Lithuanian Conceptual Colour–Emotion Associations in the Global Context of 37 Nations

. Red with anger or green with envy – such metaphors link colours and emotions. While such colour metaphors vary across languages, conceptual associations between colours and emotions have many cross-cultural similarities. Here, we took published data from 8615 participants (2172 men) coming from 37 nations (i.e., Austria, Azerbaijan, Belgium, China, Colombia, Croatia, Cyprus, Egypt, Estonia, Finland, France, Georgia, Germany, Greece, India, Iran, Israel, Italy, Japan, Latvia, Lithuania, Mexico, Netherlands, New Zealand, Nigeria, Norway, Philippines, Poland, Russia, Saudi Arabia, Serbia, Spain, Sweden, Switzerland, Ukraine, United Kingdom, and United States) and analysed Lithuanian ( n = 217) associations between colour terms and emotion concepts. Lithuanians had many associations, the most frequent being red–love , yellow–amusement , yellow–joy , and black–sadness (all endorsed by > 60% of participants). While Lithuanians associated more emotions with colours than the other participants, the Lithuanian pattern of these associations was highly similar to the global pattern ( r = .92). When compared to each other nation individually, colour–emotion association pattern similarities ranged between .65 and .89. Lithuanian patterns were the most similar to the Russian and the least similar to the Egyptian ones. Crucially, such similarities could be predicted by linguistic but not geographic distances. Nations speaking languages linguistically closer to Lithuanian also displayed more similar colour–emotion association patterns. These results support universality of colour–emotion associations and point to small but meaningful cultural differences (e


Introduction
Colours and emotions are linked in languages and traditions.Our successes are marked in green and errors in red.We send red roses to our true loves, and we speak in "coloured" language.Lithuanians fear for black days (juoda diena), get red when embarrassed (raudonuoti), work blackly (juodai dirbti; i.e., work hard), and do not want to be a white crow (balta varna; i.e., an ugly duckling) or a green cucumber (žalias kaip agurkas; i.e., novice) (Kosova & Klanauskaitė, 2015;Roch, 2015).Lithuanians are one of the few to express degrees of anger through colour -from white, to red, to blue and black (pabalęs/ įraudęs/pamėlęs/pajuodęs iš pykčio) (Sirvydė, 2007).This is unlike the English speakers, who express degrees of anger through shades of red (flushed/pink/red/scarlet with anger) (Sirvydė, 2007), also discussed in (Soriano & Valenzuela, 2009).While there are clear differences in how colour is used metaphorically in languages (e.g., Iljinska & Platonova, 2017;Kalda & Uusküla, 2019;Kosova & Klanauskaitė, 2015;Philip, 2006), empirical research in psychology has revealed many commonalities across nations (Adams & Osgood, 1973;Jonauskaite et al., 2020;Ou et al., 2018).Here, we aim to describe Lithuanian colour-emotion associations, and compare them to colour-emotion associations collected from other 36 nations.
In the seminal study, Adams and Osgood (1973) asked students from 23 countries to rate seven colour terms on the semantic differential scales loading on valence (positive-negative), arousal (arousing-calming), and power (strong-weak).1They found similar affective ratings of colours across the studied countries (e.g., black was negative and strong, red was strong and arousing).In a more recent study, Jonauskaite and colleagues (2020) assessed associations between 12 colour terms and 20 emotion concepts in representative samples of participants from 30 nations, which also included Lithuania.They reported a high degree of similarity in the patterns of associated emotions.The pattern of Lithuanian colour-emotion associations had 0.92 correspondence with the pattern of all other participants.In a subsequent study, also including Lithuania, a high degree of consistency across the lifespan, testing 16-88-year-old participants, was also observed (Jonauskaite et al., 2023).
Beyond universally understood colour-emotion associations, there are small but meaningful cultural differences.These differences might be driven by culture-specific variables such as environmental conditions or locally spoken languages (e.g., see Hupka et al., 1997;Kawai et al., 2023;Soriano & Valenzuela, 2009).Two large-scale studies supported both suppositions.Regarding environmental conditions, across 55 countries, participants living in countries closer to the equator (i.e., warmer) and with lower annual precipitation levels (i.e., dryer) were less likely to associate the colour term yellow with the concept of joy (Jonauskaite, Abdel-Khalek, et al., 2019).Regarding spoken languages, across 28 countries and 16 languages, participants whose languages labelled the PURPLE category with the cognate of purple (e.g., English) associated more positive and empowering emotions than those labelling the PURPLE category with a cognate of violet (e.g., French -violet, Lithuanian -violetinė) (Uusküla et al., 2023). 2 Even more generally, across 30 nations, lower linguistic and geographic distance predicted higher similarity in colour-emotion association patterns (Jonauskaite et al., 2020).
Previous studies looked for global cross-cultural patterns (Adams & Osgood, 1973;Jonauskaite et al., 2020Jonauskaite et al., , 2023;;Ou et al., 2018) and also identified some cultural differences (Hupka et al., 1997;Jonauskaite, Abdel-Khalek, et al., 2019;Kawai et al., 2023;Uusküla et al., 2023).Here, we took a closer look at Lithuanian colour-emotion associations and compared them to the associations obtained from 36 other nations.We tested whether geographic or linguistic closeness could predict similarity in these associations.To this end, in addition to Lithuanian participants, we recruited participants from neighbouring nations (i.e., Latvia, Poland, Russia), other European nations (e.g., Estonia, Germany, France, Switzerland), and nations located on other continents (e.g., USA, Mexico, Colombia, Nigeria, China, India, Japan, New Zealand; see all nations in Figure 1).

Figure 1
Map of the 37 studied nations, and how similar their colour-emotion associations were to the Lithuanian ones Note.NA = no data from those countries (see also Figure 4).

Participants
We took previously published data from (Jonauskaite et al., 2020(Jonauskaite et al., , 2023)).In total, there were 8615 participants (2172 men, 6389 women, 54 did not report their gender), including 217 Lithuanian participants (40 men, 177 women).Participants' mean age was 35.46 years (SD = 15.66 years, range = 15-88 years).Participants came from 37 nations and spoke 25 languages (see all demographic data in Table 1 and Figure 1).The data had been preselected, taking only participants who originally came from one of the 37 countries and who completed the study in their native language.To take Lithuania as an example, only participants who reported that their country of origin was Lithuania, their native language was Lithuanian, and who completed the study in Lithuanian were included.We did not consider their residence country, meaning that some participants might have resided in other countries.Two exceptions were Nigerian and Indian participants, who completed the study in English (the official language; see all languages in Table 1).All participants were highly fluent in their respective languages, with the self-rated mean language fluency score of 7.85 out of 8.All participants took part voluntarily and were not remunerated for their participation.The study was conducted in accordance with the principles expressed in the Declaration of Helsinki (World Medical Association, 2013) and was approved by the local ethics committee (C_SSP_032020_00003). Note.n = number of participants.Geographic distances were measured in kilometres from population centres between Lithuania and each other nation.Linguistic distances between Lithuanian and each other language were extracted from (Jäger, 2018), and ranged between 0 (i.e., identical languages) and 1 (i.e., totally dissimilar languages).Scores below 0.24 indicate linguistic relatedness.

Emotion Assessment
The Geneva Emotion Wheel (GEW, version 3.0; Scherer, 2005; Scherer, Shuman, Fontaine, & Soriano, 2013) is a self-report measure of emotion, containing 20 emotions (Figure 2).These emotions are represented along the circumference of a wheel, organized around two axes -valence (horizontal: positive vs. negative) and power (vertical: high power vs. low power).Emotions similar in valence and power are placed close to each other.Circles of increasing size connect the centre of the wheel with the circumference of the wheel,

Figure 2
The Geneva Emotion Wheel (GEW) in English and Lithuanian, with an example for RED signifying five degrees of emotion intensities (1-5).See Figure 2 for the emotion terms in English and Lithuanian and previous studies for emotion terms in the other languages (Jonauskaite et al., 2020(Jonauskaite et al., , 2023)).

Procedure
In the previous studies (Jonauskaite et al., 2020(Jonauskaite et al., , 2023)), the data were collected online on a custom-built website.Participants were given information about the study and provided informed consent.After passing the verification check, participants saw 12 colour terms in random order and were asked to associate one, several, or none of the GEW emotions with each colour term.They also rated intensity of the associated emotions.They could choose "No emotion" or "Different emotion" for each colour term (see the different emotions in Lithuanian in Table A 1).
More precisely, participants received these instructions: You will see different colour words in no particular order.For each colour word, please use the emotion wheel (see below) to indicate which emotion or emotions are for you best represented by that colour word.
Each spike in the wheel represents an emotion, for example "anger" as indicated, or a closely related emotion (e.g., irritation, a type of anger).Please rate the intensity of each emotion (one or more) that you associate with the particular colour word shown above the wheel.Smaller circles indicate weaker emotions and larger circles indicate stronger emotions.You can correct your choice by clicking on the small square at the hub of the wheel, meaning that this emotion is not associated with the colour word.
Click on "No emotion" if you do not associate any emotion with the given colour word.
If you associate that colour word with another emotion that is not displayed in the wheel, please click on "Different emotion".You will be asked to write down the emotion(s) in the pop-up window.

Data Analysis
In the previous studies, the data had been pre-cleaned by excluding participants who were too quick or too slow (i.e., took less than 3 or more than 90 min), or did not show minimal engagement with the online experiment (i.e., spent less than 20 s on the first four colour terms).Some participants had missing data on some of the colour terms, and we included them if no more than four (i.e., 33%) of colour terms had missing data.Access data here: https://osf.io/2w6gh/?view_only=e992cdbb920c433395808f34a3d4c9bd

Patterns of Colour-Emotion Associations
We calculated proportions of participants associating each colour term with each emotion concept in the following way.For each colour-emotion combination (e.g., RED and anger), we calculated the number of participants who chose the particular emotion concept (i.e., anger) and divided by the total number of participants.We repeated this procedure for all 240 colour-emotion associations (i.e., 12 colour terms x 20 emotion concepts) and combined these proportions to make the patterns of colour-emotion associations.We established colour-emotion association patterns for each nation separately as well as for all nations together (but without Lithuania).

Geographic and Linguistic Distances
We calculated geographic distances in kilometres between the centre of Lithuania vs. the centres of each other nation (see Table 1).We used population-weighted geographic centres instead of unweighted geographic centres to account for an uneven distribution of inhabitants in some countries.While such calculation had little effect on the central point of Lithuania and many other smaller countries, it affected larger countries, such as Russia, where the majority of the population is located in one part of the country.See all geographic distances to Lithuania in Table 1.
We extracted linguistic distance scores between the Lithuanian language and the national language of each other nation from Jäger (2018), capturing the phylogenetic distances between languages from lexical sources.The original linguistic distances ranged from 0 to 1, with lower linguistic distance scores indicating higher linguistic similarities.However, as the linguistic distances were not evenly spread across this range, we followed a previous publication (Jonauskaite et al., 2020) and used a power transform to the fourth power (^4) of the original distances.Languages belonging to the same language family (i.e., Indo-European) or even the same linguistic group (i.e., Baltic languages) had lower linguistic distance scores than languages from other language families (e.g., Uralic, Afro-Asiatic, Sino-Tibetan).While Jäger (2018) proposed that language pairs with distances below .7 should be considered as related, after the power transformation, such criterion became .24(i.e., 0.74^4).See all linguistic distances to Lithuanian in Table 1.

Lithuanian Colour-Emotion Associations
On average, Lithuanian participants associated 4.01 emotion concepts (SD = 4.60, range = 0-20) with each colour term.This number was higher than that on average, t(2686) = 8.56, p < .001,whereby other participants associated 3.23 emotion concepts (SD = 3.61, range = 0-20), suggesting that Lithuanians were more likely to link colours with emotions than other participants.

Lithuanian Colour-Emotion Associations in the Global Context
We used Pearson correlations to compare the Lithuanian colour-emotion association pattern (see Figure 3) with analogous association patterns of i) all the remaining participants taken together (global pattern), and ii) with patterns of each of the remaining nation.
Regarding the comparison with the global pattern, Lithuanian associations were highly correlated (r = .924,p < .001; Figure 3).Visually, the two patterns appeared highly similar, apart from the fact that Lithuanians were more likely to associate colours with emotions and thus resulted in higher proportions overall (i.e., darker cells).Still visually, it seemed that more Lithuanians linked RED with love than anger, while globally, both emotions were associated at similar frequencies with RED.It also seemed that Lithuanians associated pride and compassion with BLACK, in addition to the more common emotions like sadness, fear, anger, guilt, disappointment, and hate.Amusement was particularly strongly linked to YELLOW, in addition to joy, while globally YELLOW-joy association was more frequent.
Regarding the comparison with each of the other nations individually, the mean correlation was r = .830(see Figure 1 and Figure 4).These correlations ranged from r = .645to r = .892,suggesting a high degree of pattern similarity (1 = identical patterns).All correlations were statistically highly significant, p < .001.Lithuanian colour-emotion associations were the most similar to those of Russian, Ukrainian, Estonian, Polish, and Italian participants.Lithuanian colour-emotion associations were the least similar to those of Egyptian, Azerbaijani, and Nigerian participants (Figure 4).

Linguistic and Geographic Distances
Lastly, we used two linear regression models to predict the degree of similarity of colour-emotion association patterns by geographic and linguistic distances from Lithuania/Lithuanian. The model with geographic distances as predictors was not significant, F(1, 34) = 1.81, p = .186,R 2 adj = .023,meaning that geographic distances could not predict colour-emotion pattern similarities between Lithuania and other nations.In contrast, the model with linguistic distances as predictors was significant, F(1, 34) = 4.89, p = .034,R 2 adj = .100.Nations that spoke languages more closely related to Lithuanian also associated colours and emotion in a more similar way (Figure 5).

Discussion
Colours carry emotional meanings to many (e.g., Adams & Osgood, 1973;Fugate & Franco, 2019;Jonauskaite et al., 2020), and Lithuanians were not an exception.Lithuanians had many colour-emotion associations, the most frequent being red-love, yellow-amusement, yellow-joy, and black-sadness, all endorsed by at least 60% of participants.These associations were many-to-many rather than one-to-one, indicating that one colour carried associations with several emotions and vice versa.There were many similarities between the pattern of Lithuanian colour-emotion associations and that of the other 36 nations.Similarity to the global pattern (i.e., the pattern of all the remaining participants) was very high (r = .92).These results supported the universality of colour-emotion associations, also reported in previous empirical studies (Adams & Osgood, 1973;Johnson et al., 1986;Jonauskaite et al., 2020Jonauskaite et al., , 2023;;Jonauskaite, Wicker, et al., 2019;Ou et al., 2018;Specker et al., 2018).When compared to each other nation individually, pattern similarities ranged between 0.65 and 0.89, being the most similar to Russian, Ukrainian, Estonian, Polish, and Italian patterns of association (see a detailed study of Russian colour-emotion associations in (Griber et al., 2019).
Such pattern similarities could be predicted by linguistic distance to Lithuanian, obtained from (Jäger, 2018).Nations speaking linguistically related languages displayed more similar colour-emotion association patterns than nations speaking more distant languages.Previously, linguistic similarity was not only important for general colouremotion association patterns (Jonauskaite et al., 2020), but also for specific colours.For instance, English speakers were more likely to associate BLUE with sadness (Barchard et al., 2017), while German speakers linked envy to YELLOW (Hupka et al., 1997).This was perhaps because each language possesses metaphors linking these colours and emotions (i.e., feeling blue means to feel sad in English, and Gelb vor Neid, lit. to be yellow with envy, exists in German).In another study, emotion associations with the category PURPLE were predicted by the basic terms that participants used to label this category (Uusküla et al., 2023).
Curiously, pattern similarities were not successfully predicted by geographic distances to Lithuania.While geographic distance previously predicted the degree of joyfulness of yellow (Jonauskaite, Abdel-Khalek, et al., 2019), here the linguistic factors outweighed the geographic factors.While geographic and linguistic distances were correlated (i.e., participants living geographically closer also spoke more related languages), the two measures were not identical.Due to the past colonialisation, Indo-European languages are spoken well beyond the European continent.In the current sample, Mexican and Colombian participants spoke Spanish, while Nigerian and Filipino participants spoke English.The importance of linguistic distance suggested that colour-emotion associations might be encoded and transmitted through language.Indeed, even colour blind and blind individuals can associate colours with emotions (Jonauskaite et al., 2021;Sato & Inoue, 2016;Saysani et al., 2021), indicating that intact colour perception is not required to make such associations.
Beyond similarities, there were also some cultural differences.Lithuanians associated more emotions with colours than the others, suggesting that colours were particularly emotive to Lithuanians (also see Jonauskaite et al., 2020).Lithuanians also associated RED with love more strongly than anger, while participants in general endorsed both associations.In addition to the common associations (e.g., sadness), Lithuanians associated BLACK with compassion -a somewhat positive emotion concept in English (Scherer et al., 2013).This association could be explained linguistically, whereby compassion had been translated to Lithuanian as užuojauta.The latter word also means condolences, highlighting the link between compassion/condolences and death, and death is commonly represented by BLACK (Allan, 2009;Tham et al., 2020).Finally, based on the free responses, Lithuanians missed calmness as a potential response option, associating it with GREEN, TURQUOISE, BLUE, and WHITE.
The current study dealt with associations between colours and emotions.A priori, such research tells little about felt emotions.More studies, using different experimental designs, are necessary to understand whether colour can impact felt emotions, and if so, whether such impact goes in line with the conceptual colour-emotion associations (e.g., see Weijs et al., 2023;Wilms & Oberfeld, 2018).Likewise, the current study did not deal with colour preferences (i.e., liking or disliking specific colours (Palmer & Schloss, 2010;Pranckevičienė et al., 2009;Stanikūnas et al., 2020)).Preferences are related yet distinct affective processes from emotion (Scherer, 2005).In other words, one cannot assume that emotion associations and preferences are always congruent (i.e., not all positive colours are liked, and vice versa).Perhaps, colour preferences are more personal than colour-emotion associations, reflecting aesthetic experiences rather than learnt abstract meanings of colour.More empirical research is necessary to disentangle the two types of affective connotations.

Conclusions
Across the globe, people associate colours and emotions (e.g., Adams & Osgood, 1973;Jonauskaite et al., 2020).Lithuanians too associated colour terms with diverse emotion concepts, most of which were similar to the other 36 studied nations, in particular, those speaking linguistically related languages.These observations demonstrate that colour can be used to communicate emotions effectively and universally, making it an important tool for applied sectors (e.g., marketing, design).As there were small cultural differences, emotion communication through colour could be further tailored for a specific country.For example, red represented love more strongly than anger for Lithuanians than globally, suggesting that Lithuanians considered this colour to be more positive.As the current study dealt with conceptual associations between colour terms and emotion concepts, future studies should test whether colours can also modulate experienced emotions.Such findings would be important theoretically (i.e., how abstract associations link to experiences) and practically (e.g., health sector, including chromotherapy).

Figure 3
Figure 3Colour-emotion associations of all participants (left) and the Lithuanian participants(right)

Figure 4
Figure 4Colour-emotion association pattern similarities between Lithuanian colour-emotion associations and those of the remaining 36 nations.

Figure 5
Figure 5Similarity between the Lithuanian colour-emotion association pattern and the other nation, predicted by geographic distance (A) and linguistic distance(B)

Table 1
Demographic information of all participants, separated by nation Domicelė Jonauskaitė.Lithuanian Conceptual Colour-Emotion Associations in the Global Context of 37 Nations