A Journal Has Two Separable Properties: Emergent Thematic Identity and Competitive Position.

Abstract

For decades the quartile of a journal has been read as a statement about the journal itself: a move from Q1 to Q2 is taken to mean decline. Yet a quartile is a relative ranking within a reference set, whereas a journal also possesses properties that arise from what it actually publishes. This conceptual paper argues that conflating these two things is a category error with direct consequences for how journal metrics are interpreted. The paper advances a single hypothesis: a scientific journal possesses two operationally separable properties — an article-derived thematic identity (what it publishes) and a reference-set-dependent competitive position (where it ranks). The first is intrinsic and stable; the second is relational and mobile, moving predominantly because the surrounding competitive set recomposes rather than because the journal itself changes. The contribution is conceptual rather than algorithmic: it identifies and defends this separation, distinguishes measurement from evaluative judgment, and draws out the interpretive consequences for quartile-based evaluation. The argument is grounded in a companion empirical study whose four preregistered comparisons are presented not as independent analyses but as a single validation program: that the distinction does not require editorial categories to be wrong (Test A), that the two properties can interact yet remain non-identical (Test B), that quartile movement is, at the population level, predominantly a matter of competitive-set recomposition (Test C′), and that thematic identity is temporally stable (Test D). Their convergence supports treating thematic identity and competitive position as two separable properties of scientific journals and, once they are separated, makes a family of open, reproducible, and independently auditable interpretation metrics definable. This is a conceptual companion paper, prepared for submission to Quantitative Science Studies and not yet peer reviewed. It performs no new computation: every empirical claim is inherited by citation from the companion empirical study (Reconstructing Journal Thematic Identity from Article-Level Topics: A Reproducible Framework and a Preregistered Temporal Validation; DOI 10.5281/zenodo.21659763), whose sealed datasets, preregistrations, reproducibility.