Independent scholarship · Open accessThe monthly digest

Indexed text attachment

connections-curvature-higher-geometry.tex

Source for Connections, Curvature, and Higher Geometry: From covariant differentiation to differential cohomology

View paper · Download original file

Lines 441–540 of 4453. Source text is displayed literally.

441For a two-plane \(\Pi\subseteq T_pM\), choose a spanning pair \(X,Y\). Its sectional curvature is442\[443 K(\Pi)=444 \frac{g(R(X,Y)Y,X)}445 {|X|_g^2|Y|_g^2-g(X,Y)^2}.446\]447In the numerator, skew symmetry in each pair gives a quadratic expression on \(X\wedge Y\). Both numerator and denominator therefore scale by the square of the determinant of a change of basis of \(\Pi\). This proves independence of the spanning pair. On a surface there is only one plane at each point, giving its Gaussian curvature.448 449\begin{example}[Bi-invariant metrics]450In the Lie group example of Chapter~\ref{ch:levi}, let \(X,Y,Z\) be left-invariant fields. The commutator of their covariant derivatives acts on \(Z\) by \(\tfrac14[[X,Y],Z]\), since the Jacobi identity rewrites the difference of the two nested brackets. The last term of \eqref{eq:intrinsic-curvature} acts by \(\tfrac12[[X,Y],Z]\). Thus451\[452 R(X,Y)Z=-\tfrac14[[X,Y],Z].453\]454Taking \(Z=Y\), pairing with \(X\), and using skew-adjointness of the bracket gives \(g(R(X,Y)Y,X)=\tfrac14|[X,Y]|^2\). Positive sectional curvature in a plane therefore occurs exactly when the two directions fail to commute. On a torus all brackets vanish, and this metric is flat.455\end{example}456 457\section{The Gauss and Codazzi identities}458 459Let \(M^n\) be immersed in \(\mathbb R^{n+1}\) with induced metric. Work locally with a smooth unit normal \(N\) and write \(D\) for the ambient Euclidean derivative. Differentiating \(\langle N,N\rangle=1\) shows that \(D_XN\) is tangent. Define the shape operator \(S(X)=-D_XN\) and the second fundamental form \(\mathrm{II}(X,Y)=g(SX,Y)\). Differentiating \(\langle Y,N\rangle=0\) yields460\[461 D_XY=\nabla_XY+\mathrm{II}(X,Y)N .462\]463Here \(\nabla\) is Levi-Civita by Chapter~\ref{ch:levi}. Taking the normal component of \(D_XY-D_YX=[X,Y]\) proves that \(\mathrm{II}\) is symmetric. This decomposition can also be recorded by the splitting464\[465 \begin{tikzcd}[column sep=3em]466 TM\oplus\mathbb RN \arrow[r,"\simeq"]467      \arrow[d,"\operatorname{pr}_{TM}"']&468      M\times\mathbb R^{n+1}\arrow[dl,"\operatorname{pr}_{TM}"]\\469 TM .470 \end{tikzcd}471\]472The derivative in the ambient bundle does not preserve either summand in general, and \(S\) and \(\mathrm{II}\) measure its off-diagonal terms.473 474\begin{theorem}[Gauss and Codazzi for a Euclidean hypersurface]475One has476\begin{align*}477 R(X,Y)Z478   &=\mathrm{II}(Y,Z)SX-\mathrm{II}(X,Z)SY,\\479 (\nabla_X\mathrm{II})(Y,Z)&=(\nabla_Y\mathrm{II})(X,Z).480\end{align*}481\end{theorem}482 483\begin{proof}484Insert the decomposition of \(D_YZ\) into \(D_XD_YZ\), subtract the counterpart with \(X,Y\) interchanged, and subtract \(D_{[X,Y]}Z\). Flatness of the Euclidean derivative makes the result zero. Its tangential component is \(R(X,Y)Z-\mathrm{II}(Y,Z)SX+\mathrm{II}(X,Z)SY\). Its normal component is the coefficient485\[486 \begin{split}487 X\mathrm{II}(Y,Z)-Y\mathrm{II}(X,Z)488 &+\mathrm{II}(X,\nabla_YZ)-\mathrm{II}(Y,\nabla_XZ)\\489 &-\mathrm{II}([X,Y],Z).490 \end{split}491\]492Use \([X,Y]=\nabla_XY-\nabla_YX\) and the product rule for the covariant derivative of a covariant tensor. The coefficient is the difference of the two derivatives in the second identity.493\end{proof}494 495When \(n=2\), choosing an orthonormal basis gives \(K=\det S\). Changing \(N\) to \(-N\) negates \(S\) while preserving its determinant in dimension two. Because the right-hand definition of \(K\) uses only the metric and its Levi-Civita connection, two isometric surfaces have equal Gaussian curvature at corresponding points. This is the intrinsic content of the Theorema Egregium.496 497\begin{example}[Graph, cylinder, and sphere]498For a graph \(z=f(x,y)\) in \(\mathbb R^3\), put \(w=(1+|\nabla f|^2)^{1/2}\) and use the upward normal \(N=(-f_x,-f_y,1)/w\). Its first fundamental form has determinant \(w^2\). The second fundamental form has matrix \((f_{ij}/w)\). Thus \(K=\det(f_{ij})/w^4\). Taking \(f=xy\) gives negative curvature at every point of its graph. On the cylinder of radius \(a\), the two eigenvalues of \(S\) with outward normal are \(-1/a\) in the circular direction and \(0\) along its axis. Hence \(K=0\). The local parametrization \((s,t)\mapsto(a\cos(s/a),a\sin(s/a),t)\) exhibits the same result by pulling the metric back to \(ds^2+dt^2\). On the sphere of radius \(a\) one has \(S=-a^{-1}\operatorname{id}\), so \(K=a^{-2}\).499\end{example}500 501\section{A curvature form on an oriented surface}502 503Take a positive orthonormal frame \(e_1,e_2\) and define a one-form \(\omega\) by \(\nabla e_1=\omega e_2\). Metric compatibility gives \(\nabla e_2=-\omega e_1\). Insert these two equations into \eqref{eq:intrinsic-curvature}. The product terms in \(\omega(X)\omega(Y)\) cancel, yielding504\[505 R(X,Y)e_1=(d\omega)(X,Y)e_2,\qquad506 d\omega=-K\,dV_g .507\]508The minus sign follows from \(K=g(R(e_1,e_2)e_2,e_1)\). For the positive coframe \((\theta^1,\theta^2)\), the first structure equations are509\(d\theta^1=\omega\wedge\theta^2\) and510\(d\theta^2=-\omega\wedge\theta^1\). They follow by evaluating \(d\theta^i\) on two fields and replacing their bracket by the difference of the covariant derivatives. Rotating the frame through an angle \(\phi\) changes \(\omega\) to \(\omega+d\phi\), by differentiation of the rotation. The curvature form is thus unchanged. When local angles have only circle-valued transitions, the differentials \(d\phi\) still agree for different real lifts. In Chapter~\ref{ch:geodesics}, integration of this identity turns those frame changes into a topological integer.511 512\begin{example}[The effect of conformally scaling a plane metric]513Take \(g=e^{2u}(dx^2+dy^2)\) and the positive coframe \((e^u dx,e^u dy)\). The first structure equations give514\(\omega=-u_y\,dx+u_x\,dy\). Consequently515\[516 K=-e^{-2u}(u_{xx}+u_{yy}).517\]518For \(u=-\log y\) on the upper half-plane, this is \(-1\). For \(u=\log(2a^2)-\log(a^2+x^2+y^2)\), the Laplacian is519\(-4a^2/(a^2+x^2+y^2)^2\), giving \(K=a^{-2}\). The latter metric is the usual stereographic expression for the round sphere of radius \(a\). More generally, on a conformal chart the curvature is controlled by a scalar second-order equation. The sign of its Laplacian determines the sign of the Gaussian curvature pointwise.520\end{example}521\chapter{Geodesics, Jacobi fields, and surfaces}522\label{ch:geodesics}523 524The covariant derivative turns geodesic motion into a differential equation. A variation of one geodesic through other geodesics satisfies a new, linear equation containing curvature. On surfaces the same curvature has a global integral constrained by the topology. The local motion and the global identity thus give two ways of detecting the curvature defined in Chapter~\ref{ch:intrinsic}. We use Tu's conventions for the geodesic equation and his presentation of the surface theorem as a comparison \cite[\S\S\,13--17]{tu}.525 526\section{Energy variation and the exponential map}527 528Let \(\Gamma(s,t)\) be a smooth variation of a path \(\gamma(t)=\Gamma(0,t)\) on \([a,b]\). Put \(T(s,t)=\partial_t\Gamma(s,t)\) and \(V(s,t)=\partial_s\Gamma(s,t)\). The energy of the path \(t\mapsto\Gamma(s,t)\) is529\[530 E(s)=\frac12\int_a^b531    g_{\Gamma(s,t)}\bigl(T(s,t),T(s,t)\bigr)\,dt ,532\]533with \(s\) held fixed during integration. Both fields are sections of \(\Gamma^*TM\). The torsion-free identity \(\nabla_sT=\nabla_tV\) holds because the mixed partials of \(\Gamma\) agree and the Christoffel contributions are symmetric. Therefore534\begin{equation}535 E'(s)=\bigl[g(V,T)\bigr]_a^b-536                     \int_a^b g(V,\nabla_tT)\,dt .537 \label{eq:geo-first-var}538\end{equation}539Here metric compatibility first replaces the derivative of the squared speed by \(2g(\nabla_sT,T)\), and then gives the integration-by-parts formula. For variations fixing both endpoints, the boundary term vanishes. Every smooth field supported in a small segment of the interior serves as a variation vector there, using coordinate displacement multiplied by a cutoff. Thus \(E'(0)=0\) for every fixed-endpoint variation implies the geodesic equation along \(\gamma\) by the fundamental lemma of the calculus of variations. The reverse implication follows from \eqref{eq:geo-first-var}.540