Phraseology and Culture in English

Hans-Georg Wolf and Frank Polzenhagen
tems,” this dimension cannot be grasped by looking at the meaning of 
these items in isolation. Again, one has to look at the patterns of usage 
and consider them in the context of the cultural framework in which they 
occur.
With corpus linguistics, it has become possible to describe and com-
pare lexical frequency and lexical patterns on a broad scale. Yet, although 
Aston and Burnard (1998: 15f ) made note of the usefulness of computer 
corpora for comparing, in their terms, “geographical varieties and lan-
guages,” to our knowledge, little has been done in this direction, espe-
cially with respect to culture.
19
Our investigation of the corpora was par-
ticularly driven by our interest in keywords, and the collocations these key-
words form.
20
Some collocations occur so frequently that they can be 
considered multi-word units with a lexicalized status. A look at colloca-
tions also tells us something about the textual context in which keywords 
appear.
The corpora we used for our comparative analysis are the CEC and a 
combination of the FLOB corpus (British English) and the FROWN corpus 
(American English). These corpora roughly have the same structure, which 
should avoid skewed frequency patterns due to the predominance of certain 
lexemes in certain text types. FLOBFROWN united represent the two ma-
jor native varieties of English and can thus count as a good representation 
of anglophone “Western” culture. The CEC stands pars pro toto for (West) 
African English. FLOBFROWN (2,064,764 tokens, as calculated by Word-
Smith) has more than twice as many tokens as the CEC (898,572 tokens), 
but as Sinclair (2001: xii) has pointed out, “comparison uncovers differ-
ences almost regardless of size.” 
3.4.1. Cultural keywords 
To the extent that computer corpora represent a variety used by speakers
of a given society / culture or at least some text types produced in it, via
a corpus-based analysis we can arrive at the keywords of a society.
Wierzbicka (1997: 16) is right that “the question is not how to ‘prove’ 
whether or not a particular word is one of the culture’s keywords,
but rather to say something significant and revealing about that culture
by undertaking an in-depth study of some of them.” Wierzbicka’s (1997: 
16–17) demand to study keywords “as focal points around which en- 
tire cultural domains are organized,” and to explore these focal points

---

Fixed expressions as manifestations of cultural conceptualizations
419
in order to “show the general organizing principles which lend struc- 
ture and coherence to a cultural domain as a whole, and which often
have an explanatory power extending across a number of domains” is 
hoped to be met by relating our findings to the conceptual analysis above. 
Only, we take the conceptual system as our starting point, not the key-
words.
For our elicitation of keywords, we lemmatized most of the items impli-
cated by the cultural model across the corpora (i.e., the variant and inflected 
forms of the same word were joined to one entry). Some items were not 
lemmatized in order to facilitate the differentiation of the various senses, as 
in the case of relatives and mediums. The different senses an item may have 
were controlled as far as possible.
21
Here spirit is one exception, because in 
the CEC, it is not always clear from looking at the context if the term re-
ferred to the supernatural, the religious or the mental / emotional domain, 
and if these domains can be neatly distinguished (cf. Aston and Burnard 
1998: 15–16). Where necessary, we also checked the spelling differences in 
British and American English. 
WordSmith (Scott and Oxford University Press 1998) was the computer 
program we used. It computes keywords “by comparing the frequency of 
each word in the smaller of the two wordlists with the frequency of the 
same word in the reference wordlist” (Scott and Oxford University Press 
1998). The statistical computations were also performed by WordSmith. 
Only words that occur at least 4 times were considered. The keywords 
deemed irrelevant to this study were disregarded. 
The following findings were partially selected from earlier studies, which 
had different scopes (Wolf 2003, Wolf fc.). The tables consist of 5 col-
umns; the first column contains the keywords, the second and third column 
show the frequency of occurrence of each item in the respective corpora, 
column four the “keyness”
22
and column five the p-value. 
Table 1 reflects the salience of the community / family concept in Cam-
eroon English. Not only are terms for family and community themselves 
significantly more frequent in the CEC, as listed in a), but also terms that 
highlight the importance of the continuation of the family, as listed in b). 
Marriage is deemed a prerequisite, and the roles, duties, and moral obliga-
tions associated with raising and enlarging a family are condensed in items 
like husband, wife, parenting, maternity, procreation, and inversely, child-
less.

---

420
Hans-Georg Wolf and Frank Polzenhagen
Table 1. Keywords pertaining to the role of the community / family in the CEC.