Friday, April 15, 2011

Does the World Look Upside-Down Through Inverting Lenses, Once You're Used to Them?

For eight days in 1897, psychologist George Stratton saw the world through a pair of glasses that rotated everything 180 degrees so that up was down and left was right. At first, everything looked upside down and backwards and he stumbled about, crashing into things. By the eighth day, he had largely adapted to the inverting lenses and was able to skillfully maneuver through his environment. Although Stratton's experiment is often cited by psychologists and philosophers of perception, few have effectively addressed the question that interests me most: After adaptation, did the world still look upside-down to Stratton, or at some point did it flip rightside up?

Since after adaptation the upside-down look might no longer seem wrong, in posing this question perhaps I should avoid the word "upside-down", with its implication of error or illusion. So maybe the stays-upside-down view is better expressed as follows. Before donning the inverting lenses, let's suppose, the visual field is toovy when Stratton looks at his room (where "toovy" picks out the phenomenal quality ordinarily associated with its looking like the light is hanging down from above, etc.) and after donning the lenses for the first time, the visual field turns teavy instead (where "teavy" picks out the phenomenal quality ordinarily associated with its looking like the light is jutting up from below). Here are two possibilities for what happens during adaptation: (1.) At some point (perhaps unsteadily and fragmentedly) Stratton's visual experience flips from teavy back to toovy. (2.) Stratton's visual experience stays teavy, but he learns to associate that teavy quality with the light's actually being above him.

These seem like very different phenomenal possibilities. But which is correct? Susan Hurley and Alva Noe (2003) say that things go back to looking toovy -- that post-adaptation, Stratton's visual pheomenology is just the same as it was before donning the glasses (minus the feeling of wearing glasses and other incidentals like the narrowing of the field). They say this partly based upon, and in defense of, a controversial general theory of the relationship between skillful action and conscious experience, and partly based upon James Taylor's 1962 reports in his replication of Stratton. But Taylor's reports, too, are situated within a theoretical framework to which he appears deeply committed, and Stratton's original reports are most naturally interpreted as suggesting the opposite. For example, Stratton writes:

The harmonization of the new experience and the suppression or subordination of insistent remnants of the old were always apparent during active operations in the visual surroundings, as has been described for several of the preceding days. While I sat passively the old localization of unseen parts of my body often came back, or perhaps was the usual form in which they appeared. But the instant I began to rock my chair the new position of these parts came prominently forward, and, except in the case of my shoulders and back, readily felt more real than the old. And in walking, when hands and feet rhythmically made their appearance in the visual field, the old representation, except perhaps for some faint inharmonious sensations in the back, was fully expelled without employing any device of will or of attention whatever (1897c, p. 469).
Stratton does not say that the new experience is the same as the old, after adaptation: Rather the new experience is the reverse of the old, and adaptation is the victory of the new in the fight between them.

I don't know which way it goes, but the two possibilities seem to differ, and different observers appear to be giving different reports. The question has implications for the general question of whether skillful action shapes our phenomenology to match the world or whether, instead, phenomenology tends to stay (approximately) the same across the development of skill, only gathering new associations and behavioral contingencies.

I discuss this issue with more everyday examples of mirror reflection in earlier blog posts on the apparent location of mirror images and on the U.S. Department of Transportation's advice that "objects in mirror are closer than they appear".

Similar questions arise, of course, for inverted spectrum cases.


Kevin Reuter said...

Hi Eric,

How much do you think can more mundane examples like wearing yellow-tinted glasses tell us about the more extreme cases?

In these mundane examples, it seems that we adapt to the different phenomenology without changes to our phenomenology (I take our surprise about how we took the yellowish phenomenology to be veridical when we take off the glasses, to be good evidence in favour of that view).

Mirror cases seem to be in line with the yellow-tinted glasses case, and my intution is that we only adapt to the upside-down situation.

What does it anyway mean to say 'it looks veridical'? I don't think it means more than saying 'I take it to be veridical'.



Eric Schwitzgebel said...

Hi, Kevin, thanks for the comment!

Color adaptation might be different, due to adaptation or fatigue in the retina -- a process that wouldn't seem to have a parallel for inverting lenses. So although it's a tempting analogy, I would still be cautious about it.

The meaning of "veridical" is near the heart of the issue, too. If some resemblance relation is required, then veridicality might not be preserved through distortions. But, as you say, it might have a more purely subjective meaning.

Baker Jones said...

What's up Eric,(no pun intended)

I recently stumbled upon your blog looking for some information on metaphysics. I was blown away with some of the posts that you've written! You seem to have a very peculiar expertise in the field of philosophy. While your blog is great, my thirst for knowledge isn't quiet quenched yet...and I was wondering if you(or any other philosophers) could help guide me towards what I'm trying to find.
If you're interested in helping me, please contact me via email at, if your too busy to help, no worries!



Unknown said...


You might want to check out a book entitled ‘Living in a world transformed’ by Hubert Dolezal, 1982. He spent a summer wearing inverting goggles in a Greek village. As I recall, he did not claim that the world eventually looked right side up. There is also a paper entitled ‘The myth of upright vision. A psychophysical and functional imaging study of adaptation to inverting spectacles’, by Linden et al., Perception, 1999, 28, 469-481. They also claim that there was no return to upright vision, based on introspection, behavioral tests and brain imaging.

What seems to be going on, at least in part, is that you gradually change your visuo-motor associations. For example, you learn that to reach for something that is visually downwards you now have to execute a motion with your arm in the direction against gravity, that was earlier appropriate for reaching something that was visually upwards. This process of unlearning old associations (upwards is against gravity, downwards is along gravity) and learning to reverse them may have as a consequence that both the visual phenomenology (‘what is visually up and down?’) and motor phenomenology (‘in which direction am I reaching?’) get somewhat ambiguous and confused.


Eric Schwitzgebel said...

@ Baker: Please feel free to email me if you have some specific thoughts or questions related to my areas of specialty.

Eric Schwitzgebel said...

@ Dejan: Thanks for those great references! I am still sorting through the literature on this, so that is very helpful.

CSH said...


I wonder if this is similar to how we re-associate movement with a joystick or a mouse. For instance, typically we recognize that pulling back on a joystick will make the [plane] go up while pushing forward will make the [plane] go down. But we can reverse the y-axis and, after initially stumbling, learn the new association. Alternately, it seems intuitively obvious that moving the mouse to the top of the screen will cause the screen to move up while moving the mouse to the bottom with cause the screen to move down but in many games the axis is reversed such that "up" makes us look down and "down" makes us look up. in these cases we don't need to change our visual cues to re-associate our movement responses.

I'm not sure how that helps your overall project but it might be worth thinking about in connection with what you are doing.

Eric Schwitzgebel said...

CSH: Thanks for that thought. It's a very nice analogy!

Unknown said...

As usual, late on commenting! But for what it's worth, here are some thoughts.

First, you write: "Although Stratton's experiment is often cited by psychologists and philosophers of perception, few have effectively addressed the question that interests me most: After adaptation, did the world still look upside-down to Stratton, or at some point did it flip rightside up?" Rock in his book on perceptual adaptation carefully addresses this question. He adduces a lot of evidence that complete visual adaptation is highly unlikely for inverting prisms. (Indeed, even for modest L-R displacement, adaptation is rarely if ever complete.) It's also important to take into account the role of proprioceptive adaptation to prisms: such adaptation would not be at the level of associations and behavioral contingencies. It's quite probable that even if Stratton's visual world did not "flip," he experienced significant changes in the felt position of body parts (see Rock 1966, pp. 44-50).

It's probably best not to focus, when addressing the philosophical relevance of prism work, on the extreme cases of inversion or reversal since the experimental evidence here is quite difficult to interpret and often anecdotal. The overwhelming majority of experiments involve adaptation to modest L-R displacement, and there's lot of good evidence here for both proprioceptive and genuinely visual adaptation (with a change in phenomenology in each case). Welch's 1978 book is a fantastic resource.

Noe pays no attention to the extensive literature critical of Held's sensory reafference theory on which his own theory is partly based (see Welch for a survey of responses to Held). Looking at this literature from 60s-80s, I get the impression that almost everyone else actually involved in experimentation denied that active movement is *necessary* for adaptation, but accepted that active movement generates lots of information about the nature of the prism-induced displacement that greatly facilitates adaptation (the "information hypothesis"). In other words, active movement may indeed result in phenomenological change, but not for the reasons Noe and Held suggest. It's really quite surprising that Noe relies on Held's work, but does not address the massive criticism it received from other psychologists. See Welch 1978 chap 2. and Rock 1966 chaps. 2-4 for helpful evaluations of competing hypotheses.

Last, just because the Held-Noe theory doesn't work, that doesn't mean that motor theories of perception in general are doomed. After the demise of the sensory-reafference theory of adaptation, a lot of people went on to investigate the role of efferent factors in perception and to develop versions of the "efferent readiness" theory. Coren's 1986 paper "An Efferent Component in the Visual Perception of Direction and Extent" has a nice historical review of efferent readiness theories and also presents some striking empirical support for a version of the theory based on visual illusion evidence.

Dustin Stokes said...

Hi Eric,

Thanks for this. I know this post is a few months old, but I found it by searching for a discussion of the very question you pose.

The issue has another possible implication for the cognitive penetration of perceptual experience. Depending upon what one includes in “skilful activity”, this may be the same or different from the implication you suggest: the effect of skilful activity on phenomenology of experience. In any case, I’m writing a paper on cognitive penetrability for Philosophy Compass and at the moment discussing Paul Churchland’s use of the inverting lenses case (Churcland, ‘Perceptual Plasticity and Theoretical Neutrality: A Reply to Jerry Fodor’, Phil Sci, 1988). He takes the case to be an instance of the diachronic penetration of perception, where over time (a week or so) the subject learns new relations between movement and visual experience, re-formulates expectations accordingly and, eventually, perceives the world just as before wearing the inverting lenses. So like Hurley and Noe, Churchland’s assumption is that experience, post-adaptation, goes back to being toovy. And this assumption is important, since if the phenomenology is static across pre and post-adaptation, then the case only provides evidence for differences in judgements or beliefs about experience, rather than perceptual experience itself—in other words, cognitive penetration of cognition, not of experience. (Fodor, for what it’s worth attempts to rebut the argument not by challenging this assumption, but by (a) insisting that (relevant) cases of cognitive penetration are synchronic and (b) insisting that adaptation to inverting lenses is a recalibration of the sensory system by the sensory system, and not by cognitive states or processes.)

I think you are exactly right to point out that the evidence is ambiguous (at best) with respect to the two options (the post-adaptation phenomenology reverting to “normal” vs. the phenomenology remaining the same—teavy—across pre and post-adaptation). Do you know if anyone has published the opposing (teavy post-adaptation phenomenology) view?

I found the post very useful. Thanks and all the best.

Dustin Stokes

Eric Schwitzgebel said...

Thanks for the terrific comment, Dustin! I'll email you shortly.