Spanish colour terms

Posted: 13 Agustus 2010 in colour language
Social Science Information
DOI: 10.1177/0539018409106199
Social Science Information 2009; 48; 421
Cristina Soriano and Javier Valenzuela
Spanish colour terms
Emotion and colour across languages: implicit associations in
The online version of this article can be found at:
Published by:
On behalf of:
Maison des Sciences de l’Homme
Additional services and information for Social Science Information can be found at:
Email Alerts:
Downloaded from at Universite de Geneve on August 31, 2009
Special issue: The language of emotion – conceptual and cultural issues
Numéro spécial: Le langage de l’émotion –
questions conceptuelles et culturelles
Cristina Soriano & Javier Valenzuela
Emotion and colour across languages:
implicit associations in Spanish colour terms
Abstract. This study explores the reasons why colour words and emotion words are
frequently associated in the different languages of the world. One of them is connotative
overlap between the colour term and the emotion term. A new experimental methodology, the
Implicit Association Test (IAT), is used to investigate the implicit connotative structure of
the Peninsular Spanish colour terms rojo (red), azul (blue), verde (green) and amarillo
(yellow) in terms of Osgood’s universal semantic dimensions: Evaluation (good–bad),
Activity (excited–relaxed) and Potency (strong–weak). The results show a connotative
profile compatible with the previous literature, except for the valence (good–bad) of some of
the colour terms, which is reversed. We suggest reasons for both these similarities and
differences with previous studies and propose further research to test these implicit
connotations and their effect on the association of colour with emotion words.
Key words. Colour – Connotation – Emotion – IAT – Psycholinguistics – Semantic differential
– Semantics – Spanish
Résumé. Cette étude explore les raisons pour lesquelles les termes qui qualifient les
couleurs et ceux qui qualifient les émotions sont fréquemment associés dans plusieurs langues.
L’une de ces raisons est le chevauchement des connotations entre le terme qualifiant la couleur
et celui qui qualifie l’émotion. On utilise une méthodologie expérimentale nouvelle, le test
d’association implicite (Implicit Association Test—IAT) pour investiguer la structure
implicite des connotations des termes de couleur en Espagnol Péninsulaire, rojo (rouge), azul
(bleu), verde (vert) and amarillo (jaune) en termes des dimensions sémantiques universelles
d’Osgood: Evaluation (Bon—Mauvais), Activité (Excité-s-Relâché), Puissance (Fort—
Faible). Les résultants de l’étude montrent un profil de connotations compatible avec la
© The Author(s), 2009. Reprints and permissions:
Social Science Information, 0539-0184; Vol. 48(3): 421–445; 106199
DOI: 10.1177/0539018409106199
Downloaded from at Universite de Geneve on August 31, 2009
422 Social Science Information Vol 48 – no 3
litérature existante, à l’exception de la valence (Bon—Mauvais) de certains termes de
couleur, qui est inversée. Les auteurs émettent des hypothéses pour expliquer à la fois ces
similaritiés et ces différences par rapport aux études déjà connues et proposent d’entreprendre
d’autres recherches pour tester ces connotations implicites et leur effet sur l’association termes
de couleur—termes d’émotions.
Mots-clés. Couleur — Connotation — Emotion — IAT — Psycholinguistique —
Différentiel sémantique — Sémantique — Espagnol
Colour and emotion in the languages of the world
Emotions and colour terms are frequently associated in the languages of the
world. In English, for example, the term blue is related to sadness (to feel
blue).1 In German yellow is associated to envy (Gelb vor Neid sein, ‘to be
yellow with envy’), although in Spanish and English envy is green (cf.
Shakespeare’s green-eyed monster). Yellow in English is associated to cowardice
instead (to be yellow, yellow-bellied, to have a yellow streak). Anger
is associated to red in many languages, where we find expressions like the
German Rot sehen (‘see red’) or the French rouge de rage (‘red with anger’).
In Thai, however, a ‘body turning green’ belongs to an angry person. In
Italian one can also be verde di bile, rabbia or collera (‘green with bile,
anger or rage’; Philip, 2006: 82). An interesting case is Lithuanian, where
different degrees of anger are not expressed through different shades of red,
as happens in English (cf. flushed/pink/red/scarlet with anger), but through
different colours: white (for controlled anger), red (‘normal’ anger), blue
(very intense anger) and black (extremely intense anger) (Sirvydė, 2007). A
shade of white can also be related to anger in English (he was livid), but then
it hardly refers to ‘controlled anger’. In French, fear is associated to green
(vert de peur, ‘green with fear’), whereas green in Russian is the colour of
utter boredom (скука зеленая, ‘green boredom’) and ‘melancholy-cumyearning’
(тоска зеленая, ‘green toska’; cf. Wierzbicka, 1992). World
wide, black/dark has bad connotations, and in many languages it is directly
related to depression/pessimism (e.g. Spanish verlo todo negro, ‘see everything
black’) or to anger (black mood, Spanish estar negro, ‘be black’, i.e.
be fed up). Pink, on the contrary, has positive connotations and is widely
linked to optimism or happiness (e.g. Ukranian бачити щось в рожевому
світлі, ‘to see everything in pink light’, or English tickled pink). In the
African language Fante, ‘being jealous’ is expressed by the term anibre,
which literally means ‘eye-red’. Interestingly, inimooi in Dagbani – another
African language – also means ‘eye-red’, but refers to worry/anxiety instead
(Dzokoto & Okazaki, 2006). The heart has a colour too. In Dagbani, ‘white
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 423
heart’ is the emotion term for happiness, and in Zulu the angry heart is said
to be red (Taylor & Mbense, 1998: 202–3).
Why is this so? Colours per se do not designate anything emotional, so
why should they be associated to specific emotions or emotion terms? There
are at least four possible complementary reasons for these pairings: the
associations are based on metonymic thinking, they are based on metaphoric
thinking, colour perception creates specific emotional reactions, and colour
terms and emotion terms share the same connotative structure in the language.
Let us discuss each of these reasons in some more detail.
The first reason might be that the colour–emotion association reflects
something about the physiology of the emotional experience. This is the
case for expressions like red with anger or to be yellow. In both cases, the
colour refers to a body fluid that intervenes (or is believed to intervene) in
the experience of the emotion at stake. In the case of anger the fluid is blood,
which rushes to the neck and face areas when we feel outraged. In the case
of cowardice/fear the fluid is bile, given that the expression was coined at a
time when the liver was believed to be the seat of courage (e.g. Allan, 2009).
In both cases – and in any other linguistic expression of this sort – the
expressions can be considered to be metonymic. A linguistic metonymy is
an expression in which we refer to some entity by means of another one to
which it is related. Typical metonymic expressions reflect part-and-whole
relationships (‘there are some good heads in this department’) or cause-andeffect
relationships (‘he was livid at his son’s disobedience’, anger being the
cause of the lividness in the face). Cognitive linguists would claim that this
linguistic behaviour is an expression of systematic thought patterns called
conceptual metonymies (e.g. Kövecses & Radden, 1998; Barcelona, 2003),
like referring to the effects of something, instead of to the cause itself
(EFFECT FOR CAUSE). Therefore conceptual metonymies like EFFECT
FOR CAUSE can be considered to motivate many of the associations
between emotion and colour words. Given that emotional physiology is
roughly the same across cultures, one might expect all languages to pick up
the same colours for a given emotion. This, however, is not the case for at
least two reasons. First, because different cultures can choose different colour
aspects of a multifaceted reality (e.g. lividness vs blushing in anger).
Second, because colour choice often relies on culture-specific theories of
the causes and physiological consequences of emotions, as in the case of
envy, which in the Western world was thought to be accompanied by a rush
of bile – which is yellow-green in colour (Ogarkova, 2007: 114).
A second reason for the pairing of colour and emotion in a language may be
metaphorical thinking. Conceptual metaphors, like metonymies, are defined as
systematic (some claim automatic) patterns of thought (e.g. Lakoff, 1993). By
virtue of conceptual metaphors we represent one conceptual domain – the
Downloaded from at Universite de Geneve on August 31, 2009
424 Social Science Information Vol 48 – no 3
target – in terms of another – the source – using the knowledge we have of the
source to construe the target. For example, expressions like that’s a clear argument,
I see your point, that sheds some light on the whole issue or don’t be so
obscure instantiate in English a systematic conceptual representation of ‘understanding’
in terms of ‘seeing’ (UNDERSTANDING IS SEEING), which finds
linguistic realization in many languages of the world (e.g. English, Spanish,
Basque, French, Dutch, Italian, Greek and Romanian; Valenzuela & Soriano,
2008). It is also metaphorical thinking, and not physiology, which seems to lie
behind the pairing of black/white and envy in the Russian expressions ‘black
envy’ and ‘white envy’. One of them refers to ‘envy that is not malicious’,
while the other means ‘ill-wishing envy’. It is not difficult to guess which is
which. This conceptual association between bad and dark/black (BAD IS
BLACK/DARK) and good and white/bright (GOOD IS WHITE/BRIGHT)
is ubiquitous in language at large, not only in the realm of emotion words
(e.g. dark soul, black Monday, white lie, dark intentions, bright character, clear
mind, black mood, sombre personality, etc.). It also seems to be automatic. In
an interesting series of experiments, Meier and his colleagues (Meier, Robinson
& Clore, 2004; Meier et al., 2007) showed a compulsory systematic association
between brightness and valence: light colours automatically elicit a positive
evaluation of objects (whereas dark colours elicit a negative one), and
positive evaluation biases perceptual judgements towards brightness (whereas
negative evaluations make people judge colours as darker).2
A third reason for the pairing of colour and emotion words in a language
may be that colours themselves elicit emotional reactions. Goldstein (1942)
explicitly suggested that the body undergoes physiological reactions to
colour that are reflected in psychological experiences and functioning. For
example Elliot and colleagues (2007) have found experimental evidence that
red hinders performance on achievement tasks. However, Mehta & Zhu
(2009) have also recently shown that red enhances performance on a detailoriented
task. The two results need not be contradictory: it seems that exposure
to red enhances a sense of alertness and urgency, which interferes in
some cognitive tasks and helps in others. It is not clear what specific emotion
one would feel when presented with red, but the above evidence suggests that
the emotion should at least involve arousal. This is consistent with the sorts
of emotion words people intuitively pair with the term red – like anger, fear
or jealousy (Hupka et al., 1997) – and the emotions they associate with the
colour itself – like love (red hearts, red roses) or lust (red has been found to
enhance men’s attraction to women; Elliot & Niesta, 2008). It is also consistent
with the emotion words we explicitly observe paired with the colour term
red across languages, like anger (e.g. English, Spanish, Italian, German) and
embarrassment (e.g. Italian, English, Spanish, French).
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 425
A fourth possible reason for the association of colour and emotion in a
language may be that both colour words and emotional words share the
same (or very similar) connotative structure. Connotation is a fairly vague
term in semantics. In this article we define it as the set of meaning aspects
of a word that go beyond its mere denotation or reference. For example a
definition of red in a dictionary may say that red is ‘the hue of the long-wave
end of the visible spectrum, evoked in the human observer by radiant energy
with wavelengths of approximately 630 to 750 nanometres’.3 But surely
the term red means something more for any native speaker of English. The
word carries much more than a simple reference to the chromatic aspect of
things. Red, like any other colour term, has an emotional meaning as well.
Emotional meaning seems to be mediated by an association of colour terms
with situations or objects in real life where that coloration is present (e.g.
Leech, 1981; Allan, 2007). It could be the colouring of natural things
(blood-red, grass-green, sky-blue) or that of cultural artefacts (black robes
in mourning, red lights to signal alert, baby girls dressed in pink), and both
can lead to colour-term connotation (more or less standardized) and even to
culturally sanctioned colour symbolism.
Even though colour connotation ultimately depends on associations in
experience, some of which are the same for all human beings, the specific
value of the connotation may vary from language to language, from culture
to culture and even from person to person. An association between blood and
red is probably at stake in any culture of the world, but the specific ‘meaning’
that the term red inherits from blood depends on how blood is construed in
the first place: a sign of danger because it signals injury? A sign of victory
over enemies? A sign of violence? A sign of life and vitality? A sign of sexual
maturity? The interpretations of blood are potentially endless, and so are the
connotations of red inherited from such association (e.g. Allan, 2009).
Yet evidence has been found which argues against a hypothetical arbitrariness
of colour connotation. The specific concepts with which colours
are associated (e.g. BLOOD, WATER, SUN) and the affective value (dangerous,
beneficial, desirable, harming) of those concepts for a given person
or a given culture may vary greatly, but the ‘types’ of value themselves seem
to be fairly systematic: cross-culturally they are all related to a general evaluation
of things as ‘good’ or ‘bad’, ‘strong’ or ‘weak’, and ‘active’ or ‘passive’
(Adams & Osgood, 1973; Osgood, May & Miron, 1975). These semantic
dimensions – called Evaluation, Potency and Activity respectively – are not
exclusive of colour. Osgood and his colleagues (Osgood, Suci & Tannenbaum,
1957) discovered them in a number of concepts using the Semantic
Differential Technique, and their findings have also been replicated in
studies involving different types of stimuli (e.g. Osgood, Suci &
Downloaded from at Universite de Geneve on August 31, 2009
426 Social Science Information Vol 48 – no 3
Tannenbaum, 1957; Snider & Osgood, 1969; Mehrabian, 1972; Mehrabian &
Russell, 1974; Shaver et al., 1987; Valdez & Mehrabian, 1994).
The Osgoodian dimensions are relevant, because they concretize the
fourth reason why colour and emotion words may be associated in a language.
While the specific emotion terms associated to specific colour terms
can vary considerably across languages, in the pairs existing within each
language, colour term and emotion word should have a congruent dimensional
profile. D’Andrade & Egan came to the same conclusion when trying
to explain the surprising overlap in a study in which they asked English and
Tzeltal speakers to match emotion terms and colour chips:
At present we feel the most likely explanation of this phenomenon involves a special type
of semantic mediation based on the factors uncovered by Osgood’s semantic differential
technique. … The association of colours and emotions may come about through the use of
the semantic differential dimensions as a bridging network. That is, if a colour chip is
reacted to as ‘good’ and ‘strong’, then the emotion term which is associated to the colour
chip is an emotion which is also ‘good’ and ‘strong’. (D’Andrade & Egan, 1974: 62)
Our study: implicit connotations in Spanish colour terms
Adams & Osgood (1973) studied the connotative structure of the English
colour terms black, white, grey, red, yellow, green and blue and their equivalent
colour terms in 20 other languages of the world using the Semantic
Differential Technique. An important aspect to be taken into account is that
Adams & Osgood’s methodology requires participants to make explicit ratings
on the position of a given word (e.g. red) in a scale between two poles
(e.g. between the terms fresh and stale, or hot and cold). Factor analysis
generates the dimensions, and the specific ratings are then used to calculate
the value of each word on the dimensions. Therefore the methodology captures
explicit judgements about the connotative loadings of terms. But for
most colour emotion psychologists ‘the activation of the colour association,
as well as its influence on affect, cognition, and behaviour, is viewed as
occurring without the individual’s conscious awareness or intention’ (Elliot
et al., 2007: 156; emphasis added). Therefore it is pertinent to investigate
whether the same values in the semantic dimensions can be found at an
unconscious level as well.
We decided to use an Implicit Association Task (IAT), which, according
to the first proponents of this methodological paradigm, is useful ‘for measuring
evaluative associations that underlie implicit attitudes’ (Greenwald,
McGhee & Schwartz, 1998: 1464). This methodology has been employed to
measure evaluative conditioning (Mitchell, Anderson & Lovibond, 2003),
prejudice and prototypical evaluative judgements (White vs Black American
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 427
names – Greenwald, McGhee & Schwartz, 1998; Turkish and East Asian
populations – Gawronski, 2002), and it has proven useful in the prediction
of non-verbal (McConnell & Leibold, 2001) and sexual behaviour (Marsh,
Johnson & Scott-Sheldon, 2001).
The study was carried out in Peninsular Spanish, a language variety that
was not included in Adams & Osgood’s study (1973), although other varieties
of Spanish were. If implicit and explicit associations are the same, no big
differences should be found in the dimensional profile of the colour terms.
The specific terms selected for our study were rojo, azul, verde and
amarillo (the Spanish counterparts of red, blue, green and yellow respectively).
There were two reasons for this choice. First, the relevance of the
chromatic phenomenon they refer to, as red, blue, green and yellow are
primary colours. Second, the fact that they (and their closest equivalents in
the various languages of the world) have been hypothesized to be ‘basic’
colour terms by Berlin & Kay (1969). One of the features of basic colour
terms is that they are the first to appear across languages. It seems that red-,
yellow-, green- and blue-like terms are the first to be coined after a general
distinction between ‘dark/black’ and ‘light/white’. In other words, if a language
has only two colour terms, they tend to reflect the black/dark and
white/light distinction. If a third colour term exists, it is the equivalent of
red. After them, counterparts of either green or yellow are found. A blue-like
term appears next. Since all languages with colour terms seem to make these
distinctions, the existence of general universal RED, BLUE, GREEN and
YELLOW categories can be hypothesized. Given the order of appearance of
the terms, these categories are among the most widespread colour categories
across languages, and the colour terms that instantiate them reflect a more
universal partition of the spectrum than the colour terms in the later stages
of the Berlin & Kay model, like brown, purple, pink, orange or grey.
However, the fact that all languages seem to differentiate a part of the
chromatic spectrum through labels like blue does not mean that the conceptualization
of English blue and Spanish azul, for example, are the same.
Colour terms may differ in terms of their exact denotation, as well as in connotation.
As regards denotation, for example, Russian offers two different
lexemes to designate the same chromatic phenomenon that English labels as
blue, and the same is true for Italian, where there is blu (dark blue) and azzuro
(sky blue). Colour terms may also differ in connotation. Adams & Osgood
(1973: 144–5) found that yellow-like terms, for example, tend to be positively
marked across languages, but have a very negative evaluation in Hong Kong
Cantonese, where the colour yellow is associated to pornography.
So what are the unconscious affective connotations of Peninsular
Spanish rojo, azul, amarillo and verde in terms of ‘goodness–badness’,
Downloaded from at Universite de Geneve on August 31, 2009
428 Social Science Information Vol 48 – no 3
‘power–weakness’ or ‘excitement–relaxation’? To the best of our knowledge,
no experimental studies have been conducted so far in Peninsular Spanish to
answer this question. But there is ample psychological, linguistic and anthropological
research on the general field of colour–emotion associations in
other languages to help us generate some hypotheses. In what follows we
report on this data and the working hypotheses that stem from them.
(a) Positive and negative Evaluation
Of the three dimensions proposed by Osgood and colleagues, Evaluation is the
most controversial in the realm of colour semantics. Adams & Osgood (1973)
concluded that the semantic blue–green region is more highly evaluated than
the red–yellow region, which is closer to the neutral point (1973: 144).
Converging evidence can be found in Goldstein (1942), who claimed that the
colours green and blue are experienced as quieting and agreeable. For Wexner
(1954), the colour concept BLUE was associated to features like ‘secure/
comfortable’ and ‘tender/soothing’, which imply pleasure, in addition to low
activity. Clarke & Costall (2008) also recorded positive evaluations for green
and blue in a semi-structured interview with British students. Finally, in a
study on the connotations of English colour terms in figurative expressions,
Allan (2009) has recently observed that blue is hardly ever used in offensive
language, but in euphemistic expressions. All in all, the general colour concept
BLUE, and to a lesser extent GREEN, seem to be positively valenced.4
RED, however, is a more ambiguous case. Adams & Osgood (1973) locate
it near the neutral point, while various studies on emotion–colour association
suggest either a positive or a negative valence, depending on the concepts it
is related to. It is reported to have a positive valence when associated to love,
fertility, happiness or excitement (Wexner, 1954; Murray & Deabler, 1957;
Jobes, 1962; Pecjak, 1970; Soldat, Sinclair & Mark, 1997). The valence is
negative when associated to features like ‘contrary’ or ‘hostile’ (Odbert,
Karwoski & Eckerson, 1942; Murray & Deabler, 1957; Schaie, 1961).
The general concept YELLOW is located near the neutral point according
to Adams & Osgood (1973), who explain exceptions in terms of culture-specific
circumstance (like yellow in Hong Kong). Hupka and colleagues
(1997) view YELLOW as a slightly more negative concept, and quote
D’Andrade & Egan (1974) and Oyama, Tanaka & Haga (1963) as examples
of empirical evidence that terms with negative connotations across cultures are
systematically associated with colours in the yellow–red end of the spectrum.
Because of the above, our starting hypothesis is that the semantics of
VERDE and AZUL will be more positively marked than ROJO and
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 429
AMARILLO, although ROJO might have a positive evaluation as well. All
in all, VERDE and AZUL are expected to be the two concepts most strongly
related to the Evaluation dimension.
(b) High and low Potency
Adams & Osgood (1973) identify two colours as most closely related to the
Potency dimension: RED as the most potent and YELLOW as the least.
Neither BLUE nor GREEN is clearly defined in terms of Potency in the
data. As in the previous dimension, support for their claims is also reported
in a large number of earlier studies (1973: 145). But more recent ones have
also provided converging evidence. Hill & Barton (2005), for example, have
recently demonstrated that wearing red correlates with a greater probability
of winning in sport competitions. The reason, in their opinion, is that red
enhances dominance.
In the light of this research, we expect ROJO to be strong and AMARILLO
to be weak. Neither VERDE nor AZUL is expected to be significantly
related to the Potency dimension.
(c) High and low Activity
Of the four colours under investigation, only RED was identified as closely
related to Activity by Adams & Osgood (1973: 146). The literature on
colour–emotion, however, also mentions GREEN and BLUE as related to
Activity. According to Goldstein (1942), the colour green and (to a lesser
degree) blue are experienced as quieting. In Wexner’s study (1954) BLUE
was associated with ‘tender/soothing’. In their interviews Clarke & Costall
report many subjects associating green and blue with features like ‘calm’
and ‘peaceful’ (2008: 407). In all cases blue and green appear as low-activity
colours. In general a frequent contention in the colour–emotion tradition is
that red and other long-wavelength colours are more active than blue and the
short wavelength ones.
(d) Summary of hypotheses
A summary of the above findings renders the following hypotheses concerning
the connotative structure of the Spanish colour categories we aim to
Downloaded from at Universite de Geneve on August 31, 2009
430 Social Science Information Vol 48 – no 3
• ROJO is expected to be powerful and excited (and maybe positive).
• AMARILLO is expected to be weak.
• VERDE and AZUL are expected to be positive (and possibly relaxed).
Again, these expectations are based on research carried out on explicit
judgements and conscious associations between colour shades or colour
terms and emotion terms (or concepts), as observed in several languages.
The extent to which these predictions match the implicit connotations of
Peninsular Spanish colour terms is what we aim to investigate.
Method and materials
As required in the IAT methodology, in our study the four selected colour
terms were treated as categories (ROJO, AZUL, AMARILLO and VERDE),
for which seven terms were chosen as most representative among their most
frequent synonyms or near-synonyms in the online Spanish corpus ‘Corpus
del Español’ (over 100 million words, available at http://www.corpusdelespanol.
org/) The terms and an approximate translation into English are
shown in Table 1. The three semantic dimensions were also operationalized
as shown in Table 2.
rojo red azul blue
colorado red, blushed añil indigo
carmesí crimson celeste pale blue
encarnado red marino navy blue
escarlata scarlet turquesa turquoise
púrpura purple azulón deep blue
granate maroon azulado bluish
amarillo yellow verde green
rubio blond cetrino yellowish green
dorado golden verde césped grass green
amarillento yellowish verdoso greenish
áureo golden esmeralda emerald
gualda yellow verde oliva olive green
ámbar amber aceitunado olive coloured
Table 1
Spanish colour terms in the four colour categories
Red, Blue, Yellow and Green
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 431
The IAT is a computer-based categorization task that measures conceptassociation
strength. Participants have to classify a set of related words
(e.g. blue, indigo, turquoise) into their category (e.g. BLUE) by pressing a
computer key. Sometimes the key they have to press is also valid for a second
category (e.g. GOOD). It has been shown that implicit associations between
two concepts result in speeded responses when the two concepts are combined
in a single group for categorization purposes. So, for example, if BLUE
is associated to GOOD, it will be easier (faster) to classify a word like indigo
into an ad hoc category BLUE-or-GOOD than into a category BLUE-or-
BAD. Similarly if BLUE and GOOD are perceived to be congruent concepts,
people will make fewer errors assigning indigo to the constructed category
BLUE-or-GOOD than to BLUE-or-BAD. In sum, if GOOD and BLUE are
unconsciously related, more speed and accuracy will be expected when classifying
a word like indigo into a constructed BLUE-or-GOOD category.
Table 2
Spanish dimension terms for Evaluation,
Potency and Activity
Dimensions Terms
Evaluation Positivo: Positive: Negativo: Negative:
positivo positive negativo negative
bueno good malo bad
agradable pleasant perjudicial harmful
grato nice nocivo harming
placentero pleasurable dañino damaging
beneficioso beneficial desagradable unpleasant
provechoso advantageous destructivo destructive
Potency Poderoso: Powerful: Débil: Weak:
poderoso powerful débil weak
influyente influential sumiso submissive
fuerte strong sometido subdued
potente potent obediente obedient
asertivo assertive inhibido inhibited
dominante dominant retraído shy
dirigente leader achantado bashful, coward
Activity Excitado: Excited: Relajado: Relaxed:
tenso tense relajante relaxing
intenso intense tranquilo quiet
activo active sereno serene
estimulante stimulating flojo loose, slack
enérgico energetic tenue tenuous, faint
excitante exciting leve light, slight
vibrante vibrant inactivo inactive
Downloaded from at Universite de Geneve on August 31, 2009
432 Social Science Information Vol 48 – no 3
Any IAT investigates two categories (e.g. ROJO and AZUL) and an
evaluative dimension (e.g. Activity). The test comprises five stages. In the
first stage (see Figure 1, stage one), examples of the two investigated categories
(e.g. azul, añil, etc., for AZUL; rojo, escarlata, etc., for ROJO) are
presented in random order in the centre of a computer screen, one by one.
On the right and left top corners of the screen the names of the categories
are shown, and subjects have to classify the words appearing in the centre
as belonging to one or the other category by pressing a key situated on the
right- or on the left-hand side of the keyboard. In our experiment, subjects
used a Spanish keyboard and pressed ‘m’ for the category on the right and
‘z’ for the category on the left.
In the second stage (Figure 1, stage two), participants see examples of the
two poles in the evaluative dimension under study (e.g. words like excitante,
Stage Five
Stage One
Stage Two
Stage Three
Stage Four
IAT design – examples of stages one to five
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 433
enérgico, etc., for EXCITADO; and relajante, sereno, etc., for RELAJADO)
and they have to classify them as belonging to one of the two poles (e.g.
EXCITADO and RELAJADO) presented on the right and left top corners of
the screen.
In a third stage (Figure 1, stage three) both a category (e.g. ROJO) and
a dimensional pole (e.g. EXCITADO) appear together on one side of the
screen, and the other category (e.g. AZUL) and the other dimensional pole
(e.g. RELAJADO) on the opposite side. Examples of both categories (azul,
añil, rojo, escarlata, etc.) and both dimension poles (excitante, enérgico,
relajante, sereno, etc.) are presented in the centre of the screen and subjects
have to classify them correctly into their respective categories.
The fourth stage is identical to the first, except for the fact that the order
of the categories is reversed (see Figure 1, stage four).
In the fifth and final stage (see Figure 1, stage five), categories and poles
are again paired, but in the opposite combination from stage three.
During the experiment, subjects are informed of categorization mistakes
by an error message on the screen (for example, if they ascribe the colour
term escarlata to the category AZUL instead of ROJO). Both reaction time
(i.e. how long it takes them to decide that escarlata is a member of the category
ROJO) and accuracy (i.e. if they assign the term escarlata to the correct
category, ROJO, or mistakenly press the key for the AZUL category) are
recorded throughout the test.
The crucial stages in the design are stages three and five. When a category
(e.g. category ROJO) and one of the poles of an evaluative dimension (e.g.
pole EXCITADO in the dimension Activity) are strongly associated, their
pairing for categorization purposes is supposed to yield shorter reaction
times in the task, as well as fewer errors in the subject’s categorization
The four colour categories and the three dimensional poles were crossed to
generate all possible combinations. First we combined each colour category
with the others (6 combinations) and then each colour pair was combined
with the three dimensions (6 × 3 = 18 combinations). The order of presentation
of the two members in each colour pair and of the two poles in the
dimensions was counterbalanced throughout the experiment.
One hundred and fifteen native speakers of Spanish (most of them
students at the University of Murcia), ranging in age from 20 to 46 (mean
Downloaded from at Universite de Geneve on August 31, 2009
434 Social Science Information Vol 48 – no 3
age 24.2), participated in the study as volunteers or in exchange for
course credit. Each condition was observed by at least eight subjects.
Our results can be analysed in terms of dimensions and colours. We start
by describing the profile of each dimension. Then we analyse the specific
colour results and their connection with the literature on colour–emotion
The overall results of the study can be found in Tables 3 and 4. Table 3
reports the mean reaction times (RT) in milliseconds and their statistical
significance. The results for the Evaluation dimension have not been
included because they did not reach significance. Table 4 presents the accuracy
rates in the categorization task and their statistical significance.
In the analysis of errors, no statistically significant effects were found for
colour or dimension only, but an analysis of variance showed that their
interaction was significant (F (3,123) = 4.729, p = .004). A look at the results
shows that AMARILLO, ROJO and VERDE are perceived as positive by
our participants (p < .005), while AZUL is evaluated as negative (p < .05).5
No statistically significant differences were found in the analysis of the
RT, but the effects showed the same direction as the error analysis, which
allows us to discard a trade-off effect.
The greatest effects in terms of accuracy were found for AMARILLO (the
most positive colour) and AZUL (the most negative). These results partially
disconfirm our expectation that VERDE and AZUL would be the two colour
categories most closely associated to the Evaluation dimension, since
VERDE is not. In fact, the RTs in the positive and negative pairings for
VERDE are technically the same (a difference of 6 ms). AZUL, on the other
hand, is significantly related to Evaluation in our error analysis, but it shows
effects in the opposite direction from that expected.
Also disconfirmed was our expectation that the semantics of VERDE
and AZUL would be more positively marked than those of ROJO and
AMARILLO, since AZUL was negatively marked and there were no statistically
significant differences between VERDE, ROJO and AMARILLO in
our analysis of errors. In a colour-by-colour comparison of RTs, a paired
t-test revealed statistically significant oppositions between ROJO–AZUL
(p = .05) and AMARILLO–VERDE (p < .05).
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 435
Table 3
Mean Reaction Times (RT) ANOVA in the categorization task
ms F d.f. p
Potency: Pw W
rojo 777 889 17.875 1, 113 0,000*
amarillo 890 815 7.778 1, 113 0,006*
verde 809 807 n.s.
azul n.s.
Activity: E R
rojo 747 911 38.09 1, 103 0,000*
amarillo 831 759 12.741 1, 103 0,000*
verde 844 815 12.741 1, 103 0,000*
azul 840 766 12.741 1, 103 0,000*
Note. RTs ANOVA: Pw = powerful, W = weak, E = excited, R = relaxed
Table 4
Accuracy rates in the categorization task
% F d.f. p
Evaluation: P N
Rojo 0.94 0.91 10.504 1, 123 0,002*
Amarillo 0.94 0.90 10.504 1, 123 0,002*
Verde 0.93 0.91 10.504 1, 123 0,002*
Azul 0.91 0.95 5.897 1, 123 0,017*
Potency: Pw W
Rojo 0.94 0.89 8.985 1, 113 0,003*
Amarillo 0.90 0.93 4.509 1, 113 0,036*
Verde 0.90 0.93 4.509 1, 113 0,036*
Azul 0.91 0.91 n.s.
Activity: E R
Rojo 0.94 0.87 19.628 1, 103 0,000*
Amarillo 0.92 0.94 5.085 1, 103 0,026*
Verde 0.91 0.93 5.085 1, 103 0,026*
Azul 0.91 0.93 5.085 1, 103 0,026*
Note. Error rates ANOVA: P = positive, N = negative, Pw = powerful, W = weak, E = excited,
R = relaxed
No independent effect of colour or Potency was found, but their interaction
was statistically significant both in the analysis of errors (F (3,113) = 4.506,
p < .005) and in the RTs (F (3,113) = 8.363, p < .0005).
Downloaded from at Universite de Geneve on August 31, 2009
436 Social Science Information Vol 48 – no 3
In terms of Potency, ROJO stands out as the most powerful colour and
AMARILLO as the weakest one. The pattern is statistically significant in
terms of errors and RTs. These results confirm our hypotheses and coincide
with Adams & Osgood’s (1973) cross-cultural study. Additionally, in a
colour-by-colour comparison of RTs, a paired t-test revealed statistically
significant (or almost significant) oppositions between ROJO and all the
other colours: AMARILLO (p < .01) – the clearest case – VERDE (p < .05)
and AZUL (p = .051).
Our additional expectation that VERDE and AZUL would not be clearly
defined in terms of Potency seems to find (at least partial) support too: in the
analysis of errors, VERDE stands as significantly associated to WEAKNESS,
but no effect is found in the analysis of the RTs. As for AZUL, as expected,
none of the analyses revealed any effects.
As before, an analysis of variance showed that the interaction of colour and
Activity was significant in terms of accuracy (F (3,103) = 8.216, p < .001)
and RTs (F (3,103) = 17.74, p < .001).
The differences in terms of colour were also clear. AMARILLO, VERDE
and AZUL are significantly related to low Activity (RELAXATION), while
ROJO is significantly associated to EXCITEMENT. Both error analysis and
RTs reveal this pattern. The results confirm our hypothesis as well as Adams
& Osgood’s (1973) observations for RED. They also support our expectation
that AZUL and VERDE would be low-activity colour categories.
In a colour-by-colour comparison of RTs, a paired t-test revealed statistically
significant oppositions between ROJO and the rest of the colours
(AMARILLO: p < .001), VERDE: p < .05, and AZUL: p < .005), and
between AMARILLO and VERDE (p < .05).
ROJO (+E, +P, +A)
According to our results, ROJO is implicitly construed as a positive, highpotency,
high-activity colour. This is mostly consistent with the literature
(except for the valence, where there is no agreement). Our expectation that
ROJO would be positive was confirmed, but the strength of the association does
not match the results observed in Adams & Osgood’s semantic work. This markedness
may be the result of cultural preference (see ‘Summary and discussion’).
The term rojo is associated to ira (anger) in Peninsular Spanish in
the expression rojo de ira (‘red with anger’), and the colour is symbolically
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 437
used to represent the emotion as well (Soriano, 2005). We currently lack
information about the unconscious connotative structure of the emotion word
ira,6 but we have some information about the conscious dimensional structure
of its English counterpart anger. Fontaine and his colleagues (Fontaine
et al., 2007) have shown that anger is characterized by negative Evaluation,
high Potency and high Arousal (i.e. Activity). Except for Evaluation, the connotative
structure of the colour and the emotion word anger coincides.
High arousal is also a feature of the word fear in English, Dutch and
French (e.g. Fontaine et al., 2007), another emotion word cross-linguistically
associated to the general colour category RED (Hupka et al., 1997).
Our results show AMARILLO as characterized by positive evaluation, low
activity and low potency. These results confirm our hypothesis on the
Potency dimension. The positive evaluation, located near the neutral point
by Adams & Osgood, is in our case significantly stronger, which may be a
case of cultural preference (see ‘Summary and discussion’).
The word amarillo is not conventionally associated to any particular emotion
word in Spanish. We also lack semantic information about the dimensional
structure of Spanish emotion words. Nevertheless, the English noun
contentment has been found to be characterized by positive valence, low
arousal and medium potency (Fontaine et al., 2007). Therefore, a possibly
compatible term for amarillo in Spanish might be contento (‘content’).
In other languages, like German, Italian, Dutch or Turkish, yellow is associated
to envy (Osgood et al., 1975; Hupka et al., 1997). However, this
association is not motivated by a shared connotative structure, but rather by
metonymic reference to the colour of bile, believed to accompany the experience
of envy/jealousy and to give the complexion a greenish-yellowish
taint (Ogarkova, 2007).
VERDE (+E, ±P, –A)
For our participants, in comparison to the rest of the colour categories
VERDE is a positively tainted colour, although not so much as AMARILLO
or ROJO. On the contrary, the association with relaxation is strong. Both the
positive Evaluation and the Activity values confirmed our expectations and
match Adams & Osgood’s (1973) observations with the Semantic Differential
Downloaded from at Universite de Geneve on August 31, 2009
438 Social Science Information Vol 48 – no 3
The difference between VERDE and AMARILLO in our data seems to be
one of degree (cf. Philip, 2006, for a similar analysis in her study of Italian
colour terms.) Both colour concepts share positive Evaluation and low
Activity, but in terms of Potency, AMARILLO is always significantly weak,
while VERDE is significantly weak only when we look at error rates, but not
at RTs. Therefore VERDE seems slightly more powerful than AMARILLO.
The Spanish term verde is conventionally related to envidia (‘envy’) in
the language through the expression verde de envidia (‘green with envy’).
However, the motivation for this expression, as we saw for yellow above,
stems from the assumed participation of bile (a yellow-greenish fluid) in the
experience of envy.
AZUL (–E, ±P, –A)
The hypothesized positive valence of AZUL was disconfirmed by our data,
since it shows a significant association with the negative pole in comparison
to the other three colour categories under study (p < .05). AZUL is also
implicitly associated to relaxation.
While the relaxation was expected, the negative evaluation was not. Very
few scholars have observed a negative valence for BLUE. Soldat, Sinclair
& Mark (1997) hypothesized a relationship between BLUE and sadness, a
negative emotion, perhaps influenced by the fact that English has a lexicalized
association between the two (to have the blues, to feel blue). Philip
suggests that the negative overtones of Italian blu (dark blue) may be motivated
by the proximity of dark blue to black on the colour scale, and the
historical grouping of blu with dark colours (2006: 84–5).
Unlike the English blue, azul is not conventionally related to any emotion
term in Peninsular Spanish. However, if tristeza (‘sadness’) has the same
dimensional features as sadness – i.e. negative Evaluation (–E), low Potency
(–P) and low Arousal (–A) (Fontaine et al., 2007) – one might expect
Spanish speakers to find the association between tristeza and azul (–E, ±P,
–A) to be more natural than between tristeza and rojo (+E, +P, +A ) or tristeza
and verde (+E, ±P, –A).
Summary and discussion
Table 5 presents a summary of the dimensional profile of the four Spanish
colour categories. What our results reveal is that ROJO, AMARILLO,
VERDE and AZUL can be successfully differentiated in terms of Evaluation
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 439
(good–bad), Potency (powerful–weak) and Activity (excited–relaxed),
Evaluation and Activity being the two most descriptive dimensions in
The results also reveal a general match between implicit associations and
explicit associations (of the sort captured by Adams & Osgood) for colour
terms and connotation. Potency and Activity exhibit the same structure, but
Evaluation – the most irregular dimension in the literature – does not coincide.
A possible reason might be that the construals of Potency and Activity,
unlike that of Evaluation, are mostly dependent on physiology. Perceptions
of Potency (weak–powerful) in colour are strongly linked to brightness
(light–dark) (Valdez & Mehrabian, 1994; Gao et al., 2007). Perceptions of
Activity (excited–relaxed) are strongly linked to saturation (saturated–unsaturated)
(D’Andrade & Egan, 1974; Valdez & Mehrabian, 1994).
Therefore, saturated dark colours (red, black) should be perceived as excited
and strong, and more naturally match emotions with the same features, like
anger, lust or hate. This is reflected in language: the terms red and black
seem to be widely associated to anger cross-culturally (Hupka et al., 1997).
Also, in our own data, ROJO was implicitly perceived as excited and powerful.
Conversely, light colours like yellow should be perceived as weak. In
our data, too, AMARILLO was implicitly perceived this way, and the main
difference between VERDE and AMARILLO was that VERDE (a slightly
darker colour) was perceived as slightly stronger than AMARILLO.
People are in fact sensitive to the meaning changes conveyed by these
differences in colour shade. In an experiment on Italian colour terms used
metaphorically, Cacciari, Massironi & Corradini (2004) showed that subjects
agreed on the most congruent shades of colour for the emotional tone
of a text. Clarke & Costall (2008) also recorded spontaneous accounts of
these meaning nuances in colour during their interviews.
What about Evaluation? Brightness (dark–light) seems to have an impact
on it, too. As we saw with conceptual metaphors, brightness biases towards
positive judgements, while darkness biases towards negative ones (cf. also
Table 5
Summary of the dimensional profile for the four
Spanish colour categories
Colours E valuation Potency A ctivity
ROJO + + +
AZUL − −
Downloaded from at Universite de Geneve on August 31, 2009
440 Social Science Information Vol 48 – no 3
Valdez & Mehrabian, 1994). Therefore darker colours are typically associated
to negatively valenced emotions, like anger, hatred or depression/despair.
The difference observed between our results and Adams & Osgood’s
(1973) as regards Evaluation may mean that evaluative associations are different
at an implicit and explicit level, a possibility that requires further
investigation. But it seems to us that Evaluation may be a qualitatively different
dimension from the other two as well. Determining when something
is ‘good’ or ‘bad’ is much more general than estimating ‘the amount of
energy involved in it’ (Activity), or ‘the amount of power and control it can
exert’ (Potency). The latter are specific semantic features; ‘goodness’/‘badness’,
on the contrary, is more vague. Positive or negative connotations seem to be
crucially mediated by context and, in the case of colour, by the association
to another experiential domain with which the colour co-occurs (and/or
which is culturally sanctioned as relevant). This allows for a greater degree
of interpersonal and intercultural variability than the other two dimensions.
As Elliot and colleagues (2007: 156) explain, colour meaning may be a matter
of context. We hypothesize that, in the case of valence, this may be
especially so. For example Spaniards may regard ROJO and AMARILLO
very positively because they are the two colours of the Spanish flag and
because red and yellow outfits are typically used to represent the country in
international competitions. It is an interesting possibility that the two colours
may have lately become particularly positive in people’s minds as a
result of Spain winning the UEFA European Football Championship after
44 years, as well as the Davis Cup in tennis.
The literature on the relationship between colour and emotion is vast and
varied. Different disciplines, orientations and methodologies combine to
produce a truly heterogeneous field that tries to provide an answer to the
question of which colours are associated to which emotions in a given
culture, in a language or, possibly, all over the world. An answer to this
question requires more than linguistic analysis. But linguistic analysis is
necessary as well, because general conceptual pairings of emotions and
colours in a culture are likely to be influenced by the specific linguistic
associations of colour words and emotional words in it.
It is this association we aimed to address. We have provided several reasons
why colour and emotion terms are associated in the languages of the
world, and have given examples of cross-linguistic variability in this respect.
In addition we have suggested a new methodology for analysing connotative
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 441
overlap, one of the possible motivations for some of those associations. We
believe implicit measurements are necessary to capture the connotative structure
of colour and emotion terms, since implicit and explicit representations
(of the sort reflected by Osgood’s Semantic Differential Technique) need not
coincide, and it is the implicit unconscious representations that can give us a
better lead into the reasons for colour–emotion association. One could further
explore the unconsciousness of the associations between colour terms and the
affective dimensions testing for the existence of Stroop interference effects
between a colour term and a contradicting dimensional pole (e.g. the recognition
of the word energetic written in red vs in blue/green/yellow).
The results of our study parallel earlier findings on Potency and Activity
in colour semantics and colour–emotion association, but they diverge as
regards the evaluative sign of the Spanish colour terms. We have suggested
that colour Potency and Activity semantic ratings are likely to be more stable
across languages because of their important relation to physical properties
of colour, like brightness or saturation. Evaluation, on the other hand, is
influenced by saturation, but seems nevertheless more sensitive to context,
more dependent on external associations and a framework of reference, and
therefore more prone to cross-linguistic variability.
Further research is now necessary to establish the implicit connotative structure
of emotion terms in Peninsular Spanish, and to compare it with that of the
colour terms. This opens the field to new interesting possibilities. For example
Peninsular Spanish lacks any phraseology directly linking azul or amarillo to
specific emotion terms.7 The same is true for other colour terms, like naranja
(orange). This makes Spanish a very interesting case study for potential facilitating
effects between those colours and emotion terms with compatible dimensional
structures, as well as interference effects with incompatible ones.
All in all, the present study has suggested new ways to investigate the
connections between colour and emotion, and to render a more vivid picture
of their cross-linguistic connotative landscape.
Cristina Soriano earned her PhD in Cognitive Linguistics at the University of Murcia
(Spain) with a multidisciplinary (semantic and psycholinguistic) contrastive cognitive
account of emotional language in English and Spanish. Her research interests focus on the
metaphorical representation of concepts (primarily emotion concepts) and the psycholinguistic
investigation of theory-driven cognitive claims. She has received training in
Cognitive Semantics at the University of Berkeley and the University of Hamburg as a
visiting scholar. She left the English Philology Department of the University of Murcia,
where she taught language and translation for two years, to join the Swiss Centre for
Affective Sciences (University of Geneva), where she currently conducts psycholinguistic
research on emotion concepts. She is also the executive officer of the GRID Project, a
cross-cultural investigation of the meaning of emotion terms, and member of the ELIN
Project on emotional language in international communication. Recent publications include:
(with A. Barcelona, 2004) ‘Metaphorical conceptualization in English and Spanish’,
Downloaded from at Universite de Geneve on August 31, 2009
442 Social Science Information Vol 48 – no 3
European journal of English studies 8(3): 295–308; (with J. Valenzuela, 2009) ‘Are
conceptual metaphors accessible on-line? A psycholinguistic exploration of the CONTROL
IS UP metaphor’, in J. Valenzuela, A. Rojo & C. Soriano (eds) Trends in cognitive linguistics:
theoretical and applied models, pp. 31–50 (Frankfurt: Peter Lang); with
A. Ogarkova, in press) ‘Linguistics and emotion’, in D. Sander & K. R. Scherer (eds)
Oxford companion to emotion and the affective sciences (Oxford: Oxford University Press).
Author’s address: Swiss Centre for the Affective Sciences, 7 rue des Battoirs, CH-1207
Geneva, Switzerland. [email:]
Javier Valenzuela is a Tenured Professor of Cognitive Linguistics at the University of
Murcia (Spain). He has published extensively and has ample teaching experience in contrastive
English–Spanish linguistics, with a specialization in semantic and grammatical
contrasts. He has carried out several studies for the psycholinguistic validation of cognitive
linguistics claims, primarily concerning Construction Grammar and Conceptual Metaphor
Theory. Professor Valenzuela has participated in several multidisciplinary research projects
on semantics, cognition and metaphor financed by the Spanish Government, the Autonomic
Government of Murcia, and the Universities of Murcia and Granada. He is also involved
in an international multidisciplinary project on artificial language evolution (ALEAR –
European STREP grant). Recent publications include: (with C. Soriano, 2007) ‘Conceptual
metaphor and idiom comprehension’, in I. Ibarretxe-Antuñano, C. Inchaurralde & J.
Sánchez (eds) Language, mind and the lexicon, pp. 281–304 (Frankfurt: Peter Lang); (with
A. Rojo, 2008) ‘What can foreign language learners tell us about constructions’, in S. De
Knop & T. De Rijcker (eds) Cognitive approaches to pedagogical grammar, pp. 197–229
(Berlin; New York: Mouton de Gruyter); (in press) ‘What empirical work can tell us about
primary metaphors’, Cuadernos de filologia 14; (with A. Rojo & C. Soriano, eds, 2009)
Trends in cognitive linguistics: theoretical and applied models (Frankfurt: Peter Lang).
Author’s address: Javier Valenzuela, Department of English Philology, University of
Murcia, Spain. [email:]
The research presented in this article was supported by grants from the Spanish Ministry of
Education and Science SEJ2006–04732/PSIC and from Fundación Séneca 05817/PHCS/07.
We are most grateful to Marc Ouellet (University of Granada, Spain) for his valuable support
with data analysis, and to Gill Philips, Julio Santiago, Anna Ogarkova and Klaus Scherer for
their comments on an earlier version of this paper. All errors remain our own.
1. Italics are used for linguistic examples and specific colour or emotion terms. A translation
of their meaning (when necessary) appears in single quotation marks. Conceptual metaphors,
conceptual metonymies, as well as colour and emotion concepts are capitalized (following the
convention in cognitive linguistics).
2. For an inventory of 25 studies from various disciplines reporting on the association white–
good and black–bad, see Adams & Osgood (1973).
3. Red. (n.d.) The American Heritage® Dictionary of the English Language, Fourth Edition.
Retrieved 15 March 2009 from website:
4. For an inventory of studies reporting additional converging evidence see Adams & Osgood
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 443
5. We have collapsed the results for AMARILLO, ROJO and VERDE, since the differences
among them were non-significant.
6. The explicit meaning structure of ira and 23 other emotion terms is currently being analysed
in the GRID study. The GRID Project ( is an international
research initiative to investigate the semantic structure of 24 emotion terms in over 30
world languages. Dimensional analyses are also included. The Spanish words investigated are
tristeza (sadness), desprecio (contempt), asco (disgust), ira (anger), irritación (irritation), odio
(hate), celos (jealousy), miedo (fear), ansiedad (anxiety), estrés (stress), desesperación
(despair), interés (interest), sorpresa (surprise), gozo (joy), placer (pleasure), orgullo (pride),
felicidad (happiness), satisfacción (contentment), amor (love), decepción (disappointment),
estar herido (being hurt), compasión (compassion), culpa (guilt) and vergüenza (shame). The
scholars responsible for the Peninsular Spanish studies are Cristina Soriano (for the Southern
Spanish sample) and Itziar Alsonso-Arbiol (for the Northern Spanish sample).
7. Rojo and verde are conventionally related to anger and envy, respectively, by virtue of the
expressions rojo de ira (‘red with anger’) and verde de envidia (‘green with envy’). As we saw
earlier, these expressions are motivated by metonymic thinking, not connotative overlap.
Adams, F. M. & Osgood, C. E. (1973) ‘A cross-cultural study of the affective meanings of
colour’, Journal of cross-cultural psychology 4: 135–56.
Allan, K. (2007) ‘The pragmatics of connotation’, Journal of pragmatics 39: 1047–57.
Allan, K. (2009) ‘The connotations of English colour terms: colour-based X-phemisms’,
Journal of pragmatics 41: 626–37.
Barcelona, A. (2003) ‘Clarifying and applying the notions of metaphor and metonymy within
cognitive linguistics: an update’, in R. Dirven & R. Pörings (eds) Metaphor and metonymy
in comparison and contrast, pp. 207–77. Berlin; New York: Mouton de Gruyter.
Berlin, B. & Kay, P. (1969) Basic colour terms: their universality and evolution. Berkeley:
University of California Press.
Cacciari, C., Massironi, M. & Corradini P. (2004) ‘When color names are used metaphorically:
the role of linguistic and chromatic information’, Metaphor and symbol 19(3): 169–90.
Clarke, T. & Costall, A. (2008) ‘The emotional connotations of color: a qualitative investigation’,
COLOR Research and application 33(5): 406–10.
D’Andrade, R. & Egan, M. (1974) ‘The colours of emotion’, American ethnologist 1: 49–63.
Dzokoto, V. A. & Okazaki, S. (2006) ‘Happiness in the eye and the heart: somatic referencing
in West African emotion lexica’, Journal of black psychology 32(2): 117–40.
Elliot, A. J. & Niesta, D. (2008) ‘Romantic red: red enhances men’s attraction to women’,
Journal of personality and social psychology 95(5): 1150–64.
Elliot, A. J., Maier, M. A., Moller, A. C., Friedman, R. & Meinhardt, J. (2007) ‘Colour and
psychological functioning: the effect of red on performance attainment’, Journal of experimental
psychology: General 136(1): 154–68.
Fontaine, J., Scherer, K. R., Roesch, E. B. & Ellsworth, P. (2007) ‘The world of emotions is not
two-dimensional’, Psychological science 18(12): 1050–7.
Gao, X. P., Xin, J. H., Sato, T., Hansuebsai, A., Scalzo, M., Kajiwara, K., Guan, S. S.,
Valldeperas, J., Lis, M. J. & Billger, M. (2007) ‘Analysis of cross-cultural colour emotion’,
Colour research and application 32(3): 223–9.
Downloaded from at Universite de Geneve on August 31, 2009
444 Social Science Information Vol 48 – no 3
Gawronski, B. (2002) ‘What does the Implicit Association Test measure? A test of the convergent
and discriminant validity of prejudice-related IATs’, Experimental psychology
49(3): 171–80.
Goldstein, K. (1942) ‘Some experimental observations concerning the influence of colours on
the function of the organism’, Occupational therapy and rehabilitation 21: 147–51.
Greenwald, A. G., McGhee, D. E. & Schwartz, J. L. K. (1998) ‘Measuring individual differences
in implicit cognition: the Implicit Association Test’, Journal of personality and social
psychology 74: 1464–80.
Hill, R. A. & Barton, R. A. (2005) ‘Red enhances human performance in contests’, Nature
435(7040): 293.
Hupka, R. B., Zaleski, Z., Otto, J., Reidl, L. & Tarabrina, N. V. (1997) ‘The colours of anger,
envy, fear, and jealousy. a cross-cultural study’, Journal of cross-cultural psychology 28(2):
Jobes, G. (1962) Dictionary of mythology, folklore and symbols, vols 1–2. New York:
Scarecrow Press.
Kövecses, Z. & Radden, G. (1998) ‘Metonymy: developing a cognitive linguistic view’,
Cognitive linguistics 9(1): 37–77.
Lakoff, G. (1993) ‘The contemporary theory of metaphor’, in A. Ortony (ed.) Metaphor and
thought, pp. 202–51. Cambridge: Cambridge University Press.
Leech, G. N. (1981) Semantics: a study of meaning. Harmondsworth: Penguin (2nd edn).
Marsh, K. L., Johnson, B. T. & Scott-Sheldon, L. A. J. (2001) ‘Heart versus reason in condom
use: implicit versus explicit predictors of sexual behavior’, Zeitschrift für experimentelle
Psychologie 48: 161–75.
McConnell, A. R. & Leibold, J. M. (2001) ‘Relations among the Implicit Association Test,
discriminatory behavior, and explicit measures of racial attitudes’, Journal of experimental
social psychology 37: 435–42.
Mehrabian, A. (1972) ‘Nonverbal communication’, in J. K. Cole (ed.) Nebraska symposium on
motivation, vol. 19, pp. 107–61. Lincoln: University of Nebraska Press.
Mehrabian, A. & Russell, J. A. (1974) An approach to environmental psychology. Cambridge,
MA: MIT Press.
Mehta, R. & Zhu, R. (2009) ‘Blue or red? Exploring the effect of color on cognitive task performances’,
Science 323(5918): 1226–9.
Meier, B. P., Robinson, M. D. & Clore, G. L. (2004) ‘Why good guys wear white. automatic
inferences about stimulus valence based on brightness’, Psychological science 15(2): 82–7.
Meier, B. P., Robinson, M. D., Crawford, L. E. & Ahlvers, W. J. (2007) ‘When “light” and
“dark” thoughts become light and dark responses: affect biases brightness judgments’,
Emotion 7(2): 366–76.
Mitchell, C. J., Anderson, N. A. & Lovibond, P. F. (2003) ‘Measuring evaluative conditioning
using the Implicit Association Test’, Learning and motivation 34: 203–17.
Murray, D. C. & Deabler, H. L. (1957) ‘Colours and mood-tones’, Journal of applied psychology
41: 279–83.
Odbert, H. S., Karwoski, T. F. & Eckerson, A. B. (1942) ‘Studies in synesthetic thinking, I:
Musical and verbal associations of colour and mood’, The journal of general psychology
26: 153–73.
Ogarkova, A. (2007) ‘“Green-eyed monsters”: a corpus-based study of metaphoric conceptualizations
of JEALOUSY and ENVY in modern English’, Metaphorik 13: 87–147.
Osgood, C. E., May, W. H. & Miron, M. S. (1975) Cross-cultural universals of affective meaning.
Urbana; London: University of Illinois Press.
Downloaded from at Universite de Geneve on August 31, 2009
Soriano & Valenzuela Emotion and colour across languages 445
Osgood, C. E., Suci, G. J. & Tannenbaum, P. H. (1957) The measurement of meaning. Urbana;
London: University of Illinois Press.
Oyama, T., Tanaka, Y. & Haga, J. (1963) ‘Colour-affection and colour-symbolism in Japanese
and American students’, Japanese journal of psychology 34: 109–21.
Pecjak, V. (1970) ‘Verbal synesthesiae of colours, emotions, and days of the week’, Journal
of verbal learning and verbal behaviour 9: 623–6.
Philip, G. (2006) ‘Connotative meaning in English and Italian colour-word metaphors’,
Metaphorik 10: 59–93.
Schaie, K. W. (1961) ‘A Q-sort study of colour–mood association’, Journal of projective techniques
25: 341–6.
Shaver, P., Schwartz, J., Kirson, D. & O’Connor, C. (1987) ‘Emotion knowledge: further explorations
of a prototype approach’, Journal of personality and social psychology 52: 1061–86.
Sirvydė, R. (2007) ‘Metonymy – a sister or a stepdaughter? A case study of the colour of
anger’, Respectus philologicus 11(16): 145–53.
Snider, J. G. & Osgood, C. E., eds (1969) Semantic differential technique. Chicago, IL: Aldine.
Soldat, A. S., Sinclair, R. C. & Mark, M. M. (1997) ‘Colour as an environmental processing
cue: external affective cues can directly affect processing strategy without affecting mood’,
Social cognition 15: 55–71.
Soriano, C. (2005) ‘The conceptualization of anger in English and Spanish: a cognitive
approach’, Doctoral dissertation, Department of English Philology, University of Murcia,
Taylor, J. R. & Mbense, T. M. (1998) ‘Red dogs and rotten mealies: how Zulus talk about
anger’, in A. Athanasiadou & E. Tabakowska (eds) Speaking of emotion, pp. 191–226.
Berlin: Mouton de Gruyter.
Valdez, P. & Mehrabian, A. (1994) ‘Effects of colour on emotions’, Journal of experimental
psychology: General 123(4): 394–409.
Valenzuela, J. & Soriano, C. (2008) ‘Sensorial perception as a source domain: A crosslinguistic
study’, Paper presented at the VII International conference on Researching and
Applying Metaphor (RaAM 7), University of Extremadura, Cáceres, Spain, 29–31 May.
Wexner, L. B. (1954) ‘The degree to which colours (hues) are associated with mood-tones’,
Journal of applied psychology 38: 432–5.
Wierzbicka, A. (1992) Semantics, culture and cognition: universal human concepts in culturespecific
configurations. New York: Oxford University Press.
Downloaded from at Universite de Geneve on August 31, 2009

Tinggalkan Balasan

Isikan data di bawah atau klik salah satu ikon untuk log in:


You are commenting using your account. Logout / Ubah )

Gambar Twitter

You are commenting using your Twitter account. Logout / Ubah )

Foto Facebook

You are commenting using your Facebook account. Logout / Ubah )

Foto Google+

You are commenting using your Google+ account. Logout / Ubah )

Connecting to %s