Second-order dynamics

In brief

Pull a mass hanging on a spring down a little and let go. It does not simply drift back to its resting point and stop, the way a cooling cup of coffee settles at room temperature. It overshoots, swings past the resting point, comes back, overshoots again, and oscillates. Something about this system needs more than "how far is it from where it wants to be" to describe fully; it also needs "how fast is it currently moving, and in which direction." A rule that only knew position could never produce an overshoot, because at the resting point itself such a rule would say the rate of change is zero and the motion should stop dead. What actually happens is that the mass sails through the resting point because it is carrying velocity, momentum in the ordinary sense, and velocity is not something a first-order rule for position alone can track. Systems like this, where the evolution rule constrains an object's acceleration rather than its plain rate of change, are second-order dynamics, and this second layer of change is exactly what makes oscillation, inertia, and overshoot possible.

The full treatment

First look: why one number stops being enough

Consider the mass on the spring again, with position x measured from the resting point. If a system were first order, its rate of change dx/dt would depend only on x itself, and at x equal to zero, the rate would be forced to zero as well; the mass would simply stop wherever it reached the resting point, exactly the equilibrium behavior of first-order dynamics. But a real mass pulled down and released does not stop at the resting point. It has, at that instant, a nonzero velocity built up during the fall back toward center, and that velocity carries it through. This shows that position alone does not capture the full state of the system; velocity is an independent piece of information that must be tracked alongside it.

Deriving the spring equation from two established laws

Two well-established laws combine to produce the governing equation. Hooke's law states that a spring pulls or pushes back with a force proportional to how far it is stretched or compressed, and in the opposite direction: F = -kx, where k is a positive constant set by the spring's stiffness. Newton's second law states that force equals mass times acceleration, and acceleration is the rate of change of the rate of change of position, written d^2x/dt^2. Setting the two expressions for force equal gives m times d^2x/dt^2 equals minus k times x, or mx'' = -k*x, using the shorthand x'' for the second derivative. This equation constrains the second derivative of x, which is why it is called second order: the highest derivative appearing is the second one.

Splitting one second-order rule into two first-order rules

A second-order equation can always be rewritten as a coupled pair of first-order equations by introducing velocity v as a second tracked quantity, alongside x. The two rules become: dx/dt = v (the rate of change of position is, by definition, the velocity) and dv/dt = minus (k/m)*x (Newton's second law, rewritten to give the rate of change of velocity, that is, the acceleration). Neither rule alone is complete. The first rule needs to know v to say anything about x; the second needs to know x to say anything about v. Together, the pair (x, v) is the full state of the system, and the two coupled first-order rules acting on that pair are exactly equivalent to the original second-order equation. This is the general recipe: any second-order rule can be unfolded into two first-order rules on an enlarged state that includes both the quantity and its own rate of change.

Why the outcome is oscillation, not decay

Try a candidate solution of the form x(t) = A times the cosine of (omegat plus a phase angle), where A and the phase are set by how the mass was started, and omega is a constant to be determined. Differentiating twice, the second derivative of this cosine expression is minus omega squared times the original expression, that is, x'' = -omega^2 * x. Substituting into mx'' = -kx gives -momega^2x = -kx, which holds for all t exactly when omega^2 = k/m, that is, omega equals the square root of k over m. So a cosine wave with that specific frequency satisfies the equation exactly, for any amplitude and phase. Unlike first-order dynamics, where the only possible long-term fates are unbounded growth, decay, or a resting equilibrium, this equation produces perpetual oscillation, because the coupling between x and v continually feeds each one's rate of change from the other, rotating the pair (x, v) around rather than letting either settle or run away.

What real damping adds

Real springs and pendulums lose energy to friction and air resistance, which adds a term proportional to velocity: mx'' + cx' + k*x = 0, where c is a positive damping constant. This equation still constrains a second derivative and still needs the pair (x, v) as its full state, but depending on how large c is relative to m and k, the oscillation can shrink in amplitude over time, or the system can settle back to rest without oscillating at all. Whether damping merely shrinks the swings or removes them altogether depends on a threshold set by these three constants, and is the natural next question once the pure oscillator above is understood.

Lineage

Robert Hooke stated the proportional restoring force of a spring in 1678, publishing it first as an anagram to protect priority before revealing "ut tensio, sic vis," as the extension, so the force. Isaac Newton's 1687 Principia supplied the second law relating force to the second derivative of position, and together the two laws gave the first rigorously derived second-order equation of motion. Through the eighteenth century, Daniel Bernoulli, Leonhard Euler, and Jean le Rond d'Alembert extended the method to vibrating strings and coupled oscillators, and Joseph-Louis Lagrange later reformulated all of mechanics in terms of generalized coordinates evolving under second-order equations, generalizing the spring and pendulum far beyond their original physical settings.

The strongest case for it

The same coupled structure, a position-like quantity and a velocity-like quantity feeding each other's rates of change, describes an enormous range of physical and engineered systems: springs, pendulums swinging through small angles, the voltage and current in a circuit built from an inductor and a capacitor, and the ride height and rate of change of ride height in a vehicle's suspension. Recognizing a system as second order immediately tells an engineer to expect the possibility of overshoot and oscillation, and to look for the balance of inertia, restoring force, and damping that decides whether that oscillation grows, shrinks, or persists.

The strongest case against it

The clean spring equation and the small-angle pendulum equation are both idealizations. Hooke's law is only proportional for modest stretches; push a real spring far enough and the restoring force stops being linear in x. The pendulum's true equation of motion involves the sine of the angle, not the angle itself, and the tidy oscillator equation used here is only the small-angle approximation, valid when the sine of an angle is well matched by the angle itself measured in radians. A common misconception is assuming that any oscillating system will oscillate forever with unchanged amplitude; the pure equation derived above ignores all energy loss, and every real oscillator dissipates energy and either damps down or needs a continuous outside push to keep swinging.

Where it stands now

This is settled classical mechanics, unchanged since Newton, Hooke, and their eighteenth-century successors, and it remains the standard starting point for analyzing any inertial system before adding damping, driving forces, or nonlinearity. It is also the direct mathematical ancestor of feedback control theory, where engineers deliberately design the equivalent of extra damping or restoring terms to make a controlled system, such as a vehicle or a robotic arm, settle rather than oscillate.

Test yourself

A car's suspension can be modeled as a mass (the car body) connected to the axle by a spring and a damper. Using Newton's second law, Hooke's law for the spring force, and a damping force proportional to the vertical velocity of the body, write the second-order differential equation governing the body's height above its resting position after the car hits a bump. Rewrite your equation as a coupled pair of first-order equations using height and vertical velocity as the two state variables. Finally, state, without solving the equation exactly, what physical change to the suspension, stiffer spring, heavier damper, or lighter body, you would expect to push the system from oscillating over the bump toward settling back to rest without any overshoot at all.