In Carrasco’s experiments, an attended .20o gap is not discriminated from an “unattended” .23o gap. I think the best conclusion is that attention changes perceived size and contrast. (Recall that I am talking about spatial attention rather than attention to a property instance or an object.) Do the gaps just fail to look different or do they look the same?
In Carrasco’s main paradigm, subjects are forced to choose which stimulus is bigger (or faster or higher in contrast). In the case of an attended .20o gap as compared with an “unattended” .23o gap, subjects are as likely to choose one option as the other. In this sense these options are not discriminable. However, I mentioned that when subjects are asked instead whether the items are the same or different, the effect of attention is slightly smaller. And that may suggest that there is substantive daylight between not looking different and looking the same. (Of course this difference matters very much in some contexts, for example, as mentioned earlier, the context of the phenomenal Sorites issue; Morrison 2013.) As I mentioned earlier; Anton-Erxleben et al. and her colleagues argue persuasively that the smaller effect is due to the same/different paradigm being a less sensitive measure (2011). In what follows I will assume that the attended .20o gap looks the same in respect of size as the “unattended” .23o gap.
I put the “unattended” in quotes because I mean no commitment to the improbable claim that there is no attention on the .23o gap. There is no agreement on whether there can be conscious perception or even unconscious perception with zero spatial attention or whether zero spatial attention is even possible.[12] Indeterminacy in our concept of attention may even make this an unanswerable question. Still, I will adopt the abbreviatory convention of referring to stimuli that are not focally attended as “unattended”.
Since the apparent size of the gap differs depending on where one is attending, the question arises as to which of these various percepts of the gap gets its size right (or most nearly right) and which gets its size wrong (or most nearly wrong). Veridicality is a matter of getting things right and veridicality in perception is a matter of the world being as it appears to be. There would be no good reason to decide that the veridical percept of the gap is one in which one is attending to a spot one inch away from it; why pick one inch rather than one centimeter? (Recall that the attentional landscape of amplification and inhibition varies from place to place and from moment to moment.) The most obvious candidate for a non-arbitrary answer to the question is: the veridical percept of the gap (if there is any veridical percept) is the one in which one is attending to the gap itself.
I think of veridicality as all or none, but for the sake of accommodating different opinions I can countenance degrees of veridicality. One innocuous use of such a phrase is that if one represents a .19o gap as .20o, the percept is more veridical than if one represents it as .21o. Also we could say that other things equal, a percept that attributes a higher probability of being .21o to a .21o gap is more veridical.
But once it is stated that the most veridical percept of the gap is one in which one is attending to the gap, one wonders why one should believe this hypothesis rather than the opposite hypothesis that attention distorts by magnifying, illusorily, for the purpose of getting information and that the attended item is seen illusorily. Is the perception of the gap with less attention really illusory in the sense of a discrepancy between stimulus and perception?
In an article on Carrasco’s discovery, Stefan Treue (2004, p. 436) says this:
In summary, this study provides convincing support for an attentional enhancement of stimulus appearance. It completes a triangle of converging evidence from electrophysiology, functional brain imaging and now psychophysical findings, which argues that attention not only enhances the processing of attended sensory information but manipulates its very appearance. …attention turns out to be another tool at the visual system’s disposal to provide an organism with an optimized representation of the sensory input that emphasizes relevant details, even at the expense of a faithful representation of the sensory input. (italics added)[13]
I quote Treue not because I agree with him but in order to get a statement of that view on the table. There is no sufficient reason to accept the view that an attended perception of a gap allows us to see it as it really is rather than the view that attention in perception is like a magnifying glass, distorting for informational purposes at the cost of illusion. I can imagine considerations that might incline one towards adopting one or the other of these positions—that attention falsifies or that attention “veridicalizes”—but the adoption would be for purposes of one or another kind of utility, not as a principled reason to think that the highest degree of veridicality is really to be found in that case.
The challenge is to find a principled reason for regarding seeing a thing or place with a certain degree of attention to be more veridical than seeing it with a different degree of attention. Sufficiently decreasing attention to something can move the perception below the threshold of visibility. But not seeing something that is too small to see or to faint to see need not be a matter of illusion.
Chris Hill (in conversation) and Sebastian Watzl (forthcoming) have argued that there is an optimal level of attention and perception with all other values engender illusion. Watzl’s version of this view appeals to the idea that the function of attention is to make perceptual representations usable—as opposed to the function of perception of veridically representing the world. These functions will conflict in normal circumstances. The optimal level of attention for fulfilling the function of perception—veridicality—will be achieved in an idealized scenario of no attention, or one of equal attention to everything. This is an interesting point of view, but is contradicted by the point made earlier that veridicality conditions require a history of veridical representation.
Epistemicists about vagueness think that there can be an unknowable fact of the matter as to a sharp border between bald and not bald, a number of hairs that a bald man can have even though adding a single hair will make the man not bald. But if there can be a fact about a border even though there could be no principled reason to regard any particular border as the real one, why can’t there be a fact about what degree of attention engenders veridicality that no one could have a principled reason to accept? Epistemicists should not regard the cases as analogous since they think there is a principled reason to hold there is a fact about a border and a principled explanation for our ignorance (Sorenson 2013).
It may be objected that there is good reason to accept Treue’s point of view since after all, attention to the .20o gap makes it look, illusorily, to be the same size as the unattended .23o gap. But why not blame the illusion on the percept of the unattended gap rather than the attended gap? One can blame the mismatch, but that does not help in deciding whether attention to an individual item engenders veridicality or illusion. I think the issues are clearer when one avoids the comparative perception and just asks, say of the situation in Figure 5, whether perception of the gap can be veridical when it is cued and one is attending to it or when it is not cued and one is attending elsewhere. There is no adequate justification for one answer over the other. Some may wish to abandon the notion of veridicality as applied to perception but that would be to abandon the notion of representational content as applied to perception and so to abandon representationism. The representational content of a perception is—constitutively— the veridicality conditions. There is a strong a priori case for perceptual representation (Siegel 2010). And in any case the science of perception makes essential use of veridicality (Burge 2010).
In the discussion of the analogous issue with regard to inhomogeneities in the visual field, I noted that the sort of differences in perception caused by spatial inhomogeneities are paralleled by differences due to temporal inhomogeneities—that is variation from percept to percept due to random factors. Any two percepts of the same items at the same point in the visual field with the same degree of attention are likely to differ in apparent contrast (and other properties) due to these random factors. It is hard to see a rationale for treating spatial inhomogeneities differently from temporal inhomogeneities and it is hard to see a rationale for treating either of them differently from the inhomogeneities due to distribution of attention. Claiming that all engender illusion would make most perception illusory.
We are considering the question of whether the veridical percept of the gap is the attended one or the unattended one. But is there a well formed question here? Is it endogenous attention that counts? Or exogenous attention? We are talking about spatial attention but what if feature-based or object based attention goes counter to spatial attention? That is, one can be attending to a place but also to a property that is instantiated in another place. And is it the absolute or relative value of spatial attention that matters? That is, is it some absolute attentional value or is it the most attended place that is seen veridically? Talking on a fake cell phone drains away spatial attention, causing the subjects to miss seeing objects in the centers of their visual fields (Scholl et al. 2003). (Scholl et al. used a fake cell phone to avoid the unnecessary source of variability of features of the responses from the other end of the line.) If it is absolute value that counts then when talking on a fake cell phone (and presumably a real one too), all vision would be illusory. That is a conclusion that we would have to have some very good reason to accept.[14]
One caution: I am speaking oversimply in a number of respects in asking whether attention engenders veridicality or illusion. I mentioned the issue of whether veridicality is graded. And there is an independent issue of relativity to what property one is talking about. In the experiment pictured in Figure 5, the attended percept is certainly more likely to be veridical in respect of which side of the square the gap is on. And in the experiment pictured in Figure 6, the comparative percept—that is, the percept of the comparative size between the right and the left is distorted by attention to one side and improved by attention to the fixation point. So veridicality is certainly affected by attention—though in different ways for different properties. The question I am asking about gap size is whether a single gap is perceived more—or on the contrary, less— veridically if it is attended. More generally, there are certain properties—I have mentioned size, contrast, color saturation and others—for which attention to an individual item changes appearance of that property. Which is more veridical, the pre-change or post-change appearance?
One might think that there is a simple way to get at the issue of whether attention magnifies, illusorily, for the purposes of getting information, or whether attention makes things look more as they really are. You could just ask people how contrasty a patch is or how big a gap is and then consider whether those answers correspond better to reality when perception is attentive or inattentive. But the human ability to make such absolute judgments for at least some relevant dimensions is remarkably poor, certainly orders of magnitude worse than our ability to discriminate stimuli (Chirimuuta & Tolhurst 2005a). In particular, the uncertainty of absolute identification (absolute in the sense of the ability to say what the contrast is in percentage terms) is far larger than the effects of attention. Even if there were some sort of statistical advantage or disadvantage to conditions of attention in estimating contrast or gap size one would have to ask whether the advantage could be ascribed to better perception or to better inference from a percept that did not differ in veridicality.
I will assume in what follows that attended and unattended perception can both be veridical. Considerations of the same sort mentioned here also apply to the veridicality of perception in both the upper and lower visual field.