Article
The representation of
emotion vs. emotion-laden
words in English and Spanish
in the Affective Simon Task
International Journal of Bilingualism
15(3) 310­–328
© The Author(s) 2010
Reprints and permission:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1367006910379261
Ijb.sagepub.com
Jeanette Altarriba and Dana M. Basnight-Brown
State University of New York, USA
Abstract
An Affective Simon Task was administered to English-speaking monolinguals and Spanish–English
bilinguals, in order to assess the extent to which valence and emotionality are automatically
processed when reading a word. Participants classified words in white on the basis of valence
(positive or negative), or classified them on the basis of color (blue or green) via a key press.
Words were either emotion words (e.g., happy; anxious) or emotion-laden words (e.g., dream;
shark). Bilinguals viewed words in both English and Spanish. While only negative emotion words
produced the typical congruency effects, both negative and positive emotion-laden words produced
significant Simon effects, in monolinguals. Similar effects emerged for bilinguals, in both languages.
Results indicate that emotion word type moderates the Affective Simon Task and also provide
a demonstration of these effects in bilinguals. Overall, the data also provide further evidence
that affective coding in both a dominant and a subordinate language can influence subsequent
responses in seemingly irrelevant tasks. Results are discussed within a framework focused on the
representation of emotion in both monolingual and bilingual speakers.
Keywords
Affective Simon Task, bilingualism, emotion-laden words, emotion word processing, Simon Effect
Investigations on the processing and representation of emotion have increased in recent years,
as researchers attempt to explore the degree to which stimulus valence (positive, negative,
neutral) can be processed automatically. In a paradigm such as affective priming, for example,
a positively or negatively valenced target is facilitated when preceded by a similarly valenced
prime (see e.g., Fazio, Sanbonmatsu, Powell, & Kardes, 1986; see also Altarriba & Canary,
2004, for a bilingual demonstration of this effect). That is, positive primes facilitate responses
to positive targets, and negative primes facilitate responses to negative targets. These results
have been shown in situations in which the delay between the onset of the prime and the presentation of the target (i.e., stimulus onset asynchrony; SOA) was 300 ms (milliseconds) or
Corresponding author:
Jeanette Altarriba, Department of Psychology, Social Sciences 399, University at Albany, State University of New York,
Albany, NY 12222, USA.
Email: [email protected]
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
Altarriba and Basnight-Brown
311
less, and even in situations where the participants are asked to ignore the prime. This type of
priming is fast and efficient, involuntary, and goal-independent. In short, it is quite automatic
and modulated primarily by valence. Similarly, the emotional Stroop effect has been explored
for what it might tell us regarding the automatic capture of attention of the emotional components of word stimuli (see e.g., Eilola, Havelka, & Sharma, 2007; Sutton & Altarriba, 2008;
Sutton, Altarriba, Gianico, & Basnight-Brown, 2007). Data indicate that negative emotion
words lead to slower reaction times for color naming than positive or neutral words. Most
important, in these tasks, the valence of the stimuli is irrelevant to the primary task, yet still
exerts a powerful effect.
A related tool that has been used for both clinical and nonclinical populations is a version of the
Spatial Simon Task called the Affective Simon Task. De Houwer and Eelen (1998, Experiments 1
and 2) presented participants with nouns and adjectives that had either a positive, negative, or
neutral affective connotation (e.g., FRIEND). Participants had to attend to the grammatical status
of the words (in this case, the word is a noun). Their response was either a positive or negative
word depending on whether a noun or an adjective was presented. For example, they were instructed
to say ‘POSITIVE’ to all nouns and ‘NEGATIVE’ to all adjectives. Pairings of response and grammatical status were counterbalanced across participants. One can consider the relevant feature of
the task to be the grammatical category to which a word belonged. The irrelevant feature was the
affective connotation of the word which was not at all accentuated in the task and was to be
ignored. However, if the emotional content of the word exerts some kind of influence over processing, then it is likely that emotionality is processed automatically when engaging in the current task.
Note that the irrelevant feature and the response are both meaningfully related and both specify
affective information. However, the relevant feature is not meaningfully related to the response or
the irrelevant feature. Ultimately, the results revealed faster reaction times when word valence
matched the response (i.e., responding ‘POSITIVE’ to nouns and ‘FRIEND’ is a noun; a congruency effect), despite the fact that the task focused on grammatical categorization and participants
had been instructed to ignore the meaning of the word.
In a third experiment, De Houwer and Eelen (1998) investigated whether or not the Affective
Simon Effect would materialize if response words were items that did not describe or directly label
an emotion. Response words were employed that did not specifically label valence: FLOWER or
CANCER. The stimulus words were not semantically related to these response words. Thus, half
of the participants labeled nouns with the response ‘FLOWER’ and adjectives with the response
‘CANCER,’ and the other half of the participants had the reverse pairings. Again, there was a main
effect of congruency as reaction times were faster when a match occurred between the valence of
the response word and the grammatical category of the target item (e.g., ‘FLOWER’ was the
response for nouns, and FRIEND was a positive stimulus word). Thus, the authors concluded that
valence was activated in a manner that was involuntary and efficient and motivated a difference in
response times between congruent and incongruent conditions. This effect emerged in the third
experiment when the response words did not directly denote valence, as was the case in their first
two experiments.
De Houwer, Crombez, Baeyens, and Hermans (2001, Experiment 1) examined the Affective
Simon Effect and asked whether or not this effect generalized to cases in which responses were not
based on grammatical category, but rather on semantic category. Participants viewed target words
that named animals, such as BUTTERFLY and COCKROACH and those that named persons, such
as FRIEND and ENEMY. The vocal responses ‘POSITIVE’ and ‘NEGATIVE’ were counterbalanced across semantic category (i.e., animals and persons). A significant congruency effect emerged
in which responses that matched the valence of the target word were faster than those that
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
312
International Journal of Bilingualism 15(3)
mismatched. In a second experiment, a new condition was included that was designed to minimize
semantic processing of the target words. Within this condition, participants viewed words in either
uppercase (e.g., FRIEND and CANCER) or lowercase (e.g., honest and stupid). The vocal
responses ‘POSITIVE’ and ‘NEGATIVE’ were counterbalanced across lettercase conditions. The
same items were used within this experiment in a grammatical condition whereby ‘POSITIVE’ and
‘NEGATIVE’ were counterbalanced across grammatical category, as in previous experiments. The
results revealed the typical congruency effect that typifies the Affective Simon Effect for both
conditions, though the magnitude of the effect was greatly reduced in the lettercase conditions.
Thus, even in instances where participants are to attend to the physical features of the words, thus
minimizing or eliminating semantic processing, valence seems to play a role in responding to
emotion-related stimuli.
In a third and final experiment, De Houwer et al. (2001) extended the Affective Simon Effect to
pictures. In a semantic condition, participants were shown pictures of either manmade objects
(e.g., CHOCOLATE and TOILET) or natural objects (e.g., RIVER and FIRE). Vocal responses of
‘POSITIVE’ and ‘NEGATIVE’ were counterbalanced across the categories manmade and natural.
In a perceptual condition, pictures of the objects were shown in either black-and-white or color,
instead of words. Participants made vocal responses of either ‘POSITIVE’ or ‘NEGATIVE’ counterbalanced across the two perceptual conditions. While the Affective Simon Effect (i.e., the usual
congruency effect) emerged for pictures in the semantic condition, the effect was not demonstrated
in the perceptual condition. That is, in the latter case, there was a nonsignificant 2 ms congruency
effect. The authors suggested that in the case of the perceptual condition, participants were likely
able to ‘block out’ the influence of valence. The task simply did not activate valence sufficiently to
produce an effect. Moreover, the data indicate that participants were much faster to respond in the
perceptual condition than had been observed for all other previous demonstrations of the Affective
Simon Effect. These results replicated the work of Tipples (2001), where picture stimuli were used
though response words varied from those of De Houwer et al. (2001). Additionally, parallel results
have now been reported using spider pictures within the Affective Simon Task with high and low
fearful individuals (cf. Huijding & De Jong, 2005).
De Houwer (2003) introduced a new variant of the above Affective Simon Task. He varied the
task such that the responses themselves were no longer ‘POSITIVE’ or ‘NEGATIVE’. Instead, a
key press was used, as color discrimination became the main focus of the task. With this new variant, one can now present every target word in every color condition, rather than having them fixed
to a particular category. Participants viewed a word in white and were asked to press a computer
key (e.g., P) if the word was positive (e.g., KIND) and a different key (e.g., Q) if the word was
negative (e.g., HOSTILE). The words appeared again, only this time they appeared in either green
or blue, and the task was merely to press they key that corresponded to the color of the word (e.g.,
P for green words; Q for blue words). If the word was positively valenced (e.g., KIND in the
example above) and appeared in green, reaction times were faster – a congruency effect – since
the color now corresponded to the key that had been paired earlier with positive responses. Within
his procedures, a practice block of white words was presented first, followed by a practice block of
color words, which was in turn followed by a test block containing both white and colored words
in random order. Participants also received error feedback in the form of a red ‘+’ sign whenever
an error occurred.
Data within the above task were analyzed for reaction time to the colored words. The results
revealed that key responses to colored words were facilitated if the response had previously been
assigned to a concept sharing the same valence. (For example, KIND appeared in white and
received a ‘P’ response and then appeared in blue for a ‘P’ response; or, as a negative example,
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
Altarriba and Basnight-Brown
313
HOSTILE appeared in white and received a ‘Q’ response then appeared in green for a ‘Q’ response.)
Importantly, the magnitude of the effect was significantly larger for negative stimuli as compared
to positive stimuli. Thus, the effects were not merely the result of target–response compatibility but
rather were modulated by valence. These are responses merely based on color discrimination and
not on the valence of the words, per se, yet valence likely mediated these effects. However, do
emotion words and emotion-laden words produce similar effects? In the typical Affective Simon
Task, these word types are often mixed together, as emotion words such as HAPPY are intermixed
with emotion-laden words such as CANCER or BUTTERFLY. Additionally, how might these
effects be influenced by language in the case of bilingual speakers?
Although researchers have questioned the distinction between emotion words and emotionladen words (cf. Altarriba, 2006; Pavlenko, 2008a, 2008b), little to no empirical data exist that
outline those distinctions. Altarriba and Bauer (2004) noted that emotion words are better recalled
than abstract and concrete words. They also demonstrated positive priming effects for emotion–
emotion word pairs and abstract–emotion word pairs, but these same effects did not emerge for
emotion–abstract word pairs. That is, it appears that word type (i.e., emotion vs. abstract) moderates facilitation in word priming tasks indicating that emotion words may represent a distinct word
category as compared to abstract or other non-emotion word categories. In fact, Altarriba and
Bauer provided data that indicated that concrete, abstract, and emotion words are perceived to
reflect differences in context-availability, concreteness, and imageability (see also Altarriba, Bauer,
& Benvenuto, 1999, for similar findings). Yet, these patterns of representation and processing have
not been examined in similar ways for emotion-laden words. Altarriba and Canary (2004) used
emotion-laden stimuli and demonstrated significant, positive priming effects for moderately or
highly arousing primes paired with highly arousing targets (e.g., PRISONER–JAIL; CRIMINAL–
JAIL). However, their stimuli did not include emotion words so as to provide a basis for comparison between these two emotion word types. Thus, the current work represents the first attempt to
empirically examine processing of emotion and emotion-laden words, in a systematic framework,
within a single set of experiments.
In a similar vein, work on distinguishing emotion from emotion-laden words has not been
empirically examined across bilingual populations. Researchers have revealed, however, that the
first language learned or ‘L1’ may be seen as the more ‘emotional’ of a bilingual’s two languages.
Data supporting this contention stem from studies involving word ratings (Altarriba, 2003),
wherein Spanish words that are emotional are reported to have greater context-availability (more
easily lead to accessing a relevant context in which they could appear, from memory) in Spanish–
English bilinguals; galvanic skin responses (Harris, Ayçiçeği, & Berko-Gleason, 2003; Harris,
Berko-Gleason, & Ayçiçeği, 2005), wherein words in L1 produced stronger effects than in L2;
and questionnaire and survey data wherein participants reported that the pragmatic usage of language more often included the use of L1 for emotional contexts, themes, and topics than L2 for
those very same categories (Dewaele, 2006, 2008). Thus, even though a general pattern of results
is emerging with regard to the relative emotionality of a first language as compared to a second
language, as a field of empirical inquiry this area is very much in its infancy. The current studies,
therefore, are aimed toward providing a systematic set of studies that examine the notion of emotion word processing in L1 vs. L2 for fluent bilingual speakers.
The present research has three main aims. The first aim was to further consider the role of
valence in performance on the Affective Simon Task in the processing of emotion vs. emotionladen words. A second aim was to examine the relative influence of positive vs. negative words,
distinctly, rather than collapsing across these characteristics, as has been done in the majority of
cases investigating the above task. Finally, a third aim was to shed further light on the processing
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
314
International Journal of Bilingualism 15(3)
of emotion concepts in the L1 vs. the L2 of proficient bilingual speakers. Thus, in the current
experiments, an Affective Simon Task was administered to English-speaking monolingual and
Spanish–English bilingual participants. Within this task, participants were asked to classify words
appearing in white on the basis of valence (POSITIVE or NEGATIVE) via a key press, or to classify those words on the basis of color (blue or green) also via a key press. The words were either
emotion words (e.g., HAPPY; ANXIOUS) or emotion-laden words (e.g., DREAM; SHARK).
While monolingual participants viewed stimuli solely in English, bilingual participants viewed
words in both English and Spanish. Are Affective Simon Effects modulated by language in bilingual
speakers? If so, one might expect that effects would be larger and more robust in the native
language of a bilingual – the language touted by some to be the more ‘emotional’ of a bilingual’s
two languages (cf. Altarriba, 2003). Second, should similar congruency effects emerge for different
classes of emotional stimuli (emotion vs. emotion-laden)? Perhaps valence moderates processing
of these words in different ways so as to predict stronger or weaker effects depending on word
class. It could be the case that Affective Simon Effects emerge more strongly for emotional
stimuli since they label an emotion, as compared to emotion-laden stimuli that seem to represent
a ‘mediated’ effect between the word itself and the emotion that is elicited by the word’s connotations and denotations. Further, how are all of these effects influenced or affected by valence itself
(i.e., positive or negative)? It may be the case that negative words exert stronger influences, even
in the Affective Simon Task, and show larger congruency effects than positive words, as most of
the earlier data mentioned (e.g., emotional Stroop effects) seem to be more highly modulated by
negative vs. positive stimuli. The following four experiments were designed to examine this
series of related questions.
Experiment 1: Monolingual/emotion-laden words
Methods
Participants. Fifty-seven participants from the University at Albany, State University of New York
participated in partial fulfillment of the psychology research requirement. All were native English
speakers, with normal or normal-to-corrected vision, no reading disorders, and no color blindness.
Each participant correctly identified six plates from Ishihara’s (1939) test of color blindness in
order to verify that they could correctly distinguish colors. In addition, all participants completed
the Beck Depression Inventory II (BDI-II; Beck, Steer, & Brown, 1996) and the State-Trait Anxiety
Inventory (STAI) (Spielberger, Gorsuch, Lushene, Vagg, & Jacobs, 1983). Due to the fact that we
were interested in examining the influence of emotional stimuli in a nonclinical population, only
those participants who scored lower than 13 on the BDI and lower than 43 on the STAI were
included in the study.
Materials. Ten positive (e.g., PARTY; HUG; KITTEN) and 10 negative (e.g., JAIL; WAR;
GRAVE) emotion-laden words were selected as critical stimuli (see Appendix). All items were
matched on word length and frequency (Kučera & Francis, 1967). All items were determined to be
high in arousal (Bradley & Lang, 1999), a trait that typically characterizes an item as being emotional. Positive and negative words were matched on this attribute (see Table 1 for stimuli attributes). Valence significantly differed across the two types, with negative words being low in
valence and positive words being high in valence (according to the Bradley & Lang norms for
English). Ten practice words (five negative and five positive) that were not included in the critical
item list were also selected.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
315
Altarriba and Basnight-Brown
Table 1. Stimuli attributes for Experiments 1–4
Emotion-laden
English
Frequency
Length
Arousal
Valence
Spanish
Frequency
Length
Arousal
Valence
Emotional
Positive
Negative
Positive
Negative
FRIEND
54.40
5.50
5.48
7.38
AMIGO
141.50
6.00
5.65
7.52
FUNERAL
45.30
5.40
5.64
2.56
ENTIERRO
109.30
7.20
6.54
1.96
HAPPY
32.20
7.20
5.67
7.65
ALEGRE
39.40
7.60
5.87
7.17
ANGRY
37.80
6.30
6.15
2.66
ENOJADO
46.90
7.20
6.43
2.25
Note: English items were only presented in Experiments 1 and 2, while English and Spanish items were presented in
Experiments 3 and 4.
Procedure. Each participant completed the experiment individually. Instructions were provided in
a written format, but verbal instructions were also provided in order to verify that the participant
fully understood the task. Several practice trials were administered in order to familiarize the participant with the task. In the first practice block, individuals were told that they would be presented
with words in white only. Their task was to press the ‘P’ key if the word was positive and the ‘Q’
key if the word was negative. In the second practice block, they were instructed that words would
appear in either blue or green ink. Individuals were instructed to press the ‘P’ key if the word was
‘blueish’ in color, and to press the ‘Q’ key if it was ‘greenish’ in color. These color–key designations were counterbalanced across participants. The shades of blue and green that were used were
the same as those selected in the original Affective Simon Task work (De Houwer, 2003). These
specific shades were very close in color; therefore, the participants were told that the colors might
be hard to discriminate. Each time the participant made an error, a red ‘+’ appeared in the center of
the screen.
Once the two practice blocks were completed, one experimental block was presented. This final
block contained all 20 critical items. Positive and negative words were presented one at a time, in
random order. Each word appeared once in each of the three colors: white, blue, and green, with
the colors also appearing in random order. Therefore, a total of 60 trials (i.e., 20 words in each of
the three colors) were presented in this final block. The participants were instructed that if a word
appeared in white they should pay attention to the valence of the word and press the key that corresponded to the positive/negative evaluative response. If the word appeared in blue or green, they
should make a decision based on color. As noted earlier, instructions were counterbalanced across
participants, such that half of the individuals were instructed to press the ‘P’ key when an item was
green and the ‘Q’ key when the item was blue, and the other half to press the ‘P’ key when the item
was blue and the ‘Q’ key when the item was green.
The program for this experiment was created using E-prime software (Schneider, Eschman, &
Zuccolotto, 2002). All experiments were presented on a computer screen interfaced with an
IBM-PC computer. Words appeared in each of the three colors on a black background. Each trial
began with a fixation (+) appearing in the center of the screen. The critical item followed the fixation and remained on the screen until a response was made, or for a maximum of 3000 ms. The
inter-trial interval was 1500 ms.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
316
International Journal of Bilingualism 15(3)
Table 2. Mean response times in milliseconds (standard deviations in parentheses) and error rates for
Experiment 1 (monolingual emotion-laden)
Congruent
Error rate
Incongruent
Error rate
Effect
Positive
Negative
723 (147)
3
770 (137)
8
+47*
735 (136)
1
778 (144)
11
+43*
Note: Bold indicates significant effect.
Results and discussion
Analyses of variance (ANOVAs) were conducted separately on the latency and error rate data for
both participants and items. For participants, valence (positive, negative) and congruency (response
congruent, incongruent) were within-subjects factors. For items, congruency was a within-subjects
factor and valence was a between-subjects factor. For each individual participant, data were
trimmed so that any value 2.5 standard deviations above or below the mean was replaced with the
appropriate value (–/+ 2.5 standard deviations from the mean). A summary of mean reaction times
and error rate data appears in Table 2. Data from participants and items that fell below the 60 percent accuracy criterion were removed from the analyses. One participant and no items were deleted.
The mean BDI score for these participants was 4.13, and the mean STAI score was 26.28.
ANOVAs conducted on the response latency data revealed a main effect of congruency for both
participants and items, F1(1, 55) = 23.778, MSE = 5,242, p < .001; F2(1, 18) = 20.365, MSE =
21,265, p < .001, indicating that the congruent trials were responded to faster than the incongruent
trials. Furthermore, when the positive and negative effects were examined individually, planned
comparisons revealed significant congruency effects for both types of words: positive valence,
t(55) = 3.247, p < .01; negative valence, t(55) = 2.75, p < .01. The interaction between valence and
congruency was not significant, revealing that the +47 ms effect for positive words was not different
from the +43 ms effect for negative words. For the accuracy measure, results from the ANOVA
revealed a main effect of congruency, F1(1, 55) = 48.921, MSE = .007, p < .001; F2(1, 18) = 39.373,
MSE = .011, p < .001, as well as an interaction between congruency and valence, F1(1, 55) = 5.905,
MSE = .010, p < .05; F2(1, 18) = 7.528, MSE = .001, p = .01. Thus, overall, the results indicated that
emotion-laden words produce an Affective Simon Effect and that the effect is similar across positively and negatively valenced items in monolingual speakers. Thus, the congruency that arises out of
responding to positive stimuli, as well as negative stimuli, that has been demonstrated in the past
wherein emotion stimuli have been mixed with emotion-laden stimuli, has now been demonstrated
with a group of items that is uniformly emotion-laden. Experiment 2 examines this same effect with
a set of stimuli that is uniformly emotional – all words that label emotional states.
Experiment 2: Monolingual/emotional words
Methods
Participants. Fifty-two participants from the University at Albany, State University of New York
participated in partial fulfillment of the psychology research requirement. All were native English
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
317
Altarriba and Basnight-Brown
Table 3. Mean response times in milliseconds (standard deviations in parentheses) and error rates for
Experiment 2 (monolingual emotional words)
Congruent
Error rate
Incongruent
Error rate
Effect
Positive
Negative
712 (117)
4
729 (120)
6
+17
683 (119)
3
770 (118)
8
+87
Note: Bold indicates significant effect.
speakers, with normal or normal-to-corrected vision, no reading disorders, and no color blindness.
The same screening measures that were used in Experiment 1 (color blindness test; BDI; STAI)
were administered to those who participated in Experiment 2.
Materials. Criteria for selecting emotional stimuli for Experiment 2 were the same as those used
in Experiment 1. Ten positive (e.g., HAPPY; HOPEFUL; THANKFUL) and 10 negative (e.g.,
UPSET; REJECTED; JEALOUS) emotional words were selected (see Appendix). All items
matched on frequency, length, and arousal. Valence differed between the two groups, with positive
words being high in valence and negative words being low in valence. The emotional items used
in this experiment tended to be adjectives, which differed from the emotion-laden words used in
Experiment 1, which were all nouns. Once again, 10 words (five positive and five negative) that
were not in the critical stimuli list were selected to serve as practice items.
Procedure. The procedure was identical to that used in Experiment 1.
Results and discussion
The data were trimmed and analyzed in the same way as they were in Experiment 1. Once again,
one participant and no items were deleted because they fell below the accuracy criterion. The mean
BDI score was 4.38 and the mean STAI score was 30.29.
Results from the ANOVA conducted on the response latency data revealed a main effect of congruency, F1(1, 50) = 23.269, MSE = 5,913, p < .001; F2(1, 18) = 33.170, MSE = 791, p < .001,
indicating that overall, congruent trials were responded to significantly faster than incongruent trials.
However, planned comparisons conducted on each of the word types separately revealed a significant
congruency effect only for the negatively valenced words, t(51) = 5.646, p < .001 (see Table 3).
Furthermore, the interaction between congruency and valence was also significant, indicating that the
significant +87 ms effect for negative words significantly differed from the nonsignificant +17 ms
effect for positive words, F1(1, 50) = 14.166, MSE = 4,422, p < .001; F2(1, 18) = 16.984, MSE =
13,443, p < .01. Results from the accuracy data revealed a main effect of congruency, F1(1, 50) =
9.471, MSE = .006, p < .01; F2(1, 18) = 15.251, MSE = .001, p < .001, indicating that the participants
were more accurate in responding to congruent trials as compared to incongruent trials. Thus, overall,
it appears that unlike the effects reported in Experiment 1 with emotionally laden stimuli, emotion
word stimuli produce Affective Simon Effects only for negative words. Words that label positive
emotions do not seem to produce the congruency effects that are typical of this type of Simon effect.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
318
International Journal of Bilingualism 15(3)
In order to compare the patterns observed for the two different types of emotional stimuli,
analyses were conducted across the two experiments. These analyses revealed an interaction
between congruency and valence, F1(1, 105) = 4.737, MSE = 6,188, p < .05, as well as a triple
interaction between congruency, valence, and experiment, F1(1, 105) = 5.889, MSE = 5,264, p =
.01. More specifically, comparisons conducted on the congruency effects across experiments
revealed that the difference between the negative effects (+43 ms vs. +87 ms) was significant,
t(105) = 1.837, p < .05. One possible reason for this difference across Experiments 1 and 2 is the
notion that emotion-laden words often refer to concrete objects or items that are potentially somewhat more imageable than emotion words. Thus, the emotion-laden words are a bit more comparable to the items that have been used in the earlier research literature wherein words like
BUTTERFLY and CANCER have predominated (see e.g., De Houwer, 2003). The current findings
would indicate, therefore, that the earlier work might have yielded results that were based on the
overall influence of the inclusion of emotion-laden words within the stimulus sets rather than stemming from emotion words as a general stimulus category. That is, it could be that emotion words
produce Affective Simon Effects only when they are negatively valenced but not when they are
positively valenced. More is said about the distinctions between emotional word types and valence,
particularly with reference to comparable findings within the emotional Stroop literature (see e.g.,
Sutton & Altarriba, 2008), in the General discussion.
Experiment 3: Bilingual/emotion-laden words
Methods
Participants. Thirty-two Spanish–English bilinguals from the University at Albany, State
University of New York participated. All were native Spanish speakers with no reading disorders,
normal or normal-to-corrected vision, and no color blindness.
At the end of the experiment, each participant completed a language history questionnaire
(LHQ; Altarriba & Mathis, 1997). The LHQ was used to assess the bilinguals’ self-rated skills in
three areas: spoken, written, and conversational skills in both Spanish (L1) and English (L2).
Language history data obtained from these participants can be found in Table 4. Planned comparisons conducted on the language history questionnaire data revealed that the bilinguals rated their
English writing skills as (marginally) better than their Spanish writing skills, t(31) = 1.872, p = .07.
All other comparisons were not significantly different. Although Spanish was the first language
learned for this population, written skill in English was most likely rated as being stronger than in
Spanish because these individuals received most of their formal education from a young age in
English-speaking environments, and they were also currently attending an English-speaking
university.
Materials. The positive and negative emotion-laden stimuli used in Experiment 3 were the same
as those used in Experiment 1 with the monolingual participants. The 20 critical items were translated into Spanish so that there were a total of 40 items (10 negative English words and their 10
translations in Spanish, 10 positive English words and their translations in Spanish). The Spanish
translations of the positive emotion-laden words were matched on frequency (Sebastián, Martí,
Carreiras, & Cuetos, 2000) and length to the Spanish translations of the negative emotion words
(see Table 1). All items were high in arousal, and positive and negative items differed in valence in
Spanish (Redondo, Fraga, Padrón, & Comesaña, 2007).1
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
319
Altarriba and Basnight-Brown
Table 4. Language history data and self-reported ratings for the Spanish–English bilinguals who participated
in Experiments 3 and 4
AoA Spanish
AoA Spanish reading
AoA English
AoA English reading
Spanish spoken comp.
Spanish written skill
Spanish speaking
English spoken comp.
English written skill
English speaking
Experiment 3
Experiment 4
2.1
6.0
5.7
6.2
9.1
8.6
8.9
9.5
9.5
9.5
2.3
7.5
4.2
5.6
8.6
8.2
8.8
9.7
9.5
9.7
Procedure. The procedure was the same as that used in Experiments 1 and 2. The only difference
was that the participants were presented with two blocks of critical items, one block containing the
20 English words and another block containing the 20 Spanish words. The order of the language
blocks was counterbalanced, and all participants were presented with the two practice blocks (i.e.,
(1) white words only, (2) color words only) in the appropriate language before each experimental
block (the English word block was preceded by English practice trials and the Spanish word block
was preceded by Spanish practice trials). Within each language block, each of the critical items
appeared in random order, in each of the three colors (white, blue, and green).
Results and discussion
All data were trimmed and analyzed in the same way as they were in previous experiments. No
items and one participant were removed from the analyses as a result of high error rates. The mean
BDI and STAI scores for this group were 8.10 and 33.10, respectively.
Replicating the pattern found in the previous two experiments, the results from the ANOVA
conducted on the response latency data revealed a main effect of congruency, F1(1, 30) = 38.152,
MSE = 9,004, p < .001; F2(1, 36) = 43.375, MSE = 2,040, p < .001. Overall, participants were
faster in the congruent condition as compared to the incongruent condition. The interaction between
language and valence was significant for participants and marginally significant for items, F1(1, 30)
= 4.179, MSE = 7,597, p = .05; F2(1, 36) = 2.224, MSE = 3,275, p = .14. The difference in mean
reaction time between positive and negative items was much larger in the Spanish condition vs. the
English condition (+38 ms vs. +7 ms). Accuracy data also revealed a main effect of congruency,
F1(1, 30) = 21.380, MSE = .009, p < .001; F2(1, 36) = 47.062, MSE = .001, p < .001. Participants
were more accurate on congruent trials as compared to incongruent trials.
Planned comparisons revealed significant congruency effects in English for both the positive,
t1(30) = 3.961, p < .001, t2(9) = 6.657, p < .001, and negative items, t1(30) = 3.811, p < .001; t2(9)
= 3.608, p < .006, (see Table 5). Congruency effects in Spanish were significant for participants
and marginally significant for items for the negative items, t1(30) = 2.147, p < .05; t2(9) = 1.805,
p = .10, and significant for both participants and items for the positive items, t1(30) = 3.191, p < .01;
t2(9) = 3.211, p < .01 (see Table 5). Overall, the effects appear to be more robust in English than in
Spanish, for emotion-laden words (see Table 5). However, none of the individual congruency
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
320
International Journal of Bilingualism 15(3)
Table 5. Mean response times in milliseconds (standard deviations in parentheses) and error rates for
Experiment 3 (bilingual emotion-laden)
English
Congruent
Error rate
Incongruent
Error rate
Effect
Spanish
Congruent
Error rate
Incongruent
Error rate
Effect
Positive
Negative
759 (197)
1
852 (223)
6
+93
762 (141)
1
835 (171)
7
+73
772 (157)
2
844 (158)
7
+72
816 (142)
2
876 (215)
8
+60
Note: Bold indicates significant effect.
effects were significantly different from each other. Thus, it appears that the effects of valence and
congruency were equal across language conditions for the present group of bilingual participants,
as they processed emotion-laden stimuli. Again, these effects may indicate that emotion-laden
words, often having concrete and imageable referents, are not strongly influenced by the language
in which they appear. However, one question remains, would language (i.e., L1 vs. L2) exert an
influence in the Affective Simon Task when processing words that label emotions? Experiment 4
was designed to examine this question.
Experiment 4: Bilingual/emotional words
Methods
Participants. Thirty-four Spanish–English bilinguals from the University at Albany, State
University of New York participated. All bilinguals were from the same population as those
who participated in Experiment 3 (see Table 4). Planned comparisons conducted on the
language history questionnaire data from these participants revealed that they rated themselves
as significantly better in all three skill areas in English as compared to Spanish: spoken comprehension, t(32) = 4.215, p < .001; written, t(32) = 5.388, p < .001; speaking, t(32) = 4.417,
p < .001. Independent samples t-tests conducted to compare the participants in this experiment
to those in the previous bilingual experiment revealed no significant differences between the
two bilingual groups (i.e., in age of acquisition [AoA] of each language, reading AoA, and selfreported ratings).
Materials. The positive and negative emotional stimuli used in this experiment were the same as
those used in Experiment 2. However, as performed in Experiment 3, the 20 emotion words were
translated into Spanish so that there were a total of 40 items (10 negative English items and their
10 translations in Spanish, 10 positive English items and their translations in Spanish). All positive
and negative Spanish translations were matched on frequency, length and arousal (see Table 1).
Procedure. The procedure used was identical to that used in Experiment 3.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
321
Altarriba and Basnight-Brown
Table 6. Mean response times in milliseconds (standard deviations in parentheses) and error rates for
Experiment 4 (bilingual emotional words)
English
Congruent
Error rate
Incongruent
Error rate
Effect
Spanish
Congruent
Error rate
Incongruent
Error rate
Effect
Positive
Negative
766 (140)
2
814 (163)
8
+48
764 (159)
2
836 (172)
8
+72
848 (166)
7
861 (164)
5
+13
817 (140)
3
872 (161)
9
+55
Note: Bold indicates significant effect.
Results and discussion
All data were trimmed and analyzed in the same way as they were in the previous experiments.
One item and one participant were removed from the analyses. The mean BDI and STAI scores for
these bilinguals were 5.77 and 31.67, respectively. The data for Experiment 4 appear in Table 6.
As in all the previous experiments, the data revealed a main effect of congruency, F1(1, 32) =
5.426, MSE = 9,551, p < .05; F2(1, 35) = 21.073, MSE = 2,287, p < .001, as well as a main effect
of language, F1(1, 32) = 15.412, MSE = 36,473, p < .001; F2(1, 35) = 19.746, MSE = 2,872, p <
.001. The ANOVAs conducted on the accuracy data also revealed a main effect of congruency,
F1(1, 32) = 15.042, MSE = .008, p < .001; F2(1, 35) = 23.939, MSE = .002, p < .001, as well as an
interaction between congruency and valence, F1(1, 32) = 4.899, MSE = .007, p < .05; F2(1, 35) =
6.450, MSE = .010, p < .01. The difference in mean error rate for the congruent condition vs. the
incongruent condition was larger for negative items than for positive items (5.5 vs. 2.0).
Planned comparisons revealed significant congruency effects in English for both the negative,
t1(32) = 2.696, p < .05; t2(9) = 2.782, p < .05, and positive conditions, t1(32) = 2.148, p < .05; t2(9)
= 3.801, p < .01. In Spanish, the effect was significant only for the negative condition, t1(32) =
2.463, p < .05; t2(8) = 2.790, p < .05 (see Table 6). None of the significant effects were reliably
different from each other. Thus, the results were in line with those reported in Experiment 2 for
emotion words, as processed by monolingual participants. Namely, while negative items produced
an Affective Simon Effect, positive items did not, in English for the monolingual participants and
in Spanish for the bilingual participants. Although the current bilinguals learned Spanish as their
first language, it is likely the case that language dominance ‘switched’ for these participants, as has
been reported in other studies (see e.g., Basnight-Brown & Altarriba, 2007; Heredia, 1997), due to
the fact cited earlier that they are much more proficient in English than in Spanish and that they
more often read English words and process English words more readily than Spanish words. Thus,
English, in essence, was the more ‘emotional’ language for the current participants owing to their
greater experience in processing words in that language, and for this reason, it appears that their
Spanish results paralleled those of the English monolingual participants (Experiment 2) specifically when processing words that label emotions. In effect, for positive emotion words, Spanish
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
322
International Journal of Bilingualism 15(3)
was the ‘dominant language’ for the current bilinguals in the same way that English was the ‘dominant language’ for the monolinguals, in the second experiment. Thus, these two conditions produced highly similar results. Interestingly, the emotional component of the words presented in
Spanish in Experiment 4 did ‘capture’ the attention of these participants resulting in longer reaction
times when processing emotion words in Spanish; this finding is again likely due to the fact that
these words are less often seen ‘in print’ than their English counterparts, for the present group of
bilingual participants. More will be said regarding relative use and exposure to bilinguals’ two
languages and implications for reading and language processing in general, in the next section.
Once again, comparisons across the two bilingual experiments were performed in order to see if
emotion-laden vs. emotion words were processed differently in the L1 and L2. The results revealed
that the magnitude of the negative effects did not differ statistically. However, the difference in the
positive effects for the emotion (+13 ms) vs. emotion-laden (+72 ms) words was marginally significant, t(62) = 1.875, p = .06. Thus, emotion-laden words seemed to more readily produce Affective
Simon Effects than emotion words do, for bilinguals, as well as for monolinguals. More will be said
about the possible factors related to this finding, as well as about the relative differences between the
two types of emotional stimuli compared above, in the following section.
General discussion
Ultimately, what is being measured in the Affective Simon Task are the implicit beliefs about target
words in terms of their valence (positive or negative) and emotional quality (labeling of an emotion
or emotion-laden word). The results of the current experiments revealed that color word performance was facilitated for positive concepts when the correct response was also the response
assigned to positive white words, specifically for emotion-laden words. This finding was true for
monolinguals in English, as well as for bilinguals in English – that is, the dominant language for
each group of participants. A similar congruency effect emerged for negative concepts, for both
emotion words and emotion-laden words for monolinguals and bilinguals alike, across all language
conditions. In fact, for monolingual speakers, this latter effect was twice as strong for emotion
words as compared to emotion-laden words. Thus, Affective Simon Effects have proven to be
robust across emotion word type and across participant group, for negatively valenced items in the
current experiments. Thus, performance on the Affective Simon Task appears to be moderated by
emotion word type and by language dominance, as well. Overall, the data provide further evidence
that affective coding in language can influence subsequent responses in seemingly irrelevant tasks.
Below, the current results are discussed in greater detail in relation to what is known about language
processing differences across monolinguals and bilinguals, as well as what is known about the
processing of emotional stimuli.
Monolingual emotion word processing
An examination of the present set of data in Experiments 1 and 2 indicates that for monolingual,
English-speaking participants, emotion-laden words prompt different patterns of performance in
the Affective Simon Task as compared to words that simply label an emotion. Specifically, while
the emotion word data produce a strong congruency effect for negative words, the effect is not
significant for positive words. Up until now, the results reported regarding this effect in the literature were similar for both positive and negative stimuli (see e.g., De Houwer, 2003; De Houwer et
al., 2001). However, the previously published data stemmed from studies in which both emotionladen and emotion word stimuli were intermixed within stimulus sets. In Experiment 1 of the
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
Altarriba and Basnight-Brown
323
current article, the typical results were found with significant effects for positive and negative
stimuli that were similar. However, the stimuli were all emotion-laden in this experiment. Thus,
these findings would suggest that the inclusion of emotion-laden words in previous work may have
contributed to finding Affective Simon Effects for both positively and negatively valenced items.
The important point in the current data set is that the two types of emotional concepts (i.e., emotion
words vs. emotion-laden words) examined here do not operate equally, even when controlled for
arousal and other lexical characteristics. The current work constitutes one of the first empirical
demonstrations to directly compare these two word types in a timed, cognitive task. The reason
positive and negative emotion words moderate different behavior in word processing demonstrations are not clearly understood (cf. Schmidt & Saari, 2007). However, it is thought that negative
stimuli capture increased attention, are more likely to be fully elaborated when processed, and
enjoy a level of item distinctiveness that promotes their processing in ways that differ from positive
and neutral stimuli (Schrauf & Sanchez, 2004). The characteristics that elicit a threat-related
response from negative words appear to moderate their effects in a variety of tasks, such as the
emotional Stroop task mentioned earlier (McKenna & Sharma, 1995, 2004; Pratto & John, 1991).
Further, recall that an analysis comparing the Affective Simon Effect for negatively valenced stimuli in Experiment 1 vs. Experiment 2 yielded a significant difference in favor of stronger effects for
emotion word items as compared to emotion-laden items.
Interestingly, the Affective Simon Effect is much less pronounced for emotion-laden words, for
monolingual speakers in the current article, though it appears for both positive and negative items.
This finding combined with those just mentioned above indicates that while emotion-laden words
carry emotional components, they may not be as pronounced as they are within words that truly
label an emotion. That is, emotion-laden words, as noted earlier, may be seen as a type of ‘mediated’
emotion concept – a concept through which one comes to elicit an emotion based on one’s exposure
and past associations with a particular emotion-laden word. Level of association may moderate
this effect, implicitly as well as explicitly, and may actually prompt an Affective Simon Effect. In
fact, Pessoa and Ungerleider (2004) in their neuroimaging studies of attentional processing of
emotion-laden words conclude that these words are often processed automatically and without
explicit attention. Clearly, more research is needed on the relative similarities and differences
across different emotion word types in order to fully understand how they modulate performance
across a range of tasks.
Bilingual emotion word processing
Experiments 3 and 4 of the current study focused on the processing of emotion words vs.
emotion-laden words for bilingual speakers in English and in Spanish. As noted earlier, it
appears that again, as with monolingual speakers, emotion-laden words produced Affective
Simon Effects in all language and valence conditions, whereas emotion words did not. This
overall finding was confirmed statistically via cross-experimental analyses that were reported
earlier (i.e., Experiment 3 vs. Experiment 4). That is, it might be said that one can more likely
provide demonstrations of these effects using emotion-laden stimuli than using stimuli that
label emotional states. In fact, both positive and negative stimuli produced similar congruency
effects in English and in Spanish. Thus, regardless of language dominance, Affective Simon
Effects emerged for both languages for the current set of bilingual participants for emotionladen words (see also Duyck & De Houwer, 2008, who observed similar semantic Simon effects
in bilinguals for nonemotional stimuli). Again, as long as these words are associated to emotion
states and elicit emotional components rather automatically and implicitly, they contribute to
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
324
International Journal of Bilingualism 15(3)
finding an effect using the current task. Further, an examination of the data for bilinguals in
Experiment 4 indicates that while emotion word stimuli produced significant Affective Simon
Effects for negative words in both languages, the effects emerged only for positive words in
English, but not in Spanish. Though accounts of emotion word processing have indicated that
these words carry more emotional ‘weight’ in the native language, or the first-learned language
(cf. Altarriba, 2003), this statement must be qualified such that dominance may actually be
more predictive of behavior using emotion word stimuli than age of acquisition (Dewaele,
2008). That is, for the current bilingual participants, it is likely the case that English is more
dominant, for reasons stated earlier in this article. Thus, although these individuals can recognize positive, emotion word stimuli in Spanish, activation of those words does not necessarily
lead to differential responding in congruent vs. incongruent conditions in the current task. Note
also that this one condition – Spanish, emotional words (see Table 6) – produced, on average,
the slowest reaction times across all bilingual conditions. Thus, it is likely that while these
words captured attention, post-access processing led to the elimination of an effect in the current variant of the Simon Task.
Attentional control and bilingual advantage
The use of a variant of the Simon Task in the current study also allows one to explore the role of
executive control and inhibitory processes in both monolinguals and bilinguals. Green’s Inhibitory
Control model (1998), proposed roughly a decade ago, suggests that the active use of two languages
involves the same attentional resources that are used in tasks that are not linguistic in nature. In
recent years, much of the support for this assumption has come from research where the Simon
Task has been used in bilingual populations (see Bialystok, 2001, 2009, for reviews). Because the
Simon Task involves the presentation of congruent and incongruent trials within a single block,
executive control processes involved in task switching can be studied using this paradigm.
Bialystok and colleagues have reported an advantage for bilinguals in the Simon Task, such that
bilinguals have less difficulty with the task and thus, typically produce smaller Simon effects
(Bialystok, 2006; Bialystok, Craik, Klein, & Viswanathan, 2004). It is thought that this advantage
is the result of bilinguals being able to better inhibit irrelevant information since they consistently
do this when managing two languages.
Although bilingualism does appear to influence general cognitive abilities, it is possible that it
can do so in ways that are disadvantageous, as well as advantageous. In the current study, bilinguals
do not appear to produce smaller effects as compared to monolinguals. However, it is important to
point out that the task used in the current study is a variant of the traditional Simon effect, one that
does not typically involve linguistic processing. The current task, in contrast, does require the use
of language skills and the activation of the semantic meaning behind each word presented. It is
possible that this difference is responsible for the current outcome. After all, bilinguals have been
found to be slower in lexical decision tasks (see Basnight-Brown & Altarriba, 2007), slower to
name pictures (Kaushanskaya & Marian, 2007), and more prone to tip-of-the tongue states during
memory retrieval (Gollan & Acenas, 2004). In addition, Treccani, Argyri, Sorace, and Della Sala
(2009) recently examined whether bilinguals have better inhibitory control when using a negative
priming paradigm. Negative priming is used to describe the slower responses to stimuli that occur
when information from one trial (prime trial) is inhibited, but on the following trial (probe trial)
that information becomes relevant (Tipper, 1985). Treccani et al. (2009) reported larger negative
priming effects for bilinguals, as compared to monolinguals, suggesting that bilingualism can
influence executive control in different ways.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
Altarriba and Basnight-Brown
325
Conclusions
Performance on the Affective Simon Task appears to be moderated by emotion word type and by
language dominance, as well. Overall, the data provide further evidence that affective coding in
language can influence subsequent responses in seemingly irrelevant tasks. Further, on occasion,
negative words produced an effect that was twice as large as their emotion-laden counterparts. Do
emotion words exert ‘purer’ effects for negatively valenced items, while emotion-laden words
produced only a ‘mediated’ effect? Future research should be aimed toward examining these word
types and their influence on processing across a wide variety of tasks.
Acknowledgments
Portions of the current research were presented at the 49th annual meeting of the Psychonomic Society in
Chicago, Illinois, November 2008 and at the 7th International Symposium on Bilingualism (ISB7), Utrecht
University, Utrecht, The Netherlands, July 2009. We would like to thank Jaclyn Bleski, Megan DeStefano,
and Toni-Ann Funk for their assistance with data collection on this project.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit
sectors. Note
1. As suggested by a reviewer, arousal and valence values were obtained from the Redondo et al. (2007)
Spanish word norms. Due to the fact that several of the Spanish items used in the current study were not
present in the database (23% were not found), the mean arousal and valence values are based only on
those words that were available.
References
Altarriba, J. (2003). Does cariño equal ‘liking’? A theoretical approach to conceptual nonequivalence between
languages. International Journal of Bilingualism, 7, 305–322.
Altarriba, J. (2006). Cognitive approaches to the study of emotion-laden and emotion words in monolingual
and bilingual memory. Bilingual Education and Bilingualism, 56, 232–256.
Altarriba, J., & Bauer, L. M. (2004). The distinctiveness of emotion concepts: A comparison between emotion, abstract, and concrete words. American Journal of Psychology, 117, 389–410.
Altarriba, J., Bauer, L. M., & Benvenuto, C. (1999). Concreteness, context-availability, and imageability
ratings and word associations for abstract, concrete, and emotion words. Behavior Research Methods,
Instruments, and Computers, 31, 578–602.
Altarriba, J., & Canary, T. M. (2004). Affective priming: The automatic activation of arousal. Journal of
Multilingual and Multicultural Development, 25, 248–265.
Altarriba, J., & Mathis, K. M. (1997). Conceptual and lexical development in second language acquisition.
Journal of Memory and Language, 36, 550–568.
Basnight-Brown, D. M., & Altarriba, J. (2007). Differences in semantic and translation priming across
languages: The role of language direction and language dominance. Memory & Cognition, 35,
953–965.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). BDI-II: Beck Depression Inventory manual (2nd ed.).
Boston, MA: Harcourt, Brace.
Bialystok, E. (2001). Bilingualism in development: Language, literacy, and cognition. New York: Cambridge
University Press.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
326
International Journal of Bilingualism 15(3)
Bialystok, E. (2006). Effect of bilingualism and computer video game experience on the Simon task. Canadian
Journal of Experimental Psychology, 60, 68–79.
Bialystok, E. (2009). Bilingualism: The good, the bad, and the indifferent. Bilingualism: Language and
Cognition, 12, 3–11.
Bialystok, E., Craik, F. I. M., Klein, R., & Viswanathan, M. (2004). Bilingualism, aging, and cognitive control: Evidence from the Simon Task. Psychology and Aging, 19, 290–303.
Bradley, M. M., & Lang, P. J. (1999). Affective norms for English words (ANEW). Gainesville, FL: The NIMH
Center for the Study of Emotion and Attention, University of Florida.
De Houwer, J. (2003). The extrinsic Affective Simon Task. Experimental Psychology, 50, 77–85.
De Houwer, J., Crombez, G., Baeyens, F., & Hermans, D. (2001). On the generality of the Affective Simon
Effect. Cognition & Emotion, 15, 189–206.
De Houwer, J., & Eelen, P. (1998). An affective variant of the Simon paradigm. Cognition & Emotion, 12,
45–61.
Dewaele, J.-M. (2006). Expressing anger in multiple languages. In A. Pavlenko (Ed.), Bilingual minds:
Emotional experience, expression, and representation (pp. 118–151). Clevedon: Multilingual Matters.
Dewaele, J. M. (2008). The emotional weight of I love you in multilinguals’ languages. Journal of Pragmatics,
40, 1753–1780.
Duyck, W., & De Houwer, J. (2008). Semantic access in second-language visual word processing: Evidence
from the semantic Simon paradigm. Psychonomic Bulletin and Review, 15, 961–966.
Eilola, T. M., Havelka, J., & Sharma, D. (2007). Emotional activation in the first and second language.
Cognition & Emotion, 21, 1064–1076.
Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R. (1986). On the automatic activation of
attitudes. Journal of Personality and Social Psychology, 50, 229–238.
Gollan, T. H., & Acenas, L.-A. R. (2004). What is a TOT? Cognate and translation effects on tip-of-the
tongue states in Spanish–English and Tagalog–English bilinguals. Journal of Experimental Psychology:
Learning, Memory, & Cognition, 30, 246–269.
Green, D. W. (1998). Mental control of the bilingual lexico-semantic system. Bilingualism: Language and
Cognition, 1, 67–81.
Harris, C. L., Ayçiçeği, A., & Berko-Gleason, J. (2003). Taboo words and reprimands elicit greater autonomic
reactivity in a first than in a second language. Applied Linguistics, 24, 561–579.
Harris, C. L., Berko-Gleason, J., & Ayçiçeği, A. (2005). When is a first language more emotional?
Psychophysiological evidence from bilingual speakers. In A. Pavlenko (Ed.), Bilingual minds: Emotional
experience, expression, and representation (pp. 257–283). Clevedon: Multilingual Matters.
Heredia, R. R. (1997). Bilingual memory and hierarchical models: A case for language dominance. Current
Directions in Psychological Science, 6, 34–39.
Huijding, J., & De Jong, P. (2005). A pictorial version of the Extrinsic Affective Simon Task. Experimental
Psychology, 52, 289–295.
Ishihara, S. (1939). Tests for colour-blindness. Tokyo: Kanehara & Co., Ltd.
Kaushanskaya, M., & Marian, V. (2007). Bilingual language processing and interference in bilinguals:
Evidence from eye tracking and picture naming. Language Learning, 57, 119–163.
Kučera, H., & Francis, W. N. (1967). Computational analysis of present-day American English. Providence,
RI: Brown University.
McKenna, F. P., & Sharma, D. (1995). Intrusive cognitions: An investigation of the emotional Stroop task.
Journal of Experimental Psychology: Learning, Memory, & Cognition, 21, 1595–1607.
McKenna, F. P., & Sharma, D. (2004). Reversing the emotional Stroop effect reveals that it is not what it
seems: The role of fast and slow components. Journal of Experimental Psychology: Learning, Memory,
& Cognition, 30, 382–392.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
Altarriba and Basnight-Brown
327
Pavlenko, A. (2008a). Emotion and emotion-laden words in the bilingual lexicon. Bilingualism: Language
and Cognition, 11, 147–164.
Pavlenko, A. (2008b). Structural and conceptual equivalence in the acquisition and use of emotion words in a
second language. The Mental Lexicon, 3, 91–120.
Pessoa, L., & Ungerleider, L. G. (2004). Neuroimaging studies of attention and the processing of emotionladen stimuli. Progress in Brain Research, 144, 171–182.
Pratto, F., & John, O. P. (1991). Automatic vigilance: The attention-grabbing power of negative social information. Journal of Personality and Social Psychology, 61, 380–391.
Redondo, J., Fraga, I., Padrón, I., & Comesaña, M. (2007). The Spanish adaptation of ANEW (Affective
Norms for English Words). Behavior Research Methods, 39, 600–605.
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 user’s guide. Pittsburgh, PA: Psychology
Software Tools Inc.
Schrauf, R., & Sanchez, J. (2004). The preponderance of negative emotion words in the emotion lexicon: A
cross-generational and cross-linguistic study. Journal of Multilingual and Multicultural Development, 25,
266–284.
Sebastián, N., Martí, M. A., Carreiras, M., & Cuetos, F. (2000). LEXESP: Una base de datos informatizada
del español. Barcelona: Servicio de Publicaciones de la Universitat de Barcelona. Retrieved from http://
www.ub.es/ccil/en
Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs. G. A. (1983). Manual for the StateTrait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologist.
Sutton, T. M., & Altarriba, J. (2008). Representing emotion words in the mental lexicon: A new look at the
emotional Stroop effect. The Mental Lexicon, 3, 19–46.
Sutton, T. M., Altarriba, J., Gianico, J. L., & Basnight Brown, D. M. (2007). The automatic access of emotion:
Emotional Stroop effects in Spanish–English bilingual speakers. Cognition & Emotion, 21, 1077–1090.
Tipper, S. P. (1985). The negative priming effect: Inhibitory priming by ignored objects. Quarterly Journal of
Experimental Psychology, 37A, 571–590.
Tipples, J. (2001). A conceptual replication and extension of the Affective Simon Effect. Cognition &
Emotion, 15, 705–710.
Treccani, B., Argyri, E., Sorace, A., & Della Sala, S. (2009). Spatial negative priming in bilingualism.
Psychonomic Bulletin and Review, 16, 320–327.
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
328
International Journal of Bilingualism 15(3)
Appendix
Emotion-laden words
death
criminal
fault
flood
funeral
fever
thief
shark
snake
rape
dream
flower
friend
gift
heaven
health
wish
reward
puppy
miracle
Emotional words
muerte
delictivo
culpa
diluvio
entierro
fiebre
ladrón
tiburón
serpiente
violación
sueño
flor
amigo
regalo
cielo
salud
deseo
recompensa
cachorro
milagro
fear
angry
depressed
anxious
frustrated
hurt
lonely
mad
nervous
troubled
happy
honest
hopeful
wealthy
triumphant
proud
powerful
loyal
thoughtful
thankful
Downloaded from ijb.sagepub.com at United States International University on June 2, 2015
miedo
enojado
deprimido
ansioso
desbaratado
herido
aislado
enojado
tímido
apenado
alegre
honrado
esperanzador
rico
triunfante
orgulloso
poderoso
fiel
atento
agradecido