Metonymy as a usage event
Some properties of metonymy
Distinguishing metonymy from metaphor
In cognitive linguistics, metaphor is generally regarded as relating two separate conceptual domains via a set of unidirectional mappings from a source domain into, or ideally, onto a target domain. The general problem with this approach is related to the vagueness of the notion of domain. One difficulty with this definition is that– apart from the most generic conceptual domain, which encompasses all conceptual knowledge– every domain has a superordinate domain and thus two domains will always have at least one domain in common that contains them as proper subdomains. The difficulties arising from the notion of domain are the same for metaphor as for metonymy (see e.g. Barcelona 2002 for insightful discussion).
I propose that reference to “separate domains” should be avoided in a working definition of metaphor. Nothing would speak against the assumption that the source domain and the target domain intersect. For example, an utterance like (11) could be interpreted metaphorically:
(11) My (pet) tiger is a lion.
The conceptual frames for tiger (target) and lion (source) overlap considerably. Still, it is possible to construe a metaphoric sense for (11) where properties of the concept LION are mapped onto tigers. It thus does not seem to be a necessary condition for metaphors that the source domain and the target domain be disjunct. As for metonymy, I suggest a semiotically based conception of metaphor, which is represented in Figure 3 in Some properties of metonymy, Primary metaphors. Following Hiraga (1998), I propose that metaphor is based on an iconic relation between source meaning and target meaning. The source meaning has some frame or domain structure that is iconically replicated in the target domain. This approach is, to a certain extent, in line with traditional cognitive metaphor theory as advocated by Lakoff and Johnson (1980). The present approach differs from Lakoff and Johnson in that it treats the source meaning of a metaphor, in analogy to metonymy, as a thought vehicle (signifier) for accessing a target meaning (signified). The conception of metaphor as an iconic relationship accords with traditional accounts of metaphor in that it relies on some notion of similarity between source and target. Hiraga (1998) suggests that in metaphor the resemblance between source and target is created rather than given, i.e. even a priori dissimilar domains can be connected through metaphoric mappings. The resemblance in this case is structural rather than a “direct” imitation of (perceived) “reality” in the signifier of the linguistic sign, as in onomatopoeia or other kinds of sound symbolism. The structural resemblance is one between the two conceptual frames, the source and the target, as has been demonstrated by Lakoff and Johnson with illuminating examples. The view of metaphor as a special case of iconicity (isomorphism) avoids the pitfalls of having to come up with some criterion for distinguishing single domains from separate domains. Figure 3 diagrams in Some properties of metonymy, Primary metaphors the basic metaphoric relation. Note that the target meaning is printed in bold as in the case of metonymy, indicating that it is conceptually prominent. For convenience of graphic representation, source and target domain are represented as distinct in Figure 3 (see Some properties of metonymy Primary metaphors).
