1

Is there any reference on whether expressions of mixed order tensors, for instance:

$$a = v_iT^{i}_{.j}u^j$$

which evaluate to the same scalar (irrespective of the order of these terms above) correspond exactly to:

$$a = \begin{bmatrix} v_1 & \cdots &v_n \end{bmatrix} \begin{bmatrix} T^1_1 &\cdots &T^1_n \\ \vdots & \ddots & \vdots \\T^n_1 & \cdots & T_n^n \end{bmatrix} \begin{bmatrix} u_1 \\ \vdots \\u_n \end{bmatrix}$$

if the contracted indices obey a particular order (covariant first, contravariant second, as the matrix product above or the other way around)?

In other expressions, such as those with cartesian tensors like: $$E = \frac{1}{2}\int I_{ij}\omega_i \omega_jdV$$ can we show that any matrix combination of these terms that give the resulting matrix in the correct dimensions (1x1 in this case) will result in the correct contraction/operation (given that we can raise and lower indices as we like, as I suppose the metric is simply the identity matrix)?

Qmechanic
  • 201,751
Petrini
  • 322
  • 1
    Related: https://physics.stackexchange.com/q/20437/2451 – Qmechanic Dec 31 '20 at 00:05
  • In the case of two second order tensor contractions, the question arises if $T^i_{.j} W^j_{.i}$ can be calculated using $v_i T^i_{.j} u^j$ if $W^j_{.i} = u^j v_i$ and this outer product translates into matrix product $\mathbf{u}^T \mathbf{v}$ if we assume bold vectors are columns. – Petrini Dec 31 '20 at 00:10

0 Answers0