Pathologies on the Hilbert scheme of points (Q1985461)

R. Vakil defined singularity type as an equivalence class of pointed schemes under the relation generated by \((X,x)\sim(Y,y)\) if there is a smooth morphism \((X,x)\rightarrow(Y,y)\) [\textit{R. Vakil}, Invent. Math. 164, No. 3, 569--590 (2006; Zbl 1095.14006)]. He showed that every singularity type over \(\mathbb Z\) appears on various moduli spaces of well-behaved objects: one says that Murphy's law holds for these moduli spaces. The author investigates the Hilbert scheme of points, which is missing in Vakil's list. The main theorem states that Murphy's law holds up to retraction for \(\mathrm{Hilb}_{\mathrm{pts}}(\mathbb A^{16}_{\mathbb Z})\). The proof proceeds by a series of reductions from objects with more structure. The main role is played by a generalized Białynicki-Birula decomposition [\textit{J. Jelisiejew} and \textit{Ł. Sienkiewicz}, J. Math. Pures Appl. (9) 131, 290--325 (2019; Zbl 1446.14030)]. In order to construct the local retractions, the author uses TNT frames. Using concrete singularity types, the author shows that \(\mathrm{Hilb}_{\mathrm{pts}}(\mathbb A^{16}_{\mathbb Z})\) and \(\mathrm{Hilb}_{\mathrm{pts}}(\mathbb A^{16}_{\mathbb C})\) are non-reduced, answering questions raised by J. Fogarty [\textit{J. Fogarty}, Am. J. Math. 90, 511--521 (1968; Zbl 0176.18401)]. He also shows that not all finite schemes over finite fields lift to characteristic zero, answering a question by R. Hartshorne [\textit{R. Hartshorne}, Deformation theory. Berlin: Springer (2010; Zbl 1186.14004)]. As a corollary of the main theorem, Murphy's law holds up to retraction for \(\mathrm{Hilb}_{\mathrm{pts}}(\mathbb P^{16}_{\mathbb Z})\). Since the forgetful functor from embedded to abstract deformations of a finite scheme is smooth, the above described pathologies appear for abstract deformations of finite schemes. The author highlights that the choice of ambient dimension \(n=16\) was made for sake of transparency in the proof, but he suggests that the result may hold for \(n=6\) or even \(n=4\).