So when proving that the gradient vector of some function $g(x,y,z)$ is always orthogonal to a level set of $g$, my book uses differentiable curves which lie on this level set. As far as I understand, these curves are guaranteed to exist by the Implicit Function Theorem, but I'm having trouble understanding why.

For a level set in $\textbf{R}^2$ I understand, because the implicit function theorem says we can solve for $y$ as a function of $x$ or the other way around. So we can get a function $y=f(x)$ and then we can just define the curve $\varphi(t) = \langle t, f(t) \rangle$. But how does the same conclusion follow for a level set in $\textbf{R}^3$?

  • $\begingroup$ It goes through the same: you can solve for one of the three variables, say $z$, in terms of the other two. However many equations you have, you lose that many independent variables (in the generic situation). $\endgroup$ – Ian Oct 9 '15 at 3:04
  • $\begingroup$ I guess the reason I'm confused is because I thought if you have a parametric equation with two variables, then it is a parametric surface, not a curve. So like if you solved for z as a function of x and y then made a parametric equation out of that. $\endgroup$ – Fgilan Oct 9 '15 at 3:10
  • $\begingroup$ Yes, you get a parametric surface, not a curve, from a single scalar equation in three variables. You can then look at curves on this surface if you want. $\endgroup$ – Ian Oct 9 '15 at 3:23

The implicit function theorem shows that there is a local map $z : \Bbb R^2 \to \Bbb R$ such that $f(x, y, z(x, y))$ is constant. Any curve in the domain of $z$ in $\Bbb R^2$ is carried by the map $(x, y) \mapsto (x, y, z(x, y))$ to a curve in the level surface.


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.