Rediscovering Mechanics and Energy from Questions

This note develops Newtonian mechanics from concrete problems that might motivate it. The point is not to begin with familiar formulas and attach names to them afterward. The point is to ask what we want to predict, discover what information is missing, and introduce a definition only when it solves that problem.

The central example is a cart attached to a spring. From it, we will motivate position, velocity, acceleration, inertial mass, force, momentum, impulse, kinetic energy, and potential energy.

1. The prediction problem

Consider a cart on a nearly frictionless track, attached to a spring. Pull the cart 10 centimeters to the right and release it.

Our first question is concrete:

When will the cart first reach the spring’s relaxed position?

We want a rule that predicts the answer for different carts, springs, and release positions without repeating every possible experiment.

2. Position: what outcome are we predicting?

Choose the spring’s relaxed position as the origin:

Record the cart’s position at a sequence of times. The desired arrival time is the first time for which \(x(t)=0\).

Position is necessary, but it is not enough to predict what happens next. The cart can pass through the same position while traveling either left or right.

3. Velocity: which way is the state developing?

To distinguish those cases, measure the rate at which position changes:

Over a short interval \(\Delta t\), this gives the approximate update:

But velocity also changes while the spring acts. Knowing \(x\) and \(v\) now does not yet tell us the next velocity.

4. Acceleration: how is velocity changing?

Measure the rate of change of velocity:

Then, over a short interval,

The prediction problem has now become more precise:

Given the present spring extension and the particular cart, what acceleration should we expect?

Acceleration is still only a measured description. We have not yet explained how the spring state and the cart jointly determine it.

5. Solve the problem

Measure the cart’s acceleration at several extensions:

The measurements suggest the empirical rule

Equivalently, write the measured coefficient as \(\omega^2\):

Here \(\omega\) is not a new explanation or a universal constant. It is merely a compact way of reporting what this cart-spring pair did in the experiment.

If we release the cart from rest at \(x_0\), the measured rule predicts

Therefore the cart first reaches the relaxed position when

so for this experiment

The original prediction problem is now solved. We did not need mass or force. Those ideas become relevant only when we ask a new question:

Can one rule predict the motion after we replace the cart or the spring?

6. Inertial mass: comparing different carts

Keep the spring fixed, but repeat the acceleration measurements using another cart:

Each cart has its own predictive rule: \(a_A=-4x\) and \(a_B=-2x\). More importantly, their acceleration ratio remains the same at every tested extension:

This raises a natural question:

Can the stable difference be assigned to the carts, leaving a separate description of the spring?

Define relative inertial mass by the inverse acceleration ratio when the carts are tested under the same interaction:

If A accelerates twice as much as B, B is assigned twice the inertial mass. Assigning one reference cart a unit mass fixes the overall scale.

This definition is useful only if the same mass ratios work with different springs, collisions, and other controlled interactions. That cross-experiment consistency is an empirical discovery, not a consequence of the notation.

7. Force: the part that belongs to the interaction

Once masses have been calibrated, inspect the measurements again. At any fixed extension, they reveal

The common value depends on the spring’s extension but not on which test cart we attached. This is exactly the object we were looking for: a description of the interaction that can be reused when the cart changes. We give it the name force:

This definition was not needed to solve the one-cart problem. It earns its role only because experiments reveal that \(ma\) separates the spring from the cart. If no cart-independent common value existed, this definition would not provide the desired generalization.

For the measured spring, that common value varies linearly with extension, so we report its behavior as

The constant \(k\) characterizes the spring in our chosen mass units. The minus sign records that the acceleration points toward the relaxed position. Combining the cart-independent spring rule with the cart’s inertial mass gives

Now the solution works for any calibrated cart attached to that spring:

and hence

Force is therefore the reusable interaction-side factor that emerged from trying to generalize an already solved experiment.

8. Momentum and impulse: the accumulated effect

Consider one cart whose initial velocity \(v_1\) is known. Suppose its acceleration during the interval from \(t_1\) to \(t_2\) has been measured or modeled. Ask:

Can we predict its final velocity \(v_2\) without already knowing it?

Because acceleration is the rate of change of velocity,

Integrating the known acceleration gives the answer:

or

That completely solves the one-cart problem. Momentum and impulse are not needed yet.

Now replace the cart-specific acceleration description with the reusable interaction description discovered in the previous section:

For constant mass, the prediction becomes

The same time integral appears whenever we ask for the accumulated effect of a force history, so give it a name—impulse:

This integral calculates the total effect of the changing force over the interval—the area under its force-versus-time graph. Once \(J\) is known, dividing it by the cart’s mass gives the total velocity change.

The final-velocity prediction is now

Multiplying by \(m\) gives

This repeated endpoint quantity motivates defining momentum:

The prediction then takes the compact update form

Momentum is a property of the cart’s state, not an object-independent quantity. Impulse summarizes the force accumulated during a particular interaction, and momentum records how that interaction changed the cart.

9. The question that motivates energy

Return first to the single cart-spring pair from Section 5. We know its measured acceleration rule \(a(x)=-\omega^2x\) and release it from rest at \(x_0>0\). Ask:

What will the cart’s maximum speed be, and can we find it without calculating its position at every intermediate instant?

While the cart moves from \(x_0\) toward the center, its acceleration and velocity point in the same direction, so its speed increases until it reaches \(x=0\). We therefore want its speed at the center.

The acceleration is given as a function of position rather than time. The chain rule connects it to the change of velocity with position:

Integrating from an initial state \((x_1,v_1)\) to a final state \((x_2,v_2)\) gives

For our cart, \(v_1=0\), \(x_1=x_0\), and \(x_2=0\). Therefore,

so

That solves the one-cart problem without force, work, or energy.

Now replace the cart-specific acceleration rule with the reusable interaction description from Section 7:

Multiplying the general one-cart result by the constant mass gives

The integral calculates the accumulated effect of the force as the cart moves through space—the signed area under the force-versus-position graph. Give this quantity the name work:

The velocity-side integral can be split explicitly at the reference velocity \(0\). Using \(u\) as the integration variable,

The two terms have the same form; only the endpoint velocity changes. This motivates assigning that quantity to each state and naming it kinetic energy:

The prediction then takes the compact update form

Kinetic energy belongs to the cart’s state. Work summarizes how a particular force changes that state while the cart moves through a displacement. The factor \(\tfrac12\) is not chosen arbitrarily; it arises from integrating \(v\,dv\).

10. Potential energy: recording reversible storage

The work formula answers a question about one particular trip from \(x_1\) to \(x_2\). But calculating a new integral for every pair of endpoints repeats the same work. This suggests a new question:

Can we assign one number to each spring configuration so that the work between any two configurations is obtained by subtracting their numbers?

For the spring, the force law is

Split its work integral at the relaxed position \(x=0\):

Each term now depends on only one endpoint. Define the potential energy of a configuration, relative to the relaxed spring, by

The minus sign makes work done by the spring equal a loss of potential energy. Indeed, the split integral becomes

For \(F(u)=-ku\), the assigned state quantity is

The zero at \(x=0\) is a convenient reference; adding the same constant to every value of \(U\) would not change any prediction.

Section 9 showed that the spring’s work changes the cart’s kinetic energy:

The endpoint construction above shows that the same work is the negative change of potential energy:

Therefore,

This motivates defining the total energy of the isolated cart-spring system:

Unlike the separate values of \(K\) and \(U\), this sum remains constant during the motion.

For a cart released from rest at \(x_0\), the center speed follows immediately:

Therefore,

Energy provides a compressed route from the initial state to constraints on the final state. It can answer some questions without reconstructing every intermediate moment.

The construction works when the interaction’s work depends only on its endpoints. If different paths between the same endpoints produce different work, no single-valued potential energy can summarize the interaction in this way. When it does work, the general relation is

Different interactions give different potential functions, such as

The particle-mechanics construction is now complete. It leaves a new question:

Are kinetic and potential energy the only possible kinds of energy, or are they merely the two parts that appear in this particular description?

Energy Beyond Particle Mechanics starts from that question.