Comparison of Motivic Homotopy Theories
Tianjian Tan · 2026
Read Comparison of Motivic Homotopy Theories by Tianjian Tan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-11753.txt.
Introduction
Tan builds comparison functors between rival motivic homotopy theories. One runs from the dual of SH to localizing motives with A1-invariance imposed, in the sense of Blumberg, Gepner, and Tabuada. A second, non-A1-invariant analogue links the dual of the Annala-Iwasa-Hoyois category MS to plain localizing motives. By a Barr-Beck argument, both functors factor through module categories over a dual K-theory spectrum. Over a field with resolution of singularities, the A1-invariant functor proves fully faithful. The non-A1-invariant one does not.
Catalog record
| CREATOR | Tianjian Tan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 106846 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.11753
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Comparison of Motivic Homotopy Theories Tianjian Tan Abstract We construct a comparison functor from the dual category of motivic homotopy category 𝒮 ℋ \mathcal{SH} to the category of 𝔸 1 \mathbb{A}^{1} -invariant localizing motives Mot loc 𝔸 1 \operatorname{Mot}_{\operatorname{loc}}^{\mathbb{A}^{1}} in the sense of Blumberg, Gepner and Tabuada (with 𝔸 1 \mathbb{A}^{1} -invariance imposed). We as well construct its non- 𝔸 1 \mathbb{A}^{1} -invariant analogue: a functor from the dual category of Annala-Iwasa-Hoyois’s non- 𝔸 1 \mathbb{A}^{1} -invariant motivic homotopy category ℳ 𝒮 \mathcal{MS} to Mot loc \operatorname{Mot}_{\operatorname{loc}} . After the Barr-Beck argument, these functors factor through categories of modules over a dual version of ( 𝔸 1 \mathbb{A}^{1} -invariant) K-theory spectrum KGL ( 𝔸 1 ) \operatorname{KGL}^{(\mathbb{A}^{1})} . Over a field that admits resolution of singularities, we show that the 𝔸 1 \mathbb{A}^{1} -invariant factored functor is fully-faithful, while the non- 𝔸 1 \mathbb{A}^{1} -invariant one is not in general. 1 Introduction In the thesis, we want to compare motivic homotopy theories and and noncommutative motives in the sense of [ 3 ] . To be precise, we compare four motivic categories • non- \mathbb{A}^{1} -invariant motivic homotopy category \mathcal{MS} of Annala, Iwasa and Hoyois [ 2 ] , [ 1 ] ; • \mathbb{A}^{1} -invariant motivic homotopy category \mathcal{SH} of Morel and Voevodsky; • category of localizing motives \operatorname{Mot}_{\operatorname{loc}} of Blumberg, Gepner and Tabuada [ 3 ] ; • category of \mathbb{A}^{1} -invariant localizing motives \operatorname{Mot}_{\operatorname{loc}}^{\mathbb{A}^{1}} The comparisons between the first two, and the last two categories are given by localizations with respect to all projections X\times\mathbb{A}^{1}\rightarrow X . So we focus on the comparison between the first and the third categories, and the \mathbb{A}^{1} -invariant analogue between the second and the fourths. There is a natural candidate for the comparison functor between \mathcal{SH} and \operatorname{Mot}_{\operatorname{loc}}^{\mathbb{A}^{1}} (resp. between \mathcal{MS} and \operatorname{Mot}_{\operatorname{loc}} ), at least when we restrict the motivic homotopy category to the category of smooth schemes of finite presentation, that is, \displaystyle(Sm^{\operatorname{fp}})^{op} \displaystyle\rightarrow\operatorname{Mot}_{\operatorname{loc}}^{\mathbb{A}^{1}} \displaystyle X \displaystyle\mapsto\operatorname{L}_{\mathbb{A}^{1}}\mathcal{U}(\operatorname{perf}_{X}). resp. \displaystyle(Sm^{\operatorname{fp}})^{op} \displaystyle\rightarrow\operatorname{Mot}_{\operatorname{loc}} \displaystyle X \displaystyle\mapsto\mathcal{U}(\operatorname{perf}_{X}). The functor \mathcal{U} above is the universal localizing invariant constructed in [ 3 ] . This functoriality suggests us extending the domain to some kind of dual category of \mathcal{SH} (resp. of \mathcal{MS} ). It turns out that the desired notion of duality here is the dual object in the symmetric monoidal category of stable presentable categories \mathsf{Pr}^{\mathrm{L}}_{st} , which we will denote by (-)^{\vee} . A first example is that, for a small category \mathcal{C} , the dual of the spectra-valued presheaf category is simply \mathcal{P}sh(\mathcal{C}^{op}) , the presheaf category on \mathcal{C}^{op} . Such dualizibility is largely studied and applied in the work of Clausen-Scholze [ 5 ] , Efimov [ 6 ] , etc. To extend the natural candidate above to a functor \mathcal{SH}^{\vee}\rightarrow\operatorname{Mot}_{\operatorname{loc}}^{\mathbb{A}^{1}}, resp. \mathcal{MS}^{\vee}\rightarrow\operatorname{Mot}_{\operatorname{loc}}, we need a universal property that characterizes functors out of \mathcal{SH}^{\vee} (resp. \mathcal{MS}^{\vee} ). It turns out that just as \mathcal{SH} (resp. \mathcal{MS} ) is defined from presheaf category by first localizing with respect certain descent conditions, and then inverting the Tate circle \mathbb{P}^{1} , the dual category \mathcal{SH}^{\vee} (resp. \mathcal{MS}^{\vee} ) admits a dual description: Proposition 1.1 . \mathcal{SH}^{\vee} (resp. \mathcal{MS}^{\vee} ) is equivalent to the category constructed from the spectra-valued presheaf category \mathcal{P}sh((Sm^{\operatorname{fp}})^{op}) , by first localizing it with respect to certain codescent conditions, and then formally inverting a dual version of \mathbb{P}^{1} . To deduce the proposition above, we computed the dual category for general formal inversion \mathcal{C}[X^{-1}] for a stable, presentably symmetric monoidal, compactly generated category \mathcal{C} and X a compact object in \mathcal{C} : Proposition 1.2 (Proposition 2.14 ) . Let \mathcal{C} be a stable, compactly generated, presentably symmetric monoidal category. Let X\in\mathcal{C} be a compact object. Then we have a canonical equivalence of categories \mathcal{C}^{\vee}[(X^{dual})^{-1}]\simeq\mathcal{C}[X^{-1}]^{\vee}. Moreover, the symmetric monoidal structures are compatible. Such characterization of \mathcal{SH}^{\vee} (resp. \mathcal{MS}^{\vee} ) as formal inversion of category of cosheaves is equipped with a natural universal property of the desired form (Proposition 3.12 ): Proposition 1.3 . The composition \operatorname{L}_{(\mathbb{P}^{1})^{dual}}\circ\operatorname{L}_{\operatorname{coNis},\operatorname{co}\mathbb{A}^{1}}\circ j induces, for each presentable, stable category \mathcal{D} , a fully faithful functor \operatorname{Fun}^{L}(\mathcal{SH}(S)^{\vee},\mathcal{D})\rightarrow\operatorname{Fun}(Sm^{\operatorname{fp}}(S)^{op},\mathcal{D}) with the essential image spanned by \mathbb{A}^{1} -invariant functors F that preserve terminal object, send Nisnevich squares to pushouts, and send the fiber of fib(F(\mathbb{P}^{1})\xrightarrow{F(\infty)}F(S)) to an invertible object. Moreover, since the symmetric monoidal structures are compatible, the claim remains valid for symmetric monoidal category \mathcal{D} and symmetric monoidal functors. In particular, we have \operatorname{Fun}^{L,\otimes}(\mathcal{SH}(S)^{\vee},\operatorname{Mot}(S)_{\operatorname{loc}}^{\mathbb{A}^{1}})\simeq\operatorname{Fun}^{\otimes}_{\operatorname{Nis},\mathbb{A}^{1},\mathbb{P}^{1}}(Sm^{\operatorname{fp}}(S)^{op},\operatorname{Mot}(S)_{\operatorname{loc}}^{\mathbb{A}^{1}}) We also have a non- \mathbb{A}^{1} -invariant version (Proposition 3.19 ): Proposition 1.4 . The composition \operatorname{L}_{(\mathbb{P}^{1})^{dual}}\circ\operatorname{L}_{\operatorname{coNis},\operatorname{coebu}}\circ j induces, for each presentable, stable category \mathcal{D} , a fully faithful functor \operatorname{Fun}^{L}(\mathcal{MS}(S)^{\vee},\mathcal{D})\rightarrow\operatorname{Fun}(Sm^{\operatorname{fp}}(S)^{op},\mathcal{D}) with the essential image spanned by functors F that preserve terminal object, send Nisnevich squares and elementary blowup squares to pushouts, and send the fiber of fib(F(\mathbb{P}^{1})\xrightarrow{F(\infty)}F(S)) to an invertible object. Moreover, since the symmetric monoidal structures are compatible, the claim remains valid for symmetric monoidal category \mathcal{D} and symmetric monoidal functors. In particular, we have \operatorname{Fun}^{L,\otimes}(\mathcal{MS}(S)^{\vee},\operatorname{Mot}(S)_{\operatorname{loc}})\simeq\operatorname{Fun}^{\otimes}_{\operatorname{Nis},\operatorname{ebu},\mathbb{P}^{1}}(Sm^{\operatorname{fp}}(S)^{op},\operatorname{Mot}(S)_{\operatorname{loc}}) We check that the natural functors (Sm^{\operatorname{fp}})^{op}\rightarrow\operatorname{Mot}_{\operatorname{loc}}^{(\mathbb{A}^{1})} satisfy the above conditions, and hence define the comparison functors. A consequence of our universal properties is that the comparison functors are automatically strongly monoidal left adjoint functors, such functors factor formally through the category of modules \operatorname{Mod}_{\Phi_{\ast}(1)}(\mathcal{SH}^{\vee}) (resp. \operatorname{Mod}_{\Psi_{\ast}(1)}(\mathcal{MS}^{\vee}) ), where we use \Phi_{\ast} (resp. \Psi_{\ast} ) to denote the corresponding right adjoint functor. We are interested in the fully-faithfulness of the factored functor \operatorname{Mod}_{\Phi_{\ast}(1)}(\mathcal{SH}^{\vee})\rightarrow\operatorname{Mot}_{\operatorname{loc}}^{\mathbb{A}^{1}}, resp. \operatorname{Mod}_{\Psi_{\ast}(1)}(\mathcal{MS}^{\vee})\rightarrow\operatorname{Mot}_{\operatorname{loc}}. In fact, restricted on schemes X with sufficiently good dualizibility properties, both functors are fully faithful. A key difference appears when we consider the rigid generation of domains: working over a field k that admits resolution of singularity, \mathcal{SH}(k)^{\vee} is rigidly generated in the sense that it has a set of dualizable, compact generators. As a consequence, the restricted fully-faithfulness gives rise to a complete fully-faithfulness: Proposition 1.5 (Proposition 4.2 ) . Over a field k and after inverting the exponential characteristic e , the functor \widetilde{\Phi}:\operatorname{Mod}_{\Phi_{\ast}(1)}(\mathcal{SH}(k)[\frac{1}{e}]^{\vee})\rightarrow\operatorname{Mot}(k)^{\mathbb{A}^{1}}_{\operatorname{loc}}[\frac{1}{e}] is fully faithful on the level of mapping spectra. While in the non- \mathbb{A}^{1} -invariant case, \mathcal{MS}(k)^{\vee} is not rigidly generated and we give an example of the failure of the fully-faithfulness in this setting. Two key ingredients here are the uncountability of mapping spectra between certain (compact) objects in \operatorname{Mot}_{\operatorname{loc}} due to a theorem of Efimov ( [ 7 , Theorem 9.1] , see also Section 4.2 ), and the countability of mapping spectra between compact objects in general in \mathcal{MS}(k) : Proposition 1.6 (Proposition 4.3 ) . Let k be a countable field and e be the exponential characteristic of k . Then the full subcategory of \mathcal{MS}(k)[\frac{1}{e}]^{\vee} spanned by compact objects is countable, in the sense that the \pi_{0} of all the mapping spectra are countable. 1.1 Acknowledgements This is my master thesis at Sorbonne Université under the supervision of Vova Sosnilo. I want to express my great gratitude to him for proposing the topic of the thesis, sharing with me his ideas, answering my endless questions and giving feedback on the early version of this thesis. Moreover, I thank him sincerely for bringing me a wonderful summer thinking and discussing on these beautiful maths. I am also very grateful to Marc Hoyois and Maxime Ramzi for pointing out several mistakes and