Phenomenology and the Cognitive Sciences

, 10:481

Monstrous faces and a world transformed: Merleau-Ponty, Dolezal, and the enactive approach on vision without inversion of the retinal image


    • Department of PhilosophyNorthern Arizona University

DOI: 10.1007/s11097-011-9210-6

Cite this article as:
Bredlau, S.M. Phenom Cogn Sci (2011) 10: 481. doi:10.1007/s11097-011-9210-6


The world perceived by a person undergoing vision without inversion of the retinal image has traditionally been described as inverted. Drawing on the philosophical work of Maurice Merleau-Ponty and the empirical research of Hubert Dolezal, I argue that this description is more reflective of a representationist conception of vision than of actual visual experience. The world initially perceived in vision without inversion of the retinal image is better described as lacking in lived significance rather than inverted; vision without inversion of the retinal image affects the very content of the perceived world, including, importantly, its expressions and conducts, and not merely the orientation of this content. Moreover, I argue that the enactive approach, rather than a representationist approach, is best able to account for the perception of the world, after prolonged vision without inversion of the retinal image, as both normal and upright, yet still different from the world seen previously. Finally, in their attention to the perception of other people’s facial expressions, I argue that Merleau-Ponty and Dolezal draw out the existential significance of the enactive approach. In encountering another person, the most pressing task is generally not to observe this person’s features but, instead, to engage with this person’s expressions.


Vision without inversion of the retinal imageMaurice Merleau-PontyHubert DolezalInverted worldEnactive approach

“Nothing is more difficult,” Merleau-Ponty (1967)1 writes in the Phenomenology of Perception, “than to know precisely what we see” (italics his, 58). Yet, Merleau-Ponty argues, while knowing what the perceived world is like may not be easy, any tenable account of vision must begin from the careful description of this world. Merleau-Ponty’s assertion that our conception of vision must be grounded in the careful description of visual experience is echoed by those who, within contemporary research on consciousness and vision, advocate an enactive approach. As Varela et al. (1991), Noë (2004), and others have argued, most research on vision conceives of vision as representational; such research, therefore, tends to focus on the processes that supposedly produe the perceived world while paying little attention to the way the world is actually perceived. In not attending carefully to visual experience, though, such research often gives accounts of vision that are at odds with the very visual experiences they are meant to explain.2

In the following paper, then, I will focus on a visual experience that, while central to a significant body of research on vision, has generally not received careful attention: the visual experience accompanying vision without inversion of the retinal image.3 My motivations for focusing on this experience are threefold. First, the difference between the world perceived in vision with inversion of the retinal image and the world perceived in vision without inversion of the retinal image has traditionally been characterized as a difference in orientation. In vision without inversion of the retinal image, what is seen, it is usually asserted, is an inverted world. Yet, this depiction, I will argue, is more reflective of a representationist conception of vision than the actual visual experience itself. Drawing on the philosophical work of Maurice Merleau-Ponty and the empirical research of Hubert Dolezal, I will argue that the world initially perceived in vision without inversion of the retinal image should be described primarily as lacking in lived significance rather than as inverted. For those who undergo vision without inversion of the retinal image, it is the expressions and conduct of their perceived world, and not the orientation of this world, that is affected.

Second, insofar as the careful description of perceptual experience reveals that the world initially perceived in vision without inversion of the retinal image is lacking in significance, this description has implications, not only for the traditional depiction of this perceived world as inverted, but also for the representationist conception of vision that underlies this traditional depiction. The experience of vision without inversion of the retinal image, I will argue, offers support for an enactive, as opposed to a representationist, approach to vision. Vision must be understood as actively presenting, rather than passively re-presenting, the world.

Third, Merleau-Ponty’s and Dolezal’s discussions of vision without inversion of the retinal image, unlike many other discussions of vision without inversion of the retinal image, and indeed, unlike many discussions of normal vision, include descriptions of people and not just things. These descriptions, I will argue, can supplement the enactive-embodied approach to vision by revealing a critically important interpersonal dimension to visual experience. In the visual experiences accompanying vision without inversion of the retinal image, the lack of recognizable facial expressions and bodily comportments draws attention not only to the centrality of other people in everyday visual experience but also to our active involvement in the world we perceive.

Vision without inversion of the retinal image and Stratton

Descartes′ (1965) Optics contains perhaps the best-known report of an inverted retinal image; in “Discourse Five,” Descartes describes “taking the eye of a newly deceased man, or, for want of that, of an ox or some other large animal” (91). After preparing this eye appropriately, he writes, one can look at the back of eye and “see there, not perhaps without admiration and pleasure, a picture which will represent in natural perspective all the objects which be outside of it” (93). In this picture, however, “its parts are reversed, i.e., in a position quite the opposite to that of the objects” (97).

Following such observations, the retinal image was incorporated into the basic account of vision as an intermediary between the perceived world and the physical world that was considered its cause.4 Thus, as almost any introduction to vision continues to assert, vision begins when light from the physical world passes through the pupil at the front of the eye and forms an image of this world on the retina at the back of the eye;5 the retinal image is, with respect to the physical world, inverted, that is, if the horizontal and vertical planes of the perceived world are taken as representations of horizontal and vertical planes that are given in the physical world, these planes have, in the retinal image, been rotated 180°. Since the perceived world that accompanies the inverted retinal image is understood as having the same orientation as the physical world, normal vision is understood as vision with inversion of the retinal image.

A person can, however, be given goggles fitted with special lenses that negate, or as Merleau-Ponty puts it “correct,” this inversion (244). This “correction” is, of course, not actually the correction of a defect but, rather, the creation of an abnormality, purposefully induced in an experimental subject in order to study vision. For the person who wears these inverting lenses, light from sections of the physical world that has ordinarily been projected onto the inferior half of the retina is focused on superior half, and vice versa; in these cases, referred to as cases of vision without retinal inversion, any retinal image, rather than being inverted, would have the same orientation as the physical world.

The most influential research on vision without inversion of the retinal image was conducted by George M. Stratton (1896, 1897a, b) who performed two experiments, the first in 1896, the second in 1897. The purpose of these experiments, he writes, “…was to throw some light, if possible, on the correctness of this assumption. Is the inverted image a necessary condition of our seeing things in an upright position?” (Stratton 1896; 613). He wore inverting lenses in the first study for approximately 21.5 waking hours (Stratton 1896; 613) and, in the second study, for approximately 87 waking hours (Stratton 1897a; 343).6

Summarized briefly, Stratton reports that, at the beginning of each experiment, the world he perceives is altered and his ability to function normally severely compromised. In general, he describes the alteration to his perceived world in terms of orientation; his perceived world is, he states, compared with the world he previously perceived, rotated, as a whole, 180°. On the first day of the first experiment, for example, Stratton writes, “The course of experience was something as follows: All images at first appeared to be inverted; the room and all in it seemed upside down....All movements of the body at this time were awkward, uncertain, and full of surprises” (Stratton 1896; 614). While not all of the alterations to his perceived world that Stratton describes can be characterized as alterations of orientation, Stratton’s discussion of adaption is based on those that are. In other words, Merleau-Ponty’s summary of Stratton’s work in the Phenomenology of Perception does a good job of presenting Stratton’s own understanding of his experience, even if Stratton’s understanding is not always adequate to his experience.

Yet, as Merleau-Ponty points out in his discussion of Stratton, the contents of Stratton’s perceptual experience could not, by themselves, serve to identify his perceived world as inverted; whether one’s vision is with or without inversion of the retinal image, for example, if one looks at one’s feet, they will be adjacent to the ground and not to the sky. The mere perception of any particular thing within the world, therefore, cannot provide a fixed direction by which to judge whether the perceived world, as a whole, is upright or upside down (246). Stratton’s identification of his perceived world as inverted, then, relies on a standard that his own perceptual experience does not provide. Rather than a true description of his experience, Statton, instead, gives his pre-conceived ideas about this experience. More specifically, I would argue, Stratton imports Descartes′ account of experiencing an inverted retinal image into his own account of experiencing vision without inversion of the retinal image.

The perceived world of a person undergoing vision without inversion of the retinal image must be distinguished from the perceived world of a person who, like Descartes, observes a retinal image. The perceived world of a person who observes a retinal image is not limited to that of a retinal image; she is able to compare the orientation of the retinal image to the orientation of a larger situation whose directions have already been established. When a person observing a retinal image glances toward her feet, for example, she looks away from an altered image of the world and returns to an unaltered world. A person undergoing vision without retinal inversion, though, is denied the experience available to Descartes; when she glances toward her feet, she remains within an altered world. Her perception of her feet is precisely not embedded within a larger, unaltered perceptual experience. When discussing perceptual experiences in relation to retinal images, then, Stratton seems to forget that a retinal image is never the entirety of one’s perceived world and only rarely part of one’s perceived world. It is uncertain, therefore, what he means when he states that the perceived world of a person undergoing vision without retinal inversion is inverted, since her perceived world cannot be inverted in the same sense as a retinal image.

Stratton’s implicit conception of the perceptual experience accompanying vision without inversion of the retinal image as similar to the experience described by Descartes is not surprising. Beginning with the observation of retinal images by a single perceiver like Descartes, many accounts of vision posit the existence of such retinal images for all perceivers and understand perceptual experience in terms of these retinal images. Dennett (1991), by referring to accounts like this as versions of “Cartesian materialism,” even specifically links them to Descartes′ observations (107). Yet, as many, including some who research vision without inversion of the retinal image, have recognized, accounts of vision that posit the existence of a retinal image for all perceivers are problematic.7 Positing a retinal image that must then itself be perceived simply moves the investigation of perception farther back into the body; Noë (2004) argues, “It is incoherent to suppose that seeing an object depends on the resemblance between a picture in the eye and the object, for that presupposes that there is, as it were, someone inside the head who perceived the resemblance” (44).

Now, insofar as the present discussion of vision without inversion of the retinal image is focused on the perceptual experiences accompanying vision with inversion of the retinal image, the criticisms just reviewed, while well-deserved, might not seem particularly pertinent. That is, one might think that one could reject Stratton’s account of how a non-inverted retinal image produces an inverted perceived world while still accepting his description of the perceived world as initially inverted. In other words, it might seem as if references to retinal images could be eliminated without much effect on the present discussion.

As Thompson, Noë, and Pessoa (1999) and Noë (2004) have noted, however, the positing of a retinal image can affect not simply one’s understanding of how a particular perceptual experience is produced but also one’s understanding of what this perceptual experience is. References to retinal images, Noë (2004) argues, are just symptoms of a larger tendency to assume, “that perception is a process in the brain whereby the perceptual system constructs an internal representation of the world” (2).8 Indeed, Thompson, Noë, and Pessoa (1999) argue that Dennett does just this in his discussion of perceptual completion;9 although Dennett rejects analytic isomorphism, he neglects “precisely what phenomenologists routinely practice—the careful conceptual and phenomenological study of lived experience…As a result, Dennett seriously distorts the conceptual and phenomenological character of visual perception” (168). When perception is conceived as representational, perceptual experience itself is left unexamined on the assumption that one already knows what it is like. Thus, one might acknowledge all the criticisms of the retinal image reviewed above and still continue to misunderstand vision without inversion of the retinal image, believing that the perceptual experience that accompanies it, while not the inverted representation of a non-inverted retinal image, was the inverted representation of a non-inverted world.

Merleau-Ponty and vision without inversion of part of the retinal image

To better understand the perceptual experience that accompanies vision without inversion of the retinal image, then, I will now turn to Merleau-Ponty. Merleau-Ponty explicitly recognizes that what we see may be quite different than what we assume we see; he works to give an account of perception that is grounded in careful description of actual perceptual experience rather than in pre-conceived ideas about this experience.10 Just a few pages after discussing Stratton’s work, Merleau-Ponty describes what he perceives when standing before a face that has been turned upside down, a case of what I will call vision without inversion of part of the retinal image. He writes:

If someone is lying on a bed, and I look at him from the head of the bed, the face is for a moment normal. It is true that the features are in a way disarranged, and I have some difficulty in realizing that the smile is a smile, but I feel that I could, if I wanted, walk round the bed, and I seem to see through the eyes of a spectator standing at the foot of the bed. If the spectacle is protracted, it suddenly changes its appearance: the face takes on an utterly unnatural aspect, its expressions become terrifying [le visage devient monstrueux, ses expressions effrayantes], and the eyelashes and eyebrows assume an air of materiality such as I have never seen in them. For the first time, I really see the inverted face as if this were its ‘natural’ position: in front of me I have a pointed, hairless head with a red, teeth-filled orifice in the forehead and, where the mouth ought to be, two moving orbs edged with glistening hairs and underlined with stiff brushes. (252)

Now, given what Stratton’s account would have us expect, there should be an appreciable difference in orientation between the face Merleau-Ponty sees as he stands at the head of the bed and the face he would see if he stood at the foot of the bed. Yet, at the beginning of his description, Merleau-Ponty primarily focuses on denying any such appreciable difference. The normality of face he sees is, admittedly, qualified, and there are differences between what he sees from the head of the bed as opposed to what he would see from the foot of the bed. None of these differences, though, are characterized in terms of orientation.

Merleau-Ponty’s perception of the face as basically unaffected by his position does not, however, persist. After some time, Merleau-Ponty reports that, while what he sees may still be described as a face, it is completely different, not only from the face he saw moments earlier but also from any face he has ever seen before. What Merleau-Ponty perceives is not a face that, except for being inverted, is the same as the face he usually sees; rather, what he perceives, and here the English translation does not fully capture the significance of Merleau-Ponty’s description, is monstrous. In other words, what he perceives is strange not because it does not usually have this orientation but because it is, quite literally, not the face of a human. Thus, although Merleau-Ponty does affirm an appreciable difference between the face he perceives in vision with inversion of the retinal image and vision without inversion of part of the retinal image, it is again not the appreciable difference in orientation predicted by the Stratton’s account. Merleau-Ponty never describes the face he perceives as inverted. Rather than being an oversight, this omission is, I think, the crucial point of the description. As Merleau-Ponty writes earlier in the Phenomenology of Perception, “A landscape or newspaper seen upside down are said to represent our original view of them....It is not asked why differently arranged impressions make the newspaper unreadable or the landscape unrecognizable” (19).

Like a landscape or a newspaper seen upside down, Merleau-Ponty reveals that a face seen upside down has been deprived of its lived significance (253). In place of an abstract and formal characterization of the face as ‘inverted,’ Merleau-Ponty offers an embodied and emotional description. The difference between vision with and without inversion of part of the retinal image, he argues, cannot be adequately described geometrically. The very content of his perceived world, including the face’s features and, importantly, its expressions, is affected and not just the orientation of this content. Moreover, Merleau-Ponty understands this alteration in the perceived world as reflecting a disruption in the perceiver’s concrete and embodied engagement with the world as opposed to an inversion of her representation of the world; he writes, “Its being as an object is, therefore, not a being-for-the-thinking subject, but a being-for-the-gaze which meets it at a certain angle, and otherwise fails to recognize it” (253). Rather than conceiving of vision as, to use Ryle’s (1984) terms, a matter of ‘knowing that,’ Merleau-Ponty conceives of vision as a matter of ‘knowing how.’ A face turned upside down is experienced as a terrifying monster because one no longer knows how to approach it and respond to it competently.

Given the deeply entrenched understanding of the world perceived in vision without inversion of the retinal image as inverted, Merleau-Ponty’s descriptions may not yet seem very convincing; they are, however, well-supported by research on facial inversion. Psychologists have long recognized that inverting a face affects far more than its perceived orientation.11 Köhler (1940) notes that photographs of known or unknown persons “show a conspicuous change of appearance” when turned upside down (25). Yin (1969) states: “It is a well-known fact that pictures of human faces, when viewed upside-down, are extremely difficult to recognize” (141). Some of the earliest experiments surrounding this phenomena, Yin writes, include a 1963 study by Brooks and Goldstein, which “showed that recognition of inverted faces is worse than that of upright faces when children are asked to identify snapshots of their classmates” as well as a 1965 study by Goldstein and a 1967 study by Hochberg and Galper, which showed “That memory for inverted faces is poorer than memory for upright faces among adults” (141). P. Thompson (1980) has demonstrated that distortions which are immediately apparent in upright faces are hardly noticed when these same faces are inverted (483).12

In revealing that vision without inversion of part of the retinal image affects the lived significance, rather than the orientation, of the perceived world, Merleau-Ponty’s descriptions challenge not only the previous accounts of this perceived world as inverted, they also challenge the representationist conception of vision that motivated this depiction. In Merleau-Ponty’s case, as with Stratton’s, the physical world itself is not altered; all that is altered is the physical world’s orientation with respect to the perceiver’s eyes. That the percieved world is altered in ways other than that of orientation, therefore, despite no corresponding alterations in the physical world, offers a challenge to a representationist conception of vision.

Of course, a representationist could argue that the novelty of the experience accompanying vision without inversion of all or part of the retinal image is simply due to the novelty of inverted representations of the world. That is, Merleau-Ponty’s descriptions of the face as ‘monstrous’ or the landscape as ‘unrecognizable,’ rather than indicating a disruption in the perceiver’s concrete and embodied engagement with the world, instead, simply indicate one’s unfamiliarity with an inverted representation of the same situation. Once one has become familiar with this inverted representation, the representationist might argue, the world’s monstrosity or unrecognizability would disappear. I will address this objection at greater length in the section “Dolezal on vision without inversion of the retinal image.” For now, however, I simply want to stress that accounting for the experience in terms of an unfamiliar orientation of a familiar object is not an adequate description of the experience. What is seen is not a familiar face in an unfamiliar orientation but an unfamiliar, indeed, not even human, face.

Perhaps more troubling for Merleau-Ponty’s argument, however, are Stratton’s descriptions of apparently the very experiences of inversion that Merleau-Ponty denies. Since phenomenology must ground its claims in descriptions of lived experience, Stratton’s own descriptions of the perceived world as inverted would seem capable of immediately discrediting Merleau-Ponty’s claims.

Yet, these difficulties raised by Stratton’s descriptions are not nearly as insoluble as they might initially seem. While it is certainly true that Stratton does often describe the change to his perceived world in terms of orientation, Stratton does, nonetheless, use other terms, and, many times, Stratton’s own descriptions are at odds with his explanations. There are indications, in other words, that Stratton’s perceived world, when compared with the world he perceived previously, is not adequately described as inverted. For example, on the first day of the first experiment, Stratton (1896) writes that “although all these images were clear and definite, they did not at first seem to be real things, like the things we see in normal vision, but they seemed misplaced, false, or illusory images between the observer and the objects or things themselves” (613). On the first day of the second experiment, though, Stratton (1897a) writes that “Objects of sight had more reality in them—had more the character of ‘things,’ and less that of phantasms—than when the earlier trial began” (345). He also describes sitting “…for some time watching a blazing open fire, without seeing that one of the logs had rolled far out on the hearth and was filling the room with smoke. Not until I caught the odor of the smoke, and cast about for the cause, did I notice what had occurred” (Stratton 1897a; 345).

Descriptions of the change to his perceived world in terms other than those of orientation continue beyond the first days of both experiments; on the seventh day of the second experiment, for example, Stratton (1897b) writes that “During the walk in the evening, I enjoyed the beauty of the evening scene, for the first time since the experiment began” (466). Finally, upon removing his goggles at the end of the experiment, Stratton (1897b) describes what he saw as having “a surprising, bewildering air which last for several hours. It was hardly the feeling, though, that things were upside down” (470). These discrepancies within Stratton’s descriptions, then, should offer sufficient reason for withholding any negative judgments about Merleau-Ponty’s description until after additional cases of vision without retinal inversion are examined.

Dolezal on vision without inversion of the retinal image

Since 1897, experiments involving modification of the retinal image have continued, and a number of scientists have repeated or modified Stratton’s experiments; yet these experiments, including Dolezal’s (1982), are rarely referenced in discussions of vision without inversion of the retinal image.13 I will focus on Dolezal’s work for two reasons. First, he is aware of the obstacles that can stand in the way of a careful description of perceptual experience; he acknowledges, for example, that it is very difficult “for spectacle-wearers and experimenters alike—to distinguish between their observations and inferences and that it is these confusions that encourage verbal ambiguity when one is labeling perceptions while looking through reversing or inverting spectacles” (269).14 Second, Dolezal’s work includes extensive first-person descriptions of his perceptual experiences, including those involving other people.15

Although Dolezal’s entire study took place over 5 weeks (488 waking hours) and involved six stages, up–down reversing lenses were worn only in the third stage, which lasted 15.2 days (211 waking hours). At the beginning of stage 3, Dolezal, like Stratton, reports that his perceived world is altered and his ability to function severely compromised. Unlike Stratton, however, Dolezal does not primarily describe his perceived world as being inverted or upside down. Instead, Dolezal consistently describes the world with terms that Stratton used only occasionally; he reports, “Particularly when regarding once familiar places, objects, pictures, script, facial features, and events like smiles, I applied such labels as ‘uncanny,’ ‘weird,’ ‘novel,’ ‘surrealistic,’ ‘strange,’ ‘bizarre,’ ‘uncommon,’ and ‘surprising’” (202).16

Immediately after looking through the up-down reversing lenses for the first time, Dolezal reports, “Whatever I view has an aspect of entire irreality. I am surprised that the sea does not float toward the sky. The sea has become the sky. The waves are behaving oddly. The progress of the swimmers is irreal; it is hard for me to say whether they are coming toward the shore or going out to sea” (22). When viewing an entire person immediately after putting on the lenses, he reports that, “A man or a boy walking—I cannot be sure which—appears to be a two-legged crab moving in strange fashion” (22). Moreover, Dolezal reports that he cannot, at first, identify his friends from their faces:

Recognizing once familiar persons by their facial features (rather than by hair color, mode of dress, and time, place, or circumstance of meeting) proved to be virtually impossible at first. Even though I made use of all this information and was thus doing more than facial feature recognition as such, I confused three adult women with whom I related most and who have similar facial features. (222)

Although he does not describe any specific face in detail, his comments suggest that not only were the faces he saw not familiar, they were also not unequivocally human; he refers, for example, “to the ‘strange’ character of the observed person’s face, particularly of the eyes and eyebrows” (233) and writes that, “From the first moments of gazing through the spectacles, I became aware that I was forced to pay attention to a great many novel optical transformations that were unlike anything I had ever seen, including emotion-disturbing sights” (223).

Dolezal also reports that he cannot identify most of the facial expressions of the people he sees:

I was accurate and unfailing in noticing that someone’s facial muscles were moving, but I was not always clear about the meaning of these facial expressions. Relatively gross changes, such as a broad smile, presented no problem, though the novelty of the optical transformation ‘carrying’ that smile elicited everything from surprise to disbelief on my part....More subtle facial expressions, especially those peculiar to a particular person (e.g., those conveying consternation or mild disapproval) were either missed or not understood at all. (223)

The difficulties he reports with respect to facial expressions are also experienced with respect to conducts of the entire body; for example, he reports that the “Walking of other persons…seems to be quite comparable to the information given in a non-reversing situation by people on a dance floor performing modern, exotic-sensual dances” (334). Indeed, many of Dolezal’s descriptions, even when they are not of people, indicate alterations to the conducts of the perceived world; on the first day of stage 1, for example, he writes:

I see some children batting about a tennis ball with a pingpong size racket; what is unreal is that the shadow they appear to be batting about can hardly stay in the air; it flops down all the time as if suddenly pulled down by hidden magnets; the motion of the ball is unlike anything I’ve ever seen. (203)

By putting aside the preconceptions generated by a representationist conception of vision, Merleau-Ponty and Dolezal discover that they perceive the world very differently than they were led to expect. Like Merleau-Ponty, Dolezal stresses a change in the lived significance, in the expressions and conducts of his perceived world, rather than a change in its orientation. When looking at an inverted face, what Merleau-Ponty sees is monstrous rather than human; likewise, when looking at an inverted world, what Dolezal sees is, to echo the title of his book, a world transformed. The world Dolezal perceives in vision without inversion of the retinal image has a qualitatively different content than the world he perceives in vision with inversion of the retinal image rather than a different orientation of the same content.

Thus, while one might think that the world perceived in vision without inversion of the retinal image is turned upside down in the same manner as a retinal image, such inversion is only ever a possibility for thought and is never a possibility for perception. As Merleau-Ponty writes, “For the thinking subject a face seen ‘the right way up’ and the same face seen ‘upside down’ are indistinguishable. For the subject of perception the face seen ‘upside down’ is unrecognizable” (252). Although one can certainly rotate all or part of a retinal image 180°, what one actually perceives after this action has been performed cannot be described in the same geometrical terms as the action itself.

Yet, even if the representationist conceded that the world initially perceived in vision without inversion of the retinal image is truly novel and cannot be described merely as differently oriented than the world one perceived previously, she might maintain that this experience alone does not challenge a representationist conception of vision. The unrecognizability of people and things in vision without inversion of the retinal image, the representationist could argue, is simply due to the relative novelty of an inverted representation of the world. Much psychological research on the effect of inversion on the perception of faces, for example, argues that facial recognition requires a representation of the face to be processed or encoded in specific ways.17 This processing or encoding is, however, orientation-sensitive and does not occur when representations of faces are inverted. For an inverted face to become recognizable, then, this orientation sensitivity would need to be overcome so that the inverted representation could be processed in the same way as a non-inverted representation. One might expect such a development to occur a situation like vision without inversion of the retinal image where inverted representations of faces are the rule rather than the exception. Once these inverted representations could be processed or encoded in the same way as non-inverted representations, the world would then cease to be monstrous or unrecognizable and be seen, once again, as it was seen before, save perhaps for its orientation. Thus, even if, in the experience of a monstrous face cannot be equated with the experience of an inverted face, vision can still be understood as representational.

There is, however, at least one difficulty that this representationist argument would need to overcome. Dolezal reports that the world, when seen through inverting lenses over several days, “…comes to seem ‘normal,’ ‘upright,’ ‘erect,’ but not with respect to the remembered prespectacle appearance” (265).18 Furthermore, he postulates that, “…throughout perceptual adaption, the observer is aware of the visual differences in what he or she perceives with spectacles and what he or she remembers before donning them. Operationally, no matter how long the observer adapts, he or she should continue to be able to make factually correct same–different judgments about phenomenal experiences with and without spectacles” (24). While the world ceases to be unrecognizable, this is not because it is seen as identical, save for its orientation, to the world seen before wearing inverting lenses. Rather, the world ceases to be unrecognizable because it assumes its own, unique, lived significance; Dolezal writes, “The observer must gain and establish new perceptual meanings rather than regain and reestablish previous perceptual meanings” (24). Dolezal’s remarks suggest that adapting to vision without inversion of the retinal image is rather like adjusting to life in a new country. While one’s initial experience of this new country as foreign can be replaced with an experience of it as familiar, its particular form of familiarity will remain distinct from the other places one has felt at home. Similarly, while one can learn a new language and thus come to experience as meaningful sounds that previously lacked any meaning, this language will always be distinguishable from the other languages one speaks; as Merleau-Ponty writes, “…the full meaning of a language is never translatable into another” (187, italics his).

Dolezal’s account of adapting to vision without inversion of the retinal image, then, challenges the representationist’s claim that descriptions of the world in terms other than those of orientation are simply due to the relative novelty of an inverted representation of the world. If the world were unrecognizable due to the orientation sensitivity of the specific ways that certain representations must be processed or encoded to be recognizable, then one would expect that if the world were to become recognizable, this would require that any such orientation sensitivity be overcome. If this were to happen, though, then an inverted representation of the world would become processed or encoded in the same way as a non-inverted representation, and thus the world should be seen as identical, save perhaps for orientation, to the world perceived prior to vision without inversion of the retinal image. Yet, while Dolezal, after wearing inverting lenses for several days, comes to see the world as recognizable, he does not come to see the world as either indistinguishable from or merely differently oriented than the world he saw before.

The enactive approach and vision without inversion of the retinal image

While Merleau-Ponty and Dolezal’s descriptions run contrary to the expectations of representationist conceptions of vision, they are, I would argue, consistent with the expectations of an enactive approach to vision. In contrast to both the computational–symbolic and connectionist–dynamic approaches, the enactive approach, Varela et al. (1991) argue, offers “a critique of the adequacy of the notion of representation as the Archimedes point for cognitive science” (8).19 Rather than conceiving of the interactions between the perceiver’s body and the physical world as representing internally what is already present externally, the enactive approach, Varela et al. (1991) write, conceives of these interactions as “A history of structural coupling that brings forth a world” (206). For the enactive approach, the interactions between the perceiver’s body and the physical world are understood as creative, rather than as duplicative, processes (Varela et al. 1991, 148). One cannot assume, in other words, that the world one perceives already exists as such in the physical world.20 The conception of the physical world as already determinate must be replaced with a conception of the physical world as only coming to have the determinacy that it is perceived as having through a perceiver’s interactions with it. In other words, one must understand the perceived world as meaningful rather than given, and one must understand the perceiver as embodied participant in the world rather than as a detached observer of the world.

Furthermore, the enactive account argues that vision and bodily movement, rather than being separate processes, mutually inform one another; Noë (2004) writes, “What we perceive is determined by what we do (or what we know how to do); it is determined by what we are ready to do” (1, italics his). The world a person sees is grounded in her body’s capabilities. These capabilities encompass her personal habits, her abilities, for example, to walk, manipulate certain objects competently, and speak certain languages. Thus, for example, when Merleau-Ponty is looking at an upside down face, so long as he feels that he could, if he wanted, “walk round the bed,” he stills sees a smiling, human face. It is when this possibility of movement no longer seems possible, however, that the face becomes monstrous. These capabilities also encompass her implicit, practical knowledge of what Noë (2004) and others refer to as patterns of sensorimotor contingency, the way that her own “movement gives rise to changes in stimulation” (8). These patterns of sensorimotor contingency are distinct for each sensory modality; for example, vision and hearing, Hurley and Noë (2003) write, differ in so far as when “…you approach an object, visual field flow expands, while as you withdraw, visual field flow contracts. By contrast, as you approach the source of a sound slowly, the amplitude of the auditory stimulus increases, while as you withdraw the amplitude decreases” (146). Rather than being a matter of forming representations, vision is a matter of getting a grip on the world.21 This insight is echoed by Merleau-Ponty, who writes, “To see a face…is to take a certain hold upon it, to be able to follow on its surface a certain perceptual route with its ups and downs, and one just as unrecognizable taken in reverse as the mountain up which I was so recently toiling, and down which I am now striding my way” (253). Seeing the world requires discovering patterns of sensorimotor contingency and then using them to support one’s skillful bodily interactions with the world.

By understanding that the determinacies of the perceived world as emerging out of the relation between the perceiver and the physical world rather than as existing prior to this relation, the enactive approach, in contrast to the representationist conception of vision, can, first of all, account for why vision without inversion of retinal image affects the world’s lived significance rather than its orientation. Changing the relation between perceiver and the physical world changes the patterns of sensorimotor contingency upon which prior perception of the world was grounded. As Noë (2004) writes when discussing the effect of left/right reversing, as opposed to inverting, lenses on visual experience, “When you put on the distorting lenses, the patters of dependence between movement and stimulation are altered. This alteration has the effect of abrogating sensorimotor knowledge or skill, even though there is no change in the intrinsic character of stimulation” (8). Thus, while vision without inversion of the retinal image only changes the orientation of the perceiver with respect the physical world, this change can be expected to affect much more the perceived world’s orientation. Like left/right reversing lenses, inverting lenses would be expected to substantially, if not entirely, disrupt the perceiver’s implicit, practical knowledge of patterns of sensorimotor contingency. This disruption would, in turn, be expected to disrupt any bodily activities grounded in this knowledge thereby affecting all the lived significances that were achieved through these activities. Indeed, Noë (2004) argues that left/right reversing lenses produce a kind of experiential blindness (8).

In contrast to a representationist conception of vision, the enactive approach can also make sense of Dolezal’s description of the world he perceives after wearing up-down reversing lenses for an extended period as normal and upright and yet remaining distinguishable from the world he saw before wearing the lenses. Though others have given an enactivist account of vision that has been altered by inverting lenses (O’Regan and Noë 2001) and left/right reversing lenses (Hurley and Noë 2003), the experience that emerges after prolonged wearing of such lenses has not received much attention. Yet, it is an experience that, I would argue, the enactive account is particularly able to make sense of.

In everyday vision, the specific increases and decreases in illumination that accompany specific eye movements are dependent on the relative position of perceiver and major sources of illumination, like the sun; as O’Regan and Noë (2001) write, certain “characteristics of visual exploration of objects derive from the fact that color and brightness of the light reflected from an object change in lawful ways as the object or the light source or the observer move around, as the character of the ambient light change” (942). It is these patterns of sensorimotor contingency that ground our experience of sky and land and of objects and their shadows. This relative position, while it is not unchanging, nonetheless changes in very specific ways, ways that are themselves incorporated into the patterns of sensorimotor contingency.

Wearing inverting lenses, however, dramatically changes the position of the sun, as well as other sources of illumination, relative to the perceiver such that the specific eye movements the perceiver previously made are no longer associated with the same increases and decreases in illumination. Wearing inverting lenses, therefore, places the perceiver in a situation where she cannot depend on her already established practical knowledge of patterns of sensorimotor contingency and must, instead, establish a new practical knowledge of the particular patterns of sensorimotor contingency that apply in this new situation. Insofar as the perceiver can discover these patterns, she will, once again, be able to interact with the world skillfully and will, once again, experience the world as having a lived significance. Insofar as these patterns of sensorimotor contingency remain different from the patterns that held prior to wearing inverting lenses, though, her present experience of the world can also be expected to have a different lived significance than her previous experience. In conceiving of vision as a skillful motor activity grounded on the discovery of patterns of sensorimotor contingency, then, the enactive account can also account for the persistence of a difference other than that of orientation between the world seen during prolonged vision without inversion of the retinal image and the world seen before.

Other people

Beyond offering confirmation for an enactive approach to vision, Merleau-Ponty and Dolezal also, I would argue, draw out the existential significance of such an approach. By including people in their descriptions of vision without inversion of the retinal image, Merleau-Ponty and Dolezal attend to the interpersonal relations that occupy much of our everyday life, relations that, because they are so familiar, are easily taken for granted.

When a face is upside-down, Merleau-Ponty and Dolezal note, not only are the features of the face one perceives affected, so are its expressions. While empirical research on facial inversion has focused far more on the recognition of features than on the recognition of expressions, it is generally recognized that inversion can also render facial expressions unrecognizable. Medieval painters, for example:

…suffered frustration when they attempted to convey a properly beatific attitude toward his own martyrdom on the face of St Peter for the simple reason that the latter was crucified upside-down. Michelangelo partially solved the problem by having the subject adopt a neck-wrenching posture which, while somewhat implausible, resulted in his head being more or less upright and his feelings manifestly more apparent. (Liebert 374, qtd. in Parks et al. 1985; 748)

More recently, Köhler (1940) noted that when photographs of people are inverted “They change so much that what we call facial expression disappears almost entirely” (25). McKelvie (1995) tested whether inverting a photo depicting one of seven emotions (happy, sad, fear, anger, surprise, disgust, neutral) disrupted the classification of emotional expressions. He found that subjects were considerably less able to identify six of the emotions when the photos were inverted; the exception was the emotion of ‘happy’.22

In asserting that the expressions of an inverted face are ‘terrifying’ (252), therefore, I would argue that Merleau-Ponty is not so much describing a familiar expression as a lack of any familiar expressions. To encounter this face, a face that is so different from the human face one saw previously that it is like that of a monster, is to be unable to tell whether it is menacing or friendly, distraught, or at ease. Indeed, one could not even know where to look to find such expressions. Merleau-Ponty and Dolezal’s descriptions reveal, in other words, that a perceived face is not usually a static pattern like that on a piece of wallpaper. A perceived face is usually a dynamic activity whose most critical features, its expressions, are not, in the usual sense of the word, features at all.

Moreover, Merleau-Ponty’s descriptions suggest that one’s perception of others’ faces is intertwined with one’s own facial expressions and bodily activity. To see the face of a friend, for example, is to see someone at whom one can smile or with whom one can have a conversation, and it is in carrying out these motor activities that others’ faces assume their particular lived significance. The experiences of those who lack the ability to engage in motor activities like smiling offer some support for Merleau-Ponty’s insight and the enactive approach.23 People who have Möbius syndrome, a congenital condition in which the sixth and seventh cranial nerves are paralyzed, cannot move their eyes outward or side to side horizontally, and cannot move the muscles of facial expression. Yet, not only are those with Möbius syndrome unable to make facial expressions of their own, they also seem to have problems perceiving the facial expressions of others. One person with Möbius, for example, says that he had “difficulty in the street knowing if people were going to speak with him” (Cole 1999, 216). Cole (1999) writes that this difficulty could be generated, at least in part, by a tendency of those with Möbius not to look at others’ faces because their own “…difficulty in moving the eyes makes looking at people a stare, which is awkward” (216). Yet, he also notes that because those with Möbius do not elaborate their own facial expressions, they may be less aware of its importance to others, and be less able to interpret emotions in others” (216). The skill of forming one’s own mouth into various expressions may be critical to perceiving these expressions in others.24

The experience of vision without inversion of the retinal image draws attention, therefore, to what is usually at stake in the perception of other people. In encountering another person, the most pressing task is generally not to observe this person’s features but, instead, to engage with this person’s expressions. This task is particularly pressing because any failure in engaging with these expressions can jeopardize the perceiver herself and not simply her account of what she perceives. What she perceives, particularly when this being is another human being, is not indifferent to her; just as the perceiver’s expressions and conduct are affected by what she perceives, the expressions and conduct of what she perceives may be affected by her. Any being who confronts the perceiver makes demands from which the perceiver cannot escape responding since even the failure to respond will, itself, constitute a response. Yet, so long as the perceiver cannot tell, as in cases of vision without inversion of the retinal image, what the being before her is doing, she also cannot tell, in an important sense, what she is doing. That is, so long as she cannot perceive whether this being is snarling or smiling, she cannot tell whether her own expressions are calming or provoking. Thus, she will experience what she perceives as alarming, rather than merely curious, since this failure to perceive its expressions could very well put her in danger.25


All further references to Merleau-Ponty will be to the Phenomenology of Perception.


See, for example, Noë (2004) on the inadequacies of a “snapshot” conception of perceptual experience (35–73).


Even O’Regan and Noë (2001), in their discussion of vision without inversion of the retinal image, do not question accounts of the perceived world as inverted. Instead, drawing mainly on the work of Kohler, they focus on instances in which only fragments of the perceived world, rather than the perceived world as a whole, are described as inverted (953).


As Descartes (1965) writes, “Now, having thus seen this picture in the eye of a dead animal, and having considered its causes, you cannot doubt that an entirely similar one is formed in the eye of a live man” (97).


For a sampling of texts that contain some version of this explanation, see Gregory (1973; 21–23, 45), Bruce and Green (1990; 18, 74), and Mason and Kandel (1991; 423). For comparison of the retinal image to a camera image, as well as a discussion of some of the limitation of this metaphor, see Bruce and Green (1990; 31–32).


Stratton reported on these two studies in four articles published between 1896 and 1897 in the Psychological Review. Merleau-Ponty makes reference to the first three articles in the Phenomenology of Perception.


Dolezal (1982), for example, writes that the assumption “That visible, pictorial characteristics of the retinal image are somehow ‘looked at’ by some part of the brain during perception,” as well as the assumption, “That accurate perception necessitates a correspondence of orientation, location, shape, size, etc. of object and retinal image characteristics…deserve to be rethought carefully” (11).


While contemporary researchers may no longer conceive of vision as a process of seeing pictures in one’s head, Noë (2004) continues, “…the fundamental problem for visual theory—namely the problem of explaining how we see what we do given the imperfections of the retinal image—has basically the same shape as these antique puzzles” (46). Thompson, Noë, and Pessoa (1999) note that current discussions regarding the eye’s ‘blind spot,’ for example, usually assume that the brain must fill in the gap left by the ‘blind spot’ (37–38). Similarly, E. Thompson (1995) argues that the predominant theories of color vision begin from a “representationist model of perception” (27).


Thompson, Noë, and Pessoa (1999) argue, with respect to the idea of perceptual completion, or ‘filling in,’ for the rejection of analytic isomorphism.


See, for example, Merleau-Ponty’s discussion of the phenomenological method (viii).


For reviews of research on facial inversion, see Maurer et al. (2002) as well as Valentine (1988).


P. Thompson (1980) presented subjects with two pictures of Margaret Thatcher. In one of these pictures, Thatcher’s entire face was inverted; in the other, Thatcher’s entire face except for her mouth and eyes were inverted. When viewing the pictures, subjects noticed little difference between the two images and both appeared quite normal. Upon turning the pictures around, however, subjects a great difference between the two images and the picture in which Thatcher’s mouth and eyes have been inverted appeared extremely grotesque. Since the publication of this article, any image in which the mouth and eyes are, with respect to the rest of the image, upside-down is referred to as having been “Thatcherized,” and many subsequent experiments have employed “Thatcherized” faces.


For an overview of the work of, among others, Stratton, Kohler, Rock, Harris, Helmholtz, Held, and Howard see Dolezal (1982; 261–296).


Dolezal also reports, “…what seemed especially important to me after my own experience with gazing through up–down reversing prisms…was describing one’s observations without linguistic ambiguity, without allowing theoretical preconceptions to guide the descriptions of observations, and without mixing inferences with direct observations. Using such descriptive words as ‘upright vision,’ and ‘seeing upright’ or describing the orientation of objects as visually correct is confusing or misleading” (268). For further discussion of the difficulties in describing vision without retinal inversion, including the examination of what Dolezal identifies as “four interrelated areas of confusion” (269), see 269–271.


There is a complete absence of other people in Stratton’s account of his perceptual experiences. Wade (2003) notes this absence as ‘paradoxical’ (4).


Dolezal’s experience might be fruitfully compared with that of Lillian Kallir, who has difficulty seeing the practical meaning of the objects around her, though her perception of their color, size, and shape is intact (Sacks 2010). Kallir is unable, for example, to recognize an umbrella and, according to Sacks, “…wondered, for a half serious moment, if it was a snake” (14).


See, for examples, Farah et al. (1995) and Searcy and Bartlett (1996).


Dolezal asserts that this difference between the previously perceived world and the presently perceived world, even after adaption has occurred, also existed for Stratton; “Nowhere does Stratton say that the orientation of his FV [field of vision], relative to the orientation of his head, looked just as it had before he put on inverting lenses” (265). Furthermore, he claims that “If such a same–different criterion is applied, then it may be said that no one has yet claimed that there was any ‘flipping over’ (either sudden or gradual) of the perceived FV [field of vision], whatever ‘flipping over’ might mean” (265).


Those who criticize these approaches recognize the diverse history of experiments which they encompass; as Thompson, Noë, and Pessoa (1999) write, “This account of perception is based in a well-established empirical research program and so should not be dismissed either on conceptual grounds or simply by adducing counterexamples. It can be challenged, however, by offering an alternative theoretical and empirical framework as a rival research program.” (398).


See also Varela et al. (1991; 206. The enactive approach has been applied to the study of change blindness (O’Regan and Noë 2001; 954), inattentional amnesia (O’Regan and Noë 2001; 955), remote tactile sensing (O’Regan and Noë 2001; 956–957), and tactile visual sensory substitution (O’Regan and Noë 2001; 957–958).


Cf. Merleau-Ponty’s (1967) description of the subject of perception as a “gaze which takes a grip upon things” (253).


For a more recent study that confirms these results, see Prkachin (2003).


I would like to thank an anonymous reviewer at Phenomenology and the Cognitive Sciences for suggesting this line of thinking.


In a more recent work, Cole (2009) cites a study by Andy Calder, who found that, while those with Möbius could, in a carefully selected set of photographs, recognize facial expressions of emotion, they did show deficits in their ability to describe the facial movements that achieved these emotions; Cole writes, “…if confirmed in larger samples, it suggests that internal programmes or models of facial movement may be accessed and necessary to describe facial expressions, and without these models the ability to imagine and describe is reduced” (196).


Merleau-Ponty’s description of the expressions as ‘terrifying,’ therefore need not seem at odds with the general conclusion in empirical research on facial inversion that an inverted facial expression as a neutral facial expression. Given that this research uses pictures of faces rather than actual faces, the subject has the luxury of considering the expressions of these faces without any fear of harm.


Copyright information

© Springer Science+Business Media B.V. 2011