Pausanias Analysis

Analysis outputs and model variants

Current Tri-Marked Sentence-Level Analyses

The current paper-facing analysis uses the active three-bucket tags, then restricts the model to mythic vs. historical sentences.

VocabularyBook ScopeRhetoric MarkersSampleFeaturesAccuracyPage
Lemma Excluding 4 and 8 Included 5,424 1,000 0.692 Open
Lemma Excluding 4 and 8 Removed 5,424 1,000 0.693 Open
Lemma All books Included 7,032 1,000 0.728 Open
Lemma All books Removed 7,032 1,000 0.716 Open
Surface Excluding 4 and 8 Included 5,424 1,000 0.712 Open
Surface Excluding 4 and 8 Removed 5,424 1,000 0.698 Open
Surface All books Included 7,032 1,000 0.723 Open
Surface All books Removed 7,032 1,000 0.726 Open

People Mentions

People and Gender Mentions

480 sections and 1,734 numbered sentences processed; 3,474 mention rows extracted.

Open People Report

Morphosyntactic Stylometry

LLM Grammar Stylometry

Passage-level morphosyntactic features from the gpt-5.4-mini grammar parser, with word-frequency and character n-gram baselines.

Open Stylometry

Interactive Projection

Feature-family projection display for comparing current parsed passages and mouseover neighbor context.

Open Projection

Sentence Classifiers

Cross-validated mythic, historical, and other sentence classifiers using grammar-derived stylometric features.

Open Classifiers

Book Regression

Ridge regressors that test whether stylometric features can predict sentence book number better than a mean-book baseline.

Open Regression

Feature Trends

Book-by-book proportions for the strongest morphosyntax regression features, plus sentence-length KDE and violin diagnostics.

Open Trends

Discourse Mode Control

1,200 tagged parsed sentences across 6 discourse modes.

Open Mode Control

Complementary Robustness Checks

Semantic-Field Ablation

Reruns the main lemma model after removing kinship, reporting, memorial, and military-political vocabularies.

Open Ablation Analysis

Book-Held-Out Robustness

Trains on all but one book, then tests whether the model generalises to the held-out book.

Open Held-Out Analysis

Model Error Analysis

Confident false mythic and false historical predictions for close-reading follow-up.

Open Error Analysis

Manual Label Sensitivity

Confusion matrix, label-scenario refits, and noise stress tests using the Greta/Rosie Book 3 manual labels.

Open Sensitivity Analysis

Translation Length

Unexpectedly Long or Short Translations

Greek predictors of length residuals, plus the dual view of English terms found in longer or shorter-than-expected passages.

Open Translation Length Analysis

Residual Length vs. Mythic/Historical Strength

Exploratory comparison of length-residual terms against the main mythic/historical classifier coefficients.

Open Diagnostic Page

Place and Network Analyses

Maps and Proper-Noun Networks

Geographic map, place-pair distances, and proper-noun co-occurrence network pages.

Open Place and Network Pages

Paper Network Analyses

Mythic/historical subgraphs, bridge nouns, book-level drift, and place-person/deity networks.

Open Network Analyses

Deprecated Predictor Pages

Passage Mythic Predictors

Open Page

Passage Skepticism Predictors

Open Page

Sentence Mythic Predictors

Open Page

Sentence Skepticism Predictors

Open Page