Topic: Kernels , Subspaces , and Structural Correspondence.

Introduction: The exercise below is the "big picture" where linear algebra connects to the broader world of abstract algebra. It establishes that vector spaces aren't just isolated islands - they are part of a family of structures that follow similar rules for maps and "loss of information".
We have seen how a linear transformation `phi` acts on a basis to "move" a space. Now , we investigate the internal structure of the transformation itself. Just as some functions in algebra collaps many inputs to one (like `x^2` sending 1 and - 1 to 1) , linear transformations often "collaps" parts of the vector space. Understanding what is lost during this collaps is the key to understanding the transformations "honesty" (bijectivity).

Exercise: Let `bbb"V"` and `bbb"W"` be vector spaces over the same field `bbb"F"` , and let `phi : bbb"V" rarr bbb"W"` be a linear transformation.

(a) To what concept that we have studied for the algebraic structures of groups and rings does the concept of a linear transformation correspond?

(b) Define the kernel (or nullspace) of `phi` , and show that it is a subspace of `bbb"V"`.

(c) Describe when `phi` is an isomorphism of `bbb"V"` with `bbb"W"`.

Solution:
(a) Correspondence to Groups/Rings.
A linear transformation `phi` corresponds to a Homomorphism.
- In Groups , a homomorphism preserves the group operation `f(a + b) = f(a) + f(b)`.
- In Rings , it preserves both addition and multiplication.
- In Vector Spaces , a linear transformation is specifically a Module Homomorphism (preserving addition and scalar multiplication.

(b) The Kernel (Nullspace).
The Kernel of `phi` , denoted `ker(phi)` , is the set of vectors `v in bbb"V"` such that `phi(v) = 0_{bbb"W"}`.

Proof it is a subspace:
1. Zero Vector: Since `phi(0_{bbb"V"}) = 0_{bbb"W"}` (by linearity) , `0_{bbb"V"} in ker(phi)`.
2. Closure under addition: If `u , v in ker(phi)` , then `phi(u + v) = phi(u) + phi(v) = 0 + 0 = 0`.
3. Closure under scaling: If `v in ker(phi)` , then `phi(av) = a phi(v) = a (0) = 0`.
Thus , the kernel is a perfectly valid "subspace" living inside `bbb"V"`.

(c) Isomorphism Conditions.
`phi` is an isomorphism if and only if it is:
1. Injective (One-to-One): This happens if and only if `ker(phi) = {0}`. (No "extra" vectors are collapsed to zero).
2. Surjective (Onto): This happens if the range of `phi` is the entire space `bbb"W"`. QED

Result: The kernel measures the "failure" of a map to be an isomorphism. The bigger the kernel , the more information is "lost" as you move from `bbb"V"` to `bbb"W"`. END

Transition Note: `"The Lie Algebra Minute"`.
Note: In our last Lie Algebra exercise , we calculated `ad_e`. We found that `ad_e(e) = 0`. This means that the basis vector `e` is in the Kernel of the transformation `ad_e`. Since the kernel is not just the zero vector , `ad_e` is a "collapsing map" - it is not an isomorphism of the algebra onto itself! In the next page we will see if `ad_f` behaves the same way.