Naimark dilation

In plain terms

A textbook quantum measurement is clean. Its outcomes are mutually exclusive, each outcome corresponds to a sharp yes-or-no test on the system, and the boundaries between outcomes are exact. Real detectors aren't like that, as they're noisy, they blur neighbouring outcomes together, and they sometimes report more outcomes than the system has distinguishable states. The general description of such a detector is a generalised measurement, usually called a POVM, which is a list of effects, one for each outcome, each of which assigns a probability to its outcome without the outcomes needing to be sharply separated. On its face this is a different and worse- behaved kind of object.

Naimark's theorem says the mess is a matter of perspective. Any generalised measurement, however noisy, is an ordinary sharp measurement performed on the system together with an extra piece of apparatus. Enlarge the system to include that apparatus, a step called dilation, and the complicated measurement becomes a simple one, so the blur is what a sharp measurement of the combined system looks like when you keep track of the original system alone. A rough analogy is a shadow, as a blurred, complicated shadow on a wall can be cast by a simple solid object, and the blur belongs to the act of projecting down and not to the object. The analogy has one limit worth knowing, which is that the larger system in the theorem is a mathematical construction and it needn't coincide with the physical detector on the bench.

The consequence is that there's no separate theory of realistic measurements. There's one theory, sometimes viewed through a keyhole.

In CSD

CSD derives the Born weights as volumes, so each outcome corresponds to a region of the space of states, and the probability of the outcome is the fraction of the natural volume the region occupies. On its own that argument covers only ideal sharp measurements, and since essentially every real experiment is a generalised measurement, this would be a serious gap for a programme that claims to be about physics rather than textbooks.

Naimark's theorem closes the gap by transfer rather than by a second derivation. Dilate the measurement to the larger system, run the same volume-ratio argument there, and pull the answer back down. The generalised case then holds with no extra assumption, as nothing about the measurement needs to be typical or generic for the transfer to work.

The corpus builds the dilation explicitly and then uses it repeatedly. Symmetric informationally complete measurements, mutually unbiased bases, trine measurements, unambiguous state discrimination and weak measurements are each constructed through the same canonical dilation, and each inherits the volume statement without further work, so the generalised-measurement chapter of the reconstruction is recorded as closed on exactly this mechanism.

Mathematically

For a POVM equipped with a Naimark dilation, the Born weight of each effect equals the Fubini-Study volume of the corresponding outcome region in the dilated space, and the frequency statement transfers along with it. The result is unconditional, as no genericity hypothesis is needed, and the corpus theorem is povm_born_eq_dilated_volume_uncond.

The dilation is constructed rather than assumed. It's built canonically from the square roots of the effects, so existence comes with an explicit witness, and the instances in the corpus inherit the theorem instead of re-proving it.

The name

Mark Naimark (1909-1978) was born in Odessa, the son of an artist, and worked in Moscow at the Steklov Institute while teaching at the Moscow Institute of Physics and Technology. The dilation theorem dates from 1940. Its close cousins, due to Stinespring and Sz.-Nagy, form a small family of extension theorems discovered across a decade divided by war.

His best-known work is the 1943 Gelfand-Naimark theorem characterising C-star algebras, written with Israel Gelfand, whose Moscow seminar was one of the most productive mathematical institutions of the century, a weekly event that ran for fifty years and was famous for the vigour of its arguments.

Related entries

• Stinespring dilationEvery noisy quantum process is a reversible process on a larger system with part of that system ignored, and in CSD the larger system is the constraint surface itself, so every channel the programme uses is the environment's marginal of a volume- preserving flow.
• Born ruleQuantum probabilities aren't a law of nature in CSD, as they're the sizes of regions in the space of states, measured with a ruler nobody was free to choose.
• Luders ruleThe textbook's second rule, which says how a state changes when it's measured, isn't a second rule in CSD, as it's what the record layer does.
• Measurement flowA deterministic, volume-preserving flow carries out a measurement, and the volumes of its pointer blocks are the Born weights, for every preparation.
• Gleason and Busch theoremsThe corpus proves the standard route to the Born rule rather than borrowing it, and CSD's own derivation never uses it, so the two routes to quantum probability are genuinely independent.