Underlying the data and visualisations in this web resource is a great deal of research. The evidence for the metaphorical connections presented is made as transparent as possible through links to the Historical Thesaurus of English online for each ‘example of metaphor’ and from there into the Oxford English Dictionary online. The method for identifying the connections is discussed briefly below and in more detail in our published articles.
In order to identify areas of meaning which overlapped metaphorically, it was first important to identify areas which shared words. For this purpose, the Historical Thesaurus was divided into 415 distinct semantic areas (e.g. Light, Fear, Weapons and Armour, etc.), which are grouped into the 37 areas of experience corresponding to the second level of the Historical Thesaurus hierarchy (e.g. Matter, Emotion, Armed Hostility, etc.) and the three top-level groupings (The External World, The Mental World and The Social World). These 415 categories were then mapped against each other – all items of vocabulary in one category (the ‘start category’) were checked computationally with all the other categories. Where our queries found identical word forms then these were pulled out along with their sense and date information.
Once all of the overlapping vocabulary from the ‘start category’ was identified in every other category then it was analysed manually by a member of the Mapping Metaphor team. We were looking for lexical evidence that one area of meaning was being understood in terms of another, or had been in the past. A large percentage of the overlap reflected processes other than metaphor, such as homonymy and non-metaphorical polysemy. However, there were well over 10,000 connections in which we found evidence of metaphor. We were particularly interested in finding clusters of metaphorical overlap which provided evidence of a systematic connection between two semantic categories. For example, because of the conceptual link between weather and anger we find many of the same words in our categories of Weather and Anger, such as storm, tornado, wind and fulminate. If this was the case then we labelled this connection as ‘Strong’, while if we found a small number of lexical metaphorical connections between two categories but little evidence of a systematic connection then we labelled this as ‘Weak’. The initial coding was then checked by another member of the team and each link was examined separately in each direction (e.g. from Light to Cleverness and again from Cleverness to Light). The final stage of analysis was to decide on a final code based on the accumulated coding, comments from coders and the pinpointing of evidence from specific word senses in the Historical Thesaurus instantiating the metaphorical link. If we could not find specific evidence of metaphorical transfer in the overlapping words between two categories then the link was not regarded as ‘Metaphor’.