Measuring the Meaning of Words in Contexts: An automated analysis of controversies about Monarch butterflies, Frankenfoods, and stem cells

Co-words have been considered as carriers of meaning across different domains in studies of science, technology, and society. Words and co-words, however, obtain meaning in sentences, and sentences obtain meaning in their contexts of use. At the science/society interface, words can be expected to have different meanings: the codes of communication that provide meaning to words differ on the varying sides of the interface. Furthermore, meanings and interfaces may change over time. Given this structuring of meaning across interfaces and over time, we distinguish between metaphors and diaphors as reflexive mechanisms that facilitate the translation between contexts. Our empirical focus is on three recent scientific controversies: Monarch butterflies, Frankenfoods, and stem-cell therapies. This study explores new avenues that relate the study of co-word analysis in context with the sociological quest for the analysis and processing of meaning.

💡 Research Summary

The paper tackles the problem of how words acquire and shift meaning across different domains at the science‑society interface. While traditional co‑word analysis treats words as static carriers of meaning, the authors argue that meaning is generated in sentences and, crucially, in the broader social and cultural contexts in which those sentences appear. To capture this dynamic, they introduce two reflexive mechanisms: “metaphors,” which create abrupt, image‑driven re‑configurations of meaning, and “diaphors,” which represent more gradual, field‑specific adjustments of a shared concept.

Methodologically, the study builds a large‑scale, time‑ordered corpus spanning 1995‑2022 that includes scholarly articles, mainstream news, policy documents, and online blogs/social‑media posts. After Korean‑language preprocessing (morphological analysis, stop‑word removal, stemming), the authors extract noun‑noun, noun‑verb, and verb‑verb collocations. These collocations are embedded using a Word2Vec model, allowing the construction of a high‑dimensional semantic space. By applying temporal clustering, they track how collocation groups (clusters) evolve over time. A supervised classifier distinguishes metaphorical shifts (large jumps between clusters, often triggered by external events) from diaphoric shifts (small, continuous movements within a cluster).

Three contemporary controversies serve as case studies: monarch‑butterfly conservation, “Frankenfood” debates over genetically modified (GM) foods, and stem‑cell therapies. In the monarch‑butterfly case, early discourse linked the species to environmental‑protection vocabularies (“habitat,” “conservation”). Around the mid‑2000s, agricultural terms (“pesticide,” “GM”) entered the network, producing a diaphoric transition that reframed the debate in terms of ecosystem services and economic costs. The Frankenfood controversy illustrates a strong metaphorical shift: scientific texts originally used neutral technical language (“transgene,” “phenotype”), but media and social‑media narratives rapidly attached affective terms (“risk,” “unnatural”) to the label “Frankenfood,” creating a new, emotionally charged frame that later influenced regulatory discussions. Stem‑cell therapy shows a mixed pattern: early scholarly work emphasized regenerative biology (“differentiation,” “plasticity”), while policy and public discourse introduced ethical and commercial vocabularies (“life‑rights,” “commercialization”). Both metaphorical (life‑source imagery) and diaphoric (clinical‑application adjustments) mechanisms operate simultaneously, reshaping public acceptance and policy formulation.

Quantitative results confirm that metaphorical shifts are characterized by sudden increases in semantic distance and cross‑cluster migration, often coinciding with high‑profile events (e.g., a landmark court ruling or a major news exposé). Diaphoric shifts display smaller distance changes but persist over longer periods, reflecting the gradual integration of scientific concepts into new domains. By mapping these patterns, the authors demonstrate that automated, context‑aware co‑word analysis can reveal not only the structure of knowledge networks but also the temporal dynamics of meaning translation across domains.

The authors conclude that distinguishing between metaphors and diaphors enriches our understanding of how scientific language is negotiated, reframed, and institutionalized. Their approach offers a scalable, data‑driven tool for scholars of science communication, policymakers, and STS researchers seeking to monitor and influence the evolution of public discourse around emerging technologies.