Corpus Information

  1. •randomly selected texts from four registers of the BNC-Baby

  2. •coded for all metaphor related language using the MIPVU procedure

  3. •post-hoc cleaning up of the database to reduce the degree of annotation error *

  4. * Note: A selected number of lexical units were systematically checked by a team of three researchers after completion of the project. While this reduced the number of errors in metaphor annotation, please note that the online corpus is likely not free of error. If you find one, please inform us and we may make adjustments in a future update of the corpus.

Four registers from the BNC-Baby coded for metaphor

academic texts               49,561 lexical units            16 fragments

conversations                 48,001 lexical units            24 fragments

fiction                               44,892 lexical units            12 fragments

news texts                       45,116 lexical units            63 fragments

total                               187,570 lexical units          115 fragments

Learn more about annotating the corpus for metaphor: