Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedprecheck pass
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

The future unit normal, induced metric and second fundamental form of a spacelike hypersurface

Statement

Let (M,g,N) be a time-oriented smooth Lorentzian four-manifold (Smooth time-oriented Lorentzian four-manifolds: the geometric objects of the model) and let Σ⊆M be a spacelike embedded hypersurface in the sense of Spacelike hypersurfaces, acausal sets, Cauchy sets, domains of dependence and Cauchy horizons clause (a), so that Σ is an embedded three-dimensional submanifold (Embedded submanifolds and slice charts) and the restriction of gp to TpΣ is positive definite for every p∈Σ (Timelike, null, spacelike and causal vectors of a Lorentz vector space). In the notation of Spacelike initial data, the constraint equations, developments and maximal developments clause (a), whose claims this lemma establishes:

(a) The future unit normal. For every p∈Σ the orthogonal complement TpΣ⊥ is one-dimensional, nondegenerate for gp, and negative definite: every nonzero u∈TpΣ⊥ is timelike, gp(u,u)<0. Hence there is exactly one np∈TpΣ⊥ with gp(np,np)=−1 and gp(np,Np)<0.

(b) Smoothness along the hypersurface. The assignment p↦np is smooth along Σ: for every p∈Σ there are an open neighborhood O of p in M and a smooth vector field n~ on O (A smooth vector field is a smooth section of the tangent bundle) with n~∣Σ∩O=n∣Σ∩O, and for any two such local extensions the covariant derivative at a point of Σ in a direction tangent to Σ agrees. Thus for V∈TpΣ the vector ∇Vn∈TpM is well defined as ∇X~n~∣p for any smooth extensions n~ of n and X~ of a tangent field with X~p=V (Covariant derivative along a curve, Metric compatibility, torsion freedom and the metric-selected connection class).

(c) The induced metric and second fundamental form. The induced metric hp(V,W):=gp(V,W) is a smooth Riemannian metric on Σ, and Kp(V,W):=gp(∇Vn,W)(V,W∈TpΣ) (the second fundamental form) defines a smooth symmetric covariant 2-tensor field K on Σ (Smooth covariant 2-tensor fields on a smooth manifold) with Kp(V,W)=−gp(∇VW,np), where ∇VW is the ambient covariant derivative at p of any smooth extension of a tangent field with value W at p, in the direction of any extension with value V at p.

(d) The shape operator. The endomorphism Sp:TpΣ→TpΣ defined by hp(SpV,W)=Kp(V,W) for all W∈TpΣ is well defined, linear and self-adjoint for the positive definite form hp, and the Weingarten relation ∇Vn=SpV holds for every V∈TpΣ.

Facts & Assumptions

Given: A time-oriented smooth Lorentzian four-manifold (M,g,N) with metric-selected connection ∇ and a spacelike embedded hypersurface Σ⊆M.

[L1]

For each p∈Σ there is a slice chart (U,y) around p with Σ∩U={y4=0}, and (y1,y2,y3)∣Σ∩U are coordinates of Σ; the restriction of gp to TpΣ is positive definite (Embedded submanifolds and slice charts, Spacelike hypersurfaces, acausal sets, Cauchy sets, domains of dependence and Cauchy horizons clause (a)).

[L2]

For a smooth function f on an open set U and the metric-selected connection, there is a unique smooth vector field ξ=♯ df on U with g(ξ,X)=df(X) for every smooth field X; in components ξa=∑bgab∂bf, so ξ is smooth whenever f and g are, by smoothness of the inverse metric components (The inverse metric, musical isomorphisms and index raising, Smoothness and transformation of the inverse-metric components, Smooth time-oriented Lorentzian four-manifolds: the geometric objects of the model).

[L3]

For an observer n at p (so gp(n,n)=−1) the rest space n⊥={v:gp(v,n)=0} has dimension three and the restriction of gp to it is positive definite (The observer rest-space decomposition and the spatial projection, Observers, rest spaces and the spatial projection at an event). Consequently, if u is a nonzero timelike vector at p and v⊥u is also timelike, then applying this to the observer ±u/−gp(u,u) (Timelike, null, spacelike and causal vectors of a Lorentz vector space) exhibits a nonzero timelike vector in a positive definite space, a contradiction; hence gp(u,v)≠0 for two timelike vectors u,v, and u⊥ is positive definite.

[L4]

∇XY∣p depends only on Xp and on the values of Y along any smooth curve through p with velocity Xp; in particular two smooth fields that agree along such a curve have the same covariant derivative in the direction Xp (Covariant derivative along a curve, Existence and uniqueness of the metric-selected connection of a nondegenerate metric, and its extension to tensor fields).

[L5]

∇ is torsion-free and metric-compatible: ∇XY−∇YX=[X,Y] and X(g(Y,Z))=g(∇XY,Z)+g(Y,∇XZ) for smooth fields (Metric compatibility, torsion freedom and the metric-selected connection class, Existence and uniqueness of the metric-selected connection of a nondegenerate metric, and its extension to tensor fields).

[L6]

If smooth fields X,Y on M are tangent to Σ along Σ, then the bracket [X,Y] is tangent to Σ along Σ: in a slice chart of [L1], write X=Xi∂i+X4∂4, Y=Yi∂i+Y4∂4 with X4∣Σ∩U=Y4∣Σ∩U=0; then the fourth component of [X,Y] is X(Y4)−Y(X4), and at q∈Σ∩U this is dY4(Xq)−dX4(Yq)=0 because Xq,Yq∈TqΣ and dY4,dX4 annihilate TqΣ (Embedded submanifolds and slice charts).

Proof

technique · direct
1.1givenL1L2

(A smooth nonzero normal field on a slice chart.) Fix p∈Σ and a slice chart (U,y) as in [L1], and put f:=y4, so that Σ∩U={f=0} and dfq≠0 for q∈Σ∩U; let ξ:=♯ df on U be the metric dual field of [L2]. Then ξ is smooth on U and nonzero there, and for q∈Σ∩U and X∈TqΣ one has gq(ξq,X)=dfq(X)=0 because X is tangent to the level set {f=0}; hence ξq∈TqΣ⊥∖{0}.

2.1step 1.1givenL1L3

(The normal line is one-dimensional and negative definite.) Let q∈Σ. If v∈TqΣ∩TqΣ⊥, then gq(v,v)=0 because v is orthogonal to itself, while positive definiteness of gq∣TqΣ forces v=0; hence TqΣ∩TqΣ⊥={0}. Therefore the quotient map TqM→TqM/TqΣ restricts injectively on TqΣ⊥, so dim⁡TqΣ⊥≤dim⁡(TqM/TqΣ)=1, and ξq≠0 from step 1.1 shows that the dimension is exactly one and that dim⁡(TqΣ+TqΣ⊥)=4, that is TqM=TqΣ⊕TqΣ⊥. If u∈TqΣ⊥ is nonzero and gq(u,u)≥0, then on the direct sum TqΣ⊕Ru=TqM the form gq is a block sum of a positive definite form and the form gq(u,u)≥0 on the line Ru, hence positive semidefinite with the only possible zero values on Ru; this contradicts the existence of the timelike time-orientation vector Nq with gq(Nq,Nq)<0 (Smooth time-oriented Lorentzian four-manifolds: the geometric objects of the model). Hence gq(u,u)<0 for every nonzero u∈TqΣ⊥: the normal line is negative definite, and in particular nondegenerate.

3.1step 2.1givenL2L3

(Existence and uniqueness of the future unit normal.) Let q∈Σ. The vector ξq is nonzero and, by step 2.1, timelike; so is Nq. By [L3] two timelike vectors at q are never orthogonal, so gq(ξq,Nq)≠0, and nq:=−sign⁡(gq(ξq,Nq))−gq(ξq,ξq) ξq is a well-defined element of TqΣ⊥ with gq(nq,nq)=−1 and gq(nq,Nq)=−∣gq(ξq,Nq)∣/−gq(ξq,ξq)<0. If m∈TqΣ⊥ satisfies gq(m,m)=−1 and gq(m,Nq)<0, then m=tξq for some t∈R because the line is spanned by ξq; gq(m,m)=t2gq(ξq,ξq)=−1 gives t=±1/−gq(ξq,ξq), and the sign of gq(m,Nq)=t gq(ξq,Nq) is the product of the signs of t and gq(ξq,Nq); the condition gq(m,Nq)<0 therefore selects exactly t=−sign⁡(gq(ξq,Nq))/−gq(ξq,ξq), that is m=nq. This proves (a).

4.1step 1.1step 2.1step 3.1givenL2L4

(Smoothness of the normal along Σ.) On the slice chart (U,y) of step 1.1, the functions q↦gq(ξq,ξq) and q↦gq(ξq,Nq) are smooth on U; on Σ∩U the first is negative and the second nonzero by steps 2.1 and 3.1. Shrinking U to a smaller connected neighborhood of a given point of Σ, we may suppose g(ξ,ξ)<0 on all of U and, on the connected set Σ∩U, sign⁡(g(ξ,N)) constant; then the formula of step 3.1 defines a smooth ambient field n~:=−sign⁡(g(ξ,N))ξ/−g(ξ,ξ) on U whose restriction to Σ∩U is n. Such local extensions cover Σ and agree on overlaps because all restrict to n on Σ. Independence of the extensions: let n~1,n~2 be local extensions of n near p and let X~ be a smooth extension of a tangent field with X~p=V; choose a smooth curve γ in Σ with γ(0)=p and γ′(0)=V. Then n~1∘γ=n~2∘γ as curves in TM, so by [L4] ∇X~n~1∣p=∇X~n~2∣p. For a fixed extension n~, the map X↦∇Xn~∣p is additive and C∞(M)-linear by the affine-connection laws (Affine connection on a smooth manifold), so its value depends only on Xp; hence ∇Vn:=∇X~n~∣p is independent of both extensions and is additive in V. This proves (b).

5.1step 4.1givenL1L2L4

(h and K are smooth tensor fields.) On Σ∩U of a slice chart, the components hij=gij∣Σ∩U (i,j∈{1,2,3}) are smooth functions of the slice coordinates and the matrix (hij) is positive definite by [L1]; since Σ is covered by such charts and the local expressions agree on overlaps (both are restrictions of the smooth field g), h is a smooth symmetric positive definite covariant 2-tensor field on Σ, that is, a Riemannian metric. For K, fix p∈Σ and choose a local smooth extension n~ of n and smooth tangent extensions X~,Y~ of fields with X~p=V, Y~p=W; by step 4.1 the number gp(∇Vn,W):=gp(∇X~n~,Y~) does not depend on the three extensions. It is additive and R-bilinear in (V,W) because ∇ and g are, and it is C∞(Σ)-linear in each slot: replacing X~ by φX~ for a smooth function φ on Σ extended locally changes ∇X~n~ by a factor φ by the connection law ∇φX~n~=φ∇X~n~, and replacing Y~ by φY~ multiplies g(∇X~n~,Y~) by φ by bilinearity of g; the chart components of K in slice coordinates, Kij=g(∇∂in~,∂j), are then smooth in the slice coordinates because the connection coefficients and n~ are smooth. Hence K is a smooth covariant 2-tensor field on Σ.

6.1step 5.1step 4.1givenL4L5L6

(Symmetry and the Gauss relation.) Fix p∈Σ and V,W∈TpΣ; extend V,W to smooth ambient fields X,Y tangent to Σ along Σ, and extend n to a smooth ambient field n~ near p (step 4.1). The function g(n~,Y) vanishes identically on Σ because Y is tangent and n~ normal there; differentiating it along V at p and using [L5] gives 0=g(∇Vn~,Y)+g(n,∇VY) at p, that is Kp(V,W)=−gp(∇VW,np); the same argument with V and W exchanged gives Kp(W,V)=−gp(∇WV,np). By torsion-freedom in [L5], ∇VW−∇WV=[X,Y] at p, and [X,Y] is tangent to Σ at p by [L6]; since np is normal, gp([X,Y],np)=0. Therefore Kp(V,W)−Kp(W,V)=−gp([X,Y],np)=0, so K is symmetric and the displayed Gauss relation holds.

7.1step 5.1step 6.1givenL3L5∎

(The shape operator.) Fix p∈Σ. Since hp is a positive definite symmetric form on the three-dimensional space TpΣ, there is an hp-orthonormal basis e1,e2,e3 of TpΣ (Every finite-dimensional real or complex inner product space has an orthonormal basis); define SpV:=∑i=13Kp(V,ei) ei for V∈TpΣ. Then hp(SpV,ej)=Kp(V,ej) for each j, and since {ej} is a basis the identity hp(SpV,W)=Kp(V,W) holds for all W∈TpΣ; this also shows that SpV does not depend on the chosen orthonormal basis. The assignment V↦SpV is linear because Kp is bilinear. Self-adjointness is immediate from symmetry of K and of hp: hp(SpV,W)=Kp(V,W)=Kp(W,V)=hp(SpW,V)=hp(V,SpW). Finally, ∇Vn∈TpM satisfies gp(∇Vn,np)=12V(g(n~,n~))=0 because g(n~,n~)≡−1 on Σ and V is tangent, so ∇Vn∈np⊥; by [L3] np⊥ is three-dimensional, and it contains TpΣ, also three-dimensional, so np⊥=TpΣ and ∇Vn∈TpΣ. For every W∈TpΣ one has gp(∇Vn,W)=Kp(V,W)=hp(SpV,W); since hp is nondegenerate, ∇Vn=SpV. This proves (d) and completes the proof.

Remarks

  • Sign and orientation conventions. With signature (−,+,+,+) and future normal normalized by g(n,n)=−1, this library uses K(V,W)=g(∇Vn,W)=−g(∇VW,n). The global sign of the extrinsic curvature is a convention: Gourgoulhon defines it as the pullback of −∇n, which is −K in the present normalization (his sphere example gives Kij=−(1/R)γij where the present convention gives +(1/R)γij). Replacing n by −n reverses the sign of K. K is not claimed to be positive or negative definite, and no mean curvature sign is fixed here. Consumers must use the definition displayed in this item, not a differing textbook sign.
  • Scope. The lemma is local and kinematic: it constructs the normal from a slice chart and uses only the metric-selected connection, torsion freedom and metric compatibility. No field equation, no causal condition on the model beyond time orientation and no compactness or completeness hypothesis enters.

Depends on

Dependency tree · 2 levels

Sources