist traditions - their intentionality should be understood as constituting an aspect of how they present themselves in phenomenal consciousness, at least in so far as we are to a substantial extent able to subjectively tell which conditions on the world appear to be satisfied as part of our experiences (again, some limitations may arise due to the fact that the conditions on the world may concern particular objects). The perceptual experience of a red object, say, involves an appearance of an object as being red and should not be trusted if in fact there is no such object.

The key idea is now that the intentionality of perceptual experiences is partly self-reflexive (and thus also partly non-sensory). That is, these experiences present themselves in phenomenal consciousness as perceptual relations to external objects and should be relied upon only if they indeed involve such relations. Assuming that intentional error is a well-understood phenomenon, I believe that this suffices to adequately account for the phenomenal error characteristic of perception-like hallucinations: they involve an intentional error not only about the environment, but also about themselves and their own relationship to the environment.

4. Phenomenology and metaphysics

In the final section of the paper, I will be concerned with the metaphysical consequences of the preceding phenomenological considerations. In particular, I aim to show that the proposed intentional account of phenomenal error is neutral both on the nature and on the metaphysical structure of perceptual experiences, but establishes a serious challenge for disjunctivist views.

4.1. The metaphysical neutrality becomes revealed partly in the fact that the developed intentionalism is compatible with the views that what is essential to perceptions and hallucinations is how their common way of presenting themselves in phenomenal consciousness, or alternatively their common representational function. But it is also compatible with various disjunctivist views which claim that there is an essential difference between perceptions and hallucinations, whether in their relationality or concerning their phenomenal natures.

4.2. However, I aim to argue that the intentionalist view puts a certain constraint on any account of the nature and metaphysical structure of perceptual experiences. What we would like to have is an explanation of the intentionality common to both veridical perceptions and subjectively indistinguishable hallucinations. But for this task, it does not suffice to point to differences in relational properties or phenomenal natures because neither can establish the commonality and, in particular, the normativity of the intentionality to be explained (and it is therefore no surprise that many disjunctivists have denied the normativity of perceptual experiences). Without assuming this intentionality, on the other hand, the phenomenal error involved in hallucinations cannot be properly elucidated.

One non-disjunctivist solution is to locate the nature of mental episodes and the corresponding mental kinds on the level of their subjective (in)distinguishability - that is, on the level of how they present themselves to the subject in phenomenal consciousness (or on the underlying level of rational or functional role). The commonality and normativity of the intentional nature of all perceptual experiences may then be traced back to their shared rational role or to an even more fundamental common feature (such as their evolved representational function). Perceptual experiences may then still count as disjunctive on a lower metaphysical level in so far as veridical and hallucinatory experiences continue to differ in their relational metaphysical structures. But since this disjunctivity would not transfer to the subjectively (or functionally) understood nature of perceptual experiences, it would not bear on the most fundamental ontological divisions in the conscious mind.

* This paper is the result of a close research collaboration with Gianfranco Soldati (University of Fribourg). Thanks are due to him and also to Matthew Nudds for their very helpful comments.