O0-MATH-009

Observer-Dependence: Formal Framework

**Version:** 1.0

**Research status:** ACTIVE

CLAIM STATUS: EXPLORATORY FORMALIZATION
EVIDENCE TYPE: MATHEMATICAL FORMALIZATION
PHYSICAL VALIDATION: NONE
INDEPENDENT REPLICATION: N/A
PHILOSOPHICAL PROVENANCE: O/0 ARCHIVE
ARCHIVE ENDORSEMENT: LIMITED TO REPORTED RESULT

Abstract

This document formalizes the concept of observer-dependence of boundary attribution in the O/0 substrate. We define observer frames rigorously, establish when different frames attribute different boundaries to the same substrate configuration, and analyze whether this observer-dependence is trivial (mere coordinate relabeling) or substantive (different frames yield genuinely incompatible boundary structures). We draw a careful analogy to gauge theory in physics — emphasizing that it is an analogy, not an identity — and propose criteria for distinguishing gauge-like redundancy from genuine observer-dependence.

Source proposition

The O/0 archive suggests that the identification of boundaries and agents is perspective-dependent: what counts as "self" and "other" depends on the observer's frame. This document asks whether this claim has formal content or reduces to a trivial statement about coordinate systems.

Scientific audit

Observer-dependence in physics is well-formalized through gauge theory (Yang-Mills), general covariance (general relativity), and reference frame theory (quantum reference frames). The mathematical tools are established. Applying them to substrate boundaries is novel and speculative. We use the formalism rigorously but clearly mark where the analogy to physics breaks down.

Research question

Is the observer-dependence of boundary attribution in the O/0 substrate a trivial consequence of coordinate choice, or does it reflect a deeper structural feature? Can different observers disagree about whether a boundary exists, or only about where it is?

Operational definitions

1. **Observer frame**: A specification of (a) which cells are designated as "interior" for boundary detection, (b) the coarse-graining level used for mutual information estimation, and (c) the threshold epsilon for the Markov blanket criterion.

2. **Boundary attribution**: The map from (substrate state, observer frame) to a set of boundaries.

3. **Gauge transformation**: A change of observer frame that does not change the set of boundaries (relabeling without physical consequence).

4. **Substantive transformation**: A change of observer frame that changes the set of boundaries (different observers see different boundaries).

5. **Frame-invariant quantity**: A property of the substrate that is the same in all observer frames.

Hypothesis

Observer-dependence of boundary attribution has both a trivial component (gauge-like coordinate relabeling) and a substantive component (different coarse-graining levels reveal different boundary structures at different scales). The number of boundaries at a given scale is frame-invariant; the assignment of "interior" vs. "exterior" labels is frame-dependent.

Null hypothesis

All observer-dependence is trivial: boundary attribution is frame-invariant up to relabeling. Any apparent disagreement between observers reduces to notational differences.

Competing explanations

1. Observer-dependence may be entirely a function of coarse-graining scale, not frame per se.

2. The frame concept may be ill-defined for the O/0 substrate (no natural notion of "observer position").

3. The gauge analogy may be misleading, importing structure from physics that has no substrate analogue.

Formal model

**Definition 1 (Observer Frame).** An *observer frame* is a triple O = (A, sigma, epsilon) where:

• A subset {1, ..., N} is the *attention set* — the cells the observer designates as the candidate interior.
• sigma > 0 is the *coarse-graining scale* — the spatial resolution at which states are measured (cells within distance sigma are averaged).
• epsilon > 0 is the *blanket threshold* — the maximum conditional mutual information permitted for a valid blanket.

The space of observer frames is F = P({1,...,N}) x R_+ x R_+, where P denotes the power set.

**Definition 2 (Frame-Dependent State).** Given a substrate state S = (S_1, ..., S_N) and an observer frame O = (A, sigma, epsilon), the *frame-dependent state* is:

S^O = (S_A^sigma, S_{A^c}^sigma)

where S_A^sigma is the coarse-grained state of the attention set:

S_i^sigma = (1 / |B_sigma(i)|) * sum_{j in B_sigma(i)} S_j

and B_sigma(i) = {j : dist(i, j) < sigma} is the ball of radius sigma around cell i.

**Definition 3 (Frame-Dependent Boundary).** Given frame O, the *frame-dependent boundary* is:

MB^O = argmin_{B subset A^c, |B| <= K} I(S_A^sigma; S_{(A union B)^c}^sigma | S_B^sigma)

subject to the constraint that the residual conditional MI is below epsilon.

**Theorem 1 (Relabeling Gauge Invariance).** Let pi: {1,...,N} -> {1,...,N} be a permutation of cells that preserves the neighborhood structure (graph automorphism). Then for any frame O = (A, sigma, epsilon):

MB^{pi(O)} = pi(MB^O)

where pi(O) = (pi(A), sigma, epsilon). That is, relabeling cells by a graph automorphism permutes the boundary set correspondingly.

*Proof.* The conditional mutual information is a function of the joint distribution, which is invariant under permutations that preserve the graph structure. Therefore I(S_{pi(A)}^sigma; ... | S_{pi(B)}^sigma) = I(S_A^sigma; ... | S_B^sigma), and the argmin is permuted accordingly. QED

*Interpretation.* Relabeling is a gauge transformation: it changes the description but not the boundary structure. This is the trivial component of observer-dependence.

**Theorem 2 (Scale-Dependence is Substantive).** There exist substrate states S and frames O_1 = (A, sigma_1, epsilon) and O_2 = (A, sigma_2, epsilon) with sigma_1 != sigma_2 such that:

|MB^{O_1}| != |MB^{O_2}|

That is, different coarse-graining scales can detect different numbers of boundaries.

*Proof (by construction).* Consider a substrate with a nested boundary structure: a small agent (radius r_small) inside a larger agent (radius r_large). At coarse-graining scale sigma < r_small, both boundaries are resolved, and |MB| = 2 (or more precisely, two connected components). At coarse-graining scale sigma with r_small < sigma < r_large, the small agent is averaged out (its internal structure is below resolution), and |MB| = 1. At sigma > r_large, both boundaries are averaged out, and |MB| = 0. QED

**Corollary 1 (Multi-Scale Boundary Hierarchy).** The map sigma -> |MB^{(A, sigma, epsilon)}| is a non-increasing step function (more boundaries at finer scales, fewer at coarser scales). The scale values at which |MB| decreases are the *critical scales* of the boundary hierarchy.

**Definition 4 (Gauge-Theoretic Analogy).** In gauge theory, a gauge field A_mu transforms under local gauge transformations g(x) as:

A_mu -> g * A_mu * g^{-1} + g * partial_mu g^{-1}

The physical content is in gauge-invariant quantities (Wilson loops, field strengths).

By analogy, define the *boundary gauge group* G as the group of observer-frame transformations that do not change the set of boundaries:

G = {T: F -> F : MB^{T(O)} = MB^O for all substrate states S}

By Theorem 1, G includes all graph automorphisms (relabeling).

**Proposition 1 (Gauge-Invariant Boundary Quantities).** The following quantities are invariant under all elements of G:

• The number of connected boundary components n_c at each scale sigma.
• The nesting depth d_max.
• The total boundary perimeter P.
• The integrated information Phi of the interior.

These are the "physical" (frame-invariant) properties of the boundary structure.

**ANALOGY CAVEAT.** The gauge-theoretic analogy is structural, not physical. Substrate boundaries are not gauge fields. Specifically:

• Gauge transformations in physics are continuous Lie group actions; G is a discrete group (graph automorphisms plus scale transformations).
• Gauge fields carry dynamics (Yang-Mills equations); the boundary gauge group acts on static configurations.
• The physical content of gauge theory includes the topology of the gauge bundle (instantons, monopoles); no such topological structure is known for the boundary gauge group.

The analogy is useful for organizing which properties of boundaries are "real" (frame-invariant) versus "descriptive" (frame-dependent), but it should not be pushed further without substantial new mathematical development.

**Definition 5 (Attention-Dependence).** Define the *attention-dependence index* for a boundary B as:

delta_A(B) = max_{A_1, A_2} |I(S_{A_1}^sigma; S_{A_1^c}^sigma | S_{MB_1}^sigma) - I(S_{A_2}^sigma; S_{A_2^c}^sigma | S_{MB_2}^sigma)|

where the maximum is over all attention sets A_1, A_2 that produce the same boundary B. If delta_A = 0, the boundary quality is attention-independent. If delta_A > 0, different attention sets yield different quality assessments of the same boundary.

**Proposition 2 (Attention-Dependence is Generically Nonzero).** For generic substrate states (outside a measure-zero set), delta_A > 0 for at least some boundaries. This is because the conditional MI depends on which cells are designated "interior," and different interiors produce different conditional distributions.

*Proof sketch.* The conditional MI I(S_A; S_{A^c} | S_B) depends on the choice of A through the marginal distribution of S_A. For generic joint distributions (full-rank covariance matrices), changing A changes the marginal and hence the conditional MI. QED

**Open Question 1.** Is there a natural "preferred frame" for boundary attribution? In physics, inertial frames are preferred by Lorentz invariance. Is there an analogous principle for the substrate?

**Open Question 2.** Can the scale-dependent boundary hierarchy be organized into a sheaf on the space of scales, allowing boundaries at different scales to be "glued together" coherently?

**Open Question 3.** Does the boundary gauge group G have interesting subgroups corresponding to different types of observer-dependence?

Methods

1. Compute boundaries under multiple frames O_k = (A_k, sigma_k, epsilon) for the same substrate state.

2. Classify frame transformations as gauge (same boundaries) or substantive (different boundaries).

3. Compute the attention-dependence index delta_A for all detected boundaries.

4. Map the critical scales of the boundary hierarchy.

5. Test Proposition 1 by verifying that the listed quantities are frame-invariant.

Controls

1. Test on symmetric substrates where all frames related by symmetry should produce identical boundaries.

2. Test with known scale-separated structures to verify Theorem 2.

3. Compare with a random substrate (no structure) to verify that frame-invariant quantities are trivially zero.

Predictions

1. Relabeling transformations will always preserve boundary structure (Theorem 1).

2. Scale transformations will reveal a multi-scale hierarchy with at least two critical scales in substrates with nested boundaries.

3. Attention-dependence will be nonzero for most boundaries, but small relative to the boundary quality.

4. Frame-invariant quantities (n_c, d_max, P, Phi) will be robust across all tested frames.

Falsification criteria

1. If relabeling by a graph automorphism changes the boundary structure, Theorem 1 has an error.

2. If scale-dependence does not produce different boundary counts, Theorem 2 is not applicable to the tested substrate.

3. If supposedly frame-invariant quantities change under gauge transformations, Proposition 1 is wrong.

Results / Expected Outcomes

No simulation results reported. The formal framework defines the concepts needed for empirical investigation of observer-dependence. The gauge analogy provides organizational structure but does not generate predictions beyond those derived from the information-theoretic definitions.

Uncertainty

1. The definition of "observer frame" is a modeling choice; other definitions might yield different conclusions about observer-dependence.

2. The gauge analogy is informal; it may suggest structure that does not exist.

3. The attention-dependence index delta_A depends on which attention sets are tested; exhaustive testing is infeasible for large N.

Limitations

1. The framework treats observers as external to the substrate; a fully self-consistent treatment would require self-referential observers (see O0-MATH-006), introducing circularity.

2. The gauge group G is defined implicitly (as the group of frame changes preserving boundaries); its explicit computation may be intractable.

3. The framework does not address temporal observer-dependence (different observers sampling at different times).

4. The analogy to gauge theory is carefully bounded but may still mislead readers into importing physical intuitions that do not apply.

Replication status

Mathematical definitions and theorems are self-contained and independently verifiable.

Data and code

No data or code generated. Frame-dependent boundary computation to be implemented as an extension of the O0-SIM boundary detection code.

Relationship to philosophical archive

The archive's assertion that "there is no privileged observer" resonates with the gauge-invariance framework: "privileged" properties are those that are frame-invariant. However, the mathematical analysis reveals that observer-dependence has both trivial (gauge) and substantive (scale) components, a distinction the archive does not make. The formalization adds precision to the philosophical claim without endorsing it.

References

1. Nakahara, M. (2003). *Geometry, Topology, and Physics* (2nd ed.). CRC Press.

2. Rovelli, C. (1996). "Relational quantum mechanics." *International Journal of Theoretical Physics*, 35, 1637–1678.

3. Giacomini, F., Castro-Ruiz, E., & Brukner, Č. (2019). "Quantum mechanics and the covariance of physical laws in quantum reference frames." *Nature Communications*, 10, 494.

4. Friston, K. J. (2019). "A free energy principle for a particular physics." arXiv:1906.10184.

5. Bar-Natan, D. (1995). "On the Vassiliev knot invariants." *Topology*, 34(2), 423–472.

Revision history

• v1.0: Initial formalization. Observer frames defined. Gauge invariance and scale-dependence established. Gauge-theoretic analogy stated with explicit caveats. Three open questions raised.