I have a two-component field:

$$\phi(\vec{x}) = \left( \begin{array}{c} \phi_1(\vec{x}) \\ \phi_2(\vec{x}) \end{array} \right)$$

with $\phi^T = (\phi_1, \phi_2)$. And I am trying to evaluate:


I'm pretty sure the result of that expression is supposed to be a scalar, but I can't figure out why. $\phi$ is a 2x1 matrix, which means $\vec{\nabla}\phi$ must be the result of a matrix multiplication between a Nx2 and 2x1 matrix. You would think N would be 3, since $\vec{\nabla} = \frac{\partial}{\partial x}\vec{x} + \frac{\partial}{\partial z}\vec{z} + \frac{\partial}{\partial z}\vec{z}$, but then why are there 2 columns? Not sure what to do here so any assistance would be quite helpful.

(I'll mention that the full thing I'm trying to evaluate is $\mathcal{L} = \frac{1}{2}\left(\dot{\phi}^T\dot{\phi}-m(\vec{\nabla}\phi)^T(\vec{\nabla}\phi)\right)$. $\mathcal{L}$ is a scalar, right?)

  • 1
    $\begingroup$ It's not straight forward notation, but my guess is that $\nabla\phi$ sound be interpretet as the vector with six components, $\tfrac{\partial}{\partial x^1}\phi_1, \tfrac{\partial}{\partial x^2}\phi_1, \dots, \tfrac{\partial}{\partial x^3}\phi_2$ and then $(\vec{\nabla}\phi)^T(\vec{\nabla}\phi):=\sum_{i\in\{1,2,3\}}\sum_{j\in\{1,2\}}(\tfrac{\partial}{\partial x^i}\phi_j)^2$ $\endgroup$ – Nikolaj-K Nov 27 '13 at 10:39
  • $\begingroup$ More naturally, you would consider $\vec{\nabla}\phi$ to be a 3-component vector of 2-component vectors. It's not good to have the same index stand for two things, because it gets messy. Mathematically, of course, it's equivalent. $\endgroup$ – lionelbrits Nov 27 '13 at 12:58

When you have a matrix $\Phi = \begin{pmatrix} \phi_1\\ \phi_2\end{pmatrix}$, with one column and two rows, and its transpose matrix $\Phi^T = \begin{pmatrix} \phi_1 & \phi_2\end{pmatrix}$, with one row and two columns, the product of the two matrix $\Phi^T \Phi$ is a matrix $P$ with one column and one row:

$P =\Phi^T \Phi = \begin{pmatrix} \phi_1 & \phi_2\end{pmatrix}\begin{pmatrix} \phi_1\\ \phi_2\end{pmatrix} = (\phi_1^2+\phi_2^2)$

Because this matrix $P$ has one row and one column, it may considered as a scalar.

For your particular problem, you have :

$(\vec{\nabla}\phi)^T.(\vec{\nabla}\phi) = \sum\limits_{i=1}^N ({\partial_i}\phi)^T({\partial_i}\phi) = \sum\limits_{i=1}^N ((\partial_i \phi_1)^2 + (\partial_i \phi_2)^2)$

The first equality comes from the definition of the inner product and the gradient, and the second equality comes from the definition of the transpose operation $T$, and the manipulation of these matrices, as seen at the beginning of the answer.

| cite | improve this answer | |

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.