Indexed text attachment
differential-topology-in-higher-categories.tex
Source for Differential Topology in Higher Categories
View paper · Download original file
Lines 1599–1698 of 6119. Source text is displayed literally.
1599\end{proposition}1600 1601\begin{proof}1602After shrinking, form the transition \(g=e_0^{-1}e_1\)1603near the zero section of \(\nu_i\). It fixes the zero section,1604and its normal derivative is the identity. For \(s>0\) let1605\(g_s=\delta_{1/s}\circ g\circ\delta_s\), where \(\delta_s\)1606scales the fibers by \(s\). In a local trivialization1607\(g\) has components \((b(x,v),w(x,v))\) and \(w(x,0)=0\).1608The base component of \(g_s\) tends to \(x\), and its fiber1609component has the formula1610\[1611 \frac{w(x,sv)}{s}1612 =\int_0^1 D_vw(x,\tau s v)\,v\,d\tau .1613\]1614The equality supplies a smooth extension at \(s=0\),1615with fiber component \(v\) by normalization. It also shows1616that sufficiently small vectors, uniformly over the compact1617zero section and \(s\in[0,1]\), have images in the domain1618of \(e_0\). For positive \(s\) the transition is an embedding1619on this smaller domain, by its definition as a composition1620of embeddings. At zero it is the identity. Its derivative1621on the quotient remains the identity, and \(e_0\circ g_s\)1622gives the required family. On a region where the transition1623is the identity as a germ, every rescaled transition is1624the identity as a germ there.1625\end{proof}1626 1627The formula uses compatible transition maps of the normal1628bundle, so the extension shown in each trivialization glues1629globally. This conclusion concerns tubular germs. An isotopy1630on a prescribed radius without shrinkage requires control1631away from the zero section. There is also a precise base1632change statement for the quotient normal bundle. If1633\(f:X\to M\) is transverse to \(A\) and \(B=f^{-1}(A)\), then1634\(df\) followed by projection to \(\nu_i\) is onto along \(B\)1635with kernel \(TB\). Consequently1636\[1637 TX|_B/TB\ \cong\ (f|_B)^*\nu_i .1638\]1639This identification is natural for repeated transverse base1640change. An exponential map uses additional geometric data. Comparing1641two chosen exponentials requires a corresponding comparison1642of these choices.1643 1644\section{Collars and neat intersections}1645 1646A collar of a manifold with boundary \(M\) is an embedding1647\(c:\partial M\times[0,\epsilon)\to M\) onto a neighborhood of1648the boundary with \(c(x,0)=x\). For compact boundary such a1649collar is obtained by choosing a smooth vector field near it1650pointing strictly inward there. Use inward coordinate vectors1651in local boundary charts, and glue with a partition of unity.1652Positivity of the derivative of an inward boundary coordinate1653is invariant under changing the defining coordinate, and is1654preserved by positive sums of vector fields. Thus the glued1655field remains inward. Its positive-time flow on the1656boundary is defined for a uniform short interval by compactness.1657At time zero its derivative has image the sum of the boundary1658tangent space and the transverse inward direction. The local1659inverse function theorem in half-space charts gives a collar1660near each point. The same compactness argument as for tubular1661neighborhoods gives a common interval on which this flow is1662injective and has smooth inverse onto its image.1663 1664For a compact manifold \(M\) with boundary, suppose the1665compact submanifold \(A\) is \emph{neat}. This means1666\(\partial A=A\cap\partial M\) and near each of those points1667there are half-space charts with coordinates \((u,v,t)\),1668\(t\geq0\), in which \(A\) is defined by \(v=0\).1669One can construct a collar of \(M\) whose restriction to1670\(\partial A\) is a collar of \(A\). In the neat charts the1671vector \(\partial_t\) is tangent to \(A\) along \(A\).1672Choose all local inward fields meeting \(\partial A\)1673in this form, and choose fields on the remaining patches1674disjoint from \(A\) near those patches of the boundary.1675A partition of unity preserves tangency to \(A\) and inwardness.1676Uniqueness of solutions of the flow equation then gives1677the compatible collar near \(\partial A\).1678 1679Use product metrics on a smaller common collar and double1680\(M\) and \(A\) across their boundaries. The double is1681smooth in the signed collar coordinate, and the doubles form1682an embedded pair without boundary. Apply the normal1683exponential construction to this pair. On the product collar,1684the normal vectors to \(A\) have zero component in the collar1685direction. Their short normal geodesics have constant collar1686coordinate. Away from a smaller collar the feet of the1687vectors are a compact set separated from the boundary.1688After shrinking further, their geodesics remain in the1689original half of the double. The resulting restriction is1690a tubular neighborhood for the original neat pair, compatible1691with its boundary product structure \cite[Chapter 4]{hirsch}.1692The product condition avoids treating the inward half-line1693as an unrestricted normal direction.1694 1695\begin{example}[Collapse to a Thom space]1696Let \(A\) be compact inside a compact manifold \(M\) without1697boundary, and choose a closed radius-\(r\) disk bundle1698\(D_r(\nu_i)\) inside a tubular domain. The Thom space here