ArticlePDF Available

When “Light” and “Dark” Thoughts Become Light and Dark Responses: Affect Biases Brightness Judgments

Authors:

Abstract and Figures

Metaphors link positive affect to brightness and negative affect to darkness. Research has shown that such mappings are "alive" at encoding in that word-meaning evaluations are faster when font color matches prevailing metaphors (positive = bright; negative = dark). These results, however, involved reaction times, and there are reasons to think that evaluations would be unlikely to influence perceptual judgments, the current focus. Studies 1-3 establish that perceptual judgments were biased in a brighter direction following positive (vs. negative) evaluations, and Study 4 shows that such biases are automatic. The results significantly extend the metaphor representation perspective. Not only do evaluations activate metaphors, but such metaphoric mappings are sufficient to lead individuals to violate input from visual perception when judging an object's brightness.
Content may be subject to copyright.
When “Light” and “Dark” Thoughts Become Light and Dark Responses:
Affect Biases Brightness Judgments
Brian P. Meier
Gettysburg College
Michael D. Robinson
North Dakota State University
L. Elizabeth Crawford
University of Richmond
Whitney J. Ahlvers
North Dakota State University
Metaphors link positive affect to brightness and negative affect to darkness. Research has shown that
such mappings are “alive” at encoding in that word-meaning evaluations are faster when font color
matches prevailing metaphors (positive bright; negative dark). These results, however, involved
reaction times, and there are reasons to think that evaluations would be unlikely to influence perceptual
judgments, the current focus. Studies 1–3 establish that perceptual judgments were biased in a brighter
direction following positive (vs. negative) evaluations, and Study 4 shows that such biases are automatic.
The results significantly extend the metaphor representation perspective. Not only do evaluations activate
metaphors, but such metaphoric mappings are sufficient to lead individuals to violate input from visual
perception when judging an object’s brightness.
Keywords: metaphor, affect, evaluation, priming, judgment, perception
In one of his many speeches following the 9/11 attacks on the
Pentagon and the World Trade Center, President Bush said, “We
act now, because we must lift this dark threat from our age and
save generations to come” (Bush, 2001). Similarly, on January 6,
2003, Saddam Hussein spoke about his enemies by saying, “The
moon, the stars and the sun will, with Allah’s Grace, expose all the
schemes that they hide in the darkness of their minds and chests”
(“Saddam’s Hussein’s Speech,” 2003). Although these men are
obviously distinct in a number of ways (e.g., politics, religion,
geography), both of them used the term dark to connote negativity.
Like President Bush and Saddam Hussein, people frequently use
metaphors related to brightness to communicate evaluative infor-
mation (Gibbs, 1994; Ko¨vecses, 2000; Lakoff & Johnson, 1999).
There is general consensus that people use metaphor representation
processes, including those linked to affect (Ko¨vecses, 2000), very
frequently in their everyday discourse (Gibbs, 1994; Sweetser, 1990).
However, it is important to determine whether such metaphor pro-
cesses serve the primary purpose of communication or, alternatively,
whether such metaphor processes have a wider and more ubiquitous
influence on the representation of abstract concepts like affect. It is in
relation to the “online” (i.e., encoding-related) use of metaphor that
psycholinguists frequently differ (Gibbs, 1994). In other words, al-
though metaphor is a frequent occurrence in language, it could be
somewhat exclusive to it. Therefore, we follow Gibbs’s lead in
suggesting that more decisive evidence for the ubiquity and power of
metaphor representation requires cognitive paradigms sensitive to the
online use of metaphor-related mappings.
In theoretical support of such an online mapping, Lakoff and
Johnson (1999) argued that affective evaluations are represented,
implicitly and automatically, in terms of perceptual metaphor. Our
reading of this literature leads us to conclude that this perspective
makes at least three predictions that can be tested using social–
cognitive paradigms. One, evaluations should be faster when stim-
uli are presented in metaphor-consistent terms (e.g., a negative
stimulus should be encoded faster when in a black font color than
in a white font color). Two, evaluations should bias perceptual
judgments in a manner consistent with affective metaphor (e.g.,
perceptual judgments should be biased in a darker direction fol-
lowing negative evaluations). Three, encoding and judgment ef-
fects should occur in paradigms sensitive to automatic influences,
such as procedures associated with incidental exposure (Neely,
1991) or speeded responses (Meyer, Irwin, & Osman, 1988).
Affective Metaphor, Associative Processes, and
Perceptual Judgments
In prior work, we tested the first prediction mentioned above.
That is, we sought to determine whether evaluation latencies are
biased by irrelevant features of stimuli that happen to coincide
with affective metaphors. In one investigation (Meier, Robinson,
& Clore, 2004), we asked individuals to evaluate words with a
positive and negative meaning presented in a white or black font
color, with the valence and color manipulations independent of
Brian P. Meier, Department of Psychology, Gettysburg College; Michael D.
Robinson, Department of Psychology, North Dakota State University; L. Elizabeth
Crawford, Department of Psychology, University of Richmond; Whitney J.
Ahlvers, Department of Psychology, North Dakota State University.
Research in this article was supported by a dissertation fellowship to
Brian P. Meier from the North Dakota Experimental Program to Stimulate
Competitive Research/National Science Foundation and by Grant MH 068241
from the National Institute of Mental Health to Michael D. Robinson.
Correspondence concerning this article should be addressed to Brian P.
Meier, Department of Psychology, Gettysburg College, Gettysburg, PA
17325. E-mail: bmeier@gettysburg.edu
Emotion Copyright 2007 by the American Psychological Association
2007, Vol. 7, No. 2, 366 –376 1528-3542/07/$12.00 DOI: 10.1037/1528-3542.7.2.366
366
each other. Across six studies, we found that participants were
faster and more accurate in evaluating positive words if the words
were white (vs. black) in color, whereas participants were faster
and more accurate in evaluating negative words if the words were
black (vs. white) in color. Similar findings occurred when we
examined associations between affect and vertical position (Meier
& Robinson, 2004) and between affect and stimulus size (Meier,
Robinson, & Caven, 2006).
Given the robust nature of these previous findings, it is useful to
be clear what we have and have not shown in these studies. What
we have shown is that reaction time paradigms demonstrate robust
evidence for metaphor-related mappings at encoding involving
stimulus brightness (Meier et al., 2004), vertical position (Meier &
Robinson, 2004), and size (Meier et al., 2006). What we have not
shown is that metaphor processes bias perceptual judgments (e.g.,
brightness judgments), a central prediction of metaphor represen-
tation (Lakoff & Johnson, 1999).
Moreover, low-level perceptual properties might not be influ-
enced by top-down meaning-related cognitive systems (Marr,
1982; Pashler, 1998). It might be unlikely that affective metaphor,
activated by stimulus evaluations, would influence perceptual
judgments such as those related to brightness. Perceptual judg-
ments could thus pose a serious challenge to the perceptual pre-
dictions of Lakoff and Johnson’s (1999) thesis. Simply put, it may
be that white is white, gray is gray, and black is black, and this may
be true regardless of affect-linked metaphor.
In differentiating our previous findings from the purposes of
the present studies, an analogy to the semantic priming litera-
ture is useful (for a review, see Neely, 1991). This literature has
shown that semantically related primes facilitate (i.e., speed)
access to relevant targets, and such effects are thought to be
informative concerning the structure of semantic memory net-
works (e.g., McRae & Boisvert, 1998). In essence, our prior
findings have extended priming effects of this type to the
broader class of affective metaphors (e.g., positive is white;
Meier et al., 2004).
Yet, in the semantic priming literature, there is no presumption
that primes—for example, bread—alter perceptual judgment pro-
cesses related to targets—for example, butter. Making this sort of
leap to the realm of perceptual judgments would lead us to expect
that primes related to bread would bias individuals to perceive any
subsequently presented yellow, gooey substance as likely to be
butter. We are unaware of any such suggestions in the semantic
priming literature. This example merely reiterates the sharp dis-
tinction that should legitimately be made between spreading acti-
vation processes on the one hand and perceptual judgment pro-
cesses on the other.
Thus, no research that we know of has shown that affective
evaluations bias perceptual judgments related to metaphor, yet we
believe this is a central tenet of the metaphor representation
perspective (Lakoff & Johnson, 1999). In making a strong case for
affective metaphor, we predicted that affective evaluations would
lead individuals to judge the brightness of clearly visible shaded
patches in line with affective metaphor. If so, they should judge
targets to be lighter following positive evaluations and darker
following negative evaluations, in effect leading to biases against
the evidence provided by visual perception. The present four
studies examined this novel prediction.
Overview of Studies
In four studies, we examined the extent to which affective
evaluations would bias perceptual judgments concerning an ob-
ject’s brightness. In Study 1, we used a sequential paradigm to
determine whether residual activation from a word evaluation
would bias the judgment of a second object’s brightness. In Studies
2 and 3, we extended this analysis in a more nuanced manner by
having participants evaluate positive and negative words that var-
ied in their degree of font brightness. After each evaluation,
participants had to perform a brightness-matching task that re-
quired them to select a shaded square that matched the font
brightness of the word. In all of these studies, we predicted that
affective evaluations would bias brightness judgments in a
metaphor-consistent direction.
In Study 4, we sought to extend the results reported in Studies
1–3 by examining the automaticity of these affective influences on
perceptual judgments. Participants in this fourth study were pre-
sented with a positive or negative word followed by a square that
was lighter or darker in color. In this study, we used three proce-
dures designed to favor the automaticity of perceptual judgments.
One, there were no instructions to evaluate primes. Two, there was
a short stimulus onset asynchrony (SOA) between prime and
target. And three, the target judgment task was time limited in
nature. In short, Study 4 included procedures that should be of
specific relevance in understanding the automaticity of the present
effects. Despite the significant changes in the procedure, we again
expected affective evaluations to bias brightness judgments.
Finally, we should note that all studies used brightness judgment
paradigms that are routinely used in cognitive psychology. Spe-
cifically, both a brightness-matching task (i.e., matching the
brightness of two objects) and a two-alternative forced-choice task
(i.e., forcing participants to make one of two choices concerning an
object’s brightness) are standard ways to examine perceptual judg-
ments related to brightness (e.g., Alfonso-Reese, 2001; Floyd,
Dain, & Elliot, 2004; Maddox & Dodd, 2003; Wagner & Boyton,
1972). Thus, our dependent measures are consistent with those
used in the brightness perception literature.
Study 1
In Study 1, participants evaluated 50 positive and 50 negative
words presented one at a time in the center of the computer screen.
After each evaluation, we told participants that they would see one
of two squares that varied slightly in their level of brightness. In
actuality, there was only one square with a brightness level half-
way between white and black. Participants were asked to indicate
whether they were being shown the darker or lighter target square.
We hypothesized that the word evaluations would activate affec-
tive metaphor, which in turn would bias perceptual judgments in a
metaphor-consistent direction. For example, the square would be
judged as darker following negative primes. Findings of this type
would begin to extend the metaphor representation perspective to
perceptual judgments.
Method
Participants
Participants were 40 undergraduates (17 men and 23 women)
who volunteered for the study in exchange for extra course credit.
367
AFFECT AND PERCEPTUAL RESPONSES
Thirty-eight (95.00%) of the participants were Caucasian, 1 was
Asian, and 1 was American Indian. Participants’ average age was
19.65 years (SD 2.02).
Materials
Positive and negative words. Fifty positive (e.g., love) and 50
negative (e.g., rude) words were used in the study. These words
were identical to words used in previous studies (see Meier et al.,
2004, for a list of the words). The number of letters was similar for
the positive and negative words (F 1). We asked 8 participants
to rate the valence of the words using a 9-point scale (1
extremely negative;5 neutral;9 extremely positive). Using
word as the unit of analysis, the positive words (M 7.46, SD
0.86) were rated as significantly more positive than the negative
words (M 2.42, SD 0.70), F(1, 98) 1040.44, p .001,
p
2
.91. The absolute difference between the valence rating of
each word and the neutral midpoint was equal for the positive and
negative words (i.e., words were similarly extreme with respect to
the neutral midpoint; F 1). In Study 1, the words were presented
in a white font on a black background.
Shaded target square. We used an imaging program to create
a square that was 1.5 in. 1.5 in. (3.81 cm 3.81 cm) in size.
Using grayscale as the relevant chromatic dimension, we made the
square 50% grayscale (0% grayscale black; 100% grayscale
white). As shown in the top portion of Figure 1, the square
appeared to be light gray in color. The square was shown on a
black background.
Procedure
Participants wore headphones with a boom microphone. Partic-
ipants evaluated the 100 words one at a time. The words were
randomly selected and placed within the center of a computer
screen. Participants were told that when a word appeared, their
task was to determine as quickly and as accurately as possible
whether the word was positive (say “positive”) or negative (say
“negative”) in meaning. Immediately after each evaluation (no
preset delay), participants were told that one of two squares would
appear in the center of the screen. They were told that the squares
would vary slightly in color with one being a bit lighter in color
than the other.
In reality, only one square was shown on all trials. Although
some participants may have discerned or at least suspected that
there was only one shade of grayscale presented on all trials, we
regarded this as unlikely because we explicitly pointed to the
subtlety of the manipulation. In any case, suspicions along these
lines would likely work against the hypothesis rather than for it.
Participants were told to press the “1” button on a response box if
they believed they saw the darker target square and to press the “5”
button on a response box if they believed they saw the lighter
target square. After participants made a response, a 500-ms blank
screen was shown until the next word appeared. Participants eval-
uated each word once for a total of 100 trials.
To provide a rationale for the study, we presented participants
with the following cover story:
This experiment is concerned with your ability to evaluate words as
having either a negative or positive meaning while performing an
intervening task. A word will appear on the center of the screen. If the
word has a positive meaning, say the word “positive.” If the word has
a negative meaning, say the word “negative.” Immediately after your
evaluation, you will see one of two boxes. One box is light in color
while another box is dark in color. The difference in color between
these two boxes is small. However, most people are able to distinguish
between them. Your task is to tell us if the light or dark box is present.
If you see the light colored box, press the “5” button on the response
box. If you see the dark colored box, press the “1” button on the
response box.
Results
Primary Results
To determine whether word valence affected participants’ judg-
ments of square brightness, we coded responses by giving a dark
selection a 0 and a light selection a 1. For each participant, we
computed the mean percentage of light square responses for each
word valence. We then used a repeated-measures t test to compare
the mean percentage of light square responses by word valence.
Participants indicated that they saw the lighter of the squares
significantly more often after evaluating a positive word (M
63.05%, SD 20.59%) compared with a negative word (M
38.21%, SD 20.53%), t(39) 6.40, p .001, d 1.01.
Controlling for Word Frequency
Positive words tend to be more frequent than negative words
(Dixon, 1981; Matlin & Stang, 1978). Because we assumed that
the present effects pertained to affective metaphor rather than to
word frequency, we sought to perform an analysis in which we
were able to statistically control for word frequency effects. To
perform such an analysis, we sought to treat word (N 100) rather
than participant (N 40) as the unit of analysis. Along these lines,
Figure 1. The top row of the figure displays the target square used in
Study 1. The middle row displays the target squares used in Studies 2 and
3. The bottom row displays the target squares used in Study 4.
368
MEIER, ROBINSON, CRAWFORD, AND AHLVERS
we computed the average perceptual response (i.e., percentage of
light square responses) following each of the 100 words, averaged
across participants in the study. For each word, we also obtained a
word frequency norm from Kucˇera and Francis (1967). We then
conducted an analysis of covariance with word valence (positive
vs. negative) as the independent variable, perceptual responses as
the dependent variable, and word frequency values as the covari-
ate. Within this analysis, the main effect for word valence re-
mained significant, F(1, 97) 203.09, p .001,
p
2
.68. Thus,
the metaphor-consistent effects of evaluations on perceptual re-
sponses remained strong when controlling for word frequency.
Controlling for Object Color
Some of the words used here might correspond to objects that
are typically darker (e.g., mosquito) versus lighter (e.g., nurse)in
color. Given this fact, it is at least possible that the effect of
affective evaluations on perceptual responses might be due to
semantic associations to individual words that typically correspond
to darker or lighter objects. By contrast, we assumed that the
metaphor-consistent effect would remain significant when control-
ling for such semantic associations involving individual words.
To examine this prediction, we asked five naı¨ve judges to
determine whether the word, if it were the object in question,
would be colored in nature. For example, the word spider refers to
an object that is typically colored. By contrast, the word danger
does not refer to any particular colored object. To examine whether
the present effects would remain with potentially colored objects
excluded, we adopted a liberal criterion. Specifically, we removed
a word from the analysis if any of the five judges said the object
in question is typically colored in nature. Such criteria resulted in
the removal of 10 positive (baby, candy, clean, garden, heaven,
kiss, love, nurse, radiant, and sweet) and eight negative (cancer,
dead, devil, diseased, mosquito, poison, sour, and spider) words.
With these words removed, we performed a one-way analysis of
covariance with word as the unit of analysis (controlling for word
frequency). The effect of word valence on perceptual responses
remained significant, F(1, 79) 142.03, p .001,
p
2
.64.
Discussion
Study 1 revealed that the mere act of making an evaluation
biased the judgment of a subsequent object’s brightness. Consis-
tent with metaphors for affect and brightness, participants made a
light square judgment more often after a positive evaluation as
compared with a negative evaluation. Although the effect found in
Study 1 was rather strong (d 1.01), it is noteworthy that the task
used in Study 1 had no correct answer. Specifically, the target box
shown in Study 1 was always the same. It is possible that this
constrained paradigm, which presented no variations in target
color, may have artificially increased tendencies toward metaphor-
consistent perceptual responses. Indeed, there is reason to think
that the visual system is sensitive to gradations of brightness
(Goldstein, 1999) and therefore that a paradigm associated with
variations in brightness might offer more opportunities (relative to
Study 1) for veridical—and perhaps metaphor-unbiased—
perceptual judgments.
In addition to constraining responses, Study 1 used the terms
light and dark. These labels could introduce some response com-
patibility contributions to perceptual judgments because, according
to conventional metaphors, light means good and dark means bad
(Lakoff & Johnson, 1999). Therefore, it is somewhat of an open
question whether the results of Study 1 are due to metaphor
representation or response compatibility contributions to perfor-
mance.
To redress these possible limitations of Study 1, we used a
different paradigm in Study 2. Participants evaluated individual
words that varied in font color. Subsequent to evaluating the
words, they were asked to match the font color of the word to one
of five comparison standards. The words as well as target boxes
varied in brightness. Of further importance, there was always one
target box that exactly matched the brightness of the word in
question. Thus, there was an objectively correct answer as well as
plenty of time to perform this matching task. We also sought to
circumvent issues related to response compatibility. In Study 2,
participants’ responses were not based on light or dark. Partici-
pants were merely asked to match the color of the word to the color
of a target. Despite these procedural changes, which could indeed
reasonably eliminate influences related to affective metaphor, we
predicted that positive (relative to negative) evaluations would
result in brighter target-match responses.
Study 2
In Study 2, we used the same 100 positive and negative words
as in Study 1. However, in Study 2, we varied the font brightness
of the words by presenting them in one of three grayscale values.
After each word evaluation, participants were asked to match the
font color of the evaluated word to one of five shaded comparison
standards that varied along the brightness dimension.
Method
Participants
Participants were 25 undergraduates (9 men and 16 women)
who volunteered for the study in exchange for extra course credit.
All of the participants were Caucasian. Participants’ average age
was 19.52 years (SD 2.40).
Materials
Positive and negative words. We used the same words as in
Study 1. However, in Study 2, the words were presented in
different levels of font brightness. We used the same imaging
program from Study 1 to present the words in one of three
brightness values using a grayscale dimension (0% grayscale
black; 100% grayscale white). We randomly chose an approx-
imately equal number of positive and negative words to be pre-
sented in each brightness value (17 negative words and 16 positive
words were presented in 43.33% grayscale; 16 negative words and
17 positive words were presented in 50.00% grayscale; and 17
negative words and 17 positive words were presented in 56.67%
grayscale). By randomly assigning an approximately equal number
of words to each grayscale value, it was necessarily the case that
there was no association between word valence and font color. The
words were presented inside a white rectangle with approximately
0.25 in. (0.64 cm) of empty white space on each side of the word.
369
AFFECT AND PERCEPTUAL RESPONSES
The rectangles containing the words were presented on a blue
background.
Shaded squares. We used the same imaging program to create
five squares that were 1 in. 1 in. (2.54 cm 2.54 cm) in size.
The brightness value of the squares varied around the 50% gray-
scale midpoint using the same range of 0% grayscale (black) to
100% grayscale (white). As shown in the middle row of Figure 1,
the squares differed slightly in their appearance. From left to right,
grayscale values were 43.33%, 46.67%, 50%, 53.33%, and
56.67%, respectively. The squares were presented on a blue back-
ground.
Procedure
Participants wore headphones with a boom microphone. Partic-
ipants evaluated the 100 words one at a time. The words were
selected at random and presented in the center of a computer
screen. Participants were told that when a word appeared, their
task was to determine as quickly and as accurately as possible
whether the word was positive (say “positive”) or negative (say
“negative”) in meaning. Participants were told that the words
would vary slightly in color. Immediately after each evaluation (no
preset delay), five target squares appeared approximately 1 in.
(2.54 cm) below the word, with the word remaining on the screen.
This matching scale was identical to that presented in the middle
portion of Figure 1. Participants were told to place the mouse
cursor over the square that matched the color of the word and to
select the square by pressing the left mouse button. After partici-
pants made their selection, a 500-ms blank screen appeared until
the next word was presented. Important for present purposes,
during each trial the font brightness of each word was identical to
the brightness value of one of the five squares (i.e., the leftmost
square, the middle square, or the rightmost square). Participants
evaluated each word once for a total of 100 trials.
As in Study 1, to provide a rationale for Study 2, we presented
participants with the following cover story, which made no men-
tion of responding in a “light” or “dark” fashion (participants were
debriefed at the end of the study):
This experiment is concerned with your ability to evaluate words as
having either a negative or positive meaning while performing an
intervening task. A word will appear on the center of the screen. The
font color of this word will vary slightly from one of 5 possible
shades. You should first attend to the meaning of the word. If the word
has a positive meaning, say the word “positive.” If the word has a
negative meaning, say the word “negative.” Immediately after your
evaluation, you will see a color scale with different shades of dark-
ness. You should use the mouse to select the color shading that you
believe matches that of the word.
Results
Participants’ brightness-matching judgments were accurate on
an average of 30.48% (SD 10.40%) of trials (chance accuracy
20%). Thus, although the task was difficult, perceptual judgments
tended to be more accurate than chance, t(24) 4.90, p .001. To
determine whether word valence affected perceptual responses,
participants’ square responses were coded to resemble a Likert-
type rating scale. We coded the data in this manner because the
brightness of the squares varied in a linear and equal-interval
fashion. The responses were coded from 1 (darkest square) to 5
(lightest square). For each participant, we computed an average
perceptual response for each word valence. We then used a
repeated-measures t test to compare the average square response
by word valence. Participants were significantly more likely to
choose a lighter shaded square after evaluating a positive word
(M 3.59, SD .80) as compared with a negative word (M
2.60, SD .67), t(24) 3.64, p .001, d .73.
Control Analyses
As in Study 1, we sought to establish that the present effect was
independent of word frequency and possible associations with
stimulus color. We performed similar control analyses to those
performed in Study 1. With word frequency controlled, the effect
of word valence remained significant, F(1, 97) 88.73, p .001,
p
2
.48. In addition, the effect remained significant after delet
-
ing the 18 words that judges suggested might have some relation
to object color (and controlling for word frequency), F(1, 79)
66.56, p .001,
p
2
.46.
Analyses Related to Word Color
The positive and negative words were presented in three gray-
scales values. Two of these grayscale values were equivalent to the
lightest and darkest boxes in the target-matching task. Therefore, it
is somewhat plausible that our effects may only have occurred
when the word font color had a middle grayscale value relative to
an extreme one. However, an analysis including the variable of
word color revealed that this was not the case. Of specific impor-
tance to this analysis, there was no Word Valence Word Color
interaction in the repeated-measures analysis of variance
(ANOVA; F 1). The lack of an interaction indicates that the
main effect for valence was of equal strength for all three varia-
tions in word font color.
Discussion
Study 2 provides better evidence for the prediction that evalu-
ations activate metaphor, which in turn biases perceptual judg-
ments in a metaphor-consistent direction. That is, even within a
task in which (a) word stimuli varied in brightness, (b) the per-
ceptual judgment task had an objectively correct answer, (c) par-
ticipants’ accuracy rates were above chance, and (d) there was no
explicit reference to respond “light” or “dark,” affective metaphor
still biased perceptual responses and did so in a nonveridical
manner.
Study 3
Although Study 2 revealed that affective metaphor has implica-
tions for perceptual judgments, there is a possible alternative
explanation for the effect. In Study 2, the shaded squares were
arranged from darkest to lightest in a left-to-right manner. This
arrangement is consistent with most bipolar evaluation scales (for
a general overview of bipolar evaluation scales, see Cacioppo &
Berntson, 1994). Because of this often-used arrangement, how-
ever, it may be possible that the effect in Study 2 is not really due
to a perceptual response bias, but rather to a tendency to associate
left with negative and right with positive, somewhat independently
370
MEIER, ROBINSON, CRAWFORD, AND AHLVERS
of the specific response shades involved. To rule out this sort of
left–right bias as an explanation for our results and to replicate the
effect reported in Study 2, we thought it desirable to conduct
another study in which the arrangement of the target boxes was
reversed relative to their left–right coordinates in Study 2.
Method
Participants
Participants were 25 undergraduates (11 men and 14 women)
who volunteered for the study in exchange for extra course credit.
Twenty-one (84.00%) of the participants were Caucasian, 2 were
African American, 1 was Asian, and 1 was Hispanic. Participants’
average age was 20.02 years (SD 3.06).
Procedure
The materials and procedures for Study 3 were almost identical
to those used in Study 2. Participants evaluated the same 100
words one at a time. As in Study 2, after each word evaluation,
participants determined which shaded square matched the bright-
ness of the word. However, unlike Study 2, the target squares in
Study 3 were arranged such that the lightest square was to the left
and the darkest square was to the right.
Results
Participants’ brightness-matching judgments were accurate on
an average of 30.00% (SD 13.67%) of trials (vs. 20% for chance
accuracy). As in Study 2, participants performed at above-chance
levels, t(24) 3.66, p .001. We coded responses in the same
manner as in Study 2, meaning that lighter target responses were
given a higher numerical code. For each participant, we then
computed an average response for each word valence. We next
used a repeated-measures t test to compare the average square
response by word valence. As in Study 2, participants were sig-
nificantly more likely to choose a lighter shaded square after
evaluating positive words (M 3.61, SD 0.85) as compared
with negative words (M 2.58, SD 0.84), t(24) 3.47, p
.002, d .69.
Control Analysis
As in Studies 1 and 2, we next turned to analyses involving word
as the unit of analysis to control for word frequency and color
associations to individual words. The effect remained significant
after controlling for word frequency, F(1, 97) 109.01, p .001,
p
2
.53, and word– object color (while controlling for word
frequency), F(1, 79) 94.92, p .001,
p
2
.55.
Effects of Word Color
As in Study 2, we sought to examine whether the randomly
assigned word color mattered, with the specific idea that word
valence effects might be stronger at the 50% grayscale value. This
was not the case, as the Word Valence Word Color interaction
was not significant (F 1). Thus, the biasing effects of affect were
equally apparent regardless of the presented word color.
Discussion
The results of Study 3 reveal that those presented in Study 2
were not due to any implicit tendency to associate negative (pos-
itive) evaluations with a leftward (rightward) response. Specifi-
cally, in Study 3, we arranged the target boxes such that the
lightest of them was presented to the left of the computer screen,
whereas the darkest of them was presented to the right of the
computer screen. The findings from Study 3 replicate those of
Study 2 in suggesting that positive evaluations lead to lighter target
box judgments, regardless of the left–right arrangement of such
target boxes. We conclude that the nearly identical results of
Studies 2 and 3 are consistent with the idea that metaphor-related
mappings related to brightness are primary and quite a bit stronger
than those related to the left–right dimension (Ko¨vecses, 2000;
Lakoff & Johnson, 1999; McManus, 2002).
Study 4
The perceptual judgment data from Studies 1–3 point to a
perceptual response bias that is consistent with affective metaphor,
reliable, and large in magnitude, even when the judgment at hand
has an objectively correct response. Moreover, our perceptual
judgment tasks were consistent with those from the literature on
brightness perception (e.g., Floyd et al., 2004; Maddox & Dodd,
2003). Proponents of the metaphor representation perspective
(Gibbs, 1994; Lakoff & Johnson, 1999), however, view such
processes as automatic, and it is arguable that the tasks used in
Studies 1–3 may fall short of demonstrating the automaticity of the
present effects.
Therefore, we conducted a fourth study using three modifica-
tions of the basic paradigm. First, participants in Studies 1–3
explicitly evaluated the words, but many priming procedures (al-
though not all of them) involve incidental exposure without any
explicit meaning analysis (e.g., McRae & Boisvert, 1998). In
Study 4, we removed any instructions related to evaluating the
words. Second, some researchers have suggested that short-SOA
paradigms are useful for examining the automaticity of responses
(e.g., Neely, 1991). Therefore, we included such a short prime–
target delay (300 ms) in Study 4. Third, participants in Studies 1–3
were given as long as they needed to respond to the perceptual
target task. In Study 4, however, we thought it useful to institute a
time-limited perceptual response (400 ms) as a way to make a
further case for the automaticity of the present perceptual biases.
Method
Participants
Participants were 95 undergraduates (33 men and 49 women;
demographic data were not available for 13 participants) who
volunteered for the study in exchange for extra course credit.
Seventy-nine (83.16%) of the participants were Caucasian, 2 were
African American, and 1 participant indicated her race as “other.”
Participants’ average age was 19.87 years (SD 2.27).
Materials
Positive and negative words. We used the same positive and
negative words as in Studies 1–3. The words were presented in a
gray font (50% grayscale) on a black background.
371
AFFECT AND PERCEPTUAL RESPONSES
Shaded squares. We used the same imaging program to create
two squares that were 1.5 in. 1.5 in. (3.81 cm 3.81 cm) in size.
Using grayscale as the dimension of interest, we made one square
40% grayscale and one square 60% grayscale. As shown in the
bottom row of Figure 1, the squares were easily distinguishable as
being lighter and darker in color. The squares were presented on a
black background.
Procedure
Unlike prior studies, participants did not explicitly evaluate the
words. On each trial, a randomly chosen positive or negative word
was shown for 250 ms in the center of the computer screen.
Participants were not asked to evaluate these words, but there was
a vague reference to a memory test (which was not conducted) to
ensure some attention to the words. After the offset of the word, a
blank screen was shown for 50 ms. At this point, one of the two
squares (i.e., the lighter one or the darker one) was selected at
random and presented in the center of the screen, resulting in a
300-ms SOA from prime to target. Participants were given only
400 ms to determine if the lighter (press the “1” button) or darker
(press the “5” button) square was being shown. If participants did
not respond within this time-frame, the error message “TOO
SLOW!” appeared for 2 s. Responses within the 400-ms window
were followed by a blank screen for 1 s until the next word was
presented. The 100 words were shown in one randomized order,
and then a second one, for a total of 200 trials.
To provide a rationale for the study, we presented participants
with the following cover story:
This experiment will assess your ability to remember words while
performing a difficult intervening task. A word will appear on the
center of the screen for a fraction of a second. You should try to
remember this word for a later memory test. After a fraction of a
second, the word will be replaced by one of two colored squares. One
square will be lighter in color than the other square. In other words,
one square will be slightly dark and one square will be slightly light.
Your task is to decide, in less than half a second, whether the square
is the light one (press the 1 button) or the dark one (press the 5 button).
Results
Participants were successful in making a response in less than
400 ms on 88.91% (SD 5.26%) of the trials. Before analyzing
the nature of these responses, we first sought to determine whether
the probability of responding within the 400-ms window was
affected by word valence and target square color. We thus ana-
lyzed the percentage of responses (irrespective of which response
was made) within a 2 (word valence: positive and negative) 2
(square type: dark and light) repeated-measures ANOVA. The
main effect of word valence was not significant, F(1, 94) 1.47,
p .228. The main effect of square type was significant, F(1,
94) 11.49, p .001,
p
2
.11, which was due to the fact that
participants were more likely to respond in less than 400 ms if the
square was dark (M 89.80%, SD 5.58%) rather than light
(M 88.08%, SD 6.01%) in color. Finally, the interaction
between word valence and square type was significant, F(1, 94)
7.20, p .009,
p
2
.07.
As shown in Figure 2, participants were significantly more
likely to respond in less than 400 ms to the light square if the
preceding word had a positive (vs. negative) meaning, t(94)
2.40, p .019, d 0.25. By contrast, participants were not
significantly more likely to respond in less than 400 ms to the dark
square if the preceding word had a negative (vs. positive) meaning,
t(94) 1.10, p .275, d 0.10. Although the latter simple effect
was not significant, the interaction is consistent with prevailing
affective metaphors linking positive affect to brightness and neg-
ative affect to darkness.
We next turned our attention to the accuracy of responses made
within the 400-ms window. Overall, participants were accurate
78.31% (SD 12.10%) of the time. Although the task was
difficult, participants’ performance was above chance (50%),
t(94) 22.81, p .001. To determine whether accuracy rates
varied by valence and square type, we performed a 2 (word
valence: positive and negative) 2 (square type: dark and light)
repeated-measures ANOVA on participants’ accuracy rates. The
main effect of word valence was not significant, F(1, 94) 1.14,
p .288. The main effect of square type was significant, F(1,
94) 22.18, p .001,
p
2
.19, in that participants were more
accurate at detecting the light square (M 80.58%, SD 12.81%)
as compared with the dark square (M 75.84%, SD 13.31%).
Of most importance, the interaction between word valence and
square type was significant, F(1, 94) 15.25, p .001,
p
2
.14.
As shown in Figure 3, participants were more accurate in judging
the color of the light square if the preceding word had a positive
(vs. negative) meaning, t(94) 2.16, p .033, d 0.22. By
contrast, participants were more accurate in judging the color of
the dark square if the preceding word had a negative (vs. positive)
meaning, t(94) 3.49, p .001, d 0.36.
Figure 2. Mean percentage of responses less than 400 ms as a function of
word valence and square type (Study 4).
Figure 3. Mean accuracy rates as a function of word valence and square
type (Study 4).
372
MEIER, ROBINSON, CRAWFORD, AND AHLVERS
Control Analysis
Analyses (using word as the unit of analysis) revealed that the
accuracy effect remained significant after controlling for word
frequency, F(1, 97) 21.74, p .001,
p
2
.18 and after
deleting words associated with some mental images related to
stimulus color (and controlling for word frequency), F(1, 79)
10.30, p .002,
p
2
.12.
Effects of Repetition
Words were presented twice in Study 4 to boost cell sizes. The
repetition of words in the task gave us an additional way of
examining the automaticity of the present effects. If individuals
develop a strategy to respond according to affective metaphor, then
the effects of stimulus valence should be more apparent in the
second repetition relative to the first. Accordingly, we performed
a 2 (word valence: positive and negative) 2 (square type: dark
and light) 2 (repetition: first and second) repeated-measures
ANOVA on participants’ accuracy rates. The three-way interaction
was not significant, F(1, 94) 2.58, p .111, which means that
the effects were equally apparent in the first and second repetition
of words. Still, the interaction was close to marginal, and we
inspected the means. As shown in Table 1, affective evaluations
biased perceptual judgments to a greater extent during the first,
rather than second, repetition of words (
p
2
s .12 and .04 in the
first repetition and the second repetition, respectively). One inter-
pretation of this (albeit nonsignificant) tendency is that participants
became better at the perceptual task with practice. In addition, the
fact that the biases were at least a bit stronger during the first
repetition of words relative to the second adds to the idea that these
biasing effects are automatic in nature.
Discussion
From a theoretical standpoint, Lakoff and Johnson (1999) ar-
gued that metaphor representation processes are automatic. Study
4, using several procedures typically used to examine automatic
processes (Meyer et al., 1988; Neely, 1991), provided support for
this idea. For example, we gave Study 4 participants no goal of
evaluating the primes and also used a 300-ms SOA, as is somewhat
standard in the semantic priming literature. In addition, we time-
limited perceptual judgments. Despite the reductions in processing
time for both prime and target, and despite the fact that targets
were easily distinguishable (see the bottom row of Figure 1),
evaluations nevertheless biased perceptual judgments in a
metaphor-consistent direction.
Although we conclude that the affect–perception biases found
here are automatic, we also recognize changes in the manner in
which automaticity has been conceptualized over the years. In the
early investigations of automaticity, it was seen to be somewhat of
a monolithic entity, referring simultaneously to operations that
were quickly activated, unconscious, unintended, and ballistic in
nature (e.g., Shiffrin & Schneider, 1977). However, subsequent
research revealed that the qualities of automaticity are often inde-
pendent, so much so that few processes may be automatic accord-
ing to all four criteria (e.g., Bargh, 1994). The results of Study 4
suggest that our effects are quickly activated and unintended, two
of the four criteria of automaticity as it is classically defined. We
hasten to add that we have not shown that these effects are
unconscious or ballistic in nature, but we also note that these
criteria of automaticity have been controversial (Hollender, 1986;
Logan, 1994). Therefore, we conclude that our effects are auto-
matic in the sense of their quick activation and unintentional
nature.
General Discussion
In four studies, we sought to determine whether evaluative
processing would activate perceptual judgments of brightness, as
predicted by the metaphor representation viewpoint, but hitherto
an important gap in the literature (Rohrer, 2001). If evaluations are
guided by metaphor-linked representations in terms of stimulus
brightness, then making an evaluation should bias subsequent
perceptual responses in a metaphor-congruent (i.e., negative
dark; positive light) direction. All studies supported this basic
prediction.
In Study 1, participants were asked to determine whether they
were seeing a darker or lighter box following an evaluation. They
indicated that the same box was lighter in nature when it followed
a positive (vs. negative) evaluation. Studies 2 and 3 extended this
finding using a perceptual matching task that had an objectively
correct response. These studies made no explicit mention of “light”
or “dark” responses, and it was also the case that perceptual
judgments in the task were quite above chance. Regardless of these
factors, positive evaluations led to lighter matching responses,
whereas negative evaluations led to darker matching responses.
These results, perhaps even more strongly than those of Study 1,
support the hypothesis that affective metaphor leads individuals to
biased perceptual judgments.
Study 4 built on the prior studies while extending them in an
important way. Although the perceptual judgment task in Study 4
was easy, we time-limited responses to examine whether the
present metaphor activation phenomenon includes an automatic
component, as suggested by some metaphor representation view-
points (Gibbs, 1994; Lakoff & Johnson, 1999). Such effects did
appear to include an automatic component in that word valence
biased subsequent perceptual judgments under conditions associ-
ated with (a) the incidental exposure of primes, (b) a short SOA,
and (c) a brief response window. The data of Study 4 therefore
Table 1
Mean Accuracy Percentages by Valence, Color, and Repetition
(Study 4)
Word valence
Square color
Dark Light
MSDMSD
Repetition 1
Negative 75.73 15.39 78.20 15.78
Positive 70.84 17.76 80.89 16.10
Repetition 2
Negative 79.28 14.55 80.66 14.58
Positive 77.06 15.41 81.76 14.51
Note. Words in Study 4 were repeated twice. Mean accuracy is presented
as a function of word valence, target square color, and word repetition
variables (see text for further details).
373
AFFECT AND PERCEPTUAL RESPONSES
suggest that the present phenomenon is automatic, at least accord-
ing to the quick activation and unintentional criteria of automatic-
ity. Overall, we suggest that the results have important implica-
tions for the manner in which affect is represented.
Contributions Related to Metaphor Representation
In thinking about affective metaphor, it is easy to identify
expressions linking evaluation to stimulus brightness. For exam-
ple, Jesus Christ is the “light of the world,” whereas Satan is the
“prince of darkness.” Experimental work has sought to go beyond
the suggestion that metaphor is merely useful in speech and
language (Meier & Robinson, 2005). Specifically, this research has
sought to determine whether metaphor structures representation
itself, not just linguistic utterances.
Along these lines, the present findings are important in support-
ing three predictions of the metaphor representation perspective
(Gibbs, 1994; Lakoff & Johnson, 1999). One, although our previ-
ous studies had used the concurrent manipulation of affect and
physical cues (e.g., positive words in a white font color; Meier et
al., 2004), the present results reveal that such procedures are not
necessary to produce biases due to affective evaluation. Rather, the
mere act of making an evaluation can influence subsequent behav-
ior, even in a task that bears no obvious relation to the evaluation
task.
Two, the studies begin to address the question of whether
affective metaphor can influence perceptual judgments (Lakoff &
Johnson, 1999). We reiterate that our prior studies had not shown
such effects and that low-level perceptual judgments might be
uninfluenced by the activation of semantic meaning processes such
as those associated with evaluation (Marr, 1982; Pashler, 1998). In
this context, the large biasing effects for evaluation in the present
studies are particularly noteworthy, and we therefore conclude that
affective metaphor can lead individuals toward perceptual judg-
ments that violate data from visual sensation.
Three, the current studies help support the contention that met-
aphor mappings can be automatic in nature (Gibbs, 1994; Lakoff
& Johnson, 1999). Within all studies, there was no strategic
advantage in assuming that perceptual targets bore any relationship
to affective primes. Moreover, one could argue that metaphor-
linked perceptual responses actually placed the person at a disad-
vantage in that such effects led to systematic biases that violated
the evidence of sense receptors related to the brightness of shaded
targets. In light of these points, it is clear that the present results
represent a significant contribution to the literature. Of most im-
portance, they extend the metaphor representation perspective to
the realm of perceptual judgments, until now a neglected and
uncertain realm for this theoretical perspective (Rohrer, 2001).
Although our results do not speak to the issue of the origins of
affective metaphor, we tend to follow the suggestion that such
metaphors are built on prelinguistic sensorimotor processes
(Lakoff & Johnson, 1999; Meier & Robinson, 2005; Ramachand-
ran, 2004; Sweetser, 1990). According to such views, the pairing
of abstract concepts like affect and brightness may initially be a
sensory phenomenon in the developing human (e.g., darkness is
scary to young children). Such repeated pairings are thought to
give rise to linguistic metaphor, of both the affective and the
nonaffective varieties.
Indeed, just as affective metaphor likely relies on prelinguistic
processes, metaphor-related mappings should exert power even
when there is no specific reference to metaphor or linguistic
phrases typical of it. Clearly, the present results are consistent with
this framework in that there was no reference to metaphor or
linguistic phrases in any of the studies. Thus, metaphor-related
mappings are not merely language related in nature, but have a
deeper basis in the sensorimotor experiences that support, and
underlie, the use of metaphors common to all languages (Lakoff,
1987).
Metaphor Representation or Affective Priming?
It is worth considering the relation of the present findings to
those established within the affective priming literature (for re-
views, see Banse, 2003; Fazio, 2001). The bulk of this research has
investigated whether the brief, task-irrelevant presentation of a
positive or negative prime facilitates (e.g., speeds) the classifica-
tion of an affectively congruent target. A variety of studies have
shown that affectively congruent (i.e., good– good or bad– bad)
prime–target pairings facilitate classification of targets relative to
affectively incongruent (i.e., good– bad or bad– good) prime–target
pairings (Banse, 2003; Fazio, 2001). A second line of inquiry has
sought to determine whether affective primes influence judgments
concerning the affect of targets (Banse, 2003; Murphy & Zajonc,
1993). The affective priming literature has been centrally con-
cerned with the ubiquity and boundary conditions of the affective
priming phenomenon (Fazio, 2001; Storbeck & Robinson, 2004).
The present paradigms are somewhat compatible with the af-
fective priming phenomenon in that we, too, examined the auto-
matic consequences of affective primes. However, our paradigms
were also quite different in that the target task—related to percep-
tual judgments— bore no obvious relevance to affect. This was
especially true in Studies 2 and 3, which did not require partici-
pants to respond “light” or “dark,” and in Study 4, which made no
mention of “positive” or “negative” words. This feature of the
present tasks was quite different than target tasks typically exam-
ined in the affective priming literature (Fazio, 2001). In short, we
suggest that the relation between primes and targets in the present
studies was quite a bit broader than the affective associations
typically examined in the affective priming literature. For this
reason, we believe that our results are entirely novel in understand-
ing the biasing effects of affective primes on perceptual judgments.
Metaphor Representation or Semantic Associations?
The metaphor representation perspective may be viewed as a
“deep” theory of semantic associations common to multiple lan-
guages and cultures (Ko¨vecses, 2000; Lakoff & Johnson, 1999).
From this perspective, cultural–linguistic associations are not ar-
bitrary, but rather are strongly constrained by our physical bodies
(Gibbs, 2006). For example, it is not surprising that cultures
throughout the world associate positive affect with brightness
because human beings are diurnal creatures. Because of the nature
of our bodies and visual perception abilities, daytime simply
affords us better opportunities for foraging, perceiving present
dangers, and surviving the elements (Meier & Robinson, 2005).
Therefore, it is not surprising that we have so many linguistic
phrases linking positive affect to stimulus brightness.
374
MEIER, ROBINSON, CRAWFORD, AND AHLVERS
Could one arrive at the present predictions without reference to
the deep theory believed to underlie metaphor representation pro-
cesses? In a superficial way, perhaps, one could make these same
predictions on the basis of frequent valence– brightness pairings in
the English language (Ko¨vecses, 2000). However, making predic-
tions from this pure associative standpoint fails to recognize the
embodied constraints that determine our linguistic associations,
quite robustly across cultures. In other words, the metaphor rep-
resentation perspective (Lakoff & Johnson, 1999), with its specific
elaboration in terms of affective metaphor (Meier & Robinson,
2005), provides an explanatory framework that is simply absent
when one instead makes predictions in terms of associations bereft
of this deeper theoretical perspective.
In addition, we can state that the metaphor representation per-
spective leads to other predictions that would be difficult to handle
within a purely associative framework. As already noted, semantic
priming theories were designed to understand spreading activation
processes with semantic memory networks (Neely, 1991), but not
the perceptual judgment effects observed here. By contrast, the
metaphor representation perspective (Lakoff & Johnson, 1999),
like the embodiment perspective more generally (Gibbs, 2006),
contends that representation builds on perceptual processes and
should therefore prime judgments of the present type (e.g., Pecher,
Zeelenberg, & Barsalou, 2003; Richardson, Spivey, & Barsalou,
2003).
Broader Implications
Although the present results were specific to word evaluations
and their influence on perception, we also point out that our
findings have broader implications for the affect and emotion
literatures. Recent developments in this domain have suggested
that affect and emotion are embodied in nature (e.g., Niedenthal,
Barsalou, & Winkielman, 2005). Our findings are consistent with
this view in suggesting that emotions emerge, at least in part, from
perceptual representation processes rather than those that are
purely conceptual in nature. Moreover, we suggest that the meta-
phor representation perspective (Lakoff & Johnson, 1999), and its
elaboration in terms of affective metaphor (Meier & Robinson,
2005), should be productive in further developments in the liter-
ature.
Indeed, affective metaphor provides an explanatory framework
that can account for seemingly robust relations between mood
states and vertical selective attention. Meier and Robinson (2004)
demonstrated such a pattern of findings in the context of the
priming effects of affective evaluation on speed to recognize
higher versus lower spatial probes. More important to the broader
implications of these findings, there are a number of studies
showing that manipulations of, or naturally occurring, mood states
also appear to shift attention upward or downward in visual space
(Fisher, 1964; Meier & Robinson, 2006; Wapner, Werner, &
Kruss, 1957). For example, Wapner et al. showed that participants
receiving an F on a midterm, relative to those receiving an A on the
same midterm, subsequently bisected a vertical square lower in
visual space. Moreover, Meier and Robinson showed that de-
pressed and neurotic individuals were faster to respond to lower
visual probes versus higher visual probes, again consistent with the
idea that negative mood states appear to shift spatial attention
downward.
Although we know of no parallel findings related to brightness
perception, it is notable just how common it is to refer to depres-
sion in terms of dark perceptions, feelings, and cognitions (Meier
& Robinson, 2005). We are reluctant to state that “perceptual
therapy” would be of specific use in treating depression, but it is
worth noting that some important affect scholars have suggested
such interventions (e.g., Teasdale, 1993), and it is also clear that
light therapy can be useful in alleviating at least a certain class of
depression (Terman, Terman, & Ross, 1998).
Finally, it is important to note that the seemingly ubiquitous
associations between evaluation and stimulus brightness have po-
tential significance in understanding prejudice based on the dark-
ness of a person’s skin (e.g., as documented by Thompson &
Keith, 2001). Although affective metaphor may or may not play a
role in understanding this phenomenon, it represents one of the
obvious and intuitive extensions of the metaphor representation
perspective to very real social and emotional phenomena that may
benefit from this theoretical perspective. In sum, we suggest that
the future is bright for applications of metaphor representation to
important social and emotional phenomena.
Conclusion
The present studies sought to extend the metaphor representa-
tion perspective (Gibbs, 1994; Lakoff & Johnson, 1999) to the
real-time perceptual consequences of evaluating stimuli, whether
explicitly (Studies 1–3) or implicitly (Study 4). In all studies,
evaluative processing biased subsequent perceptual judgments in a
metaphor-consistent direction. That is, negative evaluations
primed dark perceptual judgments, whereas positive evaluations
primed light perceptual judgments. Because such effects were
replicated across different paradigms, none of which involved
metaphor-related phrases, the present results extend the metaphor
representation perspective to the realm of perceptual judgments.
We conclude by suggesting that these perceptual bias effects were
surprisingly large and robust. These not only validate the metaphor
representation perspective, but also begin to lay the groundwork
for important new directions in understanding emotion, perception,
and behavior.
References
Alfonso-Reese, L. A. (2001). Technique for estimating perceptual noise in
categorization tasks. Behavior Research Methods, Instruments, & Com-
puters, 33, 489 495.
Banse, R. (2003). Beyond verbal self-report: Priming methods in relation-
ship research. In J. Musch & K. C. Klauer (Eds.), The psychology of
evaluation: Affective processes in cognition and emotion (pp. 245–274).
Mahwah, NJ: Erlbaum.
Bargh, J. A. (1994). The Four Horsemen of automaticity: Awareness,
efficiency, intention, and control in social cognition. In R. S. Wyer, Jr.,
& T. K. Srull (Eds.), Handbook of social cognition (2nd ed., pp. 1–40).
Hillsdale, NJ: Erlbaum.
Bush, G. W. (2001). President Bush: No Nation can be Neutral in This
Conflict. Retrieved May 17, 2004, from http://www.whitehouse.gov/
news/releases/2001/11/20011106 –2.html
Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between attitudes
and evaluative space: A critical review, with emphasis on the separa-
bility of positive and negative substrates. Psychological Bulletin, 115,
401– 423.
Dixon, N. F. (1981). Preconscious processing. Chichester, England: Wiley.
375
AFFECT AND PERCEPTUAL RESPONSES
Fazio, R. H. (2001). On the automatic activation of associated evaluations:
An overview. Cognition & Emotion, 15, 115–141.
Fisher, S. (1964). Depressive affect and perception of up-down. Journal of
Psychiatric Research, 2, 25–30.
Floyd, R. A., Dain, S. J., & Elliott, R. T. (2004). Is the perception of
brightness different in poor readers? Vision Research, 44, 221–227.
Gibbs, R. W., Jr. (1994). The poetics of mind: Figurative thought, lan-
guage, and understanding. New York: Cambridge University Press.
Gibbs, R. W., Jr. (2006). Embodiment and cognitive science. New York:
Cambridge University Press.
Goldstein, E. B. (1999). Sensation and perception. Pacific Grove, CA:
Brooks/Cole.
Hollender, D. (1986). Semantic activation without conscious identification
in dichotic listening, parafoveal vision, and visual masking: A survey
and appraisal. Behavioral and Brain Sciences, 9, 1–66.
Ko¨vecses, Z. (2000). Metaphor and emotion: Language, culture, and body
in human feeling. New York: Cambridge University Press.
Kucˇera, H., & Francis, W. N. (1967). Computational analysis of present-
day American English. Providence: Brown University Press.
Lakoff, G. (1987). Women, fire, and dangerous things: What categories
reveal about the mind. Chicago: University of Chicago Press.
Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh: The embodied
mind and its challenges to western thought. New York: Basic Books.
Logan, G. D. (1994). On the ability to inhibit thought and action: A users’
guide to the stop signal paradigm. In D. Dagenbach & T. H. Carr (Eds.),
Inhibitory processes in attention, memory, and language. San Diego,
CA: Academic Press.
Maddox, W. T., & Dodd, J. L. (2003). Separating perceptual and decisional
attention processes in the identification and categorization of integral-
dimension stimuli. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 29, 467–480.
Marr, D. (1982). Vision: A computational investigation into the human
representation and processing of visual information. New York: Free-
man.
Matlin, M., & Stang, D. (1978). The Pollyanna principle: Selectivity in
language, memory, and thought. Cambridge, MA: Schenkman.
McManus, C. (2002). Right hand, left hand: The origins of asymmetry in
brains, bodies, atoms, and cultures. Cambridge, MA: Harvard Univer-
sity Press.
McRae, K., & Boisvert, S. (1998). Automatic semantic similarity priming.
Journal of Experimental Psychology: Learning, Memory, and Cogni-
tion, 24, 558 –572.
Meier, B. P., & Robinson, M. D. (2004). Why the sunny side is up:
Associations between affect and vertical position. Psychological Sci-
ence, 15, 243–247.
Meier, B. P., & Robinson, M. D. (2005). The metaphorical representation
of affect. Metaphor and Symbol, 21, 239–257.
Meier, B. P., & Robinson, M. D. (2006). Does “feeling down” mean seeing
down? Depressive symptoms and vertical selective attention. Journal of
Research in Personality, 40, 451–461.
Meier, B. P., Robinson, M. D., & Caven, A. J. (2006). Why a Big Mac is
a good Mac: Associations between affect and size. Manuscript submitted
for publication.
Meier, B. P., Robinson, M. D., & Clore, G. L. (2004). Why good guys wear
white: Automatic inferences about stimulus valence based on brightness.
Psychological Science, 15, 82– 87.
Meyer, D. E., Irwin, D. E., & Osman, A. M. (1988). The dynamics of
cognition and action: Mental processes inferred from speed/accuracy
decomposition. Psychological Review, 95, 183–237.
Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness:
Affective priming with optimal and suboptimal stimulus exposures.
Journal of Personality and Social Psychology, 64, 723–739.
Neely, J. H. (1991). Semantic priming effects in visual word recognition:
A selective review of current findings and theories. In D. Besner &
G. W. Humphreys (Eds.), Basic processes in reading (pp. 264 –336).
Hillsdale, NJ: Erlbaum.
Niedenthal, P. M., Barsalou, P. M., & Winkielman, P. (2005). Embodiment
in attitudes, social perception, and emotion. Personality & Social Psy-
chology Review, 9, 184 –211.
Pashler, H. (1998). The psychology of attention. Cambridge, MA: MIT
Press.
Pecher, D., Zeelenberg, R., & Barsalou, L. W. (2003). Verifying different-
modality properties for concepts produces switching costs. Psychologi-
cal Science, 14, 119 –124.
Ramachandran, V. S. (2004). A brief tour of human consciousness: From
imposter poodles to purple numbers. New York: Pi Press.
Richardson, D. C., Spivey, M. J., & Barsalou, L. W. (2003). Spatial
representations activated during real-time comprehension of verbs. Cog-
nitive Science, 27, 767–780.
Rohrer, T. (2001). The cognitive science of metaphor from philosophy to
neuroscience. Theoria et Historia Scientarium, 6, 27–42.
Saddam Hussein’s speech. (2003, January 6). Guardian Unlimited. Re-
trieved May 20, 2005, from http://www.guardian.co.uk/Iraq/Story/
0,2763,869642,00.html
Shiffrin, R., & Schneider, W. (1977). Controlled and automatic human
information processing: II. Perceptual learning, automatic attending, and
a general theory. Psychological Review, 84, 127–190.
Storbeck, J., & Robinson, M. D. (2004). Preferences and inferences in
encoding visual objects: A systematic comparison of semantic and
affective priming. Personality and Social Psychology Bulletin, 30, 81–
93.
Sweetser, E. (1990). From etymology to pragmatics. Cambridge, England:
Cambridge University Press.
Teasdale, J. D. (1993). Emotion and two kinds of meaning: Cognitive
therapy and applied cognitive science. Behaviour Research and Ther-
apy, 31, 339 –354.
Terman, M., Terman, J. S., & Ross, D. C. (1998). A controlled trial of
timed bright light and negative air ionization for treatment of winter
depression. Archives of General Psychiatry, 55, 875–882.
Thompson, M. S., & Keith, V. M. (2001). The blacker the berry: Gender,
skin tone, self-esteem, and self-efficacy. Gender & Society, 15, 336
357.
Wagner, G., & Boynton, R. M. (1972). Comparison of four methods of
heterochromatic photometry. Journal of the Optical Society of America,
62, 1508 –1515.
Wapner, S., Werner, H., & Krus, D. M. (1957). The effect of success and
failure on space localization. Journal of Personality, 25, 752–756.
Received February 22, 2006
Revision received August 22, 2006
Accepted November 8, 2006
376
MEIER, ROBINSON, CRAWFORD, AND AHLVERS
... In beliefs underlying several religions, black signifies sins and hell, whereas white represents God(s) and holiness. Empirical research corroborates such anecdotal evidence, showing that people tend to automatically match negative words and concepts with black, but they connect positive words and concepts with white (Meier et al. 2004(Meier et al. , 2007Sherman and Clore 2009). In sports, research shows that umpires perceive players who wear black (vs. ...
... We anticipate this aspect in line with prior research that shows that color effects, color meanings, and the influence of contextual cues on these meanings are generally automatic (Elliot et al. 2007(Elliot et al. , 2009Elliot and Maier 2012;Kareklas et al. 2019;Labrecque et al. 2013;Sundar and Kellaris 2017). Besides, prior research also indicates that logo effects can be automatic (Fitzsimons, Chartrand, and Fitzsimons 2008;Laran, Dalton, and Andrade 2011;Rahinel and Nelson 2016) and that people automatically pair negative meanings with black and positive meanings with white (Meier et al. 2004(Meier et al. , 2007Sherman and Clore 2009). ...
Article
Why, how, and when can logos with a blue positive space communicate competence versus sadness? Why, how, and when might logos with a red positive space evoke impressions of liveliness versus aggressiveness? As the current research establishes, a black background strengthens the negative meanings associated with the hue of a logo’s positive space and weakens its positive meanings. Conversely, a white background strengthens its positive meanings and weakens its negative meanings. These automatic effects occur because the hue of the positive space interacts with the color of the negative space to determine whether logos communicate positive or negative brand impressions more vividly. These effects are broadly applicable to both well-known and unknown brands, yet they are attenuated for meaningful logos and filled-frame logos. With these novel findings, this article identifies specific factors that can alter the meanings of logo hues, provides a theoretical lens for understanding the interplay of the background color and the hue of the positive space, and offers guidelines for crafting effective logos. This article also reveals which brands can benefit most from conveying negative impressions through their logos: Logos with a black (white) background enhance evaluations of brands that possess negatively (positively) valenced personality traits.
... For example, a high position in the visual-vertical dimension is associated with rationality and higher power, whereas a low position is associated with emotions and lower power (Cian et al., 2015;Sundar & Noseworthy, 2014). Brightness is associated with virtue, health, and optimism, whereas darkness is associated with evil, danger, and depression (Landau et al., 2010;Meier et al., 2007). Romantic visual stimuli psychologically activate metaphorical associations with sweetness (Yang et al., 2019). ...
Article
Full-text available
Consumers increasingly value eco‐friendly attributes of products. Packaging serves as a medium through which consumers perceive product attributes and products convey information. However, research on the impact of packaging visual complexity on a product's perceived eco‐friendliness has not been thoroughly examined. Four studies were conducted based on visual complexity metaphor theory to examine the impact of packaging visual complexity on the eco‐friendliness perception of products. The study also investigated the mediating role of naturalness and durability perception as well as the moderating effect of external environmental threats. The results indicate that the visual complexity of packaging directly or indirectly (through perceptions of naturalness and durability) affects the eco‐friendliness perception of products. Furthermore, under high environmental threats, consumers exhibit a stronger eco‐friendly perception and purchase intention towards products with simple packaging. Eco‐friendliness perception mediates the relationship between packaging visual complexity and purchase intentions. This study contributes to the literature on visual complexity metaphors and eco‐friendly packaging and complements our knowledge of how visual packaging characteristics influence the perception of product attributes. The results provide valuable insights for marketing managers on how to use the visual complexity of packaging design to effectively convey product information.
... However, the Same-target tone was actually identical to the baseline, meaning that in those trials listeners were forced to indicate a direction of magnitude change despite there being none. This procedure was similar to that adapted by Wallmark et al. (2021) from Meier et al. (2007). Crossmodal and exploratory amodal tasks. ...
Article
Full-text available
Brightness is among the most studied aspects of timbre perception. Psychoacoustically, sounds described as “bright” versus “dark” typically exhibit a high versus low frequency emphasis in the spectrum. However, relatively little is known about the neurocognitive mechanisms that facilitate these metaphors we listen with . Do they originate in universal magnitude representations common to more than one sensory modality? Triangulating three different interaction paradigms, we investigated using speeded classification whether intramodal, crossmodal, and amodal interference occurs when timbral brightness, as modeled by the centroid of the spectral envelope, and pitch height/visual brightness/numerical value processing are semantically congruent and incongruent. In four online experiments varying in priming strategy, onset timing, and response deadline, 189 total participants were presented with a baseline stimulus (a pitch, gray square, or numeral) then asked to quickly identify a target stimulus that is higher/lower, brighter/darker, or greater/less than the baseline after being primed with a bright or dark synthetic harmonic tone. Results suggest that timbral brightness modulates the perception of pitch and possibly visual brightness, but not numerical value. Semantically incongruent pitch height-timbral brightness shifts produced significantly slower reaction time (RT) and higher error compared to congruent pairs. In the visual task, incongruent pairings of gray squares and tones elicited slower RTs than congruent pairings (in two experiments). No interference was observed in the number comparison task. These findings shed light on the embodied and multimodal nature of experiencing timbre.
... When it comes to colour-emotion associations globally, previous studies linked red to love and anger, pink to love, yellow to joy, brown to disgust, and black with sadness (Fugate & Franco, 2019;Jonauskaite et al., 2020;Jonauskaite, Wicker, et al., 2019;Kaya & Epps, 2004). Darker colours were associated with more negative emotions and lighter colours with more positive emotions (Adams & Osgood, 1973;Meier et al., 2007;Specker et al., 2018). Red and black were arousing and powerful while blue and green were calming (Adams & Osgood, 1973;Kaya & Epps, 2004;Valdez & Mehrabian, 1994). ...
Article
Full-text available
Red with anger or green with envy-such metaphors link colours and emotions. While such colour metaphors vary across languages, conceptual associations between colours and emotions have many cross-cultural similarities. Here, we took published data from 8615 participants (2172 men) coming from 37 nations (i.e. and United States) and analysed Lithuanian (n = 217) associations between colour terms and emotion concepts. Lithuanians had many associations, the most frequent being red-love, yellow-amusement, yellow-joy, and black-sadness (all endorsed by > 60% of participants). While Lithuanians associated more emotions with colours than the other participants, the Lithuanian pattern of these associations was highly similar to the global pattern (r = .92). When compared to each other nation individually, colour-emotion association pattern similarities ranged between .65 and .89. Lithuanian patterns were the most similar to the Russian and the least similar to the Egyptian ones. Crucially, such similarities could be predicted by linguistic but not geographic distances. Nations speaking languages linguistically closer to Lithuanian also displayed more similar colour-emotion association patterns. These results support universality of colour-emotion associations and point to small but meaningful cultural differences (e.g., red represented love more strongly than anger for Lithuanians but not globally). Future studies should look whether colours can modulate emotions, or whether such associations are purely abstract. [Lithuanian version] Pabalti iš pykčio ar pažaliuoti iš pavydo-tai metaforos, siejančios spalvas ir emocijas. Nors tokios spalvų metaforos įvairiose kalbose skiriasi, psichologinės spalvų ir emocijų asociacijos turi daug tarpkultūrinių panašumų. Čia pasinaudojome anksčiau skelbtais 8 615 dalyvių (2 172 vyrų), kilusių iš 37 šalių (t. y. Austrijos, Azerbaidžano, Belgijos, Egipto, Estijos, Filipinų, Graikijos, Gruzijos, Indijos, Irano, Ispanijos, Italijos, Izraelio, Japonijos, Jung-tinės Karalystės, Jungtinių Amerikos Valstijų, Kinijos, Kolumbijos, Kroatijos, Kipro, Latvijos, Lenkijos, Lietuvos, Meksikos, Naujosios Zelandijos, Nigerijos, Nyderlandų, Norvegijos, Prancūzijos, Rusijos, Saudo Arabijos, Serbijos, Suomijos, Švedijos, Šveicarijos, Ukrainos ir Vokietijos), duomenimis ir analizavome, kokias asociacijas lietuviams (n = 217) kelia spalvų terminai ir emocijų konceptai. Lietuviai turėjo daug asociacijų, iš kurių dažniausios buvo rau-dona-meilė, geltona-linksmumas, geltona-džiaugsmas ir juoda-liūdesys (visas šias asociacijas pasirinko daugiau kaip 60 % dalyvių). Nors lietuviai su spalvomis siejo daugiau emocijų nei kiti dalyviai, lietuviškas šių asociacijų reljefas buvo labai panašus į pasaulinį (r = 0,92). Lyginant su kiekviena kita tauta atskirai, spalvų ir emocijų asociacijų reljefų panašumas svyravo nuo 0,65 iki 0,89. Lietuvių reljefas buvo panašiausias į rusų, o mažiausiai panašus – į egiptiečių. Šiuos panašumus buvo galima nuspėti pagal kalbinius, bet ne geografinius atstumus. Tautos, kalbančios lingvistiškai artimesnėmis kalbomis, taip pat pasižymėjo panašesnėmis spalvų ir emocijų asociacijomis. Šie rezultatai patvirtina spalvų ir emocijų asociacijų universalumą ir rodo nedidelius, bet reikšmingus kultūrinius skirtumus (pvz., raudona spalva lietuviams reiškė meilę labiau nei pyktį). Ateityje reikėtų išsiaiškinti, ar spalvos gali paveikti emocijas, ar visgi tokios asociacijos yra visiškai abstrakčios.
Article
Full-text available
Colour is an integral part of natural and constructed environments. For many, it also has an aesthetic appeal, with some colours being more pleasant than others. Moreover, humans seem to systematically and reliably associate colours with emotions, such as yellow with joy, black with sadness, light colours with positive and dark colours with negative emotions. To systematise such colour–emotion correspondences, we identified 132 relevant peer-reviewed articles published in English between 1895 and 2022. These articles covered a total of 42,266 participants from 64 different countries. We found that all basic colour categories had systematic correspondences with affective dimensions (valence, arousal, power) as well as with discrete affective terms (e.g., love, happy, sad, bored). Most correspondences were many-to-many, with systematic effects driven by lightness, saturation, and hue (‘colour temperature’). More specifically, (i) LIGHT and DARK colours were associated with positive and negative emotions, respectively; (ii) RED with empowering, high arousal positive and negative emotions; (iii) YELLOW and ORANGE with positive, high arousal emotions; (iv) BLUE, GREEN, GREEN–BLUE, and WHITE with positive, low arousal emotions; (v) PINK with positive emotions; (vi) PURPLE with empowering emotions; (vii) GREY with negative, low arousal emotions; and (viii) BLACK with negative, high arousal emotions. Shared communication needs might explain these consistencies across studies, making colour an excellent medium for communication of emotion. As most colour–emotion correspondences were tested on an abstract level (i.e., associations), it remains to be seen whether such correspondences translate to the impact of colour on experienced emotions and specific contexts.
Article
The integration of music sounds with concurrent visual scenes or objects is a occurrence in our daily lives, attracting the attention of researchers investigating how music influences our perceptions of simultaneous visuals. This study specifically investigates the role of musical pitch height in shaping our judgments of visual brightness during concurrent music–visual events. Participants were presented with pitch-modified versions of a wide range of emotional music pieces alongside various visual stimuli. The results demonstrate that lower-pitched music tends to elicit darker judgments of visual objects than higher-pitched music, when the actual visual brightness level remains constant. These findings suggest the influence of musical pitch height on introducing biases in our evaluations of visual brightness within the context of concurrent music-visual experiences, contributing to the advancement of our theoretical understanding of the complex audiovisual integration involving music in our everyday lives.
Article
Full-text available
Evaluative processes refer to the operations by which organisms discriminate threatening from nurturant environments. Low activation of positive and negative evaluative processes by a stimulus reflects neutrality, whereas high activation of such processes reflects maximal conflict. Attitudes, an important class of manifestations of evaluative processes, have traditionally been conceptualized as falling along a bipolar dimension, and the positive and negative evaluative processes underlying attitudes have been conceptualized as being reciprocally activated, making the bipolar rating scale the measure of choice. Research is reviewed suggesting that this bipolar dimension is insufficient to portray comprehensively positive and negative evaluative processes and that the question is not whether such processes are reciprocally activated but under what conditions they are reciprocally, nonreciprocally, or independently activated. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Full-text available
Using data from the National Survey of Black Americans, this study examines the way in which gender socially constructs the importance of skin tone for evaluations of self-worth and self-competence. Skin tone has negative effects on both self-esteem and self-efficacy but operates in different domains of the self for men and for women. Skin color is an important predictor of self-esteem for Black women but not Black men. And color predicts self-efficacy for Black men but not Black women. This pattern conforms to traditional gendered expectations of masculinity and femininity. Moreover, there are conditions of success that allow women to escape the effects of colorism. The impact of skin tone on self-esteem was much weaker for women from higher social class. Those who had lower self-esteem scores were dark-skinned women from working classes and dark-skinned women who were judged unattractive.
Article
In this paper I review some of the theoretical issues surrounding metaphor, and trace them through the context of the cognitive neuroscience debate. Metaphor, like all figurative language, has typically been explained as a secondary linguistic process which is a function taking place on literal language. However this explanation does not fit well with some of the recent work on right hemisphere processing of language or recent cognitive studies, both of which suggest that the figurative and literal language are processed simultaneously and share much substructure. In seeking ways to operationalize the Lakoff and Johnson view of conceptual metaphor as a constitutive cognitive phenomenon, I begin to spell out what kinds of theoretical predictions the Lakoff Johnson model would make on the neurophysiological levels af cognitive investigation. I conclude by offering some thoughts on new directions of research using these methods, and by reassessing the philosophical basis of these matters.
Book
automated social cognitive processes categorize, evaluate, and impute the meanings of behavior and other social information, and this input is then ready for use by conscious and controlled judgment and decision processes / review . . . the literature on automaticity in social cognition] / discuss the research in terms of its relevance for the] issues of awareness, intentionality, efficiency, and control (PsycINFO Database Record (c) 2012 APA, all rights reserved)(chapter)
Article
Abstract Affect is a somewhat abstract concept that is frequently linked to physical metaphor. For example, good is often depicted as light (rather than dark), up (rather than down), and moving forward (rather than backward). The purpose of our studies was to examine whether the association between stimulus brightness and affect is optional or obligatory. In a series of three studies, participants categorized words as negative or positive. The valence of the words and the brightness of the letters were varied orthogonally. In Studies 1, 2, and 3, we found that categorization was inhibited when there was a mismatch between stimulus brightness (e.g., light) and word valence (e.g., negative). Studies 4 and 5 reveal boundary conditions for the effect. The studies suggest that, when making evaluations, people automatically assume that bright objects are good, whereas dark objects are bad.
Article
Philosophers and psycholinguists have argued that abstract concepts like affect are represented via the mechanism of metaphor. This review investigates this contention, specifically within the context of social-cognition and clinical psychology research that has studied the link between affect and brightness, vertical position, and distance between the self and an object. The review will be particularly concerned with automatic and incidental linkages between affect and perception and their relevance for a variety of affective phenomena related to evaluation, mood, and emotional behavior. The cumulative data reveal that the metaphorical representation of affect has considerable merit. For this reason, the review suggests an expanded research agenda including (a) other perceptual experiences (such as those related to taste and temperature), (b) potential cultural variations, (c) neuroimaging research, and (d) the elucidation of "real world" consequences.
Article
Dedication Acknowledgements Preface 1. Introduction 2. Semantic structure and semantic changes: English perception-verbs in an Indo-European context 3. Modality 4. Conjunction, coordination and subordination 5. Conditionals 6. Retrospect and prospect References Index.