91Ó°ÊÓ

A subspace is essentially a smaller, contained vector space, which sits inside a larger vector space. It must adhere to certain rules that preserve its structure as a vector space.
To qualify as a subspace, a set must satisfy three crucial properties:

• Closure Under Addition: If you take any two vectors in the subspace and add them together, the resulting vector should still be in the subspace.


• Closure Under Scalar Multiplication: If you multiply any vector in the subspace by a scalar (a real number), the result should remain in the subspace.


• Zero Vector: The subspace must contain the zero vector, the origin of the vector space, ensuring that it includes a neutral element for addition.

These properties are critical because they ensure that operations within the subspace do not take you outside its boundary, maintaining consistency within the subspace. This structure is used extensively in linear transformations to explore ideas like the inverse image and kernel.

The concept of an inverse image in linear algebra pertains to mapping elements from one space back to their pre-mapped counterparts in another. When dealing with a linear transformation \( T: V \rightarrow W \), finding the inverse image \( T^{-1}(U) \) involves identifying those vectors in \( V \) that map into a specific subspace \( U \) of \( W \).
The inverse image \( T^{-1}(U) \) has to be a subspace of \( V \).
Using the rules of subspaces:

• When two vectors from \( T^{-1}(U) \) are added, their image via \( T \) is also in \( U \) because \( U \) is a subspace.


• Similarly, multiplying a vector from \( T^{-1}(U) \) by a scalar keeps the transformed result within \( U \), ensuring closure under scalar multiplication.


• The zero vector in \( V \) must map to the zero vector in \( U \), and this establishes the foundation of the zero element in \( T^{-1}(U) \).

Thus, the set of vectors that form the inverse image also forms a subspace of the domain \( V \), adhering to the same consistent operations as in \( U \). This establishes a strong connection between transformations and their inverse mappings.

The kernel, often referred to as the nullspace, is a foundational concept in linear algebra focused on understanding the structure of linear transformations.
Mathematically, the kernel of a transformation \( T: V \rightarrow W \) consists of vectors in \( V \) that map to the zero vector in \( W \).
You can think of the kernel as capturing how much of \( V \) collapses into \( W \) at the zero vector.
This can be visualized as follows:

• When you run the transformation, any vector that ends up at the zero vector in \( W \) is part of the kernel.


• The kernel \( T^{-1}(\{\mathbf{0}\}) \) inherently forms a subspace of \( V \). It satisfies closure under addition and scalar multiplication simply because zero maps are inherent in vector space calculations.


• It’s critical to note that when the kernel comprises only the zero vector, the transformation \( T \) is injective (one-to-one).

Understanding the kernel is pivotal to decoding a transformation’s properties, as it provides insights into the linear transformation's nullity and efficiency in mapping vectors.