Q
stringlengths
27
2.03k
A
stringlengths
4
2.98k
Result
stringclasses
2 values
Lemma 6.4. If \( \gamma \) is an admissible curve and \( V \) is a vector field along \( \gamma \) , then \( V \) is the variation field of some variation of \( \gamma \) . If \( V \) is proper, the variation can be taken to be proper as well.
Proof. Set \( \Gamma \left( {s, t}\right) = \exp \left( {{sV}\left( t\right) }\right) \) (Figure 6.5). By compactness of \( \left\lbrack {a, b}\right\rbrack \), there is some positive \( \varepsilon \) such that \( \Gamma \) is defined on \( \left( {-\varepsilon ,\varepsilon }\right) \times \left\lbrack {a, b}\right\rbrack \) . Clearly \( \Gamma \) is smooth on \( \left( {-\varepsilon ,\varepsilon }\right) \times \left\lbrack {{a}_{i - 1},{a}_{i}}\right\rbrack \) for each subinterval \( \left\lbrack {{a}_{i - 1},{a}_{i}}\right\rbrack \) on which \( V \) is smooth, and is continuous on its whole domain. By the properties of the exponential map, the variation field of \( \Gamma \) is \( V \) . Moreover, if \( V\left( a\right) = V\left( b\right) = \) 0, it is immediate that \( \Gamma \left( {s, a}\right) \equiv \gamma \left( a\right) \) and \( \Gamma \left( {s, b}\right) \equiv \gamma \left( b\right) \), so \( \Gamma \) is proper.
Yes
Theorem 6.6. Every minimizing curve is a geodesic when it is given a unit speed parametrization.
Proof. Suppose \( \gamma : \left\lbrack {a, b}\right\rbrack \rightarrow M \) is minimizing and unit speed, and let \( a = \) \( {a}_{0} < \cdots < {a}_{k} = b \) be a subdivision such that \( \gamma \) is smooth on \( \left\lbrack {{a}_{i - 1},{a}_{i}}\right\rbrack \) . If \( \Gamma \) is any proper variation of \( \gamma \), we conclude from elementary calculus that \( {dL}\left( {\Gamma }_{s}\right) /{ds} = 0 \) when \( s = 0 \) . Since every proper vector field along \( \gamma \) is the variation field of some proper variation, the right-hand side of (6.2) must vanish for every such \( V \) .\n\nThe first step is to show that \( {D}_{t}\dot{\gamma } = 0 \) on each subinterval \( \left\lbrack {{a}_{i - 1},{a}_{i}}\right\rbrack \), so \( \gamma \) is a \
Yes
Corollary 6.7. A unit speed admissible curve \( \gamma \) is a critical point for \( L \) if and only if it is a geodesic.
Proof. If \( \gamma \) is a critical point, the proof of Theorem 6.6 goes through without modification to show that \( \gamma \) is a geodesic. Conversely, if \( \gamma \) is a geodesic, then the first term in the second variation formula vanishes by the geodesic equation, and the second term vanishes because \( \dot{\gamma } \) has no jumps.
Yes
Corollary 6.9. Let \( \left( {x}^{i}\right) \) be normal coordinates on a geodesic ball \( \mathcal{U} \) centered at \( p \in M \), and let \( r \) be the radial distance function as defined in (5.9). Then \( \operatorname{grad}r = \partial /\partial r \) on \( \mathcal{U} - \{ p\} \) .
Proof. For any \( q \in \mathcal{U} - \{ p\} \) and \( Y \in {T}_{q}M \), we need to show that\n\n\[ \n{dr}\left( Y\right) = \left\langle {\frac{\partial }{\partial r}, Y}\right\rangle \n\]\n\n(6.4)\n\nThe geodesic sphere \( {\exp }_{p}\left( {\partial {B}_{R}\left( 0\right) }\right) \) through \( q \) is characterized in normal coordinates by the equation \( r = R \) . Since \( \partial /\partial r \) is transverse to this sphere, we can decompose \( Y \) as \( \alpha \partial /\partial r + X \) for some constant \( \alpha \) and some vector \( X \) tangent to the sphere (Figure 6.11). Observe that \( {dr}\left( {\partial /\partial r}\right) = 1 \) by direct computation in coordinates, and \( {dr}\left( X\right) = 0 \) since \( X \) is tangent to a level set of \( r \) . (This has nothing to do with the metric!) Therefore the left-hand side of (6.4) is\n\n\[ \n{dr}\left( {\alpha \frac{\partial }{\partial r} + X}\right) = {\alpha dr}\left( \frac{\partial }{\partial r}\right) + {dr}\left( X\right) = \alpha .\n\]\n\nOn the other hand, by Proposition 5.11(e), \( \partial /\partial r \) is a unit vector. Therefore, the right-hand side of (6.4) is\n\n\[ \n\left\langle {\frac{\partial }{\partial r},\alpha \frac{\partial }{\partial r} + X}\right\rangle = \alpha {\left| \frac{\partial }{\partial r}\right| }^{2} + \left\langle {\frac{\partial }{\partial r}, X}\right\rangle = \alpha \n\]\n\nwhere we have used the Gauss lemma to conclude that \( X \) is orthogonal to \( \partial /\partial r \) .
Yes
Corollary 6.11. Within any geodesic ball around \( p \in M \), the radial distance function \( r\left( x\right) \) defined by (5.9) is equal to the Riemannian distance from \( p \) to \( x \) .
Proof. The radial geodesic \( \gamma \) from \( p \) to \( x \) is minimizing by Proposition 6.10. Since its velocity is equal to \( \partial /\partial r \), which is a unit vector in both the \( g \) norm and the Euclidean norm in normal coordinates, the \( g \) -length of \( \gamma \) is equal to its Euclidean length, which is \( r\left( x\right) \) .
Yes
Theorem 6.12. Every Riemannian geodesic is locally minimizing.
Proof. Let \( \gamma : I \rightarrow M \) be a geodesic, which we may assume to be defined on an open interval, and let \( {t}_{0} \in I \) . Let \( \mathcal{W} \) be a uniformly normal neighborhood of \( \gamma \left( {t}_{0}\right) \), and let \( \mathcal{U} \subset I \) be the connected component of \( {\gamma }^{-1}\left( \mathcal{W}\right) \) containing \( {t}_{0} \) . If \( {t}_{1},{t}_{2} \in \mathcal{U} \) and \( {q}_{i} = \gamma \left( {t}_{i}\right) \), the definition of uniformly normal neighborhood implies that \( {q}_{2} \) is contained in a geodesic ball around \( {q}_{1} \) (Figure 6.13). Therefore, by Proposition 6.10, the radial geodesic from \( {q}_{1} \) to \( {q}_{2} \) is the unique minimizing curve between them. However, the restriction of \( \gamma \) is a geodesic from \( {q}_{1} \) to \( {q}_{2} \) lying in the same geodesic ball, and thus \( \gamma \) must itself be this minimizing geodesic.
Yes
Theorem 6.13. (Hopf-Rinow) A connected Riemannian manifold is geodesically complete if and only if it is complete as a metric space.
Proof. Suppose first that \( M \) is complete as a metric space but not geodesically complete. Then there is some unit speed geodesic \( \gamma : \lbrack 0, b) \rightarrow M \) that extends to no interval \( \lbrack 0, b + \varepsilon ) \) for \( \varepsilon > 0 \) . Let \( \left\{ {t}_{i}\right\} \) be any increasing sequence that approaches \( b \), and set \( {q}_{i} = \gamma \left( {t}_{i}\right) \) . Since \( \gamma \) is parametrized by arc length, the length of \( {\left. \gamma \right| }_{\left\lbrack {t}_{i},{t}_{j}\right\rbrack } \) is exactly \( \left| {{t}_{j} - {t}_{i}}\right| \), so \( d\left( {{q}_{i},{q}_{j}}\right) \leq \left| {{t}_{j} - {t}_{i}}\right| \) and \( \left\{ {q}_{i}\right\} \) is a Cauchy sequence in \( M \) . By completeness, \( \left\{ {q}_{i}\right\} \) converges to some point \( q \in M \) .\n\nLet \( \mathcal{W} \) be a uniformly normal neighborhood of \( q \), and let \( \delta > 0 \) be chosen so that \( \mathcal{W} \) is contained in a geodesic \( \delta \) -ball around each of its points. For all large \( j,{q}_{j} \in \mathcal{W} \) (Figure 6.14), and by taking \( j \) large enough, we may assume \( {t}_{j} > b - \delta \) . The fact that \( {B}_{\delta }\left( {q}_{j}\right) \) is a geodesic ball means that every geodesic starting at \( {q}_{j} \) exists at least for time \( \delta \) . In particular, this is true of the geodesic \( \sigma \) with \( \sigma \left( 0\right) = {q}_{j} \) and \( \dot{\sigma }\left( 0\right) = \dot{\gamma }\left( {t}_{j}\right) \) . But by uniqueness of geodesics, this must be simply a reparametrization of \( \gamma \), so \( \widetilde{\gamma }\left( t\right) = \sigma \left( {{t}_{j} + t}\right) \) is an extension of \( \gamma \) past \( b \), which is a contradiction.
Yes
Corollary 6.14. If there exists one point \( p \in M \) such that the restricted exponential map \( {\exp }_{p} \) is defined on all of \( {T}_{p}M \), then \( M \) is complete.
Null
No
Corollary 6.15. \( M \) is complete if and only if any two points in \( M \) can be joined by a minimizing geodesic segment.
Null
No
Corollary 6.16. If \( M \) is compact, then every geodesic can be defined for all time.
Null
No
Lemma 7.2. The Riemann curvature endomorphism and curvature tensor are local isometry invariants. More precisely, if \( \varphi : \left( {M, g}\right) \rightarrow \left( {\widetilde{M},\widetilde{g}}\right) \) is a local isometry, then\n\n\[{\varphi }^{ * }\widetilde{Rm} = {Rm}\]\n\n\[ \widetilde{R}\left( {{\varphi }_{ * }X,{\varphi }_{ * }Y}\right) {\varphi }_{ * }Z = {\varphi }_{ * }\left( {R\left( {X, Y}\right) Z}\right) . \]
Exercise 7.2. Prove Lemma 7.2.
No
Proposition 7.4. (Symmetries of the Curvature Tensor) The curvature tensor has the following symmetries for any vector fields \( W, X, Y \) , \( Z \) :\n\n(a) \( \operatorname{Rm}\left( {W, X, Y, Z}\right) = - \operatorname{Rm}\left( {X, W, Y, Z}\right) \) .\n\n(b) \( \operatorname{Rm}\left( {W, X, Y, Z}\right) = - \operatorname{Rm}\left( {W, X, Z, Y}\right) \) .\n\n(c) \( \operatorname{Rm}\left( {W, X, Y, Z}\right) = \operatorname{Rm}\left( {Y, Z, W, X}\right) \) .\n\n(d) \( \operatorname{Rm}\left( {W, X, Y, Z}\right) + \operatorname{Rm}\left( {X, Y, W, Z}\right) + \operatorname{Rm}\left( {Y, W, X, Z}\right) = 0 \) .
Proof of Proposition 7.4. Identity (a) is immediate from the obvious fact that \( R\left( {W, X}\right) Y = - R\left( {X, W}\right) Y \) . To prove (b), it suffices to show that \( {Rm}\left( {W, X, Y, Y}\right) = 0 \) for all \( Y \), for then (b) follows from the expansion of \( \operatorname{Rm}\left( {W, X, Y + Z, Y + Z}\right) = 0 \) . Using compatibility with the metric, we have\n\n\[ {WX}{\left| Y\right| }^{2} = W\left( {2\left\langle {{\nabla }_{X}Y, Y}\right\rangle }\right) = 2\left\langle {{\nabla }_{W}{\nabla }_{X}Y, Y}\right\rangle + 2\left\langle {{\nabla }_{X}Y,{\nabla }_{W}Y}\right\rangle \]\n\n\[ {XW}{\left| Y\right| }^{2} = X\left( {2\left\langle {{\nabla }_{W}Y, Y}\right\rangle }\right) = 2\left\langle {{\nabla }_{X}{\nabla }_{W}Y, Y}\right\rangle + 2\left\langle {{\nabla }_{W}Y,{\nabla }_{X}Y}\right\rangle \]\n\n\[ \left\lbrack {W, X}\right\rbrack {\left| Y\right| }^{2} = 2\left\langle {{\nabla }_{\left\lbrack W, X\right\rbrack }Y, Y}\right\rangle \]\n\nWhen we subtract the second and third equations from the first, the left-hand side is zero. The terms \( 2\left\langle {{\nabla }_{X}Y,{\nabla }_{W}Y}\right\rangle \) and \( 2\left\langle {{\nabla }_{W}Y,{\nabla }_{X}Y}\right\rangle \) cancel on the right-hand side, giving\n\n\[ 0 = 2\langle {\nabla }_{W}{\nabla }_{X}Y, Y\rangle - 2\langle {\nabla }_{X}{\nabla }_{W}Y, Y\rangle - 2\langle {\nabla }_{\left\lbrack W, X\right\rbrack }Y, Y\rangle \]\n\n\[ = 2\langle R\left( {W, X}\right) Y, Y\rangle \]\n\n\[ = {2Rm}\left( {W, X, Y, Y}\right) \text{.} \]\n\nNext we prove (d). From the definition of \( {Rm} \), this will follow immediately from\n\n\[ R\left( {W, X}\right) Y + R\left( {X, Y}\right) W + R\left( {Y, W}\right) X = 0. \]\n\nUsing the definition of \( R \) and the symmetry of the connection, the left-hand side expands to\n\n\[ \left( {{\nabla }_{W}{\nabla }_{X}Y - {\nabla }_{X}{\nabla }_{W}Y - {\nabla }_{\left\lbrack W, X\right\rbrack }Y}\right) \]\n\n\[ + \left( {{\nabla }_{X}{\nabla }_{Y}W - {\nabla }_{Y}{\nabla }_{X}W - {\nabla }_{\left\lbrack X, Y\right\rbrack }W}\right) \]\n\n\[ + \left( {{\nabla }_{Y}{\nabla }_{W}X - {\nabla }_{W}{\nabla }_{Y}X - {\nabla }_{\left\lbrack Y, W\right\rbrack }X}\right) \]\n\n\[ = {\nabla }_{W}\left( {{\nabla }_{X}Y - {\nabla }_{Y}X}\right) + {\nabla }_{X}\left( {{\nabla }_{Y}W - {\nabla }_{W}Y}\right) + {\nabla }_{Y}\left( {{\nabla }_{W}X - {\nabla }_{X}W}\right) \]\n\n\[ - {\nabla }_{\left\lbrack W, X\right\rbrack }Y - {\nabla }_{\left\lbrack X
Yes
Proposition 7.5. (Differential Bianchi Identity) The total covariant derivative of the curvature tensor satisfies the following identity:\n\n\[ \n\nabla {Rm}\left( {X, Y, Z, V, W}\right) + \nabla {Rm}\left( {X, Y, V, W, Z}\right) + \nabla {Rm}\left( {X, Y, W, Z, V}\right) = 0.\n\]
Proof. First of all, by the symmetries of \( {Rm} \) ,(7.6) is equivalent to\n\n\[ \n\nabla {Rm}\left( {Z, V, X, Y, W}\right) + \nabla {Rm}\left( {V, W, X, Y, Z}\right) + \nabla {Rm}\left( {W, Z, X, Y, V}\right) = 0.\n\]\n\nThis can be proved by a long and tedious computation, but there is a standard shortcut for such calculations in Riemannian geometry that makes our task immeasurably easier. To prove (7.6) holds at a particular point \( p \), by multilinearity it suffices to prove the formula when \( X, Y, Z, V, W \) are basis elements with respect to some frame. The shortcut consists of choosing a special frame for each point \( p \) to simplify the computations there.\n\nLet \( \left( {x}^{i}\right) \) be normal coordinates at \( p \), and let \( X, Y, Z, V, W \) be arbitrary coordinate basis vectors \( {\partial }_{i} \) . These vectors satisfy two properties that simplify our computations enormously: (1) their commutators vanish identically, since \( \left\lbrack {{\partial }_{i},{\partial }_{j}}\right\rbrack \equiv 0 \) ; and (2) their covariant derivatives vanish at \( p \), since \( {\Gamma }_{ij}^{k}\left( p\right) = 0 \) (Proposition 5.11(f)).\n\nUsing these facts and the compatibility of the connection with the metric, the first term in (7.8) evaluated at \( p \) becomes\n\n\[ \n{\nabla }_{W}\operatorname{Rm}\left( {Z, V, X, Y}\right) = {\nabla }_{W}\langle R\left( {Z, V}\right) X, Y\rangle\n\]\n\n\[ \n= \left\langle {{\nabla }_{W}{\nabla }_{Z}{\nabla }_{V}X - {\nabla }_{W}{\nabla }_{V}{\nabla }_{Z}X, Y}\right\rangle .\n\]\n\nWrite this equation three times, with the vector fields \( W, Z, V \) cyclically permuted. Summing all three gives\n\n\[ \n\nabla \operatorname{Rm}\left( {Z, V, X, Y, W}\right) + \nabla \operatorname{Rm}\left( {V, W, X, Y, Z}\right) + \nabla \operatorname{Rm}\left( {W, Z, X, Y, V}\right)\n\]\n\n\[ \n= \langle {\nabla }_{W}{\nabla }_{Z}{\nabla }_{V}X - {\nabla }_{W}{\nabla }_{V}{\nabla }_{Z}X\n\]\n\n\[ \n+ {\nabla }_{Z}{\nabla }_{V}{\nabla }_{W}X - {\nabla }_{Z}{\nabla }_{W}{\nabla }_{V}X\n\]\n\n\[ \n\; + {\nabla }_{V}{\nabla }_{W}{\nabla }_{Z}X - {\nabla }_{V}{\nabla }_{Z}{\nabla }_{W}X, Y\rangle\n\]\n\n\[ \n= \langle R\left( {W, Z}\right) {\nabla }_{V}X + R\left( {Z, V}\right) {\nabla }_{W}X + R\left( {V, W}\right) {\nabla }_{Z}X, Y\rangle\n\]\n\n\[ \n= 0,\n\]\n\nwhere the last line follows because \( {\nabla }_{V}X = {\nabla }_{W}X = {\nabla }_{Z}X = 0 \) at \( p \) .
Yes
Exercise 7.5. Prove Lemma 7.6, using the symmetries of the curvature tensor.
Null
No
Lemma 7.7. (Contracted Bianchi Identity) The covariant derivatives of the Ricci and scalar curvatures satisfy the following identity:\n\n\[ \operatorname{div}{Rc} = \frac{1}{2}\nabla S \]\n\nwhere div is the divergence operator (Problem 3-3). In components, this is\n\n\[ {R}_{{ij};}{}^{j} = \frac{1}{2}{S}_{;i} \]
Proof. Formula (7.9) follows immediately by contracting the component form (7.7) of the differential Bianchi identity on the indices \( i, l \) and then again on \( j, k \), after raising one index of each pair.
No
Proposition 7.8. If \( g \) is an Einstein metric on a connected manifold of dimension \( n \geq 3 \), its scalar curvature is constant.
Proof. Taking the covariant derivative of each side of (7.10) and noting that the covariant derivative of the metric is zero, we see that the Einstein condition implies\n\n\[ {R}_{{ij};k} = \frac{1}{n}{S}_{;k}{g}_{ij} \]\n\nTracing this equation on \( j \) and \( k \), and comparing with the contracted Bianchi identity (7.9), we conclude\n\n\[ \frac{1}{2}{S}_{;i} = \frac{1}{n}{S}_{;i} \]\n\nWhen \( n > 2 \), this implies \( {S}_{;i} = 0 \) . But \( {S}_{;i} \) is the component of \( \nabla S = {ds} \), so connectedness of \( M \) implies \( S \) is constant.
Yes
Lemma 8.1. The second fundamental form is\n\n(a) independent of the extensions of \( X \) and \( Y \) ;\n\n(b) bilinear over \( {C}^{\infty }\left( M\right) \) ; and\n\n(c) symmetric in \( X \) and \( Y \) .
Proof. First we show that the symmetry of \( \Pi \) follows from the symmetry of the connection \( \widetilde{\nabla } \) . Let \( X \) and \( Y \) be extended arbitrarily to \( M \) . Then\n\n\[ \Pi \left( {X, Y}\right) - \Pi \left( {Y, X}\right) = {\left( {\widetilde{\nabla }}_{X}Y - {\widetilde{\nabla }}_{Y}X\right) }^{ \bot } = {\left\lbrack X, Y\right\rbrack }^{ \bot }. \]\n\nSince \( X \) and \( Y \) are tangent to \( M \) at all points of \( M \), so is their Lie bracket. (This follows easily from Exercise 2.3.) Therefore \( {\left\lbrack X, Y\right\rbrack }^{ \bot } = 0 \), so \( \Pi \) is symmetric.\n\nBecause \( {\left. {\widetilde{\nabla }}_{X}Y\right| }_{p} \) depends only on \( {X}_{p} \), it is clear that \( \Pi \left( {X, Y}\right) \) is independent of the extension chosen for \( X \), and that \( \Pi \left( {X, Y}\right) \) is linear over \( {C}^{\infty }\left( M\right) \) in \( X \) . By symmetry, the same is true for \( Y \) .
Yes
Theorem 8.2. (The Gauss Formula) If \( X, Y \in \mathcal{T}\left( M\right) \) are extended arbitrarily to vector fields on \( \widetilde{M} \), the following formula holds along \( M \) :
Proof. Because of the decomposition (8.1) and the definition of the second fundamental form, it suffices to show that \( {\left( {\widetilde{\nabla }}_{X}Y\right) }^{\top } = {\nabla }_{X}Y \) at all points of \( M \) .\n\nDefine a map \( {\nabla }^{\top } : \mathcal{T}\left( M\right) \times \mathcal{T}\left( M\right) \rightarrow \mathcal{T}\left( M\right) \) by\n\n\[ {\nabla }_{X}^{\top }Y \mathrel{\text{:=}} {\left( {\widetilde{\nabla }}_{X}Y\right) }^{\top } \]\n\nwhere \( X, Y \) are extended arbitrarily to \( \widetilde{M} \) . We examined a special case of this construction, in which \( \widetilde{g} \) is the Euclidean metric, in Lemma 5.1. It follows exactly as in the proof of that lemma that \( {\nabla }^{\top } \) is a connection on \( M \) . Once we show that it is symmetric and compatible with \( g \), the uniqueness of the Riemannian connection on \( M \) shows that \( {\nabla }^{\top } = \nabla \) .\n\nTo see that \( {\nabla }^{\top } \) is symmetric, we use the symmetry of \( \widetilde{\nabla } \) and the fact that \( \left\lbrack {X, Y}\right\rbrack \) is tangent to \( M \) :\n\n\[ {\nabla }_{X}^{\top }Y - {\nabla }_{Y}^{\top }X = {\left( {\widetilde{\nabla }}_{X}Y - {\widetilde{\nabla }}_{Y}X\right) }^{\top } \]\n\n\[ = {\left\lbrack X, Y\right\rbrack }^{\top } = \left\lbrack {X, Y}\right\rbrack \]\n\nTo prove compatibility with \( g \), let \( X, Y, Z \in \mathcal{T}\left( M\right) \) be extended arbitrarily to \( \widetilde{M} \) . Using compatibility of \( \widetilde{\nabla } \) with \( \widetilde{g} \), and evaluating at points of \( M \) ,\n\n\[ X\langle Y, Z\rangle = \left\langle {{\widetilde{\nabla }}_{X}Y, Z}\right\rangle + \left\langle {Y,{\widetilde{\nabla }}_{X}Z}\right\rangle \]\n\n\[ = \left\langle {{\left( {\widetilde{\nabla }}_{X}Y\right) }^{\top }, Z}\right\rangle + \left\langle {Y,{\left( {\widetilde{\nabla }}_{X}Z\right) }^{\top }}\right\rangle \]\n\n\[ = \left\langle {{\nabla }_{X}^{\top }Y, Z}\right\rangle + \left\langle {Y,{\nabla }_{X}^{\top }Z}\right\rangle \]\n\nTherefore \( {\nabla }^{\top } \) is compatible with \( g \), so \( {\nabla }^{\top } = \nabla \) .
Yes
Lemma 8.3. (The Weingarten Equation) Suppose \( X, Y \in \mathfrak{T}\left( M\right) \) and \( N \in \mathcal{N}\left( M\right) \) . When \( X, Y, N \) are extended arbitrarily to \( \widetilde{M} \), the following equation holds at points of \( M \) :
Proof. Since \( \langle N, Y\rangle \) vanishes identically along \( M \) and \( X \) is tangent to \( M \) , the following holds along \( M \) :\n\n\[ 0 = X\langle N, Y\rangle \]\n\n\[ = \left\langle {{\widetilde{\nabla }}_{X}N, Y}\right\rangle + \left\langle {N,{\widetilde{\nabla }}_{X}Y}\right\rangle \]\n\n\[ = \left\langle {{\widetilde{\nabla }}_{X}N, Y}\right\rangle + \left\langle {N,{\nabla }_{X}Y + \Pi \left( {X, Y}\right) }\right\rangle \]\n\n\[ = \left\langle {{\widetilde{\nabla }}_{X}N, Y}\right\rangle + \langle N,\Pi \left( {X, Y}\right) \rangle .\n\n
Yes
Theorem 8.4. (The Gauss Equation) For any \( X, Y, Z, W \in {T}_{p}M \), the following equation holds:\n\n\[ \widetilde{\operatorname{Rm}}\left( {X, Y, Z, W}\right) = \operatorname{Rm}\left( {X, Y, Z, W}\right) \]\n\n\[ - \langle \Pi \left( {X, W}\right) ,\Pi \left( {Y, Z}\right) \rangle + \langle \Pi \left( {X, Z}\right) ,\Pi \left( {Y, W}\right) \rangle . \]
Proof. Let \( X, Y, Z, W \) be extended arbitrarily to vector fields on \( M \), and then to vector fields on \( \widetilde{M} \) that are tangent to \( M \) at points of \( M \) . Along \( M \), the Gauss formula gives\n\n\[ \widetilde{Rm}\left( {X, Y, Z, W}\right) = \left\langle {{\widetilde{\nabla }}_{X}{\widetilde{\nabla }}_{Y}Z - {\widetilde{\nabla }}_{Y}{\widetilde{\nabla }}_{X}Z - {\widetilde{\nabla }}_{\left\lbrack X, Y\right\rbrack }Z, W}\right\rangle \]\n\n\[ = \left\langle {{\widetilde{\nabla }}_{X}\left( {{\nabla }_{Y}Z + \Pi \left( {Y, Z}\right) }\right) - {\widetilde{\nabla }}_{Y}\left( {{\nabla }_{X}Z + \Pi \left( {X, Z}\right) }\right) }\right. \]\n\n\[ - \left. {\left( {{\nabla }_{\left\lbrack X, Y\right\rbrack }Z + \Pi \left( {\left\lbrack {X, Y}\right\rbrack, Z}\right) }\right), W}\right\rangle . \]\n\nSince the second fundamental form takes its values in the normal bundle and \( W \) is tangent to \( M \), the last \( \Pi \) term is zero. Apply the Weingarten equation to the other two terms involving \( \Pi \) (with \( \Pi \left( {Y, Z}\right) \) or \( \Pi \left( {X, Z}\right) \) playing the role of \( N \) ) to get\n\n\[ \widetilde{\operatorname{Rm}}\left( {X, Y, Z, W}\right) = \left\langle {{\widetilde{\nabla }}_{X}{\nabla }_{Y}Z, W}\right\rangle - \langle \Pi \left( {Y, Z}\right) ,\Pi \left( {X, W}\right) \rangle \]\n\n\[ - \left\langle {{\widetilde{\nabla }}_{Y}{\nabla }_{X}Z, W}\right\rangle + \langle \Pi \left( {X, Z}\right) ,\Pi \left( {Y, W}\right) \rangle \]\n\n\[ - \left\langle {{\nabla }_{\left\lbrack X, Y\right\rbrack }Z, W}\right\rangle \]\n\nDecomposing each term involving \( \widetilde{\nabla } \) into its tangential and normal components, we see that only the tangential component survives. Using the Gauss formula allows each to be rewritten in terms of \( \nabla \), giving\n\n\[ \widetilde{Rm}\left( {X, Y, Z, W}\right) = \left\langle {{\nabla }_{X}{\nabla }_{Y}Z, W}\right\rangle - \left\langle {{\nabla }_{Y}{\nabla }_{X}Z, W}\right\rangle - \left\langle {{\nabla }_{\left\lbrack X, Y\right\rbrack }Z, W}\right\rangle \]\n\n\[ - \langle \Pi \left( {Y, Z}\right) ,\Pi \left( {X, W}\right) \rangle + \langle \Pi \left( {X, Z}\right) ,\Pi \left( {Y, W}\right) \rangle \]\n\n\[ = \langle R\left( {X, Y}\right) Z, W\rangle \]\n\n\[ - \langle \Pi \left( {X, W}\right) ,\Pi \left( {Y, Z}\right) \rangle + \langle \Pi \left( {X, Z}\right) ,\Pi \left( {Y, W}\right) \rangle . \]\n\nThis proves the theorem.
Yes
Lemma 8.5. (The Gauss Formula Along a Curve) Let \( M \) be a Riemannian submanifold of \( \widetilde{M} \), and \( \gamma \) a curve in \( M \) . For any vector field \( V \) tangent to \( M \) along \( \gamma \) ,\n\n\[{\widetilde{D}}_{t}V = {D}_{t}V + \Pi \left( {\dot{\gamma }, V}\right)\]
Proof. In terms of an adapted orthonormal frame, \( V \) can be written \( V\left( t\right) = \) \( {V}^{i}\left( t\right) {E}_{i} \), where the sum is only over \( i = 1,\ldots, n \) . Applying the product rule and the Gauss formula, we get\n\n\[{\widetilde{D}}_{t}V = {\dot{V}}^{i}{E}_{i} + {V}^{i}{\widetilde{\nabla }}_{\dot{\gamma }}{E}_{i}\]\n\n\[= {\dot{V}}^{i}{E}_{i} + {V}^{i}{\nabla }_{\dot{\gamma }}{E}_{i} + {V}^{i}\Pi \left( {\dot{\gamma },{E}_{i}}\right)\]\n\n\[= {D}_{t}V + \Pi \left( {\dot{\gamma }, V}\right)\]
Yes
Theorem 8.6. (Gauss’s Theorema Egregium) Let \( M \subset {\mathbf{R}}^{3} \) be a 2-dimensional submanifold and \( g \) the induced metric on \( M \). For any \( p \in M \) and any basis \( \left( {X, Y}\right) \) for \( {T}_{p}M \), the Gaussian curvature of \( M \) at \( p \) is given \( {by} \)\n\n\[ K = \frac{\operatorname{Rm}\left( {X, Y, Y, X}\right) }{{\left| X\right| }^{2}{\left| Y\right| }^{2}-\langle X, Y{\rangle }^{2}}. \]
Proof. We begin with the special case in which \( \left( {X, Y}\right) = \left( {{E}_{1},{E}_{2}}\right) \) is an orthonormal basis for \( {T}_{p}M \). In this case the denominator in (8.5) is equal to 1. If we write \( {h}_{ij} = \dot{h}\left( {{E}_{i},{E}_{j}}\right) \), then in this basis \( K = \det s = \det \left( {h}_{ij}\right) \), and the Gauss equation (8.4) reads\n\n\[ \operatorname{Rm}\left( {{E}_{1},{E}_{2},{E}_{2},{E}_{1}}\right) = {h}_{11}{h}_{22} - {h}_{12}{h}_{21} = \det \left( {h}_{ij}\right) = K. \]\n\nThis is equivalent to (8.5).\n\nNow let \( X, Y \) be any basis for \( {T}_{p}M \). The Gram-Schmidt algorithm yields an orthonormal basis as follows:\n\n\[ {E}_{1} = \frac{X}{\left| X\right| } \]\n\n\[ {E}_{2} = \frac{Y-\langle Y,\frac{X}{\left| X\right| }\rangle \frac{X}{\left| X\right| }}{\left| Y-\langle Y,\frac{X}{\left| X\right| }\rangle \frac{X}{\left| X\right| }\right| } = \frac{Y - \frac{\langle Y, X\rangle }{{\left| X\right| }^{2}}X}{\left| Y - \frac{\langle Y, X\rangle }{{\left| X\right| }^{2}}X\right| }. \]\n\nThen by the preceding computation, the Gaussian curvature at \( p \) is\n\n\[ K = \operatorname{Rm}\left( {{E}_{1},{E}_{2},{E}_{2},{E}_{1}}\right) \]\n\n\[ = \frac{\operatorname{Rm}\left( {X, Y - \frac{\langle Y, X\rangle }{{\left| X\right| }^{2}}X, Y - \frac{\langle Y, X\rangle }{{\left| X\right| }^{2}}X, X}\right) }{{\left| X\right| }^{2}{\left| Y - \frac{\langle Y, X\rangle }{{\left| X\right| }^{2}}X\right| }^{2}} \]\n\n\[ = \frac{\operatorname{Rm}\left( {X, Y, Y, X}\right) }{{\left| X\right| }^{2}\left( {{\left| Y\right| }^{2} - 2\frac{\langle Y, X{\rangle }^{2}}{{\left| X\right| }^{2}} + \frac{\langle Y, X{\rangle }^{2}}{{\left| X\right| }^{2}}}\right) } \]\n\n\[ = \frac{\operatorname{Rm}\left( {X, Y, Y, X}\right) }{{\left| X\right| }^{2}{\left| Y\right| }^{2}-\langle X, Y{\rangle }^{2}} \]\n\n(In the third line, we used the fact that \( \operatorname{Rm}\left( {X, X,\cdot , \cdot }\right) = \operatorname{Rm}\left( {\cdot ,\cdot, X, X}\right) = 0 \) by the symmetries of the curvature tensor.) This proves the theorem.
Yes
Lemma 8.7. The Gaussian curvature of a Riemannian 2-manifold is related to the curvature tensor, Ricci tensor, and scalar curvature by the formulas\n\n\[ \n{Rm}\left( {X, Y, Z, W}\right) = K\left( {\langle X, W\rangle \langle Y, Z\rangle -\langle X, Z\rangle \langle Y, W\rangle }\right) \]\n\n\[ \n{Rc}\left( {X, Y}\right) = K\langle X, Y\rangle \]\n\n(8.6)\n\n\[ \nS = {2K}\text{.} \]\n\nThus \( K \) is independent of choice of frame, and completely determines the curvature tensor.
Proof. Since both sides of the first equation are tensors, we can compute them in terms of any basis. Let \( \left( {{E}_{1},{E}_{2}}\right) \) be any orthonormal basis for \( {T}_{p}M \) , and consider the components \( {R}_{ijkl} = {Rm}\left( {{E}_{i},{E}_{j},{E}_{k},{E}_{l}}\right) \) of the curvature tensor. In terms of this basis,(8.5) gives \( K = {R}_{1221} \) . By antisymmetry, \( {R}_{ijkl} \) vanishes whenever \( i = j \) or \( k = l \), so the only nonzero components of \( {Rm} \) are\n\n\[ \n{R}_{1221} = {R}_{2112} = - {R}_{1212} = - {R}_{2121} = K. \]\n\nComparing \( \operatorname{Rm}\left( {X, Y, Z, W}\right) \) with \( K\left( {\langle X, W\rangle \langle Y, Z\rangle -\langle X, Z\rangle \langle Y, W\rangle }\right) \) when each of \( X, Y, Z, W \) is either \( {E}_{1} \) or \( {E}_{2} \) proves the first equation of (8.6).\n\nThe components of the Ricci tensor in this basis are\n\n\[ \n{R}_{ij} = {R}_{1ij1} + {R}_{2ij2} \]\n\nfrom which it follows easily that\n\n\[ \n{R}_{12} = {R}_{21} = 0;\;{R}_{11} = {R}_{22} = K, \]\n\nwhich is equivalent to the second equation. Finally, the scalar curvature is\n\n\[ \nS = {\operatorname{tr}}_{g}{Rc} = {R}_{11} + {R}_{22} = {2K}. \]\n\nBecause the scalar curvature is independent of choice of frame, so is \( K \) .
Yes
Proposition 8.8. If \( \\left( {X, Y}\\right) \) is any basis for a 2-plane \( \\Pi \\subset {T}_{p}M \), then\n\n\[ K\\left( {X, Y}\\right) = \\frac{\\operatorname{Rm}\\left( {X, Y, Y, X}\\right) }{{\\left| X\\right| }^{2}{\\left| Y\\right| }^{2}-\\langle X, Y{\\rangle }^{2}}.\]
Proof. For this proof, we denote the induced metric on \( {S}_{\\Pi } \) by \( \\widetilde{g} \), and continue to denote the metric on \( M \) by \( g \) . As in the first part of this chapter, we use tildes to denote geometric quantities associated with \( \\widetilde{g} \), but note that now the roles of \( g \) and \( \\widetilde{g} \) are reversed.\n\nWe claim first that the second fundamental form of \( {S}_{\\Pi } \) vanishes at \( p \) . To see why, let \( V \\in \\Pi \\subset {T}_{p}M \), and let \( \\gamma = {\\gamma }_{V} \) be the \( M \) -geodesic with initial velocity \( V \), which lies in \( {S}_{\\Pi } \) by definition. By the Gauss formula for vector fields along curves,\n\n\[ 0 = {D}_{t}\\dot{\\gamma } = {\\widetilde{D}}_{t}\\dot{\\gamma } + \\Pi \\left( {\\dot{\\gamma },\\dot{\\gamma }}\\right) \]\n\nSince the two terms in this sum are orthogonal, each must vanish identically. Evaluating at \( t = 0 \) gives \( \\Pi \\left( {V, V}\\right) = 0 \) . Since \( V \) was an arbitrary element of \( {T}_{p}M \) and \( {II} \) is symmetric, this shows that \( {II} = 0 \) at \( p \) . (We cannot in general expect \( \\Pi \) to vanish at other points of \( {S}_{\\mathrm{{II}}} \) -it is only at \( p \) that all geodesics starting tangent to \( S \) remain in \( S \) .)\n\nNow the Gauss equation tells us that the curvature tensors of \( {S}_{\\Pi } \) and \( M \) are related at \( p \) by\n\n\[ \\widetilde{\\operatorname{Rm}}\\left( {X, Y, Z, W}\\right) = \\operatorname{Rm}\\left( {X, Y, Z, W}\\right) \]\n\nwhenever \( X, Y, Z, W \\in \\Pi \) . In particular, the Gaussian curvature of \( {S}_{\\Pi } \) at \( p \) is\n\n\[ K\\left( \\Pi \\right) = \\frac{\\widetilde{Rm}\\left( {X, Y, Y, X}\\right) }{{\\left| X\\right| }^{2}{\\left| Y\\right| }^{2} - \\langle X, Y{\\rangle }^{2}} = \\frac{{Rm}\\left( {X, Y, Y, X}\\right) }{{\\left| X\\right| }^{2}{\\left| Y\\right| }^{2}-\\langle X, Y{\\rangle }^{2}}.\]\n\nThis is what was to be proved.
Yes
Lemma 8.9. Suppose \( {\mathcal{R}}_{1} \) and \( {\mathcal{R}}_{2} \) are covariant 4-tensors on a vector space \( V \) with an inner product, and both have the symmetries of the curvature tensor (as described in Proposition 7.4). If for every pair of independent vectors \( X, Y \in V \), \[ \frac{{\mathcal{R}}_{1}\left( {X, Y, Y, X}\right) }{{\left| X\right| }^{2}{\left| Y\right| }^{2}-\langle X, Y{\rangle }^{2}} = \frac{{\mathcal{R}}_{2}\left( {X, Y, Y, X}\right) }{{\left| X\right| }^{2}{\left| Y\right| }^{2}-\langle X, Y{\rangle }^{2}}, \] then \( {\mathcal{R}}_{1} = {\mathcal{R}}_{2} \) .
Proof. Setting \( \mathcal{R} = {\mathcal{R}}_{1} - {\mathcal{R}}_{2} \), it suffices to show \( \mathcal{R} = 0 \) under the assumption that \( \mathcal{R}\left( {X, Y, Y, X}\right) = 0 \) for all \( X, Y \). For any vectors \( X, Y, Z \), since \( \mathcal{R} \) also has the symmetries of the curvature tensor, \[ 0 = \mathcal{R}\left( {X + Y, Z, Z, X + Y}\right) \] \[ = \mathcal{R}\left( {X, Z, Z, X}\right) + \mathcal{R}\left( {X, Z, Z, Y}\right) + \mathcal{R}\left( {Y, Z, Z, X}\right) + \mathcal{R}\left( {Y, Z, Z, Y}\right) \] \[ = 2\mathcal{R}\left( {X, Z, Z, Y}\right) \] From this it follows that \[ 0 = \mathcal{R}\left( {X, Z + W, Z + W, Y}\right) \] \[ = \mathcal{R}\left( {X, Z, Z, Y}\right) + \mathcal{R}\left( {X, Z, W, Y}\right) + \mathcal{R}\left( {X, W, Z, Y}\right) + \mathcal{R}\left( {X, W, W, Y}\right) \] \[ = \mathcal{R}\left( {X, Z, W, Y}\right) + \mathcal{R}\left( {X, W, Z, Y}\right) . \] Therefore \( \mathcal{R} \) is antisymmetric in any adjacent pair of arguments. Now the algebraic Bianchi identity yields \[ 0 = \mathcal{R}\left( {X, Y, Z, W}\right) + \mathcal{R}\left( {Y, Z, X, W}\right) + \mathcal{R}\left( {Z, X, Y, W}\right) \] \[ = \mathcal{R}\left( {X, Y, Z, W}\right) - \mathcal{R}\left( {Y, X, Z, W}\right) - \mathcal{R}\left( {X, Z, Y, W}\right) \] \[ = 3\mathcal{R}\left( {X, Y, Z, W}\right) \text{.} \]
Yes
Lemma 8.10. Suppose \( \left( {M, g}\right) \) is any Riemannian \( n \) -manifold with constant sectional curvature \( C \) . The curvature endomorphism, curvature tensor, Ricci tensor, and scalar curvature of \( g \) are given by the formulas\n\n\[ R\left( {X, Y}\right) Z = C\left( {\\langle Y, Z\\rangle X-\\langle X, Z\\rangle Y}\right) \]\n\n\[ \\operatorname{Rm}\left( {X, Y, Z, W}\right) = C\left( {\\langle X, W\\rangle \\langle Y, Z\\rangle -\\langle X, Z\\rangle \\langle Y, W\\rangle }\right) \]\n\n\[ {Rc} = \\left( {n - 1}\\right) {Cg} \]\n\n\[ S = n\\left( {n - 1}\\right) C\\text{.} \]\n\nIn terms of any basis,\n\n\[ {R}_{ijkl} = C\\left( {{g}_{il}{g}_{jk} - {g}_{ik}{g}_{jl}}\\right) \]\n\n\[ {R}_{ij} = \\left( {n - 1}\\right) C{g}_{ij} \]
Exercise 8.8. Prove Lemma 8.10.
No
Lemma 9.2. If \( \gamma \) is a positively oriented curved polygon in \( M \), the rotation angle of \( \gamma \) is \( {2\pi } \) .
Proof. If we use the given coordinate chart to consider \( \gamma \) as a curved polygon in the plane, we can compute its tangent angle function either with respect to \( g \) or with respect to the Euclidean metric \( \bar{g} \) . In either case, \( \operatorname{Rot}\left( \gamma \right) \) is an integral multiple of \( {2\pi } \) because \( \theta \left( a\right) \) and \( \theta \left( b\right) \) both represent the same angle. Now for \( 0 \leq s \leq 1 \), let \( {g}_{s} = {sg} + \left( {1 - s}\right) \bar{g} \) . By the same reasoning, the rotation angle \( {\operatorname{Rot}}_{{g}_{s}}\left( \gamma \right) \) with respect to \( {g}_{s} \) is also a multiple of \( {2\pi } \) . The function \( f\left( s\right) = \left( {1/{2\pi }}\right) {\operatorname{Rot}}_{{g}_{s}}\left( \gamma \right) \) is therefore integer-valued, and is easily seen to be continuous in \( s \), so it must be constant.
Yes
Corollary 9.4. (Angle-Sum Theorem) The sum of the interior angles of a Euclidean triangle is \( \pi \) .
Null
No
Corollary 9.5. (Circumference Theorem) The circumference of a Euclidean circle of radius \( R \) is \( {2\pi R} \) .
Null
No
Corollary 9.6. (Total Curvature Theorem) If \( \gamma : \left\lbrack {a, b}\right\rbrack \rightarrow {\mathbf{R}}^{2} \) is a unit speed simple closed curve such that \( \dot{\gamma }\left( a\right) = \dot{\gamma }\left( b\right) \), and \( N \) is the inward-pointing normal, then\n\n\[{\int }_{a}^{b}{\kappa }_{N}\left( t\right) {dt} = {2\pi }\]
Null
No
Theorem 9.7. (The Gauss-Bonnet Theorem) If \( M \) is a triangulated, compact, oriented, Riemannian 2-manifold, then\n\n\[{\int }_{M}{KdA} = {2\pi \chi }\left( M\right)\]
Proof. Let \( \left\{ {{\Omega }_{i} : i = 1,\ldots ,{N}_{f}}\right\} \) denote the faces of the triangulation, and for each \( i \) let \( \left\{ {{\gamma }_{ij} : j = 1,2,3}\right\} \) be the edges of \( {\Omega }_{i} \) and \( \left\{ {{\theta }_{ij} : j = 1,2,3}\right\} \) its interior angles. Since each exterior angle is \( \pi \) minus the corresponding interior angle, applying the Gauss-Bonnet formula to each triangle and summing over \( i \) gives\n\n\[ \mathop{\sum }\limits_{{i = 1}}^{{N}_{f}}{\int }_{{\Omega }_{i}}{KdA} + \mathop{\sum }\limits_{{i = 1}}^{{N}_{f}}\mathop{\sum }\limits_{{j = 1}}^{3}{\int }_{{\gamma }_{ij}}{\kappa }_{N}{ds} + \mathop{\sum }\limits_{{i = 1}}^{{N}_{f}}\mathop{\sum }\limits_{{j = 1}}^{3}\left( {\pi - {\theta }_{ij}}\right) = \mathop{\sum }\limits_{{i = 1}}^{{N}_{f}}{2\pi }.\]\n\n(9.6)\n\nNote that each edge integral appears exactly twice in the above sum, with opposite orientations, so the integrals of \( {\kappa }_{N} \) all cancel out. Thus (9.6) becomes\n\n\[ {\int }_{M}{KdA} + {3\pi }{N}_{f} - \mathop{\sum }\limits_{{i = 1}}^{{N}_{f}}\mathop{\sum }\limits_{{j = 1}}^{3}{\theta }_{ij} = {2\pi }{N}_{f} \]\n\n(9.7)\n\nNote also that each interior angle \( {\theta }_{ij} \) appears exactly once. At each vertex, the angles that touch that vertex must add up to \( {2\pi } \) (Figure 9.16); thus the angle sum can be rearranged to give exactly \( {2\pi }{N}_{v} \) . Equation (9.7) thus can be written\n\n\[ {\int }_{M}{KdA} = {2\pi }{N}_{v} - \pi {N}_{f} \]\n\n(9.8)\n\nFinally, since each edge appears in exactly two triangles, and each triangle has exactly three edges, the total number of edges counted with multiplicity is \( 2{N}_{e} = 3{N}_{f} \), where we count each edge once for each triangle in which it appears. This means that \( {N}_{f} = 2{N}_{e} - 2{N}_{f} \), so (9.8) finally becomes\n\n\[ {\int }_{M}{KdA} = {2\pi }{N}_{v} - {2\pi }{N}_{e} + {2\pi }{N}_{f} = {2\pi \chi }\left( M\right) \]
Yes
Corollary 9.8. Let \( M \) be a compact Riemannian 2-manifold and \( K \) its Gaussian curvature.\n\n(a) If \( M \) is homeomorphic to the sphere or the projective plane, then \( K > 0 \) somewhere.\n\n(b) If \( M \) is homeomorphic to the torus or the Klein bottle, then either \( K \equiv 0 \) or \( K \) takes on both positive and negative values.\n\n(c) If \( M \) is any other compact surface, then \( K < 0 \) somewhere.
Proof. If \( M \) is orientable, the result follows immediately from the Gauss-Bonnet theorem, because a function whose integral is positive, negative, or zero must satisfy the claimed sign condition. If \( M \) is nonorientable, the result follows by applying the Gauss-Bonnet theorem to the orientable double cover \( \pi : \widetilde{M} \rightarrow M \) with the lifted metric \( \widetilde{g} = {\pi }^{ * }g \), using the fact that \( \widetilde{M} \) is the sphere if \( M = {\mathbf{P}}^{2} \), the torus if \( M \) is the Klein bottle (which is homeomorphic to the connected sum of two copies of \( {\mathbf{P}}^{2} \) ), and otherwise has \( \chi \left( \widetilde{M}\right) < 0 \) .
Yes
Exercise 9.2. Prove Corollary 9.9.
Null
No
Lemma 10.1. If \( \\Gamma \) is any smooth admissible family of curves, and \( V \) is a smooth vector field along \( \\Gamma \), then\n\n\[ \n{D}_{s}{D}_{t}V - {D}_{t}{D}_{s}V = R\\left( {S, T}\\right) V.\n\]
Proof. This is a local issue, so we can compute in any local coordinates.\n\nWriting \( V\\left( {s, t}\\right) = {V}^{i}\\left( {s, t}\\right) {\\partial }_{i} \), we compute\n\n\[ \n{D}_{t}V = \\frac{\\partial {V}^{i}}{\\partial t}{\\partial }_{i} + {V}^{i}{D}_{t}{\\partial }_{i}\n\]\n\nTherefore,\n\n\[ \n{D}_{s}{D}_{t}V = \\frac{{\\partial }^{2}{V}^{i}}{\\partial s\\partial t}{\\partial }_{i} + \\frac{\\partial {V}^{i}}{\\partial t}{D}_{s}{\\partial }_{i} + \\frac{\\partial {V}^{i}}{\\partial s}{D}_{t}{\\partial }_{i} + {V}^{i}{D}_{s}{D}_{t}{\\partial }_{i}.\n\]\n\nInterchanging \( {D}_{s} \) and \( {D}_{t} \) and subtracting, we see that all the terms except the last cancel:\n\n\[ \n{D}_{s}{D}_{t}V - {D}_{t}{D}_{s}V = {V}^{i}\\left( {{D}_{s}{D}_{t}{\\partial }_{i} - {D}_{t}{D}_{s}{\\partial }_{i}}\\right) .\n\]\n\n(10.1)\n\nNow we need to compute the commutator in parentheses. If we write the coordinate functions of \( \\Gamma \) as \( {x}^{j}\\left( {s, t}\\right) \), then\n\n\[ \nS = \\frac{\\partial {x}^{k}}{\\partial s}{\\partial }_{k};\\;T = \\frac{\\partial {x}^{j}}{\\partial t}{\\partial }_{j}.\n\]\n\nBecause \( {\\partial }_{i} \) is extendible,\n\n\[ \n{D}_{t}{\\partial }_{i} = {\\nabla }_{T}{\\partial }_{i} = \\frac{\\partial {x}^{j}}{\\partial t}{\\nabla }_{{\\partial }_{j}}{\\partial }_{i}\n\]\n\nand therefore, because \( {\\nabla }_{{\\partial }_{j}}{\\partial }_{i} \) is also extendible,\n\n\[ \n{D}_{s}{D}_{t}{\\partial }_{i} = {D}_{s}\\left( {\\frac{\\partial {x}^{j}}{\\partial t}{\\nabla }_{{\\partial }_{j}}{\\partial }_{i}}\\right)\n\]\n\n\[ \n= \\frac{{\\partial }^{2}{x}^{j}}{\\partial s\\partial t}{\\nabla }_{{\\partial }_{j}}{\\partial }_{i} + \\frac{\\partial {x}^{j}}{\\partial t}{\\nabla }_{S}\\left( {{\\nabla }_{{\\partial }_{j}}{\\partial }_{i}}\\right)\n\]\n\n\[ \n= \\frac{{\\partial }^{2}{x}^{j}}{\\partial s\\partial t}{\\nabla }_{{\\partial }_{j}}{\\partial }_{i} + \\frac{\\partial {x}^{j}}{\\partial t}\\frac{\\partial {x}^{k}}{\\partial s}{\\nabla }_{{\\partial }_{k}}{\\nabla }_{{\\partial }_{j}}{\\partial }_{i}\n\]\n\nInterchanging \( s \\leftrightarrow t \) and \( j \\leftrightarrow k \) and subtracting, we find that the first terms cancel out, and we get\n\n\[ \n{D}_{s}{D}_{t}{\\partial }_{i} - {D}_{t}{D}_{s}{\\partial }_{i} = \\frac{\\partial {x}^{j}}{\\partial t}\\frac{\\partial {x}^{k}}{\\partial s}\\left( {{\\nabla }_{{\\partial }_{k}}{\\nabla }_{{\\partial }_{j}}{\\partial }_{i} - {\\nabla }_{{\\partial }_{j}}{\\nabla }_{{\\partial }_{k}}{\\partial }_{i}}\\right)\n\]\n\n\[ \n= \\frac{\\partial {x}^{j}}{\\partial t}\\frac{\\partial {x}^{k}}{\\partial s}R\\left( {{\\partial }_{k},{\\partial }_{j}}\\right) {\\partial }_{i}\n\]\n\n\[ \n= R\\left( {S, T}\\right) {\\partial }_{i}\n\]\n\nFinally, inserting this into (10.1) yields the result.
Yes
Theorem 10.2. (The Jacobi Equation) Let \( \gamma \) be a geodesic and \( V \) a vector field along \( \gamma \) . If \( V \) is the variation field of a variation through geodesics, then \( V \) satisfies\n\n\[ \n{D}_{t}^{2}V + R\left( {V,\dot{\gamma }}\right) \dot{\gamma } = 0 \n\]
Proof. With \( S \) and \( T \) as before, the preceding lemma implies\n\n\[ \n0 = {D}_{s}{D}_{t}T \n\]\n\n\[ \n= {D}_{t}{D}_{s}T + R\left( {S, T}\right) T \n\]\n\n\[ \n= {D}_{t}{D}_{t}S + R\left( {S, T}\right) T \n\]\n\nwhere the last step follows from the symmetry lemma. Evaluating at \( s = 0 \) , where \( S\left( {0, t}\right) = V\left( t\right) \) and \( T\left( {0, t}\right) = \dot{\gamma }\left( t\right) \), we get (10.2).
Yes
Lemma 10.3. Every Jacobi field along a geodesic \( \gamma \) is the variation field of some variation of \( \gamma \) through geodesics.
Exercise 10.1. Prove Lemma 10.3. [Hint: Let \( \Gamma \left( {s, t}\right) = {\exp }_{\sigma \left( s\right) }{tW}\left( s\right) \) for a suitable curve \( \sigma \) and vector field \( W \) along \( \sigma \) .]
No
Proposition 10.4. (Existence and Uniqueness of Jacobi Fields) Let \( \gamma : I \rightarrow M \) be a geodesic, \( a \in I \), and \( p = \gamma \left( a\right) \) . For any pair of vectors \( X, Y \in {T}_{p}M \), there is a unique Jacobi field \( J \) along \( \gamma \) satisfying the initial conditions\n\n\[ J\left( a\right) = X;\;{D}_{t}J\left( a\right) = Y. \]
Proof. Choose an orthonormal basis \( \left\{ {E}_{i}\right\} \) for \( {T}_{p}M \), and extend it to a parallel orthonormal frame along all of \( \gamma \) . Writing \( J\left( t\right) = {J}^{i}\left( t\right) {E}_{i} \), we can express the Jacobi equation as\n\n\[ {\ddot{J}}^{i} + {R}_{jkl}{}^{i}{J}^{j}{\dot{\gamma }}^{k}{\dot{\gamma }}^{l} = 0. \]\n\nThis is a linear system of second-order ODEs for the \( n \) functions \( {J}^{i} \) . Making the usual substitution \( {V}^{i} = {\dot{J}}^{i} \) converts it to an equivalent first-order linear system for the \( {2n} \) unknowns \( \left\{ {{J}^{i},{V}^{i}}\right\} \) . Then Theorem 4.12 guarantees the existence and uniqueness of a solution on the whole interval \( I \) with any initial conditions \( {J}^{i}\left( a\right) = {X}^{i},{V}^{i}\left( a\right) = {Y}^{i} \) .
Yes
Corollary 10.5. Along any geodesic \( \gamma \), the set of Jacobi fields is a \( {2n} \) - dimensional linear subspace of \( \mathcal{T}\left( \gamma \right) \) .
Proof. Let \( p = \gamma \left( a\right) \) be any point on \( \gamma \), and consider the map from the set of Jacobi fields along \( \gamma \) to \( {T}_{p}M \oplus {T}_{p}M \) by sending \( J \) to \( \left( {J\left( a\right) ,{D}_{t}J\left( a\right) }\right) \) . The preceding proposition says precisely that this map is bijective.
Yes
Lemma 10.6. Let \( \gamma : I \rightarrow M \) be a geodesic, and \( a \in I \) .\n\n(a) A Jacobi field J along \( \gamma \) is normal if and only if\n\n\[ J\left( a\right) \bot \dot{\gamma }\left( a\right) \text{ and }{D}_{t}J\left( a\right) \bot \dot{\gamma }\left( a\right) . \]
Proof. Using compatibility with the metric and the fact that \( {D}_{t}\dot{\gamma } \equiv 0 \), we compute\n\n\[ \frac{{d}^{2}}{d{t}^{2}}\langle J,\dot{\gamma }\rangle = \left\langle {{D}_{t}^{2}J,\dot{\gamma }}\right\rangle \]\n\n\[ = - \langle R\left( {J,\dot{\gamma }}\right) \dot{\gamma },\dot{\gamma }\rangle \]\n\n\[ = - \operatorname{Rm}\left( {J,\dot{\gamma },\dot{\gamma },\dot{\gamma }}\right) = 0 \]\n\nby the symmetries of the curvature tensor. Thus, by elementary calculus, \( f\left( t\right) \mathrel{\text{:=}} \langle J\left( t\right) ,\dot{\gamma }\left( t\right) \rangle \) is a linear function of \( t \) . Note that \( f\left( a\right) = \langle J\left( a\right) ,\dot{\gamma }\left( a\right) \rangle \) and \( \dot{f}\left( a\right) = \left\langle {{D}_{t}J\left( a\right) ,\dot{\gamma }\left( a\right) }\right\rangle \) . Thus \( J\left( a\right) \) and \( {D}_{t}J\left( a\right) \) are orthogonal to \( \dot{\gamma }\left( a\right) \) if and only if \( f \) and its first derivative vanish at \( a \), which happens if and only if \( f \equiv 0 \) . Similarly, if \( J \) is orthogonal to \( \dot{\gamma } \) at two points, then \( f \) vanishes at two points and is therefore identically zero.
Yes
Lemma 10.7. Let \( p \in M \), let \( \left( {x}^{i}\right) \) be normal coordinates on a neighborhood \( \mathcal{U} \) of \( p \), and let \( \gamma \) be a radial geodesic starting at \( p \) . For any \( W = {W}^{i}{\partial }_{i} \in {T}_{p}M \), the Jacobi field \( J \) along \( \gamma \) such that \( J\left( 0\right) = 0 \) and \( {D}_{t}J\left( 0\right) = W \) (see Figure 10.4) is given in normal coordinates by the formula\n\n\[ J\left( t\right) = t{W}^{i}{\partial }_{i} \]
Proof. An easy computation using formula (4.10) for covariant derivatives in coordinates shows that \( J \) satisfies the specified initial conditions, so it suffices to show that \( J \) is a Jacobi field. If we set \( V = \dot{\gamma }\left( 0\right) \in {T}_{p}M \), then we know from Lemma 5.11 that \( \gamma \) is given in coordinates by the formula \( \gamma \left( t\right) = \left( {t{V}^{1},\ldots, t{V}^{n}}\right) \) . Now consider the variation \( \Gamma \) given in coordinates by\n\n\[ \Gamma \left( {s, t}\right) = \left( {t\left( {{V}^{1} + s{W}^{1}}\right) ,\ldots, t\left( {{V}^{n} + s{W}^{n}}\right) }\right) .\n\nAgain using Lemma 5.11, we see that \( \Gamma \) is a variation through geodesics. Therefore its variation field \( {\partial }_{s}\Gamma \left( {0, t}\right) \) is a Jacobi field. Differentiating \( \Gamma \left( {s, t}\right) \) with respect to \( s \) shows that its variation field is \( J\left( t\right) \).
Yes
Lemma 10.8. Suppose \( \\left( {M, g}\\right) \) is a Riemannian manifold with constant sectional curvature \( C \), and \( \\gamma \) is a unit speed geodesic in \( M \). The normal Jacobi fields along \( \\gamma \) vanishing at \( t = 0 \) are precisely the vector fields\n\n\[ J\\left( t\\right) = u\\left( t\\right) E\\left( t\\right) \]\n\nwhere \( E \) is any parallel normal vector field along \( \\gamma \), and \( u\\left( t\\right) \) is given by\n\n\[ u\\left( t\\right) = \\left\\{ \\begin{array}{ll} t, & C = 0 \\\\ R\\sin \\frac{t}{R}, & C = \\frac{1}{{R}^{2}} > 0 \\\\ R\\sinh \\frac{t}{R}, & C = - \\frac{1}{{R}^{2}} < 0 \\end{array}\\right. \]
Proof. Since \( g \) has constant curvature, its curvature endomorphism is given by the formula of Lemma 8.10:\n\n\[ R\\left( {X, Y}\\right) Z = C\\left( {\\langle Y, Z\\rangle X-\\langle X, Z\\rangle Y}\\right) \]\n\nSubstituting this into the Jacobi equation, we find that a normal Jacobi field \( J \) satisfies\n\n\[ 0 = {D}_{t}^{2}J + C\\left( {\\langle \\dot{\\gamma },\\dot{\\gamma }\\rangle J-\\langle J,\\dot{\\gamma }\\rangle \\dot{\\gamma }}\\right) \]\n\n\[ = {D}_{t}^{2}J + {CJ} \]\n\nwhere we have used the facts that \( {\\left| \\dot{\\gamma }\\right| }^{2} = 1 \) and \( \\langle J,\\dot{\\gamma }\\rangle = 0 \).\n\nSince (10.7) says that the second covariant derivative of \( J \) is a multiple of \( J \) itself, it is reasonable to try to construct a solution by choosing a parallel normal vector field \( E \) along \( \\gamma \) and setting \( J\\left( t\\right) = u\\left( t\\right) E\\left( t\\right) \) for some function \( u \) to be determined. Plugging this into (10.7), we find that \( J \) is a Jacobi field provided \( u \) is a solution to the differential equation\n\n\[ \\ddot{u}\\left( t\\right) + {Cu}\\left( t\\right) = 0 \]\n\nIt is an easy matter to solve this ODE explicitly. In particular, the solutions satisfying \( u\\left( 0\\right) = 0 \) are constant multiples of the functions given in (10.6). This construction yields all the normal Jacobi fields vanishing at 0, since there is an \( \\left( {n - 1}\\right) \)-dimensional space of them, and the space of parallel normal vector fields has the same dimension.
Yes
Proposition 10.9. Suppose \( \left( {M, g}\right) \) is a Riemannian manifold with constant sectional curvature \( C \) . Let \( \left( {x}^{i}\right) \) be Riemannian normal coordinates on a normal neighborhood \( \mathcal{U} \) of \( p \in M \), let \( {\left| \cdot \right| }_{\bar{q}} \) be the Euclidean norm in these coordinates, and let \( r \) be the radial distance function. For any \( q \in \mathcal{U} - \{ p\} \) and \( V \in {T}_{q}M \), write \( V = {V}^{\top } + {V}^{ \bot } \), where \( {V}^{\top } \) is tangent to the sphere \( \{ r = \) constant \( \} \) through \( q \) and \( {V}^{ \bot } \) is a multiple of \( \partial /\partial r \) . The metric \( g \) can be written\n\n\[ g\left( {V, V}\right) = \left\{ \begin{array}{ll} {\left| {V}^{ \bot }\right| }_{\bar{g}}^{2} + {\left| {V}^{\top }\right| }_{\bar{g}}^{2}, & K = 0; \\ {\left| {V}^{ \bot }\right| }_{\bar{g}}^{2} + \frac{{R}^{2}}{{r}^{2}}\left( {{\sin }^{2}\frac{r}{R}}\right) {\left| {V}^{\top }\right| }_{\bar{g}}^{2}, & C = \frac{1}{{R}^{2}} > 0; \\ {\left| {V}^{ \bot }\right| }_{\bar{g}}^{2} + \frac{{R}^{2}}{{r}^{2}}\left( {{\sinh }^{2}\frac{r}{R}}\right) {\left| {V}^{\top }\right| }_{\bar{g}}^{2}, & C = - \frac{1}{{R}^{2}} < 0. \end{array}\right. \]
Proof. By the Gauss lemma, the decomposition \( V = {V}^{\top } + {V}^{ \bot } \) is orthogonal, so \( {\left| V\right| }_{g}^{2} = {\left| {V}^{ \bot }\right| }_{g}^{2} + {\left| {V}^{\top }\right| }_{g}^{2} \) . Since \( \partial /\partial r \) is a unit vector in both the \( g \) and \( \bar{g} \) norms, it is immediate that \( {\left| {V}^{ \bot }\right| }_{g} = {\left| {V}^{ \bot }\right| }_{\bar{g}} \) . Thus we need only compute \( {\left| {V}^{\top }\right| }_{g} \).\n\nSet \( X = {V}^{\top } \), and let \( \gamma \) denote the unit speed radial geodesic from \( p \) to \( q \) . By Lemma 10.7, \( X \) is the value of a Jacobi field \( J \) along \( \gamma \) that vanishes at \( p \) (Figure 10.5), namely \( X = J\left( r\right) \), where \( r = d\left( {p, q}\right) \) and\n\n\[ J\left( t\right) = \frac{t}{r}{X}^{i}{\partial }_{i} \]\n\n(10.9)\n\nBecause \( J \) is orthogonal to \( \dot{\gamma } \) at \( p \) and \( q \), it is normal by Lemma 10.6.\n\nNow \( J \) can also be written in the form \( J\left( t\right) = u\left( t\right) E\left( t\right) \) as in Lemma 10.8. In this representation,\n\n\[ {D}_{t}J\left( 0\right) = \dot{u}\left( 0\right) E\left( 0\right) = E\left( 0\right) \]\nsince \( \dot{u}\left( 0\right) = 1 \) in each of the cases of (10.6). Therefore, since \( E \) is parallel and thus of constant length,\n\n\[ {\left| X\right| }^{2} = {\left| J\left( r\right) \right| }^{2} = {\left| u\left( r\right) \right| }^{2}{\left| E\left( r\right) \right| }^{2} = {\left| u\left( r\right) \right| }^{2}{\left| E\left( 0\right) \right| }^{2} = {\left| u\left( r\right) \right| }^{2}{\left| {D}_{t}J\left( 0\right) \right| }^{2}. \]\n\n(10.10)\n\nObserve that \( {D}_{t}J\left( 0\right) = \left( {1/r}\right) {\left. {X}^{i}{\partial }_{i}\right| }_{p} \) by (10.9). Since \( g \) agrees with \( \bar{g} \) at \( p \) , we have\n\n\[ \left| {{D}_{t}J\left( 0\right) }\right| = {\left. \frac{1}{r}{\left| {X}^{i}{\partial }_{i}\right| }_{p}\right| }_{g} = \frac{1}{r}{\left| X\right| }_{\bar{g}} \]\n\nInserting this into (10.10) and using formula (10.6) for \( u\left( r\right) \) completes the proof.
Yes
Proposition 10.10. (Local Uniqueness of Constant Curvature Metrics) Let \( \left( {M, g}\right) \) and \( \left( {\widetilde{M},\widetilde{g}}\right) \) be Riemannian manifolds with constant sectional curvature \( C \) . For any points \( p \in M,\widetilde{p} \in \widetilde{M} \), there exist neighborhoods \( \mathcal{U} \) of \( p \) and \( \widetilde{\mathcal{U}} \) of \( \widetilde{p} \) and an isometry \( F : \mathcal{U} \rightarrow \widetilde{\mathcal{U}} \) .
Proof. Choose \( p \in M \) and \( \widetilde{p} \in \widetilde{M} \), and let \( \mathcal{U} \) and \( \widetilde{\mathcal{U}} \) be geodesic balls of small radius \( \varepsilon \) around \( p \) and \( \widetilde{p} \), respectively. Riemannian normal coordinates give maps \( \varphi : \mathcal{U} \rightarrow {B}_{\varepsilon }\left( 0\right) \subset {\mathbf{R}}^{n} \) and \( \widetilde{\varphi } : \widetilde{\mathcal{U}} \rightarrow {B}_{\varepsilon }\left( 0\right) \subset {\mathbf{R}}^{n} \), under which both metrics are given by (10.8) (Figure 10.6). Therefore \( {\widetilde{\varphi }}^{-1} \circ \varphi \) is the required local isometry.
Yes
Proposition 10.11. Suppose \( p \in M, V \in {T}_{p}M \), and \( q = {\exp }_{p}V \) . Then \( {\exp }_{p} \) is a local diffeomorphism in a neighborhood of \( V \) if and only if \( q \) is not conjugate to \( p \) along the geodesic \( \gamma \left( t\right) = {\exp }_{p}{tV},\;t \in \left\lbrack {0,1}\right\rbrack \) .
Proof. By the inverse function theorem, \( {\exp }_{p} \) is a local diffeomorphism near \( V \) if and only if \( {\left( {\exp }_{p}\right) }_{ * } \) is an isomorphism at \( V \), and by dimensional considerations, this occurs if and only if \( {\left( {\exp }_{p}\right) }_{ * } \) is injective at \( V \) .\n\nIdentifying \( {T}_{V}\left( {{T}_{p}M}\right) \) with \( {T}_{p}M \) as usual, we can compute the push-forward \( {\left( {\exp }_{p}\right) }_{ * } \) at \( V \) as follows:\n\n\[ \n{\left. {\left( {\exp }_{p}\right) }_{ * }W = \frac{d}{ds}\right| }_{s = 0}{\exp }_{p}\left( {V + {sW}}\right) \n\]\n\nTo compute this, we define a variation of \( \gamma \) through geodesics (Figure 10.9) by\n\n\[ \n{\Gamma }_{W}\left( {s, t}\right) = {\exp }_{p}t\left( {V + {sW}}\right) \n\]\n\nThen the variation field \( {J}_{W}\left( t\right) = {\partial }_{s}{\Gamma }_{W}\left( {0, t}\right) \) is a Jacobi field along \( \gamma \), and\n\n\[ \n{J}_{W}\left( 1\right) = {\left( {\exp }_{p}\right) }_{ * }W \n\]\n\nSince \( W \in {T}_{p}M \) is arbitrary, there is an \( n \) -dimensional space of such Jacobi fields, and so these are all the Jacobi fields along \( \gamma \) that vanish at \( p \) . (If \( \gamma \) is contained in a normal neighborhood, these are just the Jacobi fields of the form (10.4) in normal coordinates.)\n\nTherefore, \( {\left( {\exp }_{p}\right) }_{ * } \) fails to be an isomorphism at \( V \) when there is a vector \( W \) such that \( {\left( ex{p}_{p}\right) }_{ * }W = 0 \), which occurs precisely when there is a Jacobi field \( {J}_{W} \) along \( \gamma \) with \( {J}_{W}\left( 0\right) = {J}_{W}\left( q\right) = 0 \) .
Yes
Corollary 10.13. If \( \Gamma \) is a proper variation of a unit speed geodesic \( \gamma \) whose variation field is a proper normal vector field \( V \), the second variation of \( L\left( {\Gamma }_{s}\right) \) is \( I\left( {V, V}\right) \) . In particular, if \( \gamma \) is minimizing, then \( I\left( {V, V}\right) \geq 0 \) for any proper normal vector field along \( \gamma \) .
Null
No
Proposition 10.14. For any pair of proper normal vector fields \( V, W \) along a geodesic segment \( \gamma \) ,
Proof. On any subinterval \( \left\lbrack {{a}_{i - 1},{a}_{i}}\right\rbrack \) where \( V \) and \( W \) are smooth,\n\n\[ \frac{d}{dt}\left\langle {{D}_{t}V, W}\right\rangle = \left\langle {{D}_{t}^{2}V, W}\right\rangle + \left\langle {{D}_{t}V,{D}_{t}W}\right\rangle . \]\n\nThus, by the fundamental theorem of calculus,\n\n\[ {\int }_{{a}_{i - 1}}^{{a}_{i}}\left\langle {{D}_{t}V,{D}_{t}W}\right\rangle {dt} = - {\int }_{{a}_{i - 1}}^{{a}_{i}}\left\langle {{D}_{t}^{2}V, W}\right\rangle + {\left. \left\langle {D}_{t}V, W\right\rangle \right| }_{{a}_{i - 1}}^{{a}_{i}}. \]\n\nSumming over \( i \), and noting that \( W \) is continuous at \( t = {a}_{i} \) and \( W\left( a\right) = \) \( W\left( b\right) = 0 \), we get (10.16).
Yes
If \( \gamma \) is a geodesic segment from \( p \) to \( q \) that has an interior conjugate point to \( p \), then there exists a proper normal vector field \( X \) along \( \gamma \) such that \( I\left( {X, X}\right) < 0 \) . In particular, \( \gamma \) is not minimizing.
Proof. Suppose \( \gamma : \left\lbrack {0, b}\right\rbrack \rightarrow M \) is a unit speed parametrization of \( \gamma \), and \( \gamma \left( a\right) \) is conjugate to \( \gamma \left( 0\right) \) for some \( 0 < a < b \) . This means there is a nontrivial normal Jacobi field \( J \) along \( {\left. \gamma \right| }_{\left\lbrack 0, a\right\rbrack } \) that vanishes at \( t = 0 \) and \( t = a \) . Define a vector field \( V \) along all of \( \gamma \) by\n\n\[ V\left( t\right) = \left\{ \begin{array}{ll} J\left( t\right) , & t \in \left\lbrack {0, a}\right\rbrack \\ 0, & t \in \left\lbrack {a, b}\right\rbrack \end{array}\right. \]\n\nThis is a proper, normal, piecewise smooth vector field along \( \gamma \) .\n\nLet \( W \) be a smooth proper normal vector field along \( \gamma \) such that \( W\left( b\right) \) is equal to the jump \( \Delta {D}_{t}V \) at \( t = b \) (Figure 10.10). Such a vector field is easily constructed in local coordinates and extended to all of \( \gamma \) by a bump function. Note that \( \Delta {D}_{t}V = - {D}_{t}J\left( b\right) \) is not zero, because otherwise \( J \) would be a Jacobi field satisfying \( J\left( b\right) = {D}_{t}J\left( b\right) = 0 \), and thus would be identically zero.\n\nFor small positive \( \varepsilon \), let \( {X}_{\varepsilon } = V + {\varepsilon W} \) . Then\n\n\[ I\left( {{X}_{\varepsilon },{X}_{\varepsilon }}\right) = I\left( {V + {\varepsilon W}, V + {\varepsilon W}}\right) \]\n\n\[ = I\left( {V, V}\right) + {2\varepsilon I}\left( {V, W}\right) + {\varepsilon }^{2}I\left( {W, W}\right) . \]\n\nSince \( V \) satisfies the Jacobi equation on each subinterval \( \left\lbrack {0, a}\right\rbrack \) and \( \left\lbrack {a, b}\right\rbrack \) , and \( V\left( a\right) = 0,\left( {10.16}\right) \) gives\n\n\[ I\left( {V, V}\right) = - \left\langle {\Delta {D}_{t}V, V\left( a\right) }\right\rangle = 0. \]\n\nSimilarly,\n\n\[ I\left( {V, W}\right) = - \left\langle {\Delta {D}_{t}V, W\left( b\right) }\right\rangle = - {\left| W\left( b\right) \right| }^{2}. \]\n\nThus\n\n\[ I\left( {{X}_{\varepsilon },{X}_{\varepsilon }}\right) = - {2\varepsilon }{\left| W\left( b\right) \right| }^{2} + {\varepsilon }^{2}I\left( {W, W}\right) . \]\n\nIf we choose \( \varepsilon \) small enough, this is strictly negative.
Yes
Theorem 11.1. (Sturm Comparison Theorem) Suppose \( u \) and \( v \) are differentiable real-valued functions on \( \left\lbrack {0, T}\right\rbrack \), twice differentiable on \( \left( {0, T}\right) \) , and \( u > 0 \) on \( \left( {0, T}\right) \) . Suppose further that \( u \) and \( v \) satisfy\n\n\[ \ddot{u}\left( t\right) + a\left( t\right) u\left( t\right) = 0 \]\n\n\[ \ddot{v}\left( t\right) + a\left( t\right) v\left( t\right) \geq 0 \]\n\n\[ u\left( 0\right) = v\left( 0\right) = 0,\;\dot{u}\left( 0\right) = \dot{v}\left( 0\right) > 0 \]\n\nfor some function \( a : \left\lbrack {0, T}\right\rbrack \rightarrow \mathbf{R} \) . Then \( v\left( t\right) \geq u\left( t\right) \) on \( \left\lbrack {0, T}\right\rbrack \) .
Proof. Consider the function \( f\left( t\right) = v\left( t\right) /u\left( t\right) \) defined on \( \left( {0, T}\right) \) . It follows from l’Hôpital’s rule that \( \mathop{\lim }\limits_{{t \rightarrow 0}}f\left( t\right) = \dot{v}\left( 0\right) /\dot{u}\left( 0\right) = 1 \) . Since \( f \) is differentiable on \( \left( {0, T}\right) \), if we could show that \( \dot{f} \geq 0 \) there it would follow from elementary calculus that \( f \geq 1 \) and therefore \( v \geq u \) on \( \left( {0, T}\right) \), and by continuity also on \( \left\lbrack {0, T}\right\rbrack \) . Differentiating,\n\n\[ \frac{d}{dt}\left( \frac{v}{u}\right) = \frac{\dot{v}u - v\dot{u}}{{u}^{2}} \]\n\nThus to show \( \dot{f} \geq 0 \) it would suffice to show \( \dot{v}u - v\dot{u} \geq 0 \) . Since \( \dot{v}\left( 0\right) u\left( 0\right) - \) \( v\left( 0\right) \dot{u}\left( 0\right) = 0 \), we need only show this expression has nonnegative derivative. Differentiating again and substituting the ODE for \( u \) ,\n\n\[ \frac{d}{dt}\left( {\dot{v}u - v\dot{u}}\right) = \ddot{v}u + \dot{v}\dot{u} - \dot{v}\dot{u} - v\ddot{u} = \ddot{v}u + {avu} \geq 0. \]\n\nThis proves the theorem.
Yes
Corollary 11.3. (Conjugate Point Comparison Theorem) Suppose all sectional curvatures of \( \left( {M, g}\right) \) are bounded above by a constant \( C \) . If\n\n\( C \leq 0 \), then no point of \( M \) has conjugate points along any geodesic. If \( C = 1/{R}^{2} > 0 \), then the first conjugate point along any geodesic occurs at a distance of at least \( {\pi R} \) .
Proof. If \( C \leq 0 \), the Jacobi field comparison theorem implies that any nontrivial normal Jacobi field vanishing at \( t = 0 \) satisfies \( \left| {J\left( t\right) }\right| > 0 \) for all \( t > 0 \) . Similarly, if \( C > 0 \), then \( \left| {J\left( t\right) }\right| \geq \left( \text{constant}\right) \sin \left( {t/R}\right) > 0 \) for \( 0 < t < {\pi R} \) .
Yes
Corollary 11.4. (Metric Comparison Theorem) Suppose all sectional curvatures of \( \left( {M, g}\right) \) are bounded above by a constant \( C \) . In any normal coordinate chart, \( g\left( {V, V}\right) \geq {g}_{C}\left( {V, V}\right) \), where \( {g}_{C} \) is the constant curvature metric given by formula (10.8).
Proof. Decomposing a vector \( V \) into components \( {V}^{\top } \) tangent to the geodesic sphere and \( {V}^{ \bot } \) tangent to the radial geodesics as in the proof of Proposition 10.9 gives\n\n\[ g\left( {V, V}\right) = g\left( {{V}^{ \bot },{V}^{ \bot }}\right) + g\left( {{V}^{\top },{V}^{\top }}\right) .\n\]\n\nJust as in that proof, \( g\left( {{V}^{ \bot },{V}^{ \bot }}\right) = \bar{g}\left( {{V}^{ \bot },{V}^{ \bot }}\right) = {g}_{C}\left( {{V}^{ \bot },{V}^{ \bot }}\right) \) . Also, \( {V}^{\top } \) is the value of some normal Jacobi field vanishing at \( t = 0 \), so the Jacobi field comparison theorem gives \( g\left( {{V}^{\top },{V}^{\top }}\right) \geq {g}_{C}\left( {{V}^{\top },{V}^{\top }}\right) \) .
Yes
Theorem 11.5. (The Cartan-Hadamard Theorem) If \( M \) is a complete, connected manifold all of whose sectional curvatures are nonpositive, then for any point \( p \in M,{\exp }_{p} : {T}_{p}M \rightarrow M \) is a covering map. In particular, the universal covering space of \( M \) is diffeomorphic to \( {\mathbf{R}}^{n} \) . If \( M \) is simply connected, then \( M \) itself is diffeomorphic to \( {\mathbf{R}}^{n} \) .
Proof. The assumption of nonpositive curvature guarantees that \( p \) has no conjugate points along any geodesic, which can be shown by using either the conjugate point comparison theorem above or Problem 10-2. Therefore, by Proposition 10.11, \( {\exp }_{p} \) is a local diffeomorphism on all of \( {T}_{p}M \) . Let \( \widetilde{g} \) be the (variable-coefficient) 2-tensor field \( {\exp }_{p}^{ * }g \) defined on \( {T}_{p}M \) . Because \( {\exp }_{p}^{ * } \) is everywhere nonsingular, \( \widetilde{g} \) is a Riemannian metric, and \( {\exp }_{p} : \left( {{T}_{p}M,\widetilde{g}}\right) \rightarrow \left( {M, g}\right) \) is a local isometry. It then follows from Lemma 11. \( \dot{6} \) below that \( {\exp }_{p} \) is a covering map. The remaining statements of the theorem follow immediately from uniqueness of the universal covering space.
Yes
Lemma 11.6. Suppose \( \widetilde{M} \) and \( M \) are connected Riemannian manifolds, with \( \widetilde{M} \) complete, and \( \pi : \widetilde{M} \rightarrow M \) is a local isometry. Then \( M \) is complete and \( \pi \) is a covering map.
Proof. A fundamental property of covering maps is the path-lifting property: any continuous path \( \gamma \) in \( M \) lifts to a path \( \widetilde{\gamma } \) in \( \widetilde{M} \) such that \( \pi \circ \widetilde{\gamma } = \gamma \) . We begin by proving that \( \pi \) possesses the path-lifting property for geodesics: If \( p \in M,\widetilde{p} \in {\pi }^{-1}\left( p\right) \), and \( \gamma : I \rightarrow M \) is a geodesic starting at \( p \), then \( \gamma \) has a unique lift starting at \( \widetilde{p} \) (Figure 11.1). The lifted curve is necessarily also a geodesic because \( \pi \) is a local isometry.\n\nTo prove the path-lifting property for geodesics, let \( V = \dot{\gamma }\left( 0\right) \) and \( \widetilde{V} = \) \( {\pi }_{ * }^{-1}\dot{\gamma }\left( 0\right) \in {T}_{\widetilde{p}}\widetilde{M} \) (which is well defined because \( {\pi }_{ * } \) is an isomorphism at each point), and let \( \widetilde{\gamma } \) be the geodesic in \( \widetilde{M} \) with initial point \( \widetilde{p} \) and initial velocity \( \widetilde{V} \) . Because \( \widetilde{M} \) is complete, \( \widetilde{\gamma } \) is defined for all time. Since \( \pi \) is a local isometry, it takes geodesics to geodesics; and since by construction \( \pi \left( {\widetilde{\gamma }\left( 0\right) }\right) = \gamma \left( 0\right) \) and \( {\pi }_{ * }\dot{\widetilde{\gamma }}\left( 0\right) = \dot{\gamma }\left( 0\right) \), we must have \( \pi \circ \widetilde{\gamma } = \gamma \) on \( I \) . In particular, \( \pi \circ \widetilde{\gamma } \) is a geodesic defined for all \( t \) that coincides with \( \gamma \) on \( I \) , so \( \gamma \) extends to all of \( \mathbf{R} \) and thus \( M \) is complete.\n\nNext we show that \( \pi \) is surjective. Choose some point \( \widetilde{p} \in \widetilde{M} \), write \( p = \) \( \pi \left( \widetilde{p}\right) \), and let \( q \in M \) be arbitrary. Because \( M \) is connected and complete, there is a minimizing geodesic segment \( \gamma \) from \( p \) to \( q \) . Letting \( \widetilde{\gamma } \) be the lift of \( \gamma \) starting at \( \widetilde{p} \) and \( r = d\left( {p, q}\right) \), we have \( \pi \left( {\widetilde{\gamma }\left( r\right) }\right) = \gamma \left( r\right) = q \), so \( q \) is in the image of \( \pi \) .
Yes
Theorem 11.8. (Myers’s Theorem) Suppose \( M \) is a complete, connected Riemannian n-manifold whose Ricci tensor satisfies the following inequality for all \( V \in {TM} \) :\n\n\[ \operatorname{Rc}\left( {V, V}\right) \geq \frac{n - 1}{{R}^{2}}{\left| V\right| }^{2} \]\n\nThen \( M \) is compact, with a finite fundamental group, and diameter at most \( {\pi R} \) .
Proof. As in the proof of Bonnet's theorem, it suffices to prove the diameter estimate. As before, let \( \gamma \) be a minimizing unit speed geodesic segment of\n\nlength \( L > {\pi R} \) . Let \( \left( {{E}_{1},\ldots ,{E}_{n}}\right) \) be a parallel orthonormal frame along \( \gamma \) such that \( {E}_{n} = \dot{\gamma } \), and for each \( i = 1,\ldots, n - 1 \) let \( {V}_{i} \) be the proper normal vector field\n\n\[ {V}_{i}\left( t\right) = \left( {\sin \frac{\pi t}{L}}\right) {E}_{i}\left( t\right) \]\n\nBy the same computation as before,\n\n\[ I\left( {{V}_{i},{V}_{i}}\right) = {\int }_{0}^{L}\left( {{\sin }^{2}\frac{\pi t}{L}}\right) \left( {\frac{{\pi }^{2}}{{L}^{2}} - \operatorname{Rm}\left( {{E}_{i},\dot{\gamma },\dot{\gamma },{E}_{i}}\right) }\right) {dt}. \]\n\n(11.1)\n\nIn this case, we cannot conclude that each of these terms is negative. However, because \( \left\{ {E}_{i}\right\} \) is an orthonormal frame, the Ricci tensor at points along \( \gamma \) is given by\n\n\[ \operatorname{Rc}\left( {\dot{\gamma },\dot{\gamma }}\right) = \mathop{\sum }\limits_{{i = 1}}^{n}\operatorname{Rm}\left( {{E}_{i},\dot{\gamma },\dot{\gamma },{E}_{i}}\right) = \mathop{\sum }\limits_{{i = 1}}^{{n - 1}}\operatorname{Rm}\left( {{E}_{i},\dot{\gamma },\dot{\gamma },{E}_{i}}\right) \]\n\n(because \( {Rm}\left( {{E}_{n},\dot{\gamma },\dot{\gamma },{E}_{n}}\right) = {Rm}\left( {\dot{\gamma },\dot{\gamma },\dot{\gamma },\dot{\gamma }}\right) = 0 \) ). Therefore, summing (11.1) over \( i \) gives\n\n\[ \mathop{\sum }\limits_{{i = 1}}^{{n - 1}}I\left( {{V}_{i},{V}_{i}}\right) = {\int }_{0}^{L}\left( {{\sin }^{2}\frac{\pi t}{L}}\right) \left( {\left( {n - 1}\right) \frac{{\pi }^{2}}{{L}^{2}} - \operatorname{Rc}\left( {\dot{\gamma },\dot{\gamma }}\right) }\right) {dt} \]\n\n\[ \leq {\int }_{0}^{L}\left( {{\sin }^{2}\frac{\pi t}{L}}\right) \left( {\frac{\left( {n - 1}\right) {\pi }^{2}}{{L}^{2}} - \frac{n - 1}{{R}^{2}}}\right) {dt} < 0. \]\n\nThis means at least one of the terms \( I\left( {{V}_{i},{V}_{i}}\right) \) must be negative, and again we have a contradiction to \( \gamma \) being minimizing.
Yes
Theorem 11.9. (Rauch Comparison Theorem) Let \( M \) and \( \widetilde{M} \) be Riemannian manifolds, let \( \gamma : \left\lbrack {0, T}\right\rbrack \rightarrow M \) and \( \widetilde{\gamma } : \left\lbrack {0, T}\right\rbrack \rightarrow \widetilde{M} \) be unit speed geodesic segments such that \( \widetilde{\gamma }\left( 0\right) \) has no conjugate points along \( \widetilde{\gamma } \), and let \( J,\widetilde{J} \) be normal Jacobi fields along \( \gamma \) and \( \widetilde{\gamma } \) such that \( J\left( 0\right) = \widetilde{J}\left( 0\right) = 0 \) and \( \left| {{D}_{t}J\left( 0\right) }\right| = \left| {{\widetilde{D}}_{t}\widetilde{J}\left( 0\right) }\right| \) (Figure 11.5). Suppose that the sectional curvatures of \( M \) and \( \widetilde{M} \) satisfy \( K\left( \Pi \right) \leq \widetilde{K}\left( \widetilde{\Pi }\right) \) whenever \( \Pi \subset {T}_{\gamma \left( t\right) }M \) is a 2- plane containing \( \dot{\gamma }\left( t\right) \) and \( \widetilde{\Pi } \subset {T}_{\widetilde{\gamma }\left( t\right) }\widetilde{M} \) is a 2-plane containing \( \dot{\widetilde{\gamma }}\left( t\right) \) . Then \( \left| {J\left( t\right) }\right| \geq \left| {\widetilde{J}\left( t\right) }\right| \) for all \( t \in \left\lbrack {0, T}\right\rbrack \) .
You can find proofs in [dC92], [CE75], and [Spi79, volume 4]. Letting \( \widetilde{M} \) be one of our constant curvature model spaces, we recover the Jacobi field comparison theorem above. On the other hand, if instead we take \( M \) to have constant curvature, we get the same result with the inequalities reversed.
No
Theorem 11.10. (The Sphere Theorem) Suppose \( M \) is a complete, simply-connected, Riemannian n-manifold that is strictly \( \frac{1}{4} \) -pinched. Then \( M \) is homeomorphic to \( {\mathbf{S}}^{n} \) .
The proof, which can be found in [CE75] or [dC92], is an elaborate application of the Rauch comparison theorem together with the Morse index theorem mentioned in Chapter 10. This result is sharp, at least in even dimensions, because the Fubini-Study metrics on complex projective spaces are \( \frac{1}{4} \) -pinched (Problem 8-12).
Yes
Theorem 11.11. (Hamilton) Suppose \( M \) is a simply-connected compact Riemannian 3-manifold with strictly positive Ricci curvature. Then \( M \) is diffeomorphic to \( {\mathbf{S}}^{3} \) .
Null
No
Corollary 11.13. (Classification of Constant Curvature Metrics) Suppose \( M \) is a complete, connected Riemannian manifold with constant sectional curvature. Then \( M \) is isometric to \( \widetilde{M}/\Gamma \), where \( \widetilde{M} \) is one of the constant curvature model spaces \( {\mathbf{R}}^{n},{\mathbf{S}}_{R}^{n} \), or \( {\mathbf{H}}_{R}^{n} \), and \( \Gamma \) is a discrete subgroup of \( \mathcal{J}\left( \widetilde{M}\right) \), isomorphic to \( {\pi }_{1}\left( M\right) \), and acting freely and properly discontinuously on \( \widetilde{M} \) .
Proof. If \( \pi : \widetilde{M} \rightarrow M \) is the universal covering space of \( M \) with the lifted metric \( \widetilde{g} = {\pi }^{ * }g \), the preceding theorem shows that \( \left( {\widetilde{M},\widetilde{g}}\right) \) is isometric to one of the model spaces. From covering space theory [Sie92, Mas67] it follows that the group \( \Gamma \) of covering transformations is isomorphic to \( {\pi }_{1}\left( M\right) \) and acts freely and properly discontinuously on \( \widetilde{M} \), and \( M \) is diffeomorphic to the quotient \( \widetilde{M}/\Gamma \) . Moreover, if \( \varphi \) is any covering transformation, \( \pi \circ \varphi = \pi \) , and so \( {\varphi }^{ * }\widetilde{g} = {\varphi }^{ * }{\pi }^{ * }g = {\pi }^{ * }g = \widetilde{g} \), so \( \Gamma \) acts by isometries. Finally, suppose \( \left\{ {\varphi }_{i}\right\} \subset \Gamma \) is an infinite set with an accumulation point in \( \mathcal{J}\left( \widetilde{M}\right) \) . Since the action of \( \Gamma \) is fixed-point free, for any point \( \widetilde{p} \in \widetilde{M} \) the set \( \left\{ {{\varphi }_{i}\left( \widetilde{p}\right) }\right\} \) is infinite, and by continuity of the action it has an accumulation point in \( \widetilde{M} \) . But this is impossible, since the points \( \left\{ {{\varphi }_{i}\left( \widetilde{p}\right) }\right\} \) all project to the same point in \( M \), and so form a discrete set. Thus \( \Gamma \) is discrete in \( \mathcal{J}\left( \widetilde{M}\right) \) .
Yes
Lemma 1. Let \( S \) have density \( \alpha \) and \( 0 \in S \) . Then \( S \oplus S \) has density at least \( {2\alpha } - {\alpha }^{2} \) .
Proof. All the gaps in the set \( S \) are covered in part by the translation of \( S \) by the term of \( S \) just before this gap. Hence, at least the fraction \( \alpha \) of this gap gets covered. So from this covering we have density \( \alpha \) from \( S \) itself and \( \alpha \) times the gaps. Altogether, then, we indeed have \( \alpha + \alpha \left( {1 - \alpha }\right) = {2\alpha } - {\alpha }^{2} \), as claimed.
Yes
Lemma 2. If \( S \) has density \( \alpha > \frac{1}{2} \), then \( S \oplus S \) contains all the positive integers.
Proof. Fix an integer \( n \) which is arbitrary, let \( A \) be the subset of \( S \) which lies \( \leq n \), and let \( B \) be the set of all \( n \) minus elements of \( S \) . Since \( A \) contains more than \( n/2 \) elements and \( B \) contains at least \( n/2 \) elements, the Pigeonhole principle guarantees that they overlap. So suppose they overlap at \( k \) . Since \( k \in A \), we get \( k \in S \), and since \( k \in B \), we get \( n - k \in S \) . These are the two elements of \( S \) which sum to \( n \).
Yes
Lemma 3. Let \( k > 1 \) be a fixed integer. There exists a \( {C}_{1} \) such that, for any positive integers \( N, a, b \) with \( \left( {a, b}\right) = 1 \) , \[ \left| {\mathop{\sum }\limits_{{n = 1}}^{N}e\left( {\frac{a}{b}{n}^{k}}\right) }\right| \leq {C}_{1}{N}^{1 + o\left( 1\right) }{b}^{-{2}^{1 - k}}. \]
Null
No
Lemma 4. There exists \( \epsilon > 0 \) and \( {C}_{2} \) such that, throughout any interval \( {I}_{a, b, N} \) , \[ \left| {\mathop{\sum }\limits_{{n = 1}}^{N}e\left( {x{n}^{k}}\right) }\right| \leq \frac{{C}_{2}N}{{\left( b + j\right) }^{\epsilon }} \]
Proof. This is almost trivial if \( b > {N}^{2/3} \), for, since the derivative of \( \left| {\mathop{\sum }\limits_{{n = 1}}^{N}e\left( {x{n}^{k}}\right) }\right| \) is bounded by \( {2\pi }{N}^{k + 1} \) , \[ \left| {\mathop{\sum }\limits_{{n = 1}}^{N}e\left( {x{n}^{k}}\right) }\right| \leq \left| {\mathop{\sum }\limits_{{n = 1}}^{N}e\left( {\frac{a}{b}{n}^{k}}\right) }\right| + \left| {x - \frac{a}{b}}\right| {2\pi }{N}^{k + 1} \] \[ \leq \frac{{N}^{1 + o\left( 1\right) }}{{b}^{\frac{1}{{2}^{k - 1}}}} + \frac{{2\pi }{N}^{3/2}}{b} \leq \frac{{N}^{1 + o\left( 1\right) }}{{b}^{\frac{1}{{2}^{k - 1}}}} + \frac{2\pi N}{{b}^{1/4}}, \] by \( \mathrm{C} \), which gives the result, since \( j = 0 \) automatically. Assume therefore that \( b \leq {N}^{2/3} \), and note the following two simple facts (A) and (B). For details see [K. Knopp, Theory and Application of Infinite Series, Blackie & Sons, Glasgow, 1946.] and [G. Pólya und G. Szegö, Aufgaben und Lehrsätze aus der Analysis, Dover Publications, New York 1945, Vol. 1, Part II, p. 37]. Q.E.D.
Yes
Theorem 1.1.4 Let \( {A}_{0},{A}_{1},{A}_{2},\ldots \) be countable sets. Then their union \( A = \mathop{\bigcup }\limits_{0}^{\infty }{A}_{n} \) is countable.
Proof. For each \( n \), choose an enumeration \( {a}_{n0},{a}_{n1},{a}_{n2},\ldots \) of \( {A}_{n} \) . We enumerate \( A = \mathop{\bigcup }\limits_{n}{A}_{n} \) following the above diagonal method.
Yes
Theorem 1.1.4 Let \( {A}_{0},{A}_{1},{A}_{2},\ldots \) be countable sets. Then their union \( A = \mathop{\bigcup }\limits_{0}^{\infty }{A}_{n} \) is countable.
Proof. For each \( n \), choose an enumeration \( {a}_{n0},{a}_{n1},{a}_{n2},\ldots \) of \( {A}_{n} \) . We enumerate \( A = \mathop{\bigcup }\limits_{n}{A}_{n} \) following the above diagonal method.
Yes
Theorem 1.1.8 (Cantor) For any two real numbers \( a, b \) with \( a < b \), the interval \( \left\lbrack {a, b}\right\rbrack \) is uncountable.
Proof. (Cantor) Let \( \left( {a}_{n}\right) \) be a sequence in \( \left\lbrack {a, b}\right\rbrack \) . Define an increasing sequence \( \left( {b}_{n}\right) \) and a decreasing sequence \( \left( {c}_{n}\right) \) in \( \left\lbrack {a, b}\right\rbrack \) inductively as follows: Put \( {b}_{0} = a \) and \( {c}_{0} = b \) . For some \( n \in \mathbb{N} \), suppose\n\n\[ \n{b}_{0} < {b}_{1} < \cdots < {b}_{n} < {c}_{n} < \cdots < {c}_{1} < {c}_{0} \n\] \n\nhave been defined. Let \( {i}_{n} \) be the first integer \( i \) such that \( {b}_{n} < {a}_{i} < {c}_{n} \) and \( {j}_{n} \) the first integer \( j \) such that \( {a}_{{i}_{n}} < {a}_{j} < {c}_{n} \) . Since \( \left\lbrack {a, b}\right\rbrack \) is infinite \( {i}_{n},{j}_{n} \) exist. Put \( {b}_{n + 1} = {a}_{{i}_{n}} \) and \( {c}_{n + 1} = {a}_{{j}_{n}} \) .\n\nLet \( x = \sup \left\{ {{b}_{n} : n \in \mathbb{N}}\right\} \) . Clearly, \( x \in \left\lbrack {a, b}\right\rbrack \) . Suppose \( x = {a}_{k} \) for some \( k \) . Clearly, \( x \leq {c}_{m} \) for all \( m \) . So, by the definition of the sequence \( \left( {b}_{n}\right) \) there is an integer \( i \) such that \( {b}_{i} > {a}_{k} = x \) . This contradiction shows that the range of the sequence \( \left( {a}_{n}\right) \) is not the whole of \( \left\lbrack {a, b}\right\rbrack \) . Since \( \left( {a}_{n}\right) \) was an arbitrary sequence, the result follows.
Yes
Theorem 1.1.8 (Cantor) For any two real numbers \( a, b \) with \( a < b \), the interval \( \left\lbrack {a, b}\right\rbrack \) is uncountable.
Proof. (Cantor) Let \( \left( {a}_{n}\right) \) be a sequence in \( \left\lbrack {a, b}\right\rbrack \) . Define an increasing sequence \( \left( {b}_{n}\right) \) and a decreasing sequence \( \left( {c}_{n}\right) \) in \( \left\lbrack {a, b}\right\rbrack \) inductively as follows: Put \( {b}_{0} = a \) and \( {c}_{0} = b \) . For some \( n \in \mathbb{N} \), suppose\n\n\[ \n{b}_{0} < {b}_{1} < \cdots < {b}_{n} < {c}_{n} < \cdots < {c}_{1} < {c}_{0} \n\] \n\nhave been defined. Let \( {i}_{n} \) be the first integer \( i \) such that \( {b}_{n} < {a}_{i} < {c}_{n} \) and \( {j}_{n} \) the first integer \( j \) such that \( {a}_{{i}_{n}} < {a}_{j} < {c}_{n} \) . Since \( \left\lbrack {a, b}\right\rbrack \) is infinite \( {i}_{n},{j}_{n} \) exist. Put \( {b}_{n + 1} = {a}_{{i}_{n}} \) and \( {c}_{n + 1} = {a}_{{j}_{n}} \) .\n\nLet \( x = \sup \left\{ {{b}_{n} : n \in \mathbb{N}}\right\} \) . Clearly, \( x \in \left\lbrack {a, b}\right\rbrack \) . Suppose \( x = {a}_{k} \) for some \( k \) . Clearly, \( x \leq {c}_{m} \) for all \( m \) . So, by the definition of the sequence \( \left( {b}_{n}\right) \) there is an integer \( i \) such that \( {b}_{i} > {a}_{k} = x \) . This contradiction shows that the range of the sequence \( \left( {a}_{n}\right) \) is not the whole of \( \left\lbrack {a, b}\right\rbrack \) . Since \( \left( {a}_{n}\right) \) was an arbitrary sequence, the result follows.
Yes
Theorem 1.1.9 The set \( \{ 0,1{\} }^{\mathbb{N}} \), consisting of all sequences of 0’s and 1’s, is uncountable.
Proof. Let \( \left( {\alpha }_{n}\right) \) be a sequence in \( \{ 0,1{\} }^{\mathbb{N}} \) . Define \( \alpha \in \{ 0,1{\} }^{\mathbb{N}} \) by\n\n\[ \alpha \left( n\right) = 1 - {\alpha }_{n}\left( n\right), n \in \mathbb{N}. \]\n\nThen \( \alpha \neq {\alpha }_{i} \) for all \( i \) . Since \( \left( {\alpha }_{n}\right) \) was arbitrary, our result is proved.
Yes
Theorem 1.2.1 (Cantor) For any set \( X, X{ < }_{c}\mathcal{P}\left( X\right) \) .
Proof. First assume that \( X = \varnothing \) . Then \( \mathcal{P}\left( X\right) = \{ \varnothing \} \) . The only function on \( X \) is the empty function \( \varnothing \), which is not onto \( \{ \varnothing \} \) . This observation proves the result when \( X = \varnothing \) .\n\nNow assume that \( X \) is nonempty. The map \( x \rightarrow \{ x\} \) from \( X \) to \( \mathcal{P}\left( X\right) \) is one-to-one. Therefore, \( X{ \leq }_{c}\mathcal{P}\left( X\right) \) . Let \( f : X \rightarrow \mathcal{P}\left( X\right) \) be any map. We show that \( f \) cannot be onto \( \mathcal{P}\left( X\right) \) . This will complete the proof.\n\nConsider the set\n\n\[ A = \{ x \in X \mid x \notin f\left( x\right) \} .\n\]\n\nSuppose \( A = f\left( {x}_{0}\right) \) for some \( {x}_{0} \in X \) . Then\n\n\[ {x}_{0} \in A \Leftrightarrow {x}_{0} \notin A.\n\]\n\nThis contradiction proves our claim.
Yes
Theorem 1.2.3 (Schröder - Bernstein Theorem) For any two sets \( X \) and \( Y \) , \[ \left( {X{ \leq }_{c}Y\& Y{ \leq }_{c}X}\right) \Rightarrow X \equiv Y. \]
Proof. (Dedekind) Let \( X{ \leq }_{c}Y \) and \( Y{ \leq }_{c}X \) . Fix one-to-one maps \( f : X \rightarrow Y \) and \( g : Y \rightarrow X \) . We have to show that \( X \) and \( Y \) have the same cardinality; i.e., that there is a bijection \( h \) from \( X \) onto \( Y \) .\n\nWe first show that there is a set \( E \subseteq X \) such that \[ {g}^{-1}\left( {X \smallsetminus E}\right) = Y \smallsetminus f\left( E\right) \] \( \left( *\right) \)\n\n(See Figure 1.1.) Assuming that such a set \( E \) exists, we complete the proof as follows. Define \( h : X \rightarrow Y \) by \[ h\left( x\right) = \left\{ \begin{array}{ll} f\left( x\right) & \text{ if }x \in E \\ {g}^{-1}\left( x\right) & \text{ otherwise. } \end{array}\right. \] The map \( h : X \rightarrow Y \) is clearly seen to be one-to-one and onto.\n\nWe now show the existence of a set \( E \subseteq X \) satisfying \( \left( \star \right) \) . Consider the map \( \mathcal{H} : \mathcal{P}\left( X\right) \rightarrow \mathcal{P}\left( X\right) \) defined by \[ \mathcal{H}\left( A\right) = X \smallsetminus g\left( {Y \smallsetminus f\left( A\right) }\right) ,\;A \subseteq X. \] It is easy to check that (i) \( A \subseteq B \subseteq X \Rightarrow \mathcal{H}\left( A\right) \subseteq \mathcal{H}\left( B\right) \), and (ii) \( \mathcal{H}\left( {\mathop{\bigcup }\limits_{n}{A}_{n}}\right) = \mathop{\bigcup }\limits_{n}\mathcal{H}\left( {A}_{n}\right) \).\n\nNow define a sequence \( \left( {A}_{n}\right) \) of subsets of \( X \) inductively as follows: \( {A}_{0} = \varnothing \), and \( {A}_{n + 1} = \mathcal{H}\left( {A}_{n}\right), n = 0,1,2,\ldots \) Let \( E = \mathop{\bigcup }\limits_{n}{A}_{n} \) . Then, \( \mathcal{H}\left( E\right) = E \) . The set \( E \) clearly satisfies \( \left( \star \right) \).
Yes
Corollary 1.2.4 For sets \( A \) and \( B \) ,
\[ A{ < }_{c}B \Leftrightarrow A{ \leq }_{c}B\& B{ \nleq }_{c}A. \]
Yes
Example 1.2.5 Define \( f : \mathcal{P}\left( \mathbb{N}\right) \rightarrow \mathbb{R} \), the set of all real numbers, by\n\n\[ f\left( A\right) = \mathop{\sum }\limits_{{n \in A}}\frac{2}{{3}^{n + 1}}, A \subseteq \mathbb{N}. \]\n\nThen \( f \) is one-to-one. Therefore, \( \mathcal{P}\left( \mathbb{N}\right) { \leq }_{c}\mathbb{R} \) .
Now consider the map \( g \) :\n\n\( \mathbb{R} \rightarrow \mathcal{P}\left( \mathbb{Q}\right) \) by\n\n\[ g\left( x\right) = \{ r \in \mathbb{Q} \mid r < x\}, x \in \mathbb{R}. \]\n\nClearly, \( g \) is one-to-one and so \( \mathbb{R}{ \leq }_{c}\mathcal{P}\left( \mathbb{Q}\right) \) . As \( \mathbb{Q} \equiv \mathbb{N},\mathcal{P}\left( \mathbb{Q}\right) \equiv \mathcal{P}\left( \mathbb{N}\right) \) . Therefore, \( \mathbb{R}{ \leq }_{c}\mathcal{P}\left( \mathbb{N}\right) \) . By the Schröder - Bernstein theorem, \( \mathbb{R} \equiv \mathcal{P}\left( \mathbb{N}\right) \) . Since \( \mathcal{P}\left( \mathbb{N}\right) \equiv \{ 0,1{\} }^{\mathbb{N}},\mathbb{R} \equiv \{ 0,1{\} }^{\mathbb{N}} \) .
Yes
Example 1.2.6 Fix a one-to-one map \( x \rightarrow \left( {{x}_{0},{x}_{1},{x}_{2},\ldots }\right) \) from \( \mathbb{R} \) onto \( \{ 0,1{\} }^{\mathbb{N}} \), the set of sequences of 0 ’s and 1’s. Then the function \( \left( {x, y}\right) \rightarrow \) \( \left( {{x}_{0},{y}_{0},{x}_{1},{y}_{1},\ldots }\right) \) from \( {\mathbb{R}}^{2} \) to \( \{ 0,1{\} }^{\mathbb{N}} \) is one-to-one and onto. So, \( {\mathbb{R}}^{2} \equiv \) \( \{ 0,1{\} }^{\mathbb{N}} \equiv \mathbb{R} \) . By induction on the positive integers \( k \), we can now show that \( {\mathbb{R}}^{k} \) and \( \mathbb{R} \) are equinumerous.
Null
No
Theorem 1.3.1 If \( X \) is infinite and \( A \subseteq X \) finite, then \( X \smallsetminus A \) and \( X \) have the same cardinality.
Proof. Let \( A = \left\{ {{a}_{0},{a}_{1},\ldots ,{a}_{n}}\right\} \) with the \( {a}_{i} \) ’s distinct. By \( \mathbf{{AC}} \), there exist distinct elements \( {a}_{n + 1},{a}_{n + 2},\ldots \) in \( X \smallsetminus A \) . To see this, fix a choice function \( f : \mathcal{P}\left( X\right) \smallsetminus \{ \varnothing \} \rightarrow X \) such that \( f\left( E\right) \in E \) for every nonempty subset \( E \) of \( X \) . Such a function exists by \( \mathbf{{AC}} \) . Now inductively define \( {a}_{n + 1},{a}_{n + 2},\ldots \) such that\n\n\[ \n{a}_{n + k + 1} = f\left( {X \smallsetminus \left\{ {{a}_{0},{a}_{1},\ldots ,{a}_{n + k}}\right\} }\right)\n\]\n\n\( k = 0,1,\ldots \) Define \( h : X \rightarrow X \smallsetminus A \) by\n\n\[ \nh\left( x\right) = \left\{ \begin{array}{ll} {a}_{n + k + 1} & \text{ if }x = {a}_{k} \\ x & \text{ otherwise. } \end{array}\right.\n\]\n\nClearly, \( h : X \rightarrow X \smallsetminus A \) is one-to-one and onto.
Yes
Corollary 1.3.2 Show that for any infinite set \( X,\mathbb{N}{ \leq }_{c}X \) ; i.e., every infinite set \( X \) has a countable infinite subset.
Null
No
Example 1.3.4 Let \( X \) and \( Y \) be any two sets. A partial function \( f \) : \( X \rightarrow Y \) is a function with domain a subset of \( X \) and range contained in \( Y \) . Let \( f : X \rightarrow Y \) and \( g : X \rightarrow Y \) be partial functions. We say that \( g \) extends \( f \), or \( f \) is a restriction of \( g \), written \( g \succcurlyeq f \) or \( f \preccurlyeq g \), if domain \( \left( f\right) \) is contained in \( \operatorname{domain}\left( g\right) \) and \( f\left( x\right) = g\left( x\right) \) for all \( x \in \operatorname{domain}\left( f\right) \) . If \( f \) is a restriction of \( g \) and \( \operatorname{domain}\left( f\right) = A \), we write \( f = g \mid A \) . Let\n\n\[ \n{Fn}\left( {X, Y}\right) = \{ f : f\text{ a one-to-one partial function from }X\text{ to }Y\} .\n\]\n\nSuppose \( Y \) has more than one element and \( X \neq \varnothing \) . Then \( \left( {{Fn}\left( {X, Y}\right) , \preccurlyeq }\right) \) is a poset that is not linearly ordered.
Null
No
Example 1.3.5 Let \( V \) be a vector space over any field \( F \) and \( P \) the set of all independent subsets of \( V \) ordered by the inclusion \( \subseteq \) . Then \( P \) is a poset that is not a linearly ordered set.
In 1.3.5, Let \( C \) be a chain in \( P \) . Then for any two elements \( E \) and \( F \) of \( P \), either \( E \subseteq F \) or \( F \subseteq E \) . It follows that \( \bigcup C \) itself is an independent set and so is an upper bound of \( C \) .
No
Proposition 1.3.7 Every vector space \( V \) has a basis.
Proof. Let \( P \) be the poset defined in 1.3.5; i.e., \( P \) is the set of all independent subsets of \( V \) . Since every singleton set \( \{ v\}, v \neq 0 \), is an independent set, \( P \neq \varnothing \) . As shown earlier, every chain in \( P \) has an upper bound. Therefore, by Zorn’s lemma, \( P \) has a maximal element, say \( B \) . Suppose \( B \) does not span \( V \) . Take \( v \in V \smallsetminus \operatorname{span}\left( B\right) \) . Then \( B\bigcup \{ v\} \) is an independent set properly containing \( B \) . This contradicts the maximality of \( B \) . Thus \( B \) is a basis of \( V \) .
Yes
Theorem 1.4.1 For any two sets \( X \) and \( Y \), at least one of\n\n\[ X{ \leq }_{c}Y\text{or}Y{ \leq }_{c}X \]\n\nholds.
Proof. Without loss of generality we can assume that both \( X \) and \( Y \) are nonempty. We need to show that either there exists a one-to-one map \( f : X \rightarrow Y \) or there exists a one-to-one map \( g : Y \rightarrow X \) . To show this, consider the poset \( {Fn}\left( {X, Y}\right) \) of all one-to-one partial functions from \( X \) to \( Y \) as defined in 1.3.4. It is clearly nonempty. As shown earlier, every chain in \( {Fn}\left( {X, Y}\right) \) has an upper bound. Therefore, by Zorn’s lemma, \( P \) has a maximal element, say \( {f}_{0} \) . Then, either \( \operatorname{domain}\left( {f}_{0}\right) = X \) or range \( \left( {f}_{0}\right) = Y \) . If \( \operatorname{domain}\left( {f}_{0}\right) = X \), then \( {f}_{0} \) is a one-to-one map from \( X \) to \( Y \) . So, in this case, \( X{ \leq }_{c}Y \) . If \( \operatorname{range}\left( {f}_{0}\right) = Y \), then \( {f}_{0}^{-1} \) is a one-to-one map from \( Y \) to \( X \), and so \( Y{ \leq }_{c}X \) .
Yes
Corollary 1.4.2 Let \( A \) and \( B \) be any two sets. Then exactly one of\n\n\[ A{ < }_{c}B, A \equiv B,\text{ and }B{ < }_{c}A \] \n\nholds.
Null
No
Theorem 1.4.3 For every infinite set \( X \) ,\n\n\[ X \times \{ 0,1\} \equiv X \]
Proof. Let\n\n\[ P = \{ \left( {A, f}\right) : A \subseteq X\\text{ and }f : A \times \{ 0,1\} \rightarrow A\\text{ a bijection }\} .\n\]\n\nSince \( X \) is infinite, it contains a countably infinite set, say \( D \) . By 1.1.3, \( D \times \{ 0,1\} \equiv D \) . Therefore, \( P \) is nonempty. Consider the partial order \( \\propto \) on \( P \) defined by\n\n\[ \\left( {A, f}\right) \\propto \\left( {B, g}\right) \\Leftrightarrow A \\subseteq B\\& f \\preccurlyeq g.\n\]\n\nFollowing the argument contained in the proof of 1.4.1, we see that the hypothesis of Zorn’s lemma is satisfied by \( P \) . So, \( P \) has a maximal element, say \( \\left( {A, f}\right) \) .\n\nTo complete the proof we show that \( A \\equiv X \) . Since \( X \) is infinite, by 1.3.1, it will be sufficient to show that \( X \\smallsetminus A \) is finite. Suppose not. By 1.3.2, there is a \( B \\subseteq X \\smallsetminus A \) such that \( B \\equiv \\mathbb{N} \) . So there is a one-to-one map \( g \) from \( B \times \{ 0,1\} \) onto \( B \) . Combining \( f \) and \( g \) we get a bijection\n\n\[ h : \\left( {A\\bigcup B}\right) \times \{ 0,1\} \\rightarrow A\\bigcup B\n\]\n\nthat extends \( f \) . This contradicts the maximality of \( \\left( {A, f}\right) \) . Hence, \( X \\smallsetminus A \) is finite. Therefore, \( A \\equiv X \) . The proof is complete.
Yes
Theorem 1.4.5 For every infinite set \( X \) , \[ X \times X \equiv X. \]
Proof. Let \[ P = \{ \left( {A, f}\right) : A \subseteq X\text{ and }f : A \times A \rightarrow A\text{ a bijection }\} . \] Note that \( P \) is nonempty. Consider the partial order \( \propto \) on \( P \) defined by \[ \left( {A, f}\right) \propto \left( {B, g}\right) \Leftrightarrow A \subseteq B\& f \preccurlyeq g. \] By Zorn’s lemma, take a maximal element \( \left( {A, f}\right) \) of \( P \) as in the proof of 1.4.3. Note that \( A \) must be infinite. To complete the proof, we shall show that \( A \equiv X \) . Suppose not. Then \( A{ < }_{c}X \) . We first show that \( X \smallsetminus A \equiv X \) . Suppose \( X \smallsetminus A{ < }_{c}X \) . By 1.4.1, either \( A{ \leq }_{c}X \smallsetminus A \) or \( X \smallsetminus A{ \leq }_{c}A \) . Assume first \( X \smallsetminus A{ \leq }_{c}A \) . Using 1.4.3, take two disjoint sets \( {A}_{1},{A}_{2} \) of the same cardinality as \( A \) and \( {A}_{1}\bigcup {A}_{2} = A \) . Now, \[ X = A\bigcup \left( {X \smallsetminus A}\right) { \leq }_{c}{A}_{1}\bigcup {A}_{2} \equiv A{ < }_{c}X. \] This is a contradiction. Similarly we arrive at a contradiction from the other inequality. Thus, by 1.4.2, \( X \smallsetminus A \equiv X \) . Now choose \( B \subseteq X \smallsetminus A \) such that \( B \equiv A \) . By 1.4.4, write \( B \) as the union of three disjoint sets, say \( {B}_{1},{B}_{2} \), and \( {B}_{3} \), each of the same cardinality as \( A \) . Since there is a one-to-one map from \( A \times A \) onto \( A \), there exist bijections \( {f}_{1} : B \times A \rightarrow {B}_{1},{f}_{2} : B \times B \rightarrow {B}_{2} \), and \( {f}_{3} : A \times B \rightarrow {B}_{3} \) . Let \( C = \) \( A\bigcup B \) . Combining these four bijections, we get a bijection \( g : C \times C \rightarrow C \) that is a proper extension of \( f \) . This contradicts the maximality of \( \left( {A, f}\right) \) . Thus, \( A \equiv X \) . The proof is now complete.
Yes
Theorem 1.4.5 For every infinite set \( X \) , \[ X \times X \equiv X. \]
Proof. Let \[ P = \{ \left( {A, f}\right) : A \subseteq X\text{ and }f : A \times A \rightarrow A\text{ a bijection }\} . \] Note that \( P \) is nonempty. Consider the partial order \( \propto \) on \( P \) defined by \[ \left( {A, f}\right) \propto \left( {B, g}\right) \Leftrightarrow A \subseteq B\& f \preccurlyeq g. \] By Zorn’s lemma, take a maximal element \( \left( {A, f}\right) \) of \( P \) as in the proof of 1.4.3. Note that \( A \) must be infinite. To complete the proof, we shall show that \( A \equiv X \) . Suppose not. Then \( A{ < }_{c}X \) . We first show that \( X \smallsetminus A \equiv X \) . Suppose \( X \smallsetminus A{ < }_{c}X \) . By 1.4.1, either \( A{ \leq }_{c}X \smallsetminus A \) or \( X \smallsetminus A{ \leq }_{c}A \) . Assume first \( X \smallsetminus A{ \leq }_{c}A \) . Using 1.4.3, take two disjoint sets \( {A}_{1},{A}_{2} \) of the same cardinality as \( A \) and \( {A}_{1}\bigcup {A}_{2} = A \) . Now, \[ X = A\bigcup \left( {X \smallsetminus A}\right) { \leq }_{c}{A}_{1}\bigcup {A}_{2} \equiv A{ < }_{c}X. \] This is a contradiction. Similarly we arrive at a contradiction from the other inequality. Thus, by 1.4.2, \( X \smallsetminus A \equiv X \) . Now choose \( B \subseteq X \smallsetminus A \) such that \( B \equiv A \) . By 1.4.4, write \( B \) as the union of three disjoint sets, say \( {B}_{1},{B}_{2} \), and \( {B}_{3} \), each of the same cardinality as \( A \) . Since there is a one-to-one map from \( A \times A \) onto \( A \), there exist bijections \( {f}_{1} : B \times A \rightarrow {B}_{1},{f}_{2} : B \times B \rightarrow {B}_{2} \), and \( {f}_{3} : A \times B \rightarrow {B}_{3} \) . Let \( C = \) \( A\bigcup B \) . Combining these four bijections, we get a bijection \( g : C \times C \rightarrow C \) that is a proper extension of \( f \) . This contradicts the maximality of \( \left( {A, f}\right) \) . Thus, \( A \equiv X \) . The proof is now complete.
Yes
Proposition 1.4.8 (J. König,[58]) Let \( \\left\\{ {{X}_{i} : i \\in I}\\right\\} \) and \( \\left\\{ {{Y}_{i} : i \\in I}\\right\\} \) be families of sets such that \( {X}_{i}{ < }_{c}{Y}_{i} \) for each \( i \\in I \) . Then there is no map \( f \) from \( \\mathop{\\bigcup }\\limits_{i}{X}_{i} \) onto \( {\\Pi }_{i}{Y}_{i} \) .
Proof. Let \( f : \\mathop{\\bigcup }\\limits_{i}{X}_{i} \\rightarrow {\\Pi }_{i}{Y}_{i} \) be any map. For any \( i \\in I \\), let\n\n\[ \n{A}_{i} = {Y}_{i} \\smallsetminus {\\pi }_{i}\\left( {f\\left( {X}_{i}\\right) }\\right)\n\]\n\nwhere \( {\\pi }_{i} : \\mathop{\\prod }\\limits_{j}{Y}_{j} \\rightarrow {Y}_{i} \) is the projection map. Since for evry \( i,{X}_{i}{ < }_{c}{Y}_{i} \) , each \( {A}_{i} \) is nonempty. By \( \\mathbf{{AC}},{\\Pi }_{i}{A}_{i} \\neq \\varnothing \) . But\n\n\[ \n{\\Pi }_{i}{A}_{i}\\bigcap \\operatorname{range}\\left( f\\right) = \\varnothing .\n\]\n\nIt follows that \( f \) is not onto.
Yes
\[ {2}^{\mathfrak{c}} \leq {\aleph }_{0}^{\mathfrak{c}} \]
\[ {2}^{\mathfrak{c}} \leq {\aleph }_{0}^{\mathfrak{c}}\;\text{ (since }2 \leq {\aleph }_{0}\text{ ) } \] \[ \leq {\mathfrak{c}}^{\mathfrak{c}}\;\text{ (since }{\aleph }_{0} \leq \mathfrak{c}\text{ ) } \] \[ = {\left( {2}^{{\aleph }_{0}}\right) }^{\mathfrak{c}}\;\left( {\text{since}\mathfrak{c} = {2}^{{\aleph }_{0}}}\right) \] \[ = {2}^{{\aleph }_{0} \cdot \mathfrak{c}}\;\text{(since for nonempty sets}X, Y, Z,{\left( {X}^{Y}\right) }^{Z} \equiv {X}^{Y \times Z}\text{)} \] \[ \leq {2}^{\mathfrak{c} \cdot \mathfrak{c}}\;\text{ (since }{\aleph }_{0} < \mathfrak{c}\text{ ) } \] \[ = {2}^{\mathfrak{c}}\;\text{(since}\mathfrak{c} \cdot \mathfrak{c} = \mathfrak{c}\text{).} \] So, by the Schröder - Bernstein theorem, \( {2}^{\mathfrak{c}} = {\aleph }_{0}^{\mathfrak{c}} = {\mathfrak{c}}^{\mathfrak{c}} \) . It follows that \[ \{ 0,1{\} }^{\mathbb{R}} \equiv {\mathbb{N}}^{\mathbb{R}} \equiv {\mathbb{R}}^{\mathbb{R}} \]
Yes
Proposition 1.6.2 A linearly ordered set \( \\left( {W, \\leq }\\right) \) is well-ordered if and only if there is no descending sequence \( {w}_{0} > {w}_{1} > {w}_{2} > \\cdots \) in \( W \) .
Proof. Let \( W \) be not well-ordered. Then there is a nonempty subset \( A \) of \( W \) not having a least element. Choose any \( {w}_{0} \\in A \) . Since \( {w}_{0} \) is not the first element of \( A \), there is a \( {w}_{1} \\in A \) such that \( {w}_{1} < {w}_{0} \) . Since \( {w}_{1} \) is not the first element of \( A \), we get \( {w}_{2} < {w}_{1} \) in \( A \) . Proceeding similarly, we get a descending sequence \( \\left\\{ {{w}_{n} : n \\geq 0}\\right\\} \) in \( W \) . This completes the proof of the \
Yes
Example 1.6.3 Let \( W = \mathbb{N}\bigcup \{ \infty \} \). Let \( \leq \) be defined in the usual way on \( \mathbb{N} \) and let \( i < \infty \) for \( i \in \mathbb{N} \). Clearly, \( W \) is a well-ordered set. Since \( W \) has a last element and \( {\omega }_{0} \) does not, \( \left( {W, \leq }\right) \) is not isomorphic to \( {\omega }_{0} \). Thus there exist nonisomorphic well-ordered sets of the same cardinality.
Null
No
Proposition 1.6.5 No well-ordered set \( W \) is order isomorphic to an initial segment \( W\left( u\right) \) of itself.
Proof. Let \( W \) be a well-ordered set and \( u \in W \) . Suppose \( W \) and \( W\left( u\right) \) are isomorphic. Let \( f : W \rightarrow W\left( u\right) \) be an order isomorphism. For \( n \in \mathbb{N} \) , let \( {w}_{n} = {f}^{n}\left( u\right) \) . Note that\n\n\[ \n{w}_{0} = {f}^{0}\left( u\right) = u > {f}^{1}\left( u\right) = f\left( u\right) = {w}_{1}.\n\]\n\nBy induction on \( n \), we see that \( {w}_{n} > {w}_{n + 1} \) for all \( n \), i.e., \( \left( {w}_{n}\right) \) is a descending sequence in \( W \) . By 1.6.2, \( W \) is not well-ordered. This contradiction proves our result.
Yes
Proposition 1.7.1 (Proof by induction) For each \( n \in \mathbb{N} \), let \( {P}_{n} \) be a mathematical proposition. Suppose \( {P}_{0} \) is true and for every \( n,{P}_{n + 1} \) is true whenever \( {P}_{n} \) is true. Then for every \( n,{P}_{n} \) is true. Symbolically, we can express this as follows.
The proof of this proposition uses two basic properties of the set of natural numbers. First, it is well-ordered by the usual order, and second, every nonzero element in it is a successor. A repeated application of 1.7.1 gives us the following.
No
Proposition 1.7.2 (Definition by induction) Let \( X \) be any nonempty set. Suppose \( {x}_{0} \) is a fixed point of \( X \) and \( g : X \rightarrow X \) any map. Then there is a unique map \( f : \mathbb{N} \rightarrow X \) such that \( f\left( 0\right) = {x}_{0} \) and \( f\left( {n + 1}\right) = g\left( {f\left( n\right) }\right) \) for all \( n \) .
Null
No
Theorem 1.7.3 (Proof by transfinite induction) Let \( \\left( {W, \\leq }\\right) \) be a well-ordered set, and for every \( w \\in W \), let \( {P}_{w} \) be a mathematical proposition. Suppose that for each \( w \\in W \), if \( {P}_{v} \) is true for each \( v < w \), then \( {P}_{w} \) is true. Then for every \( w \\in W,{P}_{w} \) is true. Symbolically, we express this as\n\n\[ \n\\left( {\\forall w \\in W}\\right) \\left( {\\left( {\\left( {\\forall v < w}\\right) {P}_{v}}\\right) \\Rightarrow {P}_{w}}\\right) \\Rightarrow \\left( {\\forall w \\in W}\\right) {P}_{w}.\n\]
Proof. Let\n\n\[ \n\\left( {\\forall w \\in W}\\right) \\left( {\\left( {\\left( {\\forall v < w}\\right) {P}_{v}}\\right) \\Rightarrow {P}_{w}}\\right) .\n\]\n\n\( \\left( *\\right) \)\n\nSuppose \( {P}_{w} \) is false for some \( w \\in W \) . Consider\n\n\[ \nA = \\left\{ {w \\in W : {P}_{w}\\text{ does not hold }}\\right\} .\n\]\n\nBy our assumptions, \( A \\neq \\varnothing \) . Let \( {w}_{0} \) be the least element of \( A \) . Then for every \( v < {w}_{0},{P}_{v} \) holds. However, \( {P}_{{w}_{0}} \) does not hold. This contradicts \( \\left( \\star \\right) \). Therefore, for every \( w \\in W,{P}_{w} \) holds.
Yes
Theorem 1.7.4 (Definition by transfinite induction) Let \( \\left( {W, \\leq }\\right) \) be a well-ordered set, \( X \) a set, and \( \\mathcal{F} \) the set of all maps with domain an initial segment of \( W \) and range contained in \( X \) . If \( G : \\mathcal{F} \\rightarrow X \) is any map, then there is a unique map \( f : W \\rightarrow X \) such that for every \( u \\in W \) ,\n\n\[ f\\left( u\\right) = G\\left( {f \\mid W\\left( u\\right) }\\right) . \]
Proof. For each \( w \\in W \), let \( {P}_{w} \) be the proposition \
No
Theorem 1.7.5 (Trichotomy theorem for well-ordered sets) For any two well-ordered sets \( W \) and \( {W}^{\prime } \), exactly one of\n\n\[ W \prec {W}^{\prime }, W \sim {W}^{\prime },\text{ and }{W}^{\prime } \prec W \]\n\nholds.
Proof. It is easy to see that no two of these can hold simultaneously. For example, if \( W \sim {W}^{\prime } \) and \( {W}^{\prime } \prec W \), then \( W \) is isomorphic to an initial segment of itself. This is impossible by 1.6.5.\n\nTo show that at least one of these holds, take \( X = {W}^{\prime }\bigcup \{ \infty \} \), where \( \infty \) is a point outside \( {W}^{\prime } \) . Now define a map \( f : W \rightarrow X \) by transfinite induction as follows. Let \( w \in W \) and assume that \( f \) has been defined on \( W\left( w\right) \) . If \( {W}^{\prime } \smallsetminus f\left( {W\left( w\right) }\right) \neq \varnothing \), then we take \( f\left( w\right) \) to be the least element of \( {W}^{\prime } \smallsetminus f\left( {W\left( w\right) }\right) \) ; otherwise, \( f\left( w\right) = \infty \) . By 1.7.4, such a function exists.\n\nLet us assume that \( \infty \notin f\left( W\right) \) . Then\n\n(i) the map \( f \) is one-to-one and order preserving, and\n\n(ii) the range of \( f \) is either whole of \( {W}^{\prime } \) or an initial segment of \( {W}^{\prime } \) .\n\nSo, in this case at least one of \( W \sim {W}^{\prime } \) or \( W \prec {W}^{\prime } \) holds.\n\nIf \( \infty \in f\left( W\right) \), then let \( w \) be the first element of \( W \) such that \( f\left( w\right) = \infty \) . Then \( f \mid W\left( w\right) \) is an order isomorphism from \( W\left( w\right) \) onto \( {W}^{\prime } \) . Thus in this case \( {W}^{\prime } \prec W \) .
Yes
Corollary 1.7.6 Let \( \\left( {W, \\leq }\\right) ,\\left( {{W}^{\\prime },{ \\leq }^{\\prime }}\\right) \) be well-ordered sets. Then \( W \\preccurlyeq {W}^{\\prime } \) if and only if there is a one-to-one order-preserving map from \( W \) into \( {W}^{\\prime } \) .
Proof. Suppose there is a one-to-one order-preserving map \( g \) from \( W \) into \( {W}^{\\prime } \) . Let \( X \) and \( f : W \\rightarrow X \) be as in the proof of 1.7.5. Then, by induction on \( w \), we easily show that for every \( w \\in W, f\\left( w\\right) { \\leq }^{\\prime }g\\left( w\\right) \) . Therefore, \( \\infty \\notin f\\left( W\\right) \) . Hence, \( W \\preccurlyeq {W}^{\\prime } \) . The converse is clear.
Yes
Theorem 1.7.7 Let \( \mathcal{W} = \left\{ {\left( {{W}_{i},{ \leq }_{i}}\right) : i \in I}\right\} \) be a family of pairwise non-isomorphic well-ordered sets. Then there is a \( W \in \mathcal{W} \) such that \( W \prec {W}^{\prime } \) for every \( {W}^{\prime } \in \mathcal{W} \) different from \( W \) .
Proof. Suppose no such \( W \) exists. Then there is a descending sequence\n\n\[ \cdots \prec {W}_{n} \prec \cdots \prec {W}_{1} \prec {W}_{0} \]\n\nin \( \mathcal{W} \) . For \( n \in \mathbb{N} \), choose a \( {w}_{n}^{\prime } \in {W}_{n} \) such that \( {W}_{n + 1} \sim {W}_{n}\left( {w}_{n}^{\prime }\right) \) . Fix an order isomorphism \( {f}_{n} : {W}_{n + 1} \rightarrow {W}_{n}\left( {w}_{n}^{\prime }\right) \) . Let \( {w}_{0} = {w}_{0}^{\prime } \), and for \( n > 0 \), \n\n\[ {w}_{n} = {f}_{0}\left( {{f}_{1}\left( {\cdots {f}_{n - 1}\left( {w}_{n}^{\prime }\right) }\right) }\right) . \]\n\n(See Figure 1.2.) Then \( \left( {w}_{n}\right) \) is a descending sequence in \( {W}_{0} \) . This is a contradiction. The result follows.
Yes
Theorem 1.8.3 Every ordinal \( \alpha \) can be uniquely written as \[ \alpha = \beta + n \] where \( \beta \) is a limit ordinal and \( n \) finite.
Proof. Let \( \alpha \) be an ordinal number. We first show that there exists a limit ordinal \( \beta \) and an \( n \in \omega \) such that \( \alpha = \beta + n \) . Choose a well-ordered set \( W \) such that \( t\left( W\right) = \alpha \) . If \( W \) has no last element, then we take \( \beta = \alpha \) and \( n = 0 \) . Suppose \( W \) has a last element, say \( {w}_{0} \) . If \( {w}_{0} \) has no immediate predecessor, then take \( \beta = t\left( {W\left( {w}_{0}\right) }\right) \) and \( n = 1 \) . Now suppose that \( {w}_{0} \) does have an immediate predecessor, say \( {w}_{1} \) . If \( {w}_{1} \) has no immediate predecessor, then we take \( \beta = t\left( {W\left( {w}_{1}\right) }\right) \) and \( n = 2 \) . Since \( W \) has no descending sequence, this process ends after finitely many steps. Thus we get \( {w}_{0},{w}_{1},\ldots ,{w}_{k - 1} \) such that \( {w}_{i} = {w}_{i - 1}^{ - } \) for all \( i > 0 \), and \( {w}_{k - 1} \) has no immediate predecessor. We take \( \beta = t\left( {W\left( {w}_{k - 1}\right) }\right) \) and \( n = k \) . We now show that \( \alpha \) has a unique representation of the type mentioned above. Let \( W,{W}^{\prime } \) be well-ordered sets with no last element, and \( {A}_{n},{B}_{m} \) finite well-ordered sets of cardinality \( n \) and \( m \) respectively such that \[ {A}_{n} \cap W = {B}_{m} \cap {W}^{\prime } = \varnothing . \] Let \( f : W + {A}_{n} \rightarrow {W}^{\prime } + {B}_{m} \) be an order isomorphism. It is easy to check that \( f\left( W\right) = {W}^{\prime } \) and \( f\left( {A}_{n}\right) = {B}_{m} \) . Uniqueness now follows.
Yes
Theorem 1.8.5 The set \( \Omega \) of all countable ordinals is uncountable.
Proof. Suppose \( \Omega \) is countable. Fix an enumeration \( {\alpha }_{0},{\alpha }_{1},\ldots \) of \( \Omega \) . Then\n\n\[ \alpha = \mathop{\sum }\limits_{n}{\alpha }_{n} + 1 \]\n\n is a countable ordinal strictly larger than each \( {\alpha }_{n} \) . This is a contradiction. So, \( \Omega \) is uncountable.
Yes
Proposition 1.8.6 Let \( \alpha \) be a countable limit ordinal. Then there exist \( {\alpha }_{0} < {\alpha }_{1} < \cdots \) such that \( \sup \left\{ {{\alpha }_{n} : n \in \mathbb{N}}\right\} = \alpha \) .
Proof. Since \( \alpha \) is countable, \( \{ \beta \in \mathbf{{ON}} : \beta < \alpha \} \) is countable. Fix an enumeration \( \left\{ {{\beta }_{n} : n \in \mathbb{N}}\right\} \) of all ordinals less than \( \alpha \) . We now define a sequence of ordinals \( \left( {\alpha }_{n}\right) \) by induction on \( n \) . Choose \( {\alpha }_{0} \) such that \( {\beta }_{0} < \) \( {\alpha }_{0} < \alpha \) . Since \( \alpha \) is a limit ordinal, such an ordinal exists. Suppose \( {\alpha }_{n} \) has been defined. Choose \( {\alpha }_{n + 1} \) greater than \( {\alpha }_{n} \) such that \( {\beta }_{n + 1} < {\alpha }_{n + 1} < \alpha \) . Clearly,\n\n\[ \n\alpha = \sup \left\{ {{\beta }_{n} : n \in \mathbb{N}}\right\} \leq \sup \left\{ {{\alpha }_{n} : n \in \mathbb{N}}\right\} \leq \alpha .\n\]\n\nSo, \( \sup \left\{ {{\alpha }_{n} : n \in \mathbb{N}}\right\} = \alpha \) .
Yes
Proposition 1.10.7 (König’s infinity lemma, [57]) Let \( T \) be a finitely splitting, infinite tree on \( A \) . Then \( T \) is ill-founded.
Proof. Let \( T \) be a finitely splitting, infinite tree on \( A \) . Let \( \left( {a}_{0}\right) \) be a node of \( T \) with infinitely many extensions in \( T \) . Since \( T \) is finitely splitting (and \( e \in T),\{ a \in A : \left( a\right) \in T\} \) is finite. Further, \( T \) is infinite. So, \( \left( {a}_{0}\right) \) exists. By the same argument we get \( {a}_{1} \in A \) such that \( {s}_{1} = \left( {{a}_{0},{a}_{1}}\right) \) has infinitely many extensions in \( T \) . Proceeding similarly we get an \( \alpha = \left( {{a}_{0},{a}_{1},\ldots }\right) \) such that for all \( k,\alpha \mid k \) has infinitely many extensions in \( T \) . In particular, \( \alpha \in \left\lbrack T\right\rbrack \) , and the result is proved.
Yes
Example 1.10.3 The tree\n\n\\[ T = \\{ e\\} \\bigcup \\left\\{ {i{0}^{j} : j \\leq i, i \\in \\mathbb{N}}\\right\\} \\]\n\nis infinite and well-founded.
Null
No
Example 1.10.4 Let \( T \) be a tree and \( u \) a node of \( T \) . The set\n\n\[ \n{T}_{u} = \\left\\{ {v \\in {A}^{ < \\mathbb{N}} : u \\hat{} v \\in T}\\right\\} \n\]\n\nforms a tree. (See Figure 1.4.)
Null
No
Proposition 1.10.7 (König’s infinity lemma, [57]) Let \( T \) be a finitely splitting, infinite tree on \( A \) . Then \( T \) is ill-founded.
Proof. Let \( T \) be a finitely splitting, infinite tree on \( A \) . Let \( \left( {a}_{0}\right) \) be a node of \( T \) with infinitely many extensions in \( T \) . Since \( T \) is finitely splitting (and \( e \in T),\{ a \in A : \left( a\right) \in T\} \) is finite. Further, \( T \) is infinite. So, \( \left( {a}_{0}\right) \) exists. By the same argument we get \( {a}_{1} \in A \) such that \( {s}_{1} = \left( {{a}_{0},{a}_{1}}\right) \) has infinitely many extensions in \( T \) . Proceeding similarly we get an \( \alpha = \left( {{a}_{0},{a}_{1},\ldots }\right) \) such that for all \( k,\alpha \mid k \) has infinitely many extensions in \( T \) . In particular, \( \alpha \in \left\lbrack T\right\rbrack \) , and the result is proved.
Yes