230 Zhiying Xin
plays in metaphor studies on the one hand, and empirical discourse analysis of metaphor and
other figurative uses of language, on the other. In his view, corpora are used either to identify
genre-specific metaphors and their discursive functions or to readdress issues concerning the
nature of metaphor. From a discourse analyst‘s point of view, in uncovering the behavior of
the central metaphor (s), corpus investigations help to describe and define discursive features
of given discourses. This contributes, in particular, to revealing the stance taken by discourse
participants. The main body of Partington‘s paper is divided into two parts. In the first part,
he proposes a six-step corpus designing procedure to identify genre-specific metaphors. He
illustrates his proposal with detailed analyses of several case studies. The genres of press
editorials, leading articles of dailies, election speeches and political briefings are covered. The
analyses are of the double function to familiarize the reader with the proposed corpus
construction model, and to capture the discursive behavior and functions of dominant
metaphorical expressions. Throughout the paper the author argues for the evaluative function
of metaphors (271). They are used to assess events, people and people‘s actions. In the second
part, Partington revisits metaphor and simile. Here three large corpora are employed to
display their differences and similarities as well as their discursive functions of evaluation. He
takes like and its relevant patterns as the case study, with special focus on the two likes in
Encyclopedias are like dictionaries and Encyclopedias are like gold-mines. His findings
indicate that the second like represents greater distance between the two terms it links and is
of stronger evaluative orientations. As for the differences between simile and metaphor, in his
view, since simile takes up the final position in sentence structure, the figure as the end focus
is more emphasized than the base. In addition, simile provides more chance for comparison
expansion. In the concluding section, he identifies another difference through the discussion
over might have been. Partington believes that simile means more remoteness in the
comparison than metaphor does.
It seems nowadays an academic fashion to have a book entitled Corpus-based
approaches to X or Corpus-based analysis/studies of X. This mirrors the increasing awareness
of the significance of empirical investigations. However, many of these books, including the
collection under review, are not genuinely empirical researches. Despite the various studies
and researches, empirical methodology, including corpus research, is in fact at its initial stage.
According to Geeraerts (2006: 23-25), empirical studies should meet the following five
criteria: data-driven, quantitative methods, formulation of hypotheses, operationalization of
hypotheses and empirical circle (see also Grondelaers et al. 2007). To be tested against these
parameters, in fact, most of the contributions simply fall into the scope of corpus-illustrated
approaches (objective data but subjective analysis). To date, corpus-based investigations can
be groups along a cline according to the empiricality of their methodologies. On one extreme,
there are corpus-illustrated ones, while on the other advanced extreme, there are corpus-
driven ones (see also Geeraerts 2006). In the book under review, the majority of the
contributions are close to the corpus-illustrated extreme (e.g. in Deignan, Allan) despite the
fact that a couple of articles are fairly close to the other extreme (e.g. in Stefanowitsch,
Martin, Markert and Nissim). Of course, it is not the present volume to be blamed since this is
the current situation concerning corpus research. However, the standard would have to be
raised if such collections are meant to be academically valuable and to attract enough
attention. In fact, I had expected an eye-catching methodology-oriented book containing more
workable empirical frameworks for exploring metaphor and metonymy. Therefore, it is a