Re: linear algebra - basic determinant properties, how intuitive are they?



In article <1112846012.062175.267150@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"porky_pig_jr@xxxxxxxxxxx" <porky_pig_jr@xxxxxxxxxxx> wrote:

> In the intro course in Linear Algrebra,


Your word "the" is a bit strong. There are legions of linear
algebra texts, handling determinants in various ways.


> the concept of determinant is
> presented by starting with three elementary requirements or properties.
> Subsequently everything else (including the permutation-based formula
> and cofactors) is derived from those three properties which are
>
> (i) determinant of identity = 1
> (ii) determinant is linear in a single row
> (iii) exchanging two rows reverses the sign of determinant
> ....
> So I am a bit puzzled - where does this 3rd requiremeent arise from?....


I prefer the approach in the little book by C. C. MacDuffee, "The
Theory of Matrices" (reprinted by Chelsea), which gives a fine summary
of linear algebra in the traditional matrix-and-determinant form as it
was in 1933.

MacDuffee pp. 6-7 introduces determinants by a method due to K.
Hensel (1928). In his slightly old-fashioned language he says:

"It is desirable to have associated with every matrix A of [his algebra
of n x n matrices] a number theta(A) of [his field of scalars] which
would serve the purpose of an absolute value. This end would be
attained by finding a scalar function theta(A) of the elements a_rs of a
general matrix A such that
1. For every A, theta(A) is a non-constant rational integral
function [i.e. polynomial] of the a_rs of lowest degree such that
2. theta(AB) = theta(A) theta(B)."

Beginning from there he deduces the three properties which you
mentioned above, attributing them to Weierstrass; then derives from them
the definition in terms of even and odd permutations.

Ken Pledger.
.



Relevant Pages

  • Re: optimization with complex numbers
    ... packages) the Solver tool can solve the linear programming problem I ... because the determinant of the coefficient matrix is a nonzero complex ... whether the field of scalars is the rationals, the reals or the ... we again get the identically zero solution in 8 variables. ...
    (sci.math)
  • Re: optimization with complex numbers
    ... packages) the Solver tool can solve the linear programming problem I ... because the determinant of the coefficient matrix is a nonzero complex ... we again get the identically zero solution in 8 variables. ...
    (sci.math)
  • Re: optimization with complex numbers
    ... linear feasibility problem that can be solved using linear programming ... coefficients is nonsingular (i.e., has nonzero determinant, etc.) then ...
    (sci.math)
  • Re: vandermonde matrix question
    ... transformation is linear. ... > If I change the row order, does the determinant change? ... > I would think that I should perform row operations on ...
    (sci.math)
  • Re: Clear writing about relativity
    ... Most physicists ... algebra, ... Hence tauis not linear. ...
    (sci.physics.relativity)