DOI 10.1515/ip-2013-0026 Intercultural Pragmatics 2013; 10(4): 533 – 568 Valentina Apresjan Corpus methods in pragmatics: The case of English and Russian emotions Abstract: The present paper is a comparative corpus study of the verbal expression of emotional etiquette in American English and Russian. The study is conducted against the backdrop of certain assumptions regarding the cross-cultural centrality and marginality of emotions as formulated in the current research on cross-cultural pragmatics. The paper employs corpus-based methods to test the frequencies of the linguistic expression of diferent types of emotions in Russian and American English as encountered in diagnostic contexts of irst-person reporting. Contrary to many currently accepted theories, the present study demonstrates no absolute prevalence of positive or ethical over negative or non-ethical emotions in Russian or American English. It also disproves certain more speciic claims (the predominance of ‘pity’ in Russian), while conirming others (prominence of ‘shame’ in Russian). Certain tendencies in emotional etiquette lean toward cross-cultural universality (e.g., ‘gratitude’ as the most frequently expressed emotion), while others difer. Overall, Russian speakers tend to report more passive negative emotions (‘fear’), while English speakers prefer reporting active negative emotions (‘anger’). Russian speakers are more “self-deprecating” than English speakers, as they favor expressing ‘shame’ over ‘pride.’ At the same time, they show less empathy with the addressee, reporting more ‘contempt’-like and less ‘pity’-like emotions. The results obtained in this study can be useful for understanding and formulating culturally speciic pragmatic peculiarities and hence preferred conversational strategies in the two languages. Keywords: corpus research, emotional etiquette, emotion cluster, emotion subtype, etiquette maxim, frequency Valentina Apresjan: E-mail: vapresyan@hse.ru 1 Introduction If your boyfriend told you he wouldn’t be coming to your birthday party because of a conlicting commitment would you be reluctant to tell him that it hurt your feelings? Is it acceptable to openly admit to feeling jealousy? Is it a faux pas to put one’s venom into words? Would verbalizing anger make others raise their Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 534 Valentina Apresjan eyebrows? Is it polite to express pity verbally? Socially acceptable emotional behavior tends to difer from culture to culture and one has to learn to verbalize it as part of one’s general linguistic competence. Although this information is indispensable, it can neither be found in dictionaries nor in grammar books; instead, it has to be intuited and accumulated from one’s communication with native speakers. The present paper is a cross-linguistic study of “linguistic emotional etiquette” or speakers’ attitudes toward expressing diferent types of emotions as relected in the linguistic usage of corresponding emotion words in Russian and English. The objective of this paper is to reconstruct the linguistic emotional etiquette inherent in the Russian and English languages and outline pragmatically appropriate strategies of linguistic behavior in the realm of emotion expression in each language respectively. Based on corpus research methods,1 this study examines and challenges the main existing assumptions about the pragmatic prominence or marginality of certain emotions (e.g., ‘pity’ and ‘longing’ in Russian), as well as the possible cultural scripts postulated underlying their linguistic expression (Bulygina and Shmelev 1997: 481–495; Levontina and Zalizniak 2001: 306–309; Shmelev 2002: 300, 404–410; Wierzbicka 2002, 2006). This study considers 12 types of emotions that are ultimately ranked according to their pragmatic acceptability. The main criterion for an emotion’s pragmatic acceptability (and indirectly, its linguistic prominence or marginality) is the frequency of its linguistic expression in certain diagnostic contexts, measured in a series of specially designed corpora-based linguistic tests. It relects how comfortable language speakers feel when verbally acknowledging that they are experiencing a certain emotion. An emotion’s pragmatic acceptability is a purely linguistic parameter, inluenced but not determined by multiple extralinguistic factors, e.g., (1) whether an emotion is negative or positive,2 (2) whether it is considered ethical or non-ethical,3 and others. 1 The study uses the Corpus of Contemporary American (http://corpus.byu.edu/coca/) and the Russian National Corpus (ruscorpora.ru). Therefore, whenever making a statement about the English language, the author refers to American English as relected in the respective corpus. 2 According to Ekman’s facial display rules, positive emotions are generally more expressible across cultures, with diferent cultures allowing for varying degree of negative emotion expression. Thus American culture is considered more tolerant toward expressing negative emotions than the Japanese (Ekman 1972: 225). 3 As in Aristotle’s Nicomachean ethics, where moral virtues govern emotional behavior (Broadie and Rowe 2002). Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 535 2 Methods Comparing lexical or grammatical items between two languages is a problematic undertaking due to the frequent absence of a direct correlation between lexicons and grammars. Additionally, comparing emotional vocabulary is notoriously dificult (Wierzbicka 1999a), as this is an area of signiicant conceptual disparities with no cross-linguistic equivalency even among translatable emotion words. Measuring an emotion’s pragmatic acceptability presents another challenge, since this linguistic parameter is considerably more complex than the mere frequency of the corresponding emotion term. This paper implements two methodological solutions to tackle these diiculties: 1. Systematic comparison by entire emotion clusters rather than by single words; 2. Frequency analysis applied to specially designed representative contexts. 2.1 Comparison by emotion clusters In language, every emotion is represented by an entire “emotion cluster,” or a variety of emotion terms relecting a range of emotion subtypes – related but subtly diferent feelings.4 Emotion clusters are deined on the basis of a prototypical scenario uniting all the emotional subtypes within it.5 The prototypical scenario incorporates the general cause of the feeling, the nature of the feeling, and the wishes of the experiencer, e.g., for the ‘SADNESS’ cluster, the prototypical scenario would be as in (1): (1) ‘something bad has happened; X feels bad because of this; X wants to, but cannot change the bad situation.’ It would unite such subtypes of ‘SADNESS’ as sadness, sorrow, grief, upset, etc., with the nature and scale of the bad event and the accompanying feeling varying from emotion to emotion. Thus, emotion terms comprising a cluster would share a core semantic component while varying according to a number of important features. 4 The notion of emotion cluster and the method of cross-linguistic comparison by emotion clusters have been developed and motivated in (Apresjan 2011a, b). Emotion clusters are marked by ‘CAPITALS’ in single quotation marks, and emotion subtypes represented by actual words are given in italics. 5 The method of describing emotions through a reference to prototypical scenarios has been introduced and employed in the works of (Iordanskaja 1971; Iordanskaja 1984) and (Wierzbicka 1990; Wierzbicka 1992; Wierzbicka 1999a). Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 536 Valentina Apresjan The cause of the feeling is the main characteristic discriminating among larger emotion clusters; consider, for example ‘SADNESS’ and ‘JOY’ clusters which are opposed on the basis of whether an event is bad (‘SADNESS’) or good (‘JOY’), ‘SHAME’ and ‘PRIDE’ clusters which are opposed on the basis of whether the experiencer did something bad (‘SHAME’) or something good (‘PRIDE’), ‘FEAR’ and ‘SADNESS’ clusters which are opposed on the basis of whether the event is anticipated (‘FEAR’) or a fait accompli (‘SADNESS’). Generally, the stimulus determines the nature of the feeling: good happenings cause good feelings, and bad happenings cause bad feelings. However, there are certain exceptions to this; thus, the bad feeling of envy occurs when somebody has something good, and the pleasant feeling of Schadenfreude (a subtype of ‘JOY’), occurs when something bad happens to somebody. When the stimuli coincide, the nature of the feeling and the associated wishes provide further distinguishing grounds. Thus, Schadenfreude and the entire ‘PITY’ cluster occur in the same circumstances (when something bad happens to somebody) but cause the opposite feelings – pleasant in the case of Schadenfreude, unpleasant in the case of ‘PITY.’ Or, ‘ANGER’ and ‘HURT FEELINGS,’ which are both bad feelings caused by someone’s bad actions toward the experiencer, difer with respect to the latter’s wishes; while ‘ANGER’ involves an active wish to hurt the object of emotion in response, ‘HURT FEELINGS’ cluster (to be hurt, to be injured, to be ofended, to resent) imply a more passive attitude. Discrimination among diferent emotion subtypes within clusters is based on a variety of semantic features,6 found in all clusters, such as the speciic causes of the emotion (e.g., the sufering of the object is a prerequisite for compassion); the object of the emotion (supernatural objects for terror, higher powers for awe); the depth and strength of the emotion (wrath is deeper and stronger than irritation); the duration of the emotion ( grief lasts longer than being upset); the speaker’s attitude to the emotion (venom is negatively evaluated, while indignation is considered justiied); the symptoms and behavioral manifestations of the emotion (rage and fury involve physical aggressiveness, whereas irritation is frequently manifested verbally), and others.7 Comparing emotion clusters, as opposed to individual emotion terms, provides more balanced and comprehensive grounds for analysis. For example, if the 6 Many of semantic features discriminating among diferent emotion subtypes have been formulated in the New Explanatory Dictionary of Russian Synonyms (entries GRUST’ ‘SADNESS,’ SERDIT’SJA ‘TO GET ANGRY,’ STRAX ‘FEAR’ and others). 7 A list of emotion clusters, prototypical scenarios and emotion subtypes is in the Appendix. The list of emotion clusters and their component emotion subtypes has been originally suggested in (Apresjan 2010, 2011 a, b) and further elaborated in the present paper. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 537 English anger were compared to its Russian dictionary equivalent gnev ‘anger’ out of the context of the entire ‘ANGER’ cluster,8 one might come to the conclusion that English and Russian concepts of ‘anger’ are totally diferent. While anger describes a very wide range of feelings of difering strength and depth that can be caused by a variety of situations, gnev can only refer to a strong, sweeping, yet morally justiied feeling caused by a very serious ofense. It is experienced by a subject with a higher social status than the object of the emotion, most closely resembling the English wrath. In actual practice, the English noun anger is translated into Russian by a variety of lexical means depending on the strength, depth, cause, and nature of the feeling, as in the following examples. Note that only in one of the examples the irst dictionary equivalent gnev is used: (2) God’s anger gnev Boga wrath God-GEN ‘the wrath of God’ (3) the professor’s anger at the student who cheated on a test professor rasserdilsja na studenta professor angered-himself at student-ACC ‘The professor got mad at the student’ (4) the citizens’ anger at the raise of taxes nedovol’stvo graždan povyšeniem nalogov discontent citizen-PL-GEN rising-INSTR tax-PL-GEN ‘discontent of the citizens at the raise of taxes’ (5) the dog’s anger was apparent sobaka javno zlilas’ dog apparently raged-itself ‘The dog was snarling’ However, numerous specialized terms for diferent kinds of ‘anger’ and the absence of an umbrella term in Russian do not necessarily mean that English and Russian concepts of ‘anger’ are fundamentally diferent. Since adequate (though not word-for-word) translation is possible, it means that a similar conceptual space is covered by diferent linguistic expressions. Moreover, though the center 8 In this context, consider Wierzbicka’s apt observations on the diferences between English and Polish concepts of ‘anger’ (Wierzbicka 1999a: 32). Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 538 Valentina Apresjan of the semantic ield of ‘anger’ is lexically diferent in the English and Russian languages, its periphery demonstrates considerably less discrepancy in the emotion terms. Both languages have separate words (that are quite close translation equivalents) for the following subtypes of ‘anger’ (see also Table 1 in the Appendix): ‘strong violent uncontrolled anger,’ ‘anger as a reaction to an insigniicant unpleasant repeated stimulus,’ ‘righteous anger,’ ‘vicious anger.’ Thus, a systematic comparison of all types of ‘anger’ through the analysis of the entire ‘ANGER’ cluster would reveal signiicantly more underlying similarities and parallels between the two languages than one would discover by randomly comparing individual words. Another example of apparently unlikely linguistic parallels in emotion conceptualization that illustrates both the unexpected similarities and interesting contrasts comes from Japanese emotion term amae, discussed in detail in (Wierzbicka 1999a: 238–243). According to the study quoted in Wierbicka, amae is a certain kind of loving feeling that implies “helplessness and a desire to be loved,” a “sweet dependency.” Prototypically it is a feeling of a child toward its parents; when applied to a heteronormative romantic context, it normally describes the feelings of the female partner (for example, in Japanese pop songs). Wierzbicka’s own description of amae in terms of semantic primes supports its main semantic idea of feeling loved and protected by a strong and benevolent partner. While the feeling itself is understandable to a non-Japanese experiencer, at irst glance, it appears that lexically, this concept represents quite an exotic phenomenon. However, a similar (though by no means identical) concept seems to exist in English. Unlike traditional translations for the derived verb amaeru (to coax, to fondle), it is found in the verb to mother in its usages in romantic contexts in which it describes a protective, parent-like feeling and behavior toward a romantic partner. In combination with the verb want, it produces a semantic efect similar to that of amae, with the diference being that normally it involves a male, rather than a female, experiencer (the verb to father is not applicable to emotions, let alone to romantic feelings). Consider the following phrase, describing a desire for protective love from a romantic partner: (6) Many men want to be mothered by their wives Thus, a careful cluster analysis of the ‘LOVE’ cluster allows one to reveal unexpected aspects of how the emotion is expressed in diferent contexts. Overall, cluster analysis provides for a relatively objective perspective on the similarities and diferences between two languages in terms of their respective conceptualization of emotions. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 539 2.1.1 Cluster parallelism Each emotion cluster is deined according to a general scenario applicable to all of its subtypes – e.g., ‘ANGER’ = ‘feeling bad toward somebody when somebody did something bad and wanting to do something bad to that person’; for complete list of scenarios see Table 1 in the appendix, which summarizes the most important content of a number of emotion clusters and lists the major emotion subtypes in English and Russian.9 Cluster comparison reveals almost complete parallelism between the two languages, with infrequent exceptions, where an emotion subtype present in one language is lacking in another (e.g., the absence of horror10 subtype of ‘FEAR’ in Russian). Cross-linguistic similarities may be found also in the general principles of cluster organization. Table 1 demonstrates that most emotion clusters are organized around a neutral emotion subtype, with other possible subtypes varying according to the strength and depth of the feeling, its ethical or unethical dimensions, its physiological and behavioral manifestations, its un-/controllability, and its conventionalized social angles. Another universal factor in emotion cluster organization concerns ethical assessment: Most emotion clusters include a negatively evaluated emotion subtype where an unfavorable characterization originates from the speaker’s opinion that in the given circumstances the emotion is either inappropriate, unmotivated, or excessive. Consider the negatively lavored subtypes of ‘unethical fear’ (to poop out, to chicken out), ‘vicious anger’ (venom, spite, rancor), ‘excessive sadness’ (to be despondent, to be disheartened, to despair), ‘unethical joy’ (gloating), ‘excessive shame’ (self-conscious, shy, inhibited ), ‘excessive pride’ (to give oneself airs, to stick one’s nose up in the air), ‘inappropriate excessive pity’ (bleeding heart). 2.1.2 Cluster variation Along with all the similarities demonstrated at the beginning of Section 2.1, crosslinguistic variation is also strongly featured. Surprisingly, it is found not in the range of emotion subtypes, which are mostly invariable across languages, but in their quantitative value, though in this area, too, certain similarities exist. Each subtype comprises one or more emotion terms whose total frequency of 9 The scope of this paper does not necessitate the inclusion of an exhaustive list of all emotion subtypes; certain minor subtypes have been omitted from consideration. 10 Consider the diferential analysis of horror and fear in (Solomon 2004: 116). Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 540 Valentina Apresjan occurrence constitutes the respective “weight” of the corresponding emotion subtype within the cluster. Table 2 (see the Appendix) shows the respective “weights,” or percent ratios of diferent emotion subtypes within emotion clusters. The ‘ANGER’ cluster demonstrates one of the most striking quantitative diferences between the two languages. Conceptually, English and Russian languages distinguish among the same ‘ANGER’ types: ‘neutral anger’ (angry, anger, serdityj ‘angry,’ serdit’sja ‘to be angry’), ‘righteous anger’ (indignation, wrath, negodovanie ‘indignation,’ vozmuščenie ‘indignation’), ‘violent anger’ ( fury, rage, jarost’ ‘fury,’ bešenstvo ‘rage’), ‘vicious anger’ (venom, spite, rancor, zloba ‘venom, spite,’ zlost’ ‘anger, rancor’), and ‘anger at a repeated stimulus’ (annoyance, irritation, razdraženie ‘irritation’). However, while English speakers mostly mention neutral anger (approximately 40%) and violent anger (approximately 25%), Russian speakers reference mostly ‘righteous anger’ (more than 30%) and ‘anger at a repeated stimulus’ (almost 30%). Another example of discrepancy is in ‘SADNESS’ cluster, where English favors the neutral subtype, while predominant ‘SADNESS’ subtypes in Russian are ‘long profound sadness caused by a serious loss’ ( gore ‘grief,’ pečal’ ‘sorrow’) and ‘sadness without a cause’ (toska ‘longing, yearning for something unattainable, blues’), a tendency that is in accordance with both Wierzbicka’s observations on the Russian emotional world and the notorious Russian pessimism.11 Cultural diferences are also responsible for the discrepancies in the ‘JOY’ cluster. While it is natural for Russian speakers to “over-notice” and exaggerate negative emotions (as is the case with ‘SADNESS’), positive emotions may be “under-noticed” and downplayed, which accounts for the predominance of moderate, rational joy (udovletvorenie ‘satisfaction,’ dovol’nyj ‘content’) over other types. English, on the other hand, with its tendency to over-notice and exaggerate positive emotions, demonstrates the predominance of strong joy (happy) over all other types, including neutral joy.12 The organization of ‘SHAME’ and ‘PITY’ clusters demonstrates, on the other hand, strong parallelism between the two languages. The predominance of ‘social shame’ over ‘neutral shame’ and ‘ethical shame’ in both English and Russian is probably due to the overwhelmingly social nature of ‘SHAME’ in general. The prevalence of the compassion-type ‘PITY’ in both languages may be explained by the non-humiliating nature of that particular emotion subtype. 11 See (Wierzbicka 1990; Rancour-Laferriere 1995: 91). 12 See (Wierzbicka 1992: 299) on weakened usages of English happiness. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 541 2.2 Linguistic testing This paper relies on corpus-based methods to attain its objective; i.e., to reconstruct the linguistic emotional etiquette in Russian and English. This is conditioned by the very nature of pragmatic research, which deies the more traditional methods of linguistic experimentation or at least renders them insuicient. Consider phrases like I am gloating, which are certainly not ungrammatical, but not entrenched in actual usage. In order to assess the pragmatic acceptability of linguistic expressions, one has to study them in actual usage, hence the corpusbased statistical approach. However, merely measuring the general distribution frequencies of the corresponding emotion terms is not suicient as these frequencies only relect how oten language speakers talk about this or that emotion but not how appropriate it is to verbalize the emotion when experiencing it. For example, the term angry has the frequency of 53 per million usages, according to the Corpus of Contemporary American, while grateful has 20 per million usages. Does this mean that people feel less constrained in verbalizing their anger than gratitude? That is certainly improbable and, indeed, if one compares their irstperson usages, such as I am grateful, I was grateful vs. I am angry, I was angry, the picture changes radically. Out of the total 8,988 usages of grateful 1,247, or 14 percent, are in irst-person constructions, whereas only 800 out of 24,037, or 3 percent of the usages of angry are thus used. This means that rather than the general frequency of an emotion term only certain contexts and usages are indicative of its pragmatic appropriateness. The obvious solution to this problem is testing the frequency distribution of linguistic items that are used to express emotions experienced irst-hand rather than to talk about them as experienced by other people. However, it turns out that not all such items can function as suitable criteria for assessing a corresponding emotion’s pragmatic acceptability. For example, such common means of emotion expression as emotional interjections (wow for surprise or admiration, oh boy for surprise or frustration, yuck for disgust, etc.) are not reliable as markers of pragmatic acceptability. First, they are not unique to a single emotion type and frequently “cover” more than one emotion cluster. Second, interjections largely retain their connection with uncontrolled emotional vocalizations and are therefore sometimes uttered involuntarily. The only interjection that answers the criteria of controllability and uniqueness is the marker of ‘GRATITUDE’ thank you (or its variant thanks) and it is therefore taken into account. However, the main linguistic means of emotion expression on which the present analysis is based, is constituted by ixed constructions with terms that directly name emotions, such as I am afraid, I would like to express my gratitude, It makes me happy. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 542 Valentina Apresjan The following three constructions have been selected as linguistic indicators of an emotion’s pragmatic acceptability: I. I would like to express my X, and Russian Ja xoču vyrazit’ svoj X, where X is a noun denoting an emotion, as in English I would like to express my gratitude; Russian Ja xoču vyrazit’ svoju blagodarnost; II. I am X-ing, I am/feel/get X, and Russian Ja X-uju, Mne X-ovo, where X is a verb or an adjective denoting an emotion, as in English I am rejoicing, I am sad, I feel happy, I got angry; Russian Ja rad, Mne grustno, Ja sčastliv, Ja seržus’; III. It X-s me, It makes me X, It causes X in me and Russian Eto menja X-ujet, Eto vyzyvajet u menja X, where X is either a causative verb denoting the induction of emotion, or a noun or an adjective denoting an emotion, as in English It angers me, It makes me indignant; Russian Eto menja serdit, Eto vyzyvaet u menja negodovanie. The rationale for choosing these constructions is that they are grammatically and semantically the most universal: Nearly all emotion words may occur in them. Thus, the low frequency of emotion word usage in these constructions would point to pragmatic unacceptability, since there is no grammatical or semantic restriction on their co-occurrence. Other constructions speciic to emotion expression incorporate additional semantic components that limit the usage of emotion words in them for reasons other than pragmatics. Consider for example the construction to be in X, where X denotes an emotion: He was in shock, John was in a panic, Mary was in hysterics. The construction to be in X requires the emotion to be strong; terms denoting medium-strength and weak emotions cannot occur in it (e.g., the semantic anomaly of *She was in irritation). A possible objection might be that these constructions are stylistically heterogeneous, with the irst construction lavored as somewhat oicial and mostly used for public pronouncements and declarations, and the second and third neutral or colloquial. However, this register variation may indeed present an advantage because it means that the suggested testing provides for situational diversity in the investigation of the emotional etiquette. While it is clear that the irst construction is rarer than the second and the third, its compatibility with an emotion term points to the highest possible degree of pragmatic acceptability of that emotion. Rather than a mere acknowledgment of emotion, it is used for its oicial announcement to the audience, within language-appropriate emotional etiquette strategies. Another potential objection to the choice of diagnostic constructions might be their potential syntactic ambiguity. Indeed, with the Russian tendency for pro-drop in conversational speech, it would seem that the occasional legitimate absence of the irst-person experiencer subject might seriously interfere with the Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 543 statistics. However, a closer look reveals that this is not the case. First of all, when the verb tense in the irst-person reporting test constructions is set as present, it eliminates the need to include the irst-person pronoun in the nominative in the search of (Ja) X-uju ‘I am X-ing’ verbal constructions, since the form of the person is fully inferable from the present form of the verb. Constructions of the third type with emotional causatives and an accusative-encoded experiencer object (Eto menja X-ujet ‘It X-s me’) do not drop the pronoun and thus do not hinder the statistics either. The only two types of constructions where the irst-person pronoun is dropped and is not fully pragmatically inferable are constructions with adjectival emotion terms, encoding either nominative or dative experiencer subjects, such as Ja rad ‘I-NOM glad = I am glad’ or Mne grustno ‘I-DAT sad = I am sad’, and (b) the construction with a verb in the past tense, which shows gender, but not person: Ja radovalsja ‘I rejoiced-MASC.’ However, the pronoun is only dropped when the speaker coincides with the experiencer, which means that only the irstperson pronoun is dropped; cf. both Mne grustno ‘I am sad’ and grustno ‘Sad’, Ja grustil ‘I felt blue’ and grustil ‘Felt blue’ as possible irst-person reporting of sadness; phrases without a pronoun can only be interpreted in the irst-person, but not in the second or third. Thus, solitary adjectives and past tense verbs, such as grustno, rad, grustil cannot be possibly construed as referring to someone else’s, and not the speaker’s feelings. Therefore, correctly set search parameters eliminate any ambiguity: The total number of irst-person reporting in adjectival and past tense verbal emotion constructions is calculated by subtracting the number of second and third person constructions (in which the experiencer subject cannot be dropped), as well as constructions with noun experiencer subjects from their total number. 2.3 Diiculties: Absence of direct translational correlates and polysemy The greatest diiculties facing the statistical comparison of emotions concern the absence of direct translation equivalents (lexical gaps) and polysemy of individual emotion terms and expressions. While the issue of word-to-word correspondence, or absence thereof is important (for full detail, see Table 1, Appendix), cluster-to-cluster comparison allows for the alignment of statistical comparison between the two languages. When there is a larger number of emotion terms in a single cluster, this results in a lower frequency for each individual term, and consequently, overall expression within an emotion cluster evens out. Statistics demonstrate that the overall Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 544 Valentina Apresjan frequency of a cluster’s expression is not directly correlated to the number of individual terms in a cluster. Polysemy is another factor that may hinder objective statistical analysis. Since many emotion terms are polysemous, statistical data can yield inlated usage results. Consider, for instance, the polysemous ‘PITY’ term sorry, which expresses regret or apology rather than ‘PITY’ in the majority of diagnostic contexts (Oops, I am sorry). Its Russian correlates žal’ and žal’ko are also polysemous, referring in their two other meanings to the feelings ‘regret’ and ‘greed’ respectively. Thus, the disambiguation of polysemy is required in order to obtain reliable statistical data. Disambiguating contexts are designed individually for each polysemous word, e.g., the disambiguation of sorry in constructions with copulative verbs requires a PP with preposition for plus an NP with an animate noun, such as I am sorry for Mary, to distinguish ‘PITY’ sorry from ‘REGRET’ sorry, as in I am sorry that we have to leave so early. For its Russian correlates žal’ and žal’ko disambiguation is even trickier, consider (7a) and (7b): (7a) Mne bylo žal’ <žal’ko> tebja I-DAT was sorry you-ACC ‘I was sorry for you’ (7b) Mne bylo žal’ <žal’ko> tebja budit’ I-DAT was sorry you-ACC wake up ‘I was sorry to wake you up’ The ‘PITY’ interpretation for Russian žal’ and žal’ko requires an NP with an animate noun in accusative, as in (7a), but as (7b) demonstrates, this is not suficient, since if the NP with an animate noun in accusative is followed by an ininitive, interpretation changes to ‘REGRET’; therefore, contexts such as (7b) are excluded from the count. 2.4 Corpus-based study 2.4.1 Search parameters The study uses two freely available corpora – the National Corpus of Russian Language (ruscorpora.ru) and the Corpus of Contemporary American English (http://corpus.byu.edu). Both corpora are balanced and fairly large. In the Russian National Corpus, the search is conducted in the Main Corpus, which contains about 230 million words, with written texts of diferent genres dating from the Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 545 XVIII to the present. Corpus of Contemporary American comprises 450 million words with written texts of diferent genres dating from the 1990s to the present. In order to align the search, the time period in the Russian National Corpus is set from 1990 to the present, which limits the number of words to 79 million. Frequencies are counted per million words. The distance parameter is set lexibly, with up to two words between the subject and the predicate in all the constructions, allowing for a copula and an intensiier to it in between: Mne bylo užasno grustno ‘I-DAT was terribly sad = I was feeling terribly sad.’ 2.4.2 Statistics As stated in 2.2, for the purposes of reconstructing emotional etiquette, statistical data is collected not for all usages and occurrences of emotion terms, but only for their appearances with the irst person singular pronoun “I” in the three constructions deemed most indicative of voluntary emotion expression. Frequency distributions of emotion terms in all emotion clusters are calculated in two ways – absolutely and relatively. Absolute frequency is measured per million usages and relects the occurrence ratio of emotion terms in test constructions against the total corpus. It shows the absolute pragmatic acceptability of corresponding emotions against the entire body of language. For example, for the cluster ‘JOY’ the following emotion terms are considered: joy, joyful, joyous; happy, happiness; glad, gladden; cheer, cheerful, to cheer; content, contentment, contented; satisied, satisfaction, to satisfy; pleased, to please; delight, to delight, delighted; excited, to excite, excitement; glee, gleeful; gloating, to gloat.13 Each occurs in one or more diagnostic test constructions, e.g. I would like to express my happiness (Construction I); I am <feel, get> happy (Construction II); It makes me happy (Construction III). The frequencies of their occurrences are calculated per million usages. Further, the frequencies of all diagnostic contexts in a cluster are added to produce the total frequency, or the index of absolute pragmatic acceptability of an emotion cluster. 13 The nouns pleasure and its Russian translation correlate udovol’stvije are excluded from this list because they are not true emotion terms. They mostly pertain either to sensations caused by the performance of certain activities (to eat with pleasure, to read with pleasure), that last as long as the activity lasts, or to the activities causing these sensations (carnal pleasures, Swimming is my main pleasure). Likewise, the verb to enjoy and the noun enjoyment are not true emotion terms: They describe either a pleasant sensation (I was enjoying my breakfast, He ate with enjoyment), or permanent attitude: his enjoyment of music; I enjoy ishing. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 546 Valentina Apresjan Thus the total frequency, or the absolute pragmatic acceptability index, of ‘JOY’ cluster is approximately 33 per million usages. However, in a way, the absolute pragmatic acceptability index is a synthetic parameter as it also indirectly relects the total frequency of emotion terms under consideration, and not merely their frequency in diagnostic contexts. Stated differently, if ‘JOY’ is generally mentioned more oten than ‘ENVY’, the irst person usages of ‘JOY’ in diagnostic contexts are also likely to outnumber the irst person usages of ‘ENVY’ in diagnostic contexts. Thus the absolute pragmatic acceptability index relects not only the pragmatic acceptability of the irst person expression of a given emotion, but also, indirectly, how frequently or rarely people talk about that emotion in general. In general, “rare” emotions may be mentioned less frequently in the irst person, as well.14 The measurement of relative frequency must therefore be introduced for the purpose of capturing the purely pragmatic constituent in statistics. Relative frequency is measured in percentages and relects the occurrence ratio of emotion terms in test constructions against the total occurrence of the corresponding emotion terms in the corpus. For example, the total frequency of the ‘JOY’ cluster is approximately 323 per million usages, whereas its irst person expression in diagnostic constructions is 33 per million usages; i.e., its relative pragmatic acceptability index is approximately 10% (10% of all usages of ‘JOY’ are in the irst person). Comparatively, the ‘SADNESS’ relative frequency is the ratio of 4 per million (irst person occurrences in diagnostic constructions) to 148 (total occurrences of ‘SADNESS’ per million), which yields the relative pragmatic acceptability index of 2.7%. Thus, though the absolute frequency of irst person usages in the ‘JOY’ cluster exceeds that of ‘SADNESS’ by more than eight times, its relative frequency is only 3.7 times as high as that of ‘SADNESS.’ At the same time, the total frequency of ‘JOY’ is twice as high as the total frequency of ‘SADNESS.’ The interpretation of this statistical data reveals that English speakers generally tend to notice joy somewhat more oten than sadness, but that they express joy considerably more frequently than sadness. Based on diferent frequency distributions of emotion terms in each emotion cluster, emotion clusters are arranged hierarchically, according to the absolute or relative pragmatic acceptability of corresponding emotions. 14 However, the frequency of irst person emotion reporting is by no means determined by the general frequency of emotion terms; this parameter is suiciently independent, as demonstrated by difering frequency hierarchies in Section 3.2. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 547 3 Results and discussion 3.1 General emotion experiencing Although the main objective of this study is the reconstruction of emotional etiquette in American English and Russian based on the assessment of irst person emotion reporting, total frequencies of emotion terms in these two languages are also of interest to the researcher. These relect which emotions get more noticed by each of these two languages, and this in turn may serve as an indication of difering emotional experiences inherent in the American and Russian cultures. Table 3 below provides the total frequencies of emotion clusters in the numbers of per million usages, i.e., the total frequencies of all emotion terms comprising each of the clusters. It relects how oten speakers generally talk about diferent emotions. The table yields the following conclusions: 1. In both languages, the top half of the hierarchy is occupied by emotions with what might be called high expressive potential, that is, “noticeable” emotions that are normally accompanied by distinctive behavioral and/or physiological manifestations – four basic emotions (‘FEAR,’ ‘ANGER,’ ‘JOY,’ and ‘SADNESS’); the most conventionalized of emotions (‘GRATITUDE’), and the main ethical emotion (‘SHAME’).15 Table 3: Total frequencies of emotion clusters Emotion cluster English (per million usages) Emotion cluster Russian (per million usages) ‘Gratitude’ ‘Fear’ ‘Joy’ ‘Anger’ ‘Sadness’ ‘Shame’ ‘Pride’ ‘Pity’ ‘Disgust’ ‘Contempt’ ‘Envy’ ‘Jealousy’ 437.9 414.5 341.3 171.4 148.4 120.4 80.7 39.9 38.9 20 16.4 7.8 ‘Joy’ ‘Fear’ ‘Sadness’ ‘Anger’ ‘Shame’ ‘Gratitude’ ‘Pride’ ‘Pity’ ‘Contempt’ ‘Envy’ ‘Disgust’ ‘Jealousy’ 866.8 663.8 383.3 378.6 237.04 192.04 103.2 87.4 56.8 49.6 46.1 34.9 15 Compare data in Wierzbicka (1999b: 36), in which emotional concepts roughly corresponding to afraid, ashamed, and angry are claimed to be candidates for cross-linguistic universality. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 548 Valentina Apresjan 2. Basic emotions of ‘FEAR’ and ‘JOY’ nearly top the hierarchy in both languages, as the most important triggers of successful survival, which efectively rests on avoiding bad things (‘FEAR’) and attaining good things (‘JOY’). 3. The ith basic emotion, ‘DISGUST,’ despite its distinctive physiological manifestations, is quite low in both hierarchies, which is likely related to its losing much of its survival value in the sterile modern world, and, for etiquette considerations (see section 3.2), people rarely displaying it openly. 4. Russian speakers generally register emotion more than English speakers, especially in ‘ANGER,’ ‘SHAME,’ ‘JOY,’ ‘SADNESS,’ ‘PITY,’ ‘ENVY,’ ‘CONTEMPT,’ and ‘JEALOUSY’ clusters; 5. ‘FEAR,’ ‘PRIDE,’ and ‘DISGUST,’ though still more frequent in Russian, cut a much more balanced igure between the two languages. 6. ‘GRATITUDE’ is the only emotion that is registered in English more frequently than in Russian. 7. ‘ANGER’ and ‘SADNESS,’ though reversely positioned in the English and Russian hierarchies, are still very close in numbers within each language, suggesting that active and passive emotional reactions to bad events occur with an almost equal frequency. 3.2 Emotional etiquette in English and Russian: General tendencies While the total frequencies of emotion terms primarily relect the expressive potential of the corresponding emotions, the frequencies of irst-person expression relect the result of the interaction of emotions’ expressive potential with the maxims of emotional etiquette. The latter may prohibit or limit the expression of certain emotions, as well as promote the expression of some others. Interestingly, despite cross-linguistic diferences in the total frequencies of emotion ixation, demonstrated in Section 3.1, the frequencies of irst person reporting of emotions in the diagnostic constructions, which are indicative of their pragmatic acceptability, yield very similar results for the two languages (Table 4 below). Though the English and Russian indices of absolute pragmatic acceptability are diferent for the same emotion types, their respective hierarchization is nearly identical. This suggests that emotional etiquettes in English and Russian are remarkably alike, at least in their general tendencies. In both languages, ‘GRATITUDE’ is by far the most frequently self-reported emotion: Its pragmatic acceptability index exceeds that of the next most frequently self-reported emotion – ‘JOY’ – more than tenfold in English and almost fourfold in Russian. The etiquette maxim, which promotes the expression of emo- Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 549 Table 4: Absolute frequencies of irst-person diagnostic constructions in the corpora – absolute pragmatic acceptability indices Emotion cluster English (per million usages) Emotion cluster Russian (per million usages) ‘Gratitude’ ‘Joy’ ‘Fear’ ‘Shame’ ‘Anger’ ‘Pride’ ‘Sadness’ ‘Pity’ ‘Disgust’ ‘Envy’ ‘Contempt’ ‘Jealousy’ 322.437 24.411 11.292 6.196 6.059 4.981 3.841 3.826 1.998 1.246 0.406 0.053 ‘Gratitude’ ‘Joy’ ‘Fear’ ‘Shame’ ‘Anger’ ‘Pride’ ‘Pity’ ‘Sadness’ ‘Envy’ ‘Contempt’ ‘Disgust’ ‘Jealousy’ 126.555 33.125 29.08 20.445 8.49 4.956 4.905 4.08 3.83 2.33 1.975 0.775 tions pleasant for the addressee, ‘GRATITUDE’ the irst among them, is dubbed the maxim of pleasing. However, it is interesting that the gap between ‘GRATITUDE’ and ‘JOY’ is so much wider in English, suggesting much stricter underlying requirements about the necessity of expressing gratitude and, thus, a much stronger presence of the maxim of pleasing in the English emotional etiquette. The high position of ‘JOY’ in the hierarchy shows that in the emotional etiquettes of both English and Russian, the expression of positive emotions may be quantitatively favored over the expression of negative ones. However, the relatively low position of another positive emotion, ‘PRIDE,’ evinces the existence of the maxim of modesty in the emotional etiquettes of both languages, which moderates the tendency to express positive emotions. The position of ‘FEAR’ in emotional experiences as a whole and in the emotional etiquettes in particular is largely invariable: It is the second most talked-about emotion and the third most frequently self-reported emotion in both languages. This position of ‘FEAR’ may be triggered by its status as the most basic of emotions biologically, which renders it the strongest candidate for crosslinguistic universality. However, there still is a pronounced diference in igures: Russian ‘FEAR,’ with its pragmatic acceptability index of 29.08, closely follows ‘JOY’ with its 33.125, whereas in English ‘FEAR’ lags far behind ‘JOY,’ with an acceptability index more than twice as low. In both languages, self-reporting ‘SHAME’ occurs more frequently than selfreporting ‘PRIDE.’ However, in Russian this preference is considerably more Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 550 Valentina Apresjan explicit: The pragmatic acceptability index for ‘SHAME’ exceeds that of ‘PRIDE’ more than fourfold; in English, the ‘SHAME’ index is only 1.2 times more than the ‘PRIDE’ index. Surprisingly, ‘ANGER’ is treated as a relatively acceptable emotion by the selfreporting criterion in both languages, although here again, there is a sharp quantitative contrast between English and Russian. While the English ‘SHAME’ and ‘ANGER’ display the statistically negligible diference of 0.137 in their indices, the index of Russian ‘SHAME’ more than twice exceeds that of Russian ‘ANGER.’ The lower half of the hierarchy is occupied by ‘SADNESS,’ ‘PITY,’ ‘DISGUST,’ ‘ENVY,’ ‘CONTEMPT,’ ‘JEALOUSY’ (in Russian, ‘PITY,’ ‘SADNESS,’ ‘ENVY,’ ‘CONTEMPT,’ ‘DISGUST,’ ‘JEALOUSY’). The diferences in pragmatic acceptability indices of ‘PITY’ and ‘SADNESS’ are statistically negligible in both languages. The relatively low position of both emotions in the hierarchy underscores yet another two maxims of the English and Russian emotional etiquettes: It is not deemed acceptable to openly display an emotion that can expose either its object (the maxim of non-humiliation) or its experiencer (the maxim of self-esteem) as potentially weak or unable to change the bad situation (a weak object for ‘PITY,’ a weak experiencer for ‘SADNESS’). In addition, ‘SADNESS’ may be infrequently selfreported because openly displaying one’s sadness can be considered a ceremonial burden on the addressee who might be forced into sharing an experiencer’s feeling when explicitly presented with it. ‘DISGUST,’ ‘ENVY,’ ‘CONTEMPT,’ and ‘JEALOUSY’ ill the last four lines in the hierarchies, representing the lowest indices. The statistic infrequency of ‘DISGUST’ may be explained by its relatively low occurrence in emotional experience in general. Finally, openly displaying ‘DISGUST’ is considered impolite, for fear of ofending the object, if the object is human, or the maker of the object, if the object is food, drink, etc. The low position of ‘CONTEMPT’ relects the tendency to avoid an open display of one’s feeling of superiority over the object of emotion inherent in both English and Russian emotional etiquettes. This is the same maxim of nonhumiliation that forbids ‘PITY’ to occupy a high position in the hierarchy of pragmatic acceptability. In both languages, ‘PITY’ still rates considerably higher than ‘CONTEMPT.’ Unlike ‘CONTEMPT,’ which is clearly critical of the object of emotion, ‘PITY’ usually implies a negative valuation only of the situation in which the object of emotion inds herself. In addition, ‘PITY’ oten involves good attitude toward the object of emotion as a result of his/her bad position. It is interesting to note that the ratios of pragmatic acceptability indices for ‘PITY’ and ‘CONTEMPT’ radically difer for English and Russian: while in the English emotional etiquette, ‘CONTEMPT’ is nine times less acceptable than ‘PITY,’ in Russian, ‘CONTEMPT’ is only twice less acceptable. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 551 ‘ENVY’ and ‘JEALOUSY’ are also rarely self-reported in both languages. It is possibly due to the fact that they imply a certain “inferiority complex”, an admission of the experiencer’s weak position, which contradicts the self-esteem maxim. This inferiority complex is additionally exacerbated by a bad attitude toward the object of emotion and the distinctly unpleasant nature of the feeling itself. All these factors in combination create a negative image of the experiencer and, thus, a low acceptability index for both emotions. Here again though, there is a diference between the two languages: While ‘JEALOUSY’ occupies the last lines in both hierarchies, Russian ‘ENVY’ is greatly favored over the English one in irst-person diagnostic constructions. In Russian, ‘ENVY’ leads the foursome of the least acceptable emotions, being the most acceptable of them, with an index of 3.83 – only slightly lower than that of ‘SADNESS.’ In English, irst person ‘ENVY’ in diagnostic constructions is three times as infrequent as ‘SADNESS.’ These igures also relect the generally higher occurrence of ‘ENVY’ in the Russian emotional experience, where the number of total occurrences of ‘ENVY’ exceeds that of English by three times (see Section 3.1). Russian proneness to ‘ENVY’ correlates with Nancy Ries’s indings about the prominence of this emotion in the Russian emotional and cultural introspection (Ries 2005: 62–65). To conclude, there exist strong similarities between the English and Russian emotional etiquettes, manifested in the fact that the ordering of emotions according to the absolute frequencies of their self-reporting in these two languages is nearly identical. However, despite the same ordering of emotions, the “distances” between them are diferent, which suggests cross-linguistic variation within the generally similar schemas of emotional etiquettes and diferent “weights” of respective etiquette maxims. The next section explores cross-linguistic contrasts in the emotional etiquettes in more detail. 3.3 Emotional etiquette in English and Russian: Contrasts This section explores diferences between the English and Russian emotional etiquettes that exist within the frames of their generally similar tendencies. Table 5 below presents the percentages of irst person diagnostic constructions in the whole usage of respective emotion types in the corpus. These igures relect what portion of usages in each emotion type accrues to voluntary, controlled, purposeful expression of that emotion. They reveal diferent comparative weights and diferent interactions of the maxims, shared by the English and Russian emotional etiquettes. Comparison by percentages yields the following results. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 552 Valentina Apresjan Table 5: Percentages of irst-person diagnostic constructions in the total usages of emotion terms. Relative pragmatic acceptability indices. Emotion cluster English Emotion cluster Russian ‘Gratitude’ ‘Pity’ ‘Envy’ ‘Joy’ ‘Pride’ ‘Shame’ ‘Disgust’ ‘Anger’ ‘Fear’ ‘Sadness’ ‘Contempt’ ‘Jealousy’ 73.6% 9.6% 7.6% 7.2% 6.2% 5.1% 5.1% 3.5% 2.7% 2.7% 2% 0.67% ‘Gratitude’ ‘Shame’ ‘Envy’ ‘Pity’ ‘Pride’ ‘Fear’ ‘Disgust’ ‘Contempt’ ‘Joy’ ‘Anger’ ‘Jealousy’ ‘Sadness’ 66% 8.6% 7.7% 5.6% 4.8% 4.4% 4.3% 4.1% 3.8% 2.2% 2.2% 1% The maxim of pleasing appears to be the strongest of the maxims in the emotional etiquette in both languages. This claim is supported by the following facts. First, in both languages, the irst person reporting of ‘GRATITUDE’ considerably exceeds all other mentions of this emotion, which means that most of the times the experiencer of ‘GRATITUDE’ coincides with the speaker. This happens for the obvious reason that the most conventionalized and frequent expressions of ‘GRATITUDE’ – thank you, thanks, and the Russian correlates spasibo and blagodarju – are, in a sense, deictic words and can only refer to the feelings and intentions of the speaker herself. Thus, the prominence of ‘GRATITUDE’ on the pragmatic acceptability scale appears to be a very strong universal, which relects the absolute importance of the maxim of pleasing. Nevertheless, according to a certain diference in igures (see Table 5), the relative weight of this maxim is somewhat higher in English. Second, in both languages, irst person reporting of ‘ENVY’ constitutes the unexpectedly large share of all references to this emotion. In terms of etiquette, the relatively high igure of irst person occurrences for this fairly rare emotion may be explained by its usage as a semi-conventional way of complimenting the object for his/her achievements or possessions, which the experiencer lacks. By placing herself in the position of inferiority, the experiencer indirectly praises or even latters the object of emotion, consider: (8) “You’ve done well for yourself, Rob,” Wyatt said quietly, using his brother’s given name. “A ine woman, a steady job, a son. I envy you a little” (Miller 2008, COCA). Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 553 (9) Zaviduju vam, rebjata. Vy iskusstvom zanimaetes’, a ja . . . (S. Denisova, Tol’ko v Moskve // “Russkij reportjor,” N 45 (223), 17.11.2011). ‘(I) envy you guys. You work in art, whereas I . . .’ . . . (S. Denisova, Only in Moscow // “Russian reporter,” N 45 (223), 17.11.2011). Thus, referencing this ungenerous and therefore unseemly emotion emerges as a useful etiquette tool of indirect and thus subtler praise. To underline the harmless nature of this ‘ENVY,’ Russian speakers sometimes refer to it as belaja zavist’ ‘white envy’ or xorošaja zavist’ ‘good envy,’ Another etiquette function of ‘ENVY’ words, which contributes to their irstperson usages, is the indirect expression of ‘PITY.’ Sometimes, instead of verbalizing a potentially humiliating ‘PITY’-like emotion toward an object who inds herself in an unfavorable position, the experiencer prefers to refer to the lack of envy: (10) “You have your problems, General,” Mr. Shippen said. “I don’t envy you.” (Rinaldi 1994, COCA). (11) Ja tebe ne zaviduju . . . tebe xuže, čem mne (Griškovec 1994, RNC). ‘I don’t envy you . . . you are worse of than I.’ These two functions of ‘ENVY’ words point to the importance of the maxim of pleasing and non-humiliation maxim and their prevalence over the self-esteem maxim: In both cases, the speaker puts herself in the position, inferior to that of the object. However, the maxim of self-esteem appears to be considerably more pronounced in English than in Russian, in which the irst-person expression of emotions seems to be greatly inluenced by the oppositely directed maxim of modesty. This is supported by the ‘SHAME’ and ‘PRIDE’ igures; in English, the relative share of ‘PRIDE’ irst person expression exceeds that of ‘SHAME,’ whereas in Russian, the percentages of ‘SHAME’ irst person expression almost twice exceed those for ‘PRIDE.’ Moreover, the relative share of English ‘SHAME’ is considerably lower than that of the Russian ‘SHAME,’ whereas the share of English ‘PRIDE’ is considerably higher than that of the Russian ‘PRIDE.’ The existence of this discrepancy is also supported by the ‘ANGER’ and ‘FEAR’ igures (the former is higher in English, the latter is higher in Russian). English speakers prefer to irst-person report the active emotion of ‘ANGER,’ which implies that the experiencer views herself in the position of power, whereas Russian speakers prefer to irst-person report the passive emotion of ‘FEAR,’ which Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 554 Valentina Apresjan implies the position of weakness for the experiencer.16 Finally, the much higher relative igures for Russian ‘JEALOUSY’ (another emotion that implies the weak position of the experiencer) also support the claim of the greater importance of the self-esteem maxim for the English emotional etiquette. Interestingly, non-humiliation maxim also appears to be more important for the English emotional etiquette, as attested by the comparative ‘CONTEMPT’ igures: Russian irst person share of ‘CONTEMPT’ twice exceeds that of English. Supericially, this claim appears to be contradicted by the higher ‘PITY’ igures in English, since ‘PITY’ is an emotion, potentially humiliating for its object. However, a more diferentiated analysis reveals that the potentially humiliating subtype of this emotion, expressed by the noun pity and the verb to pity, accounts only for a small share of the irst person expression in this emotion cluster, whereas the majority of irst person usages are constituted by the neutral, nonhumiliating ‘PITY,’ expressed by the adjective sorry, and the conventional, equally non-humiliating ‘PITY’ expressed by the noun condolences. Thus the main contrast between the English and Russian emotional etiquettes consists in the diferent weights of the following maxims: the maxim of selfesteem, the maxim of non-humiliation, and the maxim of modesty. English emotional etiquette is more inluenced by the former two, whereas the latter has more impact on Russian emotional etiquette. This diference results in diferent communication strategies: Russian speakers, en masse, tend to be at the same time more self-deprecating and supercilious, whereas English speakers maintain a greater communicational parity, at least when it comes to emotion expression. 3.4 Possible directions for future research There are at least three promising trends for future research in the area of emotional etiquette. The irst, obviously, is extending the study of linguistic emotional etiquette to other languages. The research apparatus can be adjusted accordingly while retaining its central principles, namely, a corpus-based comparison of irst person emotion expression performed using entire emotion clusters. Detailed analysis of various conversational strategies pertaining to diferent emotion subtypes forms another interesting part of this research approach. 16 Consider Konstan on the Greek ‘anger,’ which could only be experienced by those suiciently powerful to take revenge on the ofender (Konstan 2007). Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 555 Diachronic changes of emotional etiquette constitute the temporal dimension of this treatise. This direction of research is closely intertwined with historical and cultural studies due to the fact that society as a whole reacts to social events and processes on the emotional scale and these reactions are to a certain extent mirrored by language. While this area of inquiry is presently under-explored, a few interesting facts may be mentioned. First, the turn of the century appears to be one of the important landmarks in the emotional etiquettes of both languages, although in diferent ways. In English, all emotions peak around the 1900s, except ‘ENVY,’ which drops around this time, whereas in Russian, the divide of the 19th and 20th centuries marks a sharp decline in ‘JOY’ and ‘PITY.’ One of the steadier emotions is ‘JEALOUSY’: In both languages, it mostly retains the same low numbers since the 1800s. Currently, in English, ‘PITY,’ ‘PRIDE,’ and ‘JOY’ are on the rise. ‘PRIDE’ and ‘SHAME’ used to be almost equal in irst-person expression ’til the 1960s, ater which ‘SHAME’ started to decline, and ‘PRIDE’ grew. In Russian, the emotions on the rise are ‘GRATITUDE’ and ‘PRIDE.’ ‘SHAME’ still exceeds ‘PRIDE’ considerably, yet it is declining, while ‘PRIDE’ has been rising since the beginning of the 2000s. ‘ENVY’ peaks sharply around the 1880s, but is presently on the decline. ‘FEAR’ had two sharp peaks – around the 1880s (possibly, as a societal reaction to the terrorist assassination of Alexander II) and the 1940s (perhaps, in response to World War II, known as the Great Patriotic War in Russia). Obviously, this area requires minute and thorough analysis in order to align language with history and culture while avoiding unfounded speculation and overgeneralization. Regardless, it is immensely promising as an interdisciplinary trend in linguistic and cultural studies. Another interesting direction of research concerns the linguistic implications of emotional etiquette. It appears that there are at least two observable trends in the “behavior” of acceptable and unacceptable emotions. The terms for acceptable emotions tend to expand their already solid linguistic presence even further, as euphemisms (e.g., the use of fear expressions, such as I am afraid to introduce bad news) or hyperboles (e.g., the use of stronger emotions, such as I am happy to refer to a much weaker emotion). On the contrary, the terms for unacceptable emotions display the opposite trend, namely, to shrink in frequency and be replaced by euphemistic expressions. For example, the linguistically “non grata” emotion of jealousy (which is nevertheless humanly understandable and frequently experienced in real life) is sometimes referred to in English by the harmless euphemistic term territorial. When used to describe human relationships, this originally biological or administrative term means simply ‘jealous,’ yet its usage allows one to speak of jealousy respectfully and to avoid all its unpleasant implications: Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 556 Valentina Apresjan (10) Guys tend to be territorial, particularly if their partners show any attention to other guys. (11) Why are guys so territorial about sex? (12) Are men or women more territorial where romantic relationships are concerned? The study of the linguistic behavior of the terms for acceptable and unacceptable emotions respectively is yet another interesting area of related research. Research in this direction is likely to provide insight not only into the understanding of linguistic etiquette, but also into the greater issues of meaning, polysemy, and metaphor. 4 Conclusion This section briely summarizes the main results of the study and places them in the context of previous research. It also considers the implications this research has for cross-cultural studies and for foreign language acquisition. 4.1 The main results of the study This study demonstrates that emotional etiquette in both English and Russian is based on the same maxims: the maxim of pleasing, the maxim of non-humiliation, the maxim of self-esteem, and the maxim of modesty. However, diferent maxims have difering “weights” in the two languages. Verbal expression of emotions by the experiencer is inluenced by two main factors: the emotion’s general expressive potential and by the maxims of emotional etiquettes. While emotions with higher expressive potential tend to be verbalized more frequently, their expression is limited by the efect of the etiquette maxims, sometimes oppositely directed. In both languages, the maxim of pleasing is the most inluential, though it has more weight in English. The main diferences between the languages can be found in the areas of the maxim of non-humiliation, the maxim of self-esteem, and the maxim of modesty. While English emotional etiquette places more importance on the former two, Russian emotional etiquette is more inluenced by the latter. As a result, Russians express more ‘SHAME,’ ‘FEAR,’ and ‘CONTEMPT,’ while English speakers express more ‘PRIDE,’ ‘ANGER,’ and ‘PITY.’ In terms of Brown and Levinson’s politeness theory (1987), there appears to be a strong contrast between Russian and English speakers with regard to the Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 557 speakers’ and the addressees’ positive politeness, which aims at positive face preservation. “Positive face” refers to one’s feeling good about oneself; speech acts that belittle either the speaker or the hearer are positive face-threatening. Russian emotion expression allows a certain degree of positive face-threatening: ‘SHAME’ expression threatens the speakers’ positive face, while ‘CONTEMPT’ expression threatens the positive face of the hearers; both emotions are quite freely expressed in Russian. English speakers, on the contrary, use emotion expression strategies that tend to preserve their own as well as the hearers’ positive face. In what concerns negative politeness, it seems that both emotional etiquettes allow a certain degree of negative face-threatening. “Negative face”-threatening involves the obstruction of one’s interlocutor’s freedom of action. Speech acts that convey the speakers’ sentiments toward the addressees’ acts or belongings threaten their negative face. Therefore, both Russian and English allow negative face-threatening, though in diferent ways – Russian allows for relatively high ‘ENVY’ expression, whereas English is more tolerant toward the expression of ‘ANGER.’ In terms of Locher and Watts’ theory of relational work (2008), Russian emotional etiquette could be construed as incorporating more extreme behavior strategies, as it is at the same time overly polite (‘SHAME’ expression) and rude (‘CONTEMPT’ expression), whereas English favors more moderate strategies of emotion expression. 4.2 Veriication of previous research In formulating the peculiarities of “Anglo” and Russian cultural scripts, Wierzbicka (2006) made claims regarding the importance of tactfulness and kindness toward the addressee in the former as opposed to the importance of directness and sincerity in the latter. This claim is supported by the indings of the present study, particularly by the greater prominence of the maxims of pleasing and nonhumiliation in the English emotional etiquette. The importance of these maxims results in relatively indirect strategies for communicating potentially ofending emotions, in contrast to a relative lack of reluctance to expressing hurtful emotions in Russian, e.g., prezrenie ‘contempt.’ Wierzbicka’s claim of the overall greater emotionality of Russian speakers (Wierzbicka 1992; Levontina 1997) has also been corroborated by the present indings (Section 3.1). Certain contentions about the prominence of particular emotions have also been substantiated by this study. Examples include the representation of ‘shame’ Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 558 Valentina Apresjan as one of the central concepts in the Russian emotional world17 (Bulygina and Shmelev 1997; Zalizniak 2000), relative rarity of sčast’je as compared to happiness – though not the entire ‘JOY’ cluster (Zaliznjak 2003), or the salience of Russian toska, ‘sadness without cause’ – though not the entire ‘SADNESS’ cluster (Wierzbicka 1990). Some other established notions of cross-cultural emotion pragmatics, such as the prominence of ‘PITY’ in the Russian linguistic worldview (Wierzbicka 1992; Levontina 2004), or a special Russian proneness to humility (Rancour-Laferriere 1995; Wierzbicka 1997; Shmelev 2005) have, on the contrary, been challenged by the data in the present research. The latter reveals the somewhat paradoxical nature of the Russian emotional etiquette. If Russian etiquette did indeed relect a tendency to humbleness, one would expect a high expression of ‘SHAME’ and a low expression of ‘PRIDE’ and ‘CONTEMPT.’ It seems, however, that this humbleness is only one-sided: While Russian speakers do indeed tend to display low self-evaluation, and, thus, have a high expression of ‘SHAME’ and a low expression of ‘PRIDE,’ they somehow consider others as even worse than themselves, which is supported by the relative prominence of ‘CONTEMPT’ in the Russian emotional etiquette. Another supposed tendency of the Russian emotional expression postulated in (Zaliznjak 2003), namely, its “bashfulness” and reluctance in expressing “exalted” emotions likewise seems to be challenged by systematic analysis of the entirety of Russian emotion clusters. While it is applicable to the Russian ‘JOY’ cluster, where neutral and rational ‘JOY’ (‘glad,’ ‘content’) greatly outweigh the stronger and more elevated subtypes (‘happiness,’ ‘delight’), ‘FEAR,’ ‘ANGER,’ and ‘SADNESS’ clusters demonstrate the opposite tendency. In the later three clusters, the expression of the “sublime” emotion subtypes, such as ‘mystical fear’ (užas), ‘righteous’ anger (negodovanije), and ‘profound lasting sadness’ ( pečal’) either exceeds or nearly equals the expression of the more mundane subtypes. 4.3 Implications for foreign language acquisition The study of emotional etiquette is instrumental in forming foreign language pragmatic competence and developing adequate conversational strategies. Understanding the emotional etiquette of a foreign language facilitates appropriate comprehension and language production. As is the case with grammar, for- 17 This tendency is, however, declining over time, which may relect gradual weakening of the modesty maxim in the Russian emotional etiquette. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 559 eign language learners are oten inluenced by the interference of L1 in the domain of pragmatics. This may result both in incorrect interpretation and production; e.g., a Russian learner of English may transfer the Russian conversational selfdeprecation into English, or an English learner of Russian may be unfavorably impressed by what he/she might perceive as conversationally aggressive position. A comprehensive corpus-based study of emotion clusters with various emotion subtypes provides useful guidance on the established patterns of emotion expression in each language. Acknowledgments: This study was carried out within the National Research University Higher School of Economics’ Academic Fund Program in 2013–2014, research grant No. 12-01-0102. References Apresjan, Valentina. 2010. Rečevye strategii vyraženija emocij v russkom jazyke [Speech strategies of emotion expression in the Russian language]. Russkij jazyk v naučnom osveščenii 2(20). 26–57. Apresjan, Valentina. 2011a. Opyt klasternogo analiza: russkie i anglijskie emocional’nye koncepty. Čast’ 1 [An attempt at cluster analysis: Russian and English emotion concepts. Part 1]. Voprosy jazykoznanija 1. 19–51. Apresjan, Valentina. 2011b. Opyt klasternogo analiza: russkie i anglijskie emocional’nye koncepty. Čast’ 2 [An attempt at cluster analysis: Russian and English emotion concepts. Part 2]. Voprosy jazykoznanija 2. 63–88. Broadie, Sarah & Christopher Rowe. 2002. Aristotle Nicomachean Ethics: Translation, introduction, and commentary. Oxford: Oxford University Press. Brown, Penelope & Stephen C. Levinson. 1987. Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Bulygina, Tatyana & Aleksey Shmelev. 1997. Jazykovaja konceptualizacija mira na materiale russkoj grammatiki [Linguistic conceptualization of the world as exempliied by the Russian grammar]. Moscow. Ekman, Paul. 1972. Universals and cultural diferences in facial expressions of emotions. In J. Cole (ed.), Nebraska symposium on motivation, 207–283. Lincoln: University of Nebraska Press. Iordanskaja, Lidija. 1971. Leksikograičeskoe opisanie russkix vyraženij, oboznačajuščix izičeskie simptomy čuvstv [Lexicographic description of Russian expressions denoting physical symptoms of feelings]. Mašinnyj perevod i prikladnaja lingvistika 16. 3–25. Iordanskaja, Lidija. 1984. Slovarnye statji bojat’sia, vostorg, voshiščat’, gnev, strah [Lexicographic entries to be afraid, delight, to delight, anger, fear]. In Tolkovokombinatornyj slovar’ sovremennogo russkogo jazyka, 158–162, 207–210, 213–214, 248–252, 806–813. Vienna. Konstan, David. 2007. The emotions of the ancient Greeks: Studies in Aristotle and classical literature. Toronto: University of Toronto Press. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 560 Valentina Apresjan Levontina, Irina & Anna Zalizniak. 2001. Human emotions viewed through Russian language. In Jean Harkins & Anna Wierzbicka (eds.), Emotions in cross-linguistic perspective. Berlin: Mouton de Gruyter. Levontina, Irina. 1997. “Dostoevskij nadryv” [Dostoevsky’s strain]. Wiener Slawistischer Almanach. Bd. 40. Wien. Levontina, Irina. 2004. Slovarnaja statja ŽALOST’ [Lexicographic entry PITY]. Novyj Objasnitel’nyj Slovar sinonimov russkogo jazyka Izdanie vtoroe, ispravlennoe i dopolnennoe. Pod obščim rukovodstvom akademika Ju.D.Apresjana [New Explanatory Dictionary of Russian Synonyms. Second edition, revised and enriched. Under the scientiic guidance of academician Ju.D.Apresjan], 327–331. Moscow. Locher, Miriam A. & Richard J. Watts. 2008. Relational work and impoliteness: Negotiating norms of linguistic behavior. In Derek Bousield & Miriam A. Locher (eds.), Impoliteness in language: Studies on its interplay with power in theory and practice, 77–101. Berlin: Mouton de Gruyter. Novyj Objasnitel’nyj Slovar sinonimov russkogo jazyka Izdanie vtoroe, ispravlennoe i dopolnennoe. Pod obščim rukovodstvom akademika Ju.D.Apresjana [New Explanatory Dictionary of Russian Synonyms. Second edition, revised and enriched. Under the scientiic guidance of academician Ju.D.Apresjan]. 2004. Moscow. 327–331. Rancour-Laferriere, Daniel. 1995. The slave soul of Russia: Moral masochism and the cult of sufering. New York: NYU Press. Shmelev, Aleksey. 2002. Russkij jazyk i vnejazykovaja dejstvitel’nost’ [Russian language and the extralinguistic reality]. Moscow: Nauka. Shmelev, Aleksey. 2005. Terpimost’ v russkoj jazykovoj kartine mira [Tolerance in the Russian linguistic worldview]. In Ključevyje idei russkoj jazykovoj kartiny mira [Key ideas of the Russian linguistic worldview]. Moscow. Solomon, Robert. 2004. In defense of sentimentality. Oxford: Oxford University Press. Wierzbicka, Anna. 1990. Duša (soul), toska (yearning), sud’ba (fate): Three key concepts in Russian language and Russian culture. In Zygmunt Saloni (ed.), Metody formalne w opisie jȩzyków słowiańskich. (Formal methods in the description of Slavic languages), 13–32. Bialystok: Bialystok University Press. Wierzbicka, Anna. 1992. Semantics, Culture and Cognition: Universal human concepts in culture-speciic conigurations. New York: Oxford University Press. Wierzbicka, Anna. 1997. Understanding cultures through their key words: English, Russian, Polish, German, Japanese. New York: Oxford University Press. Wierzbicka, Anna. 1999a. Emotions across languages and cultures: Diversity and universals. Cambridge: Cambridge University Press. Wierzbicka, Anna. 1999b. Emotional universals. Language Design 2. 23–69. Wierzbicka, Anna. 2002. Russian cultural scripts: The theory of cultural scripts and its applications. Ethos 30(4). 401–432. Wierzbicka, Anna. 2006. Anglo scripts against “putting pressure” on other people and their linguistic manifestations. In Clif Goddard (ed.), Ethnopragmatics: Understanding discourse in cultural context, 31–63. Berlin: Mouton de Gruyter. Zalizniak, Anna A. 2000. O semantike ščepetil’nosti (obidno, sovestno i neudobno na fone russkoj jazykovoj kartiny mira) [On the semantics of scrupulousness (it’s galling, one is ashamed and it’s awkward against the backdrop of the Russian linguistic worldview]. In Logičeskij analiz jazyka: Jazyki etiki, 101–118. Moscow. Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 561 Zaliznjak Anna A. 2003. Sčast’je i naslaždenije v russkoj jazykovoj kartine mira [Happiness and enjoyment in the Russian linguistic worldview]. Russkij jazyk v naučnom osveščenii 1(5). 68–84. Appendix Table 1: Emotion clusters in English and Russian Emotion cluster English Russian ‘FEAR’ = ‘feeling bad when expecting something bad to happen and being unable to change the situation’ ‘neutral fear’: to be afraid, fear, to fear ‘biological fear’: to scare, scared, fright, frightened, to frighten ‘neutral fear’: bojat’sja ‘to be afraid,’ strah ‘fear’ ‘biological fear’: pugat’sja ‘to get frightened, scared,’ ispug ‘fright, scare’ ‘rational fear’: opasenie ‘apprehension,’ opasat’sja ‘to be apprehensive’ ‘strong paralyzing mystical fear’: užas ‘terror,’ užasat’sja ‘to express terror’ ‘rational fear’: apprehension, apprehensive ‘strong paralyzing mystical fear’: terror, terriied, to terrify ‘fear mixed with disgust and shock’: horror, horriied, to horrify ‘reverent fear before a higher power’: awe, awed, to awe ‘fear before unavoidable undesirable events’: dread, to dread ‘strong uncontrolled fear resulting in hectic activity’: panic, to panic, panicked, to freak out ‘unethical fear’: to get cold feet, to chicken out ‘reverent fear before a higher power’: blagogovenie ‘awe,’ blagogovet’ ‘to awe’ ‘fear before unavoidable undesirable events’: strašit’sja ‘to dread’ ‘strong uncontrolled fear resulting in hectic activity’: panika ‘panic,’ panikovat’ ‘to panic’ ‘unethical fear’: strusit’ ‘to chicken out, to behave like a coward,’ sdrejit’ ‘to get cold feet’ Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 562 Valentina Apresjan Table 1 (cont.) Emotion cluster English Russian ‘ANGER’ = ‘feeling bad toward somebody when somebody did something bad and wanting to do something bad to that person’ ‘neutral anger’: angry, anger, to anger, mad, to madden ‘neutral anger’: serdityj ‘angry,’ serdit’sja ‘to be angry,’ serdit ’ ‘to anger’ ‘righteous motivated anger’: negodovat’ ‘to be indignant,’ negodovanie ‘indignation,’ vozmuščat’sja ‘to be indignant,’ vozmuščenie ‘indignation,’ gnevat’sja ‘to be wrathful,’ gnev ‘wrath’ ‘strong violent uncontrollable anger’: jarost’ ‘fury,’ bešenstvo ‘rage,’ besit’sja ‘to be enraged,’ besit’ ‘to enrage’ ‘anger as a reaction to an insigniicant unpleasant repeated stimulus’: razdraženie ‘irritation,’ razdražat’sja ‘to be irritated,’ razdražat’ ‘to irritate’ ‘unmotivated vicious anger’: zloba ‘venom, spite,’ zlobit’sja ‘to be venomous,’ zlost’ ‘anger, rancor,’ zlit’sja ‘to be angry, rancorous’ ‘righteous motivated anger’: indignation, indignant, wrath, wrathful ‘strong violent uncontrollable anger’: fury, furious, rage, to enrage, enraged ‘anger as a reaction to an insigniicant unpleasant repeated stimulus’: irritation, to irritate, irritated, annoyance, to annoy, annoyed ‘unmotivated vicious anger’: venom, venomous, rancor, rancorous, spite, spiteful ‘SADNESS’ = ‘feeling bad when something bad has happened and being unable to change the situation’ ‘neutral sadness’: sad, sadness, unhappy, unhappiness ‘short-lived sadness over unimportant stimulus’: upset ‘profound lasting sadness over a loss’: grief, to grieve, sorrow, sorrowful ‘neutral sadness’: grustnyj ‘sad,’ grust’ ‘sadness,’ grustit’ ‘to be sad’ ‘short-lived sadness over unimportant stimulus’: ogorčat’sja ‘to get upset,’ ogorčenie ‘being upset,’ rasstraivat’sja ‘to get upset,’ rasstrojstvo ‘being upset’ ‘profound lasting sadness over a loss’: pečal ‘sorrow,’ pečal’nyj ‘sorrowful,’ pečalit’ ‘to make sorrowful,’ pečalit’sja ‘to be sorrowful,’ skorb’ ‘sorrow, woe,’ skorbet’ ‘to be woeful,’ gore ‘grief,’ gorevat’ ‘to grieve’ Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 563 Table 1 (cont.) Emotion cluster ‘JOY’ = ‘feeling good when something good has happened’ English Russian ‘sadness caused by hopelessness’: despondence, despondent, wretchedness, wretched, dejection, dejected ‘sadness without a cause’: blues, blue ‘sadness caused by hopelessness’: unynie ‘despondence,’ unyvat’ ‘to be despondent’ ‘sadness without a cause’: toska ‘blues, yearning, longing,’ toskovat’ ‘to feel blue,’ tosklivyj ‘blue’ ‘neutral joy’: glad, joy, joyful, joyous ‘neutral joy’: rad ‘glad, joyful,’ radost’ ‘joy,’ radovat’sja ‘to be joyful, to express joy,’ radovat’ ‘to make joyful’ ‘strong emotional joy’: sčastliv ‘happy,’ sčastje ‘happiness’ ‘moderate rational joy’: udovletvorjon ‘content, satisied,’ udovletvorenie ‘content, satisfaction,’ dovolen ‘content, satisiied,’ dovol’stvo ‘content, satisfaction’ ‘excessive emotional joy’: vostorg ‘delight,’ vosxiščenie ‘delight,’ vosxiščat’ ‘to delight,’ vosxiščennyj ‘delighted’ ‘vicious joy, Schadenfreude’: zloradstvo ‘gloating,’ zloradstvovat’ ‘to gloat’ ‘strong emotional joy’: happy, happiness ‘moderate rational joy’: content, contentment, satisied, satisfaction, pleased, to please ‘excessive emotional joy’: delight, delighted, excited, excitement, to delight, to excite, glee, gleeful ‘vicious joy, Schadenfreude’: gloating, to gloat ‘DISGUST’ = ‘feeling bad when smelling, tasting or being in contact with something bad’ ‘neutral disgust’: disgust, disgusting, disgusted, to disgust ‘primarily physical disgust’: distaste ‘primarily moral disgust’: repugnance ‘strong disgust’: repulsion, repulsive, revulsion, to revolt, revolting ‘neutral disgust’: otvraščenie ‘disgust,’ otvratitel’no ‘disgusting,’ protivno ‘disgusting’ ‘primarily physical disgust’: gadlivost’ ‘creepy-crawly-ness’ ‘strong disgust’: omerzenie ‘powerful disgust’ Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 564 Valentina Apresjan Table 1 (cont.) Emotion cluster English Russian ‘disgust mixed with hatred’: abhorrence, to abhor, abhorrent, loathing, to loathe, loathesome ‘the trait of being easily disgusted’: squeamishness ‘the trait of being easily disgusted’: brezglivost’ ‘squeamishness’ ‘potentially unhealthy disgust’: aversion ‘SHAME’ = ‘feeling bad when having done something bad’ ‘neutral shame’: shame, to be ashamed ‘ethical shame’: guilt, guilty, repentance, to repent, scruples, remorse, remorseful ‘social shame’: embarrassment, to embarrass, embarrassed, mortiication, to mortify, mortiied, shy ‘conventional shame’: to regret, regret, regretful ‘social stigma’: ignominy, disgrace, disgraced ‘neutral shame’: styd ‘shame,’ stydit’sja ‘to be ashamed’ ‘ethical shame’: vina ‘guilt,’ vinovatyj ‘guilty,’ raskajanie ‘repentance,’ raskaivat’sja ‘to repent,’ sovestno ‘bothered by conscience’ ‘social shame’: smuščenie ‘embarrassment,’ smuščat’sja ‘to be embarrassed,’ smuščat’ ‘to embarrass,’ konfuz ‘discomiture,’ konfuzit’sja ‘discomited,’ nelovko ‘ill at ease, lit.: awkward,’ nelovkost’ ‘discomiture,’ neudobno ‘ill at ease, lit.: uncomfortable,’ neudobstvo ‘discomiture’ ‘conventional shame’: sožalenie ‘regret,’ sožalet’ ‘to regret’ ‘social stigma’: pozor ‘shame, disgrace’ ‘PRIDE’ = ‘feeling good when having done something good’ pride, proud gordost’ ‘pride,’ gordyj ‘proud,’ gorditsja ‘to be proud’ ‘PITY’ = ‘feeling bad when something bad happened to somebody’ ‘neutral pity’: pity, to pity, sorry ‘neutral pity’: žalost’ ‘pity,’ žalet’ ‘to pity,’ žal’, žalko ‘sorry’ ‘sharing smb.’s feelings’: sočuvstvie ‘sympathy,’ sočuvstovat’ ‘to sympathize,’ sostradanije ‘compassion,’ sostradat’ ‘to be compassionate’ ‘sharing smb.’s feelings’: sympathy, to sympathize, compassion, compassionate Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 565 Table 1 (cont.) Emotion cluster English Russian ‘conventional pity’: condolences ‘active pity’: učastie ‘active compassionate concern’ ‘conventional pity’: soboleznovanija ‘condolences’ ‘ENVY’ = ‘feeling bad when somebody has something that one wants but does not have and feeling bad toward that person’ envy, to envy, envious, jealous, jealousy zavist’ ‘envy,’ zavidovat’ ‘to envy’ ‘JEALOUSY’ = ‘feeling bad when somebody one loves another person and feeling bad toward those people’ jealous, jealousy revnost’ ‘jealousy,’ revnovat’ ‘to be jealous’ ‘GRATITUDE’ = ‘feeling good toward somebody when somebody has done something good toward them’ ‘neutral gratitude’: to be grateful, gratitude, to be thankful, thankfulness ‘rational gratitude for favor’: appreciative, appreciation, to appreciate ‘conventional expression of gratitude’: thank you, thanks ‘neutral gratitude’: blagodarnyj ‘grateful,’ blagodarnost’ ‘gratitude’ ‘rational gratitude for favor’: priznatel’nyj ‘appreciative,’ priznatel’nost’ ‘appreciation’ ‘conventional expression of gratitude’: spasibo ‘thank you, thanks,’ blagodarju ‘thank you, thanks’ ‘CONTEMPT’ = ‘feeling bad toward somebody when thinking somebody is bad and worse than oneself’ ‘neutral contempt’: contempt, contemptuous, to despise ‘neutral contempt’: prezrenie ‘contempt,’ prezirat’ ‘to despise’ ‘contempt mixed with disgust’: gnušat’sja ‘to be disdainful of’ ‘contempt mixed with disgust’: disdain, disdainful ‘mocking contempt’: scorn, scornful Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 566 Valentina Apresjan Table 2: Percent ratios of diferent emotion subtypes in emotion clusters in English and Russian ‘FEAR’ ‘Fear’ in English ‘Fear’ in Russian ‘ANGER’ ‘Anger’ in English ‘Anger’ in Russian ‘SADNESS’ ‘Sadness’ in English ‘Sadness’ in Russian Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 Corpus methods in pragmatics 567 Table 2 (cont.) ‘JOY’ ‘Joy’ in English ‘Joy’ in Russian ‘DISGUST’ ‘Disgust’ in English ‘Disgust’ in Russian ‘SHAME’ ‘Shame’ in English ‘Shame’ in Russian Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26 568 Valentina Apresjan Table 2 (cont.) ‘PITY’ ‘Pity’ in English ‘PITY’ in Russian ‘GRATITUDE’ ‘Gratitude’ in English ‘Gratitude’ in Russian ‘CONTEMPT’ ‘Contempt’ in English ‘Contempt’ in Russian Bereitgestellt von | De Gruyter / TCS Angemeldet | 212.87.45.97 Heruntergeladen am | 10.12.13 12:26