Indexed text attachment
differential-topology-in-higher-categories.tex
Source for Differential Topology in Higher Categories
View paper · Download original file
Lines 361–460 of 6119. Source text is displayed literally.
361pulling back the germ from $N$ to $M$. Composition of pullbacks362gives $D_x(gf)=D_{f(x)}g\circ D_xf$. For the identity the363derivative is the identity. Under coordinate changes the364derivation coordinates transform by the Jacobian matrices, so365gluing $U\times\mathbb R^n$ gives the smooth tangent bundle $TM$366and a smooth map $Tf:TM\to TN$. In bundle notation its defining367square is368\[369 \begin{tikzcd}[column sep=large]370 TM \arrow[r,"Tf"]\arrow[d] & TN\arrow[d]\\371 M \arrow[r,"f"'] & N.372 \end{tikzcd}373\]374This construction preserves products, allowing corner charts if375necessary. A product chart separates the two lists of variables,376giving the canonical identification $T(M\times N)\cong TM\times TN$.377For example, the tangent map of multiplication378$m:\mathbb R^2\to\mathbb R$ satisfies379$D_{(x,y)}m(v,w)=yv+xw$. This formula follows as well by multiplying380$x+\epsilon v$ and $y+\epsilon w$ in the dual numbers.381 382\section{Rank and normal forms}383 384An immersion is a smooth map with injective derivative at every385point. A submersion has surjective derivative. We use the386constant-rank theorem for boundaryless coordinate neighborhoods in387the following form \cite[Ch.~1]{hirsch}.388 389\begin{proposition}[Local normal form]390Suppose $f$ is a smooth map between boundaryless manifolds, of391constant rank $r$ near $x$. There are source and target coordinates392in which $f(u,v)=(u,0)$, with $u\in\mathbb R^r$.393\end{proposition}394 395\begin{proof}396Work near the origin and choose coordinates so the first $r$397components have an invertible minor in the first $r$ variables.398Use $(f_1,\ldots,f_r,t_{r+1},\ldots,t_m)$ as a change of source399coordinates, by the inverse function theorem. The map is now400$(u,v)\mapsto(u,h(u,v))$. Since the derivative has rank exactly401$r$, its columns in the $v$ variables, whose first $r$ entries402vanish, must themselves vanish. Restricting to a coordinate box403shows that $h(u,v)=h(u,0)$ by integration along the $v$-segments.404Subtracting $h(u,0)$ from the last target coordinates gives the405stated form. Rank zero is the same argument with an empty first406minor.407\end{proof}408 409Full rank is open, so the hypothesis applies near every point of410an immersion or a submersion in the interior. For an immersion the411form is a coordinate inclusion. For a submersion it is a coordinate412projection. Boundary versions require hypotheses on the boundary413behavior. For instance $t\mapsto t^2$ on the half-line has a414critical boundary point even though it is a submersion on the415interior. Requiring a boundary-preserving map between half-spaces416does not by itself ensure a smooth inverse up to the boundary.417 418An embedding is an immersion that is a homeomorphism onto its419image with the subspace topology. The irrational-slope map420$\mathbb R\to\mathbb R^2/\mathbb Z^2$ given by421$t\mapsto (t,\alpha t)\bmod\mathbb Z^2$ illustrates the second422condition. For irrational $\alpha$ it is an injective immersion.423There are unbounded integers $n_j$ with $\alpha n_j$ tending to424zero modulo $\mathbb Z$. Indeed the pigeonhole principle applied425to the fractional parts of $0,\alpha,\ldots,N\alpha$ gives a426nonzero integer of size at most $N$ whose phase is within $1/N$427of zero. A fixed finite set of nonzero integers cannot do this428as $N$ grows. The images of $n_j$ tend to the image of $0$, so429the inverse on the image is discontinuous.430 431A proper injective immersion between the manifolds here is an432embedding. To check that it is closed, take $y$ in the closure433of the image of a closed set. In the first-countable target434there is a sequence $y_j\to y$ in that image. The sequence435together with its limit is compact, so the preimages of its436terms lie in a compact set by properness. A convergent subsequence437of the preimages has limit in the closed set and maps to $y$.438Thus the immersion is closed, and injectivity makes the inverse439on its image continuous. For an embedded $S\subset M$, the quotient440bundle $TM|_S/TS$ is its normal bundle. A metric identifies it441with the orthogonal complement, whereas the quotient comes442directly from the embedding.443 444At the boundary a useful class is given by \emph{neat embeddings}.445In this chapter this means an embedding $f:M\to N$ with446$f^{-1}(\partial N)=\partial M$ and with derivative transverse447to $\partial N$ at each point of $\partial M$. The boundary448normal form is a coordinate inclusion preserving the first449coordinate. In fact, let $t$ and $s$ denote source and target450boundary coordinates, with $t,s\ge0$. In a neighborhood of a451boundary point the first component of $f$ is $t b(t,u)$ for a452smooth $b$ by integrating its derivative in $t$, since the453component vanishes at $t=0$. Transversality and nonnegativity give454$b>0$ after shrinking. The other target coordinates can be455linearly recombined along $\partial N$ so that the first456$\dim M$ components, including $s$, have an invertible derivative.457They become source coordinates by the inverse function theorem458on smooth extensions. The sign of $t b(t,u)$ ensures this459coordinate change respects the half-spaces. Subtracting the graph460of the remaining components gives the coordinate inclusion by a