Metaphor Identification Procedure (MIP/MIPVU)

Description

Reliably identifies metaphorically used words in user language and groups them into conceptual metaphors that reveal how users think about a product or domain.

Conceptual metaphor theory holds that we understand abstract domains through concrete ones (ARGUMENT IS WAR; TIME IS MONEY). Identifying metaphors in discourse used to be intuitive and unreliable. The Pragglejaz Group's MIP gives a step-by-step procedure: read the text, determine lexical units, establish each unit's contextual meaning, check whether it has a more basic meaning in other contexts, and mark it metaphorical if the contextual meaning contrasts with but can be understood by comparison with the basic one. MIPVU refines this and reports high inter-coder reliability. Applied to interviews and reviews, it reveals users' conceptual models (the cloud as a place, AI as a person, money as water) that shape expectations and errors.

Attributes

Loss type
Unmigrated
Effort
Weeks

Steps

  1. Select texts: interview transcripts, reviews, support chats about the domain.
  2. Apply MIP/MIPVU word by word: determine lexical units, contextual and basic meanings, and mark metaphor-related words, using a dictionary such as Macmillan for basic senses.
  3. Check reliability with two coders on a sample (kappa typically above 0.8 is attainable).
  4. Group metaphor-related words into source-target mappings (e.g. DATA IS A LIQUID: leaks, flows, streams).
  5. Compare dominant metaphors across segments and contexts, and relate them to expectations and errors.
  6. Use productive metaphors in interface language and avoid misleading ones.

Sources

  • Pragglejaz Group, MIP: A method for identifying metaphorically used words in discourse (2007), Metaphor and Symbol
  • Steen et al., A Method for Linguistic Metaphor Identification: From MIP to MIPVU (2010)
  • Lakoff & Johnson, Metaphors We Live By (1980)

Also known as

  • MIP
  • MIPVU
  • Pragglejaz procedure
  • Conceptual metaphor analysis

Categories

  • Content
  • Linguistics, Semiotics & Discourse