OT Discourse Structure Analysis¶

Narrative peak and episode boundary detection in the Hebrew Old Testament, using the Longacre (1983) framework.

Terminology:

• PEAK — the climactic paragraph/episode, marked by dense wayyiqtol + speech
• EPISODE — a coherent narrative unit, separated by disjunctive/scene-setting clauses
• BACKBONE — the wayyiqtol chain that drives the main narrative forward

Peak score formula:

peak_score = 0.4 × wayyiqtol_density + 0.3 × speech_density + 0.3 × TTR
             (all components normalised 0–1 within the book)

Questions this notebook answers:

• Where is the narrative peak of Genesis 22 (the Akedah)?
• How many episode boundaries are in Exodus 1–15?
• Which chapters of 1 Samuel have the highest wayyiqtol density?
• Where does direct speech cluster in Ruth?

Source: R. E. Longacre, The Grammar of Discourse (1983); MACULA Hebrew WLC dataset.

In [ ]:

import sys
sys.path.insert(0, '../../../src')

from bible_grammar import (
    ot_discourse_wayyiqtol_density, ot_discourse_speech_density,
    ot_discourse_lexical_diversity, ot_discourse_peak_score,
    ot_discourse_episode_boundaries, ot_discourse_narrative_profile,
    print_ot_discourse_overview, print_ot_wayyiqtol_density,
    print_ot_speech_density, print_ot_peak_score, print_ot_episode_boundaries,
    ot_discourse_density_chart, ot_discourse_peak_chart,
)
import pandas as pd

import sys sys.path.insert(0, '../../../src') from bible_grammar import ( ot_discourse_wayyiqtol_density, ot_discourse_speech_density, ot_discourse_lexical_diversity, ot_discourse_peak_score, ot_discourse_episode_boundaries, ot_discourse_narrative_profile, print_ot_discourse_overview, print_ot_wayyiqtol_density, print_ot_speech_density, print_ot_peak_score, print_ot_episode_boundaries, ot_discourse_density_chart, ot_discourse_peak_chart, ) import pandas as pd

1. Overview — Narrative Profile¶

In [ ]:

# High-level summary for Genesis
print_ot_discourse_overview('Gen')

# High-level summary for Genesis print_ot_discourse_overview('Gen')

In [ ]:

# Compare profiles across narrative books
for book in ['Gen', 'Exo', 'Rut', '1Sa', '2Sa', 'Jon']:
    p = ot_discourse_narrative_profile(book)
    print(f"{book:<5}  tokens={p['total_tokens']:>6,}  "
          f"wyy%={p['wayyiqtol_pct']:>5.1f}%  "
          f"peak_ch={p['peak_chapter']}  score={p['peak_score']}")

# Compare profiles across narrative books for book in ['Gen', 'Exo', 'Rut', '1Sa', '2Sa', 'Jon']: p = ot_discourse_narrative_profile(book) print(f"{book:<5} tokens={p['total_tokens']:>6,} " f"wyy%={p['wayyiqtol_pct']:>5.1f}% " f"peak_ch={p['peak_chapter']} score={p['peak_score']}")

2. Wayyiqtol Density — The Narrative Backbone¶

Wayyiqtol (ויקטל) is the primary narrative tense in BH, forming the backbone of sequential narrative. High density indicates fast-moving action sequences; gaps often mark episode transitions.

In [ ]:

print_ot_wayyiqtol_density('Gen')

print_ot_wayyiqtol_density('Gen')

In [ ]:

# 1 Samuel — where does David's story intensify?
print_ot_wayyiqtol_density('1Sa')

# 1 Samuel — where does David's story intensify? print_ot_wayyiqtol_density('1Sa')

In [ ]:

# Chart: wayyiqtol vs. speech density for Ruth (short book, easy to see)
ot_discourse_density_chart('Rut')

# Chart: wayyiqtol vs. speech density for Ruth (short book, easy to see) ot_discourse_density_chart('Rut')

In [ ]:

# Chart for Genesis
ot_discourse_density_chart('Gen')

# Chart for Genesis ot_discourse_density_chart('Gen')

3. Speech Verb Density¶

Density of speech-reporting verbs (אָמַר, דָּבַר, נָאַם, קָרָא) as a proxy for direct speech. Longacre identifies elevated direct speech as a key narrative peak marker.

In [ ]:

print_ot_speech_density('Gen')

print_ot_speech_density('Gen')

In [ ]:

# Ruth — a dialogue-heavy book
print_ot_speech_density('Rut')

# Ruth — a dialogue-heavy book print_ot_speech_density('Rut')

In [ ]:

# Jonah — four chapters, one prophet, lots of divine speech
print_ot_speech_density('Jon')

# Jonah — four chapters, one prophet, lots of divine speech print_ot_speech_density('Jon')

4. Narrative Peak Score¶

Composite score per chapter: 0.4 × wayyiqtol + 0.3 × speech + 0.3 × TTR (all normalised 0–1 within the book). The crimson bar in the chart marks the computed narrative peak.

In [ ]:

print_ot_peak_score('Gen')

print_ot_peak_score('Gen')

In [ ]:

# Chart — Genesis
ot_discourse_peak_chart('Gen')

# Chart — Genesis ot_discourse_peak_chart('Gen')

In [ ]:

# Exodus — peak should be near the Exodus event / Red Sea crossing
print_ot_peak_score('Exo')

# Exodus — peak should be near the Exodus event / Red Sea crossing print_ot_peak_score('Exo')

In [ ]:

ot_discourse_peak_chart('Exo')

ot_discourse_peak_chart('Exo')

In [ ]:

# 1 Samuel — expect peaks around Saul's rejection / David and Goliath (ch 17) / Saul's death (ch 31)
print_ot_peak_score('1Sa')

# 1 Samuel — expect peaks around Saul's rejection / David and Goliath (ch 17) / Saul's death (ch 31) print_ot_peak_score('1Sa')

In [ ]:

ot_discourse_peak_chart('1Sa')

ot_discourse_peak_chart('1Sa')

5. Episode Boundary Detection¶

Episode boundaries are flagged when:

1. A wayehi (וַיְהִי) scene-setting formula appears
2. A wayyiqtol chain gap (≥3 verses without wayyiqtol) is followed by a qatal

In [ ]:

print_ot_episode_boundaries('Gen')

print_ot_episode_boundaries('Gen')

In [ ]:

# Ruth — four chapters, distinct episodes
print_ot_episode_boundaries('Rut')

# Ruth — four chapters, distinct episodes print_ot_episode_boundaries('Rut')

In [ ]:

# Exodus — episode count
boundaries = ot_discourse_episode_boundaries('Exo')
print(f"Exodus episode boundary signals: {len(boundaries)}")
boundaries.head(20)

# Exodus — episode count boundaries = ot_discourse_episode_boundaries('Exo') print(f"Exodus episode boundary signals: {len(boundaries)}") boundaries.head(20)

6. Akedah Case Study — Genesis 22¶

Genesis 22 (the binding of Isaac) is a classic test case for narrative peak analysis. Longacre and others argue it is a tightly constructed narrative with the peak at vv. 9–12 (binding + angel stopping Abraham).

In [ ]:

# Zoom into Gen 22 wayyiqtol density
wyy = ot_discourse_wayyiqtol_density('Gen')
print("Genesis 22 wayyiqtol density vs. surrounding chapters:")
wyy[wyy['chapter'].isin(range(20, 25))][['chapter', 'total_tokens', 'wayyiqtol_count', 'density']]

# Zoom into Gen 22 wayyiqtol density wyy = ot_discourse_wayyiqtol_density('Gen') print("Genesis 22 wayyiqtol density vs. surrounding chapters:") wyy[wyy['chapter'].isin(range(20, 25))][['chapter', 'total_tokens', 'wayyiqtol_count', 'density']]

In [ ]:

# Peak scores for Genesis 20–25
peaks = ot_discourse_peak_score('Gen')
peaks[peaks['chapter'].isin(range(20, 25))]

# Peak scores for Genesis 20–25 peaks = ot_discourse_peak_score('Gen') peaks[peaks['chapter'].isin(range(20, 25))]

7. Ad-hoc Queries¶

In [ ]:

# Which prophetic books still have wayyiqtol narrative sections?
for book in ['Isa', 'Jer', 'Eze', 'Hos', 'Amo', 'Jon', 'Mic']:
    p = ot_discourse_narrative_profile(book)
    print(f"{book:<5}  wyy%={p['wayyiqtol_pct']:>5.1f}%  "
          f"peak_ch={str(p['peak_chapter']):>3}  score={p['peak_score']}")

# Which prophetic books still have wayyiqtol narrative sections? for book in ['Isa', 'Jer', 'Eze', 'Hos', 'Amo', 'Jon', 'Mic']: p = ot_discourse_narrative_profile(book) print(f"{book:<5} wyy%={p['wayyiqtol_pct']:>5.1f}% " f"peak_ch={str(p['peak_chapter']):>3} score={p['peak_score']}")

In [ ]:

# Torah wayyiqtol density comparison
for book in ['Gen', 'Exo', 'Lev', 'Num', 'Deu']:
    p = ot_discourse_narrative_profile(book)
    bar = '█' * int(p['wayyiqtol_pct'] * 2)
    print(f"{book:<5}  {p['wayyiqtol_pct']:>5.1f}%  {bar}")

# Torah wayyiqtol density comparison for book in ['Gen', 'Exo', 'Lev', 'Num', 'Deu']: p = ot_discourse_narrative_profile(book) bar = '█' * int(p['wayyiqtol_pct'] * 2) print(f"{book:<5} {p['wayyiqtol_pct']:>5.1f}% {bar}")

In [ ]:

# Custom: get the raw data for a specific chapter
sp = ot_discourse_speech_density('1Sa')
sp_sorted = sp.sort_values('density', ascending=False)
print("1 Samuel chapters with highest speech verb density:")
sp_sorted.head(10)

# Custom: get the raw data for a specific chapter sp = ot_discourse_speech_density('1Sa') sp_sorted = sp.sort_values('density', ascending=False) print("1 Samuel chapters with highest speech verb density:") sp_sorted.head(10)