- Email: [email protected]
Explicit and motoric dependent measures of geographical slant are dissociable: A reassessment of the findings of Durgin, Hajnal, Li, Tonge, and Stigliani (2010)
Acta Psychologica 138 (2011) 285–288
Contents lists available at ScienceDirect
Acta Psychologica j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / a c t p s y
Explicit and motoric dependent measures of geographical slant are dissociable: A reassessment of the findings of Durgin, Hajnal, Li, Tonge, and Stigliani (2010) Dennis R. Proffitt ⁎, Jonathan R. Zadra University of Virginia, United States
a r t i c l e
i n f o
Article history: Received 9 April 2010 Received in revised form 8 June 2010 Accepted 9 June 2010 Available online 4 July 2010 PsycoINFO classification: 2323 Keywords: Geographical slant Motor-dependent measures Visually guided actions Palmboards
a b s t r a c t Durgin et al. (2010) argued that the apparent accuracy of the palmboard measure of geographical slant is accidental and reflects limitations in wrist flexion that reduce palmboard adjustments by just the right amount given the perceptual overestimations upon which they are based. This account is inconsistent with findings that verbal reports and palmboard adjustments are dissociable. In addition to previous evidence found for such dissociation, Durgin et al. also found verbal/palmboard dissociations in Experiment 2. Experiments 1 and 3 of Durgin et al. lacked verbal reports and instead compared palmboard adjustments to free-hand estimates in the context of small wooden surfaces. These experiments are not relevant to the issue of verbal/palmboard dissociability. Across studies, the accuracy of Durgin et al.'s palmboard implementation is far less than that found by others (Feresin & Agostini, 2007). The design of Durgin et al.'s Experiment 5 misrepresented the experimental conditions of Creem and Proffitt (1998), and consequently, the findings of this study have no bearing on the issue at hand. © 2010 Elsevier B.V. All rights reserved.
Roughly 20 years ago, I (DRP) became interested in geographical slant perception. It seemed obvious to me that hills appeared much steeper than, in fact, they were. At that time, I could find only one published empirical study on geographical slant perception, Kammann (1967), although anecdotal descriptions were pervasive (Ross, 1974). I convinced Mukul Bhalla, Rich Gossweiler, and Jonathan Midgett that it would be worthwhile to do a systematic empirical study of this odd phenomenon. We decided that it would be good to use a number of converging dependent measures, so in addition to verbal reports, we employed a visual matching task and palmboard adjustments. Our thought was that all three of these measures would provide closely matching values. The palmboard let us down. Whereas the former two dependent measures exhibited huge overestimations – for example 5° hills were judged to be about 20° via verbal reports and the visual matching measure – the palmboard adjustments were relatively accurate (Proffitt et al., 1995). Although in this and other experiments, the palmboard adjustments provided slight overestimations of hill slants below about 20° and slight underestimations for inclines of greater magnitudes, the size of these over- and underestimations was rarely more that 5°, which seemed to us miniscule in comparison with the huge overestimations obtained via verbal reports and the visual matching task. Compared with our findings,
⁎ Corresponding author. Department of Psychology, University of Virginia, P.O. Box 400400, Charlottesville, VA 22904, United States. E-mail address: [email protected] (D.R. Proffitt). 0001-6918/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.actpsy.2010.06.003
Feresin and Agostini found far better accuracy with a more ergonomic palmboard, for slants between 4° and 18°. Durgin et al. (2010) argued that the apparent accuracy of palmboard adjustments is accidental and reflects limitations in wrist flexion that reduce palmboard adjustments by just the right amount given the perceptual overestimations upon which they are based. We considered this possibility previously: “There are two accounts for why visual awareness of geographical slant exhibits gross overestimations whereas visually guided actions do not. By one account, actions are mediated by visual awareness following a transformation that compensates for overestimations. The other account postulates that conscious awareness and action are informed by separate visual pathways, one showing evidence of bias and the other not” (Proffitt et al., 1995, pp. 425). The former alternative – that palmboard adjustments reflect a response transformation that nullifies overestimations – was eventually rejected by us as new studies yielded results supporting the latter alternative. Early on, we conducted a study in virtual reality in which participants, wearing a head-mounted display, viewed simulated hills spanning a wide range of slants up to 60° (Gossweiler et al., 1995). Palmboard adjustments maintained relative accuracy across this range with adjustments for 60° virtual hills being about 55°.
286
D.R. Proffitt, J.R. Zadra / Acta Psychologica 138 (2011) 285–288
Clearly, people are able to set the palmboard to large angles. The notion that biomechanical constraints almost exactly cancel out perceptual overestimations seemed implausible. Compelling support for the latter hypothesis, in which verbal/ visual measures and motoric adjustments are dissociable, came from a wide range of findings showing clear dissociations between these two classes of measures. First, a number of studies looked at bioenergetic influences on slant perception. In early studies, we found that manipulations of encumberment and fatigue as well as individual differences in fitness, age and health influenced the verbal/visual measures but not the palmboard adjustments (Bhalla & Proffitt, 1999). Recently, Schnall et al. (2010) found that blood glucose levels affected the verbal/visual measures of geographical slant but not palmboard adjustments. Importantly, in studies in which we predicted an influence on palmboard adjustments, such as responding from memory (which we will discuss shortly), palmboard adjustments were affected (Creem & Proffitt, 1998). These latter findings show that the null finding for bioenergetic effects on palmboard adjustments was not due to a lack of statistical power. Especially strong evidence for the dissociation between verbal/ visual and palmboard measures is seen in the following finding: The mapping between conscious awareness and motoric actions is recalibrated for long-term but not short-term changes in the physiological potential of individuals (Bhalla & Proffitt, 1999). In these experiments, participants were verbally instructed to set the palmboard to a range of angles; for example, participants were instructed to set the palmboard to 10°, 30°, 5°, and so forth. When so instructed, they set the palmboard to lesser angles than the ones given; however, the adjustment closely matched the angles that they would have set in response to viewing hills that evoked the same verbally given angle. For example, typical young adults standing at the base of a 5° hill would report that it appeared to be 20°, whereas their palmboard adjustment would be about 5°.1 When later asked to provide palmboard adjustments to a range of verbally given angles, they made a 5° adjustment when instructed to make one of 20°. Such results fit well with the notion that palmboards are accurate due to their accidentally underestimating conscious perceptions. Additional findings, however, provided evidence that the relationship between verbal and palmboard measures is not accidental in this way. Although verbal estimates (e.g., 20°) of hill slant and palmboard adjustments (5°) were consistent with verbal instructions (20°) and resulting palmboard adjustments (5°) in the case described above, this consistency was disrupted when individuals were encumbered or became fatigued. Fatigued participants, for example, would perceive the 5° hill to be steeper—about 25°; however, their palmboard adjustments remained unaffected: when given a verbal instruction to make a palmboard adjustment of 20°, they still provided a 5° response, as was the case before fatigue set in. Although viewing the hill evoked a 25° verbal judgment concomitant with a 5° palmboard adjustment, a 25° verbal instruction evoked a palmboard adjustment of 7°. Thus, in the context of viewing a hill, fatigued participants verbally stated that it was 25° and set the palmboard to 5°, whereas when asked to set the palmboard to 25°, they set it to 7°, thereby indicating a loss of internal consistency. Internal consistency was not lost for people who had adapted to long-term changes in their physiological potential. For example, when participants who were low in physical fitness viewed a 5° hill, they reported that it was 25° but 1 In this example, which will be carried through the next paragraph, the palmboard adjustment values are idealized for expository purposes. As was stated earlier, palmboard adjustments yield an overestimation of hill slants for inclines of less than 20°. Thus, when viewing a 20° hill, a young adult would have likely set the palmboard to about 8°. Moreover, they would also set it to about 8° when verbally instructed to set it to 20°. The issue under discussion is the consistency found for long-term changes in physiological state and the loss of consistency found for short-term changes. The values given for the palmboard adjustments in these examples exaggerate its accuracy for purposes of expository clarity.
made an accurate 5° palmboard adjustment. When asked to make a palmboard adjustment of 25°, they provided a 5° response. These individuals exhibited an internal consistency between their conscious representations and motoric actions (Bhalla & Proffitt, 1999). Such findings of dissociations for short- but not long-term changes in physiological potential are impossible to explain by the account that the accuracy of the palmboard is accidental and due to physiological limitations in wrist flexion. In sum, we initially entertained the possibility that palmboard adjustments were accurate due to constraints on the adjustment response that cancelled the overestimations inherent in perceiving geographical slant. The strongest evidence against this cancellation account is that verbal/visual measures and palmboard adjustments are dissociable. None of the studies presented in Durgin et al. cast doubt on the evidence for the dissociation of these response measures, and instead, they provide additional evidence that palmboard adjustments and explicit measures of geographical slant are dissociable (Experiment 2). A detailed critique of each experiment in Durgin et al. follows. Experiment 1. In Experiment 1, Durgin et al. placed a small wooden wedge with a face incline of 30° on a table in front of participants. Participants made two responses: 1) adjusting a palmboard to match the incline of the wedge, and 2) holding their hand in the air above the palmboard so that the hand's orientation felt parallel to the incline of the wedge. They found that the latter “free-hand” estimates (32.7°) were not reliably different from the actual incline of the wedge, while palmboard estimates (19.4°) were reliably less. The results from Experiment 1 generalize neither to our geographical slant studies nor to the other experiments presented by Durgin et al. First, our slant studies made use of real hills in which the hill crests were over the participants' heads. Second, in our studies, palmboard adjustments to hills of about 30° were close to accurate— not underestimated by over 10° as in the Durgin et al. study; no account was offered by Durgin et al. for this disparity in results. It may be that Durgin et al.'s palmboard was not as easy to adjust as was ours. It seems inappropriate to conclude that all palmboard adjustments made in the context of a 30° slant are underestimated when this was only the case for Durgin et al.'s experiment and not our own. Recall, moreover, that Feresin and Agostini (2007) found almost perfect accuracy with their palmboard for slants between 4° and 18°. The palmboard implementation in the Durgin et al. study is suspect. Finally, in this experiment no verbal estimates were obtained. Thus, there is no way of knowing whether the accurate free-hand responses would have been similar to verbal reports and whether the small wooden incline evoked the large overestimation of slant that is always found for real hills. When drawing generalizations from one study to another, equating experimental conditions is important. Experiment 2. In their second experiment, Durgin et al. used the same two measures as were used in Experiment 1, palmboard and free-hand responding. This study, however, was conducted outdoors in front of a real hill. In this experiment, participants also gave verbal estimates following the two non-verbal measures. An additional factor, gaze direction, was added such that participants gave two sets of estimates with regards to the hill: one while looking down at the base of the hill, and the other while looking straight ahead at the hill. The hill was not exactly uniform—the portion of the hill that participants viewed when their gaze was down was inclined by 23.5°, while the portion that subjects viewed at eye level was slightly steeper at an incline of 24.5°. Durgin et al. found that across both gaze directions the verbal and free-hand estimates were overestimated relative to the actual slope and the palmboard estimates were underestimated. The free-hand
D.R. Proffitt, J.R. Zadra / Acta Psychologica 138 (2011) 285–288
adjustments were no more accurate than those obtained with the palmboard. Moreover, the underestimation found for the palmboard was again inconsistent with our previous results that show relative accuracy for hills of this slant. For hills of comparable slant (18°), Feresin and Agostini (2007) found near perfect accuracy for participants' palmboard adjustments. Again, Durgin et al.'s palmboard implementation is suspect. Importantly, both verbal and free-hand estimates given with a downward gaze direction were significantly less than those given with a forward gaze. There was no significant effect of gaze direction, however, for the palmboard estimates. The results from Experiment 2 show a clear dissociation of palmboard adjustments from free-hand estimates and verbal reports. Only the free-hand estimates exhibited an influence of gaze angle similar to that found for verbal reports. This is evidence that the palmboard estimates are dissociable from verbal reports, whereas the free-hand estimates are not. Durgin et al. argued that the palmboard was too insensitive to detect small changes across gaze conditions; however the internal consistency findings (Bhalla & Proffitt, 1999) described above and the Creem and Proffitt (1998) findings discussed below provide evidence counter to this claim. Experiment 3. For both Experiments 3A and 3B, Durgin et al. presented small, irregularly shaped wooden surfaces anchored to a tripod that could be adjusted to various inclines a short distance in front of the participant. In 3A, participants made free-hand and palmboard estimates of the inclines. In 3B, participants reached out and placed their hand on the surfaces while palm orientation was recorded. The results for Experiment 3A showed that free-hand estimates were overestimated and that palmboard adjustments were underestimated. The experiment suffers from the same problems of generalizability as did Experiment 1: the stimuli are even further removed from the real hills used in our geographical slant studies (in addition to the small surface area of the stimuli, they were surrounded with a hemispheric enclosure of black felt, and participants wore goggles that restricted their view). As in the previous two experiments, the palmboard findings in this Durgin et al. study are inconsistent with the findings of others. This leads to two major issues with the experiment: it cannot address the primary question of dissociability because no verbal estimates were obtained, and the errors in accuracy found for palmboard estimates are inconsistent with the higher accuracy found in other studies. The results cannot be generalized to the other experiments in Durgin et al.'s paper, or to our experiments, because the relationship between free-hand and palmboard estimates cannot be evaluated in terms of perceptual overestimation without knowing the relationship to verbal estimates that would have resulted from this novel set of reduced cue stimuli. In other words, do the small wooden surfaces evoke a perceptual overestimation of their slant, or is the overestimation evident in free-hand responses due to something else, such as an artifact in its use or implementation? There were no verbal estimates, and thus, there is no way of knowing the relationship between free-hand responses and verbal reports and, subsequently, the meaning of any difference between free-hand and palmboard estimates. An additional problem in Experiment 3A is that the accuracy of the palmboard estimates are poor compared to the results obtained by us and those obtained by Feresin and Agostini (2007). This suggests that either the reduced cue stimuli (relative to the full-cue viewing of hills in our and Feresin and Agostini studies) or the different palmboard implementation used by Durgin et al. is responsible for their poorer accuracy. Experiment 3A is sufficiently different from our experiments and those of other researchers, both in stimuli and results, that the results have little bearing on the issue at hand. The reaching task in
287
experiment 3B does not add to the claims made by Durgin et al. They found that people accurately adjust the angle of their palm to match the angle of an inclined surface when reaching for it. If anything, this supports the idea that people are good at using their palms to act on surfaces of various orientations—and with a palmboard, participants are acting on a surface, while free-hand estimates do not involve any object-oriented action. Experiment 4. In the fourth experiment, participants placed their forearm on a horizontal board and then adjusted either their entire forearm (keeping the elbow on the board) or just the hand (keeping the forearm and wrist on the board) to match the orientation of a line depicting various angles between 5° and 60°. When rotating from the elbow, participants were more accurate at matching angles over 30°, and more accurate when rotating the wrist only for angles below 30°. Durgin et al. motivated Experiment 4 with the claim that palmboard adjustments involve rotating a single joint, the wrist, and that the lack of flexibility in the wrist is one cause of the “accidental accuracy” of palmboard estimates. Adjusting the palmboard, however, involves rotating the shoulder, elbow, and wrist, and as was discussed above, participants are able to easily set the board to a wide range of angles up to at least 60°. You can experience these motions of the arm and shoulder by attempting the following: Hold your hand level at waist height and position this hand about 10 inches in front of your body. Flex your wrist to its maximum and observe the slant of your hand. Now, attempt to increase this slant. This is easily achieved by slightly flexing the knees, dropping the shoulder, and bending the elbow. Both conditions in Experiment 4 placed far more limitations on the biomechanical movement needed to match orientations than does a palmboard, and thus, the results of this study do not generalize to the palmboard. Experiment 5. In Experiment 5, Durgin et al. asked participants to remember three well known hills on campus and to set the palmboard to match the incline of each hill one at a time. Participants did this either while seated with the palmboard set just above their navel, standing with the palmboard set just above their navel, or standing in the posture used in Proffitt et al. (1995) with the palmboard at waist height. Participants, who had the palmboard set just above their navel, whether seated or standing, provided higher estimates than the third group. Experiment 5 was motivated by a need to explain the memory results from Creem and Proffitt (1998), mentioned above. In the Creem and Proffitt study, participants were asked to judge hill slants from memory following a delay. With a short delay of under a minute (and while still in the presence of the hill), palmboard estimates matched normative palmboard estimates given while subjects were viewing the hill. With a longer delay (and away from the hill) palmboard estimates were significantly increased. This suggests that with a short delay actions can be guided by short-term visual guidance memory, but after a longer delay this visual guidance memory has dissipated and the action must draw on the conscious, explicit memory, which reflects overestimation. These results cannot be explained by a limitation in wrist flexion. They show that verbal reports and palmboard estimates are dissociable when responses are based upon perception but not dissociable when slant is judged from memory. These findings also demonstrate that the palmboard measure is sufficiently sensitive to reveal the presence or absence of verbal/palmboard dissociations. Durgin et al. motivated their experiment with the assertion that participants in Creem and Proffitt (1998) were seated during some palmboard estimates, and that this confounding of posture and memory accounted for the results. Prior to the submission of Durgin et al.'s paper to this journal, we informed Durgin that, to the best of our
288
D.R. Proffitt, J.R. Zadra / Acta Psychologica 138 (2011) 285–288
knowledge, the participants in Creem and Proffitt (1998) were standing when making palmboard adjustments from memory. In none of our other studies have participants ever been seated while adjusting the palmboard, and neither Creem-Regehr nor I (DRP) have any recollection of having participants sit during palmboard estimates in this experiment. Moreover, in Experiment 4 of Creem and Proffitt (1998), participants in all conditions were unquestionably standing while adjusting the palmboard from memory, yet the effects were the same. In all of our experiments, the palmboard was positioned at waist level. We never conducted a study in which the palmboard was positioned above the navel. Durgin et al.'s Experiment 5 misrepresents the experimental design of Creem and Proffitt (1998), and thus, its design is without motivation and its results have no bearing on the issues at hand. Feresin and Agostini (2007) have already shown that posture matters in palmboard adjustments. By carefully positioning their palmboard to optimize ease of use, Feresin and Agostini found far better accuracy, for slants between 4° and 18°, than were ever obtained by us. In summary, there is considerable evidence that palmboard adjustments are dissociable from verbal/visual measures. In Experiment 2, Durgin et al. provided additional evidence for their dissociability. In their general discussion, Durgin et al. proposed an account by which our bioenergetic effects on geographical slant perception are due to demand characteristics. We have addressed these claims elsewhere (Proffitt, 2010) and will not do so again here.
References Bhalla, M., & Proffitt, D. R. (1999). Visual-motor recalibration in geographical slant perception. Journal of Experimental Psychology: Human Perception and Performance, 25, 1076−1096. Creem, S. H., & Proffitt, D. R. (1998). Two memories for geographical slant: Separation and interdependence of action and awareness. Psychonomic Bulletin & Review, 5, 22−36. Durgin, F. H., Hajnal, A., Li, Z., Tonge, N., & Stigliani, A. (2010). Palm boards are not action measures: An alternative to the two-systems theory of geographical slant perception. Acta Psychologica, 134(2), 182−197. Feresin, C., & Agostini, T. (2007). Perception of visual inclination in a real and simulated urban environment. Perception, 36, 258−267. Gossweiler, R., Proffitt, D. R., Bhalla, M., & Pausch, R. (1995). A hill study: Using a virtual environment as a perceptual psychology laboratory.IEEE computer: Special issue on real applications for virtual reality March. Kammann, R. (1967). The overestimation of vertical distance and slope and its role in the moon illusion. Perception & Psychophysics, 2, 585−589. Proffitt, D. R. (2010). Affordances matter in geographical slant perception: Reply to Durgin, Baird, Greenburg, Russell, Shaughnessy, and Waymouth. Psychonomic Bulletin & Review, 16, 970−972. Proffitt, D. R., Bhalla, M., Gossweiler, R., & Midgett, J. (1995). Perceiving geographical slant. Psychonomic Bulletin & Review, 2, 409−428. Ross, H. E. (1974). Behavior and perception in strange environments. London: George Allen & Unwin. Schnall, S., Zadra, J. R., & Proffitt, D. R. (2010). Direct evidence for the economy of action: Glucose and the perception of geographical slant. Perception, 39, 464−482.
Contents lists available at ScienceDirect
Acta Psychologica j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / a c t p s y
Explicit and motoric dependent measures of geographical slant are dissociable: A reassessment of the findings of Durgin, Hajnal, Li, Tonge, and Stigliani (2010) Dennis R. Proffitt ⁎, Jonathan R. Zadra University of Virginia, United States
a r t i c l e
i n f o
Article history: Received 9 April 2010 Received in revised form 8 June 2010 Accepted 9 June 2010 Available online 4 July 2010 PsycoINFO classification: 2323 Keywords: Geographical slant Motor-dependent measures Visually guided actions Palmboards
a b s t r a c t Durgin et al. (2010) argued that the apparent accuracy of the palmboard measure of geographical slant is accidental and reflects limitations in wrist flexion that reduce palmboard adjustments by just the right amount given the perceptual overestimations upon which they are based. This account is inconsistent with findings that verbal reports and palmboard adjustments are dissociable. In addition to previous evidence found for such dissociation, Durgin et al. also found verbal/palmboard dissociations in Experiment 2. Experiments 1 and 3 of Durgin et al. lacked verbal reports and instead compared palmboard adjustments to free-hand estimates in the context of small wooden surfaces. These experiments are not relevant to the issue of verbal/palmboard dissociability. Across studies, the accuracy of Durgin et al.'s palmboard implementation is far less than that found by others (Feresin & Agostini, 2007). The design of Durgin et al.'s Experiment 5 misrepresented the experimental conditions of Creem and Proffitt (1998), and consequently, the findings of this study have no bearing on the issue at hand. © 2010 Elsevier B.V. All rights reserved.
Roughly 20 years ago, I (DRP) became interested in geographical slant perception. It seemed obvious to me that hills appeared much steeper than, in fact, they were. At that time, I could find only one published empirical study on geographical slant perception, Kammann (1967), although anecdotal descriptions were pervasive (Ross, 1974). I convinced Mukul Bhalla, Rich Gossweiler, and Jonathan Midgett that it would be worthwhile to do a systematic empirical study of this odd phenomenon. We decided that it would be good to use a number of converging dependent measures, so in addition to verbal reports, we employed a visual matching task and palmboard adjustments. Our thought was that all three of these measures would provide closely matching values. The palmboard let us down. Whereas the former two dependent measures exhibited huge overestimations – for example 5° hills were judged to be about 20° via verbal reports and the visual matching measure – the palmboard adjustments were relatively accurate (Proffitt et al., 1995). Although in this and other experiments, the palmboard adjustments provided slight overestimations of hill slants below about 20° and slight underestimations for inclines of greater magnitudes, the size of these over- and underestimations was rarely more that 5°, which seemed to us miniscule in comparison with the huge overestimations obtained via verbal reports and the visual matching task. Compared with our findings,
⁎ Corresponding author. Department of Psychology, University of Virginia, P.O. Box 400400, Charlottesville, VA 22904, United States. E-mail address: [email protected] (D.R. Proffitt). 0001-6918/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.actpsy.2010.06.003
Feresin and Agostini found far better accuracy with a more ergonomic palmboard, for slants between 4° and 18°. Durgin et al. (2010) argued that the apparent accuracy of palmboard adjustments is accidental and reflects limitations in wrist flexion that reduce palmboard adjustments by just the right amount given the perceptual overestimations upon which they are based. We considered this possibility previously: “There are two accounts for why visual awareness of geographical slant exhibits gross overestimations whereas visually guided actions do not. By one account, actions are mediated by visual awareness following a transformation that compensates for overestimations. The other account postulates that conscious awareness and action are informed by separate visual pathways, one showing evidence of bias and the other not” (Proffitt et al., 1995, pp. 425). The former alternative – that palmboard adjustments reflect a response transformation that nullifies overestimations – was eventually rejected by us as new studies yielded results supporting the latter alternative. Early on, we conducted a study in virtual reality in which participants, wearing a head-mounted display, viewed simulated hills spanning a wide range of slants up to 60° (Gossweiler et al., 1995). Palmboard adjustments maintained relative accuracy across this range with adjustments for 60° virtual hills being about 55°.
286
D.R. Proffitt, J.R. Zadra / Acta Psychologica 138 (2011) 285–288
Clearly, people are able to set the palmboard to large angles. The notion that biomechanical constraints almost exactly cancel out perceptual overestimations seemed implausible. Compelling support for the latter hypothesis, in which verbal/ visual measures and motoric adjustments are dissociable, came from a wide range of findings showing clear dissociations between these two classes of measures. First, a number of studies looked at bioenergetic influences on slant perception. In early studies, we found that manipulations of encumberment and fatigue as well as individual differences in fitness, age and health influenced the verbal/visual measures but not the palmboard adjustments (Bhalla & Proffitt, 1999). Recently, Schnall et al. (2010) found that blood glucose levels affected the verbal/visual measures of geographical slant but not palmboard adjustments. Importantly, in studies in which we predicted an influence on palmboard adjustments, such as responding from memory (which we will discuss shortly), palmboard adjustments were affected (Creem & Proffitt, 1998). These latter findings show that the null finding for bioenergetic effects on palmboard adjustments was not due to a lack of statistical power. Especially strong evidence for the dissociation between verbal/ visual and palmboard measures is seen in the following finding: The mapping between conscious awareness and motoric actions is recalibrated for long-term but not short-term changes in the physiological potential of individuals (Bhalla & Proffitt, 1999). In these experiments, participants were verbally instructed to set the palmboard to a range of angles; for example, participants were instructed to set the palmboard to 10°, 30°, 5°, and so forth. When so instructed, they set the palmboard to lesser angles than the ones given; however, the adjustment closely matched the angles that they would have set in response to viewing hills that evoked the same verbally given angle. For example, typical young adults standing at the base of a 5° hill would report that it appeared to be 20°, whereas their palmboard adjustment would be about 5°.1 When later asked to provide palmboard adjustments to a range of verbally given angles, they made a 5° adjustment when instructed to make one of 20°. Such results fit well with the notion that palmboards are accurate due to their accidentally underestimating conscious perceptions. Additional findings, however, provided evidence that the relationship between verbal and palmboard measures is not accidental in this way. Although verbal estimates (e.g., 20°) of hill slant and palmboard adjustments (5°) were consistent with verbal instructions (20°) and resulting palmboard adjustments (5°) in the case described above, this consistency was disrupted when individuals were encumbered or became fatigued. Fatigued participants, for example, would perceive the 5° hill to be steeper—about 25°; however, their palmboard adjustments remained unaffected: when given a verbal instruction to make a palmboard adjustment of 20°, they still provided a 5° response, as was the case before fatigue set in. Although viewing the hill evoked a 25° verbal judgment concomitant with a 5° palmboard adjustment, a 25° verbal instruction evoked a palmboard adjustment of 7°. Thus, in the context of viewing a hill, fatigued participants verbally stated that it was 25° and set the palmboard to 5°, whereas when asked to set the palmboard to 25°, they set it to 7°, thereby indicating a loss of internal consistency. Internal consistency was not lost for people who had adapted to long-term changes in their physiological potential. For example, when participants who were low in physical fitness viewed a 5° hill, they reported that it was 25° but 1 In this example, which will be carried through the next paragraph, the palmboard adjustment values are idealized for expository purposes. As was stated earlier, palmboard adjustments yield an overestimation of hill slants for inclines of less than 20°. Thus, when viewing a 20° hill, a young adult would have likely set the palmboard to about 8°. Moreover, they would also set it to about 8° when verbally instructed to set it to 20°. The issue under discussion is the consistency found for long-term changes in physiological state and the loss of consistency found for short-term changes. The values given for the palmboard adjustments in these examples exaggerate its accuracy for purposes of expository clarity.
made an accurate 5° palmboard adjustment. When asked to make a palmboard adjustment of 25°, they provided a 5° response. These individuals exhibited an internal consistency between their conscious representations and motoric actions (Bhalla & Proffitt, 1999). Such findings of dissociations for short- but not long-term changes in physiological potential are impossible to explain by the account that the accuracy of the palmboard is accidental and due to physiological limitations in wrist flexion. In sum, we initially entertained the possibility that palmboard adjustments were accurate due to constraints on the adjustment response that cancelled the overestimations inherent in perceiving geographical slant. The strongest evidence against this cancellation account is that verbal/visual measures and palmboard adjustments are dissociable. None of the studies presented in Durgin et al. cast doubt on the evidence for the dissociation of these response measures, and instead, they provide additional evidence that palmboard adjustments and explicit measures of geographical slant are dissociable (Experiment 2). A detailed critique of each experiment in Durgin et al. follows. Experiment 1. In Experiment 1, Durgin et al. placed a small wooden wedge with a face incline of 30° on a table in front of participants. Participants made two responses: 1) adjusting a palmboard to match the incline of the wedge, and 2) holding their hand in the air above the palmboard so that the hand's orientation felt parallel to the incline of the wedge. They found that the latter “free-hand” estimates (32.7°) were not reliably different from the actual incline of the wedge, while palmboard estimates (19.4°) were reliably less. The results from Experiment 1 generalize neither to our geographical slant studies nor to the other experiments presented by Durgin et al. First, our slant studies made use of real hills in which the hill crests were over the participants' heads. Second, in our studies, palmboard adjustments to hills of about 30° were close to accurate— not underestimated by over 10° as in the Durgin et al. study; no account was offered by Durgin et al. for this disparity in results. It may be that Durgin et al.'s palmboard was not as easy to adjust as was ours. It seems inappropriate to conclude that all palmboard adjustments made in the context of a 30° slant are underestimated when this was only the case for Durgin et al.'s experiment and not our own. Recall, moreover, that Feresin and Agostini (2007) found almost perfect accuracy with their palmboard for slants between 4° and 18°. The palmboard implementation in the Durgin et al. study is suspect. Finally, in this experiment no verbal estimates were obtained. Thus, there is no way of knowing whether the accurate free-hand responses would have been similar to verbal reports and whether the small wooden incline evoked the large overestimation of slant that is always found for real hills. When drawing generalizations from one study to another, equating experimental conditions is important. Experiment 2. In their second experiment, Durgin et al. used the same two measures as were used in Experiment 1, palmboard and free-hand responding. This study, however, was conducted outdoors in front of a real hill. In this experiment, participants also gave verbal estimates following the two non-verbal measures. An additional factor, gaze direction, was added such that participants gave two sets of estimates with regards to the hill: one while looking down at the base of the hill, and the other while looking straight ahead at the hill. The hill was not exactly uniform—the portion of the hill that participants viewed when their gaze was down was inclined by 23.5°, while the portion that subjects viewed at eye level was slightly steeper at an incline of 24.5°. Durgin et al. found that across both gaze directions the verbal and free-hand estimates were overestimated relative to the actual slope and the palmboard estimates were underestimated. The free-hand
D.R. Proffitt, J.R. Zadra / Acta Psychologica 138 (2011) 285–288
adjustments were no more accurate than those obtained with the palmboard. Moreover, the underestimation found for the palmboard was again inconsistent with our previous results that show relative accuracy for hills of this slant. For hills of comparable slant (18°), Feresin and Agostini (2007) found near perfect accuracy for participants' palmboard adjustments. Again, Durgin et al.'s palmboard implementation is suspect. Importantly, both verbal and free-hand estimates given with a downward gaze direction were significantly less than those given with a forward gaze. There was no significant effect of gaze direction, however, for the palmboard estimates. The results from Experiment 2 show a clear dissociation of palmboard adjustments from free-hand estimates and verbal reports. Only the free-hand estimates exhibited an influence of gaze angle similar to that found for verbal reports. This is evidence that the palmboard estimates are dissociable from verbal reports, whereas the free-hand estimates are not. Durgin et al. argued that the palmboard was too insensitive to detect small changes across gaze conditions; however the internal consistency findings (Bhalla & Proffitt, 1999) described above and the Creem and Proffitt (1998) findings discussed below provide evidence counter to this claim. Experiment 3. For both Experiments 3A and 3B, Durgin et al. presented small, irregularly shaped wooden surfaces anchored to a tripod that could be adjusted to various inclines a short distance in front of the participant. In 3A, participants made free-hand and palmboard estimates of the inclines. In 3B, participants reached out and placed their hand on the surfaces while palm orientation was recorded. The results for Experiment 3A showed that free-hand estimates were overestimated and that palmboard adjustments were underestimated. The experiment suffers from the same problems of generalizability as did Experiment 1: the stimuli are even further removed from the real hills used in our geographical slant studies (in addition to the small surface area of the stimuli, they were surrounded with a hemispheric enclosure of black felt, and participants wore goggles that restricted their view). As in the previous two experiments, the palmboard findings in this Durgin et al. study are inconsistent with the findings of others. This leads to two major issues with the experiment: it cannot address the primary question of dissociability because no verbal estimates were obtained, and the errors in accuracy found for palmboard estimates are inconsistent with the higher accuracy found in other studies. The results cannot be generalized to the other experiments in Durgin et al.'s paper, or to our experiments, because the relationship between free-hand and palmboard estimates cannot be evaluated in terms of perceptual overestimation without knowing the relationship to verbal estimates that would have resulted from this novel set of reduced cue stimuli. In other words, do the small wooden surfaces evoke a perceptual overestimation of their slant, or is the overestimation evident in free-hand responses due to something else, such as an artifact in its use or implementation? There were no verbal estimates, and thus, there is no way of knowing the relationship between free-hand responses and verbal reports and, subsequently, the meaning of any difference between free-hand and palmboard estimates. An additional problem in Experiment 3A is that the accuracy of the palmboard estimates are poor compared to the results obtained by us and those obtained by Feresin and Agostini (2007). This suggests that either the reduced cue stimuli (relative to the full-cue viewing of hills in our and Feresin and Agostini studies) or the different palmboard implementation used by Durgin et al. is responsible for their poorer accuracy. Experiment 3A is sufficiently different from our experiments and those of other researchers, both in stimuli and results, that the results have little bearing on the issue at hand. The reaching task in
287
experiment 3B does not add to the claims made by Durgin et al. They found that people accurately adjust the angle of their palm to match the angle of an inclined surface when reaching for it. If anything, this supports the idea that people are good at using their palms to act on surfaces of various orientations—and with a palmboard, participants are acting on a surface, while free-hand estimates do not involve any object-oriented action. Experiment 4. In the fourth experiment, participants placed their forearm on a horizontal board and then adjusted either their entire forearm (keeping the elbow on the board) or just the hand (keeping the forearm and wrist on the board) to match the orientation of a line depicting various angles between 5° and 60°. When rotating from the elbow, participants were more accurate at matching angles over 30°, and more accurate when rotating the wrist only for angles below 30°. Durgin et al. motivated Experiment 4 with the claim that palmboard adjustments involve rotating a single joint, the wrist, and that the lack of flexibility in the wrist is one cause of the “accidental accuracy” of palmboard estimates. Adjusting the palmboard, however, involves rotating the shoulder, elbow, and wrist, and as was discussed above, participants are able to easily set the board to a wide range of angles up to at least 60°. You can experience these motions of the arm and shoulder by attempting the following: Hold your hand level at waist height and position this hand about 10 inches in front of your body. Flex your wrist to its maximum and observe the slant of your hand. Now, attempt to increase this slant. This is easily achieved by slightly flexing the knees, dropping the shoulder, and bending the elbow. Both conditions in Experiment 4 placed far more limitations on the biomechanical movement needed to match orientations than does a palmboard, and thus, the results of this study do not generalize to the palmboard. Experiment 5. In Experiment 5, Durgin et al. asked participants to remember three well known hills on campus and to set the palmboard to match the incline of each hill one at a time. Participants did this either while seated with the palmboard set just above their navel, standing with the palmboard set just above their navel, or standing in the posture used in Proffitt et al. (1995) with the palmboard at waist height. Participants, who had the palmboard set just above their navel, whether seated or standing, provided higher estimates than the third group. Experiment 5 was motivated by a need to explain the memory results from Creem and Proffitt (1998), mentioned above. In the Creem and Proffitt study, participants were asked to judge hill slants from memory following a delay. With a short delay of under a minute (and while still in the presence of the hill), palmboard estimates matched normative palmboard estimates given while subjects were viewing the hill. With a longer delay (and away from the hill) palmboard estimates were significantly increased. This suggests that with a short delay actions can be guided by short-term visual guidance memory, but after a longer delay this visual guidance memory has dissipated and the action must draw on the conscious, explicit memory, which reflects overestimation. These results cannot be explained by a limitation in wrist flexion. They show that verbal reports and palmboard estimates are dissociable when responses are based upon perception but not dissociable when slant is judged from memory. These findings also demonstrate that the palmboard measure is sufficiently sensitive to reveal the presence or absence of verbal/palmboard dissociations. Durgin et al. motivated their experiment with the assertion that participants in Creem and Proffitt (1998) were seated during some palmboard estimates, and that this confounding of posture and memory accounted for the results. Prior to the submission of Durgin et al.'s paper to this journal, we informed Durgin that, to the best of our
288
D.R. Proffitt, J.R. Zadra / Acta Psychologica 138 (2011) 285–288
knowledge, the participants in Creem and Proffitt (1998) were standing when making palmboard adjustments from memory. In none of our other studies have participants ever been seated while adjusting the palmboard, and neither Creem-Regehr nor I (DRP) have any recollection of having participants sit during palmboard estimates in this experiment. Moreover, in Experiment 4 of Creem and Proffitt (1998), participants in all conditions were unquestionably standing while adjusting the palmboard from memory, yet the effects were the same. In all of our experiments, the palmboard was positioned at waist level. We never conducted a study in which the palmboard was positioned above the navel. Durgin et al.'s Experiment 5 misrepresents the experimental design of Creem and Proffitt (1998), and thus, its design is without motivation and its results have no bearing on the issues at hand. Feresin and Agostini (2007) have already shown that posture matters in palmboard adjustments. By carefully positioning their palmboard to optimize ease of use, Feresin and Agostini found far better accuracy, for slants between 4° and 18°, than were ever obtained by us. In summary, there is considerable evidence that palmboard adjustments are dissociable from verbal/visual measures. In Experiment 2, Durgin et al. provided additional evidence for their dissociability. In their general discussion, Durgin et al. proposed an account by which our bioenergetic effects on geographical slant perception are due to demand characteristics. We have addressed these claims elsewhere (Proffitt, 2010) and will not do so again here.
References Bhalla, M., & Proffitt, D. R. (1999). Visual-motor recalibration in geographical slant perception. Journal of Experimental Psychology: Human Perception and Performance, 25, 1076−1096. Creem, S. H., & Proffitt, D. R. (1998). Two memories for geographical slant: Separation and interdependence of action and awareness. Psychonomic Bulletin & Review, 5, 22−36. Durgin, F. H., Hajnal, A., Li, Z., Tonge, N., & Stigliani, A. (2010). Palm boards are not action measures: An alternative to the two-systems theory of geographical slant perception. Acta Psychologica, 134(2), 182−197. Feresin, C., & Agostini, T. (2007). Perception of visual inclination in a real and simulated urban environment. Perception, 36, 258−267. Gossweiler, R., Proffitt, D. R., Bhalla, M., & Pausch, R. (1995). A hill study: Using a virtual environment as a perceptual psychology laboratory.IEEE computer: Special issue on real applications for virtual reality March. Kammann, R. (1967). The overestimation of vertical distance and slope and its role in the moon illusion. Perception & Psychophysics, 2, 585−589. Proffitt, D. R. (2010). Affordances matter in geographical slant perception: Reply to Durgin, Baird, Greenburg, Russell, Shaughnessy, and Waymouth. Psychonomic Bulletin & Review, 16, 970−972. Proffitt, D. R., Bhalla, M., Gossweiler, R., & Midgett, J. (1995). Perceiving geographical slant. Psychonomic Bulletin & Review, 2, 409−428. Ross, H. E. (1974). Behavior and perception in strange environments. London: George Allen & Unwin. Schnall, S., Zadra, J. R., & Proffitt, D. R. (2010). Direct evidence for the economy of action: Glucose and the perception of geographical slant. Perception, 39, 464−482.