Historical Context & Motivation
Single-variable calculus, as developed by Newton and Leibniz in the seventeenth century, provided powerful tools for analyzing quantities that depend on one independent variable. However, the physical world rarely cooperates with such simplicity: temperature depends on spatial position and time, atmospheric pressure depends on latitude, longitude, and altitude, and economic output depends on labor, capital, and technology simultaneously. The need to handle such dependencies drove mathematicians toward a rigorous theory of multivariable functions—functions whose inputs are ordered tuples of real numbers rather than single scalars.
The central question this lesson addresses is deceptively simple: What does it mean for a function to accept two or three inputs, what is its natural domain, and how can we visualize its output without requiring four-dimensional graphs? The answer lies in the elegant concept of level sets—slices of the output at constant values that reduce the visualization problem by one dimension.
Core Principles & Definitions
Before diving into computations, we establish the foundational vocabulary and ideas that govern multivariable functions. These principles generalize familiar single-variable concepts—domain, range, and graph—to higher dimensions, while introducing the new idea of level sets as a dimensionally reduced portrait of a function's behavior.
Function of Two Variables
Function of Three Variables
Natural Domain
Level Curves (Contour Lines)
Level Surfaces
Visual Explanation — Level Curves of f(x, y) = x² + y²
The most illuminating way to grasp level sets is to start with a concrete example. Consider the paraboloid z = x² + y². Its graph in three-dimensional space is a bowl opening upward with vertex at the origin. To form a contour map, we slice this surface with horizontal planes z = c for several non-negative constants c and project each resulting circle down onto the xy-plane. The diagram below shows the result: a family of concentric circles centered at the origin, where the radius of each circle is √c.
Several features of this contour map generalize to all functions of two variables. First, level curves for different values of c never intersect, because a single point (x, y) cannot produce two different output values simultaneously (the function must be well-defined). Second, the spacing between adjacent level curves encodes the rate of change of the function: closely packed contours mean the function changes rapidly, while widely spaced contours indicate gradual change. These observations connect directly to the gradient vector, which points perpendicular to the level curve in the direction of steepest ascent—a topic explored in subsequent lessons on partial derivatives.
Mathematical Framework
We now formalize the definitions introduced qualitatively in Section 2 and develop the algebraic machinery for determining domains and level sets.
Determining the Natural Domain
Finding the natural domain requires identifying all algebraic restrictions that prevent the formula from producing a real number. The most common restrictions are as follows: (1) division by zero—any expression in the denominator must be nonzero; (2) even roots of negative numbers—the radicand of a square root (or fourth root, etc.) must be non-negative; (3) logarithms of non-positive numbers—the argument of ln must be strictly positive. When multiple restrictions apply simultaneously, the domain is the intersection of the individual admissible regions. Geometrically, one should describe the domain as a region in ℝ² or ℝ³ using inequalities, and, when feasible, sketch it.
Detailed Breakdown — Domains and Level Set Sketching
To solidify understanding, we examine several representative functions, determine their domains, and classify their level sets. The diagram below illustrates the domain of a function with both a square root and a denominator restriction, and then we present a classification table relating common function forms to their level curve families.
| Function Form | Natural Domain | Level Curves (c = const) |
|---|---|---|
| f(x, y) = x² + y² | All of ℝ² (no restrictions) | Circles centered at origin, radius √c, for c > 0; point at origin for c = 0 |
| f(x, y) = x² − y² | All of ℝ² | Hyperbolas with axes along x and y; two lines y = ±x when c = 0 |
| f(x, y) = ln(x² + y² − 1) | x² + y² > 1 (exterior of unit circle) | Circles of radius √(eᶜ + 1) centered at origin, for all c ∈ ℝ |
| f(x, y) = √(x − y) | x ≥ y (half-plane above line y = x) | Parallel lines y = x − c² for c ≥ 0 |
| f(x, y) = eˣ sin(y) | All of ℝ² | Curves defined by sin(y) = c·e⁻ˣ; family shifts with c |
Worked Example — Domain and Level Curves
We work through a complete example that combines domain analysis with level curve identification. Consider the function:
Level Curves vs. Level Surfaces — Strengths & Limitations
Both level curves and level surfaces serve the same fundamental purpose—reducing the dimensionality of a visualization problem by fixing the output—but they operate in different settings and come with distinct trade-offs. The table below highlights key comparisons that help clarify when and why each tool is most effective.
| Feature | Level Curves (2 Variables) | Level Surfaces (3 Variables) |
|---|---|---|
| Input dimension | (x, y) ∈ ℝ² | (x, y, z) ∈ ℝ³ |
| Output dimension | z = f(x, y) ∈ ℝ | w = f(x, y, z) ∈ ℝ |
| Graph dimension | Surface in ℝ³ (viewable) | Hypersurface in ℝ⁴ (not viewable) |
| Level set dimension | 1D curves in ℝ² (easy to draw on paper) | 2D surfaces in ℝ³ (requires 3D rendering) |
| Visualization ease | High — flat contour maps are intuitive | Moderate — 3D surfaces can be complex; cross-sections help |
| Real-world example | Topographic maps, isotherms on weather maps | Equipotential surfaces in electrostatics, medical CT isosurfaces |
| Gradient relationship | ∇f ⊥ level curve at each point | ∇f ⊥ level surface at each point |
Connection to Gradients, Optimization, and Implicit Surfaces
The concepts of domain and level set are not endpoints—they serve as the foundation for nearly every subsequent topic in multivariable calculus. Understanding how level sets connect to the gradient, optimization, and implicit functions provides motivation for the rest of the course.
| This Lesson's Concept | Advanced Extension |
|---|---|
| Level curves f(x, y) = c | The gradient ∇f is perpendicular to each level curve; the magnitude |∇f| measures how fast f changes between adjacent curves. This is the foundation of directional derivatives. |
| Level surfaces f(x, y, z) = c | ∇f at a point on a level surface gives the normal vector to the tangent plane, enabling the equation of tangent planes and normal lines to implicitly defined surfaces. |
| Domain D ⊆ ℝ² | Whether D is open, closed, bounded, or connected determines the applicability of the Extreme Value Theorem (continuous f on a closed bounded D attains its max and min). |
| Contour spacing encodes steepness | Lagrange multipliers exploit the tangency between level curves of f and the constraint g(x, y) = k to locate constrained extrema. |
| Level set as implicit relation | The Implicit Function Theorem guarantees that, under mild conditions (non-vanishing partial derivative), a level set locally defines one variable as a function of the others. |
In subsequent lessons, you will see that the gradient vector ∇f at a point (a, b) is perpendicular to the level curve f(x, y) = f(a, b) passing through that point. This remarkable geometric fact transforms level curves from a passive visualization tool into an active computational framework: the gradient tells you the direction of steepest ascent, and level curves tell you the directions along which f does not change at all. The interplay between these two perspectives is at the heart of optimization theory, fluid dynamics (streamlines vs. equipotentials), and differential geometry.
Practice Problems
Summary — Multivariable Functions & Level Sets
A function of two variables maps each point (x, y) in a domain D ⊆ ℝ² to a scalar z = f(x, y), producing a surface in ℝ³, while a function of three variables maps (x, y, z) ∈ D ⊆ ℝ³ to a scalar w = f(x, y, z), producing an unvisualizable graph in ℝ⁴. The natural domain is found by identifying all algebraic restrictions—division by zero, even roots of negatives, logarithms of non-positives—and intersecting the resulting conditions. Domains should be described both algebraically (inequalities) and geometrically (sketched regions).
Level curves (for two-variable functions) and level surfaces (for three-variable functions) are the sets of input points where the function takes a constant value c. They reduce the visualization problem by one dimension and encode critical information about the function's rate of change: closely spaced level sets indicate rapid change, while widely spaced ones indicate gradual change. Level sets for distinct output values never intersect (by the definition of a function), and the gradient vector is always perpendicular to the level set—a fact that bridges this introductory material to partial derivatives, directional derivatives, and Lagrange multipliers in subsequent lessons.