Research Article

Representation of Categories
Metaphorical Use of the Container Schema
Inge Boot and Diane Pecher
Erasmus University Rotterdam, The Netherlands
Abstract. In the present study we investigated whether the mental representation of the concept categories is represented by the container image
schema (Lakoff & Johnson, 1980). In two experiments participants decided whether two pictures were from the same category (animal or
vehicle). Pictures were presented inside or outside a frame that should activate the container schema. We found that performance to pictures was
inuenced by the frame in congruence with the metaphorical mapping (same category inside bounded region; different category not in same
bounded region). These results show that the concept categories is metaphorically represented by containers.
Keywords: conceptual metaphor, abstract concepts, grounded cognition, image schema

Recent theories have postulated that mental representations

and sensory-motor processing share mechanisms. According
to these theories, the mental representations of concepts are
determined by the sensory-motor experiences that an organism has (e.g., Barsalou, 1999; Glenberg, 1997; Goldstone,
1994). Concrete concepts are physical entities in the world
(e.g., garbage can). The body has physical experiences with
such concept by perception (e.g., seeing or touching a garbage
can) and interaction (e.g., throwing something in the garbage
can or taking a full garbage bag out of the garbage can).
According to theories of grounded cognition, such sensorymotor experiences form the mental representation of the concept (e.g., garbage can). In contrast, abstract concepts (e.g.,
ideas) are not physical entities in the world. Even though
our bodies cannot have direct physical experiences with
abstract concepts, some theories, however, provide a framework in which mental representations of abstract concepts
can still be grounded in sensory-motor experience (e.g.,
Barsalou, 1999; Barsalou & Wiemer-Hastings, 2005; Barwise
& Perry, 1983; Lakoff & Johnson, 1980; Langacker, 1986).
Lakoff and Johnsons Conceptual Metaphor Theory (CMT)
has been one of the most inuential of such frameworks. Lakoff and Johnson claim that people represent abstract concepts
in terms of concrete concepts by metaphorical mapping. In the
present paper we investigated how a particular abstract concept, categories, is represented. In the light of the CMT, we
specically investigated the idea that people represent the
abstract concept categories in terms of the concrete concept
containers. Metaphorical mapping such as those between categories and containers may explain how even abstract concepts could be formed by sensory-motor experiences (e.g.,
Barsalou, 1999; Barwise & Perry, 1983; Langacker, 1986).
Lakoff and Johnson (1980) claimed that metaphorical
sentences such as Are tomatoes in the fruit or vegetable
category? reect underlying conceptual mappings (e.g.,
categories are containers). By this they mean that the menExperimental Psychology 2011; Vol. 58(2):162170
DOI: 10.1027/1618-3169/a000082

tal structure of a concrete concept (e.g., container) is

mapped onto an abstract concept (e.g., category). During
this mapping the image schematic structure of concrete concepts is preserved. Image schemata are conceptual structures
representing spatial relations and movements in space. They
are argued to be formed by interaction of our body and
senses with objects in the world during our childhood
(Mandler, 1992) and ensuingly used to comprehend concepts in the world (Johnson, 1987; Lakoff & Johnson,
1999). For instance, activation of the container schema
should make experiences like pouring water in a cup (a container being lled) coherent. According to the CMT not only
the concrete situation as described above, but also the
abstract concept categories should also activate the container schema through metaphorical mapping (Gibbs,
1994; Johnson, 1987; Lakoff & Johnson, 1980, 1999).
Interestingly, the CMT provides a theory about the origins of metaphorical mappings. Some metaphorical mappings, called primary metaphors, are directly formed by
concrete situations (Grady, 1997 as cited in Lakoff &
Johnson, 1999). Since categories are containers is a primary
metaphor we will focus on this sort of conceptual mapping.
Primary metaphors such as categories are containers are
assumed to be formed directly through concrete experience
where the abstract and concrete concept co-occur (e.g.,
kitchen tools and socks are put in different drawers)
(Johnson, 1987, 1997; Lakoff & Johnson, 1980, 1999).
People can interact with the concrete concept (e.g., containers) while the abstract concept (e.g., categories) is present.
The correlation between the concepts in such a situation provokes a metaphorical link between the concrete and abstract
concept. Later, in situations where only the abstract concept
is present (e.g., reasoning about categories) this metaphorical
link causes the mental representation of the concrete concept
(e.g., container schema) to be used to understand the abstract
concept in that situation.
 2010 Hogrefe Publishing

I. Boot & D. Pecher: Categories are Containers

There is now a growing body of evidence in support of

the CMT, such as for the metaphor time is space (Boroditsky,
2000; Casasanto & Boroditsky, 2008), similarity is closeness
(Boot & Pecher, 2010; Casasanto, 2008), good is up
(Crawford, Margolies, Drake, & Murphy, 2006; Meier &
Robinson, 2004) and power is up (Giessner & Schubert,
2007; Moeller, Robinson, & Zabelina, 2008; Schubert,
2005; Zanolie et al., 2010), social attachment is closeness
(Williams & Bargh, 2008a) and affection is warmth
(Williams & Bargh, 2008b; Zhong & Leonardelli, 2008).
In the present study we wanted to extend these ndings to
the metaphor categories are containers. Importantly, we
wanted to investigate whether the container image schema
would be activated even in the absence of any linguistic
expression that might contain the metaphor.
According to Lakoff and Johnson metaphorical mappings
are independent from metaphorical language. The conceptual
mapping is assumed to be necessary to understand the
abstract concept, and thus should get activated in any situation in which the abstract concept needs to be understood.
In the present study we did not use metaphorical language
nor literal language. We used pictures to activate the concept
of interest. The concept categories was activated by pictures
from the same or different categories (e.g., animal and vehicle). The container image schema was activated by a black
line frame that included (in the container) or excluded (out
of the container) the pictures. In this way we minimized
effects due to activation of possible polysemous words
(e.g., in and out, see Murphy, 1996) or lexical associations.
Nonlinguistic studies so far almost always used paradigms
in which the stimuli contained features of the concrete concept
(e.g., space) as well as the abstract concept (e.g., time). The
features of the concrete concept were irrelevant to the task,
whereas the features of the abstract concept were relevant.
For instance, in a study by Casasanto and Boroditsky
(2008) participants estimated presentation times of visual
stimuli (e.g., a dot). Estimations were inuenced by the spatial
displacements of the stimuli, although displacement was taskirrelevant and uncorrelated with presentation duration. The
bigger the displacement of the stimuli on the screen the longer
participants estimated the presentation duration of the stimuli.
Thus, Casasanto and Boroditsky (2008) showed that the conceptual mapping time is space is not only active during language processing but also during processing of visual
nonlinguistic materials. They also showed that participants
performance was affected by the task-irrelevant spatial
dimension. These results suggest that people use the mental
representation of space in order to fully understand time. Similarly, in the present study we used stimuli that contained the
task-irrelevant concrete concept (e.g., container) as well as the
task-relevant abstract concept (e.g., categories).
An important question that we focused on in the present
study is whether the image schema plays a role during representation of the abstract concept or during selection of the
response. According to the CMT, image schemata are part of
the representation of the abstract concept. There is some evidence, however, that irrelevant information can sometimes
affect responses in situations with high uncertainty. For
example, participants judgments of justice in conditions
of uncertainty (e.g., when there is no information about
 2010 Hogrefe Publishing

163

the others outcome) were inuenced by irrelevant information (e.g., affect, Van den Bos, 2003; Van den Bos, Lind,
Vermunt, & Wilke, 1997). In studies investigating activation
of image schemata during processing of abstract concepts a
similar type of uncertainty might play a role as well. When
participants have to choose from many response options
(e.g., ratings on a Likert scale) and are uncertain about the
accuracy of their choice, irrelevant information may affect
responses. For instance, Schubert (2005) asked participants
to judge animal pictures on respect on a scale from 1 (not
at all) to 9 (very much). The pictures of the animals could
be presented at the top or bottom of the screen. He found
that participants gave higher ratings to powerful animals
when presented at the top compared to the bottom of the
screen. These results indicate that power is partly represented by verticality. In addition, however, participants
might have used irrelevant information (i.e., picture position) to facilitate performance (i.e., judging animals on
respect) in cases where they were uncertain about the correct
response. Because judging animals on respect is subjective
and perhaps unusual for the participants, it is possible that
the irrelevant information (position of picture) might have
inuenced the response selection in congruence with the
metaphor (e.g., animals presented at the top are more powerful than animals presented at the bottom). In sum, while it
is obvious that the conceptual metaphor (e.g., power is up)
was active during performance, it is unclear whether the
image schema affected representation or response selection.
In the present study we tried to minimize such response
uncertainty. In Experiment 1 participants decided whether
two pictures were from the same or different categories.
We used only two categories that were easy to distinguish
(animals and vehicles). Additionally, we provided feedback
to give condence about the accuracy of responses. Because
this task was very easy, and uncertainty was minimized, it was
unlikely to induce the use of irrelevant information during
response selection. The irrelevant information in our task
was the position of a rectangular frame that was presented
with the pictures. Both or only one picture could be presented
inside the frame. The frame visualized a bounded region that
should activate the container schema in its concrete meaning.
The task itself should activate the concept categories that in
turn should activate the container schema. The congruent trials were those in which the activation of the container schema
matched for the irrelevant frame and the category decision
(things from the same category are in the same bounded
region or things from different categories are not in the same
bounded region). Incongruent trials were those in which the
container schema mismatched for the irrelevant frame and
the category decision (things from the same category are
not in the same bounded region or things from different categories are in the same bounded region). Equal numbers of
congruent and incongruent trials were presented so that the
irrelevant information (frame) was not predictive or helpful
to select the correct response, which should further discourage
participants from using the irrelevant information. If we still
obtained an effect of the image schema on categorization decisions in congruence with the metaphor, this would support the
idea that the container image schema is an essential part of the
representation of the concept category.
Experimental Psychology 2011; Vol. 58(2):162170

164

I. Boot & D. Pecher: Categories are Containers

In summary, our aim of the present study was to investigate the mental representation of the concept categories. In
the light of the CMT we examined whether categories is metaphorically represented by containers. Additionally, we
examined whether this metaphorical mapping occurs during
a conceptual task on nonlinguistic stimuli and whether the
metaphorical mapping occurs during mental representation
of the abstract concept rather than during response selection.
The conceptual metaphor categories are containers is of special interest, because it has not been investigated previously.
Moreover, two important factors were controlled in order
to exclude confounding effects. First, the target pictures
were centered on the screen and the distance between them
was identical between trials. In this way eye movements
were minimized and the conceptual mapping similarity is
closeness could not inuence outcomes (see Boot & Pecher,
2010; Casasanto, 2008). Second, differences in visual complexity between trials in which both pictures were presented
inside the frame and trials in which one was outside the
frame could not affect the predicted interaction because both
congurations appeared in both the congruent and incongruent conditions.
If the concept of categories is conceptually represented
by containers, we would expect to nd an interaction. Performance for pictures from the same category should be better when presented both inside the frame than when one was
outside the frame, whereas performance for pictures from
different categories should be worse when presented both
inside the frame than when one was outside the frame.

of vehicles (using each vehicle two times), and 10 pairs of

an animal with a vehicle. In the Animal condition (both
animals or not) the vehicle pairs were not used, and in
the Vehicle condition (both vehicles or not) the animal
pairs were not used. Thus, in each task 20 pairs were presented. The pairs were presented next to each other in the
center of the screen. The relative position of the two pictures (left or right) was counterbalanced across repetitions
of the pair (two relative positions). The pairs were presented together with a frame that was a black lined square
of 8 cm by 8 cm. This frame was either presented in the
middle so that both pictures were inside the container or
moved 4 cm to the left or right from the center, so that only
one picture was in the container and one outside. For each
formed pair we created two identical slides in which the
container was in the middle, one with the container on
the right and one with the container on the left (four frame
positions of which two were identical). All slides were presented twice (two presentations). This resulted in 320 trials
for each condition (Animal or Vehicle). Additionally, 16
practice trial slides with another set of pictures of animals
and vehicles were created. The congruent trials were the
slides with two objects from the same category presented
both inside the container and slides in which the pictures
were from different categories and one was presented inside
and the other outside the container. Incongruent trials were
slides in which both pictures were animals or vehicles but
one was presented outside and the other inside the container
and slides in which the pictures were from different categories and both presented inside the container. Examples of
the four conditions are shown in Figure 1.

Experiment 1

Procedure

Method

The instruction and stimuli were presented on a computer

screen. Participants were told that two categories were used
in this experiment: animals and vehicles. Participants were
randomly assigned to the Animal or Vehicle Condition. In
the Animal Condition participants had to decide if the two
pictures presented together on the screen were both animals
or not. In the Vehicle Condition participants had to decide
whether the two pictures were both vehicles or not. They
were told to use the z-button and m-button to respond on
a QWERTY keyboard. The mapping of the responses to
buttons was counterbalanced. The pictures were presented
with the frame at different positions. Participants were told
to ignore the frame and focus on the pictures during the task.
They started with the 16 practice trials. Each trial started
with a xation point (+) presented for 500 ms. Then the target slide appeared for 1,800 ms or until a response was
given. Feedback was given for incorrect answers (Fout
Dutch translation for Incorrect) and responses slower than
1,800 ms (Te langzaam Dutch translation for Too slow)
which remained on the screen for 1,500 ms. There was no
delay between the trials. After the practice trials the 320
experimental trials were presented in random order. The
same procedure was used for the experimental trials as for
the practice trials. All 320 trials were presented in random
order without a break.

Participants and Design

Forty psychology students participated. They received
course credits or a chocolate bar as a reward for their effort.
Position of the pictures with respect to the frame and same
versus different category were manipulated within subjects,
type of categorization was manipulated between subjects. In
order to counterbalance the materials over the experimental
conditions, 20 participants decided whether two pictures
were both animals or not (Animal condition), the other 20
participants decided whether two pictures were both vehicles or not (Vehicle condition).
Materials
We selected 10 pictures of animals and 10 pictures of vehicles from Bonin, Peereman, Malardier, Meot, and Chalard
(2003), Staneld and Zwaan (2001) and Zwaan, Staneld,
and Yaxley (2002), http://leadserv.u-bourgogne.fr/bases/
pictures/), Pecher, Zanolie, and Zeelenberg (2007), and similar line drawings found on the Internet. We created 10
pairs of animals (using each animal two times) and 10 pairs
Experimental Psychology 2011; Vol. 58(2):162170

 2010 Hogrefe Publishing

I. Boot & D. Pecher: Categories are Containers

165

Figure 1. The four conditions of Experiment 1 in the Animal Task. In the Same-In Condition both animal pictures were
presented in the frame (A), in the Same-In-Out Condition one animal picture was presented in and one outside the frame
(B), in the Different-In Condition an animal and a vehicle picture were both presented in the frame (C), and in the
Different-In-Out Condition one of the pictures (animal or vehicle) was presented in and one outside the frame (D).
The pairs of pictures were horizontally centered. The frame was either presented in the center or 4 cm from the center to
the left or right on the screen.

Results
We analyzed all reaction times of correct responses within 2
standard deviations from each subjects mean. This trimming
procedure resulted in a removal of 5.6% of the correct reaction times in the Animal Condition and 4.8% in the Vehicle
Condition. The means and error rates with the within-subject
standard error of the mean (see Loftus & Masson, 1994) for
each Task Condition are shown in Figure 2.
The reaction times and error rates were submitted to a 2
(Category: Same vs. Different) 2 (Container: Both Inside
vs. One Outside) repeated measures ANOVA with Task
(animal vs. vehicle) as a between-subjects factor. In the reaction times we obtained an interaction effect of Category and
Container, F(2, 38) = 25.79, MSE = 180.3, p < .001. This
interaction effect was not signicantly different between the
Animal Task and the Vehicle Task, F < 1. With paired sample
t tests with Bonferroni correction for multiple comparison we
found that participants responded faster to pictures from the
same category that were both presented inside the container
than when one was presented outside the container,
t(19) = 7.50, SE = 3.0, p < .001, whereas participants did
not respond differently to pictures from different categories
both presented inside the container or one outside, p > .25.
Furthermore we found that participants responded faster
to pictures if they were presented both inside the container than when one was presented outside the container,
 2010 Hogrefe Publishing

F(2, 38) = 23.61, MSE = 234.9, p < .001. We also found

an overall effect of Category, F(2, 38) = 14.80,
MSE = 528.5, p < .001. The interaction effect of Category
and Task Condition, F(2, 38) = 5.03, MS = 2657.9, p < .05
shows that only in the Animal Task Condition participants
responded faster to pictures from the same category than to
pictures from different categories, F(1, 19) = 16.16,
MSE = 606.3, p < .01, while in the Vehicle Task Condition
there was no difference in reaction times between the Same
and Different Category, F(1, 19) = 1.51, MSE = 450.0,
p > .10.
In the error rates we found an interaction effect of Category and Container, F(2, 38) = 10.47, MSE = .004, p <
.005. This interaction effect was not signicantly different
for the Animal Task and the Vehicle Task, F < 1. With paired
sample t tests with Bonferroni correction for multiple comparison we found that participants responded more accurately to
pictures from the same category that were both presented
inside the container than when one was presented outside
the container, t(19) = 3.50, SE = .004, p < .005, whereas
participants did not respond differently to pictures from different categories that were both presented inside the container or
one outside the container, t(19) = 1.00, SE = .005, p > .10.
Furthermore we found an overall effect of Category,
F(2, 38) = 14.81, MSE = .000, p < .001. The interaction
effect of Category and Task Condition, F(2, 38) = 6.44,
MS = .002, p < .025 shows that only in the Animal Task
Experimental Psychology 2011; Vol. 58(2):162170

166

I. Boot & D. Pecher: Categories are Containers

Figure 2. Mean reaction times in

milliseconds and error percentages for the Animal Task Condition (A) and Vehicle Task
Condition (B) in Experiment 1.
The error bars present the withinsubject standard error of the mean
(Loftus & Masson, 1994) for each
Task Condition.

Condition participants responded more accurately to pictures

from the same category than pictures from different categories, F(1, 19) = 17.01, MSE = .000, p < .0025 while in the
Vehicle Task Condition there was no difference in error rates
between the Same and Different Category Conditions,
F(1, 19) = 1.07, MSE = .000, p < .30.
We obtained the expected interaction effect in both the
reaction times and error rates. With paired sample t test
we found that this interaction effect was predominantly present in the same category condition. Performance was better
to pictures from the same category when presented in the
frame compared to one outside the frame, whereas performance on different categories was not inuenced by different frame positions. The lack of signicance on the different
categories side was probably due to the main effects of Container and Category. Nevertheless, the interaction effect in
both the reaction times and error rates clearly shows that
the position of the frame inuences the responses on same
category and different category differently. In order to ascertain that the interaction was robust we replicated Experiment
1 with a different set of materials. This replication also
showed a signicant interaction between Category and Container, F(2, 38) = 12.01, MSE = 226.7 p < .01.1

Discussion
Experiment 1 showed that the frame (container), even though
it was irrelevant for the task, still inuenced the category decision task. We obtained an interaction effect in both the reaction times and accuracy. Performance was better to pictures
from the same category when both rather than one were presented inside the container, whereas performance was not different to pictures from different categories when both
compared to one were presented inside the container. These
1

results support the idea that the concept categories is

metaphorically represented by the image schema containers.

Experiment 2
In the next experiment we wanted to eliminate an alternative
explanation. In Experiments 1 participants had to make a
binary decision (Same/Different Category or Animal/
Nonanimal) over stimuli with binary values (Both Inside/
One Outside or simply Inside/Outside). To simplify the task,
participants might have taken advantage of these binary values by aligning them (e.g., Same Category with Both
Inside). Such alignment of polarities has been suggested
by Proctor and Cho (2006) as an explanation for a variety
of binary decision tasks. There are two reasons why such
alignment seems unlikely in the present series of experiments. First, each combination of binary values was equally
likely, so there was no benet to participants to use a specic
alignment. Second, if participants used such alignment they
should have made many errors. In contrast to this prediction,
however, the error rates were extremely low (well below
0.1% in most cases). Nevertheless, to exclude alignment
as an explanation we performed Experiment 2 in which container position was manipulated at more than two levels. In
Experiment 2 pictures from the same category and different
categories were not only presented both inside or one outside the frame (as in Experiment 1) but also both outside
the frame. If participants in Experiment 1 used alignment
of the binary values, the interaction effect between category
(same vs. different) and container (both inside vs. one outside) should disappear.
On the other hand, if the image schema affects performance, we expect to obtain again an interaction between

In Experiment 1 the pictures were from a small subcategory of Animal (four legged mammals) and Vehicle (motor driven vehicles). Thus,
the same category pictures were visually quite similar. We replicated Experiment 1 with a different set of pictures (20 line drawings of
animals and 20 line drawings of vehicles from the same sources as in Experiment 1) to extend the ndings of Experiment 1. The subset we
used in the replication were more varied pictures from different subcategories. For example, for animals, beside four-legged mammals, we
used also winged animals, an insect, sh, and reptiles and for Vehicles, beside four-wheeled transport, we used also aviations, boats, and
nonmotorized transport. By using a more varied set of pictures that were visually less similar we showed that the conceptual mapping
generalizes to broader categories.

Experimental Psychology 2011; Vol. 58(2):162170

 2010 Hogrefe Publishing

I. Boot & D. Pecher: Categories are Containers

category and frame position. Moreover, we expect the effect

of frame position to be graded. In terms of the metaphorical
mapping, two things inside the container are members of
the same category, whereas something outside the container
is not in the same category as something inside the container.
However, two things outside the container might be members
of the same or different categories (e.g., two things that are not
food could be two animals or one animal and one vehicle).
Therefore, we expect that same category decisions will be
fastest when both pictures are inside the container, slowest
when one picture is inside the container, and somewhat intermediate when both pictures are outside the container. For different category decisions the effect should be opposite. As in
the previous experiments visual complexity was equated for
the congruent and incongruent trials and thus should not affect
the predicted interaction effect.

167

Method
Participants
Thirty psychology students who did not participate in
Experiment 1 received course credits for participating.
Materials
We selected the same line drawings (10 animals and 10
vehicles) as in Experiment 1. The same 10 pairs of animals
(using each animal two times) and 10 pairs of an animal
with a vehicle were used as in the Animal Condition of
Experiment 1. The pictures were presented together with

Figure 3. The six conditions of Experiment 4. In the Same-In Condition two animal pictures were presented in the frame
(A), in the Same-Out Condition both animal pictures were presented outside the frame (B), in the Same-In-Out Condition
one animal picture was presented in and one outside the frame (C), in the Different-In Condition an animal and a vehicle
picture were both presented in the frame (D), in the Different-Out Condition both the animal picture and the vehicle picture
were presented outside the frame (E), and in the Different-In-Out Condition one of the pictures (animal or vehicle) was
presented in and one outside the frame (F). The pairs of pictures were horizontally centered. The frame was either presented
in the center, 4 cm from the center to the left or right on the screen, or 8 cm from the center to the left or right on the screen.
 2010 Hogrefe Publishing

Experimental Psychology 2011; Vol. 58(2):162170

168

I. Boot & D. Pecher: Categories are Containers

the frame in the same way as in Experiment 1 with an additional third condition in which the frame was moved 8 cm
to the left or right from the center, so that both pictures were
outside the container (altogether six frame positions of
which two were identical). Again, the relative position of
the two pictures was counterbalanced across repetitions
of the pair, but without mirroring the pictures as we did in
Experiment 1. Each slide was presented twice. This resulted
in 480 trials. Additional 18 practice trial slides with the same
set of pictures as in the practice trials of the Animal Condition of Experiment 1 were created. Examples of the six
conditions are shown in Figure 3.
Procedure
The procedure was the same as in the Animal Condition of
Experiment 1. After the 18 practice trials, the 480 experimental trials were presented in random order. Participants
could take a break after 240 experimental trials.

Results and Discussion

We analyzed all reaction times of correct responses within 2
standard deviations from each subjects mean. This trimming
procedure resulted in a removal of 4.7% of the correct reaction times. The means and error rates with the within-subject
standard error of the mean (see Loftus & Masson, 1994) for
each Task Condition are shown in Figure 4. All t tests used a
Bonferroni correction for multiple comparisons.
The reaction times and error rates were submitted to a 2
(Category: Same vs. Different) 3 (Container: Both Inside,
Both Outside, One Outside) repeated measures ANOVA. In
the reaction times we obtained an interaction effect of Category and Container, F(2, 58) = 13.44, MSE = 137.6,
p < .001. This interaction was due to signicant effects of
frame position on same category decisions. Participants
responded faster to pictures from the same category that
were both presented inside the container (M = 504) than
when one was presented outside the container (M = 531),
t(29) = 7.00, SE = 3.8, p < .001. Additionally we found

that the effect in the same category condition of the three

positions of the frame was graded. Participants responded
faster when the pictures from the same category were both
presented outside the container (M = 516) than when one
was presented outside the container (M = 531),
t(29) = 6.38, SE = 2.3, p < .001, and when they were both
presented inside the container (M = 504) than when they
were both presented outside the container (M = 516),
t(29) = 3.75, SE = 3.2, p < .005. In contrast, reaction times
to pictures from different categories were not signicantly
affected by frame position, all ps > .05.
Furthermore, we found a main effect of Container,
F(2, 58) = 34.27, MSE = 107.4, p < .001. Participants
responded faster to pictures presented both inside the container (M = 517) compared to both outside (M = 523),
t(29) = 2.62, SE = 2.08, p < .05, and one outside the container (M = 533), t(29) = 7.92, SE = 2.0, p < .001. Participants were also signicantly faster to respond to pictures
presented both outside compared to one outside the container, t(29) = 6.21, SE = 1.61, p < .001. Moreover, participants were faster to respond to pictures from the same
category (two animals) than to pictures from different categories (one animal and one vehicle), F(1, 29) = 10.04,
MSE = 928.3, p < .005.
In the error rates we found an interaction effect of
Category and Container, F(2, 58) = 5.26, MSE = .0002,
p < .01. Participants made slightly less errors to pictures
from the same category that were both presented inside
the container (M = .024) than when one was presented outside the container (M = .030), this difference approached
signicance, t(29) = 2.26, SE = .003, p < .10. The other
differences were in the predicted direction, but none of these
were signicant. Furthermore, we found a main effect of
Category, F(1, 29) = 22.53, MSE = .0003, p < .001. Participants responded more accurate to pictures from the same
category (two animals) than to pictures from different categories (one animal and one vehicle).
The present experiment showed that the frame (container) inuenced the reaction times in the category decision
task in a graded fashion across three different positions, consistent with the metaphor. These results tell us that the effect
obtained in Experiment 1 was not due to alignment of binary
values. The interaction between category and container is
still present when we used a third value. These data are in
line with the idea that we use image schemata of concrete
concepts to understand abstract concepts.

General Discussion

Figure 4. Mean reaction times in milliseconds and error

percentages for the category decision tasks in Experiment
4. The error bars present the within-subject standard error
of the mean (Loftus & Masson, 1994).
Experimental Psychology 2011; Vol. 58(2):162170

In the present study we investigate the mental representation

of categories in the light of the CMT (Lakoff & Johnson,
1980, 1999). The CMT holds that abstract concepts (e.g.,
categories) are metaphorically represented by concrete concepts (e.g., containers). Participants had to perform a category decision task on pictures that were presented together
with a container. In Experiment 1 category decisions were
affected by whether both pictures were inside the container
or not. In Experiment 2 we controlled for binary alignment
 2010 Hogrefe Publishing

I. Boot & D. Pecher: Categories are Containers

and again obtained an effect of the position of pictures relative to the container. These ndings are congruent with the
idea that the concept categories is metaphorically represented by the concept containers. Because our paradigm
used pictorial material and an easy task in which the correct
response was absolutely clear these results are not likely to
be due to activation of linguistic metaphors or use of irrelevant information for response selection.
Beside the expected interaction effect, a main effect of
Container was obtained in the reaction times for all experiments. Participants responded faster when pictures were presented inside the container than when one or both pictures
were presented outside the container. A possible explanation
is that the frame drew peoples visual spatial attention more
to what was inside the frame than to what was outside. This
main effect of Container may explain why responses to pictures from different categories were not faster when one picture was outside the container (Congruent condition) than
when the pictures were both inside the container (Incongruent condition). Probably, the effect of congruence was counteracted by the main effect of Container.
The interaction effects in the present study are consistent
with previous research investigating the metaphorical mapping of other abstract concepts with nonlinguistic materials.
Several studies have shown interactions between concrete
and abstract domains using visual stimuli that contained
both the concrete and abstract concept (Boot & Pecher,
2010; Breaux & Feist, 2008; Casasanto, 2008; Casasanto
& Boroditsky, 2008; Crawford et al., 2006; Meier, Hauser,
Robinson, Friesen, & Schjeldahl, 2007; Schubert, 2005).
These effects were congruent with the metaphor (e.g., participants judged animals as more powerful when presented
higher compared to lower on the screen). When nonlinguistic materials are used, results are no longer due to presentation of metaphorical language. Even though it might be
possible that participants verbalized the task, it is unlikely
that they verbalized the metaphor categories are containers.
Thus, the present study showed that even in a nonlinguistic
context the metaphorical mapping occurs. As we discussed
earlier, in some of these previous studies it was not clear
whether it was the mental representation of abstract concepts
or the response selection process that was affected by image
schemata. In the present study we minimized the likelihood
that the image schema affected response selection. Nevertheless, participants performance was inuenced by the frame
in congruence with the container image schema (Lakoff &
Johnson, 1999; Lakoff & Nu
nez, 2000). In addition to recent studies these ndings provide evidence for the CMT
of Lakoff and Johnson (1980, 1999).
At present it is still an open question to what extent the
CMT holds for all metaphors found in language. Only for a
number of metaphors (e.g., similarity is closeness, power is
up, and time is spatial movement) experimental evidence is
available that supports the idea that abstract concepts are
understood by conceptual metaphorical mapping, as has
been shown in the present study. Although the evidence suggests that the role of metaphors is widespread, there are still
many metaphors that have not yet been investigated. To our
knowledge, the present study is the rst to provide evidence
for the mapping of containers on categories.
 2010 Hogrefe Publishing

169

The evidence for the role of metaphors has implications

for theories on sensory-motor grounding of cognition. Such
theories propose that mental representations of concepts are
grounded in sensory-motor experience (e.g., Barsalou, 1999;
Glenberg, 1997; Goldstone, 1994). In line with this view,
the CMT framework proposes that image schemata formed
by bodily interaction are the building blocks of concrete
concepts (Johnson, 1987) as well as of abstract concepts
through the mechanism of conceptual mapping.
Although image schemata refer to basic bodily experiences, they do not represent full and rich sensory-motor
experiences. Beside the container image schema, the concept categories has other features, such as the fact that there
is visual similarity among exemplars or that exemplars of a
category have the same function. In addition to image schemata, abstract concepts might also be represented by simulations of introspective experiences and specic situations in
which the abstract concept plays a role (Barsalou, 1999;
Barsalou & Wiemer-Hasings, 2005). Together with these
proposals, the growing body of evidence for the role of
image schemata supports the idea that, like those for
concrete concepts, mental representations of abstract
concepts can be grounded in sensory-motor experiences.
Acknowledgments
This research was supported by a grant from the Netherlands
Organization for Scientic Research (NWO) to Diane Pecher.

References
Barsalou, L. W. (1999). Perceptual symbol system. Behavioral
and Brain Sciences, 22, 577660.
Barsalou, L. W., & Wiemer-Hasings, K. (2005). Situating
abstract concepts. In D. Pecher & R. A. Zwaan (Eds.), The
grounding of cognition: The role of perception and action in
memory, language, and thinking. Cambridge: Cambridge
University Press.
Barwise, J., & Perry, J. (1983). Situations and attitudes.
Cambridge: MIT Press.
Bonin, P., Peereman, R., Malardier, N., Meot, A., & Chalard, M.
(2003). A new set of 299 pictures for psycholinguistic
studies: French norms for name agreement, image agreement,
conceptual familiarity, visual complexity, image variability,
age of acquisition, and naming latencies. Behavior Research
Methods, Instruments, & Computers, 35, 158167.
Boot, I., & Pecher, D. (2010). Similarity is closeness: Metaphorical mapping in a conceptual task. Quarterly Journal of
Experimental Psychology, 63, 942954.
Boroditsky, L. (2000). Metaphorical structuring: Understanding
time through spatial metaphors. Cognition, 75, 128.
Breaux, B. O., & Feist, M. I. (2008). The color of similarity. In B.
C. Love, K. McRae, & V. M. Sloutsky (Eds.), Proceedings of
the 30th Annual Conference of the Cognitive Society (pp.
253258). Austin, TX: Cognitive Science Conference.
Casasanto, D. (2008). Similarity and proximity: When does close
in space mean close in mind? Memory & Cognition, 36,
10471056.
Casasanto, D., & Boroditsky, L. (2008). Time in the mind: Using
space to think about time. Cognition, 106, 579593.
Crawford, L. E., Margolies, S. M., Drake, J. T., & Murphy, M. E.
(2006). Affect biases memory of location: Evidence for the
Experimental Psychology 2011; Vol. 58(2):162170

170

I. Boot & D. Pecher: Categories are Containers

spatial representation of affect. Cognition & Emotion, 20,

11531163.
Gibbs, R. W. (1994). The poetics of mind: Figurative thought,
language, and understanding. New York: Cambridge University Press.
Giessner, S. R., & Schubert, T. W. (2007). High in the hierarchy:
How vertical location and judgments of leaders power are
interrelated. Organizational Behavior and Human Decision
processes, 104, 3044.
Glenberg, A. M. (1997). What memory is for?. Behavioral and
Brain Sciences, 20, 155.
Goldstone, R. (1994). Inuences of categorization on perceptual
discrimination. Journal of Experimental Psychology: General, 123, 178200.
Johnson, M. (1987). The body in the mind: The bodily basis of
reason and imagination. Chicago: University of Chicago
Press.
Johnson, C. (1997). Metaphors vs. conation in the acquisition of
polysemy: The case of SEE. In M. K. Hiraga, C. Sinha & S.
Wilcox (Eds.), Cultural, psychological and typological issues
in cognitive linguistics (152, pp. 155169). Amsterdam:
John Benjamins.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by.
Chicago: Chicago University Press.
Lakoff, G., & Johnson, M. (1999). The philosophy in the esh:
The embodied mind and its challenge to the western thought.
New York, NY: Basic Books.
Lakoff, G., & Nu?ez, R. E. (2000). How the embodied mind
brings mathematics into being. New York, NY: Basic Books.
Langacker, R W. (1986). An introduction to cognitive grammar.
Cognitive Science, 10, 140.
Loftus, G. R., & Masson, M. E. J. (1994). Using condence
intervals in within-subject designs. Psychonomic Bulletin &
Review, 1, 476490.
Mandler, J. M. (1992). How to build a baby: II. Conceptual
primitives. Psychological Review, 99, 587604.
Meier, B. P., Hauser, D. J., Robinson, M. D., Friesen, C. K., &
Schjeldahl, K. (2007). Whats up with God? Vertical space
as a representation of the divine. Journal of Personality and
Social Psychology, 93, 599710.
Meier, B. P., & Robinson, M. D. (2004). Why the sunny side
is up: Association between affect and vertical position.
Psychological Science, 15, 243247.
Moeller, S. K., Robinson, M. D., & Zabelina, D. L. (2008).
Personality dominance and preferential use of the vertical
dimension of space: Evidence from spatial attention paradigms. Psychological Science, 19, 355361.
Murphy, G. L. (1996). On metaphoric representation. Cognition,
60, 173204.
Pecher, D., Zanolie, K., & Zeelenberg, R. (2007). Verifying
visual properties in sentence verication facilitates picture
recognition memory. Experimental Psychology, 54, 173179.

Experimental Psychology 2011; Vol. 58(2):162170

Proctor, R. W., & Cho, Y. S. (2006). Polarity correspondence:

A general principle for performance of speeded binary
classication tasks. Psychological Bulletin, 132, 416442.
Schubert, T. W. (2005). Your highness: Vertical positions as
perceptual symbols of power. Journal of Personality and
Social Psychology, 89, 121.
Staneld, R. A., & Zwaan, R. A. (2001). The effect of implied
orientation derived from verbal context on picture recognition. Psychological Science, 12, 153156.
Van den Bos, K. (2003). On the subjective quality of social
justice: The role of affect as information in the psychology
of justice judgments. Journal of Personality and Social
Psychology, 85, 482498.
Van den Bos, K., Lind, E. A., Vermunt, R., & Wilke, H. A. M.
(1997). How do I judge my outcome when I do not know the
outcome of others? The psychology of the fair process effect.
Journal of Personality and Social Psychology, 72, 1034
1046.
Williams, L. E., & Bargh, J. A. (2008a). Experiencing physical
warmth promotes interpersonal warmth. Science, 322, 606
607.
Williams, L. E., & Bargh, J. A. (2008b). Keeping ones distance:
the inuence of spatial distance cues on affect and evaluation.
Psychological Science, 19, 302308.
Zanolie, K., Van Dantzig, S., Boot, I., Wijnen, J., Schubert, T.,
Giesner, S., & Pecher, D. (2010). Mighty metaphors:
Behavioral and ERP evidence that power shifts attention on
a vertical dimension. Manuscript submitted for publication.
Zhong, C., & Leonardelli, G. J. (2008). Cold and lonely.
Psychological Science, 19, 838842.
Zwaan, R. A., Staneld, R. A., & Yaxley, R. H. (2002).
Language comprehenders mentally represent the shapes of
object. Psychological Science, 13, 168171.

Received February 10, 2010

Revision received May 21, 2010
Accepted May 22, 2010
Published online August 12, 2010
Inge Boot
Psychology Department, T12-43
Erasmus University Rotterdam
P.O. Box 1738
3000 DR Rotterdam
The Netherlands
Tel. +31 10 408 2946
Fax +31 10 408 9009
E-mail ingeboot@gmail.com

 2010 Hogrefe Publishing