Frames, Configurations, and Newton's Laws

Acceleration is “absolute” in Newtonian mechanics only relative to a specific kinematic structure: a common time, Euclidean space at each time, and an affine family of inertial frames. Position and velocity change under a Galilean boost; acceleration changes only by a fixed spatial rotation. Newton’s laws therefore assume more than clocks, rulers, and a force formula—they assume which frames count as unaccelerated.

Required background. Calculus and Linear Algebra provides multivariable differentiation, linear maps, inner products, and basis changes. Helpful background. Ordinary Differential Equations provides initial-value problems, linearization, and asymptotic solution methods.

Configurations and Newtonian motions

For $N$ unconstrained point particles, a configuration is

$$q=(\mathbf x_1,\ldots,\mathbf x_N)\in Q=(\mathbb R^3)^N.$$

A time-independent holonomic constraint restricts $Q$ to a submanifold—for example, a bead constrained to a fixed sphere has $Q=S^2$. A motion is a twice-differentiable curve $q(t)$. Its velocity is tangent to $Q$; its acceleration can also have a normal component supplied by the constraint force. For example, uniform motion around a circle has constant speed but inward acceleration. Given a sufficiently regular force law and a consistent constraint model, $q(t_0)$ and $\dot q(t_0)$ determine a local solution. Time-dependent constraints, such as a moving wire, require their explicit time dependence as additional prescribed data.

This initial-data statement separates the form of the law from determinism. Writing a second-order equation does not guarantee a unique continuation: the force must satisfy suitable regularity, such as local Lipschitz control in phase-space variables. Singular collisions and deliberately non-Lipschitz forces can defeat existence or uniqueness. Nor should a curvilinear coordinate choice be confused with a new frame. Polar coordinates within one inertial frame generate coordinate terms in the component equations, but the underlying observer remains inertial; a time-dependent rotating coordinate grid changes the frame and produces genuine inertial-force terms in that description.

For one particle of constant inertial mass $m$,

$$\mathbf p=m\mathbf v, \qquad \frac{d\mathbf p}{dt}=\mathbf F.$$

The first law identifies inertial frames: an isolated particle has constant velocity. The second supplies dynamics after the force is specified. In its elementary pair-force form, the third law $\mathbf F_{ab}=-\mathbf F_{ba}$ implies total momentum conservation; angular-momentum conservation additionally requires internal forces to be central. Field-mediated interactions can carry momentum themselves, so pairwise action–reaction is not the universal conservation statement.

Galilean transformations and absolute acceleration

Inertial Cartesian coordinates are related by

$$t'=t+b, \qquad \mathbf x'=R\mathbf x+\mathbf u t+\mathbf a,$$

where $R\in SO(3)$ and $\mathbf u$, $\mathbf a$, and $b$ are constant. Differentiation with respect to the shared time gives

$$\mathbf v'=R\mathbf v+\mathbf u, \qquad \mathbf a'=R\mathbf a.$$

Thus inertial observers disagree about velocity but agree whether acceleration vanishes and agree on its magnitude. This is the precise content of “absolute acceleration” in Newtonian theory; see Tong 2013, §§1.2–1.2.1, pp. 4–7, PDF.

The result fails for a time-dependent translation. If $\mathbf x'=\mathbf x-\mathbf X(t)$, then

$$m\ddot{\mathbf x}'=\mathbf F-m\ddot{\mathbf X}.$$

The last term is an inertial force. A rotating frame likewise introduces Coriolis, centrifugal, and Euler terms. These forces restore the form of the equation in the chosen coordinates; they do not describe an additional interaction with another body.

Discriminating checks

• Setting $\mathbf u=0$ leaves a Euclidean rotation and translation; acceleration transforms as a vector.
• With $\mathbf F=0$, every inertial frame assigns a straight trajectory $\mathbf x=\mathbf x_0+\mathbf v_0t$.
• Both $m\mathbf a$ and $\mathbf F$ have dimension $MLT^{-2}$.
• For $m\mathbf a=\mathbf F$ to keep its form, a mechanical force must transform as $\mathbf F'=R\mathbf F$. Velocity-dependent forces require an explicit covariance check.

Newtonian point-particle mechanics here assumes constant masses, smooth trajectories, and forces evaluated at a common absolute time. Variable-mass systems require momentum-flux terms; impacts require impulse laws; nonholonomic constraints cannot generally be imposed merely by restricting $Q$. Relativistic boosts will also abandon absolute simultaneity.

Exercise: an accelerating laboratory

A laboratory follows $X(t)=At^2/2$ along the $x$-axis relative to an inertial frame. A free bead has $x(t)=x_0+v_0t$. Find the laboratory trajectory and the force appearing in its equation of motion.

Solution

The laboratory coordinate is

$$x'(t)=x(t)-X(t)=x_0+v_0t-\frac12At^2,$$

so $\dot x'=v_0-At$ and $\ddot x'=-A$. The laboratory therefore writes

$$m\ddot x'=F_{\rm inertial}, \qquad F_{\rm inertial}=-mA.$$

The sign is sensible: a laboratory accelerating toward $+x$ sees the free bead fall behind toward negative $x'$. The force vanishes continuously as $A\to0$, recovering a Galilean frame.

Variational Mechanics and Symmetries next replaces the componentwise force law by a scalar action and identifies its domain of equivalence.

References

• Tong, David. Dynamics and Relativity. University of Cambridge lecture notes, 2013. §§1.1–1.3, pp. 2–9, PDF.