Time in the mind: Using space to think
about time
Daniel Casasanto *, Lera Boroditsky
Department of Psychology, Jordan Hall, Bldg. 420, Stanford University, Stanford, CA 94305, USA
Received 16 September 2006; revised 5 March 2007; accepted 10 March 2007
Abstract
How do we construct abstract ideas like justice, mathematics, or time-travel? In this paper
we investigate whether mental representations that result from physical experience underlie
people’s more abstract mental representations, using the domains of space and time as a testbed.
People often talk about time using spatial language (e.g., a long vacation, a short concert).
Do people also think about time using spatial representations, even when they are not using
language? Results of six psychophysical experiments revealed that people are unable to ignore
irrelevant spatial information when making judgments about duration, but not the converse.
This pattern, which is predicted by the asymmetry between space and time in linguistic metaphors,
was demonstrated here in tasks that do not involve any linguistic stimuli or responses.
These findings provide evidence that the metaphorical relationship between space and time
observed in language also exists in our more basic representations of distance and duration.
Results suggest that our mental representations of things we can never see or touch may be
built, in part, out of representations of physical experiences in perception and motor action.
Ó 2007 Elsevier B.V. All rights reserved.
Keywords: Metaphor; Time; Space; Embodied cognition
0010-0277/$ - see front matter Ó 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.cognition.2007.03.004
*
Corresponding author. Tel.: +1 617 368 0792.
E-mail address: casasan@stanford.edu (D. Casasanto).
www.elsevier.com/locate/COGNIT
Available online at www.sciencedirect.com
Cognition 106 (2008) 579–593
1. Introduction
How do we mentally represent things that we have never experienced through the
senses, like justice, mathematics, or time-travel? One possibility is that sensory and
motor representations that result from physical interactions with the world are recycled
to support our thinking about abstract entities. Evidence for this view has come
from patterns observed in human languages. When speaking about abstract
domains, people often recruit metaphors from more concrete or perceptually rich
domains (Clark, 1973; Gruber, 1965; Jackendoff, 1983; Lakoff & Johnson, 1980; Pinker,
1989; Talmy, 1988).
For example, people often talk about time using spatial metaphors (e.g., a long
vacation, a short concert) (Alverson, 1994; Clark, 1973; Traugott, 1978). Aspects
of time are often said to be more ‘abstract’ than their spatial analogues because
we can perceive the spatial, but we can only imagine the temporal (Ornstein, 1969;
cf., Evans, 2004). Compare the following scenarios:
(a) They moved the truck forward two meters.
(b) They moved the meeting forward two hours.
The truck in sentence (a) is a physical object that can travel through space, and
whose motion we might see, hear, or feel. By contrast, in sentence (b) there is no
way to experience the meeting’s ‘motion’ through time via the senses.1
Importantly, the relationship between space and time in language is asymmetrical:
people talk about time in terms of space more often than they talk about space in
terms of time (Lakoff & Johnson, 1980, 1999). This pattern in language suggests that
our conceptions of space and time might be asymmetrically dependent: we construct
representations of time by co-opting mental representations of space, but not necessarily
the converse. Patterns in historical language change (Sweetser, 1991) and language
acquisition by children (e.g., Bowerman, 1983; Clark, 1973) likewise support
the idea that spatial representations are primary, and are later co-opted for other
uses such as time. Evidence from psycholinguistic experiments has also provided support
for this view, showing that people construct spatial representations on-line
when processing statements about time (Boroditsky, 2000, 2001; Boroditsky & Ramscar,
2002; Nu´n˜ez & Sweetser, 2006; Piaget, 1927/1969; Torralbo, Santiago, & Lupia´n˜ez,
2006; Tversky, Kugelmass, & Winter, 1991), but not necessarily the reverse
(Boroditsky, 2000).
In this paper we ask whether this asymmetric relationship between space and time
is limited to patterns in language and language processing, or whether it extends
beyond the domain of language. Is the way we think about time dependent on space
even when we’re not using language at all? Previous research on the experience of
1
Temporal representations are often more abstract than their spatial analogues, as this example
illustrates. However, some of our spatial representations may be quite abstract, as well. For example, our
conception of the Milky Way galaxy’s breadth is no more grounded in direct experience than our
conception of its age.
580 D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593
distance and duration has shown that the two are not independent (Benussi, 1913;
Bill & Teft, 1969; Cohen, 1967; Cohen, Hansel, & Sylvester, 1954; Collyer, 1977; Helson,
1930; Jones & Huang, 1982; Price-Williams, 1954; Sarrazin, Giraudo, Pailhous,
& Bootsma, 2004), but little is known about whether the relationship between the
two domains is asymmetrical, in the way that has been observed in language. The
purpose of the present study is to test whether the asymmetrical dependence of time
on space exists even at a more basic level of the human conceptual system.
This paper describes six psychophysical experiments that tested the separability of
distance and duration in human judgments. All stimuli and responses were non-linguistic.
In each task, participants viewed lines or dots on a computer screen, and
reproduced either their duration or their spatial displacement. Durations and displacements
were fully crossed, so there was no correlation between the temporal
and spatial components of the stimuli. As such, one stimulus dimension served as
a distractor for the other: an irrelevant piece of information that could potentially
interfere with task performance. Patterns of cross-dimensional interference were analyzed
to reveal relationships between spatial and temporal representations. We reasoned
that if spatial and temporal representations are symmetrically dependent on
one another, then any cross-dimensional interference should be approximately symmetric:
distance should modulate duration estimates, and vice versa. Alternatively, if
spatial and temporal representations are independent, there should be no significant
cross-dimensional interference. However, if mental representations of time are asymmetrically
dependent on mental representations of space as suggested by patterns in
language, then we should observe an asymmetrical pattern of cross-dimensional
interference: distance should affect duration estimates more than duration affects distance
estimates.
2. Experiment 1: Growing lines
2.1. Methods
2.1.1. Participants
Nine participants from the MIT community performed Experiment 1, in exchange
for payment.2
All participants gave informed consent, and all were native monolingual
speakers of English according to a language background questionnaire (i.e.,
2
A total of 72 subjects from the MIT community participated in Experiments 1–6, in exchange for
payment. Of these, 16 participants were removed from the analyses reported here for performing the
experiment incorrectly (e.g., estimating distance when they were instructed to estimate duration), or for
excessively poor performance: for each participant, duration estimates were plotted as a function of actual
stimulus duration, and distance estimates were plotted as a function of actual stimulus displacement.
Participants were excluded if the slope of their duration or distance estimates was less than 0.5, as such
poor performance (e.g., indicating that the 5-s lines lasted less than 2.5 s) was believed to result from
impatience with the repetitive task, rather than genuine inaccuracy.
D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593 581
English was the only language they learned before age 5, and was their strongest language
at time of test).
2.1.2. Materials
Lines of varying lengths were presented on a computer monitor (resolution
= 1024 · 768 pixels), for varying durations. Durations ranged from 1000 to
5000 ms in 500 ms increments. Displacements ranged from 200 to 800 pixels in 75
pixel increments. Nine durations were fully crossed with nine displacements to produce
81 distinct line types. Lines ‘grew’ horizontally across the screen one pixel at a
time, from left to right, along the vertical midline. Lines started growing 112 pixels
from the left edge of the monitor on average, but the starting point of each line was
jittered with respect to the average starting point (+/À up to 50 pixels), so that the
monitor would not provide a reliable spatial frame of reference. Each line remained
on the screen until it reached its maximum displacement, and then it disappeared.
2.1.3. Procedure
Participants viewed 162 growing lines, one line at a time, from a viewing distance
of approximately 50 cm. The word ‘‘ready’’ appeared in the center of an otherwise
blank screen for two seconds immediately before each line was shown. Immediately
after each line was shown, a prompt appeared in either the upper left or lower left
corner of the screen indicating that the subject should reproduce either the line’s displacement
(if an ‘X’ icon appeared), or its duration (if an ‘hourglass’ icon appeared).
To estimate displacement, subjects clicked the mouse once on the center of the X,
moved the mouse to the right in a straight line, and clicked the mouse a second time
to indicate that they had moved a distance equal to the maximum displacement of
the stimulus. Whereas stimuli grew from a jittered starting point on the vertical midline
of the screen, responses were initiated at a fixed starting point in either the upper
or lower left corner. Thus, the response was translated both vertically and horizontally
with respect to the stimulus. To estimate duration, subjects clicked the mouse
once on the center of the hourglass icon, waited the appropriate amount of time,
and clicked again in the same spot.
All responses were self-paced. For a given trial, subjects reproduced either the displacement
or the duration of the stimulus, never both. Response data were collected
for both the trial-relevant and the trial-irrelevant stimulus dimensions, to ensure that
subjects were following instructions.
2.2. Results and discussion
Results of Experiment 1 showed that spatial displacement affected estimates of
duration (y = 0.63x + 2503, r2
= .94, df = 7, p < .001), but duration did not affect
estimates of spatial displacement (y = 0.003x + 440, r2
= .05, df = 7, ns; Figs. 1a
and b, 2a). For stimuli of the same average duration, lines that traveled a shorter distance
were judged to take a shorter time, and lines that traveled a longer distance
were judged to take a longer time. Subjects incorporated irrelevant spatial information
in their temporal estimates, but not vice versa. This behavioral asymmetry was
582 D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593
predicted based on the asymmetrical relationship between time and space in linguistic
metaphors.
Overall, estimates of duration and displacement were highly accurate, and about
equally accurate in the two domains (effect of target displacement on estimated displacement:
y = .081x + 412, r2
= .99, df = 7, p < .001; effect of target duration on
estimated duration: y = .83x + 327, r2
= .99, df = 7, p < .001. Figs. 1c and d). The
asymmetrical cross-dimensional interference we observe cannot be attributed to a
difference in the overall accuracy of duration and displacement estimations, as no
significant difference was found (rduration À rdisplacement = 0.00, z = 0.00, ns).
Fig. 1. Grand averaged duration and displacement estimates for Experiment 1. Top: cross-domain effects.
(a) left: effect of actual line displacement on estimated duration. (b) right: effect of actual line duration on
estimated displacement. The horizontal dotted lines indicate perfect performance (i.e., because target
displacements and durations were fully crossed, for each actual displacement the average of all actual
durations was 3000 ms, and for each actual duration the average of all actual displacements was 500
pixels). The ranges of the ordinates of (a) and (b) are proportionate with respect to the total range of actual
durations and displacements. Bottom: within-domain effects. (c) left: effect of actual line displacement on
estimated displacement. (d) right: effect of target duration on estimated duration. Error bars indicate
SEM.
D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593 583
3. Experiment 2: Growing lines, selective attention
In the first experiment, participants did not know until after each line was presented
whether they would need to estimate displacement or duration. They had
to attend to both the spatial and temporal dimensions of the stimulus. Experiment
2 addressed the possibility that cross-dimensional interference would diminish if participants
were given the opportunity to attend selectively to the trial-relevant stimulus
dimension, and to ignore the trial-irrelevant dimension.
3.1. Materials and procedure
Nine participants from the MIT community performed Experiment 2, in exchange
for payment. Stimulus materials were identical to those used in Experiment 1. The
procedure was also identical, with one exception. In Experiment 1, the word ‘‘ready’’
appeared for two seconds immediately preceding each line stimulus. In Experiment 2
(and all subsequent experiments reported here), the word ‘‘ready’’ was replaced
either by the word ‘‘Space’’ next to an ‘X’ icon, or by the word ‘‘Time’’ next to
an hourglass icon. These words and symbols indicated whether the subject would
need to estimate the displacement or the duration of the next line. Line stimuli,
prompts, and responses were exactly as in Experiment 1, thus all stimuli and
responses remained entirely non-linguistic.
3.2. Results and discussion
Results of Experiment 2 replicated those of Experiment 1 (cross-domain effects:
effect of displacement on duration estimation: y = 0.74x + 2474, r2
= .92, df = 7,
p < .001; effect of duration on displacement estimation: y = 0.003x + 464, r2
= .09,
df = 7, ns. Within-domain effects: Effect of target displacement on estimated displacement:
y = 0.85x + 49, r2
= .99, df = 7, p < .001; effect of target duration on estimated
duration: y = 0.77x + 526, r2
= .99, df = 7, p < .001). Participants were able
to disregard line duration when estimating displacement. By contrast, they were
unable to ignore line displacement, even when they were encouraged to selectively
attend to duration (Fig. 2b). The cross-dimensional effect of space on time estimation
in Experiment 1 was not caused by a task-specific demand for subjects to encode spatial
and temporal information simultaneously.
Response data collected for the trial-irrelevant dimension confirmed that participants
understood the task, and were not explicitly confusing displacement with duration
(i.e., participants were not giving a spatial response when they were supposed to
give a temporal response).
4. Experiment 3: Growing lines, temporal frame of reference
Experiments 3–5 addressed concerns that spatial information in the stimulus may
have been more stable or more salient than temporal information, and that differ-
584 D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593
ences in stability or salience produced the asymmetrical cross-dimensional interference
observed in Experiments 1 and 2. One concern was that participants may have
relied on spatial information to make temporal estimates because stimuli were situated
in a constant spatial frame of reference (i.e., the computer monitor). For Experiment
3, stimuli were also situated in a constant temporal frame of reference.
Temporal delay periods were introduced preceding and following line presentations,
which were proportional to the spatial gaps between the ends of the stimulus lines
and the edges of the monitor.
4.1. Materials and procedure
Nine participants from the MIT community performed Experiment 3, in exchange
for payment. Stimulus materials and procedures were identical to those used in
Experiment 2, with the following exception. In the previous experiments, the interval
between the disappearance of the ‘ready’ screen and the appearance of the response
prompt varied with stimulus duration. In the present experiment, this interval was
fixed at 6400 ms. Stimuli were preceded and followed by a delay period, which
was proportional to spatial gap separating the ends of the line stimuli from the left
and right edges of the monitor.
4.2. Results and discussion
The same pattern of cross-dimensional interference was found in Experiment 3 as
in the previous experiments (cross-domain effects: Effect of displacement on duration
estimation: y = 0.60x + 2604, r2
= .78, df = 7, p < .001; effect of duration on displacement
estimation: y = 0.0009x + 470, r2
= .03, df = 7, ns. Within-domain effects:
Effect of target displacement on estimated displacement: y = 0.80x + 73, r2
= .99,
df = 7, p < .001; effect of target duration on estimated duration: y = 0.68x + 866,
r2
= .99, df = 7, p < .001). The availability of a constant temporal frame of reference
did not abolish the asymmetric influence of distance on time estimation (Fig. 2c).
5. Experiment 4: Growing lines, concurrent tone
Would space still influence participants’ time estimates if stimulus duration were
indexed by something non-spatial? For Experiment 4, a tone of constant frequency
and amplitude accompanied each growing line. The tone began sounding when the line
started to grow across the screen, and stopped sounding when the line disappeared.
Thus, stimulus duration was made available to the participant in both the visual and
auditory modalities, but stimulus displacement was only available visually.
5.1. Materials and procedure
Sixteen participants from the MIT community performed Experiment 4, in
exchange for payment. Stimulus materials and procedures were identical to those
D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593 585
used in Experiment 2, with the following addition. A constant tone (260 Hz) accompanied
each growing line.
5.2. Results and discussion
Results of Experiment 4 replicated those of previous experiments (cross-domain
effects: Effect of displacement on duration estimation: y = 0.55x + 2647, r2
= .90,
df = 7, p < .001; effect of duration on displacement estimation: y = 0.002x + 450,
r2
= .10, df = 7, ns. Within-domain effects: Effect of target displacement on estimated
displacement: y = 0.72x + 96, r2
= .99, df = 7, p < .001; effect of target duration on
estimated duration: y = 0.84x + 414, r2
= .99, df = 7, p < .001). Displacement
strongly influenced participants’ duration estimates but not vice versa, even when temporal
information was provided via a different sensory modality from the spatial information
(Fig. 2d).
6. Experiment 5: Moving dot
Experiment 5 was designed to equate the mnemonic demands of the spatial and
temporal dimensions of the stimulus. Rather than viewing a growing line, subjects
viewed a dot that moved horizontally across the midline of the screen. In the previous
experiments, just before each growing line disappeared participants could see its
full spatial extent, from beginning to end, seemingly at a glance. By contrast, the spatial
extent of a moving dot’s path could never be seen all at once, rather it had to be
imagined: Participants had to retrieve the dot’s starting point from memory and
compare it to the ending point in order to judge the distance that the dot traveled.
The spatial and temporal dimensions of the dot stimulus had to be processed similarly
in this regard: whenever we compute the extent of a temporal interval we must
retrieve its starting point from memory.
6.1. Materials and procedure
Ten participants from the MIT community performed Experiment 5, in exchange
for payment. Stimulus materials and procedures were identical to those used in
Experiment 2, with one exception. Rather than viewing a growing line, subjects
viewed a dot (10 · 10 pixels) that moved horizontally across the midline of the
screen, from left to right.
6.2. Results and discussion
Results of Experiment 5 replicated those of previous experiments (cross-domain
effects: Effect of displacement on duration estimation: y = 0.50x + 2452, r2
= .82,
df = 7, p < .001; effect of duration on displacement estimation: y = À0.004x + 526,
r2
= .29, df = 7, ns. Within-domain effects: Effect of target displacement on estimated
displacement: y = 0.92x + 55, r2
= .99, df = 7, p < .001; effect of target duration on
586 D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593
estimated duration: y = 0.78x + 370, r2
= .99, df = 7, p < .001). As before, we found
a strong and asymmetric cross-dimensional effect of space on time (Fig. 2e), suggesting
that the perception of a long line was not necessary to lengthen participants’ time
judgments. Representations of spatial intervals that were never perceived at a glance
but only reconstructed from memory were sufficient to modulate duration estimates.
7. Experiment 6: Stationary lines
Experiments 1–5 used moving stimuli. Is motion necessary to produce confusion
between space and time, or would the asymmetric relationship between distance and
duration still be found if static stimuli were used? In Experiment 6, participants
viewed stationary lines and estimated either their displacement from end to end or
the amount of time they remained on the screen, as in previous experiments.
7.1. Materials and procedure
Nineteen participants from the MIT community performed Experiment 6, in
exchange for payment. Stimulus materials and procedures were identical to those used
in Experiment 2, with the following exception. Rather than viewing growing lines, participants
viewed stationary lines of various (spatial) lengths, which remained on the
screen for various durations, according to the parameters used in Experiment 2.
7.2. Results and discussion
Results showed the same pattern of cross-dimensional interference found in all
previous experiments (cross-domain effects: Effect of displacement on duration estimation:
y = 0.28x + 2769, r2
= .72, df = 7, p < .002; effect of duration on displacement
estimation: y = 0.001x + 447, r2
= .10, df = 7, ns. Within-domain effects:
Effect of target displacement on estimated displacement: y = 0.84x + 32, r2
= .99,
df = 7, p < .001; effect of target duration on estimated duration: y = 0.83x + 423,
r2
= .99, df = 7, p < .001). Duration estimates were strongly and asymmetrically
dependent on the spatial length of the stimulus (Fig. 2f). This finding rules out the
possibility that motion or speed was principally responsible for the results of the previous
experiments.
An additional meta-analysis was conducted to evaluate cross-dimensional interference
effects across all six experiments. A 2 · 6 mixed ANOVA with dimension
(effect of distance on time estimation, effect of time on distance estimation) as a
within-subjects factor and Experiment (Experiments 1–6) as a between-subjects factor
compared the slopes of all cross-dimensional interference effects. Results showed
a highly significant main effect of dimension, confirming that overall the slope of the
effect of distance on time estimation (M = 0.51, SE = 0.06) was greater than the
slope of the effect of time on distance estimation (M = 0.001, SE = 0.001;
F(1,66) = 83.73, p < .0001). There was no main effect of Experiment
(F(5,66) = 1.39, ns) and importantly, no dimension · experiment interaction
D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593 587
(F(5,66) = 1.39, ns), indicating that the magnitude of the space–time asymmetry did
not differ across experiments. To test for more subtle differences between experiments,
one-way ANOVAs compared the slopes of different cross-dimensional interference
effects, considered separately. No difference in slopes was found across
experiments in the effect of distance on time estimation (F(5,66) = 1.40, ns) or the
effect of time on distance estimation (F(5,66) = 0.57, ns). In summary, no significant
differences were found in the pattern of cross-dimensional interference across experiments.
The space–time asymmetry we report appears quite robust.
8. General discussion
When Piaget (1927/1969) investigated children’s reasoning about space and time,
he found that they often based their judgments of duration on their experience of
distance. For example, when asked to judge the relative duration of two trains traveling
along parallel tracks, children often reported (erroneously) that the train that
traveled the longer distance took the longer time. Piaget concluded that children
could not reliably distinguish the spatial and temporal components of events until
about age nine. Like many contemporary results in cognitive science, our findings
suggest that Piaget was right about the phenomenon he observed, but wrong about
the age at which children resolve their confusion: apparently MIT undergraduates
cannot reliably distinguish the spatial and temporal components of their experience,
either.
There are, in principle, three possible relationships between people’s mental representations
of space and time. First, the two domains could be symmetrically dependent.
John Locke (1689/1995) argued that space and time are mutually inextricable in
our minds, concluding that ‘‘expansion and duration do mutually embrace and
Fig. 2. Summary of cross-dimensional interference effects for Experiments 1–6. The effect of displacement on
duration estimation was significantly greater than the effect of duration on displacement estimation for all
experiments: (a) Growing lines: difference of correlations = 0.75; z = 3.24, p < .001. (b) Growing lines,
selective attention: difference of correlations = 0.66; z = 2.84, p < .002. (c) Growing lines, temporal frame of
reference: difference of correlations = 0.71; z = 2.09, p < 0.02. (d) Growing lines, concurrent tone: difference
of correlations = 0.63; z = 2.60, p < 0.005. (e) Moving dot: difference of correlations = 1.45; z = 3.69,
p < 0.001. 2f, Stationary lines: difference of correlations = 0.54; z = 1.62, p < 0.05. All p-values reflect onetailed
z-tests.
588 D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593
comprehend each other; every part of space being in every part of duration, and
every part of duration in every part of expansion’’ (p. 140). Alternatively, our ideas
of space and time could be independent. Any apparent relatedness could be due to
structural similarities between essentially unrelated domains (Murphy, 1996, 1997).
A third possibility is that time and space could be asymmetrically dependent. Representations
in one domain could be parasitic on representations in the other (Boroditsky,
2000; Lakoff & Johnson, 1980, 1999).
These three possible relationships predict three distinct patterns of cross-dimensional
interference between space and time in the present experiments. If spatial
and temporal representations are symmetrically dependent on one another, then
any cross-dimensional interference should be approximately symmetric: distance
should modulate duration estimates, and vice versa. Alternatively, if spatial and temporal
representations are independent, there should be no significant cross-dimensional
interference. However, if mental representations of time are asymmetrically
dependent on mental representations of space as suggested by patterns in language,
then we should observe an asymmetrical pattern in cross-dimensional interference:
distance should affect duration estimates more than duration affects distance estimates.
Results of all six experiments unequivocally support this third possibility,
demonstrating that the asymmetric relationship between space and time found in linguistic
metaphors is also found in our more basic non-linguistic representations of
distance and duration.
Over the past century of psychophysical experimentation on space and time judgments,
two effects have been demonstrated repeatedly: the Kappa effect and the Tau
effect (Benussi, 1913; Bill & Teft, 1969; Cohen, 1967; Cohen et al., 1954; Collyer,
1977; Helson, 1930; Jones & Huang, 1982; Price-Williams, 1954; Sarrazin et al.,
2004). In a typical experiment, three light bulbs were arranged in a row and flashed
in succession, forming two spatiotemporal intervals. Participants were asked to compare
either the spatial or temporal extents of the two intervals. Often, time judgments
were found to increase as a function of the spatial separation between stimuli (the
Kappa effect), and distance judgments were found to increase as a function of the
temporal separation between stimuli (the Tau effect). At first glance, these experiments
appear similar to those we report here; the Kappa effect seems consistent with
our results, but the Tau effect appears inconsistent with our findings. Yet, our findings
are easily reconciled with these classic findings, for two reasons. First, we
hypothesize an asymmetric relationship between space and time (not a unidirectional
relationship), our hypothesis can accommodate Tau-like effects of time on space
judgments. Second, a survey of the literature suggests that Tau and Kappa effects
emerge from implicit judgments of imputed velocity, and not from influences of the
spatial or temporal components of the stimuli, per se (Jones & Huang, 1982). We
elaborate both of these points below.
8.1. Asymmetrical vs. unidirectional effects
The relationship between time and space in linguistic metaphors is asymmetrical,
but not unidirectional. It is possible, in certain cases, to talk about space in terms of
D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593 589
time. For example, we might say ‘‘we’re only a few minutes from the subway’’ to indicate
that the subway is a short distance away. Alternatively, however, this distance
information could easily be conveyed using spatial language (e.g., we’re only a few
blocks from the subway). The space–time asymmetry in language is evident not only
in terms of how frequently we talk about one domain in terms of the other, but also
in terms of how obligatory these cross-domain mappings are. Whereas temporal
metaphors for space are optional, spatial metaphors for time are so that pervasive
they would be difficult for speakers to avoid using (Jackendoff, 1983; Pinker,
1997). Based on this asymmetry in language, we predicted asymmetrical crossdimensional
interference between time and space in non-linguistic judgments. This
prediction does not entail that time can never affect spatial judgments: rather, we
predicted that for judgments on different dimensions of the same stimuli, the effect
of space on time estimation should be greater than the effect of time on space estimation.
We did not observe any significant effect of time on distance estimation,
but such a finding would still be compatible with our hypothesis so long as we also
found a significantly greater effect of distance on time estimation. We show such a
significant difference between the cross-dimensional effects of space-on-time and
time-on-space in all six experiments.
8.2. The role of imputed velocity in Tau and Kappa effects
A close examination of the literature reveals that the effects we report may be
fundamentally different from Tau and Kappa effects. Although no theory on offer
can fully explain Tau and Kappa effects (Sarrazin et al., 2004), the theory that
appears to explain the majority of available data is the ‘‘imputed velocity hypothesis’’
(Jones & Huang, 1982), according to which Tau and Kappa effects arise
because ‘‘subjects impute uniform motion to discontinuous displays’’ (pp. 128;
see also Anderson, 1974; Cohen, 1967; Collyer, 1977; Price-Williams, 1954). In
most demonstrations of the Kappa effect (cf. Price-Williams, 1954) and in all
known demonstrations of the Tau effect, participants judged the relative spatial
or temporal extents of two or more successive intervals defined by discrete stimuli
(e.g., spatiotemporally separated flashes of light). Although there was no actual or
phenomenal motion in the stimuli, participants intuitively imputed motion at a
given velocity to the flash of light as it ‘traveled’ from one bulb to the next. They
produced errors when the imputed velocity of the stimulus changed between successive
intervals, violating their intuition that it would continue to ‘travel’
between points with uniform velocity. Although experimenters explicitly manipulated
the spatial and temporal extents of stimuli, the Tau and Kappa effects may
be appropriately considered to be effects of imputed velocity on judgments of both
time and space, rather than effects of time on space judgments or space on time
judgments, per se.
Is it possible that participants imputed illusory velocity to our stimuli? This
seems unlikely in Experiments 1–5 where the actual velocity was given by the
stimuli, and even more unlikely in Experiment 6 in which there was no motion
or speed information in stimuli at all – real or implied. Furthermore, there is
590 D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593
no reason to believe that participants’ expectations of constant velocity were violated,
given that all moving stimuli moved at a constant velocity, all of our stimuli
were spatiotemporally continuous, and none of our judgments required
comparisons between successive intervals. Thus, our stimuli contained none of
the ‘active ingredients’ of stimuli used to generate the Tau and Kappa effects.
If imputing velocity to discontinuous successive intervals accounts for Tau effects,
and our stimuli do not require participants to impute velocity or judge successive
intervals, then we should not expect to find Tau-like effects (which, indeed, we do
not). By the same token, we should not expect to find Kappa-like effects in our
studies. The effects of distance on time estimation that we report are importantly
different from imputed velocity-driven Kappa effects. Experiment 6 shows that the
same asymmetric relationship of space on time is found even when static lines are
used. This result converges with Cantor and Thomas’s (1977) study showing that
for very briefly presented static stimuli (30–70 ms), spatial information influenced
temporal judgments but not vice versa (i.e., subjective duration increased as a
function of stimulus area, but subjective area did not increase as a function of
stimulus duration).
It is noteworthy that space influences temporal judgments even for the simple,
brief temporal events we studied here which could in principle be mentally represented
qua time, as proposed by interval-timer and accumulator models (Ivry &
Richardson, 2002). Thinking about time metaphorically in terms of space may allow
us to go beyond these basic temporal representations. Mentally representing time as
a linear path may enable us to conceptualize more abstract temporal events that we
cannot experience directly through the senses (e.g., moving a meeting forward or
pushing a deadline back), as well as temporal events that we can never experience
at all (e.g., the remote past or the distant future). Metaphorical mappings from spatial
paths, which can be traveled both forward and backward, may give rise to temporal
constructs such as time-travel that only exist in our imagination.
9. Conclusions
Results of six experiments showed that mental representations of duration and
displacement are asymmetrically dependent on one another. Judgments of temporal
duration depended on information about spatial extent, but not the other way
around. This pattern was predicted by the asymmetry of space–time metaphors in
language. Although the brief durations used in these studies could in principle be
mentally represented qua time, people still incorporated irrelevant spatial information
into their temporal judgments. Moreover, these effects were obtained even in
purely non-linguistic tasks: tasks that did not involve any linguistic stimuli or
responses. These findings provide evidence that the metaphorical relationship
between space and time observed in language also exists in our more basic representations
of distance and duration, and suggest that our mental representations of
things we can never see or touch may be built, in part, out of representations of physical
experiences in perception and action.
D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593 591
Acknowledgments
The authors thank Herb Clark, Steven Pinker, and the citizens of Cognation for
discussion and comments on earlier drafts, and Shima Goswami, Jesse Greene, and
Webb Phillips for programming and data collection. This research was supported by
NSF CAREER Grant #0608514 to L. Boroditsky, and by an NSF Graduate Research
Fellowship and an NSF Dissertation Improvement Award to D. Casasanto.
These studies constituted part of D. Casasanto’s doctoral dissertation at MIT.
References
Alverson, H. (1994). Semantics and experience: Universal metaphors of time in english, mandarin, hindi, and
sesotho. Baltimore: Johns Hopkins University Press.
Anderson, N. H (1974). Algebraic models in perception. In E. C. Carterette & M. P. Friedman (Eds.).
Handbook of perception (Vol. 2). New York: Academic Press.
Benussi, V. (1913). Psychologie der Zeitaufassung. Heidelberg: Carl Winter’s Universitatsbuchhandlung.
Bill, J. C., & Teft, L. W. (1969). Space-time relations: Effects of time on perceived visual extent. Journal of
Experimental Psychology, 81(1), 196–199.
Boroditsky, L. (2000). Metaphoric structuring: Understanding time through spatial metaphors. Cognition,
75(1), 1–28.
Boroditsky, L. (2001). Does language shape thought? Mandarin and english speakers’ conceptions of time.
Cognitive Psychology, 43(1), 1–22.
Boroditsky, L., & Ramscar, M. (2002). The roles of body and mind in abstract thought. Psychological
Science, 13(2), 185–189.
Bowerman, M. (1983). Hidden meanings: The role of covert conceptual structures in children’s
development of language. In D. R. Rogers & J. A. Sloboda (Eds.), The acquisition of symbolic skills.
New York: Plenum.
Cantor, N., & Thomas, E. (1977). Control of attention in the processing of temporal and spatial
information in complex visual patterns. Journal of Experimental Psychology: Human Perception and
Performance, 3(2), 243–250.
Clark, H. H. (1973). Space, time, semantics and the child. In T. E. Moore (Ed.), Cognitive development and
the acquisition of language (pp. 27–63). New York: Academic Press.
Cohen, J. (1967). Psychological time in health and disease. Springfield: Charles C. Thomas.
Cohen, J., Hansel, C. E. M., & Sylvester, J. D. (1954). Interdependence of temporal and auditory
judgments. Nature, 174, 642–644.
Collyer, C. E. (1977). Discrimination of spatial and temporal intervals defined by three light flashes: Effects
of spacing on temporal judgments. Perception and Psychophysics, 21, 357–364.
Evans, V. (2004). The structure of time:Language,meaning and temporal cognition. Amsterdam: John
Benjamins.
Gruber, J. (1965). Studies in lexical relations. Cambridge: MIT.
Helson, H. (1930). The tau effect – an example of psychological relativity. Science, 71(1847), 536–537.
Ivry, R., & Richardson, T. (2002). Temporal control and coordination: The multiple timer model. Brain
and Cognition, 48, 117–132.
Jackendoff, R. (1983). Semantics and cognition. Cambridge: MIT PRess.
Jones, B., & Huang, Y. L. (1982). Space-time dependenacies in psychophysical judgment of extent and
duration: Algebraic models of the tau and kappa effects. Psychological Bulletin, 91(1), 128–141.
Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago Press.
Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh: The embodied mind and its challenge to western
thought. Chicago: University of Chicago Press.
Locke, J. (1689/1995). An essay concerning human understanding. Amherst: Promethius Books.
592 D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593
Murphy, G. (1996). On metaphoric representation. Cognition, 60, 173–204.
Murphy, G. (1997). Reasons to doubt the present evidence for metaphoric representation. Cognition, 62,
99–108.
Nu´n˜ez, R., & Sweetser, E. (2006). Looking ahead to the past: Convergent evidence from aymara language
and gesture in the crosslinguistic comparison of spatial construals of time. Cognitive Science, 30,
401–450.
Ornstein, R. (1969). On the experience of time. Hammondsworth: Penguin.
Piaget, J. (1927/1969). The child’s conception of time. New York: Ballantine Books.
Pinker, S. (1989). Learnability and cognition: The acquisition of argument structure. Cambridge, MA, US:
MIT Press.
Pinker, S. (1997). How the mind works. New York: Norton.
Price-Williams, D. R. (1954). The kappa effect. Nature, 173(4399), 363–364.
Sarrazin, J.-C., Giraudo, M.-D., Pailhous, J., & Bootsma, R. J. (2004). Dynamics of balancing space and
time in memory: Tau and kappa effects revisited. Journal of Experimental Psychology: Human
Perception and Performance, 30(3), 411–430.
Sweetser, E. (1991). From etymology to pragmatics: Metaphorical and cultural aspects of semantic structure.
Cambridge: Cambridge University Press.
Talmy, L. (1988). Force dynamics in language and cognition. Cognitive Science, 12, 49–100.
Torralbo, A., Santiago, J., & Lupia´n˜ez, J. (2006). Flexible conceptual projection of time onto spatial
frames of reference. Cognitive Science, 30, 749–757.
Traugott, E. C. (1978). On the expression of spatiotemporal relations in language. In J. H. Greenberg
(Ed.). Universals of human language: Word structure (Vol. 3, pp. 369–400). Stanford, CA: Stanford
University Press.
Tversky, B., Kugelmass, S., & Winter, A. (1991). Cross-cultural and developmental trends in graphic
productions. Cognitive Psychology, 23, 515–557.
D. Casasanto, L. Boroditsky / Cognition 106 (2008) 579–593 593