Yp1-03(2) Idioms: Motivation and Etymology Dmitrij Dobrovol’skij and Elisabeth Piirainen Abstract

4.2. Case study: German gender-specific idioms

This section shows that there exist traces of the image component in the figurative lexicalized meaning and pragmatic features of idioms. It follows that the image component may have an effect on the use of an idiom in such a way as to restrict its usage. For instance an idiom might contain elements of a mental image which impede certain kinds of usage. There are convincing examples from gender-specific idioms, which show such restrictions. Some idioms of this type show traces of an “etymological memory”, i.e. the idiom in question “memorizes” its history and does not allow it to be used in combinatorial surroundings which are incompatible with that history. Idiom (15) serves to demonstrate how this “etymological memory” of an idiom can determine its behavior in discourse. The idiom passes on an old symbolic concept which came into being in the Middle Ages and still has an effect on this idiom in its present-day usage.

(15) German die Hosen anhaben “to wear the trousers”

‘to be the dominant partner in a marriage; it is the wife rather than the husband or partner who makes decisions in the family’

The origins of this idiom lie in the “Battle for the Trousers”, which was treated frequently in disputes, literature and paintings of the Early Modern Times. The figure of the virago, the domineering woman whose unfeminine aggressiveness was perceived as a direct threat to male authority, was a widespread stereotype at that time and treated explicitly in iconography. The “Battle for the Trousers”, women’s attempt of trying to steal their husband’s breeches which symbolized the man’s supremacy and power, is the topic of numerous proverb anthology paintings and illustrations since early fifteenth-century art, especially in Flanders, the Netherlands, Germany and adjoining regions.^vii The idiom itself was already well-known in pamphlets and moralizing texts of various languages.

At first, this idiom was only applied to a female person, or more precisely to a woman that was married or lived in a partnership. In a case study which will be briefly summarized here, it was discovered that a restriction to women is still part of the semantic structure of idiom (15) (Dobrovol’skij and Piirainen 2009: 125–130). Corpus analyses based on DeReKo (above all, its written language parts) show that the hit rate for the idiom amounts to 157 texts, 46 of which can be used for our purposes (see below for more detail). 44 texts refer to a female person; only two texts refer to a male person. This gender-specific character is due to the image component of breeches, which still has the symbolic function to represent men’s power, although different ideas are connected with trousers (breeches, pants) today and women wear trousers just like men. The image component has remained constant: It is the symbolic character of the garment, despite the fact that the source frame fixing the underlying mental image has undergone a shift towards a “modernized” concept.

Below we would like to briefly outline the approaches of how the data of analyzing gender-specifics in German idioms (Dobrovol’skij and Piirainen 2009: 116–141) have been collected. Investigations into questions concerning idiom semantics by means of corpus analyses are still rare. As Deignan (2005) emphasizes, the researcher is faced with an enormously rich resource which is too large to be processed manually. However, semantically adequate results can be obtained only on the basis of a thorough analysis of rich empirical data.
The corpus researcher who attempts to develop valid generalizations about language meaning and use has to sift through a large amount of linguistic data, looking for regularities and patterns. It follows that corpus analysis of semantic issues such as metaphor must be bottom-up rather than top-down. (Deignan 2005: 92)
In view of this fact, we have developed a method to capture the gender-specific idioms from text corpora, which consists of five steps. The first step was to compile a list of potential gender-specific idioms, based on preliminary knowledge and later completed by systematically checking many idiom dictionaries. This list consists of 107 German idioms (cf. Piirainen 2001). In a second step, these idioms have been pre-tested on the base of a smaller part of the corpora of written language from the Mannheim German corpus DeReKo (with half a billion words), which is available to any user. These preliminary tests have shown that 39 of the idioms in question did not appear in the corpora at all, and another set of 22 idioms was represented only by a too small number of texts. Idioms with fewer than three hits in the corpora were removed from the list.^viii

The third step consisted of checking the remaining 46 “gender-specific idiom candidates” against the full data of DeReKo (with approx. 3,5 billion words). The average number of hits was 349, ranging from 3410 to five.

Our task (step four) was therefore, to check 12,558 texts manually for the occurrence of the idioms in question. We refrained from analyzing the idiom mit einer weißen Weste/eine weiße Weste “(with) a white vest”, documented in 3410 contexts, because this number of texts is too large to be processed manually at the present time in the sense of Deignan (2005).^ix The remaining 9,148 contexts then have been worked through in the following way:

We marked those contexts (i) where the idiom refers to a male person, (ii) where the idiom refers to a female person and (iii) those contexts which cannot be used for our purposes because the reference of a given idiom in such context is gender-nonspecific; cf. concepts such as government, company, party etc. or expressions like people, everyone, they, the students, the Germans. Texts where the idiom was used only for the purpose of play on words have also been excluded. It turned out that the third group makes up the vast majority of all texts. Again, 36 idioms had to be removed from our list because the contexts did not provide enough information on potential gender-specific features, i.e. most relevant contexts belonged to group (iii).

Step five consisted of a careful analysis of the texts in which the remaining ten idioms occurred. It produced unambiguous results, as has already been shown with idiom (15). The other nine idioms show a similar pattern; all are gender restricted. The idioms (16–24) have image components which restrict their use to a female person so that traces of an “etymological memory” become evident. The image component can include a male animal concept (cock in (16), evoked by the constituent meaning ‘comb’), a concept of a garment, never worn by women in the past (necktie, vest (17–18)), a part of such garment (trousers pocket, vest pocket (19–21)) or equipment possessed exclusively by men (billfold, (22)), and, finally, an anatomic peculiarity of men (beard (23)). Idiom (24) is a special case. Here responsible for the usage restriction is not the image component but a kind of folk etymology, based on the polysemy of the German word Schwanz: 1. ‘tail (e.g. of a dog)’ and 2. ‘male member’.

On the basis of this data set we can now ask how far the image component of an idiom’s meaning contributes to the restrictions on its use. Coding of the data below uses the following schema: H number of hits in the corpora; U usable text passages, M number of texts with a male person and F number of texts with a female person.

(16) jmdm. schwillt der Kamm “sb.’s comb swells”

1. ‘sb. gets very angry’; 2. ‘sb. gets very proud, conceited’