Category Theory
Zulip Server
Archive

You're reading the public-facing archive of the Category Theory Zulip server.
To join the server you need an invite. Anybody can get an invite by contacting Matteo Capucci at name dot surname at gmail dot com.
For all things related to this archive refer to the same person.

Stream: learning: reading & references

Topic: Mastering Bott Periodicity

Andrius Kulikauskas (Apr 10 2024 at 10:41):

In this thread, I want to work on mastering Bott periodicity along with others who may be interested, and certainly with the help of those who can teach us. I have made a video about my own interests, which are metaphysical, and have written out the transcript with slides: Bott Periodicity Models Consciousness? Preliminary Exploration.

Bott Periodicity is related to category theory on many levels and in many variants so I hope this is a fine place to work on this! Please jump in with your related interests, comments, insights, questions, suggestions, corrections, ideas...

Andrius Kulikauskas (Apr 10 2024 at 10:52):

I myself want to have intuition on how real Bott periodicity functions as a clock with periodicity 8. My PhD is in algebraic combinatorics so I am most inclined to Clifford algebra calculations. I would like to also understand in terms of Lie theory and symmetric spaces. For me, the homotopy theory and K-theory are optional, but I have studied some of Hatcher's Algebraic Topology and ultimately that would surely add insight.

Metaphysically, my goal is to understand CPT symmetry in this context. Charge conjugation, parity, time reversal are metaphysical concepts that apparently pop out of the math. I think they relate to what I call "divisions of everything", the twosome, threesome, foursome, which I talk about in my presentation Time and Space as Representations of Decision-Making. I talk about an eight-cycle of divisions of everything, proceeding from the nullsome to the sevensome, with the eightsome (the logical square with four corners and four sides) collapsing into the nullsome (for example, if all are known and all are unknown, then the system is empty).

Charge conjugation has to do with (not) distinguishing particles (what is) and holes (what is not), thus existence. Parity has to do with participation, whether our rules apply in a mirror world, and in particular, whether a learning three-cycle (taking a stand, following through, reflecting) functions in the opposite direction. Time reversal has to do with knowledge, whether there is change, whereby a concept like How makes sense. You may disagree but all the better, as this is on my mind, and your views may surely help.

Andrius Kulikauskas (Apr 10 2024 at 11:07):

My current plan of study is:

1) Understand how to calculate the Clifford algebras $C_{0,k}$ (with generators squaring to $-1$) as matrix algebras. I think I need to use the homomorphisms for the recurrence relations.

• $\textrm{Cl}_{p+2,q}(\mathbf{R}) = \mathrm{M}_2(\mathbf{R})\otimes \textrm{Cl}_{q,p}(\mathbf{R})$
• $\textrm{Cl}_{p+1,q+1}(\mathbf{R}) = \mathrm{M}_2(\mathbf{R})\otimes \textrm{Cl}_{p,q}(\mathbf{R})$
• $\textrm{Cl}_{p,q+2}(\mathbf{R}) = \mathbf{H}\otimes \textrm{Cl}_{q,p}(\mathbf{R})$

2) Understand what are the representations (and also the $\mathbb{Z}_2$-graded representations) of the Clifford algebras. Calculate the groups $A_k=M(C_{0,k})/i^*(M(C_{0,k+1}))$ as in Dale Husemoller's book Fibre Bundles and Bott, Attiyah, Shapiro's paper Clifford Modules.

3) Do the calculations of the Lie group embedding described in the very helpful paper Michael Stone, Ching-Kai Chiu, Abhishek Roy. Symmetries, Dimensions, and Topological Insulators: the mechanism behind the face of the Bott clock. which is like a tutorial for physicists.

4) Understand in that paper how they talk about CPT symmetry and how it relates to Bott periodicity.

5) Understand how CPT symmetry relates to random matrix ensembles.

6) From there it would be great to study @John Baez 's video on Symmetric Spaces in terms of a free-forgetful adjunction.

7) I also want to understand the loop space - suspension adjunction.

Andrius Kulikauskas (Apr 13 2024 at 19:50):

I am working on my Step 1) above. I am studying José Figueroa-O'Farrill's lecture notes for his Projective Geometry course on Spin Geometry. His Lecture 2 is about Clifford algebras: classification. In Theorem 2.3, he states the three recurrence relations and he describes the homomorphism with which he proves $CL(n+2,0)\cong Cl(0,n)\otimes Cl(2,0)$ and then I will need to work out the other two myself. His notation is that the left entry counts the generators that square to $-1$, and the right entry counts the generators that square to $+1$. Just now I am realizing that here he is using the ungraded tensor product, and I think I could figure out what he is doing with that. Whereas at the beginning of the section he introduced the $\mathbb{Z}_2$-graded tensor product, which I am wrestling with, and which will be important later. That is why today I thought I should I start with his Lecture 1 on Clifford algebras: basic notions.

Andrius Kulikauskas (Apr 13 2024 at 20:03):

Lecture 1 was helpful for me to work out some basic points. In the relationship between the symmetric bilinear form and the quadratic form, I came to realize that the "symmetric" part is essential, so that $\frac{1}{2}(B(x,y)+B(y,x))=B(x,y)$. I wrote out some of the diagrams to work out that, for example, why the category of quadratic vector spaces has an initial object (zero vector space with zero quadratic form) but not a terminal object (the unique map needs to respect norms! so the terminal object can't be too small, and if it's too big, then multiple maps will be possible). Nor are there products (I think because you can try to pair generators but then you can't pair their scalars, unfortunately).

Andrius Kulikauskas (Apr 13 2024 at 20:25):

Meanwhile, I watched a couple of very good video lectures by Greg Moore, Quantum Symmetries and the 10-Fold Ways, Part 1 and Part 2. He is very concrete which helps me greatly because I am an algebraic combinatorialist. I was glad to see his concrete form for Morita equivalence, which I will need to find. I will surely relisten to the lectures with my full attention. In Part 1, he mentioned Wigner's theorem and it's connection to CPT symmetry, which is crucial for me. And it means that this is all connected to the foundations of quantum mechanics, precisely what my friend John Harland is teaching me, so that I could discuss with him his research program, which we talked about in this video, Extradynamical Evolutionary Foundation for Physics.

Andrius Kulikauskas (Apr 15 2024 at 18:39):

I watched a couple of videos by Tobias Osborne about Wigner's theorem. He talks about it in the first video of his series on Symmetries and Quantum Mechanics And he gives a proof of it in the last video of his series on Advanced Quantum Theory The point for me is that, vaguely speaking, symmetric transformations of ray space are either linear unitary operators or antilinear antiunitary operators. So that gives some context for CPT symmetry where charge conjugation C is unitary, and parity P is unitary, and time reversal T is antiunitary.

Andrius Kulikauskas (Apr 19 2024 at 21:43):

I have been learning how to express Clifford algebras as matrix algebras. I have a PhD in Math, but still, I find it hard going, taking hours to figure out things that are surely trivial to others. I envy those who can just walk down a corridor and ask a colleague for a quick explanation. Which is part of the reason I write here.

Some of the challenges are notation. For example, in José Figueroa-O'Farrill's notes, in his Chapter 2, $Cl(1,0)=\mathbb{C}$, which is to say, the left generators square to $-1$ and the right generators square to $+1$. But in his Chapter 1, when he writes $n=r+s+t$, it is the s that counts the generators that square to $+1$, and only then comes the t that counts the generators that square to $-1$, and so he writes of the Clifford algebra $Cl(s,t)$, which is the opposite as in Chapter 2. In Theorem 2.3 on the recursion relations, he writes of a quadratic vector space that is isomorphic to $\mathbb{R}^{s,t}$. I got stumped and after many hours I started to realize, for example, the subtle distinctions between the quadratic vector space and the Clifford algebra we seek. Or the input variables and the output variables in the homomorphism. I was ultimately stumped by his expression $-Q(v)$.

Andrius Kulikauskas (Apr 19 2024 at 21:46):

After several hours I switched to Husemoller's book and his Theorem 5.6 in Chapter 12. He clearly distinguishes between the $e_1,\dots e_k$ that generate $C_k$ and square to $-1$ and the $e'_1,\dots e'_k$ that generate $C'_k$ and square to $+1$. It all made much more sense.

Andrius Kulikauskas (Apr 19 2024 at 21:52):

Then I read Husemoller's Proposition 6.4 because I am trying to understand the irreducible modules over $F(n)$ and $F(n)\oplus F(n)$. I think in terms of matrices and representations (of groups!) and am just starting to think in terms of algebra representations. I have never got into modules and ring theory. I became confused what $F(n)$ meant, whether a column or a matrix. Then I read on Math Stack Exchange about simple $M_n(D)$ modules. The third answer helped me. I was confused to think that each column is a module. But finally it dawned on me that they were all isomorphic, and from that point of view, there was only one irreducible module.

Andrius Kulikauskas (Apr 19 2024 at 22:08):

Another thing that was curious to learn from Wikipedia was that direct sum means different things for groups, for rings, for algebras, and so I had to stop and wonder what it meant in my case, for example, $\mathbb{R}\oplus\mathbb{R}$ or $\mathbb{H}\oplus\mathbb{H}$. In particular, the article on Direct Sum of Modules explained that in the case of algebras such as $\mathbb{R}\oplus\mathbb{R}$, a direct sum means what category theory calls a product.

Andrius Kulikauskas (Apr 19 2024 at 22:13):

Since then I have been working to variously understand $\mathbb{R}\oplus\mathbb{R}$ as relevant for me. Here are my notes on these split-complex numbers. After many hours, I am sorting out the relationship between $Cl_{1,0}\cong \mathbb{R}\oplus\mathbb{R}$, where the Clifford algebra has identity $1$ and the direct sum has identity $(1,1)$. For quite some time I was wondering why the Clifford algebra yielded one isomorphic module, whereas the direct sum yielded two modules for either half. Then I realized that the one must not be simple, it must break down into two, like those two. These are the simple things I think about!

But I have figured out how to write out their two irreducible modules and how they come together for one irreducible modules. I will write that out and make a Math4Wisdom video.

Andrius Kulikauskas (Apr 19 2024 at 22:15):

My question, if anybody would help me here, is how do I know what the simple modules are, in this case? What is the argument or technique to establish that? I suppose it is straightforward and I may well study the module theory from scratch. But I welcome your help to save me these hours, at least for now!

John Baez (Apr 20 2024 at 18:13):

The two simple modules of $\mathbb{R} \oplus \mathbb{R}$ are these.

1) In the first, $\mathbb{R} \oplus \mathbb{R}$ acts on $\mathbb{R}$ as follows:

$(a,b)(c) = ac$

2) In the second, $\mathbb{R} \oplus \mathbb{R}$ acts on $\mathbb{R}$ as follows:

$(a,b)(c) = bc$

Andrius Kulikauskas (Apr 20 2024 at 20:01):

@John_Baez Thank you! I appreciate this very much.

This shows existence. Now I will think how to show that there are no other simple modules M. I have an idea how to do that by using these two simple modules, and creating maps with them, and showing that any other module, if it was different, would have such a submodule, thus not be simple.

I am making progress on my notes for my video on Split complex numbers.

I still want to think in what sense this all carries over for $\mathbb{H}\oplus\mathbb{H}$. And then I can think about the $\mathbb{Z}_2$ graded representations and try to calculate the groups $0$ and $\mathbb{Z}$ that show up in Bott periodicity. That is why I care about this $\mathbb{R}\oplus\mathbb{R}$. Likewise, I can think how this carries over to $M_8(\mathbb{R})\oplus M_8(\mathbb{R})$.

Andrius Kulikauskas (May 07 2024 at 15:57):

I am growing in my understanding of Bott periodicity in its various forms. I am especially interested in the embeddings of Lie groups

$O(16)\supset U(8)\supset Sp(4)\supset Sp(2)\times Sp(2)\supset Sp(2)\supset U(2)\supset O(2)\supset O(1)\times O(1)\supset O(1)$

as detailed in the paper by Michael Stone, Ching-Kai Chiu, Abhishek Roy. Symmetries, Dimensions, and Topological Insulators: the mechanism behind the face of the Bott clock. They make it very concrete.

What impresses me is how they describe this embedding by applying, one by one, any set of mutually anticommuting orthogonal complex structures $J_1, J_2, J_3,\dots$ This gave me the idea that I've been looking for to explain how Bott periodicity might model the eight-cycle of conceptual frameworks which I have documented, however tentatively.

My idea is to consider how to go back up from an embedded Lie group to the starting point, $O(16)$. How does one assemble the orthogonal complex structures to fill that gap? Each orthogonal complex structure would be a perspective and their assembly would be the cognitive framework which describes the available shifts in perspective. I think of these frameworks as "divisions of everything". The mutual anticommuting I think implies, in a sense, that there is no overlap, just as there isn't in a framework where the perspectives divide up everything.

The orthogonal complex structures are symmetry-breakings and so considering the reverse direction is removing structure and restoring symmetry. Perhaps that can be phrased as an adjunction.

I am now investigating this concretely. To go from $U(8)$ to $O(16)$ there is basically only one way and it is trivial. So that is what I call the onesome, the division of everything into a single perspective.

Next, to go from $Sp(4)$ to $O(16)$, then we take two steps, adding an orthogonal complex structure and then another. I suspect that is straightforward enough. And it would match what I call the twosome, the division of everything into two perspectives, where one perspective ("opposites coexist" as in free will) is followed by another perspective ("all is the same" as with fate).

What I am hoping is that it gets interesting going from $Sp(2)\times Sp(2)$ to $O(16)$. Suppose there were three ways to proceed, I imagine, starting with the left $Sp(2)$ or the right one or perhaps a diagonal one. Then suppose that all three ended up getting used, A, B, C, and suppose it went in a particular order, A to B to C to A. Then I would get the learning three-cycle ("take a stand", "follow through", "reflect") that I am looking for, where you can start at any point and then proceed cyclically. Similarly, I have this fantastic hope that in the further cases, the framework I seek to model do assemble themselves as we restore the original symmetry of $O(16)$ or $O(16r)$ or $O(\infty)$ if you prefer.

This may well fail but I am inspired to think that I am arriving at the mathematical concepts with which I might best model what I am trying to say about perspectives and frameworks. And if what I am saying is coherent, then there will be a mathematical model. Or otherwise I may understand, where are my observations failing me, as regards these frameworks.