#
373
British Journal of Psychology (1999), 90, 373±388 Printed in Great Britain
1999 The British Psychological Society
Role of memorization conditions in
the haptic processing of orientations
and the ` oblique eåect ’
Edouard Gentaz*
University of Geneva, Switzerland
Yvette Hatwell
University Pierre Mende[ s France of Grenoble, France
Ê
Ê
The haptic processing of vertical, horizontal, 45 and 135 oblique orientations was
studied in blindfolded sighted adults in an exploration±reproduction task. The
purpose was to determine whether the variations of the memorization conditions
between the exploration and reproduction phases would in¯uence the global
performance and the oblique eåect (lower performance in oblique orientations than
in vertical±horizontal orientations). If orientation coding depended on attentional
resources, the increase in memory constraints would aåect the haptic processing of
orientations and the oblique eåect. Memory constraints were therefore varied by
changing the length and the nature of the delay in two tasks in which previous
research has shown that the availability of gravitational cues aåected orientation
coding.
Blindfolded adults were asked to explore haptically a rod with minimal (Expt 1)
or natural (Expt 2) gravitational cues and then to reproduce the orientation of this
rod ipsilaterally after one of four memorization conditions: with 5 s or 30 s un®lled
delays, and 30 s delays ®lled with verbal or haptic interpolated tasks. When the delay
was un®lled, whatever its length (5 s or 30 s), the performance depended on the
conditions of manual exploration: the oblique eåect was absent when the
gravitational cues were minimal (Expt 1) and was present when these cues were
natural (Expt 2). By contrast, when the delay was ®lled with interpolated tasks, the
haptic oblique eåect was present whatever the conditions of manual exploration.
Taken together, these results showed that memorization conditions played a role in
the haptic processing of orientations and in the oblique eåect when the gravitational
cues were minimal during manual exploration.
This study addresses the question of human processing of vertical, horizontal and
oblique orientations in the haptic (tactual-kinesthetic) system. The experimenters
were interested in examining this spatial property because previous experiments had
shown that orientation processing diåered according to the value of the orientation
and to the perceptual system activated. In the visual system, the vertical and
* Requests for reprints should be addressed to Dr Edouard Gentaz, Faculte! de Psychologie, Universite! de Gene’ ve,
Route de Drize 9, 1227 Carouge, Switzerland (e-mail : Edouard.Gentaz! pse.unige.ch).
374
Edouard Gentaz and Yvette Hatwell
horizontal orientations are known to be perceived more accurately than the oblique
orientations. This anisotropy, which Appelle (1972) called the `oblique eåect ’, is
nearly always present, whatever the variety of tasks used (detection, discrimination,
identi®cation, recognition, memorization ; Appelle, 1972 ; Essock, 1980 ; Howard,
1982). The mechanisms underlying the processing of orientations and generating the
visual oblique eåect seem to be multi-component and occur at diåerent levels of
processing (for recent reviews, see Gentaz & Ballaz, in press; Heeley & BuchananSmith, 1990 ; Saarinen & Levi, 1995).
By contrast, in the haptic system, the oblique eåect has been explored less often and
with one type of task only: an exploration±reproduction task in which blindfolded
participants were asked to scan a rod with one hand and to reproduce its orientation
with the same or other hand. The studies showed that the oblique eåect is not always
present. Consequently, the discussion about the haptic oblique eåect is centred on its
existence and the factors which contribute to its manifestation. Lechelt and his
associates (Lechelt, Eliuk & Tanne, 1976 ; Lechelt & Verenka, 1980) asked adults to
explore a standard rod with one hand and to reproduce its orientation with the other.
A haptic oblique eåect was observed with simultaneous and successive (after 0 s or
10 s un®lled delays) reproduction of the stimulus. The results showed in addition
that absolute errors in all orientation settings were higher in the simultaneous
reproduction than in the successive reproduction after a 0 s un®lled delay (Lechelt
et al., 1976). But in oblique settings only, errors were higher in the simultaneous
reproduction than in the successive reproduction after a 10 s un®lled delay (Lechelt
& Verenka, 1980). According to these authors, the presence of the haptic oblique
eåect in the diåerent experimental conditions revealed the importance of experiential
and learning factors: because verticals and horizontals are privileged in the
` carpentered environment’, the same oblique eåect is at work in vision and haptics.
However, Appelle & Countryman (1986) questioned this interpretation. They
argued that the haptic oblique eåect demonstrated by Lechelt and his associates,
above, stemmed from the inherently diåerent scanning patterns required in the
exploration of obliques when one hand explored the standard and the other set the
response rod (` contralateral condition’) as was the case in Lechelt’s experiments.
This assumption was supported by the fact that Appelle & Countryman (1986) found
no haptic oblique eåect in an ` ipsilateral condition’ in which the same hand explored
the standard rod and, after a 5 s un®lled delay, set the response rod.
However, these studies did not control the plane in which the task was performed.
Gentaz & Hatwell (1995) asked blindfolded sighted adults to reproduce the rod
orientation after a 5 s un®lled delay in the ipsilateral and contralateral conditions,
either in the horizontal plane (like the surface of a table, as did Appelle &
Countryman, 1986), or in the frontal plane (like the surface of a blackboard on a wall,
as did Lechelt et al., 1976 ; Lechelt & Verenka, 1980), or in the sagittal plane (in the
median plane passing through the midline of the participant’s head). In this latter
plane, the same pattern of agonist} antagonist muscles and tendons is activated
when one hand explores the vertical, horizontal and oblique rods and the other hand
reproduces these orientations. By contrast, in the horizontal and frontal planes,
diåerent patterns of agonist} antagonist muscles and tendons are activated when one
hand explores an oblique rod and the other hand reproduces this oblique.
Memorization in haptics
375
Nevertheless, Gentaz & Hatwell (1995) found that the haptic oblique eåect was
present in the sagittal plane, even when the same hand explored and set the response
rod. Therefore, this pattern of results did not support Appelle & Countryman’s
(1986) interpretation.
Gentaz & Hatwell (1996) showed further that the gravitational cues provided by
the antigravitational forces developed by the scanning arm-hand system were
involved in the haptic oblique eåect, because these cues specify the gravity direction
as one axis of reference. Gravitational cues are de®ned as indirect cues taken from the
proprioceptive information contained in the pattern of tissue deformations (skin
mechanoreceptors, joints and muscles). These deformations depend on the speci®c
muscular forces necessary to maintain and move the arm-hand system during manual
exploration. The magnitude of these gravitational cues varies according to the plane
in which the task is performed. In the frontal and sagittal planes, antigravitational
forces are not the same in the diåerent directions of exploration : the bottom±up
direction requires high muscular antigravitational forces, whereas the reverse
direction requires lower muscular eåort. By contrast, in the horizontal plane, the
variability of antigravitational forces (and therefore of the gravitational cues they
generate) is minimal : because the participant’s movements are performed in front of
the participant in a direction perpendicular to gravity, they require constant
antigravitational forces.
Gentaz & Hatwell (1996) modi®ed gravity constraints by testing blindfolded
adults, in the horizontal plane, either with the forearm resting on the disc surface
supporting the rod (`supported forearm’ condition), which was similar to the
condition tested by Appelle & Countryman (1986) and Gentaz & Hatwell (1995), or
with the forearm unsupported in the air. In this latter case, greater antigravitational
forces were elicited during scanning. In the horizontal plane, the so-called `vertical ’
orientation corresponds actually to the mid-sagittal orientation, and the so-called
` horizontal’ orientation corresponds to the fronto-parallel orientation. The results
revealed that the oblique eåect was present in the unsupported forearm condition and
was absent in the supported forearm condition. Another experiment tested the three
planes, either in a ` natural’ (unsupported forearm) or in a `lightened forearm’
condition in which the gravitational cues were reduced by lightening the participant’s
forearm (through a device consisting of a set of weights and pulleys). In all planes,
the magnitude of the oblique eåect (i.e. the magnitude of the error diåerence between
vertical±horizontal orientations and oblique orientations) was lower in the lightened
condition than in the natural condition. In addition, Gentaz & Hatwell (1998)
showed that, even in those blind since birth, the variability of gravitational cues was
involved in the presence or absence of the haptic oblique eåect.
This set of ®ndings revealed the existence, in some exploratory conditions, of a
purely haptic oblique eåect. They showed in addition that this eåect was in¯uenced
by the conditions of exploration because these conditions changed the gravity
constraints of scanning. Consequently, the debate may be focused now on the nature
of the mechanisms underlying the processing of orientations and generating or not
the haptic oblique eåect. Actually, the presence or absence of this eåect seems to
indicate that diåerent modes of coding are at work in each case (Gentaz, in press;
Gentaz & Hatwell, 1996, 1998). It is proposed that the presence of the haptic oblique
376
Edouard Gentaz and Yvette Hatwell
eåect reveals that orientation is de®ned by a spatial coding relative to the geocentric
frame of reference (i.e. based on the direction of gravity) in which the vertical and
horizontal orientations are used as reference norms. This latter mode of coding
necessarily involves some spatial processing relating the haptic inputs extracted from
the currently explored region of the stimulus to the two norms. Thus, in a given
plane, any oblique orientation would be encoded in relation to the vertical and
horizontal norms, whereas the vertical or horizontal orientations would be encoded
directly. The indirect encoding of the oblique orientations, which implies the
integration of the parameters computed on the two norms, may require higher costs
of processing than the direct encoding of vertical and horizontal orientations which
refers to one norm only. The diåerence between these two types of encoding would
largely explain the occurrence of the haptic oblique eåect. By contrast, the absence
of the haptic oblique eåect would reveal that orientation is de®ned by `movement
coding’ (for review, see Millar, 1994). Each orientation would be coded in terms of
a kinesthetic sequence of movements which can be remembered. This latter mode of
coding does not necessarily involve that the haptic inputs extracted from the
explored stimulus are encoded in relation to the vertical and horizontal norms. Thus,
the possibility that the vertical, horizontal and oblique orientations are encoded in
the same manner would explain the absence of the haptic oblique eåect.
The purpose of the present study was to study the nature of these two assumed
modes of orientation coding (geocentric reference frame vs. movement coding) by
examining whether these modes depend on attentional resources. In the case of a
positive answer, this would mean that orientation coding would rely on high levels
of haptic processing. To study this question, the memorization conditions were
modi®ed in the haptic processing of vertical, horizontal and oblique orientations. In
previous studies, the delay between the exploration and reproduction phases was the
same in all experimental conditions: 5 s and un®lled (Appelle & Countryman, 1986 ;
Gentaz & Hatwell, 1995, 1996, 1998). If the modes of coding depended on
attentional resources, the increase in memory constraints would aåect the haptic
processing of orientations. In the connected ®eld of the kinesthetic processing of
distance and localization, previous studies concerning the role of memory reported
inconsistent observations. For example, Keele & Ellis (1972) observed similar
variable errors with a 0 s un®lled delay and with a 7 s un®lled delay. By contrast,
Jones & Connolly (1970) showed that absolute errors were higher after a 10 s un®lled
delay than after a 0 s un®lled delay. Similarly, Stelmach & Wilson (1970) and Posner
(1967) revealed that absolute errors were higher after a 20 s un®lled delay than after
a 0 s un®lled delay. The ®ndings about the eåect of interpolated tasks during the
delay are also inconsistent. Keele & Ellis (1972) showed that variable errors were
higher after a 7 s ®lled delay than after a 7 s un®lled delay. Similarly, Stelmach &
Wilson (1970) showed also that absolute errors were higher after a 20 s ®lled delay
than after a 20 s un®lled delay. By contrast, Jones & Connolly (1970) and Posner
(1967) observed no diåerence between a 20 s un®lled delay and a 20 s interpolated®lled delay. Further research should clarify these relations between memory
constraints and the haptic processing of spatial properties. As concerns the haptic
processing of orientations, there are no available studies in which the nature of delay
was varied.
Memorization in haptics
377
In the present study, changes in memorization conditions were proposed to
blindfolded adults. Condition 1 (5 s un®lled delay condition) replicated the conditions
proposed by Appelle & Countryman (1986) and Gentaz & Hatwell (1995, 1996) in
which the delay was 5 s and un®lled. In Condition 2 (30 s un®lled delay condition),
the delay was 30 s and un®lled. In Conditions 3 and 4, the 30 s delay was ®lled
respectively with a verbal or haptic interpolated task (in which the participants were
asked to perform an attentional demanding task). These two types of interpolated
tasks were used because some previous studies had shown that they could aåect
performances diåerently (e.g. Diewert, 1975). If the modes of orientation coding
were not stable over time, the performance would be aåected by the increase in the
un®lled delay length. Moreover, if the modes of coding required high attentional
resources, orientation processing would be aåected by the introduction of the
interpolated tasks during the delay. In order to know whether the modes of
orientation coding were aåected diåerently by the increase in the memory constraints,
the four memorization conditions were proposed after the two types of exploratory
conditions in the horizontal plane. The horizontal plane was chosen here because,
within this, the haptic oblique eåect is present when the gravitational cues are
natural, and it is absent when these cues are reduced (Gentaz & Hatwell, 1996).
Therefore, in Expt 1, the gravitational cues were minimal during manual exploration,
and orientation was de®ned in this case by movement coding. In Expt 2, the
gravitational cues were natural, and orientation was de®ned in this case by a spatial
coding relative to the geocentric reference frame.
EXPERIMENT 1
The aim of this experiment was to ascertain whether the mode of orientation coding
based on a sequence movement depended on attentional resources. In this case, the
haptic processing of orientations would be aåected by an increase in the memory
constraints during the delay. In the present experiment, the gravitational cues were
minimal during manual exploration, and it is likely that this situation generated a
movement-based coding (Gentaz & Hatwell, 1996). If this mode of coding was not
stable over time, performances should be lower in the 30 s un®lled delay condition
than in the 5 s un®lled delay condition. If this mode of coding required high
attentional resources, performances should be lower in the 30 s haptic and} or verbal®lled delay conditions than in the 30 s un®lled delay condition. By contrast, if this
movement-based coding did not depend on attentional resources, performances
should be similar in the four memorization conditions and, thus, the haptic oblique
eåect should remain absent in all conditions.
Method
Participants
In all, 32 right-handed undergraduate students in psychology and sociology (16 females, 16 males)
participated in the experiment in exchange for extra course credit. Participants were aged between 19
and 22 years. Their handedness was assessed with the use of Bryden’s 5-item (1977) hand preference
questionnaire (writing, throwing, drawing, scissors and toothbrush). In conformity with the Helsinki
378
Edouard Gentaz and Yvette Hatwell
Convention which controls and regulates human experimentation, informed consent was obtained from
all participants.
Apparatus
The apparatus was composed of a black metal disc (diameter 40 cm) equipped with a rod (25 3 1.8 cm).
This rod was mounted on the centre of the disc and could be rotated 360 around its central axis.
Magnets were ®xed inside the rod to maintain it in the desired orientation and to prevent involuntary
deviation from its position during haptic scanning. A small amount of force was required to change
intentionally its position. This rod was ®xed directly in contact with the disc (Fig. 1). The rod was
connected with a potentiometer which indicated (in degrees) to what extent its orientation deviated
from the orientation aligned with the participant’s mid-sagittal orientation. This orientation was
arbitrary labelled as the `vertical ’ orientation in the horizontal plane in the present research. The discrod was centred on the participant’s body midline. The participants sat in the front midline of the
apparatus, perpendicular to the surface of the disc, and were asked to maintain their trunk ®xed upright.
The height of the disc was adjusted with reference to the participant’s height, so that the rod was
positioned at a level under the participant’s breast.
Ê
Stimulus rod
Disc
Potentiometer
Figure 1. Lateral view of the device, with the disc positioned in the horizontal plane and the rod ®xed
directly in contact with the disc. In this situation, the participant’s forearm rested on the disc surface.
Experimental conditions and procedure
The participant was taken individually into a quiet room in which the apparatus was masked by a cloth,
and was then blindfolded with glasses which passed no visual information. The task was to scan
haptically with one hand (i.e. to move actively the arm-hand-digit system) on the oriented standard rod
and then to reproduce the same orientation with the same hand. The participant was instructed to
maintain his or her wrist and forearm resting in physical contact with the disc support during the whole
trial. In a familiarization phase, the arm-hand system was ®rst guided by the experimenter and the
Memorization in haptics
379
participant was trained to explore actively under these instructions. No time constraint was imposed in
the presentation and response phases. The orientation of the rod between the presentation and response
phases was modi®ed because the same rod orientation was used as a standard and a response alternately.
The initial orientation of the rods in the response phase was 6 15 or 6 25 relative to the standard
orientation tested during the trial and was determined at random by the experimenter.
Each participant was tested under one of the four memorization conditions introduced between the
exploration and reproduction phases. In the ®rst memorization condition (5 s un®lled delay), a 5 s
un®lled delay was introduced to replicate Appelle & Countryman’s (1986) and Gentaz & Hatwell’s
(1995) condition in which the oblique eåect was absent in previous studies. In the second condition (30 s
un®lled delay), the delay was 30 s and un®lled. In the third condition (30 s verbal-®lled delay), the delay
was 30 s and ®lled with a verbal interpolated task in which the participant was asked to recall the letters
of the alphabet in reverse order. During these three memorization conditions, the participant was asked
to move his or her hand oå the display and to maintain it in contact with his or her abdomen. In the
last condition (30 s haptic-®lled delay), the delay was 30 s and was ®lled with a haptic interpolated task
in which the participant was asked to move the index ®nger from one point to the other along a raised
path presented on a wooden board in the horizontal plane. For each memorization condition, half the
participants used the right hand, and the other half used the left hand.
Four orientations were tested in the horizontal plane and labelled arbitrary as in the previous studies :
0 (vertical ; actually mid-sagittal orientation), 1 45 (left-slanting oblique in relation to participant’s
view), 1 90 (horizontal; actually fronto-parallel orientation), and 1 135 (right-slanting oblique in
relation to participant’s view). Three additional catch (distractor) trials in other orientations (randomly
selected from 1 15 , 1 75 , 1 105 and 1 165 ) were interspersed with test trials to avoid inducing any
awareness that the tested orientations were restricted to four. Since each participant performed four
trials in each of the four orientations tested, there were 16 test trials per participant. In addition, as stated
above, three catch trials were presented randomly. No feedback was given to the participant during the
experiment. The duration of the whole session was about 40 min in the 30 s delay conditions.
Ê
Ê
Ê
Ê
Ê Ê
Ê
Ê
Ê
Ê
Results
Data analyses were carried out on the angular diåerence (in degrees) between the
positions of the standard and the response rod. The mean error (signed and absolute)
was then calculated for each of the four orientations for each participant. In order to
know whether the haptic oblique eåect was in¯uenced by the memorization
conditions, the mean absolute errors were analysed. Because absolute errors have
generally an asymmetrical distribution, a logarithmic transformation (log(x1 1)) was
carried out on each angular diåerence to ful®l the normality condition required for
the ANOVA (Abdi, 1987 ; Gentaz & Hatwell, 1996). The Lilliefors test (Lilliefors,
1967) for normality showed that the diåerence between the normal and the empirical
distribution of scores in degrees (before transformation)was signi®cant (L(calculated)
5 .1048 " L(critical) 5 .091, p ! .01), whereas the diåerence between the normal
and the empirical scores in log after transformation was not (L(calculated) 5 .0575
! L(critical)). Table 1 indicates the mean absolute errors before transformation (in
degrees} deg) and after their log transformation (log). In the following analysis of
results, the mean values are presented both in degrees and in log (mean in deg; mean
in log). A preliminary ANOVA on absolute errors in log showed that gender and
hand had no eåect and did not interact with any other factor. Consequently, results
were collapsed across gender and hand.
A 4 (memorization condition) 3 4 (orientation) ANOVA (with repeated measures
on the last factor) on absolute errors in log revealed a main eåect of orientation
(F(3,84) 5 3.5, p ! .05). As it is classically done in literature, the measurement of the
Edouard Gentaz and Yvette Hatwell
380
Table 1. Mean absolute errors (M ) in degrees (deg) and in log (log) and standard
deviations of means (SD) as a function of memorization condition and stimulus
orientation in Expt 1
Orientation
1
Vertical
Memorization condition
5 s un®lled delay
deg
log
30 s un®lled delay
deg
log
30 s verbal-®lled delay
deg
log
30 s haptic-®lled delay
deg
log
45
Ê
Horizontal
1
135
Ê
M
(SD)
M
(SD)
M
(SD)
M
(SD)
4.9
.72
(3.2)
(.21)
4.1
.69
(1.4)
(.12)
4.6
.72
(2.3)
(.16)
4.7
.73
(1.9)
(.16)
6.3
.80
(4.1)
(.25)
7.0
.82
(5.5)
(.28)
6.2
.78
(5.9)
(.25)
6.3
.85
(1.9)
(.14)
6.1
.76
(5.4)
(.29)
9.7
.99
(4.6)
(.19)
5.7
.78
(3.4)
(.23)
9.9
.99
(5.4)
(.22)
7.3
.83
(5.0)
(.30)
9.0
.99
(1.5)
(.06)
6.6
.81
(4.2)
(.29)
9.1
1.0
(1.7)
(.07)
oblique eåect was carried out by a pre-planned contrast between the vertical and
horizontal orientations on the one hand, and the 45 ±135 oblique orientations on the
other hand. The measure of the oblique eåect was legitimate because partial analyses
showed no diåerence between the vertical (M in deg 5 6.1 ; M in log 5 .78) and
horizontal (M in deg 5 5.8 ; M in log 5 .77) orientations (F(1,28) 5 .3, p " .25), nor
between the 45 (M in deg 5 7.5 ; M in log 5 .88) and 135 (M in deg 5 7.5 ; M in
log 5 .88) oblique orientations (F(1,28) 5 .2, p " .25). The magnitude of the
oblique eåect (i.e. the diåerence between the magnitude of errors when setting
vertical±horizontal and 45 ±135 obliques) was signi®cant (F(1,28) 5 9.5, p ! .01)
with more errors in the 45 ±135 oblique orientations (M in deg 5 7.5 ; M in log 5
.88) than in the vertical±horizontal (M in deg 5 6.0 ; M in log 5 .78) orientations.
The main eåect of memorization condition was signi®cant (F(3,28) 5 3, p ! .05).
There was no diåerence between the 30 s verbal-®lled delay (M in deg 5 7.9 ; M in
log 5 .88) and the 30 s haptic ®lled delay (M in deg 5 8.0 ; M in log 5 .91) conditions
(F(1,28) 5 .2, p " .25), not between the 5 s un®lled delay (M in deg 5 4.6 ; M in
log 5 .72) and 30 s un®lled delay (M in deg 5 6.5 ; M in log 5 .81) conditions
(F(1,28) 5 1.9, p 5 .18). Errors were higher in the two ®lled delay conditions than
in the two un®lled delay conditions (F(1,28) 5 6.9, p ! .05).
The orientation 3 memorization condition interaction was not signi®cant (F(9,84)
5 .85, p " .25). However, when the pre-planned decomposition of this interaction
was carried out (Fig. 2), the oblique eåect was absent in the 5 s un®lled delay
(obliques: M in deg 5 4.4, M in log 5 .71 ; vertical±horizontal : M in deg 5 4.7, M
Ê Ê
Ê
Ê
Ê Ê
Ê Ê
Memorization in haptics
381
in log 5 .72 ; F(1,28) 5 .01, p " .25) and in the 30 s un®lled delay (obliques: M in
deg 5 6.7, M in log 5 .82 ; vertical±horizontal: M in deg 5 6.3, M in log 5 .79 ;
F(1,28) 5 .38, p " .25) conditions. By contrast, the oblique eåect was present in the
30 s verbal-®lled delay (obliques : M in deg 5 9.8, M in log 5 .99 ; vertical±
horizontal: M in deg 5 5.9, M in log 5 .77 ; F(1,28) 5 9.9, p ! .05) and in the 30 s
haptic-®lled delay (obliques : M in deg 5 9.0, M in log 5 .99 ; vertical±horizontal: M
in deg 5 6.9, M in log 5 .82 ; F(1,28) 5 6.2, p ! .05). The errors in the oblique
settings were higher in the two 30 s ®lled conditions than in the two un®lled
conditions (F(1,28) 5 13, p ! .05), whereas performance in the vertical and
horizontal settings was similar in the four memorization conditions (F(3,28) 5 .28,
p " .25). This means that the presence of the oblique eåect was explained by
increased errors in the oblique settings only.
1.2
Mean angular absolute errors in log
1.1
1.0
0.9
0.8
0.7
0.6
5 s unfilled delay
0.5
30 s unfilled delay
0.4
30 s verbal-filled delay
0.3
0.2
30 s haptic-filled delay
0.1
0
0°
45°
90°
135°
Stimulus orientation
Figure 2. Mean absolute angular errors in log as a function of memorization condition and stimulus
orientation in Expt 1.
The analyses of mean signed errors are not shown in the present study because
they did not reveal signi®cant directional biases. These results are consistent with
previous studies (Appelle & Countryman, 1986 ; Gentaz & Hatwell, 1995, 1996,
1998 ; Gentaz, Luyat, Cian, Hatwell, Barraud & Raphel, 1999).
Discussion
The results showed that orientation processing was in¯uenced by the nature of
memorization conditions. The increase in the un®lled delay length did not modify the
haptic processing of orientations: performances were similar in the 5 s un®lled delay
382
Edouard Gentaz and Yvette Hatwell
and 30 s un®lled delay conditions. The haptic oblique eåect was absent in these two
un®lled delay conditions. The absence of this eåect in the 5 s un®lled delay condition
replicated Appelle & Countryman’s (1986) and Gentaz & Hatwell’s (1995, 1996)
®ndings, whereas the present study showed in addition that lengthening this un®lled
delay had no eåect. By contrast, the introduction of verbal or haptic interpolated
tasks during the 30 s delay aåected in the same manner the processing of orientations.
Indeed, errors were higher in the two ®lled delay conditions (which did not diåer)
than in the two un®lled delay conditions. This decreasing stemmed from the
increased errors of oblique settings, whereas the errors in vertical±horizontal settings
remained stable in all delay conditions. Consequently, the haptic oblique eåect was
present in the two ®lled delay conditions, although in this experiment the
gravitational cues were minimal during manual exploration in the horizontal plane.
These results suggest that the mode of orientation coding based on movement,
which is stable over time in un®lled delay conditions, requires attentional resources.
Expt 2 was further designed to extend this study to an exploratory condition in
which the gravitational cues were natural.
EXPERIMENT 2
The aim of this experiment was to examine whether the mode of orientation coding
based on the geocentric frame of reference depended on attentional resources. If this
was the case, the haptic processing of orientations would be altered by the increase
in the memory constraints during the delay. In the present experiment, the
gravitational cues were natural during manual exploration and were therefore
enhanced as compared to Expt 1. The four memorization conditions were the same
as in Expt 1. If the mode of coding based on the geocentric reference frame was not
stable over time, performances should be lower in the 30 s un®lled delay condition
than in the 5 s un®lled delay condition. If this mode of coding required high
attentional resources, performances should be lower in the two 30 s ®lled delay
conditions than in the 30 s un®lled delay condition. By contrast, if this mode of
coding did not depend on attentional resources, performances should be similar in
the four memorization conditions and, thus, the same haptic oblique eåect should be
present in all conditions.
Method
Participants
In all, 32 right-handed undergraduate students in psychology and sociology (16 females, 16 males) took
part in the experiment in exchange for extra course credit. They were aged between 18 and 22 years.
Participants’ handedness was assessed with the use of Bryden’s ®ve-item (1977) hand preference
questionnaire. Informed consent was obtained from all participants.
Apparatus and procedure
The apparatus was the same as in Expt 1, except for the position of the rod-stimulus. In this experiment,
the rod was parallel to the disc surface at 8 cm from it (Fig. 3). Therefore, antigravitational forces should
Memorization in haptics
383
be elicited in order to explore the rod. Thus, the magnitude of gravitational cues was higher in the
present study than in Expt 1. The force required to change the position of the rod, the experimental
conditions and the procedure were the same as in Expt 1.
Stimulus rod
Disc
Potentiometer
Figure 3. Lateral view of the device, with the disc positioned in the horizontal plane and the rod ®xed
8 cm from the disc. In this situation, the participant’s forearm did not rest on the disc surface.
Results
The mean absolute errors (in degrees and in log) for each condition are shown in
Table 2. As in Expt 1, a logarithmic transformation (log(x1 1)) was carried out on
each angular diåerence for each participant in each condition. A preliminary
ANOVA on the absolute errors in log showed that gender and hand had no eåect
and did not interact with any other factor. Consequently, results were collapsed
across gender and hand.
A 4 (memorization conditions) 3 4 (orientation) ANOVA (with repeated measures
on the last factor) on absolute errors in log revealed a main eåect of orientation
(F3,84) 5 18.7, p ! .01). Partial analyses showed no diåerence between the vertical
(M in deg 5 4.5 ; M in log 5 .70) and horizontal (M in deg 5 4.6 ; M in log 5 .71)
orientations (F(1,28) 5 .11, p " .25), nor between the 45 (M in deg 5 7.6 ; M in log
5 .90) and 135 (M in deg 5 7.6 ; M in log 5 .90) oblique orientations (F(1,28) 5
.12, p " .25). The oblique eåect was signi®cant (F(1,28) 5 59, p ! .01) : errors were
higher in the 45 ±135 oblique orientations (M in deg 5 7.6 ; M in log 5 .90) than
in the vertical±horizontal orientations (M in deg 5 4.5 ; M in log 5 .70). The main
eåect of memorization conditions was not signi®cant (F(3,28) 5 1.9, p 5 .16). The
orientation 3 memorization condition interaction was not signi®cant (F(9,84) 5 .67,
Ê
Ê Ê
Ê
Edouard Gentaz and Yvette Hatwell
384
Table 2. Mean absolute errors (M ) in degrees (deg) and in log (log) and standard
deviations of means (SD) as a function of memorization condition and stimulus
orientation in Expt 2.
Orientation
1
Vertical
Memorization condition
5 s un®lled delay
deg
log
30 s un®lled delay
deg
log
30 s verbal-®lled delay
deg
log
30 s haptic-®lled delay
deg
log
45
Ê
Horizontal
1
135
Ê
M
(SD)
M
(SD)
M
(SD)
M
(SD)
3.1
.58
(1.5)
(.21)
7.7
.93
(2.1)
(.10)
4.1
.67
(2.2)
(.19)
6.6
.84
(3.2)
(.21)
3.9
.68
(1.5)
(.12)
6.6
.86
(2.9)
(.17)
4.3
.70
(1.6)
(.15)
6.3
.82
(2.8)
(.23)
5.7
.76
(4.4)
(.27)
7.9
.92
(3.3)
(.18)
5.0
.74
(2.9)
(.21)
9.1
.99
(3.2)
(.11)
5.1
.78
(1.0)
(.08)
8.3
.96
(2.3)
(.11)
4.9
.73
(2.6)
(.21)
8.3
.95
(2.3)
(.12)
1.2
Mean angular absolute errors in log
1.1
1.0
0.9
0.8
0.7
5 s unfilled delay
0.6
0.5
30 s unfilled delay
0.4
30 s verbal-filled delay
0.3
0.2
30 s haptic-filled delay
0.1
0
0°
45°
90°
135°
Stimulus orientation
Figure 4. Mean absolute angular errors in log as a function of memorization condition and stimulus
orientation in Expt 2.
p " .25 ; Fig. 4). This means that the oblique eåect was signi®cant in the four
memorization conditions and that the memorization conditions did not in¯uence the
performance in the four orientation settings.
Memorization in haptics
385
As in Expt 1, the analyses of mean signed errors are not shown in the present study
because they did not reveal signi®cant directional bias.
Discussion
The results revealed that the haptic processing of orientations was not in¯uenced by
the diåerent memorization conditions. The hepatic oblique eåect was present in the
four memorization conditions. The presence of the haptic oblique eåects in the 5 s
un®lled delay condition replicated Gentaz & Hatwell’s (1966) ®ndings in which the
oblique eåect was present when gravitational cues were available in the horizontal
plane. The present research revealed in addition that this oblique eåect was equally
present in the other three memorization conditions. These results showed that the
increase in memory constraints during the delay did not aåect signi®cantly the global
accuracy of responses and did not in¯uence the haptic oblique eåect. They suggest
therefore that the mode of coding based on the geocentric reference frame, which is
stable over time, is independent from attentional resources.
GENERAL DISCUSSION
This research studied the role of memorization conditions in the haptic processing
of orientations and the oblique eåect. If orientation coding depended on attentional
resources, the increase in memory constraints would aåect performances. This
question was investigated by asking blindfolded adults to explore, in the horizontal
plane, vertical (mid-sagittal), horizontal (fronto-parallel), and 45 and 135 oblique
rod orientations with minimal (Expt 1) or natural (Expt 2) gravitational cues and to
reproduce its orientation with the same hand after diåerent memorization conditions,
with 5 s or 30 s un®lled delays, and with a 30 s delay ®lled with verbal or haptic
interpolated tasks.
First to be examined is the eåect of the un®lled delay length on orientation
processing in the two conditions of manual exploration varying the availability of
gravitational cues, and therefore varying the mode of orientation coding: with
minimal (Expt 1) or natural (Expt 2) gravitational cues. The ®ndings revealed that
the increase in the un®lled delay length (from 5 s to 30 s) did not in¯uence the
performance in the two conditions of manual exploration. On the one hand, these
results are consistent with some of the experimenters’ previous ®ndings (Gentaz &
Hatwell, 1996 ; Gentaz et al., 1999). In these studies, manual exploration was
performed in the frontal plane and with natural gravitational cues, and the un®lled
delays were 5 s (Gentaz & Hatwell, 1996) and 10 s (Gentaz et al., 1999). Results
showed that the performance was similar in these two un®lled delay conditions. On
the other hand, it is di¬cult to know whether the results of the present research are
consistent with Lechelt et al.’s (1976, 1980) ®ndings reported in the introductory test
above because these authors did not compare the successive reproduction after a 0 s
un®lled delay and the successive reproduction after 10 s un®lled delay. Taken
together, the present ®ndings revealed that the processing of orientations in the
diåerent conditions of manual exploration was stable over time whatever the length
of the un®lled delay.
Ê
Ê
386
Edouard Gentaz and Yvette Hatwell
In the present research, whatever the length of these un®lled delays (5 s or 30 s),
the haptic oblique eåect was absent when gravitational cues were minimal (Expt 1)
and was present when gravitational cues were natural (Expt 2). These ®ndings are
consistent with the assumption that the modes of orientation coding diåer according
to the kind of cues available (Gentaz & Hatwell, 1996, 1998) : ` movement coding’
occurred mainly when the spatial coding in the geocentric frame of reference was
unreliable (i.e. when gravitational cues were not available). These results are
consistent with Millar’s ` reference hypothesis’ (1994, 1997), which proposed that the
haptic perception of shape depends on the amount and type of spatial reference
information that is initially available to the participant from all sources. In this
perspective, the gravitational cues may play the role of spatial reference information.
The eåect of the introduction of interpolated tasks during the 30 s delay is now
discussed. The ®ndings showed that the verbal or haptic interpolated tasks (which
did not diåer) aåected diåerently the performance according to the conditions of
manual exploration. When the gravitational cues were minimal during manual
exploration in the horizontal plane (Expt 1), the performance was lower in the 30 s
haptic or verbal-®lled delay conditions than in the 5 s or 30 s un®lled delay
conditions. This decrease of performance in the two ®lled delays stemmed from the
increased errors in oblique settings, whereas the errors in vertical±horizontal settings
remained stable in the diåerent delay conditions. Consequently, the haptic oblique
eåect, which was absent in the un®lled delay conditions, was present in the two ®lled
delay conditions. The fact that the two interpolated tasks altered only the
performance in the oblique settings revealed that rehearsal processes were implied in
the processing of these orientations. This suggests that the mode of coding based on
movement requires high attentional resources to process e¬ciently the oblique
orientations in short-term memory. These results support the assumption that this
mode of coding would be linked to high level of haptic processing.
By contrast, when the gravitational cues were natural during manual exploration
in the horizontal plane (Expt 2), the performance did not diåer in the two ®lled delay
conditions and the two un®lled delay conditions. Consequently, the haptic oblique
eåect remained present in the diåerent memorization conditions. The fact that the
two interpolated tasks did not in¯uence the performance suggested that rehearsal
processes were minimally involved in the processing of orientations. This would
mean that the mode of coding based on geocentric frame of reference does not
require high attentional resources to process orientations in short-term memory.
Taken together, these ®ndings suggest that the mode of orientation coding based
on the geocentric frame of reference is stable over time but it is less e¬cient because
it is anisotropic: in it, the oblique eåect was present whatever the memorization
conditions. However, this mode of coding is economic because it does not depend
on attentional resources. By contrast, the mode of orientation coding based on
movement is also stable over time and it is e¬cient because it is not anisotropic:
indeed, in it, the haptic oblique eåect is absent whether there is no interference.
However, this mode of coding involves a high attentional cost. Further research
should study whether these ®ndings are also present in the other spatial planes.
In summary, the present study has shown that in blindfolded adults the
memorization conditions are involved in the haptic processing of orientations in as
Memorization in haptics
387
much as they determine, at least partially, the mode of orientation processing. Thus,
when the delay was un®lled, whatever its length, the haptic processing of
orientations depended on the conditions of manual exploration in the horizontal
plane : the oblique eåect was absent when the gravitational cues were minimal
(suggesting movement coding) and was present when these cues were natural
(suggesting geocentric coding). When the delay was ®lled with attentional
interpolated tasks, the haptic oblique eåect was present whatever the conditions of
manual exploration.
Acknowledgements
This work was supported by grants from University Pierre Mende’ s France of Grenoble and from the
Fyssen Foundation.
References
Abdi, H. (1987). Introduction au traitement statistique des donneU es expeU rimentales [Introduction to the
statistical analysis of experimental data]. Grenoble: Presses Universitaires de Grenoble.
Appelle, S. (1972). Perception and discrimination as a function of stimulus orientation : The ` oblique
eåect ’ in man and animals. Psychological Bulletin, 78, 266±278.
Appelle, S. & Countryman, M. (1986). Eliminating the haptic oblique eåect : In¯uence of scanning
incongruity and prior knowledge of the standards. Perception, 15, 365±369.
Diewert, G. L. (1975). Retention and coding in motor short-term memory : A comparison of storage
codes for distance and location information. Journal of Motor Behaviour, 7, 183±190.
Essock, E. A. (1980). The oblique eåect of stimulus identi®cation considered with respect to two classes
of oblique eåects. Perception, 9, 37±46.
Gentaz, E. (in press). Existe-t-il un `eået de l’oblique ’ dans la perception tactile des orientations?
[Is there an oblique eåect in the tactual perception of orientation ?]. L’AnneU e Psychologique.
Gentaz, E. & Ballaz, C. (in press). La perception visuelle des orientations et l’eået de l’oblique [The
visual perception of orientation and the oblique eåect]. L’AnneU e Psychologique.
Gentaz, E. & Hatwell, Y. (1995). The haptic ` oblique eåect ’ in children’s and adults’ perception of
orientation. Perception, 24, 631±646.
Gentaz, E. & Hatwell, Y. (1996). Role of gravitational cues in the haptic perception of orientation.
Perception and Psychophysics, 58, 1278±1292.
Gentaz, E. & Hatwell, Y. (1998). The haptic oblique eåect in the perception of rod orientation by blind
adults. Perception and Psychophysics, 60, 157±167.
Gentaz, E., Luyat, M., Cian, C., Hatwell, Y., Barraud, P.-A. & Raphel, C. (1999). The perception of
orientations in the visual, haptic, and somato-vestibular systems : Are there similar oblique eåects
across perceptual systems ? Manuscript submitted for publication.
Heeley, D. & Buchanan-Smith, H. (1990). Recognition of stimulus orientation. Vision Research, 30,
1429±1437.
Howard, I. P. (1982). Human visual orientation. New York : Wiley.
Jones, B. & Connolly, K. (1970). Memory eåects in cross-modal matching. British Journal of Psychology,
61, 267±270.
Keele, S. W. & Ellis, J. G. (1972). Memory characteristics of kinesthetic information. Journal of Motor
Behaviour, 4, 127±134.
Lechelt, E. C., Eliuk, J. & Tanne, G. (1976). Perceptual orientational asymmetries : A comparison of
visual and haptic space. Perception and Psychophysics, 20, 463±469.
Lechelt, E. C. & Verenka, A. (1980). Spatial anisotropy in intramodal and cross-modal judgements
of stimulus orientations: The stability of the oblique eåect. Perception, 9, 581±589.
Lilliefors, H. W. (1967). On the Kolmogorov±Smirnov test for normality with mean and variance
unknown. Journal of the American Statistical Association , 64, 387±389.
Millar, S. (1994). Understanding and representing space. Theory and evidence from studies with blind and sighted
children. Oxford : Clarendon Press.
388
Edouard Gentaz and Yvette Hatwell
Posner, M. I. (1967). Characteristics of visual and kinesthetic memory codes. Journal of Experimental
Psychology, 75, 103±107.
Saarinen, J. & Levi, D. (1995). Orientation anisotropy in vernier acuity. Vision Research, 35, 2449±2461.
Stelmach, G. & Wilson, M. (1970). Kinesthetic retention, movement extent, and information
processing. Journal of Experimental Psychology, 85, 425±430.
Received 20 October 1998; revised version received 12 February 1999