1. Introduction
Linguistics is fundamental for the study of emotion. The language we use to talk about emotion provides valuable information about affective experience itself (see Soriano, Reference Soriano, Schiewer, Altarriba and Ng2022 for an overview of affective meaning in language). One important vantage point in the study of emotion is the lexicon, because emotion words tell us about the ways different lingual communities around the world conceptualize emotional experiences.
The number and meaning of emotion words can greatly vary from language to language, suggesting a fair amount of variation in the ways different communities encode their affective experiences. Quantitatively, differences can be observed between sizeable (ca. 400–2,000 words) emotion lexicons catalogued in contemporary languages like English (Russell, Reference Russell1980; Wallace & Carson, Reference Wallace and Carson1973), Dutch (Hoekstra, Reference Hoekstra1986), Chinese (Boucher, Reference Boucher, Marsella, Tharp and Ciborowski1979; Zheng et al., Reference Zheng, Zhang, Guo, Guasch and Ferré2023) or Czech (Slaměník & Hurychová, Reference Slaměník, Hurychová and Kebza2006) and considerably smaller affective vocabularies, typically counting mere dozens of lexemes, documented in languages such as Dalabon (Ponsonnet, Reference Ponsonnet2014), Palauan (Smith & Tkel-Sbal, Reference Smith, Tkel-Sbal, Russell, Fernandez-Dols, Manstead and Wellenkamp1995) or Wolenian (Lutz, Reference Lutz1982; for further examples, see Ogarkova, Reference Ogarkova and Meiselman2021, p. 914). Likewise, in some languages, specific areas of the affective space may be either hyper- or hypocognized (Levy, Reference Levy1973) – that is, have either multiple or very scarce, if any, labels denoting nuances of a specific emotional experience (Li et al., Reference Li, Wang and Fisher2004; Russell, Reference Russell1991a; Shaver et al., Reference Shaver, Wu, Schwartz and Clark1992). Furthermore, there are also numerous accounts of lexical lacunae, or emotion words without one-word translation in other languages, like German Sehnsucht (Scheibe et al., Reference Scheibe, Blanchard-Fields, Wiest and Freund2011),Footnote 1 Greek stenahoria and ypohreosi (Panayiotou, Reference Panayiotou2004), Portuguese saudade (Neto & Mullet, Reference Neto and Mullet2014) or Japanese amae (Niiya et al., Reference Niiya, Ellsworth and Yamaguchi2006). All these observations suggest the existence of affective experiences that are significantly more salient in some communities than in others.
Qualitatively, translation equivalence of emotion words across languages is also known to be a matter of degree (Pavlenko, Reference Pavlenko2008). For example, seemingly uncontroversial translation-equivalents, such as English shame and Spanish vergüenza, were reported to have important differences in meaning (Hurtado de Mendoza et al., Reference Hurtado de Mendoza, Fernández-Dols, Parrott and Carrera2010), suggesting that the indigenous concepts denoted by those terms are not equivalent. Sometimes, such semantic differences can be traced back to cultural traits, as with ‘happiness’ in English and Japanese (Ishii, Reference Ishii, Fontaine, Scherer and Soriano2013), or ‘shame’ across cultures (Silfver et al., Reference Silfver, Fontaine, Dillen, Scherer, Fontaine, Scherer and Soriano2013). Cultural profiles can also explain semantic differences within the same language as spoken in different regions, as with the term orgoglio (‘pride’) in southern and northern Italy (Mortillaro et al., Reference Mortillaro, Ricci-Bitti, Bellelli, Galati, Fontaine, Scherer and Soriano2013), or orgulho (‘pride’) and raiva (‘anger’) in European and Brazilian Portuguese (Soares da Silva, Reference Soares da Silva2020, Reference Soares da Silva and Soares da Silva2021, Reference Soares da Silva, Schröder, Mendes de Oliveira and Tenuta2022).
While most previous studies on semantic similarities and differences in the meanings of translation equivalents have focused on comparisons of emotion words in two (or, less frequently, several) languages, one area that has rarely been tapped upon is the meaning of emotion cognates – that is, emotion words from different languages that, stemming from a common origin, have retained similar spellings, pronunciations and meanings (e.g., English joy vs. French joie). Among the few relevant studies, Soriano et al. (Reference Soriano, Fontaine, Scherer, Celle and Lansari2015) found that the French word surprise refers to an emotional experience that is comparatively more sudden, more expressive and more short-lived than the one denoted by English surprise. Likewise, in a comparative analysis of English despair and Spanish desesperación, Alonso-Arbiol et al. (Reference Alonso-Arbiol, Soriano, van de Vijver, Fontaine, Scherer and Soriano2013) established that, while the former appears to be a low arousal emotion, Spanish desesperación is high in arousal.
The issue of identifying differences in the meanings of cognate emotion terms is not trivial because, in emotion psychology, there is frequently a tacit assumption that cognate emotions words would be best translation equivalents, and emotion researchers frequently rely on such dictionary translations in their experimental or conceptual work. However, in practice, cognate terms may not refer to the same kind of affective experience, which may undermine the validity of the intended investigation.
To address this concern and to fill the aforementioned gap in research on emotion cognates, this study explores the interesting case of English frustration compared to its cognates in Spanish (frustración), French (frustration) and German (Frustration)Footnote 2. What makes this case particularly interesting is that, according to several linguistic accounts, English frustration may be a rather unique concept (Besemeres & Wierzbicka, Reference Besemeres, Wierzbicka, Blachnio and Przepiorka2009; Wierzbicka, Reference Wierzbicka1999). As argued by Anna Wierzbicka (Reference Wierzbicka1999), ‘… frustration [sic] is a highly culture-specific concept, very characteristic of modern Anglo culture, with its emphasis on goals, plans, and expected achievements’ (p. 72). This specificity of English frustration seems to be supported by its untranslatability into Greek (Panayiotou, Reference Panayiotou2004) or Russian (Pavlenko, Reference Pavlenko2008) and the claim that, in languages other than English, frustration ‘exists only as a relatively recent loan from English (Frustration in German, frustracja in Polish, frustracija in Russian, frustrasi in Bahasa Indonesia, and so on)’ (Wierzbicka, Reference Wierzbicka1999, p. 72).
And yet, according to the Cambridge and Collins bilingual dictionaries,Footnote 3 as well as the online neural translator Deepl,Footnote 4 the translation of English frustration into Spanish, French and German seems fairly straightforward, since cognate terms are available (frustración, frustration and Frustration, respectively); moreover, they are listed as first translation equivalents of the English term. In addition, as revealed by the Europarl parallel corpus (around 60 million words from European Parliament proceedings) (Koehn, Reference Koehn2005), using the OPUS automatic word alignment search engineFootnote 5 (Tiedemann, Reference Tiedemann2012), Spanish frustración, French frustration and German Frustration are also the most frequent translation equivalents of English frustration (see Table 1).
Table 1. Translations of English frustration in Europarl corpus (OPUS search engine)
n = number of occurrences in corpus.
So here the question arises: is the meaning of ‘frustration’ the same in these languages? What is the extent of the alleged specificity of English frustration? Or, in other words, in spite of them being translation equivalents, does English frustration differ significantly in meaning from its cognate terms in Spanish, French and German?
In what follows, we explore these questions in two psycholinguistic (Studies 1 and 2) and two linguistic studies (Studies 3 and 4) employing elicitation and observational methods, respectively. First, we present a psycholinguistic study designed to identify the most typical types of anger in different languages (Study 1). Second, we present the semantic profile of the different ‘frustration’ words based on self-reported ratings of the features of those words using the GRID instrument (Fontaine et al., Reference Fontaine, Scherer and Soriano2013a) (Study 2). Next, we present two types of corpus-based data: an analysis of the metaphorical profiles (Ogarkova & Soriano, Reference Ogarkova, Soriano, Musolff, MacArthur and Pagani2014) of ‘frustration’ in English and Spanish (Study 3) and an analysis of sentence co-occurrences and usage-based synonyms of ‘frustration’ in the four target languages (Study 4). We will follow up with a discussion of the findings and their possible motivation and conclude with an account of the article’s limitations and an overview of the implications of our findings for emotion psychology.
2. Study 1 (emotion labeling)
In this section, we present unpublished results from an earlier study (Ogarkova et al., Reference Ogarkova, Soriano, Lehr and Wilson2012) designed to investigate the centrality of different anger terms in different languages. The study adopted an elicitation approach and requested native speakers of different languages to label the emotion they would feel in a number of proposed scenarios. Different scenarios were built using a facet-approach (Elison, Reference Elison2005) to represent varied situations susceptible of eliciting one of four types of conflict-related emotions – namely, anger, shame, guilt and pride. We report results about anger only.
Native speakers of English (n = 11), Spanish (n = 17), French (n = 12) and German (n = 17) read five anger-eliciting scenarios and, for each of them, reported as many words in their language as they wanted in the grammatical class of their choice (nouns, adjectives, or both) to name what they would feel in those situations. The scenarios reflected different combinations of the ‘facets’ of an anger scene, such as who the wrongdoer and the disadvantaged party are (e.g., the emoter/another person/nobody in particular), the nature of the anger-eliciting event (e.g., one-time/repeated) and whether or not the wrongdoing was intentional. For instance, a scenario with no identifiable human wrongdoer was ‘My computer crashed and I was not able to finish the work to meet a crucial deadline’. Table 2 reports the most frequent words in the anger family mentioned by the participants in each language, collapsing across scenarios.
Table 2. Words elicited in an emotion-labeling task with anger scenarios (excerpt)
n = number of observations collapsing across grammatical class (words exhibit the most prevalent grammatical class observed for the root). In each language, the most frequent term and the ‘frustration’ word are highlighted in bold.
As shown in Table 2, the most frequently mentioned words were words typical for the category anger in each language: anger in English, colère in French, Wut in German and rabia in Spanish. However, ‘frustration’ was significantly more frequently mentioned in English than in any of the other three languages. In fact, English frustration was mentioned just as frequently as angry (no significant statistical difference with angry, χ2 (1) = 0.0212, p = n.s.), whereas ‘frustration’ in the other three languages was mentioned significantly less frequently than the top anger word (Spanish, χ2 (1) = 16, p ≤ 0.0001; French, χ2 (1) = 11.56, p = 0.0007; German, χ2 (1) = 13.714, p = 0.0002).
These findings are congruent with several previous observations on the cognitive availability of English frustration. In a free-listing task where English-speaking participants were asked to write down salient exemplars of the category emotion, frustration was the 11th most frequent term among 383 words listed by the respondents (Fehr & Russell, Reference Fehr and Russell1984, p. 469). In addition to being a typical emotion word, Russell and Fehr (Reference Russell and Fehr1994, p. 191) also reported frustration to be the most frequently reported term in the category anger in a free-listing task (mentioned by a third of a sample of 317 respondents).
Corpus-based analyses of language use provide congruent results revealing the salience of ‘frustration’ in English compared to other languages. ‘Frustration’ seems to be a more common word in English than in Spanish, French or German. A search for the terms in the comparable variants of the TenTen corporaFootnote 6 (Jakubíček et al., Reference Jakubíček, Kilgarriff, Kovář, Rychlý and Suchomel2013), with over 10 billion words each, using Sketch Engine (Kilgarriff et al., Reference Kilgarriff, Baisa, Bušta, Jakubíček, Kovář, Michelfeit, Rychlý and Suchomel2014) as search engine, reveals English frustration to be more frequently used than its cognate counterparts (Table 3).
Table 3. Frequency of term ‘frustration’ in four languages in comparable TenTen corpora
Together, the observations reported in this section suggest that ‘frustration’ is a more salient anger term in English compared to Spanish, French and German. The relative distance to typical anger words in each language also suggests that English frustration is closer in meaning to English anger than the cognate ‘frustration’ terms to typical anger words in the other three languages. Congruent results will be presented in the following section concerning the GRID study.
3. Study 2 (GRID profiles)
In this section, we present unpublished results of a study within the large cross-cultural framework of the GRID project (Fontaine et al., Reference Fontaine, Scherer and Soriano2013a). The GRID is an international research collaboration led by psychologists and linguists to investigate the meaning of emotion words and the structure of affective space across languages and cultures. Using the GRID questionnaire, native speakers are asked to rate the extent to which a series of proposed features are part of the meaning of emotion words in their language. The features are extracted from a variety of psychology theories of emotion and are grouped under ‘emotion components’ (the different types of experience that together ‘compose’ or give rise to an emotional episode) – namely, cognitive appraisals, physiological changes, expressions (facial, gestural, etc.), action tendencies and subjective feelings. The average ratings across participants in a given language, or across languages, enable researchers to establish semantic profiles for the terms, as well as to explore similarities and differences among them, their emotion component structure, and the dimensionality of the affective space (Fontaine et al., Reference Fontaine, Scherer, Soriano, Fontaine, Scherer and Soriano2013b).
The original GRID study investigated 24 emotion terms selected to represent all areas of affective space (Fontaine et al., Reference Fontaine, Scherer, Soriano, Fontaine, Scherer and Soriano2013b). As a follow-up, we developed a more targeted instrument (ELIN) to investigate types of emotion in the same family, and more concretely types of four conflict-relevant emotion categories: anger, shame, guilt and pride (Soriano et al., Reference Soriano, Fontaine, Ogarkova, Mejía, Volkova, Ionova, Shakhovskyy, Fontaine, Scherer and Soriano2013). The ELIN-GRID study was conducted in several languages; here, we report the findings regarding the representation of anger words specifically in Spanish (n = 83), French (n = 91), German (n = 44) and English (n = 36).
The ELIN-GRID pool of words in each language is shown in Table 4. These words (n = 28) were selected on the basis of the elicitation study reported in Section 2 (Ogarkova et al., Reference Ogarkova, Soriano, Lehr and Wilson2012), choosing the most frequently listed emotion words in each language and favoring cognate terms whose meaning we were interested in, like ‘frustration’. Their grammatical class reflected the most frequently observed for that lemma in the elicitation study.
Table 4. anger words investigated in the ELIN-GRID study in English, Spanish, French and German
The words were rated on 95 features pertaining to all emotion components, in addition to other sociocultural variables like degree of social acceptability, or frequency of the affective experience labeled by the word (see Soriano et al., Reference Soriano, Fontaine, Ogarkova, Mejía, Volkova, Ionova, Shakhovskyy, Fontaine, Scherer and Soriano2013 for details). Each feature was rated on a nine-point Likert-like scale ranging from 1 (‘extremely unlikely’) to 9 (‘extremely likely’), where five occupied the middle, neutral point of the scale (‘neither likely, nor unlikely’). Average ratings per language were mean-centered for comparison across samples (to control for differences in scale use).
To analyze the data, first a hierarchical cluster analysis (HCA) was run to investigate the internal similarity structure of each language (see Table 5). The results showed that, in all languages but English, ‘frustration’ constitutes a cluster of its own, suggesting that the word is different in meaning from the others, which are comparatively closer to one another than they are to ‘frustration’.
Table 5. HCA of anger words per language based on ELIN-GRID feature profiles
En = English, Fr = French, Ge = German, Sp = Spanish. ‘Frustration’ words in bold.
A cluster analysis of all samples together, conducted to explore the overall similarity structure of their common space (see Figure 1), confirmed this observation. At the highest level of the dendrogram, two clusters emerged: a small one with four words, including ‘frustration’ in Spanish, French and German plus English resentment, and a larger one with the remaining 24 anger words. Said differently, English frustration clustered with typical anger terms in the four languages, such as English angry and annoyed, Spanish ira and rabia, French colère, or German Wut, Zorn and Ärger (Durst, Reference Durst, Harkins and Wierzbicka2001; Oster, Reference Oster2014).Footnote 7
Figure 1. HCA of anger words across languages based on ELIN-GRID feature profiles.
Note: E = English, Fr = French, Ger = German, Sp = Spanish.
A similar pattern emerged from Pearson profile correlations that were calculated to identify the closest terms to ‘frustration’ in each language (Table 6). The results showed that, while the closest synonyms were never the most typical anger terms of each language, the proximity of ‘frustration’ to them was greater in English (angry 0.746, annoyed 0.826) than in Spanish (ira 0.271, rabia 0.416), French (colère 0.590), or German (Wut 0.142, Zorn 0.406, Ärger 0.127).
Table 6. Pearson profile correlations of ‘frustration’ words with other anger words in each language based on ELIN-GRID feature profiles
Note: Correlations of ‘frustration’ words with prototypical anger terms are in bold. EN = English, SP = Spanish, FR = French, GE = German.
To further elucidate in what specific ways English frustration was different from ‘frustration’ in the other three languages, and which features it shared with typical anger terms in the four languages, an ANOVA with cluster means as dependent variables and features as independent variables was run to determine the specific GRID features that differentiated the two clusters. A total of 19 features emerged as significantly differentiating (see Table 7). The greatest contrasts (i.e., features that scored in one direction in one cluster and the opposite direction in the other) indicate that ‘frustration’ in Spanish, French and German is characterized by feelings and expressive behaviors typical of low power (feeling weak [#1], exhibiting a slumped bodily posture [#9], lowering the head [#11]) and is more of an individual affair (experienced alone, rather than with other people [#17, 19]). By contrast, English frustration and the words in its cluster are characterized by expressive high-power behavior (speaking loud [#6] and fast [#7], lifting one’s chin [#10], pushing the chest forward [#8], exaggerating one’s emotion [#14]) and are more of a social experience (experienced with other people [#17, 18], triggered by a third person [#15]). The existence of another person as agent that causes the emotion (#15) is a prototypical anger feature (Ellsworth & Smith, Reference Ellsworth and Smith1988) characterizing only the first cluster of terms in our data. This means that the presence of an external agent doing something wrong is less likely in Spanish, French and German ‘frustration’ than in English.
Table 7. Differences between the two overall anger clusters based on ELIN-GRID feature profiles (ANOVA)
Note: Only features significant after Bonferroni correction. M1 = mean cluster 1, M2 = mean cluster 2. Bold for the most differentiating features, where each cluster behaves in opposite ways.
Other differences between both clusters are a matter of degree. In all languages, the person feels disadvantaged in some way (#2), which has an impact on their self-image (#16) and may make the person want to withdraw (#12) and not be seen (#13); but these low-power feelings and behaviors are significantly more typical in Spanish, French and German frustration words than in English. Conversely, confrontation rather than withdrawal may be a more typical response in the English sense of ‘frustration’. Regarding physiology, ‘frustration’ exhibits anger-like symptoms in all languages. Anger in general is known to induce heat and redness (Fetterman et al., Reference Fetterman, Robinson, Gordon and Elliot2011), and congruently, the features coldness (#4) and paleness (#5) are rated as atypical of ‘frustration’ in all languages. However, in English, the rejection is significantly stronger.
Together, the feature profile differences reported above suggest that English frustration is a better example of prototypical anger than its cognate terms in Spanish, French and German. More specifically, the meanings of the cluster of words to which English frustration belongs appear to be close to the conceptual prototype of anger as suggested in both psychology (Ellsworth & Smith, Reference Ellsworth and Smith1988; Kuppens, Reference Kuppens, Sander and Scherer2009; Russell, Reference Russell1991b; Russell & Fehr, Reference Russell and Fehr1989, Reference Russell and Fehr1994) and linguistics (Lakoff & Kövecses, Reference Lakoff, Kövecses, Holland and Quinn1987; Wierzbicka, Reference Wierzbicka1999, pp. 88–89; see also Fries, Reference Fries, Börner and Vogel2004; Oster, Reference Oster2014; Soriano, Reference Soriano, Fontaine, Scherer and Soriano2013; Soriano et al., Reference Soriano, Fontaine, Ogarkova, Mejía, Volkova, Ionova, Shakhovskyy, Fontaine, Scherer and Soriano2013; Weigand, Reference Weigand and Weigand1998). The anger prototype (understood as a broad category, applicable cross-culturally) is an emotion triggered by a stimulus (an event or state of affairs, but typically the actions of an external agent) assessed as bad (e.g., an injustice, an offense) for the person or for relevant others, during which the emoter feels physiologically activated and compelled to oppose the perceived wrongs, often through violence or aggression (e.g., Bender et al., Reference Bender, Spada, Seitz, Swoboda and Traber2007; Fernandez et al., Reference Fernandez, Carrera, Paez, Alonso-Arbiol, Campos and Basabe2014; Ogarkova & Soriano, Reference Ogarkova, Soriano, Schiewer, Altarriba and Ng2022). From a dimensional point of view, the emotion is characterized as negative, aroused and high in power (e.g., Scherer & Fontaine, Reference Scherer, Fontaine, Fontaine, Scherer and Soriano2013). This high power is one of the most important characteristics of anger, differentiating it from other negative and aroused emotions like fear. The results reported in this section suggest that, unlike its cognates in Spanish, French and German, the English term frustration shares more similarity with the prototypical high-power anger experience outlined above.
4. Study 3 (metaphorical profiles)
In this section, we report results from a linguistic observational study of the metaphors used to represent frustration in English and Spanish.Footnote 8 The study fits within the framework of conceptual metaphor theory (CMT) in cognitive linguistics, which assumes that patterns in figurative language use are indicative of underlying conceptual associations – referred to as conceptual metaphors – between conceptual domains. The general claim is that abstract concepts, like emotion concepts, are cognitively represented in terms of more concrete physical experiences like temperature, force or containment (Lakoff, Reference Lakoff and Ortony1993; Lakoff & Kövecses, Reference Lakoff, Kövecses, Holland and Quinn1987).
The literature on conceptual metaphors for the representation of anger in English and Spanish is abundant (Barcelona, Reference Barcelona, Labrador Gutiérrez, Sainz de la Maza and Viejo García1988; Kövecses et al., Reference Kövecses, Szelid, Nucz, Blanco-Carrión, Akkök, Szabó, Heredia and Cieślicka2015; Lakoff & Kövecses, Reference Lakoff, Kövecses, Holland and Quinn1987; Ogarkova & Soriano, Reference Ogarkova, Soriano, Musolff, MacArthur and Pagani2014; Ogarkova et al., Reference Ogarkova, Soriano, Gladkova, Piquer-Píriz and Alejo-González2018; Soriano, Reference Soriano2003, Reference Soriano2005; Suarez Campos, Reference Suarez Campos2020). Here, we report unpublished results from a study of types of anger (including ‘frustration’) in those languages (Ogarkova & Soriano, Reference Ogarkova, Soriano, Musolff, MacArthur and Pagani2014). The study constitutes a quantitative, corpus-based approach to the metaphorical expressions used with seven anger words in English (anger, rage, fury, irritation, indignation, resentment and frustration) and six such words in Spanish (ira, rabia, furia, irritación, indignación and frustración).Footnote 9
One thousand random Key Word In Context (KWIC) citations were extracted for each word from the British National Corpus (BNC) and Corpus del Español. Footnote 10 When needed to reach that number, additional citations were culled from the Bank of English and Corpus de Referencia del Español Actual (CREA).Footnote 11 The careful manual analysis of these expressions into metaphorical patterns (MP) (Stefanowitsch, Reference Stefanowitsch, Stefanowitsch and Gries2006), their classification into conceptual metaphors, and the quantification of MP tokens and types resulted in the creation of ‘metaphorical profiles’ for each word (see Ogarkova & Soriano, Reference Ogarkova, Soriano, Musolff, MacArthur and Pagani2014 for details). The metaphorical profile of a term is composed by the number of metaphorical patterns observed for each conceptual metaphor identified for that term. Since a common inventory of metaphors was observed for all anger terms (Ogarkova & Soriano, Reference Ogarkova, Soriano, Musolff, MacArthur and Pagani2014), the vector of frequencies per term allowed us to statistically explore both commonalities and differences in the way those anger types are represented via metaphor.
To explore similarities within each language, we calculated profile correlations of ‘frustration’ with the other terms in English and Spanish separately. As shown in Table 8, while English frustration enjoyed reasonably high profile correlations with the other six anger words in English (above 0.7 in most cases), the highest correlation was found with the word anger itself, with an extremely high coefficient (0.918). This means that, in terms of metaphorical behavior, English anger and frustration behave in the same way. By contrast, in Spanish, frustración has a low profile correlation with ira, a typical anger word and frequent translation of English anger. The correlation is better with more moderate forms of anger like irritación, indignación and rabia.
Table 8. Pearson profile correlations of ‘frustration’ words with other anger words in English and Spanish based on metaphorical profiles
The differences in profile correlations suggest that the metaphors used to talk about frustration and anger in English are more similar than the metaphors used to talk about frustración and ira in Spanish. To capture some of these differences we investigated the relative frequency of metaphors for ‘anger’ and ‘frustration’ in English and SpanishFootnote 12 (Table 9).
Table 9. Frequency of expressions (n tokens) of different metaphorical source domains in the representation of ‘anger’ and ‘frustration’ in English and Spanish
A series of Fisher exact tests (with Bonferroni correction to control for multiple comparisons) revealed interesting contrasts. For example, English anger and frustration exploit the metaphors opponent (1), force of nature (2) and insanity (3) to the same degree (p = ns).
This is not the case in Spanish, though. In Spanish, ira is more associated than frustration to force of nature (4) (p < 0.001) and insanity (5) (p < 0.001), whereas frustration is more associated with opponent (6) instead (p < 0.001). In addition, ira is significantly more associated with fire (7) (p < 0.001) and frustración to illness (8) (p < 0.001).
The preference of Spanish ira for expressions involving insanity, force of nature and fire and frustración for illness and opponent, underscore their very different nature, more virulent and uncontrolled in the case of ira. By contrast, English frustration is indistinguishable from English anger in the use of intense and virulent source domains like force of nature or insanity. To the extent that patterns in metaphorical language can reveal differences in conceptualization – as assumed in CMT – these linguistic differences reveal a different conceptualization of frustration in English compared to Spanish, and one that brings English frustration comparatively closer than its Spanish counterpart to prototypical anger concepts.
5. Study 4 (corpus analysis)
Our last study offers a different corpus-based analysis of the representation of ‘frustration’ across languages. In order to identify the emotion concepts that are semantically similar to ‘frustration’ in each language, we first looked at near-synonyms of the target words in the thesauruses of the online platform Sketchengine (Kilgarriff et al., Reference Kilgarriff, Baisa, Bušta, Jakubíček, Kovář, Michelfeit, Rychlý and Suchomel2014), which are built on the basis of usage patterns in corpora. A search of the word ‘frustration’ in the TenTen corpora (Jakubíček et al., Reference Jakubíček, Kilgarriff, Kovář, Rychlý and Suchomel2013) in English (EnTenTen2021), Spanish (EsTenTen18), French (FrTenTen20) and German (GeTenTen20) provided converging evidence for the specificity of English frustration discussed so far. As shown in Table 10, the closest synonym of English frustration is anger; by contrast, in the remaining three languages, no anger terms were found among the top five synonyms of ‘frustration’; instead, top positions were occupied by words from the sadness (e.g., deception, Unzufriendeheit) or fear (e.g., angustia) families.
Table 10. Top five noun synonyms of word ‘frustration’ in the four languages in Sketch Engine thesauruses (TenTen corpora)
Another approach to explore the representation of ‘frustration’ in the four languages is to look at sentence co-occurrences. Co-occurrence at sentence level can be considered a measure of semantic relatedness. It indicates words that frequently co-occur with the target word in the same sentence. Emotion words of this type are likely to designate affective experiences coherent with ‘frustration’. Accordingly, we extracted significant (log-likelihood) sentence co-occurrences for the four ‘frustration’ words from the English, Spanish, French and German News corpora in the Leipzig corpora collection.Footnote 13 The results complemented the picture afforded by the Sketchengine thesaurus in the TenTen corpora presented above.
As shown in Table 11, the most frequent emotion co-occurrence of ‘frustration’ in all languages was an anger word (anger, rabia, colère, Wut); however, in Spanish, French and German, ‘frustration’ co-occurred with a variety of other emotions too (marked in grey cells), whereas, in English, it co-occurred only with anger. This suggests that the meaning of the word ‘frustration’ is indeed coherent with anger in all languages, but the link is more specific and exclusive in English.
Table 11. Ten most frequent sentence co-occurrences of word ‘frustration’ in English, German, French and Spanish in News corpora of the Leipzig corpora collection
Note: Values in brackets indicate co-occurrence frequency counts. Gray cells indicate emotion labels (words directly naming emotions). The most frequent emotion co-occurrence in each language is highlighted in bold.
The suggestion that, in languages other than English, ‘frustration’ can refer to emotions other than anger is supported by dictionary definitions of the corresponding lexemes. In the definitions provided by the Merriam-Webster (9), Cambridge (10) and Collins (11) English dictionaries,Footnote 14 English frustration is explained through anger words (anger and annoyance). If we compare these definitions with the definitions of the verb frustrate in the Oxford English Dictionary and Middle English Dictionary of the University of Michigan,Footnote 15 the English term seems to have specialized in meaning over the years from mere ‘goal-obstruction’ to specifically ‘anger’.Footnote 16
By contrast, in Spanish, French and German, ‘frustration’ can refer to other emotions too. The reason is that, according to the definitions in the DRAE (12), Larousse (13) and Duden (14) dictionaries,Footnote 17 the meaning of ‘frustration’ in those languages profiles goal obstruction (i.e., thwarting of desires), rather than one specific emotion. When an emotion is mentioned, though, it is not ‘anger’, but rather emotions in the sadness family like ‘dissatisfaction’ (12) or ‘disappointment’ (14). Notions of ‘degradation’ (14) and ‘failure’ (12) are mentioned too.
In sum, a corpus analysis of usage-based synonyms and sentence co-occurrences of the four ‘frustration’ words, corroborated by dictionary evidence, suggests that English frustration is semantically related only to the emotion family anger, whereas its cognates in the other three languages are also associated with other emotions (in the sadness and fear families).
6. Discussion
Taken together, the observations reported in Sections 2–5 congruently support the claim that ‘frustration’ is uniquely conceptualized in English as compared to Spanish, French or German. The first reason is that it seems to be a comparatively more salient emotion concept. The second reason is that, compared to its cognates, English frustration is closer in meaning to prototypical anger, characterized by a high-power profile.
But why would English frustration, unlike its cognates in the other three languages, come to mean ‘anger’ in English? A look at emotion psychology allows us to explain this process. According to psychology appraisal theory (e.g., Ellsworth, Reference Ellsworth2013; Lazarus, Reference Lazarus1991; Scherer, Reference Scherer, Scherer, Schorr and Johnstone2001, Reference Scherer2009), different emotions are the result of different configurations of appraisals, or cognitive assessments. Among the key appraisals necessary to experience emotion, we find ‘goal congruence’, or the assessment of whether a particular event or stimulus supports or else blocks our goals (i.e., whether the emotion-eliciting event is ‘goal-conducive’ vs. ‘goal-obstructive’). Typically, appraisals of goal-conduciveness lead to positive emotions, whereas appraisals of goal-obstructiveness lead to negative ones (cf. Scherer & Moors, Reference Scherer and Moors2019). ‘Goal-obstruction’ is what the term frustration originally referred to: being impeded or blocked at somethingFootnote 18. This sense is still retained in the technical use of the term frustration in psychology (Berkowitz, Reference Berkowitz1989, Reference Berkowitz, Sander and Scherer2009).
According to Klaus Scherer’s CPM appraisal theory (Scherer, Reference Scherer, Scherer and Ekman1984, Reference Scherer, Scherer, Schorr and Johnstone2001), at a subsequent stage in the process, the brain also engages in another appraisal: our coping potential vis-à-vis the event, that is, our capacity to deal with (e.g., to adjust to or to change) the event or its consequences. In cases of goal-obstruction, an appraisal of high coping-potential leads to anger, and an appraisal of low coping-potential leads to sadness or fear.Footnote 19 To put it simply, in cases where something happens that is bad for us, we typically feel anger if we appraise that we can do something about it, and we feel sadness or fear if we appraise that we cannot.
This means that the English word frustration, whose meaning originally profiled goal-obstruction only and was emotion-agnostic, incorporated at some point an additional nuance of high coping-potential and became biased toward a specific type of emotion: anger. In the other languages, it has remained rooted in goal-obstruction only; therefore, ‘frustration’ in Spanish, English and German can refer both to anger (high coping potential) or to sadness and fear (low coping potential), or simply to underspecified negative affect (goal-obstruction, as shown in Study 4). What the GRID results in Study 2 revealed as well is that, when Spanish, French and German cognates do refer to anger, it is to a version with a low-power profile, that is, to a less prototypical form of this emotion.
But why is high coping potential, and therefore anger, salient only in the English meaning of ‘frustration’? Since cognitive appraisal preferences may be affected by culture (Bender et al., Reference Bender, Spada, Seitz, Swoboda and Traber2007, p. 199), a possible explanation may be found in the cultural traits of influential English-speaking communities like the UK and USA. According to Wierzbicka (Reference Wierzbicka1999), Anglo-Saxon cultures are characterized by ‘achievement expectations’ (p. 72). In addition, the UK and USA have marked profiles in several of Hofstede’s (Reference Hofstede2001) cultural dimensions – namely, individualism (which means that these cultures promote a focus on the person, rather than collective orientation), low power distance (so they reject power imbalance) and masculinity (they support assertiveness, strength, a ‘can-do’ attitude and competition). All these cultural traits may conspire so that, in Anglo-Saxon cultures, a goal obstruction (i.e., a threat to individual goals, which is important to individualistic communities) may be more likely met with an appraisal of high-coping potential (promoted by masculine cultures) and opposition (due to low tolerance for being disempowered). Accordingly, it is possible that, in those cultures, ‘literal frustration’ (i.e., goal obstruction) more likely leads to anger by default, and the meaning of the word frustration therefore shifts to represent the emotion experience it most frequently designates in practice.
7. Conclusion
‘Frustration’ cognates in English, Spanish, French and German are quoted as translation equivalents in dictionaries and used as translation-equivalent terms in practice. However, the data reported in this article evidence they do not mean the same.
Specifically, our findings suggest that English frustration is a clear anger word (Study 4), a good example of the anger prototype because of its high-power profile (Study 2), close in meaning to anger (Studies 2 and 3) and cognitively accessible to speakers of English and thus frequently used in the language (Study 1).
By contrast, the cognates of English frustration in the other three languages are comparatively less frequently used and less conceptually salient (Study 1), further apart in meaning from typical anger words in their respective languages (Studies 2 and 3), poor examples of the prototypical anger category because of their low-power profile (Study 2) and more open-ended as to the emotion family they may refer to (anger, sadness, fear, or simply unspecified negative affect) (Study 4).
As a way to interpret these findings, we explained how cognitive appraisals may relate to the observed semantic differences. We pose that, in all cases, ‘frustration’ refers to goal obstruction. The difference lies in the representation of another appraisal, coping potential, which we posit to be high in English and low in the other languages. Since prototypical anger is a high-power type of emotion, ‘frustration’ emerges as a typical form of anger in English only. In the other languages, it refers either to a non-prototypical form of anger (low coping potential) or to other (low-coping potential) emotions (e.g., varieties of sadness).
This study has several limitations. First, our claims regarding power and the specific appraisals of goal congruence and coping potential are based on indirect (albeit congruent) evidence. To further substantiate these claims, it would be desirable to collect direct measurements of the dimensional (especially power) and appraisal profile of ‘frustration’ in the studied languages. Complementary direct evidence of the meaning of ‘frustration’ could also be collected with a corpus-based usage-feature analysis (e.g., Glynn, Reference Glynn, Glynn and Robinson2014; Soares da Silva, Reference Soares da Silva2020) of each term in the four languages.
Another limitation is that the study describes ‘default’ or ‘average’ meanings. Yet, the word ‘frustration’ could still mean exactly the same in a given context in the four languages and be the most appropriate translation choice between them. This is underscored by the observation that the terms that appear to have a closer semantic profile to English frustration in the three other languages can vary slightly depending on the methodology used for comparison and the specific data employed (observational vs. elicited data, metaphorical construal similarity vs. usage-based synonyms, type of corpus, etc.). Therefore, we should not make sweeping statements about the non-equivalence of the different ‘frustration’ words in practice, and be prepared instead to evaluate said equivalence for translation on a case-by-case basis and in specific contexts. This, however, is fully compatible with the goal of our article, which was to raise awareness of the semantic differences that exist by default, which had not been sufficiently discussed in earlier literature.
Indeed, one of the novelties of our study is that previous observations on the ‘uniqueness’ of English frustration were made based on comparisons with languages lacking a ‘frustration’ word, such as Greek or Russian (frustration as a non-lexicalized concept). Our work extends that research to languages with cognate terms, demonstrating that, even when cognate terms are available, English frustration has a unique meaning different from ‘frustration’ in the other languages.
Another novelty of our study is the type of data employed. Earlier studies pointing out the uniqueness of English frustration were based on introspection and intuitive observations of trained linguists, who were also native speakers of the languages concerned. We complete their accounts with quantitative data from two psycholinguistic and two linguistic studies employing corpus analyses and self-report. This makes our study the first quantitative empirical investigation of the semantic uniqueness of English frustration, based on converging findings from different disciplinary and methodological approaches.
To conclude, our finding that ‘frustration’ does not mean the same across languages should constitute a word of caution to emotion researchers relying on linguistic cues for experimental or conceptual work: even cognate terms that are used as translation equivalents in practice may not refer to the same kind of affective experience in ways that can jeopardize the validity of the intended investigation.
Open access
