18 Thomas Egan
We can see in Table 2 that the two prepositions are the most commonly chosen
options in the two languages for encoding all senses of ‘betweenness’, including
the idiomatic tokens. Indeed, apart from the Motion sense and the idiomatic uses,
both prepositions are chosen to encode one and the same predication more often
than not. Figure 3 illustrates the overlap between between and entre that we see
in Table 2.Entre 18%BetweennessBetween
23%
56%Figure 3. The overlap between between and entre in coding ‘betweenness’Figure 3 cannot be taken to illustrate the overlap in the encodings of ‘betweenness’
between the two languages tout court, restricted as it is to types of ‘betweenness’
that are encoded in Norwegian by the preposition mellom (and in one relatively
small corpus at that). On the other hand, we saw in Table 1 that all the main senses
of ‘betweenness’ that are encoded in English by between may also be encoded in
Norwegian by mellom, so it is reasonable to assume that a considerable portion of
the semantic field in question is represented by the tokens which lie behind the
percentages in the figure.
One final point should be made about Figure 3, and that it is the fact that it
could not have been drawn using a 2-text corpus. If we had studied all translations
of between into French and all translations of entre into English in such a corpus,
we would still be no wiser as to the extent of the overlap. Say that x% of the tokens
of between were translated by entre, and y% of the tokens of entre by between, how
would we go about calculating the overlap? Should we add the two percentages
and divide the result by two, or should we take one as a percentage of the other?
Using the 3-text approach, on the other hand, allows us to stipulate the exact
degree of overlap, albeit based on just one set of translations.