Orð og tunga - 08.07.2019, Blaðsíða 70

58 Orð og tunga
ency in conceptual metaphors between diff erent languages, criticisms
have been raised concerning the introspective method used to obtain
the data. For example, Deignan (2005:96) claims that “there is a dis-
crepancy between the expressions which a researcher may produce
from intuition when trying to think of typical lexicalizations and the
expressions which are most frequently used in the corpus.” Further-
more, Stefanowitsch (2006) asserts the necessity of using corpus data
by cautioning that the importance (or centrality) of metaphors cannot
be measured with the introspective method. Presumably, this prob-
lem also applies to cross-linguistic studies. Partly due to the diffi culty
in quantifying the results, many contrastive analyses have focused on
the existence/non-existence of particular metaphors and metaphori-
cal expressions (Question 1), but not on their centrality to the emotion
(Question 2).
In light of this, the present study analyzes reiði metaphors based
on corpus data, and discusses not only the existence/non-existence of
metaphors (Question 1), but also their centrality to the emotion (Ques-
tion 2). To extract relevant expressions, the study adopts metaphori-
cal patt ern analysis (Stefanowitsch 2006; hereaft er, MPA). According
to Stefanowitsch (2006:66), a metaphorical patt ern is “a multi-word
expression from a given source domain (SD) into which one or more
specifi c lexical item from a given target domain (TD) have been in-
serted.” In the examples below, the emotion words are underlined
and the metaphorical collocates are italicized.
(5) a. She is boiling with anger. (Kövecses 2000:21)
b. Simmer down. (Kövecses 1990:53)
Following Stefanowitsch’s defi nition, (5a) is a metaphorical patt ern
because the sentence includes both the metaphorically used word
boiling and the emotion word anger, whereas (5b) is not a metaphori-
cal patt ern because of the lack of an emotion word. At present, it is
diffi cult to extract metaphorical expressions such as (5b), so only ex-
pressions like (5a) are analyzed with the MPA method. Stefanowitsch
showed that this method can extract almost all of the metaphors list-
ed in previous studies, as well as others.
3.2 MI score
Arguing against the corpus-based approach, Kövecses (2011:35−36)
contends that a quantitative advantage does not necessarily lead to
tunga_21.indb 58 19.6.2019 16:55:57