To build a robust model of consumer behavior, we can't rely on vague notions of 'wanting' things. We need a formal structure. Consumer theory rests on a set of axioms that define what it means to be a rational agent. These axioms aren't meant to describe how people actually think, but to provide a consistent mathematical foundation for modeling their choices.

The first two axioms establish a clear ordering of preferences. Let $X$ be the set of all possible consumption bundles. For any two bundles $x, y \in X$:

Together, completeness and transitivity guarantee that a consumer's preferences can be represented by a preference relation, $\succsim$, which is a complete and transitive ordering on the consumption set $X$. This ordering is the bedrock upon which utility functions are built.

With a preference ordering established, we need an axiom that connects these preferences to actual market behavior. This is the role of Local Non-Satiation. It's a weaker, more flexible assumption than simple monotonicity.

In simpler terms, Local Non-Satiation means you can always find a slightly different bundle that is strictly better, no matter where you are in the consumption space. There is no 'bliss point' where a consumer is perfectly content. An immediate and powerful implication of LNS is that a utility-maximizing consumer will always exhaust their budget. If they didn't, there would be some money left over, which could be used to buy a little more of something, leading to a preferred bundle. Therefore, the budget constraint is always binding: $p \cdot x = m$.

The next critical element is the shape of the preferences, defined by convexity.

The distinction between these types is critical. Strict convexity implies a preference for variety; averages are strictly preferred to extremes. This leads to unique interior solutions for the utility maximization problem. Convexity that is not strict allows for flat spots on indifference curves, where the consumer is indifferent between a range of bundles. Non-convex preferences indicate a preference for specialization, leading to corner solutions.

The ultimate goal is to derive a demand function, $x(p, m)$, which gives the optimal consumption bundle for any given price vector $p$ and income $m$. The axioms we've discussed determine whether such a function exists and what properties it has.

If preferences are complete, transitive, continuous, and strictly convex, the utility maximization problem will yield a unique optimal bundle for any given budget. This allows us to define a single-valued, continuous demand function. This is the well-behaved world of standard microeconomic models.

Non-convex preferences also cause problems, often leading to discontinuous or "jumpy" demand behavior. As prices change slightly, the optimal bundle might jump from one corner of the budget set to another, violating the continuity required for many analytical tools.

These axiomatic foundations are what allow us to model consumer choice with mathematical rigor. They define the precise conditions under which the familiar, well-behaved demand curves of introductory economics can be derived.