METAPHOR AND SYMBOL, 18(A), 255-271 Copyright © 2003, Lawrence Erlbaum Associates, Inc.
Metaphorical Expressions and Culture: An Indirect Link
Alice Deignan
School of Education University of Leeds
Lakoff (1993) argued that basic level conceptual metaphors are grounded in human experience, and are therefore likely to be found widely across different languages and cultures. However, other mappings may not be shared. It is well documented that many metaphorical expressions vary across languages, and a number of researchers have argued cultural motivations for this. Possible reasons for cross-linguistic differences in metaphor are that different cultures hold different attitudes to metaphor vehicles, or that the source domain entities and events are more salient in one culture than another. However, the corpus data discussed here suggest that rather than being a synchronic reflection of culture, metaphorical expressions are to some extent a cultural reliquary, and an incomplete one.
It is notoriously difficult to develop an operational definition of the notion of "culture," and in the literature on metaphor the term has been interpreted in various ways. Therefore, I begin by briefly outlining a few interpretations and indicating how the research described here relates to them.
In his position article on the theory of conceptual metaphors, Lakoff argued that bodily experience is universal, and therefore probably results in basic level metaphors that are widely shared by humans in different times and places (1993). These might be regarded as not being culture dependent. However, if "culture" is understood in its broadest sense, it could be argued that every aspect of human experience is filtered through it. This view is taken by Gibbs (1999), who does not accept that apparently "universal" bodily experience is interpreted in the same ways across cultures. He argued:
Requests for reprints should be sent to Alice Deignan, School of Education, University of Leeds, Leeds LS2 9JT United Kingdom. E-mail: a.h.deignan@education.leeds.ac.uk
256 DEIGNAN
One cannot talk about, or study, cognition apart from our specific embodied interactions with the cultural world, (and this includes the physical world, which is not separable from the cultural world in the important sense that what we see as meaningful in the physical world is highly constrained by our cultural beliefs and values). (Gibbs, 1999,p.153)
From another perspective, "culture" can be perceived as including the dominant ideologies of a community. In his introduction to this special issue, Boers drew on Lakoff' s work (1987) to suggest that the figurative language of a community might be seen as "a reflection of that community's conventional patterns of thought or world views." The notion that metaphors encapsulate a community's views has been widely developed and explored in the literature on metaphor and ideology (for a discussion of some studies in this field, see Deignan, 2000).
Both Gibbs' (1999) all-embracing view of culture and the ideological perspective on it outlined previously are important issues, but they are too broad to be tackled within the approach taken here. Instead, this article follows the interpretation of Lakoff (1993) put forward by Boers (this issue) in the introduction. He pointed out the logical entailment of his division of experience into universal and specific, which is that "unlike general physical experience, specific experiential domains are more likely to be culture-dependent and thus to vary from place to place." Here, this narrower perspective on the relation between culture and metaphor is explored.
In this article, I shall first report some cross-linguistic studies, considering ways in which linguistic metaphors can be explained by cultural differences in "specific experiential domains." I then add an analysis of corpus data from the Bank of English,1 which aimed to investigate these mechanisms further. The analysis suggests that a simplistic view of metaphors as a reflection of culture raises several problems, which are discussed in the final section.
METAPHOR VARIATION ACROSS CULTURES Evidence of Cross-Linguistic Variation
A number of studies have sought out cross-linguistic differences in metaphor use, often focusing on domains thought likely to differ, because of known cultural differences. Through such studies it has been demonstrated that different languages do exhibit different patterns of figurative language use. The differences are of several kinds. In the most extreme case of variation, metaphors that are frequent in one language are rare or nonexistent in another, as is the case for some metaphors in
lrThe Bank of English is a corpus of approximately 450 million words of written and spoken English, owned by HarperCollins Publishers and held at the University of Birmingham, England. A sample of approximately 56 million words was searched for the studies reported here, through CobuildDirect.
METAPHOR AND CULTURE 257
English and Spanish (Deignan, Lima, & Lopez-Mora, 1998). In other cases, similar metaphors are used in two languages, but are many times as frequent in one of the languages, as is found for some metaphors shared by the English, French, and Dutch (Boers & Demecheleer, 1997). Frequency aside, some shared metaphors may draw on the same source domain but differ in details across languages. This is the case for English and Chinese: Both languages use heat as a metaphor for anger, but English metaphors suggest heated liquid, whereas Chinese metaphors suggest hot gas (Yu, 1995). Kovecses (2002) gave a number of examples of such variations, taken from a range of languages. Furthermore, metaphors may differ in entailments: The metaphor of parenthood is used to talk about business in both English and Farsi, but in English a "parent company" is one that controls another, whereas in Farsi it is one that supplies raw materials to a company that uses them in manufacturing (Henderson, 1986). Examination of the examples produced in these and similar studies suggests that two closely connected factors underlie the differences between languages, each of which will further be discussed:
1. different cultures may hold different folk beliefs about attributes of the source domain; and
2. the source domain may be less salient in different cultures.
Folk Beliefs About Attributes of the Source Domain
The suggested argument is that if the members of a particular community hold particular beliefs about, say, the connection between a part of the body and an emotion, we might expect to find metaphors in which the body part stands for the emotion in their language. To take another example, if members of a particular culture hold a particular attitude toward a particular animal, then that animal might be used to stand metaphorically for a particular quality in their language. There is evidence both for and against this hypothesis.
The hypothesis seems to work as a way of accounting for many English metaphors that are derived from the source domain of the human body and used to talk about emotions. Fernandez (1996, pp. 122-127) discussed the medieval theory of humors, in which the four elements, fire, water, air, and earth, were linked to physiological types and to personality traits. She argued that the belief that particular emotions originated in particular parts of the body led to associations found in many English idioms, such as the link between the head and brain with reason, and the heart with emotion. Corpus investigations into lexical items such as "heart" suggest that these links are present in current metaphorical use. The following Bank of English citations show its use to stand for feelings toward others:
"She loved his brilliance and his generous heart."
"Her sincerity and unhappiness were clear and his heart went out to her."
258 DEIGNAN
In the case of animal metaphors, however, the hypothesis does not seem to hold up as well. MacArthur Purdon investigated the attitudes held by native speakers of Spanish and English toward various animals (2001). She then examined idioms and metaphors containing animal lexis, and found that there was not a consistent relation between the qualities her informants associated with a particular animal and the figurative expressions in which the word for that animal appeared. For example, both English and Spanish speakers associate dogs with faithfulness, and in both languages the animal is associated with generally positive qualities. However, the figurative use of the lexeme dog does not reflect this. The following Bank of English citation is typical, in that dog is used to denote a despised entity:
"It said much for the guy that he knew the car was a dog, didn't want to drive it, but he did the job."
The most frequent idiomatic expression involving the noun "dog" in the Bank of English is "let sleeping dogs lie," in which dogs connote unpredictable reactions and aggression. "Dog" is also used as a verb with the figurative meaning of "pursue" or "affect, repeatedly," in citations such as:
"It seemed to fit with the feeling of dread that had been dogging her all day." "Her career has been dogged by drink and anorexia."
Here it seems that the notion of faithfulness is transferred from the literal meaning, in the sense that an entity that "dogs " someone follows them tenaciously for a long period of time. However, the positive connotation usually attached to the quality of faithfulness, and to speakers' attitudes toward dogs, according to MacArthur Purdon (2001), is reversed, and the metaphor is only used to talk about negatively viewed entities. Other entities that dog people in the corpus include:
bad luck, bureaucratic obstacles, controversy, disagreements, dismal figures on unemployment, injury
The adjective "dogged" breaks this pattern of negativity to some extent:
"[They] had gained an enviable reputation for dogged determination and skill."
"Dogged" is used to talk about the quality of not giving up in the face of resistance, and this seems to be presented relatively positively by this metaphor; the collocates enviable and skill in the previous citation are typical of the corpus data. However, having overtones of dullness rather than dynamism, being dogged could not be considered an entirely positive quality.
METAPHOR AND CULTURE 259
The example of "dog/dogs/dogging/dogged" suggests that the attributes or evaluational stance associated with the literal sense of a word are sometimes not present or are even reversed in its metaphorical sense. This means that the existence of different cultural values and attitudes will not always provide an explanation for cross-linguistic differences in metaphorical meanings.
Salience of the Source Domain
In his introduction, Boers (this issue, p. 233) suggested that particular source domains "may not be (equally) available for metaphorical mapping in all cultures." Geographical or cultural differences might mean that for speakers of one language a particular domain is salient and thus drawn on as a source domain for metaphors. For instance, Boers and Demecheleer (1997) found that the closely related languages English, French and Dutch used similar metaphors to talk about economics, but that the frequency of use varied. In particular, they searched for examples of national stereotypes and found that gardening metaphors were three times as frequent in their English corpus as in their French corpus, whereas food metaphors were nearly three times as frequent in the French corpus as the English one. They speculate that this is a reflection of the relative interest in these activities in French and English culture.
Deignan and Potter (1999) searched for similar stereotypes in body part metaphors in their English and Italian corpora. Although they did not find overwhelming evidence of these stereotypes, they did find a few suggestive tendencies. For instance, they noted a tendency for metonyms associated with the mouth to focus on the target domain of speech in the English corpus, and on the target domain of eating in the Italian corpus. Examples of eating expressions include the Italian, "essere di bocca buona" (literal English translation: to be of good mouth). Here, the mouth stands metonymically for eating, and one of its meanings is to describe someone who habitually eats well. In a metaphorical extension, the expression is also used to describe a person who is easily pleased. In English there are a few instances of similar figurative expressions, such as "be in good/bad taste," but far more numerous are citations such as the following:
"The ability to speak several languages is an asset, but the ability to keep your mouth shut in any language is priceless."
Here "keep one's mouth shut" stands metonymically for not speaking. Such expressions are also found in Italian, but were relatively less frequent than expressions where eating is an aspect of the metaphorical or metonymic meaning.
Boers (1999) found that health metaphors were more frequent in his corpus of "The Economist" in editions from the winter months. In northern Europe this is a time when people are more likely to suffer from illness, due to the hostile climate at
260 DEIGNAN
that time, and therefore are likely to be preoccupied with their health to a greater extent than during the summer months.
These examples suggest that the existence or relative salience of an entity in a culture, or during a particular period of time, will affect its use as the source domain of a metaphor. It would seem highly likely that this is a factor that can partly explain cross-cultural differences in metaphorical meaning and use, different source domain entities having different levels of salience for different communities. However, even more important than the salience of single entities may be language users' shared awareness of prototypical events and behavior in their culture. A number of examples are now discussed.
Deignan, Lima, and Lopez-Mora (1998) found several clusters of English metaphors not used in Spanish, and Spanish metaphors not used in English. Their English metaphors are from the domain of horse-racing, and their Spanish metaphors are from bull-fighting and religion. The factor discussed previously, the existence or relative prominence of the source domain, would seem to account for these, each of the three source domains they explored being stereotypic ally associated with British and Spanish culture, respectively. However, it seems worth noting that all their examples are phrasal, rather than single word metaphors, and that each of them seems to refer not to a single entity, but to a prototypical situation or sequence of events. Examples of English horse-racing metaphors are:
"He is too young to be in the running for the job of Prime Minister." "The Green Party was running neck and neck with the Communists." "Ourfirst-past-the-post voting system punishes small parties."
Each expression refers to a particular micro-event in a horse race, in the first case to horses being evaluated as potential winners or otherwise, and in the second and third to a point in a race where competition is intense, toward the finish. These metaphors are not just mapping the participants in a horse race onto the target domain, they are mapping sequences of events and knowledge associated with them.
This is seen even more vividly in Castilian Spanish bull-fighting metaphors, which map onto the target domain of argument.
"Ver los tows desde la barrea" = (literally) "to watch the bulls from the fence" (remain neutral in an argument, sit on the fence) "Echar un capote" = (literally) "to throw a cape" (come to someone's aid in an argument, throw someone a lifeline)
"Tirarse al ruedo" = (literally) "to jump into the bull-ring" (to join in an argument)
"Poner un par de banderillas" = (literally) "to put in a pair of banderillas" [small pointed sticks] (to be unpleasant or irritating to someone, to have a dig)
METAPHOR AND CULTURE 261
"Estar para el arrastre " = (literally) "to be ready to be dragged away " (to be extremely tired)
In some cases people with only an outsider's knowledge of the source domain might not understand what is alluded to. For instance, "poner un par de banderillas" refers to the practice at the start of a bull-fight of baiting the bull with colored, pointed sticks, which are jabbed into its skin, especially around the shoulders. This is used as a metaphor for being deliberately irritating or unpleasant. "Estar para el arrastre " refers to the disposal of a dead bull at the end of a fight, when it is ceremoniously dragged away.
Spanish also draws on religion for a number of idiomatic expressions, which include:
"Quedarse para vestir santos " = (literally) "to stay and dress saints " (to be left on the shelf; i.e., normally of a woman, to remain unmarried) "No ser santo de tu devocion" = (literally) "not to be the saint of your devotion" (not be the type of thing that you like)
Of interest, Lima and Lopez-Mora suggest that the closest translation of the latter metaphor, "no ser santo de tu devocion," in British English would be "not be your cup of tea," evoking a very British cultural stereotype (personal communication). Both of these religious metaphorical idioms draw on Catholic images not readily accessible to the non-Catholic majority of British English speakers, of praying through particular saints.
Moon (1998) discussed "idiom schemas," a notion that she links to frame semantics and schema theory (Minsky, 1975). The corpus data she examined show that the traditional view of idioms as frozen units is not borne out: Evidence of language in use shows them to be far more variable than generally believed. She argued that this suggests that rather than a surface phenomenon, they are actually realizations of underlying "cultural stereotypes, or stereotyped situations, where evaluations, connotations, and images are givens, constrained by contextual ideology" (Moon, 1998, p. 165). Although Moon uses this notion primarily to account for the variation she found in surface forms of what are clearly the same idioms, by suggesting that they are manifestations of the same idiom schema, the notion is also helpful in considering the metaphorical idioms discussed previously. As Moon defined it, an idiom schema refers to a small fragment of folk experience, often a prototypical event. This notion could cover the sight of the dragging away of a dead bull, or the finishing of a horse race. Idiom schemas are generally rich in connotations, and, as Moon pointed out, will often have a particular evaluation associated with them. This seems to be true of the metaphorical idioms discussed previously, which have a vividness that is difficult to convey in a literal paraphrase. (Indeed, many of the most apt English translations suggested for Spanish metaphors by Lima and Lopez-Mora are themselves metaphorical.)
262 DEIGNAN
It seems then that both the factors discussed in this section may have a part to play in the development of culturally specific metaphors. To explore these further, I conducted a series of corpus studies, which are reported in the following sections. They were initially conducted to gain an impression of the relative importance of the factors discussed so far, but they also led to several unexpected observations.
INITIAL CORPUS-BASED STUDY
Method
The aim of the initial corpus-based study was to answer the following questions:
• What nonliteral uses are found in the data?
• What linguistic forms do they take?
• To what extent might these be specific to British English?
• Do the factors outlined previously appear to account for the culturally-specific nonliteral uses found in the data?
The corpus researcher is faced with huge amounts of data, which normally have to be accessed by searching for single lexemes. Partington (1998) described the process as follows: "The researcher has an intuition about language, checks this against the data the corpus provides, and this checking process frequently suggests other avenues of research to be taken, often entirely unsuspected at the outset" (p. 1). The dilemma is often where to start; the choice of one particular lexeme as the initial focus often seems arbitrary and unscientific. For this study, the concordances for "horse/horses/horsing" were chosen (there were no citations of "horsed" in the data). There were two reasons for choosing this particular lexeme: First, research previously cited had suggested that horse-racing is an important and possibly culture- specific metaphor in English, and second, animal metaphors have been the topic of a good deal of interest. The expressions found in these data were then discussed with native speakers of a number of other languages. This study raised some interesting questions, and as follow-up studies I also examined the concordances for "car/cars," "cart/carts," "warhorse/warhorses," "workhorse/workhorses," and some words for items associated with horses, such as "bridle" and "reins."
Where the corpus contained more than 500 citations of a word form, 500 citations were randomly selected for study. The words were studied for literal or figurative meaning, and collocational and syntactic patterns were noted. Where the whole concordance was less than 500 citations, uses that occurred only once in the concordance were disregarded, because it was felt that these were probably idiosyncratic, and therefore not of interest to a study of metaphors shared by communities. Where the 500 citation sample was not the whole concordance, one-off uses in
METAPHOR AND CULTURE 263
the sample were checked against the whole concordance and were disregarded if they did not occur more than once in the entire concordance.
It has been extensively argued that corpus data can bring insights into language in use not available through intuition (e.g., Sinclair, 1991). Many of the studies previously discussed have drawn on corpus data. However, they have tended to concentrate on domains or lexical sets specified in advance-that is, on specific meanings or groups of meanings considered potentially interesting from a cultural perspective. This study differs from those in that it starts with language data rather than semantic patterns. Although the lexeme central to the study, HORSE, was chosen because it seemed likely that it would show up interesting, culturally-specific patterns of meaning, I approached the data in the way typically taken by corpus lexicographers. This meant that I examined every corpus citation in my sample, not limiting the analysis to exponents of the meanings I was initially interested in. I then attempted to develop a classification of all the nonliteral meanings in the sample, and generalizations that could apply to the whole sample, the only exception being, as previously noted, uses that occur only once.
Even though the analysis of a single semantic area cannot in itself form the grounds of generalizations about language and culture, the results of this study reflected some linguistic and semantic patterns found in other corpus studies, as well as suggesting directions for future research. In the rest of this section, I discuss the findings of the initial study, in relation to the questions listed previously.
What Nonliteral Uses Are Found in the Data?
In the analysis of figurative uses, metaphor was not separated from other tropes such as metonymy. Similes have been grouped separately, partly because their distinctive grammatical structure makes this easy to do, and partly because it can be argued that they are not full members of the group, because they involve a literal comparison. Table 1 gives the nonliteral uses found, with unedited corpus examples.
What Linguistic Forms do They Take?
In linguistic terms, the patterns of figurative use found in the concordance of the lexeme HORSE are as follows:
• The noun and the verb have different figurative uses: Examination of the verb forms showed that it appears to have no literal use, whereas the metaphorical meaning of the verb is not found in the noun citations.
• Different patterns of figurative use emerge for different inflections: separating citations into singular and plural inflections showed that there is only one nonliteral use, "dark horse/horses," found in both.
TABLE 1
Non-literal Meanings of "HORSE" From a 56 Million Word Sample of
The Bank of English
Meaning or Expression
Similes
Eats like a horse
Work like a horse
Strong as a horse
Other non-literal expressions: horse (singular) High horse
One-horse town
From the horse's mouth
Shut the stable door after the horse has bolted
Put the cart before the horse
Dark horse
Other non-literal expressions: horses (plural) Dark horses
Wild horses wouldn't drag ... Horses for courses
Drive a coach and horses through something
Other non-literal expressions: horsing Horsing around
Example
"And she admits she eats like a horse." A typical breakfast is a whole box of cereal with a pint of milk, four slices of toast, a huge bowl of fruit salad and a quart of orange juice.
"Nobody wants to be unemployed, so everybody works like a horse hoping that the high quality of his work will keep him his job."
It had been absurd, pretending to faint like that. "Everyone knew she was as strong as a horse."
"Solicitors, erm, they sit on their high horse and they 're disdainful of anybody that hasn't got what they would term as legal status."
"He arrives in the one-horse town of Red Rock."
"The participants in the discussion wanted answers from the horse's mouth."
"The review is welcome but it is shutting the stable door after the horse has bolted."
"Slavery and racial ideologies have indeed been related in many societies around the world, but to say that slavery was based on race is to put the cart before the horse."
"Steven King has emerged as the latest dark horse for next month's state surf lifesaving titles at Kurrawa."
"... World Cup dark horses Spain."
"Wild horses would not drag the secrets of the
confessional from a priest." "Because he has lost so many players he has
been forced to adopt a horses for courses
approach in Sweden." "The legislation setting up the lottery has driven
a coach and horses through gambling
regulations."
"Viera then started horsing around and splashed up his right foot, inches from my face."
Note: The Bank of English is a corpus of approximately 450 million words of written and spoken English, owned by HarperCollins Publishers and held at the University of Birmingham, England. A sample of approximately 56 million words was searched for these studies, through CobuildDirect.
264
METAPHOR AND CULTURE 265
• There is a large number of fixed and semifixed expressions: The verb form always appears in the phrasal verb "horse around." Where horse is used in similes, it occurs in the phrases "like a horse " or "as a horse." In other nonliteral citations of the noun, it also occurs in strings that are semifixed.
These are typical of tendencies found in other concordance studies (e.g., the figurative use of "SHOULDER," Deignan, 1999; and meanings of "QUALITY," Moon, 1987; and "YIELD," Sinclair, 1991). These points have been underresearched and underemphasised in the metaphor literature in comparison with semantic aspects of metaphor, but there are reasons why they may be of considerable importance. Language in use is the central source of nonintuitive evidence we have to research the nature of metaphor: As Kovecses (2002) remarked, "it is the metaphorical linguistic expressions that reveal the existence of conceptual metaphors" (p. 6). Here we see a high level of fixedness, and a tendency for figurative uses of lexemes to settle into distinctive syntactic and lexical patterns. It was suggested previously that fixed expressions may be the surface realizations of idiom schemas. Several writers have argued that the tendency toward fixedness and chunking might suggest that language users often make choices at the level of phrases rather than at word level (Erman & Warren, 2000; Sinclair, 1991).
To What Extent Might These Expressions Be Specific to British English?
A comprehensive research study would consist of consulting comparative corpora and a large number of native speakers to determine which metaphorical expressions are unique to which languages. This would be enormously informative, but at this stage, an initial trial study only was undertaken. This involved 14 teachers of English as a second language, all of whom speak a language other than English as their first language and are bilingual in English. The first languages spoken are: Korean (4 speakers), Chinese (3 speakers), Arabic (3 speakers), Japanese (2 speakers), Russian (1 speaker), and Castilian Spanish (1 speaker). Informants were given explanations of the nonliteral expressions previously discussed, and asked whether a translation equivalent exists in their first language. They also were asked whether there are other figurative expressions with a similar structure and meaning.
As expected, responses suggested that almost none of the English figurative expressions are shared by any of the 6 other languages spoken by informants. Closest are the similes; Russian shares "work like a horse," "strong as a horse," and "eat like a horse," though "wolf" is more frequent than "horse" in the last expression. Other languages draw on different animals; for instance, "eat like a dog" (Korean), or "like a tiger/cow" (Chinese). In Japanese, a hard worker works "like an ant." Most of the other figurative expressions listed in the previous table have no equivalents, the exception being "dark horse," which is used in Russian, Chinese,
266 DEIGNAN
and Japanese, the Chinese and Japanese informants suggesting that it originally entered their language as a borrowing from English. Russian also shares "put the cart before the horse." "Shut the stable door after the horse has bolted" can be translated as "fix the fence after losing the cow" in Korean, and "make the fence tight after the sheep has been stolen" in Chinese. Castilian Spanish has the expression "Don't look a gift horse in the mouth." This expression is known to native users of English, but as it did not occur in the English concordance sample studied it would seem to be relatively rare.
Do the Factors Previously Outlined Appear to Account for the Culturally-Specific Nonliteral uses Found in the Data?
Having established that most of the metaphorical expressions found in the data studied seem, from this initial survey, to be specific to English, the question is to what extent the factors discussed previously might account for their having developed in British English and not in the other languages cited.
The first factor was "folk beliefs about attributes of the source domain," which would seem a logical account of similes. There seems to be a widespread tendency to draw on the animal kingdom as a source of similes for human behavior, and it could be argued that the divergence in different communities' beliefs about each animal's attributes accounts for the difference in similes found. However, it is debatable whether British speakers would produce "hungry" and "hard-working" as among the most important qualities they associate with horses, and further informant testing would be needed to establish whether this is the case.
With regard to the second factor, salience of the source domain, the fact that there is a number of nonliteral expressions in which horses feature might be ascribable to their salience in British culture. With the exception of the similes found, the figurative uses do not take a form that could be reduced to the classical "X is Y" notion of metaphor. Rather, the noun uses tend to allude to situations in which horses feature. This means that in most cases the metaphorical interpretation seems to be derived from a knowledge of a source domain situation as a whole, rather than to attributes of horses in particular. For example, the interpretation of the metaphorical meaning of the expression "to put the cart before the horse " depends on knowledge that a horse is typically used to pull (rather than push) a cart. Knowledge of prototypical events seems to explain the bulk of the nonliteral expressions found in the sample. Most are highly suggestive of Moon's notion of idiom schemas.
It has been argued in this section that both the factors listed have an influence on the development of language-specific metaphors, but that the second may account for the greater number of types of culturally specific figurative language. However, approaching the data from this corpus perspective also raised doubts as to the soundness of the simplistic model of metaphor as a reflection of culture that I have put forward so far in this article.
METAPHOR AND CULTURE 267
PROBLEMS WITH A CULTURAL APPROACH TO METAPHOR
In this section, I deal with two issues raised by the initial concordance study and cross-linguistic comparison. First, consideration of the results of these studies suggests that it might be difficult to predict which metaphorical meanings become conventionalized in a particular language, even considering cultural attitudes and stereotypes. Furthermore, it also suggests that the aspects of the source domain that are drawn on to create metaphorical expressions may not be those aspects that are most salient to current speakers. This second point was the starting point for a second analysis of concordance data, described later in this section.
To illustrate the first issue, the discussion here is limited to Castilian Spanish. This is because of all the informants whose contributions were discussed in the previous section, the Spanish speaker's culture is probably the closest to British English culture. This means we might expect the least amount of divergence in metaphorical expressions, and the cross-linguistic comparison, as will be explained, is thus the most extreme example of unpredictability. The following expressions are some of those found in British English but not in Castilian Spanish.
"drive a coach and horses through something"
"shut the stable door after the horse has bolted
"put the cart before the horse "
"one horse town"
"wild horses wouldn't drag ..."
These English metaphors can be explained in cultural terms using the factors discussed previously: They reflect the salience of the source domain in British culture, in particular relating to the use of horses as transport and work animals. In the last case, interpretation also depends on knowledge of the attributes of wild horses: their frightening power and strength. However, unlike the metaphors from the source domains of horse-racing, bull-fighting, and religion discussed previously, these do not seem to evidence a true cultural difference. Like Britain, Spain used horses as a means of transport and for work up until the last century. The knowledge of attributes and prototypical events, and the salience of the source domain, are therefore likely to be closely equivalent in the two cultures. There is no clear reason why particular expressions should have developed and remained in one language and not the other. Until more is known about the development of metaphor and idioms, a degree of arbitrariness has to be accepted; cultural differences do not explain all linguistic differences.
It previously was suggested that many of the language-specific nonliteral uses of "HORSE" can be explained by reference to shared knowledge of prototypical events and situations. This was illustrated by the expression "put the cart before
268 DEIGNAN
the horse," which denotes misplacing of priorities. What can be construed of the etymology of the other metaphorical expressions found suggests the same type of mapping, as the following examples show: "One-horse town," used to denote a very small and uneventful settlement, presumably refers to the fact that in such a settlement very few people owned horses and there is therefore little movement. "Drive a coach and horses through something" draws on awareness of a moving coach and horses as a large and powerful force. Less transparent is "be on one's high horse," which derives from the use of war-horses or chargers by the wealthy, who were thus is a position to look down, literally and metaphorically, on others (Shorter Oxford English Dictionary). Although all the previous expressions are widely understood, and their origins seem transparent in most cases, it is probable that the events and situations that they refer to are not directly experienced by the majority of current British English speakers. They refer to the use of horses as working animals, as a means of transport, and in battle, uses that have largely died out in the industrialized world. This observation led to a number of further corpus studies, some typical findings of which are now discussed.
Stubbs (2001) suggested that a possible means of tracing speakers' experience of an entity is to analyze the uses of the lexeme in a large, contemporary corpus (p. 100). It is not argued here that corpus data can reveal the average British English speaker's internal representation of the entity, but they do show how it is read, spoken and written about in a sample of language that is claimed to be representative. To examine the hypothesis put forward previously, that current speakers do not frequently write and speak about literal horses in association with their former uses, the collocates of literal horse/horses were examined. It was found that the most frequent lexical collocates (as opposed to grammatical words such as prepositions and determiners) in the corpus, in order of frequency, are: show, racing, riding, trials. Citations in which horse/horses are used literally were then examined ("horsing" is not used literally), and it was found that there is little reference to horses as modes of transport, working or fighting animals. The central topics are horse-racing and the associated betting, horse-riding as a rural hobby, and, connected to this, keeping horses as pets. The following citations are typical:
"At Essex the previous day Renwick had ridden another of her horses, Baby Baby, to victory."
"I like going out and playing sport, horse riding, er that's the way I like to enjoy myself."
The current literal citations suggest that British speakers talk of horses in connection with their leisure, whereas most of the metaphorical uses of horse suggest their use in nonleisure transport, in work or war. It seems reasonable to assume that over the last century there has been a shift in the role of horses, from beasts of work to creatures who we share our leisure time with, and it therefore seems that our lit-
METAPHOR AND CULTURE 269
eral use of the word reflects contemporary perceptions, whereas the metaphorical uses are largely fossilized.
It was shown earlier however that the source domain of horse racing does generate some metaphors, such as those used to talk about politics. They did not appear in the initial analysis because although they draw on a horse-racing schema, they do not generally include the lexical item horse. Concordances for these expressions were examined, as were the concordances for other lexis from the horse schema, such as bridle, saddle, and rein. The evidence suggested that horse racing generates a fairly small set of metaphors in relation to the agricultural and transport uses of the field, although of course a wider corpus study would be needed to confirm this, and to explore other source domains.
The former roles of horses in transport and work are now performed mainly by the car and other motorized vehicles. The car plays a far larger role in the lives of current speakers than does the horse, and this is reflected in the fact that citations of car/cars are nearly four times as frequent in the corpus as citations of horse/horses. It might be therefore expected that there would be a great number of metaphors referring to cars, but this is not the case. In the sample 1000 citations of car and 500 citations of cars searched, there were no nonliteral uses. Words for components of cars, such as gears and brakes, do occur in metaphorical expressions, but these are relatively infrequent, and may be part of a wider machinery metaphor, which also produces expressions such as "be a cog in the wheel" and "throw a spanner (or wrench) in the works" and is not specific to cars.
It is not clear why certain source domains should generate large numbers of metaphorical expressions, whereas others that are apparently closer to our contemporary lives are almostunused as a source of metaphors. The most obvious possibility is that there is a lapse of time between entities becoming salient and their adoption as metaphor vehicles, and similarly they may remain as metaphor vehicles long after their literal referents have fallen out of use. The fairly widespread use of the computer as a metaphor, for example, for the brain, may be a counterargument, but it could be claimed that this is largely a technical use. Another possible explanation is that we are drawn to the picturesque in seeking metaphor vehicles: Although we are said to be in love with our cars, we do not find them suitably quaint to draw on in creating figurative language as we did horses.
CONCLUSION
This article has considered the role that culture may play in determining the content and form of metaphorical expressions. A number of examples of cross-cultural differences were cited, and several reasons for these differences were put forward. All these examples seem to have been uncovered by examining metaphors from source domains known in advance to be more salient to one culture than another. In my anal-
270 DEIGNAN
y sis, I examined metaphorical expressions that had not been selected in this way, and found that known cultural differences seem to play a much smaller part in explaining cross-linguistic differences. On this evidence, I would argue that even though there are a number of cross-domain mappings generated by unique cultural stereotypes, in English and presumably other languages, these do not account for most of the differences in metaphorical expressions found across languages.
I also argued that many of the metaphorical expressions that I found in my study derive from historical situations. It is well known that some English idioms, such as "kick the bucket," became fossilized many years ago and are largely opaque to current speakers. These studies suggest that many, more transparent, metaphorical expressions are also historical in that they allude to knowledge that is still shared as part of our cultural repository, but no longer directly experienced.
REFERENCES
Boers, F. (1999). When a bodily source domain becomes prominent: The joy of counting metaphors in the socio-economic domain. In R. W. Gibbs & G. J. Steen (Eds.), Metaphor in cognitive linguistics (pp. 47-56). Amsterdam: John Benjamins.
Boers, F. (2003). Applied linguistics perspectives on cross-cultural variation in conceptual metaphor. Metaphor and Symbol, 18, 231-238.
Boers, F., & Demecheleer, M. (1997). A few metaphorical models in (Western) economic discourse. In W-A. Liebert, G. Redeker, & L. Waugh (Eds.), Discourse and perspective in cognitive linguistics (pp. 115-129). Amsterdam: John Benjamins.
Deignan, A. (1999). Corpus-based approaches to metaphor. In L. Cameron & G. Low (Eds.), Researching and applying metaphor (pp. 177-199). Cambridge, England: CUR
Deignan, A. (2000). Persuasive uses of metaphor in discourse about business and the economy. In C. Heffer & H. Sauntson (Eds.), Words in context: A tribute to John Sinclair on his retirement. Birmingham, England: University of Birmingham.
Deignan, A., Lima, A., & Lopez-Mora, R. (1998, April). Metaphor, Culture and the Classroom. Paper presented at the 31st IATEFL conference, University of Manchester, England.
Deignan, A., & Potter, L. (1999, July). A corpus study of metaphors and metonyms in English and Italian. Paper presented at the Third Seminar on Researching and Applying Metaphor, University of Tilburg, The Netherlands.
Erman, B., & Warren, B. (2000). The idiom principle and the open choice principle. Text, 20, 29-62.
Fernandez, C. (1996). Idioms and idiomaticity. Oxford, England: Oxford University Press.
Gibbs, R. W. (1999). Taking metaphor out of our heads and putting it into the cultural world. In R. W. Gibbs & G. J. Steen (Eds.), Metaphor in cognitive linguistics (pp. 146-166). Amsterdam: John Benjamins.
Henderson, W. (1986). Metaphor in economics. In M. Coulthard (Ed.), Talking about text (pp.
110-127). Birmingham, England: University of Birmingham. Kovecses, Z. (2002). Metaphor: A practical introduction. New York: Oxford University Press. Lakoff, G. (1987). Women, fire and dangerous things: What categories reveal about the mind. Chicago:
Chicago University Press. Lakoff, G. (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor and Thought
(pp. 202-251). New York: Cambridge Unoversity Press.
METAPHOR AND CULTURE 271
Mac Arthur Purdon, F. (2001, April). Making semantically opaque metaphors transparent in FL instruction: Descriptive versus explanatory adequacy of the concept of attribute saliency. Paper presented at the Fourth Seminar on Researching and Applying Metaphor, University of Tunis, Tunisia.
Minsky, M. L. (1975). A framework for representing knowledge. In P. Winston (Ed.), The psychology of computer vision (pp. 211-227). New York: McGraw-Hill.
Moon, R. E. (1987). The analysis of meaning. In J. Sinclair (Ed.), Looking up: The COBUILD project in lexical computing (pp. 86-103). London: Collins ELT.
Moon, R. E. (1998). Fixed expressions and idioms in English: A corpus based approach. Oxford, England: Clarendon Press.
Partington, A. (1998). Patterns and meanings: Using corpora for English language research and teaching. Amsterdam: John Benjamins.
Sinclair, J. (1991). Corpus, concordance, collocation. Oxford, England: Oxford University Press.
Stubbs, M. (2001). Words and phrases: Corpus studies of lexical semantics. Oxford, England: Blackwell.
Yu, N. (1995). Metaphorical expressions of anger andhappiness in Chinese and English. Metaphor and Symbol, 10, 59-92.
Copyright © 2003 EBSCO Publishing