Sources & disclaimers.

What this site claims — and what it doesn't

All interpretations presented here reproduce the standard experimental predictions of quantum mechanics; on ordinary experiments, no measurement can tell them apart. The one qualified exception is the family of objective-collapse models, which modify the dynamics slightly and are therefore testable in principle — so far, every such test has come back null, tightening the allowed parameter space.

The interactive scenes are conceptual representations, not simulations of the underlying physics. Curves stand for amplitudes, dots for detection events, diverging tracks for branches; none of it is drawn to physical scale, and the wave pictures are representations of a mathematical object, not of a material wave.

Where an interpretation is presented, its claims are attributed (“according to this interpretation…”). Nothing on this site asserts that consciousness causes collapse; “observation” and “measurement” refer to physical interactions that leave records, not to a person looking.

Copenhagen is presented as a family of related positions rather than one uniform doctrine; the same is true, to a lesser degree, of the other interpretations.

Known simplifications in the interactive models

• §The double-slit detection pattern is sampled from a stylized intensity curve (a cosine-squared fringe under a Gaussian envelope, or the sum of the two single-slit patterns with the which-path detector on) — not from solving the Schrödinger equation.
• §The pilot-wave trajectories are illustrative paths consistent with the qualitative behaviour of de Broglie–Bohm trajectories, not numerically integrated guidance-equation solutions.
• §The measurement scene's collapse timing under objective collapse is a random draw for teaching purposes; real GRW/CSL rates depend on mass and configuration and are vastly slower for microscopic systems.
• §Branching in the many-worlds visuals shows two branches; realistic measurements produce enormous families of decohering branches.
• §The superposition scene shows a two-outcome system; amplitudes are real-valued in the visual (no phase), so interference is illustrated only in the double-slit scene.
• §The Bell scene draws its quantum correlations from the exact singlet prediction, E(θ) = −cos θ, but samples them a pair at a time — so short runs wander around the curve, exactly as a real experiment does.
• §The dashed curve is not an envelope over all local theories. At any single angle a local model can match the quantum curve exactly; what Bell bounds is the four-setting CHSH combination, which is why the comparison is made there.
• §The CHSH value is estimated from binned correlations on the assumption that only the angle between the detectors matters, and three of the four test settings fall in the same bin — so its errors compound. The figure reports a standard error and only calls a violation at three of them.
• §The evolution packet widens in step with the clock. A free Gaussian packet actually widens as the square root of one plus the square of the elapsed time, so the early part of the run is drawn faster than it really is.
• §Under objective collapse the localization moment is drawn from a uniform range for pacing. Real GRW and CSL hit times are exponentially distributed, at a rate set by the number of particles rather than by the state.
• §The “local hidden variables” source in the Bell scene is one specific deterministic local model, chosen because it does as well as any local theory possibly can (it reaches the CHSH bound of 2 and no further). Bell's theorem rules out every local hidden-variable theory, not only the one drawn here.
• §Because that local model sits exactly on the bound, short runs of it will sometimes measure a little above 2. That is sampling noise, not a violation — real experiments need large samples and careful error bars for the same reason.
• §The entanglement scene assumes an ideal pair and perfect detectors. Real Bell tests must also close the detection and locality loopholes, which is what made the 2015 experiments significant.
• §With the which-path detector on, the model draws the plain sum of the two single-slit patterns — one broad band with the fringes gone, as a real run gives. What it cannot show is partial which-path information, which yields partial fringes: this detector is all or nothing.
• §The cat scene's coherence time is an illustrative scaling — one decade of time lost per decade of particles, anchored arbitrarily — not a calculation. It follows the shape of Zurek's estimate, in which decoherence time falls inversely with mass, but not its magnitudes: Joos and Zeh put a ten-micrometre dust grain in air below 10⁻³⁰ seconds, and the heaviest object yet shown to interfere is a molecule of about two thousand atoms.
• §The evolution scene shows a one-dimensional packet spreading with no potential and no phase structure; measuring draws a position from that packet.