A more difficult witness
Four ways nature reveals that a particle has something inside
A photon cannot be parked on a table. It cannot be brought to rest. It has no charge distribution to map. So how can anyone claim that light has no inside?
Begin the interrogationPart 1 interrogated the proton. It left three kinds of evidence: a size, a spectrum of resonances, and debris shaped by quarks and gluons.
Electron: aim probes at it. Photon: make light interact, then check the outcome.Same verdict · different evidence
An electron can be slowed, steered, and examined in a controlled beam. A photon never offers a rest-frame portrait. Both are called elementary because every appropriate test still works without hidden machinery.
WITNESS AElectron
Scattering, excitation searches, and precision measurements still describe the electron as one pointlike degree of freedom.
WITNESS BPhoton
No extra polarization, massive excitation ladder, or unexplained photon interaction has been required.
THE BOTTOM OF SCIENCE · PART 2Aim probes at the electron.
Make photons interact.One can be slowed and mapped. The other must be tested in flight.
Four ways an inside leaks out
Size is only the first clue. Parts can sit at different positions, move relative to one another, redistribute momentum, and modify an interaction long before a detector sees dramatic wreckage.
A short-wavelength probe scatters from charge at many positions. Those waves reinforce or cancel, so the measured rate changes with angle and probe scale. From that change, physicists infer how the charge is distributed.
Internal degrees of freedom rearrange only at particular energies. A sweep reveals resonance peaks: the quantum notes of a composite object.
Fragments, jets, momentum sharing, and correlations reveal how the original object carried its energy—even when confinement hides isolated parts.
At low energy, the standard and hidden-structure predictions may nearly overlap. At higher energy, an extra short-distance interaction can change rates or angles more strongly, bending the data away from the standard curve.
The Particle Interrogation Room
Each test has the same logic: predict what hidden parts would do, look at the experiment, then compare the two.
Stored constituents should also reveal themselves through matching form factors, resonances, or recurring momentum patterns.
Near a nucleus, photon energy supplies the pair’s mass and motion while the nucleus takes recoil. QED predicts the event without stored electrons.
One dramatic event is not enough. The strongest case for constituents is when shape, resonances, breakup patterns, and precision measurements all point to the same inner pieces.
Photon, Hit it hard: The electron and positron are created in the interaction.
Test 2 · the particle rings
Push a trolley and it rolls faster. Strike a bell and particular notes appear because its parts move relative to one another. Quantum composites have their own version: heavier internal arrangements that appear as peaks at specific collision energies.
Fire pions at protons. Change the total energy of the collision and count how often each outcome occurs.
Collisions pile up here. At this fixed invariant mass, the quark–gluon system can form the short-lived Δ(1232) baryon.
PDG · Δ baryonsIf an electron had rearrangeable parts, collisions could repeatedly produce a heavier excited electron at one invariant mass. Searches reconstruct the outgoing electron and photon; no e* peak has appeared in the tested models.
A photon’s energy changes with the observer. A true internal excitation would instead create a peak at a fixed invariant mass in collision products. No such photon ladder has been found.
Red photon, blue photon: same particle. Run towards red light and it arrives bluer. Its energy is definite once we name a frame and conserved in that frame, but another observer assigns another value.
A resonance survives the change of viewpoint. Every observer agrees that the Δ is heavier than the proton. That invariant extra mass is the signature of a new internal arrangement.
RESONANCE ≠ SPEEDAn excited state is the object’s insides moving differently.A faster object carries more motion. An excited object carries a new internal arrangement.
Test 3 · breakup drama
A collision can expose components that were bound together. It can also spend energy to manufacture entirely new particles. Detectors see the ending; physics must decide which story produced it.
In a nuclear field, pair production begins just above 1.022 MeV. The photon vanishes; the electron and positron usually continue roughly along its old direction and curve oppositely in a magnetic field. The nucleus takes a small recoil.
NIST · pair-production thresholdFinal particles appeared; stored constituents were not recovered.Energy is currency, not inventory. An unlit filament is not a box of visible photons. Switch it on: electrical energy heats the wire, and the filament creates light.
Deep-inelastic scattering measures how quarks and gluons share the proton’s momentum. The same momentum pattern predicts the directions and energies of jets. Proton form factors and resonances independently point to that same quark–gluon interior.
Pair production needs 1.022 MeV for the two rest masses, plus motion; the nucleus takes recoil. QED predicts the threshold, rate, and tracks without putting an electron and positron inside the photon. No photon form factor or resonance supplies a second clue that they were stored there.
BREAKUP VERSUS CREATIONDid the collision reveal old parts—or pay for new ones?The difference is corroboration: do several independent tests point to the same hidden pieces?
Test 4 · the numbers drift
Suppose the parts are too small to resolve directly. They can still add a faint interaction. Compare the measured number with the point-particle prediction; a mismatch that grows with energy is the clue.
An electron has mass and charge. It can be held in a trap, steered in a beam, and struck by other particles. Hidden parts could change its magnetic response or make scattering peel away from the point-electron prediction.
Fan and colleagues measured the electron’s magnetic moment with 0.13-parts-per-trillion uncertainty. Comparing it with the Standard Model tests the theory at about one part in 10¹² and depends on an independent measurement of the fine-structure constant. Fan et al. · PRL
A 2025 PDG benchmark excludes an excited electron below 5.6 TeV when its mass equals the compositeness scale. The number belongs to that model; it is not an electron radius. PDG 2025
A photon cannot be stopped and has no static charge map. So make photons collide or convert, then count what comes out. Hidden parts could add a third polarization, a new threshold or resonance, or an energy-dependent change in the rate.
ATLAS measured light-by-light scattering across diphoton invariant masses from 6 to 100 GeV in lead-ion data at 5.02 TeV per nucleon pair. The energy and angle distributions followed the tested Standard Model prediction. ATLAS
The quantum-cloud complication
The clean classroom picture—a bare dot travelling through empty space—is useful and incomplete. Calculations of an interacting particle include the response of the fields coupled to it. Turning those terms into a little swarm that escorts the particle is where confusion starts.
The calculation includes terms in which an electron emits and reabsorbs photons; charged-particle pairs contribute too. These quantum corrections help predict its magnetic moment with extraordinary precision. They describe the interacting electron’s measurable behaviour, not tiny gears bound inside a casing.
A photon can couple to a charged pair. In high-energy collisions, a nearly on-shell photon can be described with quark and gluon distributions; LEP measured these photon structure functions. Their pointlike contribution is calculable from the photon’s couplings. A hadron-like contribution is modelled and constrained by data. Neither describes a permanent crew inside every photon. L3 at LEP
Two electrons moving at different speeds still have the same intrinsic mass, charge, and interactions: they are identical particles in different states of motion. Two ordinary protons are identical in the same sense even though each contains quarks and gluons. Put enough invariant energy into that quark–gluon system and a short-lived baryon resonance such as the Δ can form; that is not an ordinary proton merely travelling faster.
Signals of a deeper layer
A deeper layer must leave a repeatable pattern: several measurements should point to the same new mass or energy scale.
The current experimental summary
Finite size
Resonances
Constituent dynamics
No resolved radius
No e* resonance
Precision rules hold
No extra polarization
No unexplained resonance
Tested interactions hold
This interrogation zoomed in on three examples. In the Standard Model, quarks, neutrinos, the other charged leptons, gluons, the W and Z bosons, and the Higgs are also treated as elementary.
Proton data require quarks and gluons with internal motion. Electron and photon data require no internal degrees of freedom at the scales tested. In the Standard Model, they are elementary.
Fourteen useful objections
No experiment has found permanent smaller ingredients inside a photon. The Standard Model treats it as an elementary excitation of the electromagnetic field. A repeatable new form factor, polarization, threshold, resonance, or interaction pattern would be evidence for a deeper layer.
A photon can be prepared as a wave packet spread over a region, but that spread is not a little casing or measured charge radius. A photon has no rest frame and no ordinary static charge distribution. Experiments therefore constrain its behaviour and possible new interactions rather than photographing an edge.
We collide photons, convert them in matter, measure their polarization, and compare interaction rates and angles with precise quantum-electrodynamic predictions. A rest-frame portrait is only one kind of test; moving particles still leave exact statistical patterns.
No. Blue and red photons are the same particle species with different energy and momentum in a named frame. Run towards red light and you measure a higher frequency. An excited state instead has a particular extra invariant mass, which every observer agrees on. The proton has such resonances; the photon has no established ladder of heavier internal versions.
No. In a nuclear field, pair production begins just above 1.022 MeV. The nucleus takes recoil while the photon energy creates the electron, positron, and their motion. The same logic lets many collisions create particles that were absent from the incoming objects.
It is a hypothetical heavier partner that would reveal hidden electron constituents rearranging internally. Searches assume specific production and decay rules, so their limits are model-dependent. The 2025 Particle Data Group review reports no excited electron below 5.6 TeV in one standard benchmark.
Parts can move relative to one another, producing particular allowed energy arrangements. Those arrangements appear as resonance peaks. Simply accelerating a pointlike particle changes its motion continuously; it does not manufacture a ladder of internal notes.
A single spray would be ambiguous. The evidence is the repeated pattern: jet numbers, directions, energies, correlations, scaling in deep-inelastic scattering, and agreement across many processes with one quark-and-gluon theory. The same theory predicts when and how the sprays should appear.
The strong interaction confines quarks. Pulling one away stores enough energy in the colour field to create new quark–antiquark pairs. Detectors receive colour-neutral hadrons clustered into jets, not an isolated quark.
Usually no. “Virtual particle” is bookkeeping language for intermediate contributions in a quantum calculation. Constituents are persistent degrees of freedom tied together in a bound identity. Quantum fluctuations can change measured interactions without being little objects stored inside a shell.
No. “Identical” means the same intrinsic mass, charge, spin magnitude, and interactions—not the same motion or location. Two electrons at different speeds are the same species in different states. Two ordinary protons are also identical in that sense even though protons are composite. A short-lived baryon resonance has a different invariant mass and is a different state or species. Field identity tells us what kind of particle it is; scattering and spectroscopy tell us whether that kind has parts.
Yes. Any hidden dynamics must begin beyond current sensitivity, imitate all successful Standard Model predictions at accessible energies, and avoid producing the form factors, resonances, contact interactions, or excited states already sought.
A reproducible extra polarization, a new fixed-mass state tied consistently to photon interactions, an unexplained threshold, or an energy-dependent departure in processes such as light-by-light scattering would force a rethink. Several independent clues pointing to the same new scale would make the strongest case.
Different machines improve different questions. More LHC data sharpen rare high-energy photon and excited-lepton searches. A clean high-energy lepton collider would improve precision and contact-interaction tests. A future muon or very-high-energy hadron collider could reach heavier hidden dynamics directly.