A Tutorial on Rigorous Scientific Inference: Bayesian Model Selection, the Zero-Patch Standard, and the Deductive Primacy of Axioms Illustrated by the Voynich Manuscript

1. The Methodological Principles

Scientific reasoning rests on a small set of explicit, falsifiable methodological principles [3]. These are not optional preferences; they are the minimal logical commitments required for inference to be coherent, probabilistically sound, and empirically testable. Every subsequent claim in this tutorial follows deductively from them.

Principle 1. Evidence Supremacy (Ontological Primacy of E)

The evidence vector E — the reproducible, quantitative invariants of the observations — holds absolute primacy. No hypothesis H is adequate unless the likelihood $P\left(E\right|H)$ is intrinsically high without post-observation adjustments. Falsifiable by: Existence of an alternative $H^{\prime }$ that achieves strictly lower surprise $S=-lnP\left(E\right|H^{\prime })$ with equal or fewer unconstrained parameters.

Principle 2. Deductive Primacy of Priors

Priors must be deduced bottom-up from the topology of E (structural invariants: entropy ratios, distributional shapes, clustering, repetition patterns) as the minimal model that maximizes intrinsic $P\left(E\right|H)$. Postulated priors (top-down assumptions imposed prior to examining E) are admissible only provisionally and must be rejected if they require patching to survive. Falsifiable by: A postulated H yielding higher surprise on held-out data $E^{\prime }$ than a deduced alternative. We separate exploratory deduction of candidate priors (based on low-level invariants) from confirmatory evaluation: priors and hyperparameters used for confirmatory scoring are fixed before evaluating on reserved folios or held-out structural tests.

Principle 3. Zero-Patch Standard (Rejection of Unconstrained Auxiliary Parameters)

A “patch” is defined as an auxiliary parameter $\theta$ introduced to a hypothesis Hafter observing contradictory evidence E, where $\theta$ lacks independent prior constraints. Such patching is inadmissible because the marginal likelihood penalizes unconstrained volume. Mathematically, a model class that permits unconstrained patching functions indistinguishably from a theory generator—a meta-model capable of fitting any noise pattern, rendering it unfalsifiable and scientifically void.

$$P\left(E\right|H)=\int P\left(E\right|H,\theta )P\left(\theta \right|H)d\theta \approx P\left(E\right|H,{\theta }_{\mathrm{best}})·{\displaystyle \frac{\delta \theta }{\Delta \theta }}$$

where $\Delta \theta$ is the prior range of the parameter. A large unconstrained $\Delta \theta$ creates a massive penalty (the Occam factor). Consequently, under standard model-selection criteria (e.g., BIC) heavily patched models tend to be disfavored relative to simpler, non-patched competitors.

By “zero-patch” we mean model classes where structure is fixed a priori or deterministically deduced from low-level invariants; models claiming ad-hoc post-hoc flexibility will incur substantial model-complexity penalties unless those degrees of freedom are independently constrained. Falsifiable by: Direct computation of $\Delta S$ or $\Delta \mathrm{BIC}$ showing that a patched model outperforms a simpler structural model on unseen data.

Principle 4. Honest Invalidation Requirement

Every hypothesis H must be accompanied by pre-specified operational tests (falsification gates) that predict quantitative outcomes on E or future $E^{\prime }$. Failure of a test requires rejection of H, not modification of H or the test. Falsifiable by: Systematic salvage via patching across replications or meta-analyses.

Principle 5. Parsimony as Predictive Power

Among models with comparable $P\left(E\right|H)$, the one with the fewest unconstrained degrees of freedom has higher expected predictive accuracy on future data $E^{\prime }$. This follows from the free-energy principle and Occam’s razor formalized in information theory [4,5]. Falsifiable by: Cross-validation showing simpler models generalize better.

These five principles constitute a closed, self-consistent foundation.

2. The Patching Fallacy: Mathematical Anatomy

Why is patching mathematically inadmissible? Consider a base hypothesis H that yields low likelihood $P\left(E\right|H)=ϵ\ll 1$. To force a fit, a researcher introduces an auxiliary patch $\theta$ with prior range $\Delta \theta$ (e.g., “The scribe used an arbitrary set of 500 abbreviations”) with prior range Δθ.

The marginal likelihood is the integral over the parameter space:

$$P\left(E\right|H)=\int P\left(E\right|H,\theta )P\left(\theta \right|H)d\theta \approx P\left(E\right|H,{\theta }_{\mathrm{best}})·{\displaystyle \frac{\delta \theta }{\Delta \theta }}$$

where $\delta \theta$ is the width of the peak where the fit is good. If the patch $\theta$ is unconstrained ($\Delta \theta$ is large), the ratio ${\displaystyle \frac{\delta \theta }{\Delta \theta }}$ becomes vanishingly small. This acts as an “Occam Factor” or penalty term [4].

This penalty is not a stylistic preference; it is a direct consequence of probability theory. Every unconstrained parameter introduced to save a model exponentially dilutes its predictive power. In terms of the Bayesian Information Criterion (BIC), each parameter contributes a cost of ln n [6].

We reinforce this with two foundational results from computability theory:

• Predictive Capacity (Vapnik): A model with free parameters (patches) possesses a high Vapnik-Chervonenkis (VC) dimension [7], allowing it to “shatter” (fit) diverse datasets including noise. In contrast, a model with VC dimension zero is structurally rigid: its hypothesis class contains only a single function, forcing exact entailment or falsification.
• Algorithmic Probability (Solomonoff): The universal prior probability of a hypothesis is 2^−L, where L is the length of the shortest program that generates the data [8]. Every patch adds code complexity ΔL, exponentially decreasing the model’s intrinsic probability.

Thus, a fundamental ontology must minimize both VC dimension and algorithmic complexity to remain admissible.

3. Case Study: The Voynich Manuscript

[The Evidence Vector]

The Voynich Manuscript (VMS) serves as an ideal high-dimensional test case because standard hypotheses have been maintained for decades solely through extensive patching. We apply the principled framework to evaluate competing explanations for the anomaly.

The Voynich Manuscript is dated to the early 15th century on vellum—an expensive material requiring the hides of approximately 30 calves and months of preparation—indicating substantial investment and coordinated effort by its creator(s). Across six centuries, no credible authorship claim, commercial exploitation, decoded solution, or coherent motive for elaborate hoaxing has emerged [9], which together constrain the prior toward genuine but misunderstood encoding over deliberate deception.

3.1. The Evidence Vector (E)

We do not re-estimate Voynich text statistics in this work. Instead, we treat a small set of repeatedly reported, transcription-insensitive corpus-level regularities as an empirical evidence vector E. Concretely, E is assembled from published measurements and long-established observations based on standard EVA-family transcriptions and community tokenization conventions [10], beginning with Currier’s foundational characterization of systematic heterogeneity [11] and Tiltman’s early analysis [12]. Our contribution is to show that, conditional on these published constraints, many commonly assumed priors over plausible generative mechanisms become inconsistent, and that more appropriate priors can be deduced from E.

Use of literature values (no new corpus measurement): Each component below is taken as reported in the cited sources, including their stated transcription snapshot, preprocessing, and uncertainty estimation (when provided). Where multiple sources report compatible values, we treat the resulting range as defining a robust constraint.

1.

$E_{hapax}$ (Extreme Uniqueness): Prior quantitative analyses report that approximately $58.2\%\pm 0.8\%$ of word types are singletons (hapax legomena) [13,14]. This non-linguistic profile is distinct from natural language Zipfian curves, as confirmed by recent quantitative audits [15].

2.

$E_{morph}$ (Rigid Morphology): Published morphological characterizations report that roughly $98\%\pm 0.5\%$ of tokens conform to a constrained Prefix–Root–Suffix (or comparable) template. This structure, famously identified as “Crust-Mantle-Core” by Stolfi [?], persists across standard EVA-style segmentation [16].

3.

$E_{sect}$ (Sectional Disjointness): Reported vocabulary overlap between major manuscript sections is low; typical summaries give an average Jaccard overlap $J<0.3$, consistent with Currier’s observation of distinct “languages” [11] and recent clustering analyses [17,18].

4.

$E_{delim}$ (Positional Rigidity): Multiple analyses report glyph(s) with near-zero positional variance (e.g., restricted to line-initial position) [19], indicating strong layout-conditioned constraints incompatible with fluid orthography.

5.

$E_{label}$ (Contextual Compression): Published comparisons of illustration labels vs. running text report systematic reduction: labels preferentially use a strict subset of the morphological components found in body text [?].

4. Deducing the Prior

4.1. Rejection of Postulated (Patched) Models

Under Principle 3, we evaluate standard hypotheses:

• Natural Language ($H_{lang}$): Under the Zero-Patch Standard and given $E_{hapax}$ and related constraints, standard natural-language generative priors require substantial unconstrained auxiliary assumptions (e.g., extreme abbreviation, polyglot switching) to achieve adequate fit [20]. Assessment: Strongly disfavored under Zero-Patch assumptions.
• Cipher ($H_{cipher}$): Under the Zero-Patch Standard and given $E_{morph}$, simple substitution or straightforward cipher models cannot account for the observed morphology and spectral periodicity [21] without additional unconstrained mechanisms. Assessment: Strongly disfavored under Zero-Patch assumptions.

4.2. The Deduced Prior: Structured Reference System ($H_{ref}$)

Applying Principle 2, we deduce the model directly from the topology of E. What class of information system naturally exhibits rigid morphology, high uniqueness, and sectional segregation? Real-world analogs, such as medieval inventories, relational databases, or indices, exhibit similar statistical fingerprints: high hapax rates in unique identifiers, repetitive metadata, and partitioned vocabularies.

$$H_{ref}:\mathrm{The}\mathrm{text}\mathrm{is}\mathrm{a}\mathrm{Structured}\mathrm{Reference}(\mathrm{Database}/\mathrm{Inventory}).$$

This model naturally accounts for the evidence vector under the Zero-Patch constraint, without introducing ad-hoc auxiliary parameters. We therefore present it as a leading hypothesis rather than a definitive historical solution:

1. Prediction of $E_{morph}$ and $E_{hapax}$ (The Primary Key): A reference system consists of distinct entities. Each entity requires a unique identifier (The Root). Metadata (Prefix/Suffix) is repetitive. A list of 1,000 distinct recipes requires 1,000 unique keys. The high Hapax rate is not an anomaly; it is a requirement of a database.

2. Prediction of $E_{sect}$ (Thematic Partitioning): A database of “Herbs” uses different unique keys than a database of “Stars.” Sectional disjointness is intrinsic.

3. Prediction of Local Repetition (Relational Pointers): In a Relational system (e.g., ingredients in recipes), the Root acts as a re-entrant pointer. If “Root A” = “Basil”, it must repeat whenever Basil is referenced. This explains local repetition without invoking narrative grammar.

4. Prediction of $E_{delim}$ (Serialization): The glyphs restricted to line-starts are not phonetic. They function as Record Delimiters or Item Markers in the data stream. Their zero variance is a feature of syntax, not orthography.

5. Prediction of $E_{label}$ (Lossless Compression): In the body text, a token requires full metadata: [Class] + [ID] + [State]. On an illustration, the visual context provides the [Class]. The label therefore undergoes lossless compression, stripping the Prefix to display only the [ID]. This matches the observed brevity of labels.

6. Prediction regarding “Sorting”: The text is not sorted alphabetically. This implies it is sorted by a different column in the database: Semantic Sorting. The adjacency of entries implies semantic proximity, not lexical proximity.

5. Formal Specification of the Deduced Model ($H_{ref}$)

To move beyond qualitative debate, we provide a formal specification of the deduced prior ($H_{ref}$). This moves the hypothesis from a narrative claim to a testable mathematical object.

5.1. Distinction Between Calibration and Patching

It is crucial to distinguish between Parameter Estimation (Calibration) and Patching.

• Calibration: Determining the value of a constant required by the deduced structure (e.g., measuring G in $F=G{\displaystyle \frac{m_1{m}_2}{r^2}}$). This fixes the specific realization of the model but does not alter its complexity class.
• Patching: Introducing new structural terms or auxiliary rules to force a fit (e.g., adding $+k·r^3$ to the gravity equation because the data deviates). This increases model complexity to absorb error.

The specification below allows for calibration of distributions (the “constants” of the system) derived from the evidence vector E, but strictly forbids the addition of post-hoc structural patches.

5.2. Model Components and Topology

Let the manuscript glyph inventory be the finite set $\mathcal{G}$. We posit a deterministic partitioning rule (segmentation) $\varphi :{\mathcal{G}}^{*}\to \{P,R,S,D\}$ that maps glyph sequences to four disjoint functional classes:

• P: Prefix sequences (Metadata/Classifiers).
• R: Root sequences (Primary Identifiers/Keys).
• S: Suffix sequences (Status/State markers).
• D: Delimiters (Record separators).

This partitioning is not arbitrary; it relies on the consensus statistics of E (e.g., EVA transcription data), where R corresponds to the high-entropy, high-hapax core of the token, and $P/S$ correspond to the low-entropy, repetitive periphery.

5.3. The Generative Template

A Token T is strictly defined as the concatenation:

$$T=\pi ·r·\sigma$$

where $\pi \in P^{*}$ (optional prefix), $r\in R$ (mandatory root), and $\sigma \in S^{*}$ (optional suffix).

A Section is defined by a partition of the root space R. While natural systems exhibit noise, $H_{ref}$ predicts that the root vocabulary is partitioned into semantic domains $R_1,R_2,\dots ,R_m$ such that the overlap between sections is indistinguishable from noise:

$$J(R_i,R_j)<ϵ\mathrm{for}i\ne j$$

where J is the Jaccard index and $ϵ$ represents the noise floor of the system (e.g., misplaced folios or generic “stop-word” roots).

5.4. Likelihood and Zero-Patch Constraint

The likelihood of the corpus $\mathbf{W}$ under $H_{ref}$ is:

$$logP\left(\mathbf{W}\right|H_{ref})=\sum _{t\in \mathbf{W}}logp\left(t\right|\mathrm{Sec}\mathrm{tion})$$

where $p\left(t\right)$ is determined by the calibrated frequency tables of $P,R,S$.

Rigidity as the Defense Against Patching: It is imperative to state that the functional form of the template is fixed. The template $T=\pi ·r·\sigma$ is not a flexible schema; it is a rigid constraint analogous to a physical law. While we calibrate the distributions $P\left(\pi \right),P\left(r\right),P\left(\sigma \right)$ (akin to determining the gravitational constant G), the structure itself admits zero degrees of freedom. Unlike standard linguistic models which introduce “nulls,” “abbreviated forms,” or “polyglot switches” whenever a token deviates from expected grammar, $H_{ref}$ forbids the addition of structural terms. Any deviation contributes directly to the error term. The defense against patching is the rigidity of the template: if the data requires a fourth slot or a permuted order to achieve high likelihood, $H_{ref}$ is falsified, not patched.

5.5. Specific Falsifiers for $H_{ref}$ (The Kill List)

While Appendix A provides general tests for any prior, we provide here the specific empirical outcomes that would immediately invalidate $H_{ref}$. This is the explicit criteria for falsifying our deduction:

1.

Low Root Uniqueness: If the set R (Roots) is found to follow a standard natural-language Zipfian curve (small core vocabulary) rather than the reported high-hapax profile ($\sim 50-60\%$ uniqueness), the “Primary Key” interpretation collapses.

2.

High Intra-Token Entropy: If Mutual Information $I(P;R)$ is high (implying grammatical agreement or vowel harmony), the orthogonality of ID vs. Metadata is disproven.

3.

Delimiter Mobility: If glyphs identified as D are shown to have high positional variance (scattering randomly), the record-structure hypothesis fails.

4.

Significant Sectional Overlap: If $J(R_i,R_j)\gg ϵ$, the thematic partitioning hypothesis fails.

5.

Failure of Label Compression: If labels do not show systematic stripping of P (Prefixes) relative to the body text, the “Contextual Compression” prediction fails.

6. Conclusion

The Voynich Manuscript illustrates a broader methodological lesson: interpretive impasses are often artifacts of incorrect priors maintained through patching. However, this framework does not invalidate the detailed statistical work of the past century; rather, it provides the analytical context to integrate it.

Viewed mathematically, the history of VMS research resembles a Taylor expansion: an attempt to approximate a complex topology through an accumulating series of local adjustments (patches). Researchers correctly identified "null words" to explain entropy dips, "fixed slots" to explain positional rigidity, and "micro-dialects" to explain sectional disjointness. These were not errors; they were accurate descriptive terms in an approximating series.

Our contribution is to propose a concrete generative function as a parsimonious explanation. By deducing the prior from the evidence, we show how many previously ad-hoc patches map naturally onto a single structural class (the proposed $H_{ref}$). The features that appear anomalous under a linguistic interpretation can be reinterpreted as expected consequences of a reference-system architecture. We present $H_{ref}$ as the leading hypothesis consistent with available corpus invariants; it remains open to revision or replacement should new evidence or independently constrained alternative models emerge.