Does a Solution Even Exist?

The question we never stopped to ask

Up to now you have been a solver. You separated variables, ran the integrating factor, hunted the hidden potential of an exact equation — every time, you started from the quiet assumption that a solution was sitting there waiting to be found. That assumption deserves to be put on trial. Given an initial value problem y' = f(x, y) with y(x0) = y0, is there any function at all that threads the given point and obeys the equation? This is the existence question, and it is logically prior to every recipe you have learned: a method that computes a formula is worthless if no solution exists for it to compute.

Here is the catch that makes the question urgent: most differential equations have no closed-form solution at all. The handful of equations you can crack with separation or the linear recipe are the lucky exceptions, not the rule. For the vast majority — the ones numerical and qualitative methods later in this ladder are built to handle — you will never write down a formula. So you cannot answer "does a solution exist?" by simply producing one. You need a theorem that certifies existence without ever exhibiting the solution itself.

A slope field you can always walk along

The cleanest way to feel why a solution should exist is to picture the slope field. The equation y' = f(x, y) does not hand you a curve, but it does plant a tiny arrow at every point of the plane: at (x, y) the prescribed slope is exactly f(x, y). A solution is nothing more than a curve that stays tangent to those arrows everywhere it goes. So the existence question becomes wonderfully physical: starting from (x0, y0), can you always keep walking in the direction the local arrow points?

Intuition says yes, provided the arrows do not jerk around wildly. If the direction field varies smoothly — or even just continuously — then as you take a tiny step, the next arrow points almost where the last one did, and you can follow them in an unbroken curve. If instead f had a violent jump, the arrows could contradict each other and trap you with nowhere consistent to step. This picture is exactly the content of Peano's existence theorem, which converts that intuition into a guarantee.

Peano's theorem: continuity is enough

Here is the statement, stripped to its essentials. If f(x, y) is continuous on some rectangle of the plane containing the starting point (x0, y0), then the initial value problem y' = f(x, y), y(x0) = y0 has at least one solution defined on some interval around x0. That is the whole hypothesis: continuity of f. No smoothness, no special form, no separability — just an unbroken direction field. The Peano existence theorem is generous precisely because it asks for so little.

Two pieces of fine print matter, and honesty demands we say them out loud. First, the theorem only promises a solution on some interval around x0, possibly a short one — it makes no claim that the solution survives for all x. How far it actually reaches is the subject of the interval of existence, the final guide in this rung. Second, and more startling, Peano promises "at least one" solution. It is silent on whether that solution is unique. Continuity buys you existence and nothing more.

Where existence is not enough

The gap Peano leaves open is not a theoretical nicety — there are honest, simple equations where it bites. The classic example is y' = y^(2/3) with y(0) = 0. The right-hand side f(x, y) = y^(2/3) is perfectly continuous everywhere, so Peano cheerfully guarantees a solution exists. And it does: the constant function y(x) = 0 sits at the start and never moves. But it is not alone. The function y(x) = (x/3)^3 also passes through (0, 0) and satisfies the equation, and so does a whole family that lingers at zero for a while and then peels away. One initial condition, infinitely many solutions.

This is the failure of uniqueness, and it is not a pathology cooked up to be cruel. In the slope field picture, the trouble is that the arrows near the x-axis steepen so abruptly — f(x, y) = y^(2/3) is continuous but its slope in y blows up as y approaches 0 — that the curve at the origin genuinely faces a fork. The field is continuous enough for a path to exist, but not regular enough to force a single path. The failure of uniqueness here is the slope field's honest answer, not a flaw in our reasoning.

The missing ingredient, and what comes next

So the diagnosis is sharp. Continuity of f delivers existence but can leave uniqueness wide open. To close the fork, we need to control not just how big f is but how fast it can change as y varies — to forbid the arrows from steepening too violently. That extra control has a name you will meet in the very next guide: the Lipschitz condition, a quantitative cap on how much f(x, y) can change when y changes. It is exactly what y^(2/3) fails to satisfy at the origin.

Add the Lipschitz condition to continuity and you upgrade from "a solution exists" to "exactly one solution exists" — the celebrated Picard-Lindelof theorem. Better still, its proof is constructive: it does not merely assert the unique solution, it builds it as the limit of a sequence of ever-improving guesses. That construction, Picard iteration, is the engine of guide 3 in this rung, and it is the same fixed-point idea that quietly powers many numerical methods.

  f continuous            =>  a solution EXISTS        (Peano)
  f continuous + Lipschitz =>  exactly ONE solution     (Picard-Lindelof)

  example  y' = y^(2/3),  y(0)=0 :
     continuous  -> a solution exists  (e.g. y = 0)
     NOT Lipschitz at y=0 -> also y = (x/3)^3 ... uniqueness fails

Keep the honest framing in mind as you climb. Existence theorems certify that a solution is there to be approximated; they almost never hand you a formula. That is not a weakness of the theory — it is the reason the rest of this subject, the numerical schemes and the qualitative phase-plane reasoning, exists at all. Knowing a unique solution lives somewhere in the slope field is precisely the license you need to go chase it with a computer and trust the answer.