Thank you to the contributors of this blog for inviting me to write here!
I've been reading some recent work in philosophy of perception (a field I am fairly unfamiliar with), and I found the accounts of hallucination quite striking. Here's a passage from Susanna Schellenberg's Experience and evidence:
"consider a perceiver and a hallucinator. Percy, the perceiver, accurately perceives a white cup on a desk. Hallie, the hallucinator, suffers a subjectively indistinguishable hallucination as of a white cup on a desk; that is, it seems to her that there is a white cup where in fact there is none. What evidence do Percy and Hallie have for believing that there is a white cup on a desk? I will argue that Hallie has some evidence for her belief, but that Percy has more evidence than Hallie."
In the rest of this intricate article, Schellenberg demonstrates elegantly the metaphysical primacy of good cases of perception over bad ones. But what is striking is the suggestion that someone can simply sit down (or so I imagine the unfortunate Hallie) and hallucinate a white cup in front of her. Imagine what the world would be like if we really had such hallucinations on a regular basis!
So why do philosophers of perception so often speak about hallucinations as if they are common, and as if they are phenomenologically identical to veridical perception?
Probably they do this to test theories of perception. Katalin Farkas calls hallucinations where perception and hallucination are phenomenologically indistinguishable "philosophical hallucinations". Such hallucinations are largely constrained to philosophical thought experiments - they rarely (if ever) occur. People often realize they are in fact hallucinating. It is simply not the case that members of the non-clinical population can quietly sit down and perceive a cup that isn't there.
Not all philosophers of perception are interested in philosophical hallucinations. Some are interested in the hallucinations as they actually occur. For instance, Maurice Merleau-Ponty based his account of hallucination on clinical data, and attempted to explain why hallucinators often can distinguish between hallucinations and veridical perception, due to the different phenomenology of these experiences (see this paper by Romdenh-Romluc).
Imagine a world where philosophical hallucinations really did occur. Not frequently, but often enough for us to doubt the testimony of perception. To understand out what such a world would look like, I'd need a novella or novel in speculative fiction. However, we can already briefly speculate what it would do to our experience. In such a world, perception would bear a striking analogy to testimony (even more so than Reid claimed): in the case of testimony, we trust what others tell us, but sometimes we are deceived. Most cases of false testimony are indistinguishable to us from cases of true testimony. The same would be true for perception.
Although philosophical work on perception often uses philosophical hallucinations as a way to probe our intuitions about accounts of perception, it seems to me theories of perception would differ from the present ones if philosophical hallucinations occurred frequently. For one thing, perception is regarded as a basic source of knowledge, but if we regularly saw or heard things that weren't there, this would be less evident. We'd rely more on background information (for instance, witnessing something like Magritte's Golconda, we would disbelieve what we saw). As a result, we'd disbelieve a lot of real but improbable things we saw.
In such a world, social confirmation would become very important because it would be the most reliable way to validate our perceptual experiences. To be quite speculative, I think philosophers would have been more attuned to the social dimension of knowledge.
Interestingly, Merleau-Ponty recognized the importance of this social dimension. He noted that "an important difference between perceptions and hallucinations is that whilst perception presents its object as being situated in the intersubjective world, hallucination does not" (Merleau-Ponty, cited in Romdenh-Romluc). This quality of "publicness" would be even more important if we could not distinguish between the phenomenology of hallucinations and of veridical perception.