Corpus & Data

The Plutarch Observatory corpus is an annotated resource for studying metaphors of mind and cognition across Plutarch’s extant works. It links corpus-scale patterns to individual Greek passages, so researchers can move between distant reading, close reading, and historically sensitive interpretation.

Corpus scope

112 works by Plutarch

  • Covers both the Moralia and the Parallel Lives.
  • Includes 1,469 paragraph-level segments with metaphorical content.
  • Contains 1,787 annotated metaphorical mappings related to mind and cognition.

Distribution

Moralia and Lives

  • Moralia: 1,455 mappings across 1,164 annotated segments.
  • Lives: 332 mappings across 305 annotated segments.
  • Average density: 1.22 metaphor instances per annotated segment.

Annotation inventory

Scenes and subframes

  • 609 distinct source scenes and 381 distinct target scenes.
  • 1,174 source subframes and 382 target subframes.
  • 2,345 source lexical units and 985 target lexical units.

Source corpus and selection

Research corpus

Moralia and Lives together

The project was designed around Plutarch’s Moralia and Parallel Lives, rather than treating these bodies of work separately. This allows metaphorical patterns of cognition to be compared across ethical, biographical, political, philosophical, and literary contexts.

Authenticity

Working boundaries

The grant description defines the core source material as the generally authentic writings of the Moralia together with the extant paired and unpaired Lives. Spurious and fragmentary works are not treated as the immediate basis for corpus construction.

Current dataset

Annotated records

The numbers on this page describe the current Observatory dataset: the passages, mappings, scenes, subframes, and lexical units that have been coded and standardized for exploration. They should not be confused with a claim that every line of the project source corpus contains a coded metaphor.

Annotation model

Level 1

Scenes

Scenes capture broad source-target mappings, such as CONTAINER → MIND or DISEASE → BAD_STATE_OF_THE_MIND. They make it possible to compare recurring metaphor families across works.

Level 2

Subframes

Subframes preserve the fine-grained structure of a metaphor in context. For example, a governance scene may involve source-side elements such as ruler, laws, obedience, conflict, or civic disorder.

Unit of analysis

Paragraph-level segments

The corpus uses paragraph-level segments anchored in Plutarch’s standard reference system. This gives enough context for extended and layered metaphors without detaching them from their textual setting.

Annotation workflow

Phase 1

Framework and coding fields

The team refined the metaphor framework on selected passages and translated theoretical questions into concrete annotation fields.

Phases 2-3

Iterative validation

Pilot annotation used two roles: “thinkers”, who developed conceptual interpretations, and “coders”, who translated those interpretations into structured records. Disagreements helped refine criteria.

Phase 4

Large-scale annotation

Coders annotated the remaining corpus, segmented passages, recorded source and target scenes, and documented lexical evidence.

Phases 5-6

Standardization and analysis

The team standardized scene and subframe labels, organized lexical units, checked consistency, and prepared the data for exploratory analysis.

What each record makes inspectable

Passage context

From pattern to Greek text

Each coded fragment is linked to its work and traditional paragraph numbering, allowing researchers to locate passages in standard editions and interpret them in context.

Metaphor data

Mappings and lexical evidence

Records expose source and target scenes, hierarchical scene paths, metaphor type, source and target lexical units, lemmata, and scene-internal subframes.

Scholarly transparency

Notes and provenance

Coder notes and provenance remain visible so users can assess annotation decisions in context instead of relying only on abstract labels or aggregate counts.

Quantitative overview

Frequent targets

Mind, reason, and mental states

The most frequent target scenes are MIND (164), REASON (120), BAD_STATE_OF_THE_MIND (58), LEARNING (56), and KNOWLEDGE (51).

Frequent sources

Container, disease, and formation

The most frequent source scenes are CONTAINER (97), DISEASE (70), MOULDING (56), GOVERNANCE (51), and LIGHT (50).

Co-occurrence

Recurring pairings

Common source-target pairings include CONTAINER → MIND, DISEASE → BAD_STATE_OF_THE_MIND, and MOULDING → EDUCATION.

Exploratory network analysis

The corpus can also be studied as a network in which scenes are nodes and annotated mappings are weighted edges. This makes it possible to identify conceptual hubs and communities of related metaphors. Initial analysis identifies six major communities, with MIND, REASON, CONTAINER, DISEASE, GOVERNANCE, and MOULDING functioning as important organizing nodes.

Access and reuse

Works

Work-first navigation

The Works module supports navigation by Plutarchan work, with work-level metadata, fragment counts, and stable links to specific fragments and metaphor records.

Metaphor Explorer

Scenes, filters, and passages

The Explorer lets users filter by source and target scenes, inspect distributions as tables and plots, and move directly from a pattern to its passages.

Lexical and catalogue search

Forms, lemmata, and paths

Search supports coded lexical units, lemmata, subframes, metaphor type, scene paths, and glob-style patterns. Lemmatization and source evidence remain visible for philological checking.

The dataset is also intended for machine-readable reuse. Quantitative views should be read as exploratory summaries: they improve transparency and help detect patterns, but interpretation still requires returning to the Greek passages and the annotation notes.