In mathematics, a linear map is a rule that connects two vector spaces while preserving their core arithmetic structure. This concept appears everywhere from solving systems of equations to describing rotations and scaling in computer graphics. At its core, a linear map ensures that addition and scalar multiplication behave consistently before and after the transformation.

Thinking of vectors as arrows or points in space helps visualize what a linear map does. When you apply such a map, straight lines remain straight, the origin stays fixed, and the grid of coordinates is transformed in a perfectly uniform way. This predictability makes linear maps a powerful tool for modeling relationships where proportional changes in input lead to proportional changes in output.

Definition and Core Properties
A linear map between two vector spaces must satisfy two fundamental conditions for every pair of vectors and any scalar. First, the map preserves vector addition, meaning the image of a sum equals the sum of the images. Second, it preserves scalar multiplication, so stretching a vector before mapping yields the same result as mapping and then stretching.

These two rules can be combined into a single statement about linear combinations. For any vectors and scalars, the map of a combination is the same combination of the maps. This property ensures that the structure of the entire vector space is maintained in the target space, which is why linear maps are also called linear transformations.
Additivity and Homogeneity

Additivity requires that for any vectors u and v, the map of u + v is the map of u plus the map of v. This captures the idea that combining inputs and then transforming them gives the same result as transforming each input separately and then combining. Visually, this prevents the map from introducing bends or curves that would break the straightness of lines.
Homogeneity states that for any vector v and scalar c, the map of c v equals c times the map of v. This condition controls how the map handles changes in magnitude, ensuring that directions are preserved or reversed consistently. Together, additivity and homogeneity guarantee that the transformation respects the linear structure of the spaces involved.
Matrix Representation

When vector spaces are finite-dimensional, every linear map can be described by a matrix. Choosing bases for the domain and codomain allows us to write the map as a rectangular array of numbers that encodes how each basis vector is transformed. Multiplying this matrix by a coordinate vector yields the coordinates of the image under the linear map.
This matrix representation makes computations concrete and links abstract linear algebra to numerical algorithms. Different choices of bases lead to different matrices for the same map, but the underlying transformation remains unchanged. The matrix also reveals key properties such as rank, nullity, and invertibility, which are essential for understanding the behavior of the linear map.
Examples and Applications

Simple geometric operations in the plane provide clear examples of linear maps. Rotating every point around the origin by a fixed angle is a linear transformation, as is scaling all coordinates by a positive factor. Reflections across lines through the origin also satisfy the defining properties, because lines through the origin remain fixed as sets.
In physics and engineering, linear maps model relationships where output is directly proportional to input. Systems described by linear differential equations often rely on linear operators that act on function spaces. In data science, techniques such as principal component analysis use linear maps to project high-dimensional data onto lower-dimensional subspaces while preserving variance.




















Derivatives as Linear Maps
In calculus, the derivative at a point can be viewed as a linear map that best approximates a function near that point. This linear map, called the differential, captures the instantaneous rate of change in every direction. It takes a small displacement vector and outputs the corresponding change in the function value to first order.
This viewpoint unifies many concepts in multivariable calculus, where higher-dimensional derivatives become linear transformations between spaces of different dimensions. The ability to approximate nonlinear behavior by linear maps is foundational for optimization, stability analysis, and numerical methods.
Kernel and Image
The kernel of a linear map consists of all vectors that are sent to the zero vector, forming the set of directions collapsed by the transformation. The image is the set of all possible outputs, representing the subspace spanned by the transformed basis vectors. Both the kernel and image are themselves vector spaces, providing insight into the structure and limitations of the map.
Understanding the kernel helps identify dependencies and redundancies in the mapping, while the image reveals the effective dimensionality of the output. These concepts are formalized in the rank-nullity theorem, which links the dimensions of the kernel and image to the dimension of the domain space, offering a powerful tool for analyzing linear transformations.
Grasping what a linear map is opens the door to a wide range of mathematical ideas and practical techniques. By focusing on how addition and scaling interact with transformations, this concept provides a clear and flexible framework for studying relationships between vectors and spaces. Exploring further reveals deeper connections that continue to shape modern mathematics and its applications.