Generalizing: Curl of a Function

  • Thread starter Thread starter pbandjay
  • Start date Start date
  • Tags Tags
    Curl Function
pbandjay
Messages
117
Reaction score
0
Hello. I have tried to search all day for the answer to my question but could not find anything.

Is there a way to generalize the curl of a vector field to dimensions greater than three? It seems pretty straightforward to find the gradient and divergence of higher dimensional vector fields, but my calculus book only offers the definition: curl F := ∇×F, which appears to only work if F has three components. Or am getting something wrong?

I am also only in my first real analysis course, so I have not come across this kind of topic in that class yet.
 
Physics news on Phys.org
Not easily. The problem is that the best way to generalize the cross product itself is to use the "alternating tensor". That is the operator represented by \epsilon_{i_1i_2\cdot\cdot\cdot i_n} equal to 1 if i_1i_2\cdot\cdot\cdot i_n is an even permutation of 1 2 3 \cdot\cdot\cdot\ n, -1 if it s an odd permutation, and 0 if neither of those is true.
For example, when n= 3, \epsilon_{ijk} has 27 entries. \epsilon_{123}= \epsilon_{231}= \epsilon_{312}= 1, \epsilon_{132}= \epsilon_{213}= \epsilon_{321}= -1, and the other 21 entries are all 0.

With that w_i= \sum_{j=1}^3\sum_{k=1}^3\epsilon_{ijk}u_jv_k gives w_1= \epsilon_{123}u_{2}v_{3}+ \epsilon_{132}u_3v_2= u_2v_3- u_3v_2
w_2= \epsilon_{213}u_1v_3+ \epsilon_{231}u_3v_2= -u_2v_3+ u_3v_2
w_3= \epsilon_{312}u_1v_2+ \epsilon_{321}u_2v_1= u_1v_2- u_2v_3
precisely the cross product.


But, in n dimensions, \epsilon has n indices. In order to get a vector result, we would have to multiply n-1 vectors, not just two.
 
In a general space whose dimension is not necessarily three, the generalization of curl is a Hodge dual of an exterior derivative of the vector field. The vector field is a 1-form, its exterior derivative is a 2-form, and its Hodge dual is a (n-2)-form. Which is a scalar (a number) for n=2, a vector for n=3, a higher-rank tensor for n>=4.
 
Thank you for the help! Looks like I have a few things to learn about the topic.
 
Back
Top