The notion of deliberate metaphor use is dependent on a number of assumptions. This paper aims to elucidate these assumptions, their core concepts, and the most important relations between all of these. Focusing on metaphor in written or printed texts, central claims will include the following:

  1. All metaphor is a matter of language use, and is therefore a matter of utterances exhibiting dimensions of expression (‘language’), conceptualization (‘thought’), and interaction (‘communication’).
  2. All language use is a matter of discourse, which is a matter of the use of text in code in context.
  3. All discourse is a matter of genre knowledge, and therefore exhibits typical properties for distinct classes of genre events and their related instantiations in documents (like written or printed texts).
  4. Since all genre events are based in genre knowledge, they involve goal-directed action on the part of the language user, which is realized in intentional language use, including loose, vague, indirect and ambiguous language use. It should be noted that goal-directed action and language use as intentional are not necessarily conscious, which is a different phenomenon. This makes all metaphor in language use intentional by definition in this pragmatic sense of the term ‘intentional’
  5. However, some metaphor is more intentional than most metaphor: some metaphors are used deliberately as metaphors, in that they are intended to draw the addressee’s attention to the source domain as a separate domain of reference in the utterance which is used as a starting point for setting up a non-literal comparison as part of the meaning of the utterance.
  6. This momentary attention to the source domain in deliberate metaphor use as a distinct domain of reference serves as an alien perspective on the target domain as the locally dominant domain of reference and can be intentionally used for a wide range of discourse functions in relation to the text, code, and context properties of the genre event.
  7. Corpus work shows that claims 5 and 6 do not apply to the bulk of metaphor in language use.
  8. In terms of processing, the bulk of metaphor in language use may not require source domain access to construct the meaning of the metaphorical expression in the utterance, as it may instead be processed by lexical disambiguation and conceptual categorization. This produces a paradox of metaphor, suggesting that most metaphor is not processed metaphorically, whereas only some metaphor (deliberate metaphor)
  9. We need a three-dimensional model of metaphor in language use as proposed in DMT to resolve the paradox of metaphor, as it is ‘metaphor in communication’ (deliberate metaphor) that is predicted to be always processed metaphorically.
  10. All of these claims are equally applicable to other, more multimodal manifestations of discourse and can be researched in various ways by different disciplines to different academic effects.