Bartek Chomanski (University of Miami)
1. Seeing “What” without Seeing “Where”
Consider the scene before your eyes. There’s probably a computer screen you see right in front of you; a coffee cup to your right, within easy reach; a wall behind the computer screen; a stack of papers to grade on your left. In a typical visual experience, you see objects at particular, egocentrically specified locations in external space. But, for broadly Kantian reasons, this (seeing objects at locations in space) doesn’t seem like a merely contingent feature of our experience. It appears unimaginable, perhaps even inconceivable, that we could ever visually experience objects without experiencing them as occupying space at all. Yet, this position has recently come under sustained attack by John Schwenkler (2012). Schwenkler enlists the results of some experiments carried out on a patient with Balint’s syndrome (Friedman-Hill, Robertson, & Treisman, 1995) to show that they constitute a counterexample to the claim that the experience of space is necessary for visual spatial awareness of objects.
One of the best-studied patients with Balint’s syndrome is known in the literature as RM. RM’s performance in various experiments seems to suggest that his ability to perceive spatial objects does not require the ability to perceive “absolute space” (Schwenkler, 2012). In particular, RM seems unable to locate the objects he’s perceiving anywhere in space, and yet, he’s able to accurately identify the objects’ shapes. This has led Schwenkler to postulate that RM doesn’t perceive the objects as in space at all.
In this paper I argue that Schwenkler’s conclusion is premature. I sketch an alternative explanation of the data which preserves the necessity claim and which can account for a variety of RM’s results reported in other experiments, beside the one that Schwenkler appeals to. Craig French (2015, forthcoming) has also recently offered alternative explanations of RM’s results that don’t necessitate Schwenkler’s conclusion. My interpretation appears to differ from French’s (2015), and can be read as a further development of his (forthcoming). In Section 2, I present Schwenkler’s argument; in Section 3 I interpret RM’s results in terms of indeterminate spatial content of experience and show that this interpretation doesn’t require Shcwenkler’s radical conclusion. In Section 4 I review other experimental results on RM and suggest that they appear to support my interpretation over Schwenkler’s.
2. Does RM fail to Perceive Space?
The argument proposed by Schwenkler seems to be premised on the idea that RM fails to perceive objects’ locations and orientations, but he does perceive the objects themselves. Schwenkler writes:
RM had suffered a pair of strokes that resulted in severe damage to his posterior parietal cortex, and as a consequence was unable to localize the objects he saw. For example, when shown a display with a target at one of five locations along the vertical or horizontal meridians and told to report whether the target’s position was up, down, or center (in the vertical blocks) or right, left, or center (in the horizontal ones), RM averaged only 70% correct across all conditions. Similarly, when instructed to judge the relative position – left or right in one block, up or down in the other – of an „X‟ with respect to that of an „O‟ that was also presented on the screen, RM was correct approximately 50% of the time, a performance no better than chance (Friedman-Hill et al., 1995). RM could detect the target stimuli well enough; he just couldn’t tell where they were, either on the screen itself or with respect to other things.
Moreover, RM explicitly denies awareness of the objects’ locations. He is also unaware of their orientations. Schwenkler concludes that RM did not, at the time of the experiment, perceive the objects as arranged in a larger spatial framework: “his experience was of a wholly unimaginable sort: not of oriented spaces that were cut off at the boundaries of the things that occupied them, but of things without spatial locations or orientations. RM perceived shapes that did not appear to be in space at all.” If Schwenkler is correct, then RM’s experience falsifies the thesis that experience of spatial objects requires a representation of space as some sort of overarching framework.
However, the conclusion that Schwenkler draws does not actually follow. The inability to perceive where the objects are, or how they are oriented does not entail the lack of awareness of these objects as being within an overarching spatial framework. RM’s results are consistent with the view that he does perceive objects as in space, but the kind of information about their location he is able to consciously access is severely limited. In the next section I outline the interpretative framework that I think captures RM’s results reported by Schwenkler but does not entail his radical conclusion. After that, I offer some tentative reasons why this interpretation should be preferred.
3. Determinacy and Spatial Vision
My interpretation of RM’s predicament is simple. On my view, rather than failing to experience the object as in space at all, RM’s experience of the objects’ spatial locations within the larger space is indeterminate with respect to most spatial relations.
Indeterminacy in the content of visual experience is not at all an alien idea. It is appealed to when explaining how our peripheral vision represents the environment (e.g. Nanay (2015), Stazicker (2011)), and when explaining what the experience of Type-2 Blindsight patients is like (Brogaard (2015)). Furthermore, it appears that it is not odd to ascribe some indeterminacy to the spatial content of location experiences in ordinary conscious vision. Consider the case described by Susanna Siegel (2006):
Consider two experiences of seeing a rabbit, in both of which a rabbit looks to be in a certain direction and at least distance D away from the speaker [sic]. Let us suppose that there really is a rabbit (that looks the way the experience characterizes it) in that direction and at that distance away, but that, in one case, the rabbit is at L1, whereas in the other, it is just slightly to the left, at L2. Now, if the rabbit is far enough away, it seems plausible to suppose that these experiences could be phenomenally indistinguishable from one another. The question then arises whether either is falsidical with respect to location. If we hold constant everything else about the two situations besides the location of the rabbit seen, then it seems implausible to classify one as falsidical with respect to location and the other not. If both experiences are veridical, then the experience will be indeterminate with respect to whether the rabbit is at L1 or L2.
In general, Siegel’s point can be expressed thus: if two experiences of objects which are in fact at different locations are phenomenally indistinguishable with respect to the objects’ experienced location, then the experience is indeterminate with respect to location. Siegel shows that it’s plausible that small-scale indeterminacy is involved in ordinary vision. But we can put the indeterminacy described by Siegel to further use. Consider an auditory experience of the following sort: You have your eyes closed and you hear two people talking. The spatial content of your auditory experience doesn’t specify the spatial relation between the two voices’ locations. You can’t tell who’s to the right of whom. That is, you seem to perceive two objects without being able to perceive how they’re spatially related to one another, what their orientation is, etc. Your experience is wholly indeterminate as to how the two sources of sound are aligned on the left-right axis. That is, it is wholly indeterminate with respect to some spatial relation. If you were asked where the sources of the sounds you’re hearing are, with respect to one another, you’d be at a loss. But it wouldn’t thereby follow that you would experience the voices as not in space at all.
The same indeterminacy doesn’t arise for normal vision, however. Normally, when you see two objects, you experience where they are with respect to one another on the left-right axis.
So, it looks like some experiences are better are representing space (and at representing objects’ locations in particular) than others: the better (at representing space) an experience is, the more determinate spatial content it has. Vision is typically better than audition in that respect. It represents spatial properties and relations in a more determinate way than audition.
I propose that we take RM’s experience to involve just such indeterminacy extended to most (if not all) ways of visually experiencing spatial locations of objects. As the results that Schwenkler appeals to suggest, RM’s experiences of X’s and O’s are (sometimes) phenomenally indistinguishable, even though sometimes X is to the left of the O, and other times it’s to the right. Hence, it makes sense to say that RM sees those objects as occupying indeterminate spatial locations. Since the indeterminacy with respect to the experienced location doesn’t preclude seeing objects as in space, RM’s experience is not an experience of objects as not in space at all. Hence, Schwenkler’s conclusion is not mandatory.
To make more concrete the idea that complete indeterminacy in the awareness of location doesn’t preclude awareness of the object as in space, consider the following story: If all I know about the location of a bear is that it’s outside (that’s all I heard on the radio, say), and then you’d ask me where the bear is, I could reasonably say that I don’t know, especially if I assumed you meant a specific spatial location, or a specific relation to some other place, e.g. our favorite spot in the woods, or even with respect to my own point of view or your body or whatever. Further, if I kept switching radio stations only to find out from their reports that the bear is outside (perhaps each formulated in a different way, or sometimes in a different language), and you kept asking me for (what I’d assume would be) its precise location or orientation, I’d continue denying knowledge of that. It would not, however, mean that I’m not aware of the bear as in space at all. My representation of the bear’s spatial relations would be impoverished but not nonexistent. I would be aware of the bear as occupying an indeterminate location. The data Schwenkler cites (RM’s poor results when determining the locations of objects) is consistent with RM’s possessing such an impoverished capacity to represent space as such.
I take it that anyone in RM’s situation would be confused by not being able to perceive the precise location of objects one used to be able to localize precisely. This inability would presumably lead some rational persons to make reports that RM is quoted as making (“I can’t see where it is”); it would, it seems, make a rational person particularly reluctant to guess the location of the stimulus.
Sure, RM’s representation of space is severely impaired. But, arguably, it could be that it’s impaired in roughly the way my representation of the space surrounding the bear is. In particular, RM can experience the objects as occupying determinable locations, without perceiving determinate locations, just like I know that the bear is at some determinable location without knowing where it is. The difference is, of course, that I’m learning about the bear through testimony, and RM is learning about the object(s) through sensory experience. But that doesn’t seem to show that what (the kind of information) we learn has to be different.
In other words, an alternative interpretation of RM’s case is that he has a conscious representation of objects (perhaps serially rather than simultaneously) as occupying determinable, but not determined, locations in space, perhaps against some undifferentiated background. Hence, RM’s experience is not a counterexample to the thesis Schwenkler sets out to refute. It’s possible to represent an object as in space without representing its location in a determinate way. The inability to localize an object in space is not sufficient to show that the object is not experienced as occupying space.
Overall, the experimental results and the patient’s reports Schwenkler appeals to are consistent with the idea that RM’s visual experience might represent location broadly enough that little information useful to do well in the experimental paradigm could be extracted from this representation. Schwenkler seems to assume that to represent something visually as in space, we need to represent its precise location. But this isn’t a necessary truth. Spatial representation is varied. The confusing nature of RM’s reports might reflect the unusual poverty of his visual experience of space (especially visual experience of location), rather than the complete absence of the representation of space as such.
3.1 Lack of Orientation Awareness Does not Entail Lack of Awareness of Space
Above I discussed the first reason one might be reluctant to ascribe conscious representation of space to RM, namely his inability to localize objects. In this section, I discuss the second reason why RM is taken to lack visual awareness of space: his inability to get the orientation of seen objects right. RM is able to recognize objects (such as letters or short words); hence, he is able to represent fairly detailed intrinsic spatial relations (those holding between parts of the object) without being able to recognize the objects’ orientations. This leads Schwenkler to conclude that RM doesn’t perceive the objects as in space at all.
However, correctly identifying a complex object does not appear to require correctly identifying its orientation even in ordinary experience. Ordinary humans, as it turns out, are fairly good at categorizing pictures (telling what the picture is a picture of) even when the pictures are presented to far peripheral vision. E.g. subjects performed way above chance even for stimuli appearing at eccentricity 51.5° when the task was to categorize the target stimulus as belonging to a certain kind, e.g. animal (Thorpe, Gegenfurtner, Fabre‐Thorpe, & Buelthoff, 2001). There is thus a good reason to believe that we’re able to recognize the objects even in such unfavorable conditions.
Yet, results of a different experiment, ran on visually unimpaired subjects (Sally & Gurnsey, 2003) provide evidence that, when an object is presented to peripheral vision, we are not very good at determining its orientation. In Sally’s and Gurnsey’s experiment, the subjects’ judgments of the spatial orientation of the stimulus were poor (at chance) when the stimulus was removed from foveal vision even by as little as 8°. This suggests that the ability to correctly categorize a complex stimulus and the ability to correctly report stimulus orientation could come apart even in ordinary (peripheral) vision. But it doesn’t seem right to say that the stimuli that are correctly categorized are experienced as not in space at all. Hence, RM’s results in which he was able to correctly identify objects without being able to correctly determine their orientation don’t show that he perceived the complex objects as not in space.
Consequently, it looks like RM’s results do not require us to accept Schwenkler’s conclusion. RM is able to perceive space.
4. Other Experiments Suggest that RM is Able to Perceive Space
Additionally, Schwenkler’s interpretation of what RM’s experience is like is inconsistent with a range of other experimental results and observations made on RM. Firstly, Robertson et al. (1997) report that even though RM did poorly in identifying which of the array of objects was in motion, he nonetheless “frequently reported seeing motion in displays that included no moving objects … He would complain that the stationary letter whose color or identity he was naming was drifting about on the screen.” Now, it does seem impossible to experience the motion of an object without experiencing a change of the object’s position relative to some reference frame. RM couldn’t have experienced the letters moving, had he not experienced them as changing locations. It is difficult to see how such an experience could be accounted for while maintaining that RM is unable to see the moving object as in space.
Another set of results that appears inconsistent with Schwenkler’s view is the data on reaching in RM (Baylis & Baylis, 2001). In the experiments, RM is asked to reach to one of the twelve locations at which a point of light appears. He is better than chance, though his results still are low for the task that’s relatively easy. The results of these experiments suggest that RM’s grasp of locations in egocentric space is fairly good. But being able to locate an object in egocentric space seems to imply some ability to experience the object in relation to oneself. And hence, as occupying the same spatial framework.
However, care should be taken when drawing conclusions about spatial perception from the tasks requiring the subject to perform an action. Given the empirical discoveries concerning the two visual systems hypothesis (Goodale & Milner, 1992), the reaching data should not be used as evidence for what kind of experience RM is a subject of. This is because, roughly speaking, the system guiding visual action and that responsible for conscious vision function largely independently of one another, and the results from experiments examining visually-guided action should not be used to form conclusions about visual experience (or, more generally, visual perception).
Interestingly, Baylis & Baylis have also examined RM’s ability to report on perceived stimuli without engaging in reaching behavior. In one experiment, RM was required to merely name (without reaching) the location at which the stimulus appeared (in egocentric terms) and, again, his errors, though multiple and more frequent than in the reaching task, are mostly mistakes concerning depth (RM gets the direction—left, right, center—correct much more often than chance, but he frequently seems incapable of judging correctly whether the objects are nearer or farther away from him). Mistakes in direction but not the depth of the stimulus were rarer, mistakes in both, rarer still. There then seems to be a good case for thinking that RM perceives the object at least in relation to himself, and even that, in part, the perception is accurate (after all, he was above chance, though not as good as regular perceivers, in determining whether the object appeared on the left or right). There is, then, at least an implicit representation of one object (himself) spatially related to another (images on the screen). And for this, it seems that representing these objects as occupying a single spatial framework is required.
Similarly, another set of experiments conducted on RM by Phan et al. (2000) suggests, in the experimenters’ own words, that RM “has some ability to localize stimuli with respect his own body” and that he is able to use an egocentric frame of reference to do so. This, in turn, suggests that RM can experience objects as in a place that’s distinct from the one his body occupies. Plus, the use of an egocentric frame of reference implies some ability to represent objects as occupying the same space as his body does.
All these results point to the idea that RM must have some representation of the spatial locations of the stimuli he sees, albeit each suggest some impoverished representation. But Schwenkler’s claim concerns different results in a different task. So how does the above relate to the experimental results that Schwenkler appeals to? Why should the experiments just reported have any bearing on Schwenkler’s interpretation?
It might be said that the experiments Schwenkler bases his interpretation on were made earlier than the ones conducted by Baylis & Baylis and Phan et al. RM’s condition may have improved by the time he was examined the next time. Or, it might be that the nature of the tasks was different enough that the experiences evoked were correspondingly different. These points are well-taken. It seems to me, however, that one reason that may be offered to offset the strength of these points is that we may explain all the experimental data (i.e. those appealed to by Schwenkler, and those related above) by postulating impaired, but not non-existent representation of space. In this way, RM’s improvement over time would be one of degree (increasingly more detailed representation of spatial location) rather than in kind (some representation of space vs. no representation of space). We would have to give divergent explanations of the two sets of results, if we work along the lines that Schwenkler recommends. A (more) unitary explanation seems better than a non-unitary explanation. A less radical explanation seems better than a more radical one.
One could also reply that the interpretation I’m offering makes RM’s experience impossible to imagine. But this makes it no worse off than Schwenkler’s who says, after all, that RM’s experience is not the kind we could picture to ourselves at all.
Experimental results gathered from RM threaten to undermine the Kantian thesis that visual awareness of space is required for visual awareness of objects. I have offered a sketch of an alternative interpretation of these results. My interpretation seems able to explain the data taken to refute the thesis, plus it enables us to explain other results of experiments on RM in terms of gradual improvement of his perceptual capacities, rather than as a radical improvement in kind. The existence of an alternative interpretation shows that Schwenkler’s radical claim concerning the nature of spatial representation is not mandatory.
Baylis, G. C., & Baylis, L. L. (2001). Visually misguided reaching in Balint’s syndrome. Neuropsychologia, 39(8), 865-875.
Brogaard, B. (2015). Type 2 blindsight and the nature of visual experience. Consciousness and Cognition, 32, 92-103.
Foley, R. (2015). The case for characterising type-2 blindsight as a genuinely visual phenomenon. Consciousness and Cognition, 32, 56-67.
French, C. (2015). Bálint’s Syndrome and the Structure of Visual Experience. Retrieved from http://craigafrench.github.io/assets/CFBalintsPaper.pdf
French, C. (forthcoming). Object seeing and spatial perception. In F. Dorsch, F. Macpherson & M. Nida-Rümelin (Eds.), Phenomenal Presence: Oxford University Press.
Friedman-Hill, S., Robertson, L. C., & Treisman, A. (1995). Parietal contributions to visual feature-binding: Evidence from a patient with bilateral lesions. Science, 269, 853-855.
Goodale, M. A., & Milner, A. D. (1992). Separate visual pathways for perception and action. Trends in Neurosciences, 15(1), 20-25.
Lee, G. (2014). Experiences and their parts. In D. Bennett & C. Hill (Eds.), Sensory integration and the unity of consciousness (pp. 287-322). Cambridge, Mas.: The MIT Press.
Nanay, B. (2015). Perceptual content and the content of mental imagery. Philosophical Studies, 172(7), 1723-1736.
Robertson, L. C. (2003). Binding, spatial attention and perceptual awareness. Nat Rev Neurosci, 4(2), 93-102.
Robertson, L. C., Treisman, A., Friedman-Hill, S., & Grabowecky, M. (1997). The interaction of spatial and object pathways: Evidence from Balint’s syndrome. Journal of Cognitive Neuroscience, 9(3), 295-317.
Sally, S. L., & Gurnsey, R. (2003). Orientation discrimination in foveal and extra-foveal vision: Effects of stimulus bandwidth and contrast. Vision Research, 43(12), 1375-1385.
Schwenkler, J. (2012). Does visual spatial awareness require the visual awareness of space. Mind & Language, 27(3), 308-329.
Siegel, S. (2006). Subject and object in the contents of visual experience. Philosophical Review, 115, 355-388.
Stazicker, J. (2011). Attention, Visual Consciousness and Indeterminacy. Mind & Language, 26(2), 156-184. doi: 10.1111/j.1468-0017.2011.01414.x
Thorpe, S. J., Gegenfurtner, K. R., Fabre‐Thorpe, M., & Buelthoff, H. H. (2001). Detection of animals in natural images using far peripheral vision. European Journal of Neuroscience, 14(5), 869-876.
Weiskrantz, L. (2002). Prime-sight and blindsight. Consciousness and Cognition, 11(4), 568-581.
 Lynn Robertson’s gives the following characterization of Balint’s syndrome: “Balint’s syndrome is a neuropsychological disorder that results from damage to both parietal lobes. Clinically, it includes three main symptoms: simultanagnosia (the inability to see more than one object at a time); optic ataxia (the fixation of gaze with severe problems in voluntarily moving fixation); and optic apraxia (the inability to reach towards the correct location of perceived objects)” (2003)
 It seems possible, however, to represent motion as such, without having to represent an object moving against some background. Blindsighted subjects provide one example (Foley, 2015; Weiskrantz, 2002). Noticing motion somewhere in your peripheral vision might count as another example (Lee, 2014).
 But, to engage in pure speculation, one might perhaps think that my interpretation offers us a way of grasping imaginatively RM’s predicament thus: perhaps his experience in his entire visual field is not unlike what we experience in the periphery. While this is pure speculation, it’s interesting to note that the responses of people tasked by Thorpe et al. with categorizing images shown to their peripheral vision indicate similar unwillingness to guess as do RM’s reports.