Abraham–Lorentz Equation and Its Pathologies

The previous article closed the chapter's energy books and left a loaded weapon on the table: Newton's law with the self-force installed, $m\mathbf{a}=\mathbf{F}_{\text{ext}}+m\tau\dot{\mathbf{a}}$ . This article pulls the trigger — solves the equation honestly, as a dynamical system — and watches classical electrodynamics malfunction in the most instructive way any theory has ever malfunctioned. The jerk term makes the equation third-order in time: position and velocity no longer fix the future, and the extra solution family it unlocks contains a free particle whose acceleration grows as $e^{t/\tau}$ , manufacturing energy from nothing. Exorcise those runaways with a boundary condition at future infinity — the only cure the equation permits — and the survivor is stranger still: the charge begins accelerating before the force is applied, over the window $\tau\approx6\times10^{-24}$ s. Causality or runaway-freedom; the equation will not sell you both. Griffiths files these results, with a connoisseur's relish, under "the ointment in the fly" — and the article's job is to make that verdict precise rather than merely scandalous. The pathologies are solved exactly (they take one integrating factor), their scales are priced (the preacceleration displacement for a laboratory step field is $10^{-29}$ m — fourteen orders below a proton radius), and their meaning is extracted: the equation is announcing the boundary of its own jurisdiction, the timescale $\tau\sim r_e/c$ below which a classical point charge is not a consistent object. The modern resolution is then executed, not just cited: the Landau–Lifshitz reduction of order, which demotes the equation back to second order, keeps every verified prediction — the linewidths, the damping — and deletes the pathologies at the stroke of a substitution. This closes the E9 radiation chapter. The next article opens a new one — and its subject, hiding in plain sight through everything just done, is the symmetry of spacetime itself.

Key takeaways

The Abraham–Lorentz equation $m\mathbf{a}=\mathbf{F}_{\text{ext}}+m\tau\dot{\mathbf{a}}$ is third-order: it demands a third initial datum beyond position and velocity, and the extra solution mode is the trouble.
Force-free solution: $\mathbf{a}(t)=\mathbf{a}_0e^{t/\tau}$ — a runaway that self-accelerates forever, violating energy conservation unless $\mathbf{a}_0=0$ is imposed by hand.
Killing runaways requires a future boundary condition, which converts the equation into $\mathbf{a}(t)=\frac{1}{m\tau}\int_t^\infty e^{-(t'-t)/\tau}\mathbf{F}_{\text{ext}}(t')\,dt'$ : the acceleration now reads the force's future.
For a step force this yields preacceleration: motion begins a time $\sim\tau$ before the force arrives — acausal in principle, $10^{-29}$ m in practice.
Verdict: causality and runaway-freedom cannot both be satisfied; the classical point charge fails self-consistency on scales $\lesssim\tau=2r_e/3c$ , exactly where quantum electrodynamics takes over.
The Landau–Lifshitz reduction — substitute $m\dot{\mathbf{a}}\approx\dot{\mathbf{F}}_{\text{ext}}$ inside the small term — restores second-order dynamics, keeps all first-order-in- $\tau$ physics, and has no runaways and no preacceleration; it is the working equation of strong-field physics.
Key idea: taken literally, the third-order equation offers a choice between runaways and preacceleration; taken as an effective theory valid above $\tau$ (Landau–Lifshitz), it is correct, causal, and confirmed.

The core insight

Radiation reaction enters as a force proportional to the jerk $\dot{\mathbf{a}}$ , making the equation third-order. This single fact generates both its successes (correct average energy balance) and its pathologies (runaways and preacceleration).

Where this lives today

The Abraham–Lorentz pathology complex has become physics' favorite teaching case for effective theory — the discipline of knowing a model's scale of validity and refusing to over-solve it. The runaway/preacceleration dilemma appears, in near-identical mathematical form, wherever higher-derivative corrections tempt literal treatment: gravitational radiation reaction in binary-inspiral modeling (the post-Newtonian equations harbor the same spurious modes, tamed by the same reduction of order), higher-derivative gravity theories, and the ghost fields of quantum field theory — all managed with the LL playbook this article executed. Dirac's 1938 covariant analysis (the Abraham–Lorentz–Dirac equation, this equation's relativistic parent) and its Landau–Lifshitz reduction are, respectively, the advanced chapter's subject and the working core of the strong-field codes. Experimentally, the once-academic equation now has a beamline. The 2018 Gemini campaigns — GeV laser-wakefield electrons colliding with a counter-propagating petawatt pulse — measured electron energy loss in the radiation-reaction-dominated regime and tested classical-LL against quantum-corrected models; the data already disfavor the uncorrected classical treatment in the strongest collisions, exactly as the $\hbar$ -aware reading of the pathologies predicts. Next-generation facilities (multi-10-PW: ELI-NP, Apollon, ZEUS, and the planned 100-PW class) list "radiation reaction and the classical-quantum boundary" as headline science. The fly, after a century as a thought experiment, is finally in the apparatus.

Origins · 1 minPaul Dirac and how this was discovered

Paul Dirac

1902–1984 · Cambridge — who took the self-force relativistic and stared down its pathologies without blinking

Dirac's 1938 paper "Classical theory of radiating electrons" is the equation's coming-of-age. Dissatisfied with Lorentz's shell-model derivation ("it is not a permissible mathematical procedure to assume the electron has a structure"), he rederived the self-force covariantly for a true point charge, using only conservation of the stress-energy tensor on a tube around the worldline plus the elegant splitting of the self-field into half-retarded-plus-advanced (divergent, mass-renormalizing) and half-retarded-minus-advanced (finite, the radiation field) — producing the Abraham–Lorentz–Dirac equation and the modern renormalization-first viewpoint in one stroke. Characteristically, Dirac neither hid the runaways and preacceleration nor apologized for them: he catalogued the solutions, noted the acausal window $\sim\tau$ , and remarked that a theory's misbehavior at scales beyond its evidence was a finger pointing at future physics. It pointed well: the divergence-subtraction he performed classically became, in the hands of his successors (and to his own famous distaste), the renormalization program of QED — where the electron's self-interaction finally gets a finite, predictive, Nobel-decorated treatment. The fly in the classical ointment turned out to be the larva of quantum field theory.

History

1902

— Abraham derives the rigid electron's self-force: the jerk term enters physics

1904

— Lorentz's shell expansion isolates the universal reaction force; divergent mass absorbed

1912

— Schott's acceleration energy explains the averaged-vs-instantaneous ledger

1938

— Dirac's covariant point-charge derivation: the ALD equation, runaways and preacceleration catalogued

1948

— QED's renormalization completes, quantum-mechanically, the subtraction Dirac performed classically

1951

— Landau and Lifshitz publish the reduction-of-order equation, the pathology-free working form

1960

s — Rohrlich and others codify the effective-theory reading: validity above τ, by construction

2018

— Gemini laser–beam collisions probe the radiation-reaction-dominated regime experimentally

The physical picture

Count degrees of freedom — the whole drama is visible from there. Newton's second law is a machine that converts two initial data (where, how fast) into a unique future. Add a term proportional to $\dot{\mathbf{a}}$ and the machine demands three data: where, how fast, and how hard currently accelerating. Nature does not supply the third; nothing in the preparation of an electron fixes its initial acceleration independently of the force. The equation has one solution-dimension more than the physics has facts, and that orphan dimension must contain something. What it contains is the homogeneous solution of $m\mathbf{a}=m\tau\dot{\mathbf{a}}$ : pure exponential growth, $e^{t/\tau}$ . Where does the runaway's energy come from? The equation launders it through the mass renormalization of the previous article. The divergent negative bare mass hiding behind the observed $m$ is an unbounded energy reservoir; a literal point charge can fall forever down its own self-energy well, paying for both the growing kinetic energy and the growing Larmor bill. The runaway is not a bug in the algebra — it is the honest dynamics of an object whose self-energy was infinite before renormalization papered it over. The shell model with $\varepsilon>c\tau$ (equivalently, an electron larger than its classical radius) has no runaways at all; shrink past that size and they appear. The pathology is the point limit's revenge. And the cure defines the disease's other face. To amputate the runaway, one imposes boundedness at future infinity — a final condition, not an initial one. Differential equations solved backward from the future produce presents that depend on what is about to happen: preacceleration. The mathematics is impeccable; the ontology is on fire; and the fire is confined to a region — times of order $\tau$ , displacements of order $10^{-29}$ m — where the classical point electron had no business claiming jurisdiction anyway.

Figure

The extra mode and the impossible choice — three data in, causality or boundedness out, never both

The math: the equation as a dynamical system

Restrict to one dimension (the vector case factorizes) and write the equation in the form solvers prefer:

\dot{a}-\frac{a}{\tau}=-\frac{F_{\text{ext}}(t)}{m\tau}. \qquad (1)

where: this is a first-order linear ODE for the acceleration — the fastest way to see both the equation's tractability and its illness. The homogeneous solution is $e^{+t/\tau}$ : a growing exponential, because the sign of the self-force term makes the acceleration anti-damped. Every pathology below is a corollary of that plus sign, and every solution below is one integrating factor away.

Derivation AGeneral solution, and the runaway mode

The goal is the complete solution space of (1) — exhibiting exactly where the extra freedom lives and what it costs. Strategy: integrating factor $e^{-t/\tau}$ , then read the two-parameter-family structure (one constant from this first-order equation for $a$ , plus the ordinary two from integrating $a$ up to position) against the two data nature supplies. The algebra is three lines; the interpretation is the article.

Step 1. Multiply (1) by $e^{-t/\tau}$ and recognize a total derivative:

\frac{d}{dt}\Big(e^{-t/\tau}a\Big)=-\frac{e^{-t/\tau}}{m\tau}F_{\text{ext}}(t).

Step 2. Integrate from $t$ to a reference time $T$ :

a(t)=e^{t/\tau}\left[e^{-T/\tau}a(T)+\frac{1}{m\tau}\int_t^{T}e^{-t'/\tau}F_{\text{ext}}(t')\,dt'\right].

Why the structure already convicts the equation: whatever reference is chosen, $a(t)$ carries a free constant multiplying $e^{t/\tau}$ — the orphan mode. Prescribing initial data $a(0)$ generically excites it.

Step 3. The force-free case. With $F_{\text{ext}}=0$ : $a(t)=a_0e^{t/\tau}$ . Unless $a_0=0$ exactly, a free electron's acceleration doubles every $4.3\times10^{-24}$ s — in one femtosecond it grows by $e^{160{,}000{,}000}$ . The mode is not merely unphysical; it is the most violently unphysical object in classical physics, and it must be executed by fiat.

Step 4. Execute it: demand $e^{-T/\tau}a(T)\to0$ as $T\to\infty$ (acceleration grows slower than $e^{t/\tau}$ in the far future — boundedness suffices). Step 2 then collapses to the unique non-runaway solution:

a(t)=\frac{1}{m\tau}\int_t^{\infty}e^{-(t'-t)/\tau}\,F_{\text{ext}}(t')\,dt'.

Why this is the fateful line: the surviving solution computes today's acceleration as an exponentially weighted average of the force over the FUTURE window $[t,t+\text{few}\,\tau]$ . The equation is now integro-differential with a future kernel — runaway-free by construction, and acausal by construction. The choice made in this step is the only lever the equation offers, and neither of its positions is innocent. *Verified: substitution of Step 4 back into (1) checks by differentiation under the integral (the boundary term at $t'=t$ supplies $-F/m\tau$ ); matches Griffiths §11.2.3 and Jackson §16.4 conventions (note Jackson writes the kernel with $s=(t'-t)/\tau$ ).*

Go deeper · Derivation B · 1 minThe step force — preacceleration solved exactly

Derivation BThe step force — preacceleration solved exactly

The goal is the canonical exhibit: a constant force switched on at $t=0$ , $F_{\text{ext}}(t)=F\,\Theta(t)$ , run through the non-runaway solution — producing the acceleration profile whose $t<0$ tail is the famous preacceleration. Strategy: the future integral of Derivation A Step 4 evaluates in closed form on each side of the switch; then we price the acausality in laboratory numbers and read the verdict.

Step 1. Before the switch ( $t<0$ ): the integrand is nonzero only for $t'>0$ , so

a(t)=\frac{1}{m\tau}\int_0^{\infty}e^{-(t'-t)/\tau}F\,dt'=\frac{F}{m}\,e^{t/\tau}\qquad(t<0).

The charge is already accelerating — exponentially ramping up as the force's arrival approaches, with time constant $\tau$ .

Step 2. After the switch ( $t>0$ ):

a(t)=\frac{1}{m\tau}\int_t^{\infty}e^{-(t'-t)/\tau}F\,dt'=\frac{F}{m}\qquad(t>0),

— the ordinary Newtonian value, exactly, with no overshoot and no relaxation: all the anomaly lives before zero.

Step 3. Price the crime. The pre-switch velocity acquired: $v(0^-)=\int_{-\infty}^0\frac{F}{m}e^{t/\tau}dt=\frac{F\tau}{m}$ ; the pre-switch displacement $\frac{F\tau^2}{m}$ . For an electron meeting a fierce laboratory step of $E=10^6$ V/m: $F/m=1.8\times10^{17}$ m/s², pre-velocity $1.1\times10^{-6}$ m/s, pre-displacement $7\times10^{-30}$ m — fourteen orders of magnitude below a proton radius, thirty below anything measurable. The acausality is a matter of principle with no laboratory jurisdiction.

Step 4. The verdict, stated precisely. Within the literal third-order theory: initial-value causality $\Rightarrow$ runaways (energy nonconservation); runaway-freedom $\Rightarrow$ response precedes stimulus by $\sim\tau$ . There is no third setting. Since $\tau=2r_e/3c$ is exactly the scale where the point-charge idealization was already forfeit (the shell model's expansion parameter, the renormalization's hidden cutoff), the honest reading is not "electrodynamics is acausal" but "the classical point electron fails self-consistency at its own smallest scale, and announces it politely, in exponentially small print." Verified: reproduces Griffiths §11.2.3 (Eqs. 11.85–11.89, the worked step-force example) and Jackson §16.4; continuity of $a$ at $t=0$ checks from both sides; the pre-velocity $F\tau/m$ integrates from the exact profile.

Go deeper · Derivation C · 2 minThe Landau-Lifshitz reduction — deleting the pathologies at the stroke of a substitution

Derivation CThe Landau-Lifshitz reduction — deleting the pathologies at the stroke of a substitution

The goal is the modern working equation: since the self-force term is a first-order-in- $\tau$ correction, it has no right to change the order of the dynamics — so evaluate it on the zeroth-order motion. Strategy: substitute Newton's own law into the jerk, obtain the second-order Landau–Lifshitz (LL) equation, verify on two touchstones (the oscillator linewidth; cyclotron damping) that all confirmed physics survives, and confirm the pathologies do not.

Step 1. The substitution. To zeroth order in $\tau$ , $m\mathbf{a}=\mathbf{F}_{\text{ext}}$ , hence $m\dot{\mathbf{a}}=\dot{\mathbf{F}}_{\text{ext}}$ along the trajectory. Insert into the small term:

m\mathbf{a}=\mathbf{F}_{\text{ext}}+\tau\,\frac{d\mathbf{F}_{\text{ext}}}{dt}=\mathbf{F}_{\text{ext}}+\tau\Big[\frac{\partial\mathbf{F}_{\text{ext}}}{\partial t}+(\mathbf{v}\cdot\nabla)\mathbf{F}_{\text{ext}}\Big],

the total derivative unpacked along the motion; for the Lorentz force the chain rule adds the useful explicit piece $\tau\frac{q}{m}\,\mathbf{F}_{\text{ext}}\times\mathbf{B}$ . The equation is now second order: two data in, unique causal future out, error $O(\tau^2)$ — precisely the order at which the original equation was never trustworthy anyway.

Step 2. Touchstone one: the oscillator. $F_{\text{ext}}=-m\omega_0^2x$ gives $\tau\dot F=-m\tau\omega_0^2v$ : the equation becomes $\ddot x+\omega_0^2\tau\,\dot x+\omega_0^2x=0$ — the damped oscillator with $\gamma_{\text{cl}}=\omega_0^2\tau$ , exactly the linewidth physics the previous article confirmed against spectroscopy. Nothing verified is lost.

Step 3. Touchstone two: cyclotron damping. For $\mathbf{F}_{\text{ext}}=q\mathbf{v}\times\mathbf{B}$ , the LL correction $\tau\frac{q}{m}\mathbf{F}\times\mathbf{B}=\tau\frac{q^2}{m}(\mathbf{v}\times\mathbf{B})\times\mathbf{B}$ has a component $-\tau\frac{q^2B^2}{m}\mathbf{v}_\perp$ : a genuine drag. Its power drain, $\tau q^2B^2v_\perp^2/m$ , equals the Larmor power with $a=qv_\perp B/m$ identically (check: $\frac{\mu_0q^2}{6\pi c}\frac{q^2v^2B^2}{m^2}=\tau\frac{q^2B^2v^2}{m}$ by the definition of $\tau$ ), reproducing the 2.6 s radiative cooling at 1 T computed in the Larmor article — now from the force side of the books.

Step 4. The pathologies, audited out. Force-free: LL reads $m\mathbf{a}=0$ — no runaway exists to amputate. Step force: $\dot F$ is a delta function, giving a velocity jump of $F\tau/m$ at $t=0$ (the preacceleration's integrated effect relocated to the causal side, at the honest limit of the model's time resolution) and Newtonian motion elsewhere — no motion before the force, ever. The reduction keeps the physics, deletes the disease, and openly declares its warranty: valid when the force varies slowly on the scale $\tau$ , which is everywhere classical physics itself is valid. Verified: LL form matches Landau & Lifshitz §76 (nonrelativistic reduction of their Eq. 76.3); the oscillator and cyclotron reductions match the previous articles' independently derived rates; the relativistic LL equation (same logic run on the covariant ALD equation) is the standard of strong-field plasma codes.

Figure

The step-force acceleration profile — all the anomaly lives before zero, ramping up with time constant tau

Limiting cases

case	result	meaning
$\tau\dot a\ll a$	Newtonian dynamics $+O(\omega\tau)$ corrections	the entire classical world; both AL and LL agree here
$F_{\text{ext}}=0$ , literal equation	$a=a_0e^{t/\tau}$	the runaway: doubling every $4.3\times10^{-24}$ s
runaways banned	$a(t)=\frac{1}{m\tau}\int_t^\infty e^{-(t'-t)/\tau}F\,dt'$	future-reading kernel: preacceleration window $\sim\tau$
step force, banned runaways	$a=\frac Fm e^{t/\tau}$ ( $t<0$ ), $\frac Fm$ ( $t>0$ )	pre-displacement $F\tau^2/m\sim10^{-29}$ m for lab fields
extended charge, $\varepsilon>c\tau$	no runaway solutions exist	the pathology is the point limit's property, not electrodynamics'
Landau–Lifshitz	second-order, causal, correct to $O(\tau)$	the working equation; pathologies absent by construction
oscillator via LL	$\gamma_{\text{cl}}=\omega_0^2\tau$	linewidth physics preserved exactly
$\hbar$ on	QED at scale $\hbar/\tau\sim100$ MeV... (Compton scale intervenes first)	quantum theory takes over three orders before $\tau$ : at $\lambda_C/c\approx137\,\tau$

Common trapDo not solve the third-order equation literally, and do not report its pathologies as predictions

Two disciplined refusals define competent use of this subject. First: never integrate $m\mathbf{a}=\mathbf{F}_{\text{ext}}+m\tau\dot{\mathbf{a}}$ forward as an initial-value problem — numerically or analytically. The runaway mode grows by $e$ every $6.3\times10^{-24}$ s; any roundoff, any physical perturbation, any honest uncertainty in the unphysical third initial datum detonates it. (Numerical codes that tried are a small graveyard of the literature.) Use Landau–Lifshitz, or the future-integral form if the point of the exercise is the pathology itself. Second: runaways and preacceleration are diagnostics of domain failure, not physics. The electron does not preaccelerate; the classical point-electron model preaccelerates, on scales where it had already forfeited validity — note that the Compton wavelength, where quantum mechanics takes the wheel, is $137$ times larger than $c\tau$ . The correct citation is: "the classical theory is internally inconsistent below $\sim\tau$ , consistently with its known replacement by QED there." Anything stronger is a category error in both directions.

Worked examples

1 · Pricing the acausality — the preacceleration budget for a real field (confidence builder)

Setup. An electron sits at rest; at $t=0$ a uniform field $E=10^6$ V/m (a strong laboratory electrostatic field) switches on. Using the non-runaway solution, compute the preacceleration window, the velocity already acquired at $t=0$ , and the displacement before the force exists.

Step 1. The window. The profile $a(t<0)=\frac{F}{m}e^{t/\tau}$ has time constant $\tau=6.27\times10^{-24}$ s: the charge stirs, in any meaningful amplitude, only during the last few $\tau$ before the switch — call it $2\times10^{-23}$ s of sin.

Step 2. Velocity at the switch: $v(0)=\dfrac{F\tau}{m}=a\tau=(1.76\times10^{17})(6.27\times10^{-24})=\mathbf{1.1\times10^{-6}\ m/s}$ — a micron per second, acquired acausally.

Step 3. Displacement before the force: $x(0)\simeq\dfrac{F\tau^2}{m}=a\tau^2=\mathbf{7\times10^{-30}\ m}$ . Against reality's yardsticks: $10^{-14}$ of a proton radius; $10^{5}$ Planck lengths; and the field's own switch-on, however engineered, cannot be timed to better than attoseconds ( $10^{-18}$ s) — five orders of magnitude coarser than the entire acausal episode.

Step 4. Read the numbers as the theory's defense counsel. No experiment, actual or conceivable within classical physics, resolves the window; and every physical "step" force has a rise time enormously longer than $\tau$ , over which the future-integral solution is indistinguishable from causal Newtonian response. The acausality is a real property of the equations and an unfalsifiable one of the world — the cleanest possible sign that it belongs to the model's pathology, not to nature's.

The lesson: always convert a conceptual scandal into meters and seconds before deciding how scandalized to be.

2 · Landau–Lifshitz at work — radiative spiral in a magnetic field (exam level)

Setup. An electron gyrates in $B=1$ T with initial perpendicular speed $v_0=0.01c$ . Apply the LL equation to find the trajectory's decay law and timescale, and verify consistency with the Larmor-article energy route.

Step 1. Assemble the LL force. With $\mathbf{F}_{\text{ext}}=q\mathbf{v}\times\mathbf{B}$ (static, uniform: $\partial_t$ and $(\mathbf{v}\cdot\nabla)$ terms vanish), the correction is $\tau\frac{q}{m}\mathbf{F}_{\text{ext}}\times\mathbf{B}=\tau\frac{q^2}{m}(\mathbf{v}\times\mathbf{B})\times\mathbf{B}$ . Expanding the triple product with $\mathbf{v}\perp\mathbf{B}$ : $-\tau\frac{q^2B^2}{m}\mathbf{v}$ — a clean linear drag.

Step 2. The reduced dynamics: $m\dot{\mathbf{v}}=q\mathbf{v}\times\mathbf{B}-\tau\frac{q^2B^2}{m}\mathbf{v}$ . The magnetic term steers (no work); the drag saps: $\frac{d}{dt}\big(\tfrac12mv^2\big)=-\tau\frac{q^2B^2}{m}v^2$ , i.e. exponential energy decay with

t_{\text{damp}}=\frac{m}{2\tau q^2B^2/m}\cdot 2=\frac{m^2}{\tau q^2B^2}=\frac{3\pi\varepsilon_0m^3c^3}{q^4B^2}=\mathbf{2.6\ s\ at\ 1\ T}.

The electron spirals inward, orbit radius shrinking as $e^{-t/2t_{\text{damp}}}$ (radius goes as $v$ ), completing $\sim\omega_c t_{\text{damp}}/2\pi\approx7\times10^{10}$ turns on the way — the gentlest death spiral in physics.

Step 3. Cross-examine against the energy route. The Larmor article computed, from pure power bookkeeping, the same $2.6$ s at 1 T. Here the number emerges from the force side — the LL drag's work rate reproduces $P_{\text{Larmor}}$ identically (the definition of $\tau$ makes $\tau q^2B^2v^2/m\equiv\mu_0q^2a^2/6\pi c$ with $a=qvB/m$ ). Force ledger and energy ledger agree to the digit: the reduction preserved the physics.

Step 4. And cross-examine the pathologies. Set $B\to0$ : the LL equation reads $m\dot{\mathbf{v}}=0$ — the free electron coasts, no runaway available even in principle. The same model that damps the cyclotron orbit correctly is structurally incapable of the third-order diseases. That pairing — full retention of verified physics, categorical absence of pathology — is why LL is the production equation in every strong-field plasma and astrophysics code.

The lesson: reduction of order is not an approximation bolted onto a broken theory; within the theory's honest domain it is the theory, stated at its own accuracy.

Going further

References. Griffiths, Introduction to Electrodynamics, 4th ed., §11.2.3 · Jackson, Classical Electrodynamics, 3rd ed., §16.2–16.4 · Landau & Lifshitz, The Classical Theory of Fields, §75–76 · Dirac, "Classical theory of radiating electrons," Proc. R. Soc. A 167, 148 (1938) · Rohrlich, Classical Charged Particles, 3rd ed.

Provenance. Coefficient $\mu_0q^2/(6\pi c)$ and $\tau=\mu_0q^2/(6\pi mc)$ verifiable against Griffiths 4th ed. Eqs. 11.80–11.81 and Jackson 3rd ed. §16.2 (Gaussian form $2q^2\dot a/3c^3$ ); electron value $\tau\approx6.27\times10^{-24}$ s recomputable from $r_e=2.818\times10^{-15}$ m via $\tau=\tfrac23r_e/c$ . Runaway and step-force preacceleration solutions reproduce Griffiths §11.2.3 Eqs. 11.85–11.89; Landau–Lifshitz reduction from Landau & Lifshitz §76. A verifier should re-derive the integrating-factor general solution, check the step-force profile's continuity at $t=0$ from both sides, and confirm the LL cyclotron damping time $3\pi\varepsilon_0m^3c^3/q^4B^2$ against the Larmor-route value. Addendum for this edition: per the pair's division of labor, the force's derivations (energy balance with Schott fine print; the Lorentz shell with renormalization) are owned by Radiation Reaction Force and recalled, never re-derived; this edition owns the equation-as-dynamical-system analysis (Derivations A–B), the executed LL reduction with its two touchstones (Derivation C), and the effective-theory verdict; the covariant ALD derivation (Dirac 1938 route), the classical self-energy/4-3 program, the runaway taxonomy at depth, and the full linewidth theory are deferred by name to the E11 advanced articles. Worked numbers for this edition (preacceleration budget at $10^6$ V/m: window $\sim2\times10^{-23}$ s, $v(0)=1.1\times10^{-6}$ m/s, $x(0)=7\times10^{-30}$ m $=8\times10^{-15}$ proton radii; runaway e-folds $1.6\times10^8$ per fs; LL cyclotron $t_{\text{damp}}=2.6$ s at 1 T, $7\times10^{10}$ turns) recomputed for this edition.

Builds on: Radiation Reaction Force · Larmor Formula · Advanced Potentials and Causality · Fields of a Moving Point Charge Leads to: Postulates of Special Relativity · Jefimenko's Equations

Is the runaway solution just a mathematical artifact, or does it have physical content?

It has exactly one piece of physical content: it marks where the model's energy books are unbounded below. The renormalized point charge hides an infinite negative bare mass; the runaway is the dynamics discovering that reservoir and spending it — growing kinetic energy and Larmor output funded by an ever-more-negative unobservable sector. Two facts confirm this reading. An extended charge model with radius $\varepsilon>c\tau$ has a finite self-energy and provably no runaway solutions — the disease switches on precisely at the point limit. And in QED, where the electron's self-energy is renormalized consistently, no runaway counterpart survives. So: not an algebra slip, but a faithful symptom of an inconsistent idealization — which is why it cannot be cured within the model, only excised (future boundary condition) or side-stepped (reduction of order).

Why does the future-integral solution not violate causality observably — could a cleverly timed force reveal preacceleration?

Try to design the experiment and watch it dissolve. To see motion before the force, you need (i) a force whose onset is sharp on the scale $\tau\sim10^{-23}$ s — but any physical force is carried by fields that themselves have rise times set by their sources, at best attoseconds ( $10^5$ times too slow), and over such rise times the future kernel just smooths imperceptibly; (ii) a detector resolving displacements of order $F\tau^2/m\sim10^{-29}$ m — thirty orders beyond any interferometer, and far below the electron's own quantum position uncertainty in any bound or beam state ( $\gtrsim10^{-12}$ m); and (iii) the electron to be a classical point particle during the window — which quantum mechanics already denies at scales $137$ times larger. The acausality is thus triply protected from observation. That is not an excuse but the diagnosis: an inconsistency with no observable jurisdiction is a statement about the model's domain, and the domain boundary it draws ( $\sim\tau$ ) is exactly where we independently know the model ends.

If the Landau-Lifshitz equation is "the same theory to first order in tau," why do people call the Abraham-Lorentz equation wrong and LL right?

Because "same to $O(\tau)$ " is precisely the theory's entire warranty, and the two equations differ only in the unwarranted orders. The derivation of the self-force (energy balance, or the shell expansion) is itself controlled only to first order in $\tau$ ; the third-order equation's exact solutions — runaways, preacceleration profiles — are statements about $O(\tau^2)$ -and-beyond structure that the derivation never certified. LL is the unique second-order equation agreeing with AL at the certified order; its solutions stay within the certified domain by construction. So the accurate language: AL and LL are the same physics; AL over-commits mathematically and pays with pathologies; LL commits exactly as far as the derivation warrants. (There are exotic force profiles where the iterated reduction is subtle — delta kicks, the fine print of Step 4 — but none where honest classical physics distinguishes them experimentally.)

What exactly did Dirac's 1938 treatment add beyond Abraham and Lorentz?

Three permanent things. First, covariance: the Abraham–Lorentz–Dirac equation is the exact special-relativistic self-force, with the extra $u^\mu(a^\nu a_\nu)/c^2$ term that keeps the four-velocity normalized — indispensable for storage rings and strong-field work, and the parent from which this article's equation is the $v\ll c$ shadow. Second, point-particle rigor: no shell, no shape assumptions — just stress-energy conservation on a world-tube plus the retarded/advanced field split, with the divergence isolated covariantly and absorbed into the mass: renormalization as a principled subtraction rather than a model artifact. Third, the honest catalogue: Dirac solved for the pathologies rather than hiding them, initiating the effective-theory reading this article ends on. The price of his rigor: the pathologies are provably intrinsic to the classical point charge, not curable by better classical modeling — which is the sharpest form of the argument that the resolution had to come from outside classical physics.

The preacceleration reminds me of the advanced potentials. Is there a real connection?

A genuine kinship, running through Dirac's derivation. The self-force's finite part is built from the difference of retarded and advanced self-fields — the combination that is regular on the worldline — so the advanced solutions the causality article quarantined re-enter, legitimately, as a computational device for the radiation field. The preacceleration then has a satisfying reading: excising the runaway forces the solution to depend on the force's future within $\sim\tau$ , which is precisely the worldline-local residue of the advanced field's future-reading character. Wheeler–Feynman absorber theory makes the family relation explicit: in their time-symmetric electrodynamics, the absorber's response converts half-retarded-plus-half-advanced into effective retardation plus exactly the radiation reaction force — the self-force emerging as the universe's advanced echo. None of this licenses observable acausality (the budgets of worked example 1 stand); it locates the equation's acausal whisper in the same mathematical drawer where classical electrodynamics keeps all its time-symmetric bookkeeping.

After renormalization, runaway excision, and reduction of order — what is actually left of the classical theory of the electron?

A surprisingly dignified estate. What is gone: the ambition of a complete classical model of the electron — structure, self-energy, and dynamics all classical, all consistent, all the way down. Abraham's electromagnetic-mass program died here. What remains: a controlled effective theory, quantitatively superb in its domain — radiation damping in accelerators (designed daily with these equations), classical linewidths, antenna reaction, laser-plasma dynamics up to the quantum boundary — plus a set of concepts (renormalization, effective validity ranges, reduction of order, pathology-as-diagnosis) that turned out to be the methodological skeleton of modern field theory. The classical electron taught physics how to fail correctly: how to extract every valid prediction from a theory while reading its inconsistencies as coordinates of the next theory's border post. As the chapter closes, that border post has a name — the scale where $\hbar$ enters — and the road onward from here runs first through relativity, whose electrodynamic origins the next article finally makes explicit.