Aneta Pavlenko (henceforth A.P.) is a pioneer in the study of “languages and emotion” in general and “emotions and bilingualism” in particular. If I voice in this commentary some disagreements, I do so in order to engage in a constructive discussion, conscious of how much she has done in this area, and also, of how much we share in our respective perspectives on emotion, languages, and bilingual lives.
The question of the place of “emotion and emotion-laden words in the bilingual lexicon” is an important and neglected area of research into bilingualism. As I see it, however, this area has suffered not only from a lack of empirical studies, but also – and perhaps even more so – from insufficient attention to conceptual issues.
To begin with, to study the place of “emotion and emotion-laden words” in the bilingual lexicon we need to know very well which area of the lexicon we have in mind and how exactly this area should be delimited. A.P. addresses these questions, but in my view, her answers are not entirely satisfactory.
A.P. notes, rightly, that in comparative work on emotion lexicons “[t]he key problem . . . is the continuing lack of clarity as to what counted as emotion words in particular studies” (p. 147). She also notes that “[t]hese issues still plague the emotion lexicon research and scholars continue to differ in ways they define emotion, emotion-related, and emotion-laden words” (p. 148).
In my view, however, a methodology which can solve these problems already exists, and has been tested in numerous studies of emotions across languages and cultures (see, e.g., my Emotions across languages and cultures: Diversity and universals, 1999). I am referring here to what is known as the “NSM” approach (from “Natural Semantic Metalanguage”). NSM is a system for meaning representation based on and interpretable through natural language – any natural language. It comprises a mini-lexicon of 65 or so indefinable primes (such as feel, think, body and happen), which evidence suggests are present as identifiable word-meanings in all languages, and a simple and universal mini-grammar (for an extensive bibliography, see the NSM homepage, www.une.edu.au/bcss/linguistics/nsm/).
According to A.P., “emotion and emotion-laden terms should be considered as a separate class of words in the mental lexicon” (p. 147), but she doesn't really make clear on what basis she identifies this class.
What A.P. calls “emotion words” are defined as words that “directly refer to particular affective states (‘happy’, ‘angry’) or processes (‘to worry’, ‘to rage’) and function to either describe (‘she is sad’) or express them (‘I feel sad’)” (p. 148). To my mind, however, this is an example of defining one unknown via another unknown. The word affective is a technical term of Anglophone psychology which has no clear meaning until it is explained in intelligible words of ordinary language. Without such an explanation, affective is not any clearer than emotion.
According to the NSM approach, complex, obscure and technical words can only be effectively explained when their definitions are anchored in simple and self-explanatory words (such as feel, think, body, and happen). Moreover, since such simple and self-explanatory words have their counterparts in all languages, by phrasing research into “emotions” in terms of such words we can at the same time free it from an Anglocentric bias, inevitable in research relying on technical English.
Relying on the methodology outlined above, cross-linguistic research into emotions has discovered a template which plays an important role in the lexicon of numerous languages – possibly, all languages – and, consequently, in the mental lexicon of their speakers. This template can be represented as follows:
As discussed in detail in my Emotions book (Wierzbicka, Reference Wierzbicka1999), the meanings of English adjectives like angry and sad are instances of this template; and so are the meanings of Russian verbs like serdit'sja (roughly, “to be angry”) and grustit' (roughly, “to be sad”), despite the lack of exact correspondence between these English and Russian words.
A.P. distinguishes between “affective states” (e.g. angry, sad, happy) and “affective processes” (e.g. to worry, to rage; in Russian, serdit'sja), and she rightly assumes that the meaning of an adjective like angry is different, in some respects, from that of a verb like serdit'sja (cf. Wierzbicka, Reference Wierzbicka and Landsberg1995). But to show what the differences in question really are one needs sufficiently fine-grained analytical tools. One also needs such tools to identify what, for example, the English adjective angry and the Russian verb serdit'sja share, and why they both qualify as “emotion words”.
The template presented here provides such an answer: all words which satisfy this template – and only such words – belong to the same semantic class, which can be labelled “emotion words”. It is an empirical finding that many, and perhaps all, languages do have words or phrasemes which embody this template in their meaning. We have, therefore, an objective, non-arbitrary basis for claiming that there is indeed in many, if not all, languages a distinct class of “emotion words”. The validity of this claim depends both on its broad empirical basis and on its conceptual clarity, explicitness and precision.
Although A.P. does not define her class of “emotion words” (at least not in simple and self-explanatory terms) we can assume that what she has in mind is in fact a category corresponding to the template presented here, and therefore a non-arbitrary conceptual category with cross-linguistic validity.
The same cannot be said, however, about the proposed category of “emotion-laden words”. This category, which is said to include words like jerk, loser, cancer, and air crash, is semantically highly heterogeneous. A.P. divides this category into subcategories such as (a) “swearwords and expletives”, (b) insults, (c) childhood reprimands, (d) endearments, (e) aversive words, and (f) interjections. But I think this is a very mixed grouping, without any shared semantic components.
“Swearwords” like shit, “insults” like idiot, “endearments” like darling, and interjections like Ouch! do share a semantic component, which can be represented in NSM as “when I say this, I feel something”, so they can be seen as a unitary category in a broad sense, but there is no such component in the meaning of words like spider, death or kill (another example of “emotionally-laden words” cited in A.P.'s article).
Many people may of course show an “aversive emotional reaction” to words like spider, death, and kill, but there is no “I feel” component in the meaning of these words. Kill differs in this respect from murder (which does have such a component), spider from vermin, and die from croak. For reasons of space, in what follows, I will not discuss the heterogeneous collection of “emotion-laden words” anymore, focussing instead on the relatively homogeneous one of “emotion words” or “emotion concepts”.
A.P. defines “emotion concepts” as “prototypical scripts that are formed as a result of repeated experiences and involve causal antecedents, appraisals, physiological reactions, consequences, and means of regulation and display” (p. 150). The notion of “prototypical scripts” which involve “appraisals” is consistent with the results of empirically-based cross-linguistic semantics based on the NSM framework. This is the idea on the basis which dozens of “emotion terms” like sad and angry, and their closest semantic counterparts in many other languages, have been explicated in my Emotions book and in the work of other NSM researchers (cf. Goddard and Wierzbicka, in press).
As I see it, however, the causal antecedents are parts of the prototypical cognitive scenarios. If the technical word “appraisals” stands for “ways of thinking”, then a cognitive scenario shows what the way of thinking associated with a given emotion concept is. Whatever the antecedents of, say, anger on any particular occasion, the idea that there is a causal link between, roughly speaking, “angry feelings” and “angry thoughts” is part and parcel of the concept of “anger”. More precisely:
She was angry (at that time) =
she felt something at that time
like people can feel when they feel something bad
because they think like this about something:
“this someone did something bad
I don't want things like this to happen
I want to do something because of this”
she felt like this because she thought like this
This is all there is to the English concept of “anger” (encoded in the English word anger). Of course when somebody is angry, there are some “causal antecedents, physiological reactions, consequences, and means of display”, but none of these can be specified as parts of the concept “anger”. But the prototypical cognitive scenario spelled out above identifies this English-specific concept sufficiently.
A.P. notes that “in any pair of languages, specific emotion concepts may overlap completely, partially, or not at all”, and she further distinguishes between what she calls “nesting”, “split”, and “differentiation” (p. 152). These distinctions are helpful and well justified. But “concept comparability” assumes a tertium comparationis – a common measure. NSM, with its universal grid of concepts and its universal mini-grammar, gives us such a common measure.
A.P. refers to the memoir of a Russian actress Koreneva, who while living with her American husband in the United States “came to rely on the Anglo concept of ‘frustration’ in her thinking and behavior” (p. 151).
I have discussed the uniqueness of the English concept “frustration” in my Emotions book:
[F]rustration is a highly culture-specific concept, very characteristic of modern Anglo culture, with its emphasis on goals, plans, and expected achievements. In other languages, the concept of “frustration” exists only as a relatively recent loan word from English (Frustration in German, frustracja in Polish, frustracija in Russian, frustrasi in Bahasa Indonesia, and so on); and as such, it has been spreading. (Wierzbicka Reference Wierzbicka1999, p. 72)
But to establish the differences between the English concept of “frustration” and its closest counterparts in Russian we need to go beyond comments explaining “frustration” in Russian as “čuvstvo neudovletvorenija, smešannoe s dosadoj, kotoroe voznikaet posle bol'šix ožidanij”, and in English as: “frustration – a feeling of dissatisfaction mixed with vexation/annoyance that appears after great expectations” (p. 151).
But “dosada” is as unique in Russian as “frustration” is in English. Furthermore, the attempted definition misses the crucial point that the expectations of the “frustrated” person concern, specifically, what he or she can achieve (and not, for example, what can happen to him or her).
A Greek–English bilingual quoted by A.P comments: “‘Frustration’ is such an amazing word, the lack of it in a language is so amazing . . .” (p. 154). Such metalinguistic comments by bilingual persons carry great weight as testimony of lived experience, and A.P.'s work in eliciting and collecting such comments is extremely valuable. In addition, however, we need rigorous comparisons of concepts in the bilingual lexicon based on a suitable methodology, and this is not addressed, or even mentioned, in A.P.'s article.
The important contribution of bilingual persons lies in their testimonies of their lived experience, not in their ad hoc definitions of the English concepts which strike them as unfamiliar and even “amazing”. If we want to seriously compare concepts in the bilingual lexicon, we need to pin down the semantic components correctly and precisely. We can do so using nsm, as the following explication of frustration illustrates:
Frustration (e.g. she (X) felt frustration)
she felt something at that time
like people can feel when they feel something bad
because they think like this about something:
“I want something to happen
I thought before that it would happen if I did something
I am doing it now
I have been doing it for some time
this thing is not happening because of this”
she felt like this at that time because she thought like this at that time
In closing, let me quote A.P.'s statement from a pre-publication version of her article: “future studies, guided by the more precise understanding of the structure of emotion concepts, could examine both the relationship between the concepts of particular languages (concept comparability) and the relationship between the concepts in the bilingual mind (concept encoding)”. There is indeed a great need for future studies of emotions in the bilingual lexicon and, I would add, in bilingual lives (cf. Besemeres, Reference Besemeres2002; Besemeres and Wierzbicka, eds. Reference Besemeres and Wierzbicka2007), and any such future studies will have to build on Aneta Pavlenko's pioneering work in this area.
At the same time, in searching for that more precise understanding of the structure of emotion concepts it would be good to take into account cross-linguistic semantic studies of emotion concepts which have already been done, and also, to address the issue of methodology, which, arguably, is the key to the “more precise understanding of the structure of emotion concepts” (Wieizbicka, in press) in the bilingual lexicon and, consequently, in bilingual lives. Thus, together with my admiration for A.P.'s work in the area of emotions and bilingualism, I would like to express some frustration (what a good word!) at the persistent reluctance of students of emotion to engage with the methodological issues which have been raised in NSM work for several decades – starting with my “Semantic primitives” (Wierzbicka, Reference Wierzbicka1972) and “The semantic structure of words for emotions” (Wierzbicka, Reference Wierzbicka, Jakobson, van Schooneveld and Worth1973). These issues will simply not go away.
You have
Access
Aneta Pavlenko (henceforth A.P.) is a pioneer in the study of “languages and emotion” in general and “emotions and bilingualism” in particular. If I voice in this commentary some disagreements, I do so in order to engage in a constructive discussion, conscious of how much she has done in this area, and also, of how much we share in our respective perspectives on emotion, languages, and bilingual lives.
The question of the place of “emotion and emotion-laden words in the bilingual lexicon” is an important and neglected area of research into bilingualism. As I see it, however, this area has suffered not only from a lack of empirical studies, but also – and perhaps even more so – from insufficient attention to conceptual issues.
To begin with, to study the place of “emotion and emotion-laden words” in the bilingual lexicon we need to know very well which area of the lexicon we have in mind and how exactly this area should be delimited. A.P. addresses these questions, but in my view, her answers are not entirely satisfactory.
A.P. notes, rightly, that in comparative work on emotion lexicons “[t]he key problem . . . is the continuing lack of clarity as to what counted as emotion words in particular studies” (p. 147). She also notes that “[t]hese issues still plague the emotion lexicon research and scholars continue to differ in ways they define emotion, emotion-related, and emotion-laden words” (p. 148).
In my view, however, a methodology which can solve these problems already exists, and has been tested in numerous studies of emotions across languages and cultures (see, e.g., my Emotions across languages and cultures: Diversity and universals, 1999). I am referring here to what is known as the “NSM” approach (from “Natural Semantic Metalanguage”). NSM is a system for meaning representation based on and interpretable through natural language – any natural language. It comprises a mini-lexicon of 65 or so indefinable primes (such as feel, think, body and happen), which evidence suggests are present as identifiable word-meanings in all languages, and a simple and universal mini-grammar (for an extensive bibliography, see the NSM homepage, www.une.edu.au/bcss/linguistics/nsm/).
According to A.P., “emotion and emotion-laden terms should be considered as a separate class of words in the mental lexicon” (p. 147), but she doesn't really make clear on what basis she identifies this class.
What A.P. calls “emotion words” are defined as words that “directly refer to particular affective states (‘happy’, ‘angry’) or processes (‘to worry’, ‘to rage’) and function to either describe (‘she is sad’) or express them (‘I feel sad’)” (p. 148). To my mind, however, this is an example of defining one unknown via another unknown. The word affective is a technical term of Anglophone psychology which has no clear meaning until it is explained in intelligible words of ordinary language. Without such an explanation, affective is not any clearer than emotion.
According to the NSM approach, complex, obscure and technical words can only be effectively explained when their definitions are anchored in simple and self-explanatory words (such as feel, think, body, and happen). Moreover, since such simple and self-explanatory words have their counterparts in all languages, by phrasing research into “emotions” in terms of such words we can at the same time free it from an Anglocentric bias, inevitable in research relying on technical English.
Relying on the methodology outlined above, cross-linguistic research into emotions has discovered a template which plays an important role in the lexicon of numerous languages – possibly, all languages – and, consequently, in the mental lexicon of their speakers. This template can be represented as follows:
someone feels something
like people can feel when they feel something because they think like this: [. . .]
As discussed in detail in my Emotions book (Wierzbicka, Reference Wierzbicka1999), the meanings of English adjectives like angry and sad are instances of this template; and so are the meanings of Russian verbs like serdit'sja (roughly, “to be angry”) and grustit' (roughly, “to be sad”), despite the lack of exact correspondence between these English and Russian words.
A.P. distinguishes between “affective states” (e.g. angry, sad, happy) and “affective processes” (e.g. to worry, to rage; in Russian, serdit'sja), and she rightly assumes that the meaning of an adjective like angry is different, in some respects, from that of a verb like serdit'sja (cf. Wierzbicka, Reference Wierzbicka and Landsberg1995). But to show what the differences in question really are one needs sufficiently fine-grained analytical tools. One also needs such tools to identify what, for example, the English adjective angry and the Russian verb serdit'sja share, and why they both qualify as “emotion words”.
The template presented here provides such an answer: all words which satisfy this template – and only such words – belong to the same semantic class, which can be labelled “emotion words”. It is an empirical finding that many, and perhaps all, languages do have words or phrasemes which embody this template in their meaning. We have, therefore, an objective, non-arbitrary basis for claiming that there is indeed in many, if not all, languages a distinct class of “emotion words”. The validity of this claim depends both on its broad empirical basis and on its conceptual clarity, explicitness and precision.
Although A.P. does not define her class of “emotion words” (at least not in simple and self-explanatory terms) we can assume that what she has in mind is in fact a category corresponding to the template presented here, and therefore a non-arbitrary conceptual category with cross-linguistic validity.
The same cannot be said, however, about the proposed category of “emotion-laden words”. This category, which is said to include words like jerk, loser, cancer, and air crash, is semantically highly heterogeneous. A.P. divides this category into subcategories such as (a) “swearwords and expletives”, (b) insults, (c) childhood reprimands, (d) endearments, (e) aversive words, and (f) interjections. But I think this is a very mixed grouping, without any shared semantic components.
“Swearwords” like shit, “insults” like idiot, “endearments” like darling, and interjections like Ouch! do share a semantic component, which can be represented in NSM as “when I say this, I feel something”, so they can be seen as a unitary category in a broad sense, but there is no such component in the meaning of words like spider, death or kill (another example of “emotionally-laden words” cited in A.P.'s article).
Many people may of course show an “aversive emotional reaction” to words like spider, death, and kill, but there is no “I feel” component in the meaning of these words. Kill differs in this respect from murder (which does have such a component), spider from vermin, and die from croak. For reasons of space, in what follows, I will not discuss the heterogeneous collection of “emotion-laden words” anymore, focussing instead on the relatively homogeneous one of “emotion words” or “emotion concepts”.
A.P. defines “emotion concepts” as “prototypical scripts that are formed as a result of repeated experiences and involve causal antecedents, appraisals, physiological reactions, consequences, and means of regulation and display” (p. 150). The notion of “prototypical scripts” which involve “appraisals” is consistent with the results of empirically-based cross-linguistic semantics based on the NSM framework. This is the idea on the basis which dozens of “emotion terms” like sad and angry, and their closest semantic counterparts in many other languages, have been explicated in my Emotions book and in the work of other NSM researchers (cf. Goddard and Wierzbicka, in press).
As I see it, however, the causal antecedents are parts of the prototypical cognitive scenarios. If the technical word “appraisals” stands for “ways of thinking”, then a cognitive scenario shows what the way of thinking associated with a given emotion concept is. Whatever the antecedents of, say, anger on any particular occasion, the idea that there is a causal link between, roughly speaking, “angry feelings” and “angry thoughts” is part and parcel of the concept of “anger”. More precisely:
She was angry (at that time) =
she felt something at that time
like people can feel when they feel something bad
because they think like this about something:
“this someone did something bad
I don't want things like this to happen
I want to do something because of this”
she felt like this because she thought like this
This is all there is to the English concept of “anger” (encoded in the English word anger). Of course when somebody is angry, there are some “causal antecedents, physiological reactions, consequences, and means of display”, but none of these can be specified as parts of the concept “anger”. But the prototypical cognitive scenario spelled out above identifies this English-specific concept sufficiently.
A.P. notes that “in any pair of languages, specific emotion concepts may overlap completely, partially, or not at all”, and she further distinguishes between what she calls “nesting”, “split”, and “differentiation” (p. 152). These distinctions are helpful and well justified. But “concept comparability” assumes a tertium comparationis – a common measure. NSM, with its universal grid of concepts and its universal mini-grammar, gives us such a common measure.
A.P. refers to the memoir of a Russian actress Koreneva, who while living with her American husband in the United States “came to rely on the Anglo concept of ‘frustration’ in her thinking and behavior” (p. 151).
I have discussed the uniqueness of the English concept “frustration” in my Emotions book:
[F]rustration is a highly culture-specific concept, very characteristic of modern Anglo culture, with its emphasis on goals, plans, and expected achievements. In other languages, the concept of “frustration” exists only as a relatively recent loan word from English (Frustration in German, frustracja in Polish, frustracija in Russian, frustrasi in Bahasa Indonesia, and so on); and as such, it has been spreading. (Wierzbicka Reference Wierzbicka1999, p. 72)
But to establish the differences between the English concept of “frustration” and its closest counterparts in Russian we need to go beyond comments explaining “frustration” in Russian as “čuvstvo neudovletvorenija, smešannoe s dosadoj, kotoroe voznikaet posle bol'šix ožidanij”, and in English as: “frustration – a feeling of dissatisfaction mixed with vexation/annoyance that appears after great expectations” (p. 151).
But “dosada” is as unique in Russian as “frustration” is in English. Furthermore, the attempted definition misses the crucial point that the expectations of the “frustrated” person concern, specifically, what he or she can achieve (and not, for example, what can happen to him or her).
A Greek–English bilingual quoted by A.P comments: “‘Frustration’ is such an amazing word, the lack of it in a language is so amazing . . .” (p. 154). Such metalinguistic comments by bilingual persons carry great weight as testimony of lived experience, and A.P.'s work in eliciting and collecting such comments is extremely valuable. In addition, however, we need rigorous comparisons of concepts in the bilingual lexicon based on a suitable methodology, and this is not addressed, or even mentioned, in A.P.'s article.
The important contribution of bilingual persons lies in their testimonies of their lived experience, not in their ad hoc definitions of the English concepts which strike them as unfamiliar and even “amazing”. If we want to seriously compare concepts in the bilingual lexicon, we need to pin down the semantic components correctly and precisely. We can do so using nsm, as the following explication of frustration illustrates:
Frustration (e.g. she (X) felt frustration)
she felt something at that time
like people can feel when they feel something bad
because they think like this about something:
“I want something to happen
I thought before that it would happen if I did something
I am doing it now
I have been doing it for some time
this thing is not happening because of this”
she felt like this at that time because she thought like this at that time
In closing, let me quote A.P.'s statement from a pre-publication version of her article: “future studies, guided by the more precise understanding of the structure of emotion concepts, could examine both the relationship between the concepts of particular languages (concept comparability) and the relationship between the concepts in the bilingual mind (concept encoding)”. There is indeed a great need for future studies of emotions in the bilingual lexicon and, I would add, in bilingual lives (cf. Besemeres, Reference Besemeres2002; Besemeres and Wierzbicka, eds. Reference Besemeres and Wierzbicka2007), and any such future studies will have to build on Aneta Pavlenko's pioneering work in this area.
At the same time, in searching for that more precise understanding of the structure of emotion concepts it would be good to take into account cross-linguistic semantic studies of emotion concepts which have already been done, and also, to address the issue of methodology, which, arguably, is the key to the “more precise understanding of the structure of emotion concepts” (Wieizbicka, in press) in the bilingual lexicon and, consequently, in bilingual lives. Thus, together with my admiration for A.P.'s work in the area of emotions and bilingualism, I would like to express some frustration (what a good word!) at the persistent reluctance of students of emotion to engage with the methodological issues which have been raised in NSM work for several decades – starting with my “Semantic primitives” (Wierzbicka, Reference Wierzbicka1972) and “The semantic structure of words for emotions” (Wierzbicka, Reference Wierzbicka, Jakobson, van Schooneveld and Worth1973). These issues will simply not go away.