• home
• about
• new
• essays
• AI
• book

John Abramson is retired and lives in the Lake District in Cumbria, England. He obtained an MSc in Transpersonal Psychology and Consciousness Studies in 2011 when Les Lancaster and Mike Daniels ran this course at Liverpool John Moores University. In 2015, he received an MA in Buddhist Studies from the University of South Wales. He can be contacted at johnabramson@btinternet.com

Where the Map Must Meet the Territory

A Reply to Frank Visser's "The Map Is Not the Territory"

John Abramson / Claude

Frank's challenge to me, pressed through this whole exchange, has been simple: stop interpreting and predict something. This reply places two bets on the table. Both can be settled by experiment—against me. One has already faced a first test, in a laboratory deep under the Gran Sasso mountain, and survived it; a rival research programme did not. The other comes with its escape hatch welded shut in advance, by my own hand. On the way to the bets I concede two of Frank's points in full, because he is right about them, and I ask him for exactly one thing in return. Nothing here asks the reader to adopt a worldview. It asks that a disagreement, narrowed this far, be settled the way disagreements should be: by theorems and by laboratories.

First concession: the argument so far accommodated, it did not predict

Frank asks whether my five quantum puzzles are five independent problems solved by one mechanism, or five places where the same interpretation was applied after the fact. He is right that the framework came first and was brought to the puzzles second. A theory that predicts in advance earns more credibility than one that explains after the fact. I concede this fully—for the argument I have made so far. The five dissolutions are readings of known facts, and I claim no more for them than accommodation can support. The framework can also make testable commitments about experiments not yet done; I set those out below.

Frank also warns that a sufficiently broad framework can dissolve any paradox just by changing the vocabulary. True—and this is the main danger for a proposal like mine. My answer is not a promise but a test I apply to my own work, in both directions. First: a structural concept stays in an argument only if deleting it makes the argument collapse. Second: work done by ordinary science is never credited to the framework. The evidence that this test cuts is what it refuses. Evolution is the clearest case: I examined it, found that Darwinian theory is complete on its own terms, and claimed nothing. Several other tempting extensions have been examined and refused this year on explicit counterexamples. A skeleton key opens every door. This framework keeps declining doors. That does not prove it true. It shows it is selective.

Where the disagreement lives

Frank's central alternative is this: perhaps there is just one world, and it is simply stranger than classical intuition allows. No second level needed. He calls this ontological economy, and he is right that whoever adds to the ontology carries the burden of proof.

Here is why I think that option is not the modest middle position it appears to be. Two sets of theorems close it in from both sides.

First side. The no-go theorems—Kochen-Specker, Bell, and their modern unification—prove that there is no consistent way to assign definite values to all physical quantities across all measurement contexts. Not "we haven't found one." There provably is none. So a world of jointly definite, locally resident facts is too small to contain what quantum theory describes. That is a theorem, not a preference.

Second side. Frank's own title suggests the natural escape: perhaps the quantum state is only map, not territory—an accurate description, not a real thing. But this exact question stopped being a metaphor in 2012. Pusey, Barrett and Rudolph proved that any theory reproducing quantum statistics must treat the quantum state as a feature of reality, not as a mere state of knowledge. (The theorem needs one assumption—that independent preparations are really independent—and I flag that honestly; but the assumption is weak and the result has stood for over a decade.)

Put the two together. The state is real, and it cannot live among the definite local facts. Real, and homeless in the manifest world. That is the situation my two-level reading names. One theorem supplies each half.

So I want to ask Frank for one thing in return for the concessions above: state the one-strange-world alternative at the same level of detail as the view it is meant to replace. My proposal is fully specified in the paper, with its costs itemised.[1] When "one strange world" is specified, I believe it goes one of three ways. Either it quietly restores jointly definite facts in one arena—closed by theorem. Or it becomes one of the known interpretations (Everett, treating the state as law rather than thing, moving reality into configuration space)—each of which pays a known price that my paper sets out. Or it ends up saying: there is a real, non-spatial global object, and there are definite spatiotemporal outcomes, and they are not the same kind of thing—which is my view in different words. Until it is specified, "one strange world" is not a rival ontology. It is a placeholder. And an unspecified placeholder should not enjoy the presumption of economy over a specified proposal.

Second concession: the final step is interpretive

Even granting all of the above, Frank's instinct that there is daylight between "the classical picture fails" and "reality has two levels" is correct. The theorems force non-classicality. They do not, by themselves, force the two-level description over "one non-classical whole."

So let me state the extra premise I rely on, in one sentence, with its failure condition. The premise: a mathematically indispensable, experimentally certified global object that provably cannot reside among the definite local facts should be treated as real and as belonging to its own level, not demoted to bookkeeping. It fails if the state can be shown dispensable or merely representational—which is exactly what the Pusey-Barrett-Rudolph result constrains, and why that theorem matters here.

That is my working method throughout: every step not forced by a theorem is carried as a named assumption with a stated failure condition. On Frank's related point—that mathematics existing does not make it a physical mechanism—I simply agree, and the paper says so. The mathematics describes the level; the physical no-go results are what argue it exists.

What the framework forbids

Frank distilled his challenge into one line: "show me what the second realm predicts that the first realm cannot."

First, a distinction, so that this section does not take back the concession above. The five dissolutions were accommodation: readings of facts already known. What follows are forward commitments: statements about experiments not yet done, which can fail. Accommodation looks backward; a bet looks forward. I conceded the first. I now offer the second.

And let me be exact about what can and cannot be offered. An interpretation that leaves the quantum formalism untouched cannot predict a number that quantum mechanics does not. What it can do is forbid things—and forbidding is predicting, because forbidden things can show up in a laboratory. My framework forbids two things, one against each of its main rivals.

Against the collapse theories (GRW, CSL, Diósi-Penrose). Those theories modify quantum dynamics, so they predict physical traces: faint spontaneous radiation, anomalous heating, noise. On my view, projection is not a mechanism operating inside the manifest world—nothing crosses, nothing diffuses—so there are no such traces to find. Here I must be honest about tenses. The sharpest test so far predates my paper, so I cannot claim to have predicted it: the Gran Sasso underground experiment (Donadi and colleagues, Nature Physics, 2021) measured the radiation that the parameter-free Diósi-Penrose model requires, found none, and ruled that model out. What I can say is this: my framework entails that null result—had the radiation been found, my view would be dead on arrival. So Gran Sasso is a test my framework has already survived and the collapse programme did not. And the bet now runs forward: I commit that the null results will continue: any confirmed detection of the radiation or heating predicted by dynamical collapse models would refute the framework. Each such experiment is a fresh chance to kill my view. I claim no foresight about Gran Sasso; I claim only that the framework is inconsistent with the radiation that experiment ruled out.

Against Everett. On my view, projection is a real, completed event. Once genuine registration has occurred, it cannot be coherently undone: no experiment will ever revive interference across a completed registration. Once a registration that meets the (still to be published) criterion has occurred, interference across that registration will not be recoverable. Everett's picture says the opposite—with enough control, any measurement is reversible in principle.

Each bet is shared with one rival against the other. Together they are mine alone: a real boundary, with no mechanism inside the manifest world. That is the corridor my framework lives in, and experiment is narrowing it from both sides. Find collapse noise, and I am wrong one way. Coherently reverse a completed registration, and I am wrong the other way.

One debt remains, and I state it rather than hide it. The second bet becomes sharp only when the framework says, in advance, exactly what counts as registration. Until that criterion is published, I am not entitled to explain away any future revival by saying the registration was not real—that would be the immunizing move Frank rightly condemns, and I bind myself against it now. Producing that criterion is the framework's central research task. It is also where the specific threshold Frank asks for would come from: derived from the framework, not improvised for an exchange. If that task is never completed, the framework remains what Frank says it currently is—an interpretation, though one with two standing bets on the table.

Accepting the invitation

Frank asked to see the full argument, and the paper is on its way to him. One request when reading it: almost all of it is scaffolding. The load-bearing wall is the reclassification of the no-go theorems (§7.2) together with the legitimacy test that governs every application. If that section fails, the framework fails. If it holds, the two-level reading is no longer optional decoration. I would rather Frank spend his fire there than anywhere else.

So the position is now simple. One claim of mine to check, in one section of one paper. One alternative of Frank's to specify, so it can be checked the same way. Two standing bets that experiment can settle against me. And one criterion owed, named as the debt it is. That seems to me exactly where an honest disagreement should stand.

NOTES

[1] John Abramson, "Quantum Paradox as Category Error: A Consilience Argument for a Two-Level Ontology", unpublished paper.