Phraseology and Culture in English

part of a wider study on interlanguage pragmatics. 
Different varieties of English have been explored in terms of the occur-
rence of modal items by Collins (1991), who studied the distribution of 
necessity and obligation modals in American, Australian and British Eng-
lish. All of these studies demonstrate a degree of variation in terms of the 
distribution and frequency of different modal items and the structures that 
surround the modals in question. 
Corpus-based studies such as these, as well as those investigations that 
are based on elicited and invented data, have contributed to a better under-
standing of the pragmatic and discoursive functions of modal items, and 
they have demonstrated in detail the multi-functionality of individual mo-
dal items (Schiffrin 1987). However, a look at any stretch of naturally oc-
curring discourse suggests that modal items often form part of relatively 
stable expressions (e.g. could you just..) and that they tend to cluster with 
other modal items. As such, they behave like a large proportion of language 
as a whole in that they are part of phraseological units which display a high 
degree of co-selection (Sinclair 1991). 

---

Definitely maybe: Modality clusters and politeness in spoken discourse 
259
3. Multi-word units and modality clusters 
As outlined above, previous research on modal items has mainly focused 
on a description of individual words, both in the area of politeness theory 
and in corpus-based descriptions. 
However, a number of mainly corpus-based studies have highlighted the 
occurrence of particular patterns when modality items are being used. Farr 
and O’Keeffe (1996), for example, identify the recurrence of ‘pronouns +
modals’ and ‘modals + pronouns’ in their discussion of the word would.
Kennedy (1996) uses the 100 million word British National Corpus as the 
basis for his discussion of modal verb phrase structures of different com-
plexity. 
However, the area of modality clusters remains relatively under-explored. 
Möllering (2001) studies the modal particle eben in a range of corpora of 
spoken German and finds that this particle takes on a particular temporal 
function when it collocates with the modal particle ma, the shortened form 
of mal often found in naturally occurring spoken discourse. A particular 
meaning of eben that carries a temporal function comes about when it col-
locates with ma. Thus, ma and eben combine in a stable construction to cre-
ate a new meaning. Similarly, Barron (2003) studies the meaning of a num-
ber of German modal participles in terms of two-word units. The evidence 
presented in these two studies suggests, therefore, that certain modal clus-
ters create a separate meaning. 
We can then distinguish broadly between three types of clustering. The 
first type is the kind of clustering that occurs as part of relatively stable 
phrases, as discussed by Kennedy (1996). Examples such as ‘Could you 
just + verb phrase’ would come under this category. The second type of 
clustering is a mere accumulation of different modal items, often inter-
spersed by other parts of speech. The third type describes a stable cluster of 
modal items that combine to form a particular meaning and that might be