February 29th, 2012, 04:26 PM  #1 
Newbie Joined: Feb 2012 Posts: 4 Thanks: 0  Lorentz matrix problem
I looked at a copy of "Problem Book in Quantum Field Theory" by Voja Radovanovic, and it has a problem where you need to prove properties of a matrix based on the property that it doesn't change the "spacetime interval" of a vector. There's an argument in the solution of the problem that I don't understand, and I'm hoping someone can help me with it. I'll summarize the problem and proof in matrix language instead of index notation. Given a matrix , and a vector , define . Show that if for every x, then The proof is short: he just substitutes the definition, then does a little algebra to get . Then he claims that since this is true for *every* x, that . The last part I don't understand. For example, for *every* x, but certainly those matrices are not equal. I don't know if I caught an error in the proof or if I'm just not understanding something about it. Does anyone have a more detailed proof of this that shows how you get from to ? 
February 29th, 2012, 04:37 PM  #2 
Senior Member Joined: Feb 2012 Posts: 628 Thanks: 1  Re: Lorentz matrix problem
In the notation , is that g times the dot product , or is it the dot product of and , or is it something else entirely? Obviously the interpretation g times x times x makes no sense (under matrix multiplication, anyway).

February 29th, 2012, 08:41 PM  #3 
Newbie Joined: Feb 2012 Posts: 4 Thanks: 0  Re: Lorentz matrix problem
It is (gx) * x.

February 29th, 2012, 08:50 PM  #4  
Newbie Joined: Feb 2012 Posts: 4 Thanks: 0  Re: Lorentz matrix problem
Here's the original text for those who enjoy index notation or want to see it in situ. Quote:
 
March 1st, 2012, 11:17 AM  #5 
Newbie Joined: Feb 2012 Posts: 4 Thanks: 0  Re: Lorentz matrix problem
Well, I managed to figure this out after a bit more pondering. It's necessary to use the properties of g on both sides of the equation. For arbitrary matrices A and B, where for all x, you can show that their entries on the main diagonal must be equal by picking each unit vector in the basis and substituting it for x. For example, picking in yields Now that we know the diagonals are equal, you can show by picking each vector in the basis with two 1's in it (e.g. (0, 1, 1, 0)) that . For example, picking in yields which implies since for all indices n. Now going from the general case to the special case where B = g, and i ? j, then , so the entries off diagonal are antisymmetric. Finally, we know that is symmetric, since g is symmetric and . The only way that the offdiagonal entries can be both symmetric and antisymmetric is if they are zero, and the ondiagonal entries must be equal, so as expected. The same proof seems like it would work for any diagonal matrix substituted for g, so there's probably some slightly more sophisticated linear algebra that would make short work of this through some decomposition or something. I think the more general statement would be for any matrix A and diagonal matrix D, if for all x, then . Is this an obvious linear algebra thing? Because that proof in the book glossed over it like it was nothing. 

Tags 
lorentz, matrix, problem 
Thread Tools  
Display Modes  

Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
matrix problem  golomorf  Linear Algebra  0  February 17th, 2013 10:47 AM 
3d Matrix problem  Marvin174  Linear Algebra  0  March 19th, 2012 01:09 PM 
Derivation of the Lorentz Transformation  ben.g95  Physics  0  October 12th, 2011 04:20 PM 
Matrix problem.  halloweengrl23  Linear Algebra  2  December 3rd, 2010 02:00 PM 
Matrix problem  jlfr  Linear Algebra  2  March 30th, 2008 07:02 PM 