theHigherGeometer

rimcræftiga |
bespoke constructions in categorified geometry since 2010 |
dude

Blog: https://thehighergeometer.wordpress.com/
Coffee fund: https://buymeacoffee.com/highergeometer
Location: Adelaide, Australia

Public Key

npub1d60hhaertjrlmhefe54s00hfyga38wangrqthcu3zdnnnfm3myase5mhwm

NIP-05 Address

highergeometer_at_mathstodon.xyz@momostr.pink

Profile Code

nprofile1qqsxa8mm7u34eplamu5u62c8hm5jywcnhwe5ps9muwg3xeee5acajwcpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dspqzlvc

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1d60hhaertjrlmhefe54s00hfyga38wangrqthcu3zdnnnfm3myase5mhwm

Show more details

Last Notes

2026-08-31 12:54:34 UTC

Writing notes for an intro lecture on Lie groupoids tomorrow for grad students, the first of two, and it turns out that I will have trouble keeping the material to a manageable amount. I want to get at the end of the second one to a statement of the slice theorem for proper Lie groupoids, which says how there are nice local models that are products of boring transitive examples (in the orbit direction) and the action groupoids from a unit open ball in a representation G → O(n) of a compact Lie group G, for the transverse direction. Certainly not the proof, though!

2026-08-08 23:35:45 UTC

- reply

This was from my latest publishing agreement, I didn't opt for publishing with Springer with an Creative Commons license, because I decline to pay them money when the arXiv version is free for anyone to read anyway, and I link to it from the published abstract.
@npub1nf4…nqe4
https://media.mathstodon.xyz/media_attachments/files/117/062/501/514/177/677/original/96b47a620dc0b338.png

2026-08-08 22:25:32 UTC

- reply

It's in the publishing agreement that you would have/will sign, they don't email you personally to force it.
@npub1nf4…nqe4

2026-08-08 12:49:01 UTC

From @npub1nf4…nqe4 's https://arxiv.org/abs/2509.22584
I also, based on my experience last year, will avoid Springer Nature in future, unless it's unavoidable (eg a student's paper that really has nowhere else to go), or I make a one-time exception for a bucket-list journal: Comm. Math. Phys.
https://media.mathstodon.xyz/media_attachments/files/117/059/955/898/756/561/original/a83ac55e7f0e0a69.png

2026-07-27 05:16:39 UTC

"The situation isn’t quite like the situation in chess, where powerful chess programs got to the point they could beat humans, but human to human chess competition survived. Where we’re going could be analogous to chess playing where everyone is using, with or without acknowledging it, some sort of chess program."
— @npub1u9z…448w (https://www.math.columbia.edu/~woit/wordpress/?p=15787)
I'm kinda sick of people comparing the use of AI in math to the use of engines chess. They really, really aren't comparable. It's more like the use of steroids in sport, except there are no regulations whatsoever, and a small handful of companies are giving them free to some people, selling them to others, people talk about it openly and even boast of using them. Or maybe more like little motors in bicycles in the Tour de France, because it's external technology, not a self-modification thing (at least, not in the way steroids directly boost one's muscle mass etc). Mathematics isn't a supervised and highly-regulated activity.
I actually do use AI to discuss bits of mathematics I want to learn, where it's not obvious what source I can go to. I ask it questions about small things that are pre-existing, or small calculations in areas I'm not familiar with, as a way to absorb the ideas. Very often it's a painful process of me telling the AI it's actually wrong, or not really answering my questions, but I guess this type of thing is getting rarer (I can't quantify, my usage is sporadic).

2026-04-22 00:59:48 UTC

- reply

Oh, I thought it was from the IAS or similar. Looks very official and fancy ^_^. I wouldn't have been so blatantly critical if I'd know it was yours! 😅

2026-04-21 23:52:17 UTC

- reply

I would feel much better about that otherwise rather nice poster if the truncated decimal representation were instead written 0.78130...
Because reading it literally as written you have "[...] 0.78130 IS IRRATIONAL!"

2026-04-21 02:06:07 UTC

- reply

The section of the Grassmann number page cites a book I can't read, but it seems very much like a very imprecise and wibbly-wobbly physics picture (not to mention the discussion of how many fermions you have/need, and the specific mentions of 1, 2, 4 odd dimensions, which are super-specific to supersymmetric theories (under the names N=1, N=2, N=4 SYM etc)

2026-04-21 01:47:55 UTC

I think the definition here is wrong: https://en.wikipedia.org/wiki/Supermanifold#Concrete:_as_a_smooth_manifold it claims the "one-dimensional space of Grassmann numbers" are essentially given by the exterior algebra on a countably infinite-dimensional vector space.
I think the page is inconsistent with itself, and the target of the link has other issues: I don't believe you can take an infinite-dimensional alternating algebra and magically "employ a topology" that somehow reduces it to be finite-dimensional.
https://media.mathstodon.xyz/media_attachments/files/116/440/143/692/526/711/original/34de840a105613d0.png
https://media.mathstodon.xyz/media_attachments/files/116/440/163/331/715/506/original/a55154534173234c.png

2026-04-18 08:06:34 UTC

- reply

Someone called the ChatGPT proof when it first landed a "Proof from The Book", but almost surely the one by @npub1gve…sduy that is much shorter and cleaner should be described as that, right?
I wouldn't exactly say "completing the proof" since the first AI-generated proof was subsequently formalised in Lean (and given the provenance, I believe the owners of that process know how to make sure this is a legit formalisation). But certainly "a complete understanding" is due to the people who simplified the proof.
@npub1nf4…nqe4

2026-04-18 00:06:11 UTC

- reply

FWIW, here's the short proof by @npub1gve…sduy at the link, it's very much "better" than the GPT one, I think, and certainly conceptually clearer.
https://media.mathstodon.xyz/media_attachments/files/116/422/736/893/256/905/original/bf6971d5e027c6b8.jpg

2026-04-12 03:53:59 UTC

- reply

I hope there is a physical system that gets discovered one day, most likely in material science or solid state physics, that is necessarily modelled by nonabelian 2-bundles. I believe there are known examples that use bundle gerbes, to the extent of articles in glossy experimental science magazines with images generated from measuring things in a lab. it's only taken 25-30 years to get to that point, from Michael Murray's original paper in the JLMS. So at some point in the next decade or so, who knows?

2026-04-08 08:34:37 UTC

- reply

I think it's ok as a technical crutch, but actually doing the Wick rotation using imaginary time is nontrivial. Witness @npub1u9z…448w slow burn trying to come to grips with how this is supposed to work on a serious level of modern math physics, even without fancy proposals for universe origins.
@npub18zu…xdht @npub1nf4…nqe4

2026-04-08 00:43:19 UTC

- reply

He worked on both the geometry of numbers https://en.wikipedia.org/wiki/Geometry_of_numbers as well as mathematical physics later in his life. When he was just 18 he won (jointly, due to his youth) a big prize from the French Academy of Sciences. Was a close friend of Hilbert and was in the Göttingen mathematics department after Zurich, which was a position taken out of necessity to have a job. Also, Carathéodory was his student, also a later Göttingen member.

2026-04-08 00:28:43 UTC

Hilbert, by Constance Reid.
A great book, showing David Hilbert's passage from a bold young and ambitious mathematician to an old man surrounded by the ruin of the mathematics department in Göttingen in the 1930s. This helped me place a lot of names of contemporaries, and I can appreciate Minkowski's truncated career much better, I had no idea how big a deal he was in this whole circle, nor that he died early. The author treats the mathematics very well even though she's not trained in it, and from a modern standpoint it helps me connects back from post-1930s work to the previous generation's revolutionary developments.
Electronic version: https://doi.org/10.1007/978-1-4612-0739-9
#Read2026
https://media.mathstodon.xyz/media_attachments/files/116/366/227/433/424/499/original/e35ca4d2d7271b78.jpg

2026-03-31 16:23:11 UTC

- reply

your suggestion made me think of this Far Side cartoon:
https://static1.cbrimages.com/wordpress/wp-content/uploads/2023/05/the-far-side-didnt-wash-hands.jpg

2026-03-30 08:15:58 UTC

- reply

Well, I heard Kevin talk about problems in the Langlands program where Jim Arthur claimed big results in a number of really meaty "forthcoming" papers and people took his word, and then it turned out there were big problems. In that kind of mathematics, the problem wasn't that it wasn't formalised, but that people are perhaps getting a bit too confident.....

2026-03-30 01:15:54 UTC

- reply

I take you point on Kevin's comment ( @npub1vzx…ncd7 if he logs in and sees this), but a) Kevin likes making hyperbolic public statements to stir the pot b) never forget the journalist and then their editors in the process and c) what Kevin in currently doing is actually stopping to re-think how best to prove FLT, in a way that's novel and more cleanly abstracted and packaged. Perhaps this type of novel re-thinking is not happening every day in number theory, and so it feels really fun. Obviously, category theorists abstract structures and proofs all the time, so this process feels like an absolute no-brainer to us. (Nota Bene: I can also make hyperbolic statements to stir the pot).
Actually looking at the mathematics he is doing currently is really interesting, because it's making it much clearer and leaner (pun not intended), or at least exposing where the really hard core theorems are. Ideally those can subsequently get the same treatment, and people don't just cite a 150-page paper from the 70s or 80s that makes enormous effort to prove a very elementary statement as a black box, but think through if it can be improved.

2026-03-28 06:00:45 UTC

- reply

It's true that every 2-element set has a unique Z/2-torsor structure, but in this case we are talking about a "physical" set, and the action is (conceivable as) "physical" and meaningful, arising from an action of a reflection symmetry group of space, so it's not like I just said "the set {The Beatles, my cat} is a Z/2-torsor", or "the set {Ø, |N} is a Z/2-torsor".

2026-03-27 20:14:38 UTC

- reply

so {e^- parity ,e^+ parity} is a torsor?

2026-03-26 23:16:25 UTC

- reply

Gives a new meaning to the phrase "a proper 1420" 🙃 More like the infamous "Parker square" due to @npub1g3p…jcyf , than the good mug of drink as in the original saying.

2026-03-18 05:07:16 UTC

- reply

The book "On the Sphere and Cylinder" was not lost, I believe. It was Archimedes' book "The Method" that was *only* saved in the palimpsest, others were known in translation possibly in fragmentary form.
But having the palimpsest version gives an additional Greek source for the text, perhaps older than other copies.

2026-03-18 05:03:11 UTC

- reply

"the manuscript changed hands several times before reaching its current owner" hides a lot of the 20thC history. The palimpsest was moved around a couple of times under unknown circumstances. Here's some history from Wikipedia from after the time the photos were taken:
"The manuscript was still in the Greek Orthodox Patriarchate of Jerusalem's library (the Metochion of the Holy Sepulchre) in Constantinople in 1920.[8] Shortly thereafter, during a turbulent period for the Greek community in Turkey that saw a Turkish victory in the Greco-Turkish War (1919–22) along with the Greek genocide and the forced population exchange between Greece and Turkey, the palimpsest disappeared from the Greek church's library in Istanbul.
Sometime between 1923 and 1930, the palimpsest was acquired by Marie Louis Sirieix, a "businessman and traveler to the Orient who lived in Paris."[8] Though Sirieix claimed to have bought the manuscript from a monk, who would not in any case have had the authority to sell it, Sirieix had no receipt or documentation for a sale of the valuable manuscript. Stored secretly for years by Sirieix in his cellar, the palimpsest suffered damage from water and mold.
[...]
Sirieix died in 1956, and, in 1970, his daughter began attempting quietly to sell the valuable manuscript. Unable to sell it privately, in 1998, she finally turned to Christie's to sell it in a public auction,..."
The missing page newly discovered was doctored possibly in 1942, according to the CNRS press release, which puts it in Paris during WW2. Needless to say, the situation may have not been straightforward.

2026-03-17 00:32:15 UTC

Would people buy an nLab-branded t-shirt?

2026-03-10 02:57:31 UTC

- reply

Ah, what a good feeling. I'm glad I'm not the only one for whom this never gets old
@npub1vlc…mzqq

2026-02-06 04:18:39 UTC

- reply

If you had to ask me, the key idea is that there are two ways of measuring the topology of certain higher-dimensional shapes described by polynomial equations: one that involves calculus (so something vaguely analogous to how electromagnetic fields are described), and one that only involves cutting out curved "slices" of the shape using more polynomial equations. The conjecture is that the latter, which can on the face of it measure at least *some* of the features that the calculus-based approach can see, actually captures everything.

2026-02-06 04:14:00 UTC

- reply

In Devlin's book The Millennium Problems he spends nearly three pages at the start of the chapter on the Hodge conjecture explaining that it's really difficult to explain. He says "To the layperson, the very inaccessibility of the problem is perhaps its most interesting feature." And he had a lot of experience explaining mathematics to the general public, but nonetheless resorted to something halfway between understandable and accurate: so more or less neither.

2026-02-05 10:01:29 UTC

- reply

I've been keeping half an eye on what he has been doing with his collaborators over the past decade or so, it's definitely fun to watch. Having a sketch of Connes' current ideas at the level of 19th century tech (sweeping a few things under the rug that got proved in the early 20th century) was a fascinating exercise. It reminds me of when arithmetic geometers give talks that give an on-ramp to a key modern conjecture in terms of some really concrete classical problem, like the congruent number problem leading to elliptic curves and then Tunnell's theorem, which is conditional on the BS-D conjecture.

2026-02-05 06:14:51 UTC

New from Alain Connes. His "Letter to Riemann" is fun.
https://arxiv.org/abs/2602.04022
He is honest about the fact we cannot prove the RH, and that it's not clear if his preferred approach will be the way, but also:
"Our discovery of a large class
of functions directly related to the Weil quadratic form and with zeros provably on the critical line, combined with the extraordinary numerical evidence linking truncated Euler products to the actual zeros of zeta, suggests that Riemann’s original insights may contain more power than previously realized. The accuracy achieved using only primes less than 13—with errors as small as \(2.6 \times 10^{-55}\)—cannot be dismissed as coincidence"
Apparent numerical coincidences in number theory like this at very minimum demand an explanation.

2026-01-22 00:11:50 UTC

- reply

I don't read the last one as 'quintessentially American'. The internet tells me the earliest reference recorded by the OED is rather old:
"The OED's earliest citation for touched as an adjective, meaning slightly crazy, is from John Humfrey's 1672 book, Authority of Magistrate of Religion: 'I count this man now as one in a Fever, that is, touched in his head, and who can help such a conceit?'"

2026-01-21 03:26:21 UTC

- reply

No, he's opening up his book by description of natural numbers and then different kinds of numbers that arise in solving problems (squares, 'roots' etc)
https://archive.org/details/algebraofmohamme00khuwuoft/page/5/mode/1up

2026-01-21 03:08:21 UTC

"When I consider what people generally want in calculating, I found that it always is a number."
—Abu Ja'far Muhammad ibn Musa Al-Khwarizmi, https://en.wikipedia.org/wiki/Al-Jabr ca.820
TIL this is the opening sentence of the book proper (after the author's preface).

2026-01-20 06:54:46 UTC

- reply

Good thing I'm keeping myself under that limit already....

2026-01-20 05:37:13 UTC

"I’ve decided to push back on the pressure to publish by making a rule for myself: I will no longer publish more than seven papers per year. I’ve published a median of 15 papers a year for the past five years, so that means I will have to halve my output. "
🤯
https://doi.org/10.1038/d41586-025-04061-w
(From a member of the research committee of Australia's NHMRC, the main medical research funding body here)

2025-12-16 00:01:23 UTC

- reply

There plenty more instances of **phrase** in the article. There's also really badly made TikZ figures that I don't think any author would be happy to leave as-is.

2025-12-15 21:52:44 UTC

- reply

ok, the paper has obvious signs of LLM-generation that the author hasn't proof-read and fixed. There's Markdown syntax mixed into the TeX in the image.... on MO this would be a fast close.
https://media.mathstodon.xyz/media_attachments/files/115/725/791/615/015/010/original/4e40df22ce7d3123.jpg

2025-12-15 21:39:36 UTC

- reply

I wrote originally "if you mean Huawei, then note that they also fund research into topos theory and do things like this: https://www.ihes.fr/en/laurent-lafforgue-en/"
But it's not. That's not to make any judgement call on the paper (or the company) either way.

2025-11-28 04:21:09 UTC

- reply

"one-hot" (sic) encoded, on top of using \(\aleph(i)\) for a set of nearest neighbours. Whyyyyyy. And that model. Wow. Very statistics.
The author being from https://www.college-cn.com/Anhui/1428/ doesn't fill me with confidence
cc [@infotainment.bsky.social](https://bsky.brid.gy/r/https://bsky.app/profile/infotainment.bsky.social )
https://media.mathstodon.xyz/media_attachments/files/115/625/373/708/714/817/original/8a32321b08948ac3.png
https://media.mathstodon.xyz/media_attachments/files/115/625/383/452/874/531/original/94a9ec272c5cfaf2.png

2025-11-28 00:42:50 UTC

- reply

I agree about the super-general and -abstract bit. The key point is that one needs a) compactness of the base (so really some kind of finiteness condition) and b) the ample invertible sheaf, which is really about admitting an open immersion to some scheme of the form Proj(S).

2025-11-27 21:17:12 UTC

- reply

The Brauer group can be defined in two ways: Azumaya algebra objects up to iso, where you say they are equivalent if there is a suitable bimodule, and Azumaya objects up to iso, where you mod out by the submonoid (under \(\otimes\) of things of the form \(End(E)\). It is the latter approach that this definition takes seriously.
I seem to recall that to go from the bimodule definition to the one I'm using you can take one of the vector bundles to be the underlying vector bundle of A, using the property \(End(A) \simeq A\otimes A^{opp}\) (and vice versa for \(B\), or else the underlying vector bundle of the bimodule. This was all 10+ years ago, but I know it works out.

2025-11-27 20:29:59 UTC

- reply

https://mathstodon.xyz/@highergeometer/115618036587381377 has the definition in the picture with two slides. This easily implies the bimodule definition (there's a left A and right B action on A \otimes End(E)...) The other direction is less obvious. In any case, even if the definitions imply each other, the data is different.

2025-11-27 13:14:04 UTC

- reply

"It actually looks the same to me: you talk about Morita equivalences of Azumaya bundles, but those should be the bimodules that are invertible (up to isomorphism)."
Aha, but I'm not using bimodules, and my 1-morphisms are not invertible even up to iso.
And the theorem about etale H^2 and Br is only about certain schemes: eg separated quasicompact schemes with ample line bundle

2025-11-27 11:53:09 UTC

- reply

Cool, I looked it up, and some of it sounded familiar. It was this thread, right: https://categorytheory.zulipchat.com/#narrow/channel/266967-deprecated.3A-mathematics/topic/Azumaya.20algebras/near/368619176 ?
And for anyone watching, the pause-point for the n-category café posts was here, in part 6 : https://golem.ph.utexas.edu/category/2023/10/grothendieckgaloisbrauer_theor_5.html
I think I'm being much more low-brow, and taking a lot of algebraic ingredients for granted. As noted I'm deliberately using a *different* categorification of the Brauer group to a symm. mon. (2,1)-category, rather than the usual one with bimodules at 1-arrows. It's for no other reason but that I know how to adapt Gabber's proof using this 2-category, and not the bimodule one!
I can imagine that the construction I give generalises a lot, for instance, as you say, to the real and the super vector bundle setting. And even more generally, I'm sure. It felt to me like some kind of double comma construction at the time, and I wrote it as \(\mathbb{E} \downarrow \mathbb{E}\), which was on my office board for a while.
I'd be happy to talk more about this or your ideas.

2025-11-27 01:03:10 UTC

- reply

A vector bundle where the fibres are Azymaya algebras. For complex vector bundles these are all kinda boring fibres (since they are all matrix algebras), but as a bundle they are nontrivial.

2025-11-26 23:26:34 UTC

- reply

I should add the bits in little font in the theorem were conjectural as I hadn't sat down and proved those facts. This was where I stalled, but also my postdoc at the time finished and other still-ongoing projects were prioritised.

2025-11-26 21:07:12 UTC

- reply

that was my original idea, about a decade ago, but sadly this project fell by the wayside. It very much feels like a double cat!
https://media.mathstodon.xyz/media_attachments/files/115/618/021/749/213/549/original/ebf8fc29779ff107.jpg
https://media.mathstodon.xyz/media_attachments/files/115/618/027/870/086/617/original/a2355cf553444605.jpg

2025-11-26 13:13:54 UTC

I had a thought that perhaps I should reframe my unpublished, descent-theoretic proof of the Grothendieck–Serre theorem (on the Brauer group of a compact 'nice' space) using symmetric monoidal \((\infty,2)\)-categories or something.

2025-11-26 02:08:55 UTC

- reply

Does the department you are in have an admin phone number? If forced you could add that.

2025-11-24 00:43:48 UTC

Riemann Hypothesis proof claimers: here's my proof bro, all the mathematicians didn't see it because it's simple and they are blinded by BiG mAtH. It only took 5 pages!!
Actual analytic number theorist who proved the weak Goldbach conjecture: here's an explicit estimate that uses zeroes of the zeta function...
https://media.mathstodon.xyz/media_attachments/files/115/601/887/390/133/278/original/21b4c9bd16d00538.png

2025-11-20 13:36:07 UTC

- reply

I think the result near the bottom here must be the result that Paseka–Vetterlein are using, though the proof is in the mystery 2003 paper (listed as [8] in the bibliography)...
Let me try cutting and pasting:
"Каждый оператор x ∈ B(H) представляется в виде конечной суммы x = ∑ x_k, где каждое x_k есть произведение не более чем двух проекторов при dim H = ∞ и не более чем трех проекторов при 2 ≤ dim H < ∞."
which machine translates to:
"Each operator x ∈ B(H) is represented as a finite sum x = ∑ x_k, where each x_k is a product of at most two projectors for dim H = ∞ and at most three projectors for 2 ≤ dim H < ∞"
Well, there you go. But we can't see the proof!
https://media.mathstodon.xyz/media_attachments/files/115/582/269/248/393/461/original/66e05ee037e0ba10.png

2025-11-20 13:28:26 UTC

- reply

Hmmm. That 2005 paper has this citation:
Бикчентаев А. М. О представлении линейных операторов в гильбертовом пространстве в виде конечных сумм произведений проекторов // Докл. РАН.. 2003. Т. 393, № 4. С. 444–447.
https://www.mathnet.ru/php/archive.phtml?wshow=paper&jrnid=dan&paperid=1659&option_lang=eng
which *is* the citation in the Paseka–Vetterlein paper. It's listed as being in MathSciNet, but not @npub1w8x…d8qh
Annoyingly, you can't get a pdf!
@npub1nf4…nqe4

2025-11-20 13:10:37 UTC

- reply

The closest paper I can find is this
A. M. Bikchentaev, “On representation of elements of a Von Neumann algebra in the form of finite sums of products of projections”, Sibirsk. Mat. Zh., 46:1 (2005), 32–45 mathnet mathscinet zmath; Siberian Math. J., 46:1 (2005), 24–34
https://www.mathnet.ru/eng/person18192
If any database would list this paper, it would be in math-net.ru
😠

2025-11-20 12:38:42 UTC

- reply

HMMMMMMM
I'm not sure this paper exists.....
A. M. Bikchentaev, "On the representation of linear operators in a Hilbert space as finite sums of products of projections" (in Russian), Dokl. Ross. Akad. Nauk 393 (2003), 444–447.
Consulting @npub1w8x…d8qh I can't see a paper in that journal in that volume with those page ranges...

2025-11-20 12:30:54 UTC

- reply

Yes, and they need to cite a non-obvious paper for that part!
A. M. Bikchentaev, On the representation of linear operators in a Hilbert space as finite sums of products of projections (in Russian), Dokl. Ross. Akad. Nauk 393 (2003), 444–447.
(this is not in zbMath with a few quick searches, and Google Scholar isn't helping. MathSciNet might have it, but I'm busy and not in the mood to look)

2025-11-20 09:55:12 UTC

- reply

Oh, and also this one!
C. Heunen, A. Kornell, N. van der Schaaf, Axioms for the category of Hilbert spaces and linear contractions, Bulletin of the London Mathematical Society 56 (2024) https://arxiv.org/abs/2211.02688

2025-11-20 09:52:28 UTC

RE: https://mastoxiv.page/@arXiv_mathCT_bot/115581087904495544
A different setup and conclusion compared to:
C. Heunen, A. Kornell, Axioms for the category of Hilbert spaces, Proc.
Nat. Acad. Sciences 119 (2022) https://arxiv.org/abs/2109.07418
M. Di Meglio, C. Heunen, Dagger Categories and the Complex Numbers: Axioms for the Category of Finite-Dimensional Hilbert Spaces and Linear Contractions, Applied Categorical Structures 33 (2025) https://arxiv.org/abs/2401.06584
and
S. Lack, S. Tobin, A characterisation for the category of Hilbert spaces,
Appl. Categor. Struct. 33 (2025) https://arxiv.org/abs/2412.03776
#note1emm…ch76

2025-11-16 23:34:08 UTC

- reply

I can't speak for @npub1nf4…nqe4 but here's a talk by a mathematical physicist that I know: https://researchers.ms.unimelb.edu.au/~dridout%40unimelb/seminars/120917.pdf

2025-11-12 11:18:07 UTC

- reply

Oh, it was definitely pronounced as "th" in Old English, the writing at the time using ð and þ shows that. The -s ending was a Northern dialect thing (from the Old Norse influence), and it was eventually adopted further south. Shakespeare mixed the use of -s and -eth, so matters were definitely shifting strongly by that point at the latest. I saw one person claim in a discussion that by the late 1600s the -s/-es ending had mostly won out, except for certain common words like hath and doth.

2025-11-12 09:04:03 UTC

"There is an important passage in Richard Hodges’ Special Help to Orthography (1643) that deserves quoting in full, along with an extract from his homophone lists:
(a) . . . wee use to write thus, leadeth it, maketh it . . . &c Yet in our ordinary speech . . . we say leads it, makes it . . . Therefore, whensoever eth, cometh in the end of any word, wee may pronounce it sometimes as s and sometimes like z, as in . . . bolteth it and boldeth it, which are commonly pronoun’ct, as if they were written thus, bolts it, bolds it . . .
(b) cox, coks, cocketh; clause, claweth, claws; courses, courseth, corpses; fleas, fleaeth, flayeth; Mr Knox, he knocketh, many knocks; reasons, reasoneth, raisins"
I'd love to track down a scan of this 17thC book, when I'm at my computer...

2025-11-04 09:53:21 UTC

- reply

probably not, which why I'm less confident. Maybe if I say something like the integral cohomology is torsion-free, maybe we get a projective Z[G]-module or whatnot.

2025-11-04 07:25:05 UTC

Hmm. I have a compact simple Lie group G, and some homogeneous space X for it. I need to know the G-invariants in H^k(X,Z). If one is calculating real cohomology then I could believe that because one can calculate de Rham cohomology using invariant forms (in this case), then the whole cohomology group is G-invariant. For integral cohomology it doesn't seem ludicrous, but I'd have to think about it.
Also I think I saw a claim that properties such as this group has imply taking G-invariants is an exact functor. I guess this follows from Schur's lemma...

2025-10-24 11:46:10 UTC

- reply

Oh, good to know

2025-10-24 07:10:30 UTC

Hey @npub1qrh…usg2 this link gives me an internal server error: https://arxiv.org/abs/alg-geom/9306001 (is a link a search engine gave me) but https://arxiv.org/abs/alg-geom/9306001v1 works.

2025-10-23 22:54:06 UTC

- reply

Thanks!
@npub1whr…qc2t

2025-10-23 22:53:14 UTC

- reply

Indeed:
"Is this LLM-generated? "
"I'm not sure if this is real, but the abstract says machine-verified."
"It seems a pretty serious attempt though- it’s not just some random crank paper."
"The fact that they have formalized the proof should mean it will be quicker to verify whether or not this is indeed it."
@npub1nf4…nqe4

2025-10-23 21:47:37 UTC

- reply

Oh dear, some people there don't seem to realise just how trash this is.

2025-10-23 20:16:29 UTC

- reply

if it was on the actual site it's gone now, as far as I can tell.

2025-10-22 23:50:05 UTC

- reply

These days, my mind is a bit split when I see the letter æ because for Old English it's the short a like in cat.
So my brain is suggesting at least three vowel sounds for the ash: short a, something like in 'day', and then the long e (as in the old encyclopædia).
@npub1whr…qc2t

2025-10-22 09:43:50 UTC

An image caption on a BBC article (about Philip Pullman and his comments on AI ingestion of authors' works):
"Dafne Keen, with her daemon Pan, played Lyra in the BBC's His Dark Materials series, which ran from 2019 to 2022"
I guess they do not really mean that Dafne Keen has a real-life daemon and together they played the character Lyra 🤣

2025-10-22 04:21:45 UTC

Oh dear. AI-generated Lean code and a document saturated with bullet-point lists
https://arxiv.org/abs/2510.17829
Don't bother. Cross-listed to math.CT, but should be in math.GM.

2025-10-20 23:02:48 UTC

- reply

Oh, I can appreciate that, but in a context discussing physics and electromagnetism, I was being carefully pedantic to not be conflating the mathematical object with the physical concept!

2025-10-20 22:29:16 UTC

- reply

The query on the entry to the right of "source" is tantalising. I'm sure Urs would have words to say for that lacuna! At the very least, it should be something like a de Rham current, I imagine.
@npub10mu…2h9c @npub18a6…szuj @npub1zz2…9mhx

2025-10-20 22:23:55 UTC

- reply

It's notation from the original Wu–Yang paper. Amusingly, it seems to only be in this table, in the rest of the paper they write only \(U_1\) to match \(SU_2\). I suspect it might have been a conflation of the notations \(U_1\) and \(U(1)\)!
https://media.mathstodon.xyz/media_attachments/files/115/408/819/568/263/640/original/6f2a6e61e88a47e6.png

2025-10-20 00:10:35 UTC

In honour of the passing of Chen-Ning (Frank) Yang, here is the "dictionary" that he published with Tai Tsun Wu, that is an absolute foundation stone for the area of mathematical physics that I work adjacent to:
https://en.wikipedia.org/wiki/Wu%E2%80%93Yang_dictionary

2025-10-15 05:58:44 UTC

- reply

Seems like a @npub12cu…d38p question

2025-10-13 23:21:21 UTC

- reply

Since it seems not to be true, what implication does it have for the particle physics ideas?

2025-10-09 09:34:52 UTC

- reply

90s anime fans will probably immediately think of this https://nz.pinterest.com/pin/336081190962419072/

2025-10-08 02:15:04 UTC

- reply

From the point of view of planar ternary rings, I can understand the maps of shape x |-> ax+b, but I'm not sure where the division comes from. Hmm. But checking the Wikipedia page for PTR, multiplication is a loop, so you have (left and) right inverses, so there is an element 1/(cx+d). OK.
The nLab page says a projective line is called Desarguesian if you have even just a restricted version of that 3-transitivity on points, whereby you have a group that fixes two points (possibly the coinciding) and can move any point distinct from those (that) to any other. The grammar of the definition there is not clear if it means this has to be true for one case, or always. And you definitely get weirder examples than this. I wonder if OP^1 is a 'Desarguesian line'?

2025-10-07 21:09:14 UTC

- reply

Regarding the projective plane/incidence geometry stuff, the condition analogous to Pappian and Desarguesian that gets you a Moufang plane (https://en.wikipedia.org/wiki/Moufang_plane, OP^2 is the prototypical example, I believe) is "little-Desarguesian". This holds iff the planar ternary ring is an "alternative division ring" (of which the octonions are an example).
Axiomatic/synthetic projective *lines* are less easy, and you get more exotic examples
https://ncatlab.org/nlab/show/projective+line#synthetic_projective_lines
analogous to how you get non-Desarguesian planes, but no non-"Desarguesian" projective n-space for n≥3.
I think Buekenhout's definition of projective line is instructive to consider for lines isomorphic to OP^1.
@npub1zz2…9mhx

2025-10-04 06:38:01 UTC

- reply

I have no expectations! The argument that one shouldn't expect such a bundle from the screenshot seems to me insufficient: it just points out the elements of OP^2 aren't literally O-lines, so no obvious analogue of the canonical bundle on say CP^2 falls out immediately. But this isn't a proof that no nontrivial O-line bundle exists on OP^2!
The obvious thing to try to start is classify rank‐4 complex vector bundles on OP^2, which surely is possible, say by obstruction theory using the known cell-structure of the Cayley plane.
@npub1nf4…nqe4

2025-10-03 23:42:06 UTC

- reply

on the nLab Urs pointed out a source (https://ncatlab.org/nlab/files/DrayManogue15-12-5.pdf#page=3) that says the literal elements of OP^2 are not octonionic lines, but with the recent discussion you had about octonionic things and Jordan algebras one might wonder if there is a better way to think about this.
Alternatively, one could ask if there is a map OP^2 -> BGL(2,H) that is not null-homotopic, hence classifying a nontrivial rank-2 quaternionic vector bundle, and if this has any structure that makes it "look like" an octonionic line bundle.
Of course, GL(1,O) doesn't really exist, so one needs to be creative....
@npub1zz2…9mhx
https://media.mathstodon.xyz/media_attachments/files/115/312/653/761/399/020/original/8ef091da66d1a6ba.png

2025-10-03 12:47:03 UTC

Is there a rank-8 real vector bundle (or rank 2 quaternionic vector bundle) on the octonionic projective plane, with any additional structure that makes it "octonionic"?
Cc @npub1nf4…nqe4 @npub1zz2…9mhx

2025-09-28 10:11:00 UTC

- reply

My main difficulties are that the mindset of the area is very different to mine; the techniques are a mix of custom-tailored examples; sometimes hard measure theory, sometimes immediately restricting to considering special cases like separable or étale groupoids, or something technical about the interior of the bundle of automorphism groups, when the theorems available are more general, because the final intended result only works for that one case.
There's not a really principled category theoretic treatment, outside big theorems where it's still only implicit (like the C*-algebra really depending on the underlying topological stack, if one considers a bicategory of C*-algebras and Morita equivalences)

2025-09-28 10:04:53 UTC

- reply

The term "reduced" is not about reducing the original C*-algebra, as you surmise, but it's because you do a completion of the more naive algebra of functions on the groupoid with a different norm to the one that gives the "full" C*-algebra of the groupoid.
I believe that the concept of Cartan subalgebra of a C*-algebra is what gives you a a locally compact topological groupoid, and such that the algebra is the full C*-algebra of that groupoid. And if you have a groupoid C*-algebra, then it has a Cartan subalgebra.
Such a subalgebra is a maximal abelian sub-C*-algebra with a retraction, satisfying properties. It's definitely possible to have many different Cartan subalgebras, including ones that give rise to very different groupoids.
It's not known how to tell if a given random C*-algebra has a Cartan subalgebra. I believe the known examples that fail to have a Cartan subalgebra are reduced C*-algebras of non-amenable loc.comp. topological groups.

2025-09-28 07:16:36 UTC

I had a random thought: given that at least some C*-algebras arise as that attached to a locally compact groupoid (which determines a topological stack), and the reduced C*-algebra is a quotient of that algebra, does that mean that it corresponds to something like a substack? And for those reduced C*-algebras that don't have a Cartan subalgebra (a necessary and sufficient condition to be a full C*-algebra of some loc.cpt groupoid), does this mean they are coming from something like a non-topological substack?

2025-09-27 12:55:25 UTC

- reply

Yes. And I found an answer by Derek Holt on math.SE, who is a senior finite group theorist, who also says ATLAS introduced the notation, so I think I can take it as a fact.

2025-09-27 12:52:24 UTC

- reply

I was under the impression this was called "ATLAS notation" in the literature. This author of MO question claims it is the source of the notation: https://mathoverflow.net/q/489747/
"The ATLAS of finite groups and representations thereof introduced the dot notation 𝐺=𝐴.𝐵 if 𝐴 can be embedded in 𝐺 such that 𝐺/𝐴=𝐵. Clearly 𝐴.𝐵 can mean many different non-isomorphic groups."
[@g_merzon](https://mathstodon.xyz/@g_merzon )

2025-09-27 10:13:42 UTC

- reply

well, to me the ATLAS notation that uses . and ⋅ and : for what look like binary constructions on groups, but which are highly underdetermined is even weirder (using such notation for an arbitrary, unspecified normal extension of groups??). But that's a different Conway conversation.
@npub1nf4…nqe4

2025-09-27 10:07:57 UTC

- reply

Looking just at the Conway–Smith proof of Theorem 14 there, I would have assumed it was a way of implicitly bracketing the product (to get around non-associativity, and needing to use lots of parentheses), and the snippet from the Russian edition seems to confirm that.

2025-09-24 08:15:21 UTC

https://travisdoesmath.github.io/s6/
funky visualisation of the nontrivial element in Out(S_6)

2025-09-24 02:47:59 UTC

@npub1255…9rve
https://web-cdn.bsky.app/profile/davidkbutler.bsky.social/post/3lyimmjpoxk2b
My honest answer:
It always made sense, from the moment I first saw it with @npub1nf4…nqe4 explaining it in a blog post. It was the coolest thing I had seen, to the point I physically printed out part of the webpage with the definition of category and functor (this was in 2001) and had it on hand to excitedly read it out loud to my friends while waiting outside the lecture theatre for Maths I to start.
What I didn't do, however (to the best I recall!), is get all smug about it and ask lecturers or tutors if such-and-such was a functor, as a way of showing off.
I realise I am atypical in these respects and I don't expect others to have the same experience.

2025-09-22 08:06:34 UTC

- reply

Peter Scholze imagined them, and he and Clausen can apparently give the first Chern class of the underlying (perfectoid) vector bundle for these Kähler differentials, in the so-called twistor-\(\mathbb{P}^1\) formalism that is analogous to working on the Fargues–Fontaine curve in the non-Archimedean setting, but they don't have a formalism in which to construct the actual object.

2025-09-22 06:09:37 UTC

"So, Dustin, are these Kähler differentials on \(\mathbb{Q}\) over \(\mathbb{F}_1\) in the room with us now?"
🙃
https://media.mathstodon.xyz/media_attachments/files/115/246/453/160/114/125/original/867d6a9894bb0178.png

2025-09-21 10:54:58 UTC

- reply

Strongly compact cardinals are listed here: https://neugierde.github.io/cantors-attic/Upper_attic and the page on them has lots of information, for people who can read serious set theory but haven't looked at them before.
Spoiler: they are waaaay above assuming there is a proper class of inaccessible cardinals (the Grothendieck–Tarski axiom of universes).

2025-09-21 10:48:06 UTC

- reply

If there were genuine higher-geometric content—in the style I study—lurking in the standard model I would be ecstatic. In the meantime, I guess BRST quantisation (which secretly, or not, uses \(L_\infty\)-algebras) might count as close enough to being useful for SM physics?
@npub1zz2…9mhx

2025-09-21 08:17:36 UTC

- reply

Given a map of semisimple Lie groups \(H \to G\), you can consider the \(\mathbb{Z}\)-linear map
\[ H^4(BG,\mathbb{Z}) \to H^4(BH,\mathbb{Z})\]
and that is what you define to be the index.
In the case that they are both simple and simply-connected, you get an injective map \(\mathbb{Z}\to \mathbb{Z}\). For \(G\) simple and \(H\) a semisimple subgroup (again simply-connected), you get a list of positive integers, indexed by the simple summands of the Lie algebra \(\mathfrak{h}\).
Because \(H^4(BU(1),\mathbb{Z})\simeq \mathbb{Z}\) you still get something meaningful even with \(U(1)\)-factors in \(H\).
That the subgroup of interest is quotient of \(SU(3)\times SU(2)\times U(1)\) (with \(H^4 \simeq \mathbb{Z}^3\) by \(\mathbb{Z}/6\) means there may well be some nonsense involving extra factors.

2025-09-20 09:46:53 UTC

- reply

Thanks for keeping me in the loop!
@npub1zz2…9mhx might be able to guess a question that interests me off the bat: what are the Dynkin indices for the (Lie algebras of the) stabliser subgroups for the various F_4 homogeneous spaces in this story? Particularly, what is the index of su(3) + su(2)+ u(1) < f_f? (this is no longer a positive integer, since the subalgebra is not simple....and it's not obvious what "index" should be when there are u(1) summands. But from a Lie group perspective, you can think about the restriction of the 2-group extension String_G to a U(1) < G, and then see what 2-group extension of U(1) (what Ganter calls a "categorical circle") you get)
@npub10tf…rt7p

2025-09-15 09:25:15 UTC

- reply

ah, ok, thanks for interpreting.
@npub10tf…rt7p @npub1zz2…9mhx

2025-09-15 09:14:24 UTC

- reply

I can only guess the so(9) bivectors is somehow the tensor square of some so(9) representation, or perhaps the tensor square of so(9). I don't know what that works out to be...
I think it would help me if the answer were less telegraphic and didn't assume standard physics abuses of notation/terminology in representation theory. Also, writing not with unicode will help, because I suspect some symbols have been mangled.
@npub10tf…rt7p @npub1zz2…9mhx

2025-09-12 00:18:55 UTC

- reply

Ah, was it this talk?
https://mathstodon.xyz/@johncarlosbaez/115174390161940101