Matrix concept Questions (invertibility, det, linear dependence, span)

Click For Summary
SUMMARY

The discussion centers on the relationships between matrix invertibility, determinants, linear independence, and the span of column vectors in square matrices. It is established that if matrix A is invertible, then its determinant is non-zero, indicating that the columns of A are linearly independent and span the matrix. The conversation highlights the importance of understanding these concepts through the lens of linear algebra, as well as the geometric interpretation of determinants as volumes spanned by column vectors. The participants suggest consulting resources like Axler's linear algebra book for deeper insights.

PREREQUISITES
  • Understanding of matrix theory and properties
  • Familiarity with determinants and their properties
  • Knowledge of linear independence and vector spaces
  • Basic concepts of linear transformations and their implications
NEXT STEPS
  • Study the properties of determinants, including the homomorphism property
  • Learn about the rank-nullity theorem and its implications for linear transformations
  • Explore the geometric interpretation of determinants and linear independence
  • Consult Axler's "Linear Algebra Done Right" for comprehensive understanding
USEFUL FOR

Students and professionals in mathematics, particularly those studying linear algebra, as well as educators seeking to clarify the connections between matrix properties and their geometric interpretations.

Sunwoo Bae
Messages
60
Reaction score
4
Homework Statement
For n*n matrix, I would like to know how matrix A being invertible means that det(A) is not 0, that A is linearly independent, and that columns of A spans the matrix.
Relevant Equations
NA
I have a trouble showing proofs for matrix problems. I would like to know how

A is invertible -> det(A) not 0 -> A is linearly independent -> Column of A spans the matrix

holds for square matrix A. It would be great if you can show how one leads to another with examples! :)

Thanks for helping out.
 
Physics news on Phys.org
Really any book on linear algebra has proofs of these facts. Have you consulted one of those? I like Axler's book on linear algebra :)
 
The answer to these questions could be given easily on some higher level. E.g. the first inclusion follows from the fact that the determinant is a group homomorphism. I assume that you didn't want to hear this as explanation. Hence I have to ask you, what you already know.

What is a matrix to you?
What is the determinant to you?
What does "A is linearly independent" mean, if not that the column vectors are linearly independent?

Edit: Do you know the formulas ##\det(A\cdot B)=\det(A)\cdot \det (B)## and ##n= \dim \operatorname{im}(A) + \dim \operatorname{ker}(A)##?
 
Last edited:
fresh_42 said:
The answer to these questions could be given easily on some higher level. E.g. the first inclusion follows from the fact that the determinant is a group homomorphism. I assume that you didn't want to hear this as explanation. Hence I have to ask you, what you already know.

What is a matrix to you?
What is the determinant to you?
What does "A is linearly independent" mean, if not that the column vectors are linearly independent?

Edit: Do you know the formulas ##\det(A\cdot B)=\det(A)\cdot \det (B)## and ##n= \dim \operatorname{im}(A) + \dim \operatorname{ker}(A)##?

I am aware of the the first formula, but not the second one.

The trouble I am having with linear algebra is that when I am given a question, I would know how to solve for determinants and inverse or decide if a matrix is invertible or not. But I do not know how one is related to the other..
 
You can imagine the determinant as the volume of the object spanned by its column vectors. If they are linearly independent, then they span a parallelepiped with positive volume. If not, then they span a hyperplane. But a lower dimensional object in a higher dimensional space has no volume.

The algebraic method is faster, but delivers no insights. If ##A## is invertible, then this means there is a matrix ##B## such that ##A\cdot B = 1##. As the determinant respects multiplication we get from this ##\det(AB)=\det(A)\det(B)=\det(1)=1## and so that ##\det(A)## is a divisor of ##1##, i.e. especially not equal ##0##.

If you have only a formula for the determinant, then you should prove this homomorphism property first, e.g. per induction.

This is why I asked you, what determinant means to you.

"A is linearly independent" is not enough. What does that mean? If I take what you wrote, then I see a vector ##A## in the vector space of linear functions on ##\mathbb{R}^n##. As a single vector is always linear independent as soon as it is different from the zero vector, this statement is trivially true and has absolutely nothing to do with the determinant of ##A##. Hence I assumed, that you meant something else. However, the
closest possibility to interpret what you might have meant, is to translate it by "the column vectors of ##A## are linear independent in ##\mathbb{R}^n##. But the I have ##n## linear independent vectors, and of course they span an ##n-##dimensional space. So where is the problem?

This is why I asked you about the meaning of "A is linearly independent". Linear independence always requires a reference: Where do the vectors live? When did you switch from matrix to vector? Matrices can only be linear independent if considered as vectors in some vector space.

The rank formula is the best way to see the implication from ##A## to its column vectors.

The remaining implication ##\det(A)\neq 0 \Longrightarrow ## "column vectors of ##A## are linearly independent" is best seen the other way around. Assume a non trivial linear dependence of the column vectors and see what is does to the determinant. One can handle this with the properties of the determinant, or the properties of ##A## as linear mapping.

This is why I asked you what ##A## means to you: a number scheme or a linear function?

Here is a good read about the geometry of such things:
https://arxiv.org/pdf/1205.5935.pdf
I don't know whether it will answer your questions above, but it will definitely help you to imaginate the objects you are dealing with.
 

Similar threads

  • · Replies 25 ·
Replies
25
Views
3K
Replies
15
Views
2K
Replies
4
Views
2K
  • · Replies 4 ·
Replies
4
Views
1K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
Replies
2
Views
2K
Replies
4
Views
2K
Replies
2
Views
2K