Projection and Invariance in Scientific Explanation

Harry Sticker

arxiv: 2603.19323 · v4 · submitted 2026-03-17 · ⚛️ physics.hist-ph

Projection and Invariance in Scientific Explanation

Harry Sticker This is my paper

Pith reviewed 2026-05-15 10:22 UTC · model grok-4.3

classification ⚛️ physics.hist-ph

keywords projectioninvariancescientific explanationperspectival structurepluralismrenormalization grouplevel-relative realismequivalence classes

0 comments

The pith

Projection from complex systems to equivalence classes reveals invariants by omitting within-class variation.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that any scientific representation must omit variation to function and that this omission occurs through projection, a mapping that partitions states into equivalence classes and makes patterns visible. This process is both revelatory, granting access to invariants, and constitutive, defining the concepts needed to express them. It distinguishes vertical projections, where earlier theories survive as recoverable limiting cases, from horizontal ones, where the invariants depend on the specific projection. A sympathetic reader would care because the account explains the continued use of Newtonian mechanics and the persistence of pluralism in mature sciences without treating either as a temporary defect.

Core claim

Projection is a principled mapping from underlying complexity to a descriptive space that partitions states into equivalence classes, omits within-class variation, and makes patterns visible that would otherwise be lost. It is simultaneously revelatory and constitutive: it makes genuine invariants tractably accessible while bringing into being the concepts through which they become expressible. Vertical cases allow earlier projections to survive as limiting cases with recoverable omission, while horizontal cases render omission constitutive so that invariants are accessible only at the level of the projection that defines them.

What carries the argument

Projection: a principled mapping from underlying complexity to a descriptive space of equivalence classes that omits within-class variation to reveal invariants.

If this is right

Persistent pluralism in mature sciences arises because different horizontal projections access different invariants.
The renormalization group functions as a systematic implementation of the invariant-tracking criterion.
Higher-level projections reveal genuine structural features of the world, supporting level-relative realism.
Scientific progress consists in refining projections rather than eliminating omission.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

No description that omitted nothing could explain anything, since it would be identical to its subject.
Attempts to reduce all phenomena to a single projection would necessarily miss invariants visible only under other projections.
The vertical-horizontal distinction could be applied to theory choice in fields such as economics or systems biology to classify when approximations are limits versus when they are constitutive.
A direct test would involve checking whether new projections developed in complex-systems research exhibit the same revelatory-constitutive duality.

Load-bearing premise

The distinction between vertical and horizontal projections can be drawn in a non-circular way that tracks which omissions are recoverable versus constitutive without depending on the invariants the projection is supposed to reveal.

What would settle it

A documented case in which a projection classified as horizontal has its invariants remain fully detectable and unchanged after a vertical refinement that recovers all previously omitted variation would falsify the constitutive character of horizontal omission.

read the original abstract

Any representational enterprise must omit variation in order to function. NASA still uses Newtonian mechanics, though Einstein superseded Newton, and the standard picture of scientific progress cannot explain how. A description that omitted nothing would be identical to its subject and would explain nothing. This paper argues that omission is not a defect but the central structural feature of any enterprise that builds representations from incomplete information. The key concept is projection: a principled mapping from underlying complexity to a descriptive space that partitions states into equivalence classes, omits within-class variation, and makes patterns visible that would otherwise be lost. Projection is simultaneously revelatory and constitutive: it makes genuine invariants tractably accessible while bringing into being the concepts through which they become expressible. The paper distinguishes vertical cases, in which earlier projections survive as limiting cases of more refined successors with recoverable omission, from horizontal cases, in which omission is constitutive, and invariants are accessible only at the level of the projection that defines them. The framework accounts for persistent pluralism in mature sciences, treats the renormalization group as a systematic implementation of the invariant-tracking criterion, and defends a level-relative realism on which higher-level projections reveal genuine structural features of the world. The deepest claim is an inversion of the standard picture: perspectival structure is not a concession to complexity but the condition for invariant detection. A world rich in invariants cannot be exhausted by a single projection.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Sticker's paper inverts the usual view by treating projection as necessary for spotting invariants rather than a limitation, but the vertical/horizontal distinction looks circular without an independent test.

read the letter

The main takeaway is that this paper treats omission in scientific representation as the feature that makes invariants detectable, not a defect to be minimized. Projection works by mapping complex states to equivalence classes that drop within-class variation, and Sticker claims this move is both revelatory and constitutive: it surfaces genuine invariants while creating the concepts needed to express them. The vertical cases keep earlier projections as recoverable limits, like Newtonian mechanics under Einstein, while horizontal cases make the omissions definitional so that invariants only appear at that specific level. This setup is meant to explain why pluralism stays around in mature physics and why renormalization tracks invariants systematically. It also supports a level-relative realism where higher projections capture real structure. That inversion of the standard picture is the clearest new move, and the abstract shows it handles effective theories without forcing everything into one ultimate description. The argument is coherent on its own terms and engages the right literature on explanation and theory change. The soft spot is the circularity risk in the vertical/horizontal split. Classifying an omission as recoverable seems to require already knowing which invariants survive the limit or mapping, yet those invariants are supposed to be what the projection reveals. The abstract gives no separate syntactic or behavioral criterion that avoids this loop, so the account of pluralism and realism rests on an interpretive distinction that may not hold up independently. Concrete cases are illustrated but not mechanically checked against an external test. This is for philosophers of science who work on realism, pluralism, and effective theories. A reader looking for a unified conceptual tool to reorganize those debates would find it useful, even if they end up disagreeing with the inversion. It deserves peer review because the core framing is clear and the engagement with renormalization and level-relative realism is substantive enough to warrant referee input on the classification issue.

Referee Report

2 major / 2 minor

Summary. The paper argues that scientific representations necessarily omit variation via 'projection'—a mapping from underlying complexity to equivalence classes that reveals invariants while constituting the concepts through which they are expressed. It distinguishes vertical projections (earlier descriptions survive as limits with recoverable omissions, e.g., Newtonian mechanics) from horizontal ones (omissions are constitutive and invariants accessible only at that level), accounts for persistent pluralism, treats the renormalization group as systematic invariant-tracking, and inverts the standard view by claiming perspectival structure is required for invariant detection rather than a concession to complexity.

Significance. If the central inversion holds, the framework offers a unified conceptual tool for understanding level-relative realism and pluralism in mature sciences, with the renormalization-group application providing a concrete bridge to physics practice. The interpretive approach supplies no new formal axioms or data but organizes existing cases around the recoverable/constitutive omission distinction; its value lies in reframing omission as revelatory rather than defective.

major comments (2)

[Abstract and §3] Abstract and §3 (Vertical/Horizontal distinction): the partition into vertical cases (recoverable omissions, earlier projections survive as limits) and horizontal cases (constitutive omissions) is defined in terms of the invariants each projection reveals, yet no antecedent, non-circular criterion—such as purely syntactic limit behavior or independent equivalence-class test—is supplied for classifying a given omission as recoverable versus constitutive. This renders the classification dependent on the very invariant structure the projection is said to make tractable.
[§4] §4 (Renormalization group as invariant-tracking): while the RG is presented as a systematic implementation of the invariant-tracking criterion, the text does not demonstrate that the fixed-point analysis supplies an independent test of the vertical/horizontal distinction rather than presupposing it; the claim that RG 'implements' the criterion therefore risks restating the framework rather than providing external confirmation.

minor comments (2)

[§2] Notation for 'projection' is introduced without a compact formal definition (e.g., as a function P: StateSpace → EquivalenceClasses); adding one would clarify subsequent claims about equivalence-class partitioning.
[Introduction] The NASA Newtonian-mechanics example in the abstract is repeated in the introduction without additional detail on which invariants remain recoverable; a brief table contrasting recoverable versus constitutive omissions across two cases would improve readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful and constructive report. The comments identify places where the vertical/horizontal distinction and the RG discussion require sharper formulation. We have revised the manuscript to supply an explicit, non-circular criterion for the distinction and to clarify the independent role of fixed-point analysis. Our responses to the major comments follow.

read point-by-point responses

Referee: [Abstract and §3] Abstract and §3 (Vertical/Horizontal distinction): the partition into vertical cases (recoverable omissions, earlier projections survive as limits) and horizontal cases (constitutive omissions) is defined in terms of the invariants each projection reveals, yet no antecedent, non-circular criterion—such as purely syntactic limit behavior or independent equivalence-class test—is supplied for classifying a given omission as recoverable versus constitutive. This renders the classification dependent on the very invariant structure the projection is said to make tractable.

Authors: We agree that an antecedent, non-circular criterion strengthens the framework. In the revised §3 we now define the distinction via the existence of a limiting procedure: an omission counts as recoverable (vertical) precisely when there is a continuous scale parameter such that, as the parameter tends to the coarser value, the finer projection converges uniformly to the coarser one in a suitable state-space norm, with the omitted variation becoming arbitrarily small. This limit test is stated in purely syntactic terms before any invariants are identified. The revealed invariants are then read off from the resulting equivalence classes. The revised text includes a short formal statement of this criterion together with the Newtonian-to-relativistic example to illustrate its application. revision: yes
Referee: [§4] §4 (Renormalization group as invariant-tracking): while the RG is presented as a systematic implementation of the invariant-tracking criterion, the text does not demonstrate that the fixed-point analysis supplies an independent test of the vertical/horizontal distinction rather than presupposing it; the claim that RG 'implements' the criterion therefore risks restating the framework rather than providing external confirmation.

Authors: We accept that the original wording left the independence of the RG test implicit. The revised §4 now separates the steps: the beta-function flow and the spectrum of relevant operators are computed directly from the renormalization procedure and yield the scaling dimensions without prior reference to the vertical/horizontal label. Only after the fixed point is obtained do we classify the case according to the limit criterion introduced in §3. We illustrate this with the Ising-model fixed point, where the flow itself determines the critical exponents and thereby shows that the omission of microscopic details is constitutive (horizontal) rather than recoverable. A short paragraph has been added to make the logical order explicit. revision: partial

Circularity Check

1 steps flagged

Projection defined via invariants it reveals; vertical/horizontal distinction presupposes the invariants used to classify omissions as recoverable vs. constitutive

specific steps

self definitional [Abstract]
"The key concept is projection: a principled mapping from underlying complexity to a descriptive space that partitions states into equivalence classes, omits within-class variation, and makes patterns visible that would otherwise be lost. Projection is simultaneously revelatory and constitutive: it makes genuine invariants tractably accessible while bringing into being the concepts through which they become expressible. The paper distinguishes vertical cases, in which earlier projections survive as limiting cases of more refined successors with recoverable omission, from horizontal cases, in wh"

Projection is defined as the operation that renders invariants tractable; the vertical/horizontal taxonomy is then defined by whether those same invariants survive as recoverable limits or are constitutive only at the projection's own level. The criterion for classifying any given omission therefore presupposes the invariant structure that the projection is claimed to discover, so the distinction is equivalent to its own output by construction.

full rationale

The abstract supplies the core definitions and the load-bearing distinction. Projection is introduced as the mapping that 'makes genuine invariants tractably accessible' while the vertical/horizontal partition is drawn precisely by whether 'omissions [are] recoverable' or 'invariants are accessible only at the level of the projection that defines them.' No antecedent, non-circular criterion (syntactic, limit-behavioral, or otherwise independent of the invariants) is stated for deciding recoverability. Consequently the classification of projections reduces to the invariant structure the projections are said to reveal, rendering the central inversion (perspectival structure as condition for invariant detection) self-referential by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The paper rests on the domain assumption that every representational enterprise must omit variation, treats projection as a new organizing concept rather than a fitted parameter, and introduces no new physical entities.

axioms (1)

domain assumption Any representational enterprise must omit variation in order to function.
Stated in the opening sentence of the abstract as the starting point for the entire argument.

invented entities (1)

projection no independent evidence
purpose: A principled mapping from underlying complexity to a descriptive space that partitions states into equivalence classes and omits within-class variation.
Introduced as the key structural feature; no independent empirical test or falsifiable prediction is supplied in the abstract.

pith-pipeline@v0.9.0 · 5528 in / 1332 out tokens · 26804 ms · 2026-05-15T10:22:57.722749+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/AbsoluteFloorClosure.lean absolute_floor_iff_bare_distinguishability unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

A projection is a principled mapping from underlying complexity to a structured descriptive space that (1) partitions underlying states into equivalence classes, (2) omits variation within those classes, and (3) thereby makes certain patterns visible that would otherwise be lost.
IndisputableMonolith/Foundation/AlexanderDuality.lean alexander_duality_circle_linking unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

P₂ is vertically related to P₁ when there exists a specified limiting transformation under which the equivalence classes of P₁ are recoverable from those of P₂, and the invariants I₁ are recovered as limits of the invariants I₂.
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

The renormalization group reveals this by iterating the partitioning operation—applying successive coarse-grainings to the system and asking what remains invariant across them.

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.