Erroneously categorized metaphors
In passing let us mention that some of the examples are “fake metaphors”: there are in fact cases of antanaclasis, i.e., the use of a word in two different senses, such as in the following examples, the antanaclastic word bolded:
- (29) “Oh, Jason, take me!” she panted, her breasts heaving like a college freshman on $1-a-beer night.
- (30) He fell for her like his heart was a mob informant and she was the east river.
- (31) The plan was simple, like my brother-in-law Phil. But unlike Phil, this plan just might work.
Here, as was the case in the metaphors that describe inherently funny referents, we face a possible complication: the connection between the two unrelated senses of the word may create a (failed) metaphor connecting two unrelated domains. If that is the case, then these cases of antanaclasis become failed metaphors by default and then there exists a prima facie plausibility in labeling them as metaphors (however “failed ones”).
Discussion of failed metaphors
The explanation for failed metaphors is both pragmatic and semantic. On the pragmatic side, it relies on an extension of the principle of cooperation (CP), which has been justified on independent grounds elsewhere (Attardo 1999; Eisterhold, Boxer and Attardo 2006). Essentially, this entails the postulation of a “principle of non-cooperation” (NCP), which governs the violation of the principle of cooperation. Applied to metaphors, the NCP prescribes avoidance of excessive or unnecessary detail (overdone metaphors), avoidance of multiple metaphors (mixed metaphors), avoidance of exaggeration, etc. Most (all?) of these rules follow directly from the tenets of the NCP, or perhaps from a straightforward extension of the CP to metaphorical discourse (e.g., do not exaggerate). Conversely, it should be stressed that failed metaphors are not completely pragmatic in nature: un-metaphors and overdone metaphors are clearly also semantic (no mapping, unwarranted mappings).