THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016
Vol. 69, No. 9, 1831–1841, http://dx.doi.org/10.1080/17470218.2015.1100642
The weight of expectation: Implicit, rather than explicit,
prior expectations drive the size–weight illusion
Gavin Buckingham and Aimee MacDonald
Department of Psychology, School of Life Sciences, Heriot-Watt University, Edinburgh, UK
(Received 17 August 2015; accepted 22 September 2015; first published online 11 December 2015)
In the size–weight illusion, small objects feel heavier than identically weighted larger objects. This illusion is thought to be a consequence of how one’s prior expectations can influence conscious perception
—lifters expect the large object to outweigh the small object and subsequently experience it as feeling
lighter than they expected it to be. Here, we directly examined how a familiar object’s identity can affect
how heavy someone expects it to be, and how these expectations will influence subsequent perceptions
of heaviness. We describe two novel weight illusions induced with familiar objects. In one condition,
participants judged the weight of a set of similar-size objects with very different natural weights (a polystyrene sphere, a tennis ball, and a cricket ball), which had all been adjusted to weigh the same amount
as one another. In this condition, participants experienced a small, but reliable, weight illusion, with the
lightest looking ball feeling heavier than the heaviest looking ball. In the other condition, participants
judged the weights of a different set of balls, which were different sizes, but similar natural weights, to
one another (a golf ball, a foam soccer ball, and an inflated beach ball). Again, participants experienced a
perceptual illusion, but in the opposite direction. Surprisingly, participant’s perceptions matched, rather
than contrasted with, their explicit expectations such that, even though they expected the golf ball to
outweigh the beach ball they perceived the golf ball as feeling heavier than the beach ball. The effect
of object mass appeared to dominate the effect of conscious expectations, suggesting that contrasting
expectations of heaviness are not necessary to experience weight illusions and that current models of
this robust perceptual effect must be revised.
Keywords: Perception; Weight illusions; Object identity; Lifting; Expectations.
It is well established that an individual’s perception
of heaviness does not perfectly reflect the mass of
the object they are lifting. One famous example
of this stark separation between conscious perception and sensory input can be experienced with
the “size–weight illusion” (SWI), where a small
object is typically reported to feel substantially
heavier than an identically weighted, but otherwise
similar-looking, large object (Charpentier, 1891;
Ellis & Lederman, 1993). Analogous misperceptions of weight can be experienced by manipulating
objects’ apparent material properties instead of their
volume: In the “material-weight Illusion” (MWI),
an object that appears to be made from a lowdensity material will be judged as heavier than an
identically weighted object that appears to be
made from a high-density material (Ellis &
Lederman, 1999; Seashore, 1899). These weight
Correspondence should be addressed to Gavin Buckingham Department of Psychology, School of Life Sciences, Heriot-Watt
University, Edinburgh EH14 4AS, UK. E-mail: g.buckingham@hw.ac.uk
The authors would like to thank I. Sperandio and several anonymous reviewers for their comments on an earlier draft of this manuscript, and P. Dimitriou for his help with preparation of the experimental stimuli.
© 2015 The Experimental Psychology Society
1831
BUCKINGHAM AND MACDONALD
illusions are robust, persistent, and cognitively
impenetrable (Buckingham, 2014).
Despite over 100 years of research, no explanation has been able to fully describe what causes
these illusory misperceptions of weight. Early theories proposed that the illusion had a sensorimotor
origin, stemming from mismatches between efferent and afferent information (Davis & Roberts,
1976). This early interpretation of an expectationdriven effect has been rejected in recent years,
with several studies demonstrating that the magnitude of the SWI and MWI is unrelated to gripping
and lifting forces over repeated trials (Buckingham,
Cant, & Goodale, 2009; Flanagan & Beltzner,
2000; Grandy & Westwood, 2006). In terms of
the MWI, scientists have tended to favour cognitive explanations—individuals experience a contrast
to their explicit expectations of how heavy each of
the stimuli will be, such that the heavy-looking
stimulus feels lighter than it was expected to be
and vice versa. Indeed, this reasonably uncontentious explanation receives strong support from
Ellis and Lederman’s (1998) golf ball illusion,
showing that only individuals with concrete expectations about weight differences between stimuli (a
real golf ball and an equally weighted practice golf
ball) experience a subsequent weight illusion.
Indeed, a range of cognitive factors have been
found to influence perceptions of heaviness under
a variety of contexts (Dijker, 2008; Schneider,
Parzuchowski, Wojciszke, Schwarz, & Koole,
2015; Schneider, Rutjens, Jostmann, & Lakens,
2011).
In the context of the SWI, however, it is less
clear what causes the illusory weight difference
between differently sized objects with the same
mass. Explanations of the cause of the SWI fall
into one of two categories: (a) the bottom-up,
direct perception of another variable, which lifters
interpret as heaviness, or (b) the top-down modulation of perceptual experience. Bottom-up explanations reject the idea of illusions as a failure of
the perceptual system, reclassifying the illusory
weight difference as the detection of some other
action-relevant variable that varies between the
stimuli. Candidate variables include the detection
of difference in an object’s density (Drewing &
1832
Tiest, 2013; Grandy & Westwood, 2006), inertia
tensor (Amazeen & Turvey, 1996), or some reflection of our perceptual expertise in detecting an
object’s “throwability”—the relationship between
volume and mass, which allows some items to be
thrown further than others—(Zhu & Bingham,
2011). By contrast, the top-down explanation is
conceptually similar to that put forward for the
MWI and golf ball illusion, with cognitive (rather
than sensorimotor) expectations interacting with
sensory information from the hands and arms to
drive the illusion that the objects have different
weights from one another. Thus, in the SWI,
lifters expect the large object to outweigh the
small object, leading to the opposite percept. It is
assumed that these expectations are built up
through our repeated prior experiences of the positive correlation between volume and mass in objects
that appear to be made from the same material. In
this context, the illusion can be taken to reflect how
our perceptual system alerts us to stimuli that
deviate from our prior experiences of how heavy
an object is likely to be (Baugh, Kao, Johansson,
& Flanagan, 2012).
A large body of evidence exists to support the
role of expectations in the illusory weight difference
experienced by lifters in typical SWI paradigms.
Flanagan and colleagues demonstrated that the
magnitude of the SWI can be altered by altering
lifters’ expectations through perceptual learning
(Flanagan, Bittner, & Johansson, 2008). They
gave participants thousands of trials worth of
experience at lifting objects with an inverted
density relationship (where the small objects actually weighed more than the large objects) and
noted that when lifters interacted with identically
weighted versions of these objects, they experienced
a reduced illusion, no illusion at all, or an inverted
illusion (depending on how much expectationaltering experience they had had with the inverted
stimuli). The influence of prior expectations on
weight perception has been compounded by
recent work where the SWI was induced in a
single object (Buckingham & Goodale, 2010). In
that study, participants lifted an unchanging
medium-sized cube without visual feedback after
a short preview period of another object, priming
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
WEIGHT ILLUSIONS WITHOUT EXPECTATIONS
them to expect to lift a larger or smaller cube than
they eventually interacted with. When participants
expected to lift the small cube, the medium-sized
cube they lifted felt substantially heavier than it
did when they expected to lift the large cube.
Neither of these studies can be easily explained in
terms of bottom-up, affordance-like effects,
suggesting that top-down effects must play a role
in inducing at least a portion of the illusory
weight difference experienced in the SWI (see
Buckingham, 2014, for further discussion on this
point).
The role of cognitive expectations in subsequent
weight perception is not well understood. For one,
it is not clear exactly which of the many factors that
can make one object heavier than another are incorporated with our sensory feedback to induce the
illusion. The simplistic nature of typical weight illusion experiments, where a single object property is
varied, often overlooks the potentially complex
interactions that may influence how heavy an individual may expect a particular object to be. Recent
work has provided evidence that lifters’ explicit
expectations do not drive perceptions of heaviness
when size and material are varied in the same set
of stimuli (Buckingham & Goodale, 2013). In
the study, participants lifted large and small cubes
that appeared to be made from (a) polystyrene
and (b) aluminium, all of which had been adjusted
to weigh the same amount as one another.
Participants unsurprisingly reported that they
expected a large difference in weight between the
large and small aluminium cubes, whereas they
expected a comparatively small difference in
weight between the large and small polystyrene
cubes. However, when lifting these stimuli and
judging their weights, they reported that the small
polystyrene cube outweighed the large polystyrene
cube by the same amount as for the aluminium
set of objects. In short, participants experienced
an identical-magnitude SWI, regardless of their
divergent initial expectations about how heavy
each object should be.
The current work was designed to directly
examine how a familiar object’s identity can affect
how heavy someone expects it to be, and how
these expectations will influence subsequent
perceptions of heaviness. It is well established that
our expectations associated with an object’s identity
will influence how we interact with it
(Hermsdörfer, Li, Randerath, Goldenberg, &
Eidenmüller, 2011). To investigate how expectations can influence how heavy something feels,
two sets of similarly weighted stimuli were
created. If explicit expectations are a critical factor
in weight perception, each of these sets of stimuli
would be expected to induce markedly different
patterns of weight illusions. The first set consisted
of three balls of approximately the same size,
which are normally different weights from one
another (a polystyrene ball, a tennis ball, and a
cricket ball). These similarly sized balls were
adjusted to weigh approximately the same amount
as one another. In this case, participants should
have robustly different expectations about how
heavy they expect each of the balls to be, which
should manifest as an illusory difference in weight
between the identically weighted balls (analogous
to the MWI). The second set consisted of three
sports balls that naturally had very similar masses,
but markedly different sizes (a golf ball, a toy
soccer ball, and an inflated beach ball). When
lifting these balls and judging their weights, participants should expect little or no difference between
how heavy each object will be, and thus experience
no subsequent weight illusion (i.e., judge them all
as having the same weight). If subjects did experience a weight illusion with this latter set of balls,
it would have to stem from factors other than participants’ prior expectations.
EXPERIMENTAL STUDY
Method
Participants
Seventy-two participants were recruited to take part
in a weight-perception study (14 males, 58 females;
mean age = 21.5 + 5.6 years). Sixty-three participants were right-handed, and nine were lefthanded, determined by self-report of writing
hand. Participants were recruited through the university student research participation system in
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
1833
BUCKINGHAM AND MACDONALD
Figure 1. The three balls judged by participants in the (A) same size, adjusted weight (SS-AW) and (B) different size, natural weight (DSNW) conditions. To view this figure in colour, please visit the online version of this Journal.
return for course credit. All participants reported
normal or corrected-to-normal vision and did not
report any muscular or cutaneous problems.
Participants gave written informed consent prior
to testing, with all procedures approved by the
Heriot-Watt University ethics committee.
Stimuli
Participants judged the weight of six different stimulus balls on an arbitrary 100-point scale in relation to
an 82-g reference object (a 5 × 5 × 5-cm wooden
cube, which they were told weighed 50 out of
100). In the “same size, adjusted weight” (SSAW) condition, participants judged the weight of
three balls with approximately the same size,
adjusted to have the same mass as one another—a
cricket ball, a tennis ball (increased in mass to
match the weight of the cricket ball), and a polystyrene ball (also increased in mass to match the
tennis and cricket ball; see Figure 1A). In the “different size, natural weight” (DS-NW) condition, participants judged the weight of three different-size,
same-mass balls—a golf ball, a cotton-filled toy
soccer ball, and an inflated beach ball (see
Figure 1B). The polystyrene and tennis balls were
adjusted by cutting open and filling with various
quantities of lead, keeping an approximately central
weighting. There were no visible indications that
the balls had been altered in any way. Care was
taken not to handle the balls in front of the participant before or during the experiment, so as to give
no indication that the balls were weighted differently
1834
from their standard form. Original and adjusted
properties of all stimuli are presented in Table 1
below.
Procedure
Participants sat in front of a table with their eyes
closed and their dominant hand outstretched with
their palm facing upwards. Participants were asked
to avoid resting their arm on the table surface and
to keep their eyes closed until an object had been
placed in their hand. Prior to judging the weights
of each set of experimental stimuli, the experimenter
placed the reference weight in the participant’s hand
and informed them that this object represented a
weight of 50 on a 100-point scale. After removing
the reference weight, participants then made a
verbal judgement about how heavy each of the
balls looked, prior to touching them (“expected heaviness”). Then, after being allowed to experience the
reference weight for a second time, the first ball was
placed on the hand of the participant, at which point
they opened their eyes and gave a rating of how
heavy the ball felt on the 100-point scale (“perceived
heaviness”). Participants first judged the weight of
each of the three balls three separate times in randomized triplets in one of six pseudorandomized
orders. After a short break, participants were then
once again given the reference weight before
judging the weight of the three balls three times in
the other condition. Presentation order of the two
conditions was counterbalanced across subjects. In
total, 18 perceived heaviness ratings were given by
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
WEIGHT ILLUSIONS WITHOUT EXPECTATIONS
Table 1. The physical properties of the stimuli
Original ball
Condition
Same size, adjusted weight (SS-AW)
Different size, natural weight (DS-NW)
Experimental ball
Ball type
Weight
(g)
Diameter
(cm)
Density
(g cm–3)
Weight
(g)
Diameter
(cm)
Density
(g cm–3)
Polystyrene
Tennis
Cricket
Golf
Soccer
Beach
3
55
159
45
52
52
6.7
6.7
7.0
4.3
8.6
27.5
0.02
0.35
0.89
1.08
0.16
0.005
151a
157a
159
45
52
52
6.7
6.7
7.0
4.3
8.6
27.5
0.97a
1.01a
0.89
1.08
0.16
0.005
Note: The diameter (cm), weight (g), and density (g cm–3) of each of the balls, in their original form, and for the stimulus balls used in
the experiment.
a
Adjusted from original form.
each participant, and the experiment took approximately 10 min to complete. Pre-lift-off expected
heaviness ratings and average perceived heaviness
ratings for the balls in each condition were examined
in separate one-factor repeated measures analyses of
variance (ANOVAs), and significant effects were
examined with Bonferroni-corrected post hoc
t tests. Where necessary, Greenhouse–Geisser tests
were used to correct for inhomogeneity of variance
in the F tests.
Results
Same size, adjusted weight (SS-AW) condition
The omnibus analysis of the pre-lift-off heaviness
ratings indicated that participants expected the
three balls in this condition to have substantially
different weights from one another (polystyrene:
15.5 + 13.3, tennis: 55.5 + 15.1, cricket: 79.7
+ 15.0), F(1.8, 130.6) = 528.7, p , .001, η2 =
.88 (Figure 2A). The cricket ball was predicted to
outweigh both the tennis, t(71) = 20.1, p , .001,
and the polystyrene ball, t(71) = 28.7, p , .001,
and the tennis ball was predicted to outweigh the
polystyrene ball, t(71) = 14.0, p , .001. Thus, participants’ expectations were in line with the balls’
real-world properties (Table 1).
The omnibus analysis of the post-lift-off heaviness judgements indicated that that participants
experienced the balls as being different weights
from one another (polystyrene: 71.8 + 11.1,
tennis: 75.0 + 11.1, cricket: 66.6 + 12.2), F
(1.7, 121.6) = 23.5, p , .001, η2 = .25;
Figure 2B). However, these ratings of heaviness
did not reflect participants’ expectations of heaviness. Post hoc t tests indicated that the tennis ball
was judged as heavier than both the cricket ball, t
(71) = 7.5, p , .001, and the polystyrene ball, t
(71) = 2.9, p , .05, while the polystyrene ball was
judged as heavier than the cricket ball, t(71) =
3.6, p , .005. No correlation was observed
between the expected heaviness and the average
perceived heaviness for the polystyrene ball
(r = .13, n = 72, p = .28), the tennis ball (r = .10,
n = 72, p = .39), or the cricket ball (r = .04,
n = 72, p = .77).
Different size, natural weight (DS-NW) condition
The omnibus analysis of the pre-lift-off heaviness
ratings indicated that participants did not expect
the three balls in this condition to weigh the
same amount as one another (golf: 59.0 + 22.5,
soccer: 38.3 + 18.9, beach: 16.0 + 14.0), F(1.6,
115.0) = 160.3, p , .001, η2 = .69 (Figure 3A).
In contrast to the real-world properties of the
stimuli, participants expected the golf ball to be
heavier than the foam soccer ball, t(71) = 8.8, p ,
.001, and the beach ball, t(71) = 14.9, p , .001,
and the foam soccer ball was predicted to outweigh
the beach ball, t(71) = 12.0, p , .001. Thus, even
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
1835
BUCKINGHAM AND MACDONALD
Figure 2. Participants’ expected heaviness (A) and perceived heaviness (B) of the similarly sized balls that had been adjusted to weigh the same
amount in the same size, adjusted weight (SS-AW) condition. Error bars show standard error of the mean.
though the three balls in this condition naturally
weighed similar amounts to one another (Table
1), participants expected them to have markedly
different weights.
The omnibus analysis of the perceptual ratings
of heaviness after lift-off also yielded a significant
main effect of ball type, indicating that the balls
were judged as having different weights from one
another (golf: 50.1 + 18.0, soccer: 31.7 + 14.9,
beach: 31.7 + 14.9), F(1.8, 126.4) = 235.8, p ,
.001, η2 = .77 (Figure 3B). Bonferroni-corrected
1836
post hoc t tests indicated that the golf ball was
judged as heavier than both the foam soccer, t(71)
= 12.4, p , .001, and the beach ball, t(71) =
18.7, p , .001, and that the foam soccer ball was
judged as heavier than the beach ball, t(71) =
11.6, p , .001. Furthermore, significant correlations between expected heaviness and the
average perceived heaviness were observed for the
golf (r = .31, n = 72, p , .01), soccer (r = .34,
n = 72, p , .005), and beach (r = .26, n = 72,
p , .05) balls. Thus, participants’ perceptions of
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
WEIGHT ILLUSIONS WITHOUT EXPECTATIONS
Figure 3. Participants’ expected heaviness (A) and perceived heaviness (B) of the differently sized balls in the different size, natural weight
(DS-NW) condition. Error bars show standard error of the mean.
heaviness were in line with, rather than a contrast
to, their expectations of heaviness.
Comparison of weight illusion in SS-AW and
DS-NW conditions
Participants experienced illusory weight differences
between the stimuli in both experimental conditions. To quantify the magnitude of these
weight illusions, we subtracted the heaviness
rating for the ball with the lowest average heaviness
rating from that for the ball with the highest
average heaviness rating separately for each participant in the SS-AW and DS-NW conditions
(Figure 4). These individual illusions were then
averaged and compared with a paired-sample
t test. In contrast to our initial hypotheses, participants experienced a significantly larger illusion in
the DS-NW condition (where the heaviest
feeling object felt 35.7% heavier) than in the SSAW condition (where the heaviest feeling object
felt only 13.6% heavier), t(71) = 9.98, p , .001,
Cohen’s d = 1.17. The overall magnitude of these
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
1837
BUCKINGHAM AND MACDONALD
Figure 4. The magnitude of the illusory weight difference (highest rating – lowest rating) in each condition. SS-AW = same size, adjusted
weight; DS-NW = different size, natural weight. Error bars show standard error of the mean.
weight illusions did not correlate with one another
(p = .73).
Finally, in order to rule out the influence of any
order effects (i.e., whether lifting adjusted stimuli
influenced participants’ perception of the unaltered
objects), we compared the magnitude of the illusion
in the first set of balls lifted between the groups
with an independent-samples t test. The results
of this between-group comparison mirrored the
within-group comparison outlined above—individuals experienced a significantly larger illusion when
lifting the DS-NW balls (an illusory weight difference of 42%) than they did when lifting the SSAW balls (an illusory weight difference of 14.5%,
t(70) = 8.6, p , .001).
Discussion
The current work examined how familiarityinduced expectations of heaviness influenced subsequent perceptions of heaviness with a novel
weight illusion. In typical weight illusion studies,
participants’ perceptions of an object’s weight will
contrast with their expectations of its heaviness.
Thus, in the SWI, small objects feel heavier than
identically weighted large objects because of the
well-founded prior expectations that small things
1838
will be less heavy than larger things of the same
type. Here, we directly examined the role of expectations in weight perception by creating sets of familiar stimuli aimed at inducing strongly or weakly
divergent expectations of heaviness. Two sets of
balls were presented to participants; one was
created to elicit strongly divergent expectations of
heaviness across the set (the SS-AW condition),
whereas the other was created to elicit similar
expectations of heaviness across the set (the DSNW condition). The balls in the SS-AW set
were roughly the same volume, whereas the balls
in the DS-NW set were obviously different
volumes. The balls in the SS-AW set were adjusted
to have similar weight as one another, whereas the
balls in the DS-NW set naturally weighed approximately the same amount as one another. We predicted that, if expectations of heaviness drive
subsequent perceptions of heaviness, the balls in
the SS-AW condition should feel very different
in weight from one another (i.e., lifters would
experience a robust weight illusion), whereas the
balls in the DS-NW condition should feel similar
in weight to one another (i.e., lifters would experience little or no weight illusion).
In contrast to our predictions, lifters experienced
only a modest weight illusion when judging the
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
WEIGHT ILLUSIONS WITHOUT EXPECTATIONS
weight of the balls in the SS-AW condition,
despite robustly different expectations of heaviness
for each object (Figure 2). At first glance, the illusion in this condition appears to be analogous to the
MWI, where objects that appear to be made from
low-density material feel heavier than objects that
appear to be made from high-density materials.
However, lifters’ perceptions did not obviously contrast with their expectations in the way they do with
typical MWI paradigms (Harshfield & DeHardt,
1970). Here, the tennis ball was judged to be
heavier than both the cricket and polystyrene
balls, in spite of participants’ expectations that it
would weigh substantially less than the cricket
ball and substantially more than the polystyrene
ball. This finding is unlikely to be due to actual
differences in object mass between the tennis and
polystyrene balls, which are well below the 10%
difference required to detect differences in weight
when holding objects (Ross & Di Lollo, 1970). It
is, however, possible that this unexpected deviation
from the MWI is due to slight size differences
between the stimuli (Table 1), as well as friction/
colour cues (Flanagan & Bandomir, 2000;
Walker, Francis, & Walker, 2010), overriding any
effect of prior expectations. Another explanation
for this finding could relate to differences in participants’ experience with each one of the stimuli influencing their perceptions in an unexpected way, as it
is likely that our sample had more experience with
tennis balls than they did with either polystyrene or
cricket balls. The degree to which the precision of
an individual’s prior expectation for a particular
object influences subsequent perceptions of heaviness is an interesting topic of study for future
research.
This influence of object size over and above that
of expected heaviness on perceptions of heaviness
was clearly evident in the DS-NW condition,
where participants erroneously expected the
(small) golf ball to be heavier than the (large)
beach ball (Figure 3A). This erroneous expectation
created a unique set of circumstances—with lifters
expecting the smallest object of the set to weigh
more than the largest. Despite this initial expectation, and in stark opposition to usual contrastive
nature of all weight illusions described to date,
participants perceived the golf ball as weighing significantly more than the soccer ball, which they in
turn judged as weighing significantly more than
the inflated beach ball. The illusion that participants experienced when hefting these objects was
significantly larger than the one elicited through
lifting the stimuli in the SS-AW condition
(Figure 4). Thus, the illusion-inducing effects of
object volume far outstrip those associated with
familiar object identity, leading participants to
experience a normal-looking SWI in the absence
of contrasting expectations of heaviness.
These findings are particularly important in
the context of understanding what factors contribute to the perceptions of heaviness underlying
the SWI. Previous cognitive models of the SWI
have suggested that the illusory weight difference
stems from a contrast to lifters’ initial expectations, reflecting a unique way in which prior
expectations are incorporated with sensory input
to drive conscious perception (Buckingham &
Goodale, 2010; Ernst, 2009; Flanagan &
Beltzner, 2000). However, little is known about
the nature of the expectations that influence our
perceptions in such a way. The current work
allows us to rule out the proposal that explicit
expectations of heaviness drive subsequent perceptions of heaviness, causing the illusory misperceptions of weight seen in the SWI. Instead the
findings from both our conditions indicate that
implicit expectations related to object size is a
more dominant factor than prior explicit expectations. These findings effectively refute the suggestion that the illusion-causing priors in the
SWI stem from size–weight associations for individual families of objects (Flanagan et al., 2008),
as participants in our study clearly understood
that all the objects in our study were made
from different materials.
On the face of it, as the strongest illusion was
seen in conditions where size and density were
varied between the stimuli, these findings might
well be taken as support of the notion that the
SWI reflects a perceptual skill for selecting the
object that one can throw the furthest—the ecologically relevant property of “throwability” (Zhu &
Bingham, 2011). There is, however, strong
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
1839
BUCKINGHAM AND MACDONALD
evidence to suggest that the SWI is caused, at least
in part, by prior expectations (Buckingham &
Goodale, 2010; Flanagan et al., 2008). The findings from the current work suggest that the SWI
does not reflect a contrast to the explicit prior
knowledge that participants can articulate when
asked how heavy they expect something to be.
Instead, we propose that the illusory weight difference experienced when lifting SWI-inducing
objects reflects a more implicit form of prior knowledge and contrasts with each object’s deviation
from the average size–weight relationship of all
the objects that one typically experiences. As, presumably, a robust positive correlation exists
between volume and mass across hand-held
objects in general, individuals should experience
the same magnitude of SWI regardless of what
type of object they are lifting—a proposal that is
consistent
with
our
previous
findings
(Buckingham & Goodale, 2013). The broad generalizability of this function is, to our knowledge,
unique in perception (Fahle, 2005). This simple
mechanism may exist to facilitate rapid categorical
judgements about an object’s properties once it
has been lifted. This hypothesis can be easily falsified by studies demonstrating a situation where,
without training, small objects feel lighter than
equally weighted large objects; to date no such situation has been reported.
The current work is the first evidence that a
weight illusion can be induced by varying object
identity, rather than by manipulating object size
or material properties. Objects of the same size
with different prior expectations of heaviness elicited a relatively modest illusion, in line with the
MWI. By contrast, the stimuli that varied in size
across object category induced a far larger perceptual effect, in line with the SWI. The difference
in the magnitudes of these novel illusions suggests
that size cues are a more dominant factor for the
perception of heaviness than learned cues to mass
(see also Buckingham & Goodale, 2013). The
current work also adds to the growing evidence
that size is a special property that may processed
separately from, and given a higher value than,
other cues to object weight such as object identity
or apparent material.
1840
REFERENCES
Amazeen, E. L., & Turvey, M. (1996). Weight perception and the haptic size–weight illusion are functions
of the inertia tensor. Journal of Experimental
Psychology: Human Perception and Performance, 22
(1), 213–232. http://doi.org/10.1037/0096-1523.22.
1.213
Baugh, L. A., Kao, M., Johansson, R. S., & Flanagan, J.
R. (2012). Material evidence- interaction of welllearned priors and sensorimotor memory when
lifting objects. Journal of Neurophysiology, 108,
1262–1269. http://doi.org/10.1152/jn.00263.2012
Buckingham, G. (2014). Getting a grip on heaviness perception: A review of weight illusions and their probable
causes. Experimental Brain Research, 232(6), 1623–
1629. http://doi.org/10.1007/s00221-014-3926-9
Buckingham, G., Cant, J. S., & Goodale, M. A. (2009).
Living in a material world: How visual cues to
material properties affect the way that we lift objects
and perceive their weight. Journal of Neurophysiology,
102(6),
3111–3118.
http://doi.org/10.1152/jn.
00515.2009
Buckingham, G., & Goodale, M. A. (2010). Lifting
without seeing: The role of vision in perceiving and
acting upon the size weight illusion. PLoS ONE, 5(3),
e9709. http://doi.org/10.1371/journal.pone.0009709
Buckingham, G., & Goodale, M. A. (2013). Size
matters: A single representation underlies our perceptions of heaviness in the size-weight illusion. PLoS
ONE, 8(1), e54709. http://doi.org/10.1371/journal.
pone.0054709
Charpentier, A. (1891). Analyse expérimentale quelques
éléments de la sensation de poids. Archives de
Physiologie Normales et Pathologiques, 3, 122–135.
Davis, C., & Roberts, W. (1976). Lifting movements in
the size-weight illusion. Perception & Psychophysics,
20, 33–36.
Dijker, A. J. M. (2008). Why Barbie feels heavier than
Ken: The influence of size-based expectancies and
social cues on the illusory perception of weight.
Cognition, 106(3), 1109–1125.
Drewing, K., & Tiest, W. M. B. (2013). Mass and density
estimates contribute to perceived heaviness with weights
that depend on the densities’ reliability. In World
Haptics Conference (WHC), 2013 (pp. 593–598).
IEEE. Retrieved from http://ieeexplore.ieee.org/xpls/
abs_all.jsp?arnumber=6548475
Ellis, R. R., & Lederman, S. J. (1993). The role of haptic
versus visual volume cues in the size-weight illusion.
Perception & Psychophysics, 53(3), 315–324.
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
WEIGHT ILLUSIONS WITHOUT EXPECTATIONS
Ellis, R. R., & Lederman, S. J. (1998). The golf-ball illusion: Evidence for top-down processing in weight
perception. Perception, 27(2), 193–201.
Ellis, R. R., & Lederman, S. J. (1999). The materialweight illusion revisited. Perception & Psychophysics,
61(8), 1564–1576.
Ernst, M. O. (2009). Perceptual learning: Inverting the
size-weight illusion. Current Biology, 19(1), R23–25.
http://doi.org/10.1016/j.cub.2008.10.039
Fahle, M. (2005). Perceptual learning: Specificity versus
generalization. Current Opinion in Neurobiology, 15(2),
154–160. http://doi.org/10.1016/j.conb.2005.03.010
Flanagan, J. R., & Bandomir, C. A. (2000). Coming to
grips with weight perception: Effects of grasp configuration on perceived heaviness. Perception &
Psychophysics, 62(6), 1204–1219.
Flanagan, J. R., & Beltzner, M. A. (2000). Independence
of perceptual and sensorimotor predictions in the
size-weight illusion. Nature Neuroscience, 3(7), 737–
741. http://doi.org/10.1038/76701
Flanagan, J. R., Bittner, J. P., & Johansson, R. S. (2008).
Experience can change distinct size-weight priors
engaged in lifting objects and judging their weights.
Current Biology, 18(22), 1742–1747. http://doi.org/
10.1016/j.cub.2008.09.042
Grandy, M. S., & Westwood, D. A. (2006). Opposite
perceptual and sensorimotor responses to a sizeweight illusion. Journal of Neurophysiology, 95(6),
3887–3892. http://doi.org/10.1152/jn.00851.2005
Harshfield, S., & DeHardt, D. (1970). Weight judgment
as a function of apparent density of objects.
Psychonomic Science, 20, 365–366.
Hermsdörfer, J., Li, Y., Randerath, J., Goldenberg, G.,
& Eidenmüller, S. (2011). Anticipatory scaling of
grip forces when lifting objects of everyday life.
Experimental Brain Research, 212(1), 19–31. http://
doi.org/10.1007/s00221-011-2695-y
Ross, J., & Di Lollo, V. (1970). Differences in heaviness
in relation to density and weight. Perception &
Psychophysics, 7(3), 161–162. http://doi.org/10.3758/
BF03208648
Schneider, I. K., Parzuchowski, M., Wojciszke, B.,
Schwarz, N., & Koole, S. L. (2015). Weighty data:
Importance information influences estimated weight
of digital information storage devices. Frontiers in
Psychology, 5, Article No. 1536. http://doi.org/10.
3389/fpsyg.2014.01536
Schneider, I. K., Rutjens, B. T., Jostmann, N. B., &
Lakens, D. (2011). Weighty matters: Importance literally feels heavy. Social Psychological and Personality
Science, 2(5), 474–478. http://doi.org/10.1177/
1948550610397895
Seashore, C. E. (1899). Some psychological statistics II.
The material weight illusion. University of Iowa
Studies in Psychology, 2, 36–46.
Walker, P., Francis, B. J., & Walker, L. (2010). The
brightness-weight illusion. Experimental Psychology,
57(6), 462–469. http://doi.org/10.1027/1618-3169/
a000057
Zhu, Q., & Bingham, G. P. (2011). Human readiness to
throw: the size–weight illusion is not an illusion when
picking the best objects to throw. Evolution and
Human Behavior, 32(4), 288–293. http://doi.org/10.
1016/j.evolhumbehav.2010.11.005
THE QUARTERLY JOURNAL OF EXPERIMENTAL PSYCHOLOGY, 2016, 69 (9)
1841