Phraseology and Culture in English

2. Previous research on collocations in varieties of English 
The literature on World Englishes contains many informal observations to 
the effect that the role of phraseology, collocations and idioms in variety 
identification might have been underestimated in research to date. Deplor-
ing the absence of relevant systematic research, Crystal, for example, notes: 
“Collocations, however, are likely to prove one of the most distinctive do-
mains of varietal differentiation.” (2003: 162) He goes on to give a number 
of examples from the literature, for example take light for cut the power 
supply. His very choice of examples, however, points towards the gaps in 
current research on the topic. Thus, it is certainly true that the expression 
just mentioned is occasionally encountered in West African English. How-
ever, the interesting thing about it is not the mere fact of its existence but 
rather the details of its use in context – questions such as how often it oc-
curs (both in absolute terms and, more importantly, in relation to its inter-
nationally intelligible synonym), which genres or types of text it concen-
trates in, which type of writer uses it for which audience, and how its use is 
evaluated in the community. In all likelihood, such an analysis would show 
that – like most alleged lexico-grammatical features of West African Eng-
lish – this use is not really a distinctive feature of the variety (seen as a 
stable decontextualised system), but an option available to West African 
users of English for specific communicative purposes in specific discourse 
contexts.
In a study on collocations in science writing, Christopher Gledhill has 
defined phraseology as “the preferred way of saying things in a particular 
discourse” (2000: 1) and at the same time proved the usefulness of
a genre-specific approach to collocations in English. As I hope to show
in the present contribution, a similar approach can work for the study
of regional variation in written English. It will lead to an empirical- 
ly more comprehensive picture of regional variability in standard Eng-
lish and redress an unfortunate bias in previous scholarship which has 
tended to over-emphasise grammatical regionalisms, which are few in 
number and relatively unimportant, and lexical regionalisms, many of 
which tend to be diachronically unstable, at the expense of phraseological 
ones.
Given the importance of the phenomenon, it is surprising to see that there 
are relatively few previous studies, often conducted in isolation from each 
other. It seems that two types of phraseological phenomena have received 
privileged treatment in the literature on variation: 

---

Varieties of English around the world
441
(a) 
collocations with a prominent grammatical function (i.e. “colligations” 
in the terminology proposed by Firth), and 
(b) 
idioms, that is complex units whose meanings are not compositional 
and which exhibit a high degree of structural fixedness. 
Frequently mentioned examples of the first phenomenon are variety-specific 
preferences in the choice between different to, different than and different
from, or between have a look, walk etc. and take a look, walk.
3
With regard 
to idioms, it has been common knowledge that the English-speaking world 
divides into an American-dominated “baseball zone”, in which idioms such 
as (not) get to first/second base are current, and a “cricket” zone largely
co-extensive with the former British Empire, in which knock someone for 
six is used and understood (cf. Crystal 1999). A comprehensive survey of 
the use of (international and local) idioms in East African English is pro-
vided in Skandera (2003). 
For the area in between, that is the prototypical collocations with lexical 
rather than grammatical function, compositional meaning, and a medium 
degree of fixedness, there is less previous research to rely on. An important 
contribution to the study of the phenomenon is made in Biber et al.’s Long-
man Grammar, which, among other things, amasses a wealth of informa-
tion on contrasts between British and American adverbial usage, for example 
with regard to adverb + adjective collocations of the type too bad or real
good (more common in American English) or very nice (more frequent in 
British English – see 1999: 545 for details). 
Some recent work in cognitive linguistics is relevant to the present in-
vestigation because it explicitly aims beyond the boundaries of traditional 
descriptive linguistics and focusses on the language-culture interface. In sev-
eral publications spanning a period of two decades, Anna Wierzbicka has 
pinpointed important context-bound discursive aspects of variation in varie-
ties of English – so far mainly confined to Australian English (Wierzbicka 
1986 and 2001) and Singaporean English (Wierzbicka 2003), but in princi-
ple transferable to other regions, as well. As the following list of concerns 
indicates, Wierzbicka does not explicitly study collocations, nor use ma-
chine readable corpora: 
The areas of language which are particularly revealing in this respect [i.e. in 
investigations of the language-culture interface] are, as usual, terms of ad-
dress, conversational routines, discourse markers, names for categories of 
people perceived as distinctive and words and expressions designating values 

---

442
Christian Mair 
and human attitudes perceived as either ‘good’ or ‘bad’ – especially words 
which can be regarded as cultural ‘key words.’ (2003: 329) 
Cognitive research and the corpus-based study of collocational variability can 
thus be seen as complementary approaches to the same phenomenon, cul-
turally motivated lexical variability in world Englishes.
4
An interesting com-
plement to Wierzbicka’s study therefore is Ooi 2000, who in a comparative 
analysis of Singaporean and British digitised databases has shown ten col-
locations, for instance filial piety, to be characteristically frequent in the for-
mer variety. 
The major problem in the systematic corpus-based study of phraseologi-
cal phenomena is succinctly summarised by Stubbs: 
Word frequency lists are a standard resource for linguists. However, although 
lists of well-known phrases are available in many taxonomies and dictionar-
ies of collocations, only very limited frequency data are available. There are 
two obvious reasons for this lack of data on the frequency of phrases. First, 
although the phraseological nature of language has been thoroughly docu-
mented by corpus studies, there is still a tendency, following hundreds of 
years of lexicographic tradition, to think of individual words, rather than 
phrases, as the basic units of language. Second, since there are severe prob-
lems in defining phrasal units in corpora, it is difficult to know what to count. 
Indeed, it is doubtful if there could be a definitive ‘phrase frequency list’, 
since the units in question are so variable, and can be defined at such differ-
ent levels of abstraction. (2002: 215–216) 
On the practical level, such difficulties of definition are reflected in the great-
er computational volume and software requirements needed for the system-
atic analysis of collocations in a corpus, and a more indirect link between 
the statistical patterns uncovered and significant linguistic facts (see, e.g., 
Scott 2001 for a demonstration of this point).