<p>A widely held view in biology is that higher Linnaean categories or ranks (genera, families, etc.) ought to be regarded as ‘mere conventions’. That is typically understood to mean that their utility and application stem from (relatively) recent conventions that have been adopted for their usage in biological classification. There are no underlying theories or robust ontological concepts of these higher ranks, in contrast to taxonomic groupings like species or populations (disputed as those underlying concepts or theories may be). Absent these conventions, the application of higher ranks are in important ways arbitrary, reflecting their lack of theoretical and conceptual grounding. Indeed, some taxonomists have argued that this lack of grounding justifies adopting rank-free taxonomies. Michael Devitt (<CitationRef CitationID="CR15">2023a</CitationRef>) challenges this entrenched view, calling it an exaggerated skepticism. Instead, he takes the continued use of higher Linnaean ranks as evidence that they reflect “minimal category concepts,” which do genuine theoretical and explanatory work by marking out a relative level of taxonomic hierarchy. We disagree. Here, we unpack what it means to treat Linnaean categories as mere conventions, arguing they are better understood as informational proxies for phylogenetic hierarchies. Yet, Devitt is right to focus on this case, as it provides an exemplar of how scientists retain the utility of scientific practices that outlive their underlying initial theoretical and conceptual justification. We conclude by offering an alternative explanation for this in the context of a pragmatic account of science that treats Linnaean classification as a biological formalization.</p>