Available online at www.sciencedirect.com
Cognition 106 (2008) 433–443
www.elsevier.com/locate/COGNIT
Brief article
Evidence for the activation of sensorimotor
information during visual word recognition:
The body–object interaction eVect
Paul D. Siakaluk a,¤, Penny M. Pexman b, Laura Aguilera a,
William J. Owen a, Christopher R. Sears b
a
Department of Psychology, University of Northern British Columbia, Prince George, BC, Canada V2N 4Z9
b
Department of Psychology, University of Calgary, Calgary, AB, Canada T2N 1N4
Received 5 July 2006; revised 15 November 2006; accepted 10 December 2006
Abstract
We examined the eVects of sensorimotor experience in two visual word recognition tasks.
Body–object interaction (BOI) ratings were collected for a large set of words. These ratings
assess perceptions of the ease with which a human body can physically interact with a word’s
referent. A set of high BOI words (e.g., mask) and a set of low BOI words (e.g., ship) were created, matched on imageability and concreteness. Facilitatory BOI eVects were observed in lexical decision and phonological lexical decision tasks: responses were faster for high BOI words
than for low BOI words. We discuss how our Wndings may be accounted for by (a) semantic
feedback within the visual word recognition system, and (b) an embodied view of cognition
(e.g., Barsalou’s perceptual symbol systems theory), which proposes that semantic knowledge
is grounded in sensorimotor interactions with the environment.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Embodied cognition; Lexical processing; Visual word recognition
*
Corresponding author. Tel.: +1 250 960 6120; fax: +1 250 960 5744.
E-mail address: [email protected] (P.D. Siakaluk).
0010-0277/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.cognition.2006.12.011
434
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
1. Introduction
What do humans know and how is this knowledge acquired? How is this knowledge accessed when recognizing an object, understanding a concept, or comprehending a story? Traditionally, cognitive scientists sought to answer these questions
principally by analyzing internal cognitive processes, such as the manner in which
symbolic representations are manipulated via rules (e.g., Cowart, 2004). Although
this approach is indisputably important, much of the research it has generated has
either ignored or down-played the importance of our sensorimotor interactions with
the environment. A growing Weld within cognitive science, known as embodied cognition, examines how sensorimotor interactions with the environment are integral in
the acquisition of knowledge and to the development of cognitive processes that bear
on that knowledge (Barsalou, 1999; Clark, 1997; Cowart, 2004; LakoV & Johnson,
1999; Pecher & Zwaan, 2005; Wilson, 2002).
According to Wilson (2002), the embodied cognition framework incorporates six
distinct claims regarding the manner in which the mind, the body, and the external
world interrelate to help organisms successfully cope with environmental demands.
Two of these claims are directly relevant to the research presented in this paper. The
Wrst is that “(c)ognition is for action. The function of the mind is to guide action, and
cognitive mechanisms such as perception and memory must be understood in terms
of their ultimate contribution to situation-appropriate behavior” (p. 626). The second claim is that “oV-line cognition is body based. Even when decoupled from the
environment, the activity of the mind is grounded in mechanisms that evolved for
interaction with the environment – that is, mechanisms of sensory processing and
motor control” (p. 626).
Our research examined whether sensorimotor information gained through body–
object interaction is relevant in a task as abstract and removed from motor activity as
visual word identiWcation. If such sensorimotor information can be shown to inXuence visual word recognition, then this result would provide support for the claims of
the embodied cognition perspective outlined above. As Pecher and Zwaan (2005)
argued, “it is crucial for the embodied framework to demonstrate that cognition is
grounded in bodily interactions with the environmentƒ thus, it needs to be shown
that sensorimotor patterns are activated when concepts are accessed.” (p. 3). Our
research was motivated by this theoretical issue.
In her discussion of the claim that oV-line cognition is body based, Wilson (2002)
suggested that many on-line cognitive processes that originally evolved to help an
organism directly deal with its environment may have been co-opted for diVerent,
less direct functions. These cognitive processes may also be used oV-line, to run mental simulations of possible ways of interacting with the environment, avoiding the
need to actually carry out each possible action. Recent studies have investigated these
cognitive processes using one of two general approaches. First, researchers have used
methodologies that employ relevant sensorimotor activities, just prior to or at the
time of testing, that are proposed to be integral to the required responses. For example, experiments have shown that speciWc arm movements (e.g., pulling or pushing
movements) inXuence responses when items have a particular valence (i.e., positive or
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
435
negative) (Markman & Brendl, 2005), or are congruent with a particular implied
action during sentence comprehension (Glenberg & Kaschak, 2002). This type of
approach has also been applied in virtual reality environments. In one such study,
Christou and BülthoV (1999) reported superior recognition of scenes when participants had been directed to actively (vs. passively) explore those scenes in a virtual
reality environment. Similarly James et al. (2002) reported faster recognition of
objects following active rotation (vs. passive viewing of object rotations). As these
examples demonstrate, sensorimotor activities, such as arm movements or active
exploration of virtual environments, facilitate cognitive processes underlying
responding.
The second approach is to examine sensorimotor eVects using methodologies that
do not employ relevant sensorimotor activities prior to or at the time of testing. For
example, LakoV and Johnson (1999) provided detailed accounts of the ways in which
abstract metaphors are understood through knowledge gained via sensorimotor
interactions with the world. As an example, consider one particularly vivid metaphor,
Bad Is Stinky, as in “This movie stinks” (p. 50; emphasis in the original). This metaphor is comprehensible because of prior sensorimotor experience with foul-smelling
objects (e.g., rotting food) and the desire to avoid them. A second example of this
approach (although seldom discussed as embodied cognition eVects) is the research
examining the eVects of imageability and concreteness in visual word recognition
tasks (e.g., Cortese, Simpson, & Woolsey, 1997; James, 1975; Kroll & Mervis, 1986;
Strain, Patterson, & Seidenberg, 1995). (Because imageability and concreteness are
conceptually similar and highly correlated, we will consider them jointly). A great
deal of research has shown that words with referents that can be more easily imaged/
sensed have richer mental representations that enable them to be recognized more
rapidly in word recognition tasks. Imageability/concreteness eVects have also been
reported in word association tasks (de Groot, 1989), and deWnition tasks (Sadoski,
Kealy, Goetz, & Paivio, 1997).
Although it is diYcult to separate precisely the eVects of sensory experience and
motor experience on cognitive processing, we argue that imageability/concreteness
ratings primarily gauge sensory experience. If this is the case, then an important and
interesting question is whether variables that capture primarily motor experience also
inXuence conceptual processing. A few recent studies have demonstrated that motor
information, as gauged by variables measuring object manipulability (Magnie, Besson, Poncet, & Dolisi, 2003; Myung, Blumstein, & Sedivy, 2006; Wolk, Coslett, &
Glosser, 2005), inXuences object recognition. To our knowledge, only one study has
examined the eVects of motor information on word recognition. Myung et al. (2006)
examined the eVects of motor information on priming using an auditory lexical decision task. Related primes and targets shared manipulation features. Manipulation
features were deWned as “general actions on an object that involve body movements
and typically are associated with its intended usage” (p. 228). For example, keys and
screwdrivers share manipulation features at a general action pattern level because
they require similar wrist movements. Myung et al. reported a signiWcant priming
eVect: target words preceded by a word that shared common manipulation features
(e.g., key-screwdriver) were responded to faster than target words preceded by a word
436
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
that did not share common manipulation features (e.g., bar-screwdriver). These
results suggest that sensorimotor information was activated in a priming paradigm
and facilitated auditory lexical decisions.
1.1. The present research
The present research builds on the Myung et al. (2006) Wndings. We examined
the eVects of sensorimotor knowledge on recognition of individual words (we did
not use a priming paradigm). That is, we examined whether sensorimotor knowledge inXuences word recognition, even when that knowledge is not preactivated
by related primes.
To create the stimuli for our study, we asked a group of undergraduate
students to rate words as to the ease or diYculty with which a human body can
physically interact with each word’s referent. These ratings allowed us to extract a
variable we call body–object interaction (BOI). We based this rating scheme on
Cortese and Fugett’s (2004) rating scheme for collecting imageability ratings.
They had participants rate words “as to the ease or diYculty with which they
arouse mental images” (p. 387). After collecting the ratings we then selected items
such that the set of high BOI words were matched to the set of low BOI words for
both imageability and concreteness (as well as several other lexical and semantic
variables). Thus, any eVects of BOI should be attributable to sensorimotor
information that has been gained through prior bodily interactions with the
objects.
We assumed that eVects of sensorimotor experience (BOI eVects) could be
explained in terms of feedback activation from semantics to orthography and to
phonology. The notion of feedback activation in the visual word recognition system has recently been invoked to explain a number of eVects (Hino & Lupker,
1996; Hino, Lupker, & Pexman, 2002; Pecher, 2001; Pexman & Lupker, 1999;
Stone, Vanhoy, & Van Orden, 1997; Strain et al., 1995). The feedback activation
account has the following important assumptions. First, the system is fully interconnected, with feedforward and feedback connections between sets of orthographic, phonological, and semantic units (e.g., Coltheart, Rastle, Perry, Langdon,
& Ziegler, 2001; Plaut, McClelland, Seidenberg, & Patterson, 1996). Second, the
inXuence of semantic feedback is determined by the type of feedback connections
between the semantic units and either the orthographic units or the phonological
units. That is, to the extent that the mappings between semantic and other sets of
units are many-to-one, there will be a facilitatory inXuence of semantic feedback
(relative to the situation where the same connections are one-to-one). Third, lexical decision task (LDT; is it a word?) responses are based primarily on the activation of the orthographic units and phonological lexical decision task (PLDT; does
it sound like a word?) (and naming) responses are based primarily on the activation of the phonological units.
Within this framework, we predicted that eVects of BOI would be facilitatory. The
rationale for this prediction comes from a number of previous research Wndings, all
of which suggest that semantic richness (e.g., high imageability/concreteness, high
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
437
number of semantic features, high number of word senses) tends to facilitate word
identiWcation (James, 1975; Pexman, Lupker, & Hino, 2002; Rodd, Gaskell, &
Marslen-Wilson, 2002). To the extent that high BOI words also have richer semantic
representations, they should generate stronger feedback activation from semantics to
both orthography and to phonology, and thus they should be identiWed more rapidly
than low BOI words.
We examined the eVects of BOI using a LDT, to assess the inXuence of BOI on
feedback activation from semantics to orthography, and using a PLDT, to assess the
inXuence of BOI on feedback activation from semantics to phonology. Typically,
semantic eVects are larger in the LDT if the task is made more diYcult by including
pseudohomophones (e.g., brane) as foils (e.g., James, 1975; Pexman & Lupker, 1999).
We therefore included pseudohomophone foils in the LDT to make the task maximally sensitive to BOI eVects. In this context, the prediction we derived from the
embodied cognition perspective was that high BOI words would be identiWed faster
than low BOI words.
2. Method
2.1. Participants
Sixty undergraduate students from the University of Northern British Columbia
participated in the experiments for bonus course credit: 30 participants each in the
LDT and in the PLDT. All participants were native English speakers and reported
normal or corrected-to-normal vision.
A separate group of 25 participants rated the stimuli for body–object interaction, and another group of 25 participants rated the stimuli for number of features.
These participants were drawn from the same population as the individuals who
participated in the LDT and PLDT experiments, and also received bonus course
credit.
2.2. Stimuli
An initial pool of 234 words was Wrst created. Each word had only one entry in the
ITP Nelson Canadian Dictionary (1997) and all had noun deWnitions listed Wrst. The
words were randomly ordered in two questionnaires. For the body–object interaction
questionnaire, participants were instructed to rate each word according to the ease or
diYculty with which a human body can physically interact with the word’s referent.
For the number of features questionnaire, participants were instructed to rate each
word according to the number of features possessed by the word’s referent (the
instructions for this questionnaire were those used by Toglia & Battig (1978)). For
the body–object interaction questionnaire a scale from 1 to 7 was placed to the right
of each word, with 1 indicating low body–object interaction and 7 indicating high
body–object interaction. For the number of features questionnaire the same scale
was placed next to each word, with 1 indicating low number of features and 7
438
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
Table 1
Mean characteristics for word stimuli
Word type
BOI
Freq
SubF
N
PFI
NumF
NumS
NumA
SemD
Image
Conc
High BOI
Low BOI
5.3
3.3
17.3
17.3
3.5
3.6
7.1
6.4
3.0
3.0
3.4
3.7
5.7
4.8
14.3
13.3
307.7
307.8
6.3
6.2
5.9
5.8
Note. BOI, body–object interaction; Freq, print frequency; SubF, subjective familiarity; N, number of
orthographic neighbours; PFI, phonological feedback inconsistency; NumF, number of features; NumS,
number of senses; NumA, number of associates; SemD, semantic distance; Image, imageability; Conc,
concreteness.
indicating high number of features. Participants were encouraged to use the entire
scale in making their ratings, and were instructed that all the words were nouns and
to base their ratings on this fact.
Based on these ratings, 48 words were selected for use in the experimental tasks
(hereafter the experimental items). Twenty-four of the words were rated as being high
in body–object interaction (e.g., mask) and the other 24 words were rated as being
low in body–object interaction (e.g., ship). These two word groups were matched for
length, printed frequency, subjective familiarity, number of orthographic neighbours,
phonological feedback inconsistency, number of features, number of senses, number
of associates, semantic distance, and, importantly, both imageability and concreteness (all ps > .15). The descriptive statistics for the experimental words are presented
in Table 1.
Forty-eight pseudohomophones with extant word bodies (e.g., brane), matched to
the experimental items on length, were used in both the LDT and the PLDT. In addition, 96 nonwords were used in the PLDT. The stimuli are listed in Appendix A.
2.3. Apparatus and procedure
For both the LDT and the PLDT the stimuli were presented on a colour VGA
monitor driven by a Pentium-class microcomputer running DirectRT software
(2006). A trial was initiated by a Wxation marker that appeared at the center of the
computer display. The Wxation marker was presented for 1000 ms and was then
replaced by a letter string. For the LDT, participants were instructed to decide
whether each item was a real English word or not. For the PLDT, participants were
instructed to decide whether each item sounded like a real English word or not. For
both tasks, ‘yes’ responses were made by pressing the “?” key and ‘no’ responses were
made by pressing the “z” key on the computer keyboard. Participants were instructed
to respond as quickly and as accurately as possible. Response latencies were measured to the nearest ms. Stimulus order was randomized separately for each participant. The intertrial interval was 2000 ms.
For the LDT, each participant Wrst completed 20 practice trials, consisting of 10
words similar in normative frequency to the experimental items and 10 pseudohomophones. For the PLDT, each participant Wrst completed 20 practice trials, consisting
of 5 words similar in normative frequency to the experimental items, 5 pseudohomophones, and 10 nonwords.
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
439
3. Results
Response latencies less than 250 ms or more than 2500 ms were treated as outliers and removed from the dataset. In addition, for each participant, response
latencies greater than 2.5 SD from the cell mean of each condition were also
treated as outliers and removed. A total of 51 observations (3.54% of the data)
were removed from the LDT and 50 observations (3.47% of the data) were
removed from the PLDT. The mean response latencies of correct responses and
mean error percentages for all stimuli are presented in Table 2. In the subject
analyses (F1) BOI was a within-subject manipulation; in the item analyses (F2)
BOI was a between-item manipulation.
3.1. LDT
In the analysis of the response latency data, the 35 ms BOI eVect was statistically
signiWcant, F1(1, 29) D 14.39, MSe D 1278.89, p < .005, 2 D .33; F2(1, 46) D 5.45,
MSe D 2558.94, p < .05, 2 D .10. Responses to high BOI words were faster than
responses to low BOI words. There was no BOI eVect on error rates (both Fs < 1).
3.2. PLDT
In the analysis of the response latency data, the 36 ms BOI eVect was statistically
signiWcant, F1(1, 29) D 12.80, MSe D 1485.47, p < .005, 2 D .30; F2(1, 46) D 7.35,
MSe D 1964.13, p < .01, 2 D .13. Like the lexical decision responses, phonological lexical decisions were faster to high BOI words than to low BOI words. Error rates were
not analyzed because there were too few errors for a proper analysis (less than 1%
overall).
Table 2
Mean response latencies (in milliseconds) and standard errors, and mean response error percentages and
standard errors
Stimulus
LDT
PLDT
M
SE
M
SE
Response latencies
High BOI
Low BOI
BOI eVect
Pseudohomophones
Nonwords
628
663
35
817
–
17.1
21.8
727
763
36
1032
1288
22.0
26.9
Response errors
High BOI
Low BOI
BOI eVect
Pseudohomophones
Nonwords
3.36
3.42
0.06
6.08
–
30.3
–
0.8
0.8
0.9
–
0.57
1.00
0.43
10.47
9.31
35.5
45.0
0.3
0.4
1.6
1.2
440
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
4. Discussion
The purpose of this study was to determine whether sensorimotor knowledge, as
measured by BOI, inXuences the processing of written words. We observed a facilitatory BOI eVect – faster responses to high BOI words than to low BOI words – in both
the LDT and the PLDT.1 These results, like the results of Myung et al. (2006; Experiment 1), suggest that lexical semantics (conceptual knowledge that is accessed in
word recognition tasks) includes information about sensorimotor experience with
objects. This requires an expansion of the traditional notion of lexical semantic content, to include information gained through nonverbal, primarily motor, experience.
The semantic feedback framework, introduced in Section 1.1, provides an explanation for the mechanism by which BOI facilitates visual word recognition. According to
this framework, words with relatively richer semantic representations (measured by
BOI in the present study) provide stronger feedback activation to orthography and to
phonology, allowing faster responding. What has to be explained, of course, is how
sensorimotor information is incorporated in semantic memory. On this issue, an
embodied cognition framework, such as that provided by Barsalou’s (1999) perceptual
symbol systems theory, provides one account for how sensorimotor knowledge, as
measured by BOI, may be stored in and retrieved from memory.
Barsalou (1999, 2003a, 2003b; Barsalou, Simmons, Barbey, & Wilson, 2003) proposed the perceptual symbol systems theory to describe how semantic knowledge is
grounded in sensorimotor experience. According to the theory, knowledge is
acquired through sensorimotor experience and retrieval of that knowledge involves
simulation or partial reenactment of the sensorimotor states implicated at encoding.
Semantic knowledge is thus represented in terms of simulators. Activation for a particular concept involves activating a subset of the knowledge that is represented in a
simulator and running a simulation that reenacts some of what is known about that
concept. As Barsalou (2003a) put it,
“Consider the category of CARS. Visual information about how cars look is
integrated with auditory information about how they sound, olfactory information about how they smell, motor information about driving them, somatosensory information about feeling the ride in them, and emotional information
associated with speed, dangerous situations, etc. The resulting representation is
a distributed systemƒ that establishes knowledge about CARS.” (p. 1180)
Thus, sensorimotor information is proposed to be incorporated as conceptual
knowledge. As such, the BOI eVects we have observed are quite compatible with perceptual symbol systems theory.
1
The high BOI words and the low BOI words were matched as closely as possible on all of the lexical
and semantic variables listed in Table 1, but there were still small diVerences between the two sets of words
on a few of these variables. Statistically equating the two sets of words on each of these variables via analyses of covariance (ANCOVA) demonstrated that these diVerences were not important. The eVect of BOI
was statistically signiWcant in all of the ANCOVAs with one minor exception: when concreteness was the
covariate in the analysis of the LDT response latencies the BOI eVect was marginally signiWcant (p D .06).
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
441
4.1. Conclusion
In the present work, we report a new semantic eVect: the BOI eVect. This eVect
underscores the need for theories of visual word recognition to include sensorimotor
knowledge as an essential form of knowledge contained in lexical semantics. Our
results provide support for the idea that oV-line cognition is to a signiWcant extent
embodied – that is, inXuenced by prior sensorimotor experience with the world.
Acknowledgements
This research was supported by Natural Sciences and Engineering Research Council of Canada (NSERC) discovery grants to P.D.S., P.M.P., and C.R.S. We thank Shirley Fortin, Kim Foster, Keri Locheed, and Linda Kerswell for their help in data
collection. We also thank three anonymous reviewers for their very helpful comments.
Appendix A. Items used in the experiments
A.1. High BOI words
belt brick couch crown crumb dish drum fence Xute gift grape lamp mask pear
pipe purse rope skirt stool suit tape thorn tool vest.
A.2. Low BOI words
cake cliV cloud clown creek dirt ditch dorm Xame Xood juice kite lace leaf mist
pond seed shelf ship silk smog torch tribe tube.
A.3. Pseudohomophones
berd boal boan bote brane crain dait doar drane gaim gard goast gote groop gurl
hoam hoap hoze jale jerm jirk joak klaim koast nale noat nurve rane rong rore roze
scail sheat shurt skalp skarf sleap smoak stawl stoar swet teath thret tode treet tutch
werk wheet.
A.4. Nonwords
bame beal besh bime binch bope bram brame brank brate bulch chate cheen clace
clirp crong cruss dack dake dawk dreeb dunch duss fage Wlt Wtch Xane Xang Xef Xet
foom fulk fung gake gick glank gless grabe grafe gurse hain hape hean helt hife hine
jick jote kine kooce loke ludge meep merch moach nent nerbe pake pame pape pell
petch pilk pleap poote potch pribe prog pung rame rask rell scaV scug shate shink
slirt soat spale spen spoop stort strup tain talt tane tark thurn tinch toin trake treen
trine turt vank yelf.
442
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
References
Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral & Brain Sciences, 22, 577–660.
Barsalou, L. W. (2003a). Abstraction in perceptual symbol systems. Philosphical Transactions of the Royal
Society of London: Biological Sciences, 358, 1177–1187.
Barsalou, L. W. (2003b). Situated simulation in the human conceptual system. Language & Cognitive Processes, 18, 513–562.
Barsalou, L. W., Simmons, W. K., Barbey, A. K., & Wilson, C. D. (2003). Grounding conceptual knowledge
in modality-speciWc systems. Trends in Cognitive Sciences, 7, 84–91.
Christou, C. G., & BülthoV, H. H. (1999). View dependence in scene recognition after active learning. Memory & Cognition, 27, 996–1007.
Clark, A. (1997). Being there: Putting brain, body, and world together again. Cambridge, MA: MIT Press.
Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: a dual route cascaded model
of visual word recognition and reading aloud. Psychological Review, 108, 204–256.
Cortese, M. J., & Fugett, A. (2004). Imageability ratings for 3,000 monosyllabic words. Behavior Research
Methods, Instruments, and Computers, 36, 384–387.
Cortese, M. J., Simpson, G. B., & Woolsey, S. (1997). EVects of association and imageability on phonological mapping. Psychonomic Bulletin & Review, 4, 226–231.
Cowart, M. (2004). Embodied cognition. Internet encyclopedia of philosophy. Available from http://
wwww.iep.utm.edu/.
de Groot, A. M. B. (1989). Representational aspects of word imageability and word frequency as assessed
through word association. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15,
824–845.
DirectRT Software [Computer software]. (2006). Available from http://www.empirisoft.com/
DirectRT.aspx.
Glenberg, A. M., & Kaschak, M. P. (2002). Grounding language in action. Psychonomic Bulletin & Review,
9, 558–565.
Hino, Y., & Lupker, S. J. (1996). EVects of polysemy in lexical decision and naming: an alternative to lexical access accounts. Journal of Experimental Psychology: Human Perception and Performance, 22,
1331–1356.
Hino, Y., Lupker, S. J., & Pexman, P. M. (2002). Ambiguity and synonymy eVects in lexical decision, naming, and semantic categorization tasks: interactions between orthography, phonology, and semantics.
Journal of Experimental Psychology: Learning, Memory, and Cognition, 28, 686–713.
ITP Nelson Canadian dictionary of the English language: An encyclopedic reference. (1997). Scarborough,
ON: ITP Nelson.
James, C. T. (1975). The role of semantic information in lexical decisions. Journal of Experimental Psychology: Human Perception and Performance, 1, 130–136.
James, K. H., Humphrey, G. K., Vilis, T., Corrie, G., Baddour, R., & Goodale, M. A. (2002). “Active” and
“passive” learning of three-dimensional object structure within an immersive virtual reality environment. Behavior Research Methods, Instruments, & Computers, 34, 383–390.
Kroll, J. F., & Mervis, J. S. (1986). Lexical access for concrete and abstract words. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 12, 92–107.
LakoV, G., & Johnson, M. (1999). Philosophy in the Xesh: The embodied mind and its challenge to western
thought. New York: Basic Books.
Magnie, M., Besson, M., Poncet, M., & Dolisi, C. (2003). The Snodgrass and Vanderwart set revisited:
norms for object manipulability and for pictorial ambiguity of objects, chimeric objects, and nonobjects. Journal of Clinical and Experimental Neuropsychology, 25, 521–560.
Markman, A. B., & Brendl, C. M. (2005). Constraining theories of embodied cognition. Psychological Science, 16, 6–10.
Myung, J.-Y., Blumstein, S. E., & Sedivy, J. C. (2006). Playing on the typewriter, typing on the piano:
manipulation knowledge of objects. Cognition, 98, 223–243.
Pecher, D. (2001). Perception is a two-way junction: feedback semantics in word recognition. Psychonomic
Bulletin & Review, 8, 545–551.
P.D. Siakaluk et al. / Cognition 106 (2008) 433–443
443
Pecher, D., & Zwaan, R. A. (2005). Introduction to grounding cognition. In D. Pecher & R. A Zwaan
(Eds.), Grounding cognition: The role of perception and action in memory, language, and thinking (pp. 1–
7). Cambridge, UK: Cambridge University Press.
Pexman, P. M., & Lupker, S. J. (1999). Ambiguity and visual word recognition: Can feedback explain both
homophone and polysemy eVects? Canadian Journal of Experimental Psychology, 53, 323–334.
Pexman, P. M., Lupker, S. J., & Hino, Y. (2002). The impact of feedback semantics in visual word recognition: number-of-features eVects in lexical decision and naming tasks. Psychonomic Bulletin & Review, 9,
542–549.
Plaut, D. C., McClelland, J. L., Seidenberg, M. S., & Patterson, K. (1996). Understanding normal and
impaired word naming: computational principles in quasi-regular domains. Psychological Review, 103,
56–115.
Rodd, J. M., Gaskell, M. G., & Marslen-Wilson, W. D. (2002). Making sense of semantic ambiguity:
semantic competition in lexical access. Journal of Memory and Language, 46, 245–266.
Sadoski, M., Kealy, W. A., Goetz, E. T., & Paivio, A. (1997). Concreteness and imagery eVects in the written composition of deWnitions. Journal of Educational Psychology, 89, 518–526.
Stone, G. O., Vanhoy, M., & Van Orden, G. C. (1997). Perception is a two-way street: feedforward and
feedback phonology in visual word recognition. Journal of Memory and Language, 36, 337–359.
Strain, E., Patterson, K., & Seidenberg, M. S. (1995). Semantic eVects in single-word naming. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 21, 1140–1154.
Toglia, M. P., & Battig, W. F. (1978). Handbook of semantic word norms. Hillsdale, NJ: Erlbaum.
Wilson, M. (2002). Six views of embodied cognition. Psychonomic Bulletin & Review, 9, 625–636.
Wolk, D. A., Coslett, H. B., & Glosser, G. (2005). The role of sensory-motor information in object recognition: evidence from category-speciWc visual agnosia. Brain and Language, 94, 131–146.