The Role of Context in Translating Colour Metaphors: An Experiment on English into Estonian Translation

Contemporary theory on metaphor states that metaphor is conceptual, conventional, and part of the ordinary system of thought and language. It has been argued that metaphors can become a translation problem, since transferring them from one language and culture to another may be restricted by linguistic and cultural differences. We investigated how colour metaphors are translated from English into Estonian. To understand how metaphors are translated, a cognitive empirical study was carried out with 21 colour metaphors. The experiment was conducted with two separate groups of volunteers. The first group participated in a context-based translation task, the second in a context-free one. The experiment indicates that colour metaphors are culture specific. It also revealed that context plays a crucial role in the comprehension and translation of colour metaphors. The more novel and original the metaphor is, the more varied are the translation strategies used by the participants (e.g. yellowbellied person). Differences in translation choices were obvious between translators and non-translators. Qualitative differences appeared as translators were more target culture oriented and non-translators more source culture oriented, for example.

. However, less attention has been paid to the topic of colour metaphor translation, and so this is an appropriate ground for further development. Available studies that have been conducted include the language pairs English-Chinese (Li 2011), English-Persian (Shabani 2008, Ahmadi andKetabi 2011), English-Arabic (Chatti 2014), English-Turkish (Kolahdouz 2015), and English-French (Bennett 1981). It is important to note that these examples included analysis of existing translations or bilingual dictionaries, but there is limited evidence from empirical analysis of colour metaphor translation. The current study aims to contribute to the analysis based on empirical data. Language should be explored in relation to its speakers and to the cognitive processes activated when using language (De Knop 2014: 73), and so we found it important to research colour metaphor translation with cognitive tasks and not to base our analysis on different corpora.
It has been argued that metaphor plays an important role in human cognition and social activities (He 2011(He : 1804, and the impetus for studying metaphor meticulously was the publication of Metaphors We Live By (Lakoff & Johnson 1980). Lakoff's cognitive view of metaphor was regarded as a breakthrough in metaphor studies since it considered metaphor as a way of thinking. Cognitive linguistic scholars argue that metaphor is fundamentally a conceptual process of mapping one domain of experience onto another (Lakoff and Johnson 1980: 5).
Glucksberg and colleagues propose that understanding a metaphor requires two kinds of world knowledge (1997: 59). First, enough knowledge is needed about the topic to allow an appreciation of which kinds of characterisation are relevant and meaningful, meaning the dimensions of the within-category variation of the topic concept must be understood. Second, enough must be known about the metaphor vehicle to know the kinds of things it may epitomise. Although a considerable number of publications have dealt with colour terms, very few have paid due attention to the metaphorical nature of colours (e.g. Wyler 1992: 143-148), especially from a cross-cultural perspective (Amouzadeha et al. 2012: 238, Bazzanella et al. 2016, Hamilton 2016. Niemeier (1998: 141) stresses that "analysing colour terms and their meaning extensions is a challenging endeavour indeed since such a lot of extra-linguistic knowledge comes to the surface". In the causal construction under study this extra-linguistic knowledge can be recognised in the associations related to the colour adjective. Consequently, colour terms and their use in particular expressions have to be studied with reference to the speakers of a given language (De Knop 2014: 84-85).
In translating colour metaphors from one language into another it is crucial to understand the intersection between linguistics, translation and culture, as each of these elements plays an important role in any understanding of how the others behave and what their ties with each other are.
Beyond doubt, the most challenging cases of translation of colour metaphors are those where no oneto-one correspondence can be found in any language other than the original, such as the English "yellowbellied". It is difficult to find an appropriate translation that is equivalent for culturally loaded metaphors since translation equivalents usually do not share all the specific semantic connotations and associations found in other languages (Philip 2011, Chatti 2014, and so other strategies for translation must be chosen. It must be remembered that the translation unit for such linguistic forms is meaning and not the original word or phrase itself. Colours in different languages and cultures may convey different connotative interpretations. This is particularly true when it comes to translating culturally-charged expressions like metaphors, idioms, and proverbs, where literal translation rarely provides an accurate target text. Colour metaphors embody a symbolic meaning established by convention that belongs to the community's cultural heritage. A crosscultural analysis of colour symbolism is essential for conveying the exact meaning of colour metaphors cross-linguistically. Indeed, the same colour may convey different meanings in different cultures, depending on the beliefs, norms and constraints which prevail in each culture (Chatti 2014: 5, see also House 2006. On top of these cultural aspects, modern metaphor theory emphasises the usefulness of context (Glucksberg 2003, Kolahdouz 2015, Kövecses 2003, 2005, Olivera & Sacristán 2001: 76, Philip 2003: 315-317, Wyler 1992. Estes and Glucksberg (1998: 338) conclude from their experiments that context can differentially activate features of combined concepts, and that it may do so through semantic priming.
De Knop (2014: 74) also stresses the importance of context, stating that it is interesting that the associations with colours do not depend only on the speech communities in which colour terms are used, but also on the linguistic context. In Western countries "red" is often associated with love and passion, but in the context of a causal construction it associates rather with anger (MacLaury et al. 1997: 77, Uusküla 2011: 151-152, Vaňková 2007. Wierzbicka argues that colour perception is, by and large, the same for all human groupings (1990: 103), but colour conceptualisation is different in different cultures, although there are also some striking similarities (MacLaury, Paramei, Dedrick 2007, Paulsen, Uusküla, Brindle 2016, Wierzbicka 1990. The traditional approach to metaphor translation has been prescriptive in nature. Toury calls a prescriptive approach normative in translation studies (1985), meaning it imposes criteria that stipulate how translation should be made so it can be kept as close to the original as possible. The traditional theory of metaphor translation interprets metaphors as a figure of speech (Dagut 1976, Newmark 1981, van den Broeck 1981, but a descriptive approach follows the lines of cognitive linguistics (Tirkkonen-Condit 2001, Snell-Hornby 1988 and says what translations are actually like (Schäffner 2004(Schäffner : 1255, allowing researchers to draw implications about how translations are actually made, not how they should be made. One of the many implications of this approach is that the translation is situated within the target culture system, and considered according to its significance and acceptability (Toury 1995: 36-39). Samaniego Fernandez (2013) focuses on metaphor and on the positive influence the cognitive approach to metaphor has exerted on descriptive translation studies. She argues that the notion of metaphor imported from cognitive linguistics has led to a more realistic study of metaphor translation that has allowed researchers to explain cases that were traditionally disregarded as anomalous or incorrect renderings.
The present study aims to observe and describe how colour metaphors are translated from English into Estonian by volunteer participants. The research was carried out on translation tasks from English into Estonian, where some colour metaphors occurred in a context while others were presented in a context-free setting so the role of context could be observed. The English colour categorisation system maps fairly easily onto Estonian as both languages have developed exactly 11 basic colour terms (Berlin & Kay 1969, Davies & Corbett 1995, Sutrop 2000. We apply cognitive analysis in order to investigate the performance of translators, rather than to analyse whether the translations are carried out well or less well, except in obvious cases of mistranslation. We intend to use our analysis to construct a classification of colour metaphor translation strategies which could partially serve as a guideline for translators if needed. Second, we observe the role of the context and how it affects cases of obvious mistranslation. Third, we discuss whether volunteer participants' cultural competence and linguistic proficiency alter the translation strategies used and whether people who have lived in English-speaking countries grasp the context better than people without such an experience. Moreover, we are interested in whether the difference between professional translators and non-translators is obvious from their translation choices. Van Doorslaer argues that translation education matters when decisions are made in translation (2015: 239). The process of translation will be described elsewhere, as we do not study the reaction times or other measurable components of the translation process in the present article.

Material and methods
Cognitive research was carried out with two separate groups of participants (for further details see Table  2). Some of them translated a text containing colour metaphors, while others translated single phrases of colour metaphors without any context. The identity of the participants remained anonymous and the results were generalised across all the participants. All the participants agreed to donate their knowledge on a volunteer basis and they filled in a participant form. They were allowed to stop at any time during the task if they did not want to proceed.
Both tasks focused on translating 21 English colour metaphors into Estonian. The metaphors were selected by the authors based on their conventionality in the source language (e.g. yellow-bellied person or red tape) or their conventionality in both the source and the target language (e.g. rose-coloured glasses or give the green light). The chosen metaphors contained basic colour terms like red, non-basic objectrelated colour terms like emerald, and non-basic colour terms derived through semantic shift like silver. The metaphors chosen included lexicalised metaphors and non-lexicalised culture-specific novel metaphors (see Table 1 for details). They also contained some adjectives associated with colour or classically used as colour modifiers, but that are not regarded as colour terms per se, such as dark in the expression dark ages. Two metaphors in our sample contained the word colour itself, true colours and flying colours (as in Sandford 2012: 203).
In the context-free translation task, the colour metaphors were presented to the subjects in one column on a white sheet of paper and forty-eight volunteers participated in the context-free part of the translation task (see Table 2 for details). The mean age of the participants was 30.0 years, the youngest being 20 years old and the oldest 53. One fifth of the participants had lived in an English-speaking country for an average of two years and eight months, and those countries were the United States, the United Kingdom and Australia.
In the second, context-based translation task the participants had to use screen recording software while translating a text of 445 words. The text consisted of 24 sentences, of which 12 contained colour metaphors. There were altogether 21 colour metaphors in the test (as listed in Table 1). The British National Corpus was used to compile the text for the test (shown in Appendix 1). The sentences were edited to some extent to provide coherent and cohesive context. Thirty-three volunteers took part in the context-based part of the study (further data in Table 2). The average age of the participants was 34.2 years, the youngest being 23 years old and the oldest 64. One third of the sample had lived in an English-speaking country for an average of 2 years, and these countries were the United States, the United Kingdom, Australia, Canada and Ireland.
All participants were L1 Estonian speakers. The main criterion for eligibility was C-level proficiency in English in the Common European Framework of Reference for Languages (CEFR), based on self-assessment. The translators in the sample were people with at least two years of translation experience. Professional translators had experience in translating economic, technical, legal or fictional texts or had specialised in audiovisual translation, localisation or interpreting. The participants were not informed about the specific nature of the task until the immediate beginning of the test in order to encourage spontaneity. Both groups of participants were allowed to consult electronic materials and dictionaries while taking the test.
After doing the translation task, the subjects filled in the background questionnaire. Besides standard demographics such as gender, age and proficiency in English, they were also requested to report whether they had lived in an English-speaking environment. Participants were also asked about any problems that occurred during the translation tasks and were asked for some self-reflection on the task. Additionally, eleven semi-structured interviews were carried out after the translation task with participants in the context-free part of the task and fourteen for the context-based part of the study.

Results
Each colour metaphor was analysed separately. We report on the strategies used for colour metaphor translation metaphor by metaphor in sub-sections 3.1-3.5, and the translation strategies are illustrated in Figures 1-5. We have chosen five different metaphors to illustrate different strategies.
We noted the use of five main approaches, which can be combined into a classification of colour metaphor translation strategies: 1. reproducing the colour metaphor in the target language with a colour metaphor; 2. replacing the colour metaphor without a colour word/colour metaphor in the target language; 3. converting the colour metaphor to a sense or paraphrase; 4. omission, if the colour metaphor is redundant or untranslatable; 5. obvious mistranslation of colour metaphor, where the original meaning gets lost in translation.

Rose-coloured glasses remain rose-coloured
The English colour metaphor "to look through rose-coloured glasses" means to see things in a positive light: to see only the pleasant things about a situation and not notice the things that are unpleasant (Cambridge Dictionary). This phrase is also lexicalised in Estonian with a similar meaning and a similar colour word: roosad prillid 'pink glasses', läbi roosade prillide nägema või vaatama (EKSS) as in example (1). (1) läbi roosade prillide nägema või vaatama through-prep pink-pl.gen glasses.pl.gen see-inf look-inf 'looking through rose-coloured glasses' In glossing roosa into English we can use the equivalent pink, as both roosa and pink are basic colour terms (Berlin andKay 1969, Sutrop 2000). One-to-one correspondence of rose-coloured would be roosivärvi, which means colour of a rose.
The participants in both the context-free study and the context-based one translated the phrase quite similarly using the colour word rose (roos in Estonian). There were no comprehension or translation problems related to this metaphor, and nor did any participants omit the colour metaphor in either part of the test. However, cases of semantic priming seemed to occur, especially in the context-free translation task. As the colour word rose can refer to both an abstract colour and a flower, this colour metaphor was interpreted as glasses in the colour of the flower rose (roosikarva prillid 'rose-coloured glasses' or roosivärvi prillid 'rose-coloured glasses'). In addition, there was one case of translation in each part of the study where the metaphor was preserved without a colour word as läbi lillede vaatama 'looking through flowers', which could also stem from similar semantic priming, as this expression refers to euphemistic, mild or indirect speech in Estonian. Strategies used to translate colour metaphor rose-coloured glasses are summarised in Figure 1.

The English 'blue in the face' can become blue, green, red, white, black or grey
The metaphor blue in the face is used where someone is exhausted from anger, strain, or other great effort (idioms.thefreedictionary.com). In Estonian, different colour words are used to refer to a similar emotion, with näost siniseks minema 'to go blue in the face' as in (2) (2) näost siniseks minema face-sg-elA blue-Trl go-inf 'to go blue in the face' and näost roheliseks minema 'to go green in the face' for getting angry (Õim 2007), and näost punaseks minema 'to go red in the face' from anger, strain and similar (Õim 2013).
The participants in the study also used different colour words, in both the context-free and the contextbased translation tasks. This was the colour metaphor for which the largest number of different colour words were used in translation, with blue, green, white, black and grey all used.
In the context-free part of the study blue and green were used, but white also appeared among the translation equivalents. In cases where the metaphor was preserved in translation without a colour word, the most frequently used choice was näost ära olema literally 'to be away from the face' 'not quite like yourself from the face' as in example (3) (3) näost ära olema face-sg-elA away from-ADV be-inf 'to be not quite like yourself from the face' To paraphrase the colour metaphor, the participants chose between vihane 'angry' and kurnatud 'exhausted' or 'pingutusest üleväsinud', literally 'overtired from exertion' in Estonian. In the context-free part vihast kahvatu 'pale from anger' and näost kaame 'whey faced' also appeared. Paraphrasing gave the translation equivalents of vihane 'angry', pingutusest kurnatud 'exhausted from great effort', pingutusest väsinud 'tired from making effort'. In the context-based translation task, the colour words blue and green were used most frequently, but the colour word black (must in Estonian) also emerged to translate the colour metaphor as in example (4) (4) kuni maa on must until-prep ground-sg.nOM be-3.sg black 'until the ground is black' meaning so much effort has been made that even the grass turns from green to black.
The colour word grey was used in a metaphor for hair turning grey from extreme effort as in example (5) (5) kuni oleme peast hallid until-prep be-1pl head-sg-elA grey-pl 'until our hair turns grey' The translation equivalents where no colour words were preserved in Estonian were relatively ingenious, such as 'until we are out of air' (kuni õhk saab otsa) and 'we become breathless' (pingutame hingetuks), or very culture specific in nagu Vändrast saelaudu 'like producing wooden boards in Vändra', which is a reference to a famous sawmill in Estonia by which the participant compared the legislation process to the excessive manufacturing of consumer goods. Paraphrasing the colour metaphor resulted in translations lõputult 'endlessly' and nii palju, kui tahame 'as much as we want'.
However, what seems to be noteworthy is the only case of mistranslation of this colour metaphor in the experiment, which occurred in the context-free task. One of the participants translated it vägivallatunnustega 'someone with visible marks of violence' which suggests suffering an injury resulting in bruises, which are most commonly described in Estonian by the word sinikas, a blue mark, which is derived from the colour word blue -sinine. This lets us conclude that different associations might emerge in the translation of colour metaphors when there is no supporting context. Strategies used to translate colour metaphor blue in the face are summarised in Figure 2.

A yellow-bellied can become someone with a fat belly
In English, to be yellow-bellied means to be cowardly or easily scared. If you are yellow-bellied, you are not brave (Vocabulary.com). In Estonian, however, there is no similar metaphor with a colour word, and neither is yellow associated with cowardliness or any other negative meaning in Estonian. Rather, yellow seems to be the most positively perceived colour in Estonian culture. Therefore it is not surprising that the participants did not reproduce the metaphor in the target language with a colour word. The most frequently used translation strategy was to replace the colour metaphor with a metaphor without a colour word, notably argpüks (cowardly), a compound word of arg (cowardly) and püks(id) literally 'pants', which could be translated back into English as chicken-hearted. This was the most common translation equivalent used in both parts of the study. Likewise, participants who preferred to paraphrase the meaning of the metaphor into Estonian translated it as in example (6)   (6) keegi, kes pigem kardab, kui kahetseb someone who rather be afraid-3.sg than regret-3.sg 'someone who would rather be safe than sorry' However, it was interesting to find that the context might also serve as a negative factor and mislead the participants, as the subjects who encountered the metaphor for the first time in the context-based part of the study and decided not to look up for the meaning in the dictionary or online, tended to translate the metaphor incorrectly. Translation equivalents included käpardlik 'fumbling', loru 'goofy', aegunud 'outdated', kahepalgeline 'two-faced', ära hellitatud 'spoiled, used to getting their own way', omakasupüüdlik 'selfish, opportunistic', suure rasvase kõhuga 'with a big fat belly', and tundliku kõhuga 'with a sensitive belly'.
The tendency to misunderstand and furthermore to mistranslate the metaphor also occurred in the context-free translation task. Retrospective interviews with both groups revealed that this was a new metaphor for several participants. This meant they faced comprehension difficulties and had to look up the meaning in order to be confident about the translation. Participants who did not know the meaning and decided not to check were misled. This resulted in obvious mistranslations where yellow-bellied became someone with a big, fat and yellow belly (suure kollase rasvase kõhuga inimene in Estonian). The novelty of this colour metaphor is one of the reasons for the translation errors.
The second type of the cases of obvious mistranslation in the context-free translation task was where the opposite meaning to the original in the target language was used in Estonian as in example (7) (7) julge hundi rind on rasvane brave-gen wolf-.sg.gen chest be-3.sg greasy 'brave like a wolf' This metaphor was among the ones that was most frequently omitted in the test. Retrospective interviews revealed that this was also because it was a novel metaphor. Participants who looked up the meaning found some sources saying that being yellow-bellied meant being originally from Lincolnshire (The Phrase Finder). As this was too confusing, they decided to leave the phrase untranslated in order to avoid erroneous translations. Most of the participants in the context-free part of the study emphasised that supporting context would have helped them to translate the colour metaphors.
Strategies used to translate colour metaphor yellow-bellied person are summarised in Figure 3.

Red tape can refer to various different things from borderlines to blue movies in Estonian
The English colour metaphor red tape refers to rigid or mechanical adherence to bureaucratic rules and regulations, especially those involving unnecessary paperwork (Phrases.org.uk). In Estonian, the most commonly used gloss is bürokraatia 'bureaucracy' (Keeleveeb) or bürokraatlikkus 'bureaucracy-like', or paberlik asjaajamine 'paper procedures' (Hanko and Liiv 1998: 596). However, in English-Estonian dictionaries one possible translation equivalent is a literal translation of red tape as punane lint. Although the modern metaphor has departed from its original English meaning it has a clear etymological source, because Red ribbon was once used by government employees and lawyers to tie up bundles of legal documents. Before any official business could be done, the red tape had to be cut. (Idiom Origins).
In Estonian the most accurate translation would be bürokraatia 'bureaucracy', which refers to narrowminded formal procedures and attitudes.
In the context-based translation task, 27 of the 33 participants translated red tape as bureaucracy. One of the translators used a literal translation as punane lint with the colour word red, which is a dictionary equivalent but does not convey the original meaning in the target culture. Metaphorical translation referred to juuksekarva lõhki ajama 'splitting hairs' as in example (8) (8) juuksekarva lõhki ajama hair-sg.prT to split 'splitting hairs' while paraphrases included liiga palju seadusega reguleeritud 'overregulation in the law' as in example (9) (9) liiga palju seadusega reguleeritud too much law-COM regulate-pAss.pTCp 'overregulated by the law' There were three cases of mistranslation in which the phrase was translated as allajoonimine 'underlining', varjamine 'hiding', and ilustamine 'glamorising'.
In the context free translation task the most commonly used translation equivalent was also bürokraatia 'bureaucracy' or liigne bürkraatia 'too much bureaucracy'. The colour metaphor was paraphrased as ametlikud reeglid 'official rules' or reeglid, mis tunduvad ebavajalikud 'rules that seem unnecessary'. Mistranslations included piirjoon 'borderline' or 'a borderline that cannot be crossed', which could be the result of semantic priming to national administrative borderlines, perhaps during the Soviet era. Other mistranslations were täiskasvanute film 'blue movie' and keelatud 'something forbidden'. These might be the cases of colour associations where red is interpreted as a colour forbidding or warning against certain types of behaviour, for example in traffic. Equally, punane joon 'red line' or punane teip, which is literally 'red sticking tape', punane vaip 'red carpet', and punane piir 'red border' were examples of mistranslation of the colour metaphor red tape.
Strategies used to translate colour metaphor red tape are summarised in Figure 4.

Silver lining can become silver mountain in Estonian
Kolahdouz considers the colour metaphor silver lining as meaning something valuable that helps to ease difficult situations (2015: 5). This is a phrase said to emphasise that every difficult or unpleasant situation has some advantage (Cambridge Dictionary). The English-Estonian idiom dictionary presents a translation equivalent of lootusekiir 'a ray of hope', pointing out that it comes from the proverb every cloud has a silver lining (Hanko and Liiv 1998: 356). There is no reference to a similar metaphor with a colour word in Estonian, so the results of the translation tasks were not easy to predict, like with the colour metaphor yellow-bellied person.
In the context-based translation task, the original colour word silver was preserved once, when the metaphor was translated literally as hõbedane vooder 'silver lining'. It can be argued whether this conveys the original meaning of the metaphor in Estonian sufficiently for the reader to understand the reasoning behind it. The most popular translation strategy, in 70% of the cases, was to preserve metaphor but without a colour word. The translation solutions were lootuskiir, eredam valgus pilveäärel, õnnelik lõpp and tulevik võib olla helgem, which could be translated back into English as 'a ray of hope', 'more vivid light on the edge of a cloud', 'happy ending' and 'the future may be brighter'. To paraphrase the meaning into Estonian, olukorda on võimalik lahendada 'it is possible to solve the situation' was used as in example (10) (10) olukorda on võimalik lahendada situation-sg.prT be-3.sg possbile solve-inf 'it is possible to solve the situation' There was one case of mistranslation, where the subject translated the metaphor as võltsreaalsus 'false reality'. However, in the context-free translation task obvious mistranslation occurred in a quarter of cases. The example translations of silver lining included petlik mulje 'false impression' and hõbedane taust 'silver background'. Like in the context-based part of the experiment, lootusekiir 'ray of hope' was presented as a translation equivalent, preserving a metaphor in the target language. A more creative example where a metaphor was preserved was terake suhkrut 'a tiny grain of sugar'. Paraphrasing without a supporting context resulted in the translation equivalent igas halvas on midagi head 'there is something good in every bad thing'.
Strategies used to translate colour metaphor silver lining are summarised in Figure 5.

Discussion
The phenomenon of metaphor has been widely discussed within the discipline of translation studies, predominantly with respect to translatability. It has been argued that metaphors can become a translation problem, since transferring them from one language and culture to another "may be hampered by linguistic and cultural differences" (Schäffner 2004(Schäffner , 1253. Bennett emphasises that it is especially difficult to consider colour metaphors in different languages as different languages identify different things around us, filter our experience of them, and characterise things differently (Bennett 1981, 273). Within this study we aimed to describe how English colour metaphors are translated into Estonian. We carried out context-based and context-free translation tasks with volunteer participants to draw conclusions about the importance of context in the process of metaphor comprehension and translation. In addition to that, we aimed to elucidate whether language and cultural competence influence the translation of colour metaphors.
We based our study on descriptive translation theory and cognitive metaphor theory. Our interest lies in actual language use and so we conducted cognitive translation tasks. To generalise the results and describe how participants translated English colour metaphors into Estonian we compiled a descriptive classification of translation strategies that was explained in the Results section. Rather than focusing on quantified data, which might not be representative at such a small scale, we focus our discussion on specific examples that present relevant information for qualitative analysis and suggest grounds for further research. An overview of the translation strategies used is presented in Table 3.
It is important for the aims of this article to focus on the obvious cases of mistranslation. The data collected suggest that in a context-free task, where only single phrases were presented to participants without any additional context, obvious mistranslation occurred more than two and a half times as often as in the context-based task. It is clear from our data that the colour metaphors that are lexicalised in both languages with a similar colour word are translated similarly in both parts of the experiment. A good example is the colour metaphor rose-coloured glasses, where the most frequent translation strategy was to preserve the metaphor in translation with a colour word. Such metaphors do not seem to pose problems for comprehension or for translation, with the single exception of translating blue in the face in the context-free part of the current study.
On the other hand, the more unusual the metaphor, the more varied the strategies to translate it. Such metaphors seem to be more difficult to translate than ones that are lexicalised. Yellow-bellied person and silver lining exemplify this in this study, as the strategies used are more versatile and cases of obvious mistranslation start to occur. There are also cases in which dictionaries might give confusing results, and there are two such examples in our study. The first interpreted a yellow-bellied person as someone from Lincolnshire, which was often mentioned in post-experiment interviews as a finding from explanatory dictionaries that caused confusion, and the second was the translation of the colour metaphor red tape, where the literal translation presented in a bilingual dictionary provided confusing information causing obvious mistranslations that did not convey the original meaning. Reasons for omitting the metaphor in translation were manifold. In the context-based part of the study participants explained during the interviews that they had managed to find other ways of presenting the original meaning in the target text and decided not to reproduce the metaphor or to paraphrase its meaning. In the context-free translation task, omissions were caused by insecurity about the meaning of the colour metaphor. In one interview a participant was confused about the online meaning of a yellow-bellied person being from Linconshire, which the participant said made even less sense than his own suggestion for the meaning, and in the end omitted it. The greatest difference in the use of translation strategies was in cases of obvious mistranslation, which were caused by the novelty of the colour metaphors and a lack of translation experience or training.
It has been argued before that dictionaries are often inaccurate when it comes to colour concepts (Tavast and Uusküla 2015;Wyler 1992: 116-122). As dictionaries rely on the intuition of the dictionary compilers it is not surprising that some cases might lead to mismatches and misunderstanding with actual language use for specific metaphors. This also calls for further attention from dictionary compilers, who could improve the information given in bilingual dictionaries in the future. We recommend that compilers of bilingual dictionaries pay attention to presenting translation equivalents to culturally specific colour metaphors. In academic literature, it is not always encouraged to interpret the literal translation of metaphors as inaccurate, but rather to see it as presenting novel metaphors. It has been argued that if a metaphor triggers different associations in two cultures, a literal translation should be avoided, unless the culture specificity of the source language is to be emphasised, in which case the addition of an explanation is advised (see Schäffner 2004, Bazzanella et al. 2016 for further discussion). We agree with Schäffner and argue, however, that using literal translations to translate colour metaphors, especially when there is no supporting context, hinders comprehension and misguides the reader, though word-for-word translation can hardly be expected to get the full meaning across (Meriläinen et. al 2016: 118, see also Samaniego Fernández 2013 for further discussion). The use of a literal translation for the metaphor red tape in our study serves as a clear example of that. From the data gathered during the context-free and context-based translation tasks it is clearly evident that context plays an important role in translating culturally specific items like colour metaphors. This is apparent from the analysis of the translation strategies used in both parts of the experiment and it was stated by the majority of the participants during the interviews conducted after the translation task. Furthermore, it was also highlighted by the participants in the context-free part of the study, as several participants mentioned that context would have helped them in understanding as well as in translating the phrases that they had not encountered before or for which they were not certain about the meaning.
However, it should be noted that there are cases where context can hinder comprehension of the colour metaphor, like with the example of yellow-bellied, which was mistranslated in several texts in the sense of a lazy, goofy or out-dated person. Such obvious mistranslations occurred because, as stated in postexperiment interviews, some of the participants were encountering the metaphor for the first time and decided not to check its meaning. They stated that they drew conclusions about its meaning solely from the context. This aspect of colour metaphor translation deserves further analysis and more empirical research.
Translating can also trigger semantic priming (Estes and Glucksberg 1998: 338). That some metaphors triggered semantic priming can be discerned, especially in the context-free translation task, where rosecoloured glasses in English was translated as roosikarva or roosivärvi prillid, meaning (literally 'rosecoloured glasses' in the sense 'glasses of the colour of the flower rose'. Equally, red tape in English triggered a word-to-word translation into punane teip 'red masking tape'. One of the aims of the study was to analyse whether professional translators had different results from the participants with no previous knowledge of translation. It has been claimed that both translation training and experience influence the decisions made during translating (Van Doorslaer 2015: 239).
Qualitative differences appeared, as translators were more target culture oriented and non-translators more source culture oriented for example. In post-experiment interviews, the translators claimed that they focused on their readers and aimed to present translation equivalents that would be easier to understand in the target culture. We noticed that experienced translators also offered more creative translations than people with no previous experience and translation training did.
The analysis also revealed that translators approached the process differently, and in the context-based translation task for example, they often validated the phrasing in the target language by searching for how that phrase was used in Estonian. In the context-free translation task, they wrote down several possible translation choices so they could select between different possible translation solutions. This allows us to reach the conclusion that translation experience and training matters.
Moreover, it was evident that cultural competence influences how participants comprehend and transfer colour metaphors from one language into another. One-fifth of the participants in the context-free part of the study and one-third of the participants in the context-based part live or have lived in a source language context. Participants who live or have lived in an English-speaking country faced fewer difficulties in comprehending and translating colour metaphors. No obvious mistranslation occurred among these participants.

Conclusion
Researching culture-specific concepts empirically allows language in use to be analysed, and running an experiment on colour metaphor translation provided several interesting results. Unfortunately, gender differences could not be discerned because the male sample was small, but intriguingly, age, education and knowledge of foreign languages had no substantial impact on the translation outcome. However, participants who live or have lived in an English-speaking country faced fewer difficulties in translating colour metaphors. This indicates that the language environment supports comprehension of culture-specific phenomena like metaphors. The importance of context should not be underestimated in translation. The study emphasised that the translation of colour metaphors can often result in a fallacy being produced, because of the linguistic and cultural differences between languages. Several examples from our database illustrate both cultural differences between languages and the importance of context in translation, as a yellow-bellied person in English can become 'someone with a fat belly' in Estonian or red tape could be interpreted as 'an adult movie'. The most important conclusion is that the more novel and original the metaphor is, the more varied are the strategies used when translating colour metaphors. In addition, the opposite applies, since the subjects translated the more lexicalised metaphors using similar strategies and similar colour words. Our study also confirms that professional translators and participants with translation training face fewer difficulties in metaphor translation tasks. Moreover, it was also evident that professional translators approach metaphor translation tasks differently, so it is clear that both translation education and experience matter. Further research is necessary for conclusions to be drawn about the translation process of colour metaphors.