Diachronic Data Analysis Supports and Refines Conceptual Metaphor Theory

TL;DR

This paper presents a statistical, data-driven approach to analyzing metaphors over time, empirically testing longstanding theories and exploring their systematic features to enhance NLP understanding.

Contribution

It introduces the first empirical investigation of metaphor features and integrates metaphor theory into quantitative NLP frameworks.

Findings

Empirical validation of longstanding metaphor theories

Identification of systematic features of metaphors

Data-driven insights into metaphor evolution

Abstract

As a contribution to metaphor analysis, we introduce a statistical, data-based investigation with empirical analysis of long-standing conjectures and a first-ever empirical exploration of the systematic features of metaphors. Conversely, this also makes metaphor theory available as a basis of meaning emergence that can be quantitatively explored and integrated into the framework of NLP.