Could someone show why for the Hessian to be well defined ($d_{p}^{2}f(v,w) = L_{v}L_{w}f$) we need $p$ to be a critical point.


Let $H_p(U, V) = UV f(p)$. Then $H_p(U, gV) = U(gV f) = (Ug)(Vf) + gUVf = gH_p(U, V)$ as $Vf=0$ at $p$. Thus $H_p$ is $C^\infty$ linear and defines a quadratic form. Note that similar argument shows that $H_p$ is symmetric

  • $\begingroup$ i am assuming V is supposed to be the Jacobian, since you mention Vf = 0, and what is U? $\endgroup$ – masszz Nov 5 '13 at 7:58
  • $\begingroup$ Sorry about the notation, $U$ and $V$ are both vectors fields. $\endgroup$ – user99914 Nov 5 '13 at 8:00
  • $\begingroup$ And $Vf = L_V f$. $\endgroup$ – user99914 Nov 5 '13 at 8:01
  • $\begingroup$ got it. Thanks you very much. One more thing, i am asked to find the Hessian in local coordinates, now i just wrote the Hessian matrix (matrix of second partials), is this generally a correct notation? $\endgroup$ – masszz Nov 5 '13 at 8:04
  • $\begingroup$ You are correct only when you are at a critical point. $\endgroup$ – user99914 Nov 5 '13 at 8:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.