Independent scholarship · Open accessThe monthly digest

Indexed text attachment

algebraic-and-differential-topology-in-higher-categories.tex

Source for Algebraic and Differential Topology in Higher Categories

View paper · Download original file

Lines 9561–9660 of 10051. Source text is displayed literally.

9561\begin{tabular}{crrrrrrrrr}9562\toprule9563$k$ & $0$ & $1$ & $2$ & $3$ & $4$ & $5$ & $6$ & $7$ & $8$\\9564\midrule9565$\pis_k$ & $\Z$ & $\Z/2$ & $\Z/2$ & $\Z/24$ & $0$ & $0$ & $\Z/2$ & $\Z/240$ & $(\Z/2)^2$\\9566generator & $\iota$ & $\eta$ & $\eta^2$ & $\nu$ & & & $\nu^2$ & $\sigma$ & $\eta\sigma,\varepsilon$\\9567\bottomrule9568\end{tabular}9569\end{center}9570 9571\noindent Here $\eta,\nu,\sigma$ are the stable classes of the three Hopf maps of \Cref{rem:hopfone}. The source9572proves the columns $k=0,1,2$ (\Cref{thm:hopfdeg,thm:first,thm:second}) and reports that Rokhlin obtained $k=3$ by9573the same method, while ``for $k>3$ the problem so far defies solution by our method'' \src{II, Introduction}.9574The remaining columns, and several dozen more, were found by the algebraic methods the source attributes to the9575French school \cite{Ser51,Tod62,Rav86}: Serre proved that all the groups with $k>0$ are finite, and the Adams9576spectral sequence computes them from the cohomology of the Steenrod algebra.9577 9578Two features of the table were already visible to the geometric method. The groups in degrees $3$ and $7$ are9579the images of $J$ (\Cref{thm:J}), of orders $24$ and $240$: framed spheres. The classes $\eta^2$ and $\nu^2$ in9580degrees $2$ and $6$ are not in the image of $J$; they are products of spheres with product frames, and they are9581detected by the Arf and Kervaire invariants (\Cref{rem:kervaire}).9582 9583\section{What became of the method}9584 9585The main method computes homotopy groups from manifolds. It stalled at $k=3$ because the classification of9586framed manifolds up to cobordism requires, in each dimension, a way to simplify an arbitrary manifold, and9587beyond surfaces the obstructions to simplification are as hard as the groups being sought. Fifteen years later9588the direction of the argument was reversed. Kervaire and Milnor \cite{KeMi63} applied surgery to framed9589manifolds whose boundary is a homotopy sphere, took the stable stems as known from algebra, and deduced the9590classification of smooth structures on spheres: the group of exotic spheres in dimension $k$ is, up to the9591image of $J$ and the Kervaire invariant, the stable stem $\pis_k$. The dictionary of \Cref{thm:pontryagin} has9592been read from right to left ever since.9593 9594In the other direction, the statement that framed manifolds are the homology theory of the initial ring9595has been extended from groups to categories: the cobordism hypothesis (\Cref{rem:cobhyp}) is9596\Cref{thm:sphere} with manifolds with corners in place of closed manifolds and duality data in place of9597inverses.9598 9599\section*{Notes}9600For spectra and the smash product see \cite{EKMM97,Lur17}; for the stable stems, \cite{Rav86}. The values in the9601table and in \Cref{fig:stems} are quoted from the literature and not computed here; the orders of the image of9602$J$ are computed from Bernoulli numbers. \chk{R5} Rokhlin's computation of $\pis_3$ is the origin of his theorem9603on the signature of four-manifolds.9604 9605\section*{Exercises}9606\begin{exercise}\label{exer:hur}9607Show that the map $\pis_k(X_+)\to H_k(X)$ induced by $\sS\to\HZ$ sends a framed manifold $g\colon M\to X$ to9608$g_*[M]$, and that it is an isomorphism for $k=0$.9609\end{exercise}9610\begin{exercise}\label{exer:eta2}9611It is known that $\eta^3\ne0$. Using the table, show that $\eta^3=12\nu$.9612\end{exercise}9613\begin{exercise}\label{exer:defect}9614Prove that the angle defects of a closed polyhedral surface with triangular faces add up to $2\pi\chi$.9615\end{exercise}9616\begin{exercise}\label{exer:J}9617Which entries of the table for $k\le8$ are entirely in the image of $J$, according to \Cref{thm:J}?9618\end{exercise}9619%% ---- end ch/12-stems.tex ----9620%% ---- begin ch/13-coda.tex ----9621\chapter{Assessment}\label{ch:coda}9622 9623\section{What the language did}9624 9625\emph{It identified the two monographs as one construction.} Chains are the free $\Z$-module on a space9626(\Cref{thm:leftadjoint}); the suspension spectrum is the free $\sS$-module (\Cref{thm:sphere}); the first is9627obtained from the second by extension of scalars (\Cref{thm:tworings}). Simplicial cycles and framed manifolds9628are the cycles of the two theories, and the table in \Cref{ch:stems} lists the parallel, row by row.9629 9630\emph{It said why the first monograph is easy and the second is hard.} The coefficients of the first theory,9631the homotopy groups of $\HZ$, are $\Z$ in degree zero. Those of the second are the stable stems. Everything the9632source proves in Part One is linear algebra over a ring whose homotopy is trivial; everything in Part Two is9633the first three homotopy groups of a ring whose homotopy is not.9634 9635\emph{It sorted the proof of the fixed point theorem into functors.} The fixed point index is a trace in9636spectra, the Lefschetz number a trace in $\DD(\Z)$, and they agree because a symmetric monoidal functor preserves9637traces and the two units have the same $\pi_0$ (\Cref{thm:doldpuppe}).9638 9639\emph{It explained an error.} The quadratic form on a framed surface fails to be linear because the sphere9640spectrum is not a product of Eilenberg--MacLane spectra (\Cref{rem:kinvariant}). The announcement of 19389641amounts to the assumption that it is.9642 9643\emph{It placed the constructions of Part One.} The polyhedron is a colimit, subdivision a functor with a right9644adjoint, the cone a contraction, the prism a product with an interval, and invariance the statement that9645chains extend to a functor on homotopy types (\Cref{prop:colim,prop:sd,prop:cone-op,thm:prism,thm:infty}).9646 9647\section{What it did not do}9648 9649\emph{It proved nothing.} Every theorem here is in the source or in the literature. The proofs given are those9650that are short; the others are indicated by their structure. In particular the general position arguments of9651\Cref{ch:position}, on which the whole of Part Two rests, and the analysis of cobordisms of surfaces in9652\Cref{ch:arf}, are the source's and are not reproduced.9653 9654\emph{It did not compute a stable stem.} The higher-categorical description of $\sS$ as a free object is not a9655method of calculation. The values in \Cref{fig:stems} come from the Adams spectral sequence and its successors.9656The script accompanying this volume checks the invariants that detect the first three stems, on examples; it9657does not and could not verify that they are complete.9658 9659\emph{It did not replace the geometry.} That a framed circle in $\R^4$ has an invariant in $\Z/2$ is, in the9660language of this volume, the statement $\pis_1=\Z/2$; one still has to see why, and the source's reason, a loop