Objects in a linear world
If addition and scaling are the only allowed ways to combine elements, what sort of objects can live in such a world at all?
What laws make linearity into a genuine structure?
The previous article drew the boundary of the linear world. It told us what must be excluded: interaction terms, powers, and all the nonlinear combinations that make expressions harder to control.
A stable structure needs laws as well as boundaries. Once only addition and scaling remain, the real question is what laws make that restricted world hold together.
The first requirement is closure.
If u and v are allowed elements, then u + v must again be an allowed element. If a is a scalar and v is allowed, then a·v must also stay inside the same world.
Closure gives the world stability. It turns the set into a setting where the defining operations keep us inside the same mathematical space.
Closure is the first layer. Coherence adds the next.
If addition is going to represent the accumulation of independent contributions, then it must behave coherently. Grouping leaves the result unchanged, and order leaves it unchanged as well. There must also be a neutral element, usually written 0, that leaves each vector in place under addition. And every element must have an additive inverse, so that contributions can balance each other when the structure demands it.
So the additive side keeps the same disciplined shape as before: addition must be associative and commutative, there must be a zero element, and every element must have an additive inverse.
Scaling has its own coherence conditions.
Scaling by 1 leaves each vector in place. Scaling by a and then by b agrees with scaling once by the product ab. And scaling cooperates with the additive structure. If we first add two elements and then scale, we should get the same result as scaling each one first and then adding:
a·(u + v) = a·u + a·v
Likewise, if we first add two scalars and then let them act on one element, that should agree with adding the two resulting contributions:
(a + b)·v = a·v + b·v
These are exactly what make superposition trustworthy. Separate contributions stay separate as they combine. Scaling spreads through the sum instead of creating new interaction terms.
V × V ──+──▶ V
│ │
a· a·
│ │
V ───────▶ V
Read it this way. The objects in the world can be added, and scalars can act on them. The allowed operations cooperate. If we scale contributions and then combine them, or combine first and then scale, the result agrees in the precise way the linear laws demand.
Once those laws are imposed over a chosen field of scalars, the resulting object is a vector space.
That name can mislead at first. The definition speaks only of a set whose elements may be added and scaled coherently. Pictures of arrows, coordinates, lengths, and geometry come later.
This is why the concept matters so widely. The same structure appears whenever effects combine by accumulation and respond proportionally to rescaling. The linear object is the mathematical form of that disciplined behavior.
It also marks an important simplification relative to the polynomial world. In a polynomial ring, multiplying terms can create new mixed expressions of higher degree. In a vector space, the laws keep every combination within addition and scaling, so composition stays transparent.
That gives the linear world its stability. It is algebra after interaction has been deliberately removed and coherence has been tightened around what remains.
So the objects are now clear: sets equipped with addition and scalar multiplication satisfying the linear laws. The next step is to identify the right transformations between them.
If a linear object is defined by coherent addition and scaling, what should a map between two such objects be required to respect before it deserves to count as a genuine linear transformation?