Transferability of mathematical judgment across profiles

Determine whether rules concerning question selection, mathematical taste, and movement between local and global viewpoints transfer broadly across mathematicians in AI-assisted research collaborations.

Background

The paper proposes separating broadly transferable components of a research profile from components that reflect an individual mathematician’s judgment. It suggests that verification and provenance rules may transfer more readily than preferences governing which questions to pursue or how to move between mathematical viewpoints. The transferability of these individual-level forms of judgment remains unresolved.

References

Rules for verification and provenance may transfer broadly. Whether the same is true of question selection, mathematical taste, or movement between local and global viewpoints is less clear.

— The Mathematician Leads: Building a Research Profile in the Age of Frontier AI  (2608.28369 - Hofer, 28 Aug 2026) in Section 5, “Neighboring ideas and the missing layer”

However, to state explicitly, this gap is not proven to be the following. It is not shown to be tacit knowledge as a person holds it. It is not separated into Collins's somatic and collective kinds; although the fact that more people passing judgement correlates with a bigger gap points to CTK, nothing explicitly tests this.

— Verifiable, Articulable, and Tacit Components of Preference  (2610.03025 - Spangher et al., 2 Oct 2026) in Limitations, paragraph “Whether this is tacit knowledge”