The eminent 19th century physiologist and vision scientist Johann Purkinje wrote:
Often I have been surprised that blinking the eyes does not disturb sight, assuming that during the blink complete darkness must occur. On closer observation however, I found that the visual field of the open eye, with all its lights and images for a short time remains before the mind, after the eyelids are closed. The more attentively I comprehend a single not very extensive image, the longer I am able to hold it with closed eyes before the mind. This afterimage is precisely to be distinguished from dazzling images [Blendungsbilde; afterimages caused by bright light?]. The afterimage can only be held longer by free activity and disappears as soon as the will diminishes, though it can by the same action be called back again; dazzling images float automatically before the mind, disappearing and appearing again for objective reasons.
[Translations of other related passages are posted in the Underblog here.]
Later psychologists in the introspective tradition (right through Brown's 1965 influential review of the literature on afterimages) often took for granted that such momentary positive images exist, though Purkinje's remarks about the the involvement of the will and the distinction from "Blendungsbilde" were largely neglected.
Close your eyes. Blink a few times. What do you think? Are there "Purkinje afterimages" of this sort?
I think I sometimes experience a positively colored "dazzling image" that quickly fades -- for example, if I look for five seconds at the flourescent light overhead, I seem to experience an afterimage of its bright bars for a fraction of a second after I close my eyes, and before the usual "flight of colors" sets in. Maybe I also have the experience after staring at my computer screen awhile. But I can't find such a momentary image when I'm looking in ordinary conditions at ordinary reflective surfaces -- my bookcase, for example.
I wonder how much Purkinje, in this passage, is driven by a perceived need to explain the lack of interruption in the blink. To the extent that's what's driving him, I'm suspicious -- partly because I think it possible that a blink does momentarily disturb visual experience, and partly because Dennett has convinced me (in his marvellous book Consciousness Explained) that often we don't notice even fairly large gaps and lacunae in our visual experience (temporal gaps, blind spots, impoverished information from the periphery) if there's no particular epistemic need that's going unfulfilled.
There are also temporal issues. Brown describes the "first positive phase" of the afterimage as starting after about 50 milliseconds and continuing for about 50 milliseconds. That's pretty fast! Given the speed of neural processing, our experience in general probably runs some tens of milliseconds behind environmental input (at least unexpected input). So there are issues, at least, about how one draws the line between ordinary visual experience, possibily slightly delayed relative to a brief and unexpected stimulus, and "afterimages", conceived as something distinct from that.
I also worry that once the existence of such images was admitted into the introspective literature, later authors would be reluctant to deny their existence, from fear of being seen as less introspectively expert than those who make such fine judgments about their afterimages: "Well Purkinje saw it, and I see it. If you don't, maybe you're just not well enough attuned to your own experiences, well enough practiced at introspection!"