In the context of subject ontogeny research, that is, the study of how subjects change or do not change through revision of indexing languages, there a number of methodological questions. The nature of semantics, ideas, and subjects is a long and rich discussion in the context of knowledge organization (Raganathan, 1967; Wilson, 1968; Foskett, 1969; Hjørland, 1992; Mai, 2001; Olson, 2002; Zeng, Zumer, and Salab,a 2010; Dutta, 2015). This discussion is kept current by Birger Hjørland's reference work on the matter (Hjørland, 2017). However, the discussion predates the founding of the discipline of library and information science. Ranganathan cites the early Vedic work (fl. 1500-600 BCE) on subjects (Ranganathan, 1957). Plato (c. 380 BCE) lists those subjects that are required for a good education of future leaders and contemplates their nature (Emlyn-Jones and Preddy 2013). Following Plato's dialogues the Western encyclopedists, and ultimately bibliographers, write subjects into their work, sometimes reflecting on them, but often not (c.f., Diderot and d'Almbert 1751-1772). Our contemporary preoccupation with subjects and how they remain the same or change is linked directly to our philosophical assumptions. In this context we often talk about realism and nominalism (Bosewell, 1986). Realism considers subjects (categories, ideas, semantics) to be “the footprints of reality... they exist because humans perceive a real order of the universe and name it,” (Bosewell, 1982 p.91). The nominalists subscribe to “the belief that categories are only the names... of things agreed upon by humans, and that the ‘order' people see is their creation rather than their perception,” (Bosewell, 1982 p. 91). There are refinements and caveats to these two extremes, and there are different and various ways realism or nominalism manifests (e.g., in the mind, external to the mind, in action, or in culture). Further, it is rare to see any researcher hold one of these beliefs absolutely. It is possible to hold a realist position and assume the same phenomenon is named differently in different context and in different points in time. And here is where the methodological question comes back to subject ontogeny work. Namely, does research done in one or the other camp use different methods to arrive at a description of how a subject changes over time in the life of an indexing language?
