Emotion
2011, Vol. 11, No. 1, 29 –37
© 2011 American Psychological Association
1528-3542/11/$12.00 DOI: 10.1037/a0022017
Long-Lasting Attentional Influence of Negative and Taboo Words in an
Auditory Variant of the Emotional Stroop Task
Julie Bertels and Régine Kolinsky
Elise Pietrons and José Morais
Université Libre de Bruxelles, and Fonds de la Recherche
Scientifique (F.R.S. - FNRS), Brussels, Belgium
Université Libre de Bruxelles
Using an auditory adaptation of the emotional and taboo Stroop tasks, the authors compared the effects
of negative and taboo spoken words in mixed and blocked designs. Both types of words elicited carryover
effects with mixed presentations and interference with blocked presentations, suggesting similar longlasting attentional effects. Both were also relatively resilient to the long-lasting influence of the preceding
emotional word. Hence, contrary to what has been assumed (Schmidt & Saari, 2007), negative and taboo
words do not seem to differ in terms of the temporal dynamics of the interdimensional shifting, at least
in the auditory modality.
Keywords: emotional Stroop, taboo Stroop, spoken words, carryover effects
According to McKenna and Sharma (2004), the frequent lack of
attentional bias with the mixed design is probably due to the
involvement, in the EST, of a slow or long-lasting attentional
component in addition to the fast short-lived interference mechanism that intervenes during stimulus presentation. The slow component would indeed lead to an RT delay not only on the current
(emotional) trial, but also— or even exclusively— on the subsequent (e.g., neutral) trial, hence masking the within-trial fast interference effect. With blocked presentations, given that word
valence is constant across trials, online within-trial interference is
not masked by the occurrence of between-trials long-lasting effects; hence, the observed attentional bias would reflect the combined influence of fast and slow effects. The term carryover effect
was first used by Waters and collaborators (Waters, Sayette, Franken, & Schwartz, 2005; Waters, Sayette, & Wertz, 2003) to refer
to these slow effects and to contrast them with the fast influence on
attention that emotional stimuli operate within a single trial.
However, there is an alternative, nonexclusive explanation for
the fact that interference is usually observed with blocked but not
mixed EST designs: The blocked presentation of emotional stimuli
induces a congruent mood; specifically, negative words may cause
prolonged rumination effects (e.g., Holle, Neely, & Heimberg,
1997; Richards, French, Johnson, Naparstek, & Williams, 1992),
delaying RTs in the negative relative to the neutral blocks.
It is interesting that several studies have reported consistent
interference linked to taboo relative to neutral words in healthy
participants in both mixed and blocked designs of the taboo Stroop
task (TST; Eilola, Havelka, & Sharma, 2007; MacKay & Ahmetzanov, 2005; MacKay et al., 2004; Schmidt & Saari, 2007; Siegrist, 1995; Taylor, Kornblum, Lauber, Minoshima, & Koeppe,
1997). Given the divergent results between the EST and TST
depending on presentation design, Schmidt and Saari (2007) proposed that what is usually referred to as the emotional Stroop
effect would in fact only reflect the slow long-lasting influence of
negative words on attention, whereas the taboo Stroop effect
would result from a fast short-lasting effect of what these authors
considered to be particularly threatening stimuli.
The emotional Stroop task (EST) has often been used to investigate attentional biases linked to emotional information. In this
task, a neutral or emotionally charged word is presented during
each trial, and the participant has to name its physical color as fast
as possible. The “Stroop” label may seem inappropriate (e.g.,
Algom, Chajut, & Lev, 2004) because there is no potential
(in)congruence between the emotional valence of the carrier word
and its color. However, the affective dimension of the word,
although task irrelevant, interferes with color naming: Reaction
times (RTs) are delayed on emotional relative to neutral trials. This
is an instance of the interference effects considered to reflect fast
attentional capture by the emotional content of the word, called
attentional biases or, more commonly in this case, emotional
Stroop effects.
Attentional biases toward negative, but not positive, words have
been observed using the EST (e.g., Dalgleish, 1995; McKenna &
Sharma, 1995; Pratto & John, 1991; Ruiz-Caballero & Bernandez,
1997; but see, e.g., Borkenau & Mauer, 2006; Dresler, Mériau,
Heekeren, & van der Meer, 2009). Nevertheless, in healthy participants not selected on the basis of their anxiety level, these
biases are usually not observed with emotional words (either from
one specific emotional category or from several) mixed with
neutral words within the same block of trials, but only with
blocked presentation, namely with words of a single emotional (or
neutral) category in each block (see Bar-Haim, Lamy, Pergamin,
Bakermans-Kranenburg, & van IJzendoorn, 2007, and Phaf &
Kan, 2007, for meta-analyses, but see, e.g., McKenna & Sharma,
1995).
Julie Bertels and Régine Kolinsky, Research Unit in Cognitive Neurosciences, Université Libre de Bruxelles, Bruxelles, Belgium; Fonds de la
Recherche Scientifique, Bruxelles, Belgium; Elise Pietrons and José Morais, Research Unit in Cognitive Neurosciences, Université Libre de Bruxelles, Bruxelles, Belgium.
This work was supported by Fonds de la Recherche Scientifique Grant
1.5705.06. We thank Nadia Benboutayab for recording the material.
Correspondence concerning this article should be addressed to Julie
Bertels, Université Libre de Bruxelles, Av. F.D. Roosevelt, 50 – C.P. 191,
B – 1050 Bruxelles, Belgium. E-mail: [email protected]
29
30
BERTELS, KOLINSKY, PIETRONS, AND MORAIS
A crucial difference between the effects of negative and taboo
words on attentional orienting would thus consist in their temporal
course. However, this distinction rests on insufficient evidence:
Whereas negative and taboo words have been compared in blocked
presentations (Eilola et al., 2007; Schmidt & Saari, 2007) and in
mixed presentations in which either taboo or negative words were
presented with neutral words in separate blocks (Schmidt & Saari,
2007), to our knowledge, effects of these words have never been
contrasted directly in a mixed, within-subject design with taboo,
negative, and neutral words presented within the same block.
Moreover, carryover effects of taboo words have never been
directly assessed, so that suggesting that slow effects would be the
privilege of negative words is taking a shortcut.
In the present study, we used a hybrid version of the EST and
TST in which we presented negative, positive, taboo, and neutral
words. The aim was to contrast the attentional effects of emotional
words of different valences in both mixed and blocked designs.
More specifically, we investigated the occurrence of fast and slow
effects of these words in a mixed design in which all types of
emotional words (plus neutral words) were presented in the same
block as a within-subject factor. This design provided the opportunity not only to directly compare fast and slow effects of words
of different emotional valence, but also to investigate slow effects
of emotional words on emotional words of either the same or
different emotional valence.
Contrary to previous studies in which linguistic EST and TST
have used written words, we presented spoken words. The auditory
equivalent of the physical color was the identity of the speaker
uttering the word: On each trial, participants had to identify the
speaker as fast as possible from four potential speakers by pressing
one of four buttons labeled with the speakers’ first names. In
addition to increasing the social relevance of the task given that
speaker identification is part of daily life activities, there is a clear
ecological benefit resulting from the auditory presentation.
Indeed, audition is considered to be the main alerting system
(Scharf, 1998): In many daily life situations, people hear a danger
before seeing it. As such, it has been shown that auditory signals
have a more important alerting power than visual signals (Harvey,
1980; Posner, Nissen, & Klein, 1976) and that the localization of
an auditory event could influence the localization of a visual
stimulus, whereas the reverse is not true (Spence & Driver, 1997).
Moreover, regarding verbal stimuli, oral language is a more fundamental and widespread cognitive ability than written language
(see, e.g., Morais & Kolinsky, 2001; Morais, Macedo, & Kolinsky,
2004) because it is acquired before reading and writing and is
phylogenetically more ancient. According to Wurm, Vakoch,
Strasser, Calin-Jageman, and Ross (2001), this is a strong argument for using spoken language in studies that investigate the
impact of emotional stimuli. Still, although the impact of emotional prosody on cognitive processing has been investigated (e.g.,
Brosch, Grandjean, Sander, & Scherer, 2008; Sander et al., 2005;
and using an auditory variant of the EST: Wurm et al., 2001;
Wurm, Labouvie-Vief, Aycock, Rebucal, & Koch, 2004), to our
knowledge, only one study has investigated the influence of the
emotional content of spoken words. Bertels, Kolinsky, and
Morais (2010) used an auditory adaptation of the dot-probe task
(MacLeod, Mathews, & Tata, 1986) to explore the influence of
emotional words on spatial attentional orienting. That study
confirmed the interest of using spoken words to investigate
attentional biases in participants not selected on the basis of
their personality traits: These exhibited spatial attentional biases that are usually not observed with written words.
Coherently, here, we expected to observe attentional biases to
negative and taboo words, namely delayed speaker identification
latencies for these words relative to neutral words. No attentional
bias was expected for positive words (McKenna & Sharma, 1995;
Pratto & John, 1991). In the mixed design, carryover effects were
estimated by taking into account word valence not only in a
specific trial N (N words), but also in the preceding trial N–1 (N–1
words). We expected to observe carryover effects of negative
words, as such slow effects previously have been observed using
visual materials (Ashley & Swick, 2009; Frings, Englert, Wentura,
& Bermeitinger, 2010; Kunde & Mauer, 2008; McKenna &
Sharma, 2004; Waters et al., 2003, 2005). Predictions relative to
taboo words were less clear given that the carryover effects of
these words have never been investigated. Nevertheless, if
Schmidt and Saari (2007) were right in postulating slow effects
only for negative words, interference linked to taboo words would
be observed with both mixed and blocked presentations, whereas
interference linked to negative words would not be observed with
mixed presentations, due to carryover effects, but only with
blocked presentations.
Method
Participants
Seventy-nine native French-speaking students of the Université
Libre de Bruxelles took part in the experiment (51 women; six
left-handed), ranging from 18 to 41 years of age (M ⫽ 23 years).
The results of 15 participants (nine women), all right-handed, were
discarded from the analyses: six because their average RTs were
superior to 2 standard deviations above the overall average performance, eight because their error rates exceeded the same threshold, and one on the basis of both criteria. They were paid for their
participation.
Previous studies investigating attentional biases in the visual
modality have pointed to the robustness of this phenomenon only
in selected groups of participants such as anxious (e.g., Mogg &
Marden, 1990), repressives (e.g., Dawkins & Furnham, 1989), and
depressed people (e.g., Bradley, Mogg, Millar, & White, 1995).
Therefore, we used personality questionnaires (filled in after the
task) assessing anxiety and depression levels (the Spielberger
Trait-State Anxiety Inventory [Spielberger, 1983] and the Beck
Depression Inventory [Beck, Steer, & Brown, 1996], respectively),
as well as willingness to be socially desirable (the Marlowe–
Crowne Scale of Social Desirability [Crowne & Marlowe, 1960])
to check for any correlation between these scores and the emotional effects. Table 1 displays the average scores obtained from
these questionnaires.
Material and Apparatus
Stimuli consisted of 20 negative, 20 positive, and 20 emotionally neutral words (for similar proportions in mixed designs, see,
e.g., McKenna & Sharma, 1995, and Mauer & Borkenau, 2007), as
well as 20 taboo words. These mono- or bisyllabic words came
from the database created by Bertels, Kolinsky, and Morais (2009).
AUDITORY EMOTIONAL AND TABOO STROOP
Table 1
Average Standard Scores Obtained on the STAI-Y (Spielberger,
1963) and on the SDS (Crowne & Marlowe, 1960) and Average
Raw Scores (/63) Obtained on the BDI-II (Beck et al., 1996) in
Mixed and Blocked Designs (Standard Deviations Are in
Parentheses)
Questionnaire
Mixed
design
(n ⫽ 32)
Blocked
design
(n ⫽ 32)
Average
(n ⫽ 64)
STAI–Y
State anxiety
Trait anxiety
SDS
BDI–II
46.41 (7.27)
44.13 (9.10)
48.13 (5.00)
7.81 (6.58)
49.38 (10.78)
49.56 (12.65)
43.91 (7.85)
8.03 (7.72)
47.89 (9.24)
46.84 (11.27)
45.98 (6.89)
7.92 (7.13)
Note. STAI–Y ⫽ Spielberger Trait-State Anxiety Inventory (Form Y);
SDS ⫽ Marlowe–Crowne Scale of Social Desirability; BDI–II ⫽ Beck
Depression Inventory (2nd ed.). The displayed scores are standard scores
for STAI and SDS, while they are raw scores for the BDI.
The four types of words were matched according to oral frequency,
number of phonological neighbors, and phonological uniqueness
point (cf. database Lexique 3.01; New, Pallier, Ferrand, & Matos,
2001; see the Appendix).
Four French-speaking actors (two women, two men) pronounced the words with a neutral tone. Words were digitally
recorded on a Sony MiniDisc and then transferred on a Macintosh
Powerbook G3 via the interface Digidesign DIGI 002 Rack; they
were cleaned, normalized, and synchronized with Protools/
Digidesign 6.2.2 software. Mean word length was 627 ms (SD ⫽
112 ms).
Participants were tested individually wearing headphones in
front of a computer screen. Stimulus presentation and data collection were controlled using the E-Prime button box and 1.1.4.1
software (Schneider, Eschman, & Zuccolotto, 2002) running on an
Asus laptop.
Procedure
Each trial started with a 1,000-ms fixation cross displayed in the
middle of the screen. An auditory word was then presented, uttered
by one of the four speakers. Participants were required to identify
the speaker by pressing one of the four keys of the button box as
quickly and accurately as possible. The first name of each speaker
was written on the top of the corresponding button (Marcha,
Claire, Tim, and Alex). Participants had 5,000 ms to answer. The
interval between the response and the next trial was 500 ms,
leading to a 1,500-ms response–stimulus interval.
The experiment started with three 20-trial practice blocks, during which participants learned to identify the four speakers (i.e., to
match a voice with a first name). They received feedback and
correction in case of error. Next, they were presented with four
80-trial experimental blocks, receiving feedback—a beep—when
the answer was wrong. Practice words were not used in the
experimental blocks.
Participants were randomly assigned to the blocked and mixed
conditions (32 participants by condition; 22 and 20 women in the
blocked vs. mixed condition, respectively).
31
In the blocked condition, all words of a block had the same
emotional valence (neutral, negative, positive, or taboo). Hence,
the same words were repeated 4 times in each block, each time
uttered by a different speaker. At least 10 other words separated
two repetitions of the same word.
In the mixed condition, all words were presented once in each
block in a pseudorandom order: A word of the same emotional
type or said by the same speaker was never presented more than 3
times in a row. One fourth of the words of a specific emotional
type were pronounced by a specific speaker. Doing this, the
emotional valence and the identity of the speaker were not correlated, and each block included the same number of stimuli pronounced by a specific speaker. Blocks differed by the assigned
pseudorandom order and by the speaker pronouncing a specific
word (each word was pronounced by a different speaker in each
block).
In both conditions, block order and response key assignment
were counterbalanced across participants using a Latin square
design.
After the experiment, participants completed the three personality questionnaires listed above.
Results
Emotional and Taboo Stroop Effects With Mixed and
Blocked Presentations
A 2 (presentation: blocked design, mixed design) ⫻ 2 (participant’s gender) ⫻ 4 (word valence: neutral, negative, positive,
taboo) repeated measures analysis of variance (ANOVA) was run
on the RTs. The participant’s gender was considered given previous studies revealing divergences in emotional Stroop effects
between men and women (e.g., Egloff & Schmukle, 2004; Sass et
al., 2010; Smith & Waterman, 2005). Design and gender were
between-subjects factors, and word valence was a within-subject
factor. Average RTs per word valence are presented in Figure 1
separately for the blocked and mixed designs.
Gender was the only significant main effect, F(1, 60) ⫽ 4.066,
p ⬍ .05, with women responding faster than men (735 vs. 790 ms).
The word valence by presentation interaction was also significant,
F(3, 180) ⫽ 8.24, p ⬍ .001.1 As expected, word valence was
significant in both the blocked and mixed designs, F(3, 90) ⫽
3.963, p ⫽ .01, and F(3, 90) ⫽ 6.063, p ⫽ .001, respectively, but
affected RTs quite differently (see Figure 1).
Planned comparisons revealed that, with blocked presentations,
both taboo and negative words led to longer RTs than neutral
words, F(1, 30) ⫽ 4.669, p ⬍ .04, and F(1, 30) ⫽ 9.566, p ⬍ .005,
with no bias occurring for positive words, F ⬍ 1. The positive
attentional biases associated with negative (24 ms) and taboo
words (34 ms) were not correlated with any of the personality
scores (⫺.18 ⬍ rs ⬍ .19, ps ⬎ .10). Similarly, the nonsignificant
attentional bias for positive words (7 ms) was not correlated with
any of these scores (⫺.05 ⬍ rs ⬍ .19, ps ⬎ .10).
With mixed presentations, planned comparisons showed that
taboo words led to shorter RTs compared with neutral words, F(1,
1
When depression, social desirability, and trait- and state-anxiety T
scores were considered as covariates, this interaction remained significant,
F(3, 165) ⫽ 6.831, p ⬍ .001.
32
BERTELS, KOLINSKY, PIETRONS, AND MORAIS
Figure 1. Mean (⫾SE) reaction times (RTs) for each word valence, with the blocked (n ⫽ 32) and mixed (n ⫽
32) presentations.
30) ⫽ 4.755, p ⬍ .04, with positive words only tending to elicit
longer latencies, F(1, 30) ⫽ 3.492, p ⫽ .07. No difference was
found between negative and neutral words, F ⬍ 1. The negative
attentional bias for taboo words (i.e., the difference between RTs
on taboo and neutral words, ⫺15 ms) was not correlated with any
of the personality scores (⫺.13 ⬍ rs ⬍ .00, ps ⬎ .10). Similarly,
the nonsignificant attentional biases for positive and negative
words (⫺15 and 0 ms, respectively) were not correlated with any
of these scores (⫺.17 ⬍ rs ⬍ .06, ps ⬎ .10).
No correlation was found between the mean attentional bias for
each individual emotional word (calculated by subtracting the
mean RT on neutral words from the mean RT on each individual
emotional word, separately for blocked and mixed presentations)
and word oral frequency (blocked: r ⫽ ⫺.202, mixed: r ⫽ ⫺.071,
ps ⬎ .10), number of phonological neighbors (blocked: r ⫽ ⫺.125,
mixed: r ⫽ ⫺.193, ps ⬎ .10), or phonological uniqueness point
(blocked: r ⫽ .039, mixed: r ⫽ .112, ps ⬎ .10).
The mean error rate was 3.04% (range: 0.000 –7.188%). Repeated measures ANOVAs on the error rates corrected by the
arcsine function did not show any significant effect or interaction
( ps ⬎ .10).
Carryover Effects With Mixed Presentations
To track carryover effects with mixed presentations, we looked
for the occurrence of attentional biases, taking into account word
valence not only on trials N but also on trials N–1. Although this
analysis is not orthogonal to the preceding one, its relevance is
supported by various studies suggesting a major involvement of
slow effects in emotional Stroop situations (e.g., Waters et al.,
2003, 2005). Moreover, it provides the opportunity to further
explore the unexpected results observed with our mixed design.
We thus ran a 4 (word valence at trial N) ⫻ 4 (word valence at trial
N–1) repeated measures ANOVA on RTs at trial N.2 Both were
within-subject factors. If carryover effects did occur, we would
expect to observe a significant interaction between word valence
on trial N and word valence on trial N–1.
This was indeed the case. In addition to the main effect of word
valence on trial N, F(3, 93) ⫽ 5.133, p ⫽ .002, already observed
in the previous analysis, there was a main effect of word valence
on trial N–1, F(3, 93) ⫽ 11.586, p ⬍ .001. It is important to note
that these effects were qualified by an interaction between these
two factors, F(9, 279) ⫽ 6.108, p ⬍ .001.
This interaction was decomposed using Bonferroni-adjusted
paired samples t tests.3
First, we compared the size of the fast within-trial effects of
emotional words. Indeed, when the N–1 word was neutral, any
difference observed on the subsequent trial N between RTs on
emotional words (separately for the three types of emotional
words) and RTs on neutral words may be attributed to the fast
within-trial component. These comparisons showed that after an
N–1 neutral word, RTs on negative, positive, or taboo N words did
not differ significantly from RTs on neutral N words ( ps ⬎ .10; see
white bars in Figure 2). These nonsignificant attentional biases
(⫹23, ⫹21, and ⫹ 13 ms, respectively) were not correlated with
any of the personality scores (⫺.107 ⬍ rs ⬍ .177, ps ⬎ .10).
Second, we assessed the size of the slow between-trials carryover effects of emotional words presented on trial N–1 on RTs on
trial N. The size of these effects was estimated as the difference
between RTs on neutral N words following emotional words
(separately for negative, positive, or taboo trials) and RTs on
neutral N words following neutral words. The presentation of N–1
negative or taboo words relative to N–1 neutral words delayed RTs
2
Given that neither the main effect of participant’s gender nor any
interaction involving this factor was significant ( ps ⬎ .10), we did not
consider it in this second analysis.
3
The p values were multiplied by 15 given that 15 t tests were made. We
compared the case in which a neutral word was followed by an emotional
word with the case in which it was followed by a neutral word (three
comparisons). Also, for each word valence at trial N, we compared the case
in which it was preceded by an emotional word with the case in which it
was preceded by a neutral word (4 ⫻ 3 comparisons).
AUDITORY EMOTIONAL AND TABOO STROOP
33
Figure 2. Mean (⫾SE) reaction times (RTs) with mixed presentations (n ⫽ 32), separately for each word
valence in trial N–1 as a function of word valence in trial N. Horizontal lines denote the significant comparisons.
on neutral words presented at trial N (⫹50 and ⫹ 46 ms, respectively, see (a) and (b) in Figure 2), t(31) ⫽ 4.142 and 5.419, ps ⬍
.01. No effect of the presentation of N–1 positive words was
observed ( p ⬎ .10). These attentional biases were not correlated
with any of the personality scores (⫺.194 ⬍ rs ⬍ .177, ps ⬎ .10).
Similarly, the nonsignificant attentional bias for positive words
(⫹8 ms) was not correlated with any of these scores (⫺.049 ⬍
rs ⬍ .068, ps ⬎ .10).
Given that words of different emotional valences were mixed in
the same block of trials, we had the opportunity to assess, in
addition to carryover effects of emotional words on neutral words,
the influence of an emotional N–1 word on the responses to
emotional N words. To our knowledge, this never has been investigated before. For each type of emotional N word, we compared
the case in which it was preceded by an emotional word with the
case in which it was preceded by a neutral word. Negative N–1
words delayed responses to positive N words (⫹68 ms), t(31) ⫽
4.466, p ⬍ .003 (see (c) in Figure 2). All the other comparisons
were not significant ( p ⬎ .10). In particular, RTs on negative and
taboo N words were not affected by the previous presentation of an
emotional (relative to a neutral) N–1 word ( ps ⬎ .10). The significant attentional bias was not correlated with any of the personality scores (⫺.183 ⬍ rs ⬍ .103, ps ⬎ .10). Similarly, the nonsignificant attentional biases were not correlated with any of these
scores (⫺.213 ⬍ rs ⬍ .271, ps ⬎ .10).
Discussion
We created an auditory hybrid version of the EST and TST to
investigate the influence of the task-irrelevant emotional dimension of spoken words on the processing of an unrelated taskrelevant dimension of the same stimulus, here the speaker’s identity. We used both blocked and mixed presentations and examined,
in the mixed design, the temporal course of this emotional influence by investigating the effects of the presentation of an emotional word on the response to the subsequent word. The influences
of negative, positive, and taboo emotional valences were consid-
ered in this study. To our knowledge, this is the first study that has
systematically investigated long-lasting, carryover effects of taboo
words in a mixed design, and that allowed a direct comparison
with the effects of negative and positive words presented in the
same block. This creates a quite ecologically valid emotional
context given that everyday conversations mostly consist of a
mixture of words of various emotional contents (e.g., negative,
neutral, and taboo; negative, positive, and neutral, etc.). Moreover,
the fact that different types of emotional words were presented in
the same block gave us the opportunity to assess interemotional
carryover effects, whereas previous studies exclusively have assessed carryover effects of emotional stimuli on subsequent neutral
stimuli. Another major interest of our study consisted of the
presentation of spoken rather than written words. In addition to the
fact that oral language is a more fundamental and widespread
cognitive ability than written language (see, e.g., Morais & Kolinsky, 2001; Morais et al., 2004) and that auditory signals have a
more important alerting power than visual signals (Harvey, 1980;
Posner et al., 1976), the use of spoken words allowed using an
ecologically valid speaker identification task through which attentional effects of taboo words were assessed in their most frequent
(oral) vehicle.
As predicted, interference of negative and taboo words, but not
of positive words, was observed with blocked presentations. Thus,
the first contribution of the present study is to show that emotional
and taboo Stroop effects are not only visual effects, but can also be
observed with auditory stimuli. Although further studies should
aim at comparing more directly and precisely the effects of visual
and auditory stimuli, this finding suggests that the occurrence of
such effects depends on the activation of semantic, probably
modality-independent, representations.
With mixed presentation, results were more complex, given that
RTs depended both on the emotional valence of the word at a
specific trial and on the emotional valence of the word presented
at the previous trial. In summary, no fast effect of emotional words
34
BERTELS, KOLINSKY, PIETRONS, AND MORAIS
was observed. Instead, negative and taboo words slowed RTs in
ensuing trials containing neutral words.
We first discuss the absence of any fast effect of negative and
taboo words with the mixed presentation. Then, we consider their
carryover effects, and finally, taking the data of both the blocked
and mixed presentations into account, comment on potential effects of the emotional context.
No Fast Effect of Negative and Taboo Words
As noted above, data from the mixed condition of our study do
not support the existence of any fast within-trial effect of the
emotional valence of the words on the processing of the taskrelevant dimension of the stimuli. Indeed, following the presentation of a neutral word, participants were not slower to identify the
speaker uttering an emotional rather than a neutral word.
The absence of any fast effect of negative words is coherent with
previous findings from studies investigating this effect of negative
stimuli in the visual modality, either with written words (McKenna &
Sharma, 2004; but see Ashley & Swick, 2009) or with pictures
(Kunde & Mauer, 2008). We have thus extended these findings to the
auditory modality. However, Frings et al. (2010) recently showed that
pseudorandom designs consisting of controlling the succession of
neutral and emotional words actually create a contingency between
the occurrence of neutral and emotional words that masks fast effects
of the emotional words. Although in our mixed presentation we
indeed prevented the succession of more than three words of the same
emotional type, the fact that we used four emotional categories (instead of two as in Frings et al.) allowed not inducing a robust pattern
of contingencies. Nevertheless, further studies should take this possibility into consideration by presenting emotional and neutral trials
completely randomly.
Contrary to negative words, the temporal course of the effects
linked to taboo words has never been investigated. Nevertheless, it
is widely recognized that taboo words do lead to fast effects (Eilola
et al., 2007; MacKay & Ahmetzanov, 2005; MacKay et al., 2004;
Schmidt & Saari, 2007; Siegrist, 1995; Taylor et al., 1997). As a
matter of fact, interference effects consistently have been observed
with taboo words in blocked and mixed designs, a situation revealing the influence of fast effects in theory. Based on this
observation, Schmidt and Saari (2007) proposed that the taboo
Stroop effect would be distinct from the emotional Stroop effect in
that it would result from a short- rather than long-lasting influence
of these words. However, given that no effect of taboo words was
observed here after the presentation of a neutral word, namely in
a situation in which only within-trial fast effects can occur, our
auditory results are inconsistent with this idea. Further studies in
the visual modality should aim to explore systematically the existence of fast effects of taboo words. Whatever the conclusion of
such studies, it is worth noting that our results are more liable to
reflect everyday functioning given that taboo words are more
frequent in speech than in written language.
Carryover Effects of Negative and Taboo Words
Although negative and taboo words did not induce fast withintrial effects, they did affect processing speed on the next trial,
hence demonstrating that they generate slow long-lasting effects.
As a matter of fact, carryover effects of both types of words were
observed in the mixed condition. Both negative and taboo words
delayed RTs on subsequently presented neutral words. Negative
(but not taboo) words also delayed RTs on subsequently presented
positive words. Remarkably, these carryover effects lasted for
more than 1 s, given that the mean word length (627 ms) and the
response–stimulus interval (1,500 ms) led to a long intertrial
interval. Usually, for long intertrial intervals, slow effects tend to
fade, at least in the visual modality (McKenna & Sharma, 2004).
What do these slow long-lasting effects reflect? Although there
is no general agreement about the underlying psychological mechanisms, various authors assume that attentional disengagement can
be identified as the main locus of the effect (McKenna & Sharma,
2004; Phaf & Kan, 2007; Strauss, Allen, Duke, Ross, & Schwartz,
2008; Waters et al., 2003, 2005; see also White, 1996). The idea
that attentional biases are due to difficulties in disengaging attention from the emotional content of stimuli was first formulated in
the context of spatial attentional orienting. Although facilitation of
responses to a probe presented at the same spatial location where
an emotional stimulus had been presented right before has been
interpreted in terms of attentional vigilance (i.e., an effect at the
level of the engagement of attention; for a meta-analysis, see
Bar-Haim et al., 2007), various authors suggest that this effect
would actually reflect difficulties in disengaging attention from
that location (e.g., Fox, Russo, Bowles, & Dutton, 2001; Fox,
Russo, & Dutton, 2002; Koster, Crombez, Verschuere, & De
Houwer, 2004). Even if no spatial attentional shift is involved in
the emotional Stroop situation, this idea has been transferred to the
dimensional orienting of attention, namely to attentional orienting
toward a specific dimension of the stimulus (e.g., Garner, 1962,
1974). Once attention has been initially captured by a stimulus,
participants should disengage from it to process the next stimulus.
In particular, in the case of a stimulus containing an emotional,
task-irrelevant dimension, attentional disengagement from this dimension would be difficult, hampering the processing of the taskrelevant dimension of the next stimulus. Our results suggest that
(at least when stimuli are spoken words) such attentional disengagement problems arise with threatening stimuli (negative or
taboo), but not with positive ones.
It is interesting that the present results also suggest that these
threatening stimuli are themselves relatively immune to such carryover effects. Indeed, responses to negative and taboo words were not
at all affected by the emotional valence of the preceding word. Yet, as
pointed out by Frings et al. (2010), the impact of priming processes
should moderate these conclusions: The successive presentation of
two words of the same emotional valence would create a kind of
“spreading inhibition,” making the target less accessible than when it
is preceded by a neutral word, causing less interference. Nevertheless,
this idea is difficult to reconcile with the pattern of data observed in
the present study given that only slow effects and no fast interference
of negative or taboo words were found.
Thus, although there is no fast within-trial effect of threatening
stimuli, their emotional valence seems both to induce long-lasting
attentional disengagement effects (observed on the next trial) and
to modulate the impact of the previous trial on the current trial.
Potential Effects of the Emotional Context
Interference effects were observed for both negative and
taboo words in the blocked design. In this condition both fast
AUDITORY EMOTIONAL AND TABOO STROOP
and slow effects can occur, but the interference effects reported
in the literature are usually interpreted as reflecting mostly the
occurrence of slow effects (see, e.g., McKenna & Sharma,
2004). However, the pattern of carryover effects observed in the
present study with mixed presentation is not consistent with this
idea. Indeed, as already commented, with mixed presentations,
negative and taboo words only delayed RTs on subsequent
neutral trials, not on subsequent negative or taboo trials. Hence,
carryover effects similar to the ones observed with mixed
presentation cannot account for the interference effects observed with blocked presentations. Although further research
should aim at better understanding the origin of these effects, it
may be the case that the successive presentation of many
negative words induced a congruent, negative mood, responsible for RT slowing in the emotionally homogeneous blocks
(Holle et al., 1997; Richards et al., 1992). This may have
occurred in the taboo blocks as well, insofar as both negative
and taboo words are negatively loaded (see the Appendix and
Bertels et al., 2009).
Similarly, one may argue that congruent mood explains the
results observed with mixed presentations. Indeed, in the mixed
design, all types of emotional (and neutral) words were presented. This gave us the opportunity to assess carryover effects
of emotional words on subsequent emotional words, and allowed us to compare the effects of different types of emotional
words in a within-subject design without repeating the same
neutral words across successive blocks of trials (which may
induce habituation effects) or using a different set of neutral
words for each type of emotional words. But, in the mixed
design, the proportion of words of negative valence (either
negative or taboo words) was higher than the proportion of
neutral words or positive words, probably creating a negative
emotional context that could have induced a negative mood.
Note that most other studies mixing only negative (or only
taboo) words with neutral words in the same block of trials used
the same proportion of negative words (i.e., half of the trials) as
our study. Hence, the fact that our data are quite coherent with
the results observed in these studies (e.g., McKenna & Sharma,
2004) does not eliminate the possibility that negative mood
induced the effects in all cases.
However, at least one piece of evidence supports the idea that
the specific emotional context used here was not mainly responsible for the pattern of results observed with mixed presentations. If one assumes that the high proportion of negative
words would have induced a negative mood bias that slowed
RTs (Holle et al., 1997; Richards et al., 1992), this should have
been the case not only in the negative and taboo homogeneous
blocks of the blocked presentation but also overall in the mixed
presentation. Contrary to this idea, overall RTs did not differ
between presentations in the repeated measures ANOVA, F ⬍
1, and the RT difference between mixed versus blocked designs
did not reach significance for either neutral or positive words,
t(62) ⫽ 1.208 and 1.488, respectively, ps ⬎ .10.
Thus, although further studies will have to assess more
precisely the impact of the specific emotional contexts created
by mixing words of various emotional valences, we think that
the advantages of the specific mixed design used in the present
study prevailed over this potential limitation.
35
Conclusion
Contrary to the idea that the effects of negative and taboo words
differ primarily according to their temporal course, with taboo
Stroop effects resulting from fast attentional effects and emotional
Stroop effects reflecting slow effects of negative stimuli (Schmidt
& Saari, 2007), the present results show that the effects of both
types of words largely overlap, at least when their vehicle is
spoken. As a matter of fact, neither negative nor taboo words
elicited fast effects. Rather, both had long-lasting influences on the
processing of nonthreatening stimuli, and both were themselves
relatively resilient to carryover effects of preceding emotional
words. Finally, both types of words produced interference effects
when presented in blocks.
In conclusion, the present results suggest that at least with
respect to auditory presentations, negative and taboo words do not
differ in terms of the temporal dynamics of the interdimensional
shifting. Future studies should aim at establishing whether the
same holds true for visual presentations, for which the taboo
Stroop effect seems more reliable than the emotional Stroop effect.
References
Algom, D., Chajut, E., & Lev, S. (2004). A rational look at the emotional
Stroop phenomenon: A generic slowdown, not a Stroop effect. Journal
of Experimental Psychology: General, 133, 323–338.
Ashley, V., & Swick, D. (2009). Consequences of emotional stimuli: Age
differences on pure and mixed blocks of the emotional Stroop. Behavioral and Brain Functions, 5(14). doi:10.1186/1744-9081-5-14
Bar-Haim, Y., Lamy, D., Pergamin, L., Bakermans-Kranenburg, M. J., &
van IJzendoorn, M. H. (2007). Threat-related attentional bias in anxious
and non-anxious individuals: A meta-analytic study. Psychological Bulletin, 133, 1–24.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck
Depression Inventory (2nd ed). San Antonio, TX: Psychological Corporation.
Bertels, J., Kolinsky, R., & Morais, J. (2009). Norms of emotional valence,
arousal, threatening valence and shocking valence for 80 spoken
words—Comparison between neutral and emotional tones of voice.
Psychologica Belgica, 49, 19 – 40.
Bertels, J., Kolinsky, R., & Morais, J. (2010). Emotional valence of spoken
words influences the spatial orienting of attention. Acta Psychologica,
134, 264 –278.
Borkenau, P., & Mauer, N. (2006). Processing of pleasant, unpleasant, and
neutral words in a lateralised emotional Stroop task. Cognition & Emotion, 20, 866 – 877.
Bradley, B. P., Mogg, K., Millar, N., & White, J. (1995). Selective
processing of negative information—Effects of clinical anxiety, concurrent depression, and awareness. Journal of Abnormal Psychology, 104,
532–536.
Brosch, T., Grandjean, D., Sander, D., & Scherer, K. R. (2008). Behold the
voice of wrath: Cross-modal modulation of visual attention by anger
prosody. Cognition, 106, 1497–1503.
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability
independent of psychopathology. Journal of Consulting Psychology, 24,
349 –354.
Dalgleish, T. (1995). Performance on the emotional Stroop task in groups
of anxious, expert, and control subjects: A comparison of computer and
card presentation formats. Cognition & Emotion, 9, 341–362.
Dawkins, K., & Furnham, A. (1989). The colour naming of emotional
words. British Journal of Psychology, 80, 383–389.
Dresler, T., Mériau, K., Heekeren, H. R., & van der Meer, E. (2009).
Emotional Stroop task: Effect of word arousal and subject anxiety on
emotional interference. Psychological Research, 73, 364 –371.
36
BERTELS, KOLINSKY, PIETRONS, AND MORAIS
Egloff, B., & Schmukle, S. C. (2004). Gender differences in implicit and
explicit anxiety measures. Personality and Individual Differences, 36,
1807–1815.
Eilola, T. M., Havelka, J., & Sharma, D. (2007). Emotional activation in
the first and second language. Cognition & Emotion, 21, 1064 –1076.
Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening
stimuli draw or hold visual attention in subclinical anxiety? Journal of
Experimental Psychology: General, 130, 681–700.
Fox, E., Russo, R., & Dutton, K. (2002). Attentional bias for threat:
Evidence for delayed disengagement from emotional faces. Cognition &
Emotion, 16, 355–379.
Frings, C., Englert, J., Wentura, D., & Bermeitinger, C. (2010). Decomposing the emotional Stroop effect. Quarterly Journal of Experimental
Psychology, 63, 42– 49.
Garner, W. R. (1962). Uncertainty and structure as psychological concepts. New York: Wiley.
Garner, W. R. (1974). The processing of information and structure. Potomac, MD: Erlbaum.
Harvey, N. (1980). Non-informative effects of stimuli functioning as cues.
Quarterly Journal of Experimental Psychology, 32, 413– 425.
Holle, C., Neely, J. H., & Heimberg, R. G. (1997). The effects of blocked
versus random presentation and semantic relatedness of stimulus words
on response to a modified Stroop task among social phobics. Cognitive
Therapy and Research, 21, 681– 697.
Koster, E. H. W., Crombez, G., Verschuere, B., & De Houwer, J. (2004).
Selective attention to threat in the dot probe paradigm: Differentiating
vigilance and difficulty to disengage. Behaviour Research and Therapy,
42, 1183–1192.
Kunde, W., & Mauer, N. (2008). Sequential modulations of valence
processing in the emotional Stroop task. Experimental Psychology, 55,
151–156.
MacKay, D. G., & Ahmetzanov, M. V. (2005). Emotion, memory, and
attention in the taboo Stroop paradigm: An experimental analogue of
flashbulb memories. Psychological Science, 16, 25–32.
MacKay, D. G., Shafto, M., Taylor, J. K., Marian, D. E., Abrams, L., &
Dyer, J. R. (2004). Relations between emotion, memory, and attention:
Evidence from taboo Stroop, lexical decision, and immediate memory
tasks. Memory & Cognition, 32, 474 – 488.
MacLeod, C., Mathews, A., & Tata, P. (1986). Attentional bias in emotional disorders. Journal of Abnormal Psychology, 95, 15–20.
Mauer, N., & Borkenau, P. (2007). Temperament and early information
processing: Temperament-related attentional bias in emotional Stroop
tasks. Personality and Individual Differences, 43, 1063–1073.
McKenna, F. P., & Sharma, D. (1995). Intrusive cognitions: An investigation of the emotional Stroop task. Journal of Experimental Psychology, 21, 1595–1607.
McKenna, F. P., & Sharma, D. (2004). Reversing the emotional Strop
effect reveals that it is not what it seems: The role of fast and slow
components. Journal of Experimental Psychology: Learning, Memory,
and Cognition, 30, 382–392.
Mogg, K., & Marden, B. (1990). Processing of emotional information in
anxious subjects. British Journal of Clinical Psychology, 29, 227–229.
Morais, J., & Kolinsky, R. (2001). The literate mind and the universal
human mind. In E. Dupoux (Ed.), Language, brain and cognitive development: Essays in honor of Jacques Mehler (pp. 463– 480). Cambridge,
MA: MIT Press.
Morais, J., Macedo, C., & Kolinsky, R. (2004). Contrainte et expérience
dans l’apprentissage de la lecture. In L. Ferrand & J. Grainger (Eds.),
Psycholinguistique cognitive: Essais en l’honneur de Juan Segui [Cognitive psycholinguistics: Essays in honor of Juan Segui]. (pp. 269 –288).
Bruxelles: De Boeck.
New, B., Pallier, C., Ferrand, L., & Matos, R. (2001). Une base de données
lexicales du français contemporain sur internet: LEXIQUE. [A lexical
database of contemporary French on the internet: LEXIQUE]. L’Année
Psychologique, 101, 447– 462.
Phaf, R. H., & Kan, K.-J. (2007). The automaticity of emotional Stroop: A
meta-analysis. Journal of Behavior Therapy and Experimental Psychiatry, 38, 184 –199.
Posner, M. I., Nissen, M. J., & Klein, R. M. (1976). Visual dominance: An
information-processing account of its origins and significance. Psychological Review, 83, 157–171.
Pratto, F., & John, O. P. (1991). Automatic vigilance: The attentiongrabbing power of negative social information. Journal of Personality
and Social Psychology, 61, 380 –391.
Richards, A., French, C. C., Johnson, W., Naparstek, J., & Williams, J.
(1992). Effects of mood manipulation and anxiety on performance of an
emotional Stroop task. British Journal of Psychology, 83, 479 – 491.
Ruiz-Caballero, J. A., & Bernandez, J. (1997). Anxiety and attention: Is
there an attentional bias for positive words? The Journal of General
Psychology, 124, 194 –211.
Sander, D., Grandjean, D., Pourtois, G., Schwartz, S., Seghier, M. L.,
Scherer, K. R., & Vuillemier, P. (2005). Emotion and attention interactions in social cognition: Brain regions involved in processing anger
prosody. NeuroImage, 28, 848 – 858.
Sass, S. M., Heller, W., Stewart, J. L., Levin Silton, R., Edgar, J. C., Fisher,
J. E., & Miller, G. A. (2010). Time course of attentional bias in anxiety:
Emotion and gender specificity. Psychophysiology, 47, 247–259.
Scharf, B. (1998). Auditory attention: The psychoacoustical approach. In
H. Pashler (Ed.), Attention (pp. 75–117). Hove, England: Psychology
Press.
Schmidt, S. R., & Saari, B. (2007). The emotional memory effect: Differential processing or item distinctiveness? Memory & Cognition, 35,
1905–1916.
Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime reference
guide. Pittsburgh, PA: Psychology Software Tools.
Siegrist, M. (1995). Effects of taboo words on color-naming performance
on a Stroop test. Perceptual and Motor Skills, 81, 1119 –1122.
Smith, P., & Waterman, M. (2005). Sex differences in processing aggression words using the emotional Stroop task. Aggressive Behavior, 31,
271–282.
Spence, C., & Driver, J. (1997). Audiovisual links in exogenous covert
spatial orienting. Perception & Psychophysics, 59, 1–22.
Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory
(Form Y) (“Self-evaluation questionnaire”). Palo Alto, CA: Consulting
Psychologists Press.
Strauss, G. P., Allen, D. N., Duke, L. A., Ross, S. A., & Schwartz, J.
(2008). Automatic affective processing impairments in patients with
deficit syndrome schizophrenia. Schizophrenia Research, 102, 76 – 87.
Taylor, S. F., Kornblum, S., Lauber, E. J., Minoshima, S., & Koeppe, R. A.
(1997). Isolation of specific interference processing in the Stroop task:
PET activation studies. NeuroImage, 6, 81–92.
Waters, A. J., Sayette, M. A., Franken, I. H. A., & Schwartz, J. E. (2005).
Generalizability of carry-over effects in the emotional Stroop task.
Behaviour Research and Therapy, 43, 715–732.
Waters, A. J., Sayette, M. A., & Wertz, J. M. (2003). Carry-over effects can
modulate emotional Stroop effects. Cognition & Emotion, 17, 501–509.
White, M. (1996). Automatic affective appraisal of words. Cognition &
Emotion, 10, 199 –211.
Wurm, L. H., Labouvie-Vief, G., Aycock, J., Rebucal, K. A., & Koch,
H. E. (2004). Performance in auditory and visual emotional Stroop tasks:
A comparison of older and younger adults. Psychology and Aging, 19,
523–535.
Wurm, L., Vakoch, D., Strasser, M., Calin-Jageman, R. J., & Ross, S.
(2001). Speech perception and vocal expression of emotion. Cognition
& Emotion, 15, 831– 852.
AUDITORY EMOTIONAL AND TABOO STROOP
37
Appendix
Values of Emotional Dimensions and Lexical Factors for Each Emotional
Category of Words
Emotional
valence
(1–7)
Emotional
category n Mean
Neutral
Negative
Positive
Taboo
20
20
20
20
4.444
1.821
6.225
2.787
Arousal
(1–7)
Threatening
valence
(1–5)
Shocking
valence
(1–5)
SD
Mean
SD
Mean
SD
Mean
SD
1.181
1.553
1.191
1.588
3.820
5.923
2.245
5.248
1.011
1.144
1.466
1.370
1.400
4.253
1.275
3.043
0.741
0.970
0.661
1.242
1.083
2.665
1.175
3.553
0.268
1.309
0.448
1.195
Mean oral
frequency
Mean
SD
Mean number
Mean
of
phonological
phonological uniqueness
neighbors
points
Mean
SD
Mean
SD
67.964 75.070 10.100 8.372 3.950 1.356
73.895 126.61
6.400 6.524 4.600 1.095
71.904 107.11
7.600 6.193 4.500 0.889
36.675 76.773 8.300 7.449 3.800 1.473
Received December 19, 2008
Revision received May 18, 2010
Accepted September 10, 2010 䡲