Does the argument from illusion show that there are no differences between the visual experiences involved in veridical perception, illusion, and hallucination?

 

No, it does not!   In fact, the argument from illusion assumes that there are no differences between the visual experiences involved in veridical perception, illusion, and hallucination.  And the argument fails precisely because that assumption can be challenged, and denied, by the Direct Realists at whom the argument is directed.

Direct Realist theories of perception maintain that in veridical perception we are directly and immediately aware of really existent external physical objects in our environment.  Indirect Realist theories of perception argue instead that we are directly and immediately aware of something intermediate (frequently called "sense-data"), and hence only indirectly of really existent external physical objects in our environment.  The argument from illusion is presented by indirect realists as demonstrating that direct realism is wrong, and that in veridical perception there must be something intermediate between the perceiving mind and external reality.

The argument from illusion dates back to Hume (Hume, David; Enquiry Concerning Human Understanding, 1758, section XII.1) who framed the first attempt.  In modern form, it proceeds roughly as follows:-

(i)         when we are the subject of an illusion, what we visually experience is not what is really there (in the Müller-Lyer Illusion, for example, the lines are not different lengths as they appear to be), and when we are the subject of an hallucination, what we visually experience is not really there at all (there really aren't any pink elephants dancing on the table);

(ii)        therefore, what we are visually experiencing, what we are immediately aware of, in the cases of illusion and hallucination cannot be really existent external physical objects in our environment.  It must be something else (like "sense-data").

(iii)        there is no discernable phenomenological difference between the visual experiences involved in veridical perception, illusion, and hallucination;

(iv)       therefore, there is no reason to suppose that in the case of veridical perception we are visually experiencing (immediately aware of) a really existent external physical object in our environment.

(v)        therefore, the direct realist conception of perception must be false.

You'll notice that the argument's conclusion, that what we have a visual experience of in the case of veridical perception must be something that is not really existent external physical objects in our environment, is based on the assumption at step (iii) that veridical perception can be treated as similar in all meaningful respects with the erroneous visual experiences of illusion, and the fabricated visual experiences of hallucinations.  And particularly can be so treated for the purposes of identifying what those visual experiences are direct awareness of.

This key assumption does have some supporting rationale that when looking at the argument from a broader perspective can be considered to be a part of the argument.  There is no dispute over the empirical evidence that in the case of many illusions the subject cannot change the way that they perceive the situation, even though they know that their visual awareness is wrong.  In the case of the Müller-Lyer Illusion, for example, no amount of careful measurement and self-persuasion can change the fact that the lines stubbornly continue to look different lengths.  Or that a change in mental state -- particularly one's set of beliefs -- can change the visual experience of other illusions although the external objects do not change.  In the Duck-Rabbit Illusion, for example, you might see only the rabbit and not be able to visually experience the duck until it is suggested that it is there.  There is undoubtedly (and undisputedly) the same persuasive immediacy to illusory visual experiences as there is to veridical visual experiences.  In the absence of other evidence, the argument from illusion suggests, it seems reasonable to suppose that veridical perceptions and illusory perceptions are phenomenologically indistinguishable from the subject's perspective.

There is a key hidden premise, however, behind how the argument from illusion proceeds from the subjective phenomenology of the two kinds of non-veridical visual experiences.  The reasoning assumes that any visually experienced properties must be objective properties of the thing being visually experienced.  This "objective property" assumption shows up clearly in step (ii) above where it is argued that there must be something other than the external environment that we experience when we experience an illusion or an hallucination.  In applying the hidden premise that there is actually such a thing as seeing an illusory situation "as it really is", this critical step in the argument proposes that if I am seeing the illusion, then there must be something that really is as the illusion appears to be so I can see it "as it really is".  It is assumed that if the Müller-Lyer lines are really of equal length (an obviously non-problematic stipulation), and if I see them as of unequal length, then I can not be seeing them "as they really are".  What I must be seeing is some intermediary thing that I can see "as it really is" -- something that has lines of unequal length.  And in applying the premise to hallucinations, this critical step in the argument proposes that if I am seeing the hallucination, then there must be something that really is as the hallucination appears to be so I can also be seeing it "as it really is".  If I see non-existent pink elephants, then there must exist something that displays the properties of pink elephants so that I can be seeing it "as it really is".

However, this hidden "objective property" premise underlying the argument from illusion gives Direct Realists the opportunity to demonstrate that despite apparently similar subjective phenomenology, illusions are a quite different kind of visual experience from hallucinations.  Direct Realism argues that the visual experience of an illusion does not involve inherent observer-independent properties of the thing perceived, but rather relational properties involving the external object, the conditions of perception, and the visual apparatus (including the mind and its set of beliefs) of the perceiver.  The Müller-Lyer lines can only be seen as unequal -- given an experiencer with our kind of visual processor.  They simply cannot be seen as equal.  There is no possibility of there being an experience of this illusion "as it really is" if "as it really is" is taken to be an experience of equal lines.  The shape of the swizzle stick is straight, but its visual appearance (the character of the visual experience it generates) varies according to the physics of light refraction.  If I am seeing it as bent, then I am seeing it "as it really is" -- given the physics involved.  If I can see the rabbit but not the duck, then I am still seeing the picture "as it really is" -- given my current set of beliefs.  Thus, according to Direct Realism illusory visual experiences are in fact veridical - given the viewing conditions.  Hence Direct Realism can argue that veridical perception actually includes illusions and hence obviously there will be no differences between the visual experiences involved in veridical perception and illusion.

With illusions removed from the argument, that leaves only hallucinations.  So does the argument from illusion show that there are no differences between the visual experiences involved in veridical perception and hallucination?   Again, the answer is "No".  The reason hasn't changed.  The argument from illusion still assumes that there are no differences between the visual experiences involved in veridical perception and hallucination.

The argument from hallucination is a much stronger line of reasoning than is the argument from illusion.  (In point of fact, the "argument from illusion" is often called the "argument from hallucination" in recognition of the weakness of the reasoning from illusions.)  There is no question about the fact that when experiencing an hallucination or a dream, there is nothing in the real physical world that in any way correlates with the subject's visual experience.  The empirical evidence unquestionably is that many subjects of hallucinations (or dreams) do not distinguish between their fabricated visual experience and veridical perception.  Even in the grip of the most outrageous hallucination, many subjects will operate on the basis of the reality of their hallucinatory visual experiences in the same way that they would in the presence of veridical visual experiences.  However, it must be emphasized that this is not as strong as suggesting (as is done for illusions) that they could not, if they chose, distinguish between the two.  The empirical evidence is merely that they often do not.  Clearly then, the phenomenology of veridical perceptions and hallucinations are at least sometimes operationally undistinguished from the subject's perspective, even if it remains possible that they might be distinguishable should they so choose, and even if sometimes the subjects can so distinguish.  (Sometimes, when subjects see pink elephants, they know that there are no such things as pink elephants, even if the visual hallucinations are very convincing.)   Yet it is the claim of the argument from hallucination that not only are veridical visual experiences sometimes operationally undistinguished from hallucinatory visual experiences, but there is in fact no discernable difference between the two.  The empirical evidence would not seem to support this stronger claim.  That lack of persuasive evidence is not of immediate concern to the indirect realist, however, since their argument is based on a claim of principle rather than fact.  The argument claims that it is possible in principle that hallucinatory visual experiences are indistinguishable from veridical visual experiences.  But the empirical evidence at least raises some doubts about the reasonableness of that claim.  Doubts that warrant further examination.

To further examine the reasonableness of this claim, we must consider hallucinations in light of what we know of biology in general and evolution in particular.  The range of sensory acuity involved in perceptual experiences runs the scale from photo- and chemically sensitive bacteria to the binocular tri-colour vision of the primates, the auditory range of bats, the olifactory prowess of dogs, etc..  And the capabilities of the "experiencers" involved in perceptual experiences also runs the scale from unicellular bacteria to Homo sapiens.  Any explanation of veridical sensory experiences must be equally applicable, in general principles, to that entire landscape, and all of the five senses involved.  The implication is that any explanation of sensory experiences - veridical or hallucinatory - in humans must be exportable, in at least general terms, to that entire virtual landscape, unless an explanation can be provided as to why human sensory experiences are different.  (Unless, of course, it is argued that any creature in that landscape is just as susceptible to hallucinations as are humans.)   Since the claim that there are no differences between veridical sensory experiences and hallucinatory sensual experiences offers no opportunity for such a "uniqueness" explanation, there is a profoundly important reason for supposing that veridical perception is fundamentally different in its essential nature from an hallucination. 

Veridical perception is the means by which we manage to flourish despite the slings and arrows that the environment throws in our way.  Hallucinations are but static, noise, or unwanted feedback in the processing that is a sensory experience.  If this is a valid way of considering hallucinations, then it would be profoundly unreasonable that there should be no discernable differences between the experiences involved in veridical perception and hallucination.  Perhaps the correct approach is to view the two kinds of visual experience as processes in time rather than as snap-shot instances of experience.  For given that approach, veridical visual experiences are clearly quite different and are obviously distinguishable from hallucinations by the subject.  Considered as an element of our interaction with the environment over time, hallucinations are like false beliefs.  Veridical sensory experiences aren't likely to get you killed, and they are likely to get you what you want.  Hallucinations aren't likely to get you what you want, and are likely to get you killed.  In other words, given the successful evolutionary history of Homo sapiens, one is forced to conclude that the proportion of hallucinatory experiences that really are phenomenologically indistinguishable from veridical experiences must be very insignificant.  They must be so rare, in fact, that there can have been no meaningful evolutionary pressures to develop a means of avoiding their threat (otherwise they wouldn't be indistinguishable).

In summary then, the argument from illusion does not show that there are no differences between the visual experiences involved in veridical perception, illusion, and hallucination, it assumes that there are no differences.  Even when the scope of the discussion is expanded to encompass the rationale for that assumption, I have shown that the assumption is not reasonable.  In the case of illusions, it can be argued that the assumption, while true, is based on a false understanding of the nature of illusions.  And in the case of hallucinations, it can be argued that the assumption is not only unreasonable, but also almost certainly false.

 

References

Merriam-Webster Online Dictionary (http://www.m-w.com/cgi-bin/dictionary)

Crane, Tim, "The Problem of Perception", The Stanford Encyclopedia of Philosophy (Winter 2006 Edition), Edward N. Zalta (ed.), URL=<http://plato.stanford.edu/archives/win2006/entries/perception-problem/>.

[Up] [Home] [Next]