Greek Preposition Analysis — NT and LXX¶

This notebook demonstrates bible_grammar.greek_prepositions, which provides frequency, case-binding, and collocate analysis for Greek prepositions in both the NT (MACULA Nestle1904) and the LXX (CenterBLC/Rahlfs).

Case binding — Greek prepositions govern one or more cases depending on meaning. Case is not stored on the preposition token itself; it is determined by an adjacency join to the word immediately following (word_num + 1). Results are normalized to Title case (Accusative, Genitive, Dative).

Data:

• NT: ~10,900 preposition tokens (MACULA Nestle1904)
• LXX: ~55,000 preposition tokens (CenterBLC Rahlfs 1935)

Functions:

Function	Purpose
greek_prep_frequency(corpus)	Frequency table of all prepositions
greek_prep_by_book(lemma, corpus)	One preposition across all books
greek_prep_distribution_table(corpus)	Major preps by book group
greek_prep_cases(lemma, corpus)	Case-binding profile for one prep
greek_prep_collocates(lemma, corpus, case)	Top collocates, filterable by case
compare_greek_preps(l1, l2, corpus)	Side-by-side collocate comparison
nt_lxx_compare(lemma)	NT vs. LXX case-binding comparison

In [ ]:

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

import pandas as pd
pd.set_option('display.max_rows', 60)
pd.set_option('display.max_colwidth', 45)

from bible_grammar.greek_prepositions import (
    greek_prep_frequency, greek_prep_by_book, greek_prep_distribution_table,
    greek_prep_cases, greek_prep_collocates, compare_greek_preps, nt_lxx_compare,
    NT_MAJOR_PREPS, LXX_MAJOR_PREPS, PREP_GLOSS,
)

import sys sys.path.insert(0, '../../../src') import pandas as pd pd.set_option('display.max_rows', 60) pd.set_option('display.max_colwidth', 45) from bible_grammar.greek_prepositions import ( greek_prep_frequency, greek_prep_by_book, greek_prep_distribution_table, greek_prep_cases, greek_prep_collocates, compare_greek_preps, nt_lxx_compare, NT_MAJOR_PREPS, LXX_MAJOR_PREPS, PREP_GLOSS, )

1. Overall Frequency¶

The top 4 (ἐν, εἰς, ἐκ/ἐπί, πρός) account for ~60% of all NT prepositions. The LXX has a similar profile but ἐπί is proportionally larger (13.1% vs. 8.1%) and ἕως is more frequent than in the NT.

In [ ]:

print('=== NT ===')
display(greek_prep_frequency(corpus='nt', top_n=15))

print('=== LXX ===')
display(greek_prep_frequency(corpus='lxx', top_n=15))

print('=== NT ===') display(greek_prep_frequency(corpus='nt', top_n=15)) print('=== LXX ===') display(greek_prep_frequency(corpus='lxx', top_n=15))

2. Distribution by Book Group¶

How do preposition profiles vary across corpus sections?

In [ ]:

print('NT — by book group:')
display(greek_prep_distribution_table(corpus='nt'))

print('LXX — by book group:')
display(greek_prep_distribution_table(corpus='lxx'))

print('NT — by book group:') display(greek_prep_distribution_table(corpus='nt')) print('LXX — by book group:') display(greek_prep_distribution_table(corpus='lxx'))

3. Case Binding Profiles¶

Greek prepositions can govern one case (monocase) or multiple cases with different meanings (bicase/tricase). The most interesting are the bicase prepositions:

• διά + genitive = through (means/agency); διά + accusative = because of (cause)
• ἐπί + genitive = on (resting); ἐπί + dative = on (based on); ἐπί + accusative = onto (motion)
• κατά + genitive = against; κατά + accusative = according to
• μετά + genitive = with; μετά + accusative = after
• παρά + genitive = from (alongside); παρά + accusative = beside (contrary to)

In [ ]:

# Monocase prepositions
for lemma in ['ἐν', 'εἰς', 'ἐκ']:
    print(f'\n{lemma} ({PREP_GLOSS.get(lemma, "")})')
    display(greek_prep_cases(lemma, corpus='nt'))

# Monocase prepositions for lemma in ['ἐν', 'εἰς', 'ἐκ']: print(f'\n{lemma} ({PREP_GLOSS.get(lemma, "")})') display(greek_prep_cases(lemma, corpus='nt'))

In [ ]:

# Bicase prepositions — theologically significant
for lemma in ['διά', 'κατά', 'μετά', 'παρά']:
    print(f'\n{lemma} ({PREP_GLOSS.get(lemma, "")})')
    display(greek_prep_cases(lemma, corpus='nt'))

# Bicase prepositions — theologically significant for lemma in ['διά', 'κατά', 'μετά', 'παρά']: print(f'\n{lemma} ({PREP_GLOSS.get(lemma, "")})') display(greek_prep_cases(lemma, corpus='nt'))

In [ ]:

# Tricase: ἐπί
print('ἐπί NT:')
display(greek_prep_cases('ἐπί', corpus='nt'))

print('ἐπί LXX:')
display(greek_prep_cases('ἐπί', corpus='lxx'))

# Tricase: ἐπί print('ἐπί NT:') display(greek_prep_cases('ἐπί', corpus='nt')) print('ἐπί LXX:') display(greek_prep_cases('ἐπί', corpus='lxx'))

4. NT vs. LXX Case-Binding Comparison¶

Does a preposition's case distribution shift between the LXX (translated from Hebrew) and the NT? This can reveal how Septuagintal Greek influenced NT usage.

In [ ]:

for lemma in ['διά', 'ἐπί', 'κατά', 'μετά']:
    print(f'\n=== {lemma} ===')
    display(nt_lxx_compare(lemma))

for lemma in ['διά', 'ἐπί', 'κατά', 'μετά']: print(f'\n=== {lemma} ===') display(nt_lxx_compare(lemma))

5. Collocate Analysis — What Follows a Preposition?¶

Use greek_prep_collocates() to find top words following each preposition. Filter by case to isolate one semantic sense.

In [ ]:

# ἐν + dative collocates (NT)
print('ἐν + Dative collocates (NT):')
display(greek_prep_collocates('ἐν', corpus='nt', case='Dative', top_n=15))

# ἐν + dative collocates (NT) print('ἐν + Dative collocates (NT):') display(greek_prep_collocates('ἐν', corpus='nt', case='Dative', top_n=15))

In [ ]:

# διά + genitive (agency/means) vs. + accusative (cause)
print('διά + Genitive (NT) — through/by means of:')
display(greek_prep_collocates('διά', corpus='nt', case='Genitive', top_n=15))

print('διά + Accusative (NT) — because of:')
display(greek_prep_collocates('διά', corpus='nt', case='Accusative', top_n=15))

# διά + genitive (agency/means) vs. + accusative (cause) print('διά + Genitive (NT) — through/by means of:') display(greek_prep_collocates('διά', corpus='nt', case='Genitive', top_n=15)) print('διά + Accusative (NT) — because of:') display(greek_prep_collocates('διά', corpus='nt', case='Accusative', top_n=15))

In [ ]:

# εἰς + accusative — destination (NT vs LXX)
print('εἰς + Accusative collocates (NT):')
display(greek_prep_collocates('εἰς', corpus='nt', case='Accusative', top_n=15))

print('εἰς + Accusative collocates (LXX):')
display(greek_prep_collocates('εἰς', corpus='lxx', case='Accusative', top_n=15))

# εἰς + accusative — destination (NT vs LXX) print('εἰς + Accusative collocates (NT):') display(greek_prep_collocates('εἰς', corpus='nt', case='Accusative', top_n=15)) print('εἰς + Accusative collocates (LXX):') display(greek_prep_collocates('εἰς', corpus='lxx', case='Accusative', top_n=15))

6. Compare Two Prepositions Side by Side¶

Both εἰς and πρός can mark direction/goal. How do their objects differ?

In [ ]:

compare_greek_preps('εἰς', 'πρός', corpus='nt', top_n=20)

compare_greek_preps('εἰς', 'πρός', corpus='nt', top_n=20)

In [ ]:

# ἀπό vs. ἐκ — both mark separation/source
compare_greek_preps('ἀπό', 'ἐκ', corpus='nt', top_n=20)

# ἀπό vs. ἐκ — both mark separation/source compare_greek_preps('ἀπό', 'ἐκ', corpus='nt', top_n=20)

7. Single-Book Focus¶

Zoom into Romans for Pauline preposition usage.

In [ ]:

print('Preposition frequency in Romans:')
display(greek_prep_frequency(corpus='nt', book='Rom', top_n=12))

print('διά case binding in Romans:')
display(greek_prep_cases('διά', corpus='nt', book='Rom'))

print('Preposition frequency in Romans:') display(greek_prep_frequency(corpus='nt', book='Rom', top_n=12)) print('διά case binding in Romans:') display(greek_prep_cases('διά', corpus='nt', book='Rom'))

In [ ]:

# Compare Gospels vs. Pauline διά usage
print('διά + Genitive (Gospels):')
display(greek_prep_collocates('διά', corpus='nt', case='Genitive', book_group='Gospels', top_n=10))

print('διά + Genitive (Pauline):')
display(greek_prep_collocates('διά', corpus='nt', case='Genitive', book_group='Pauline', top_n=10))

# Compare Gospels vs. Pauline διά usage print('διά + Genitive (Gospels):') display(greek_prep_collocates('διά', corpus='nt', case='Genitive', book_group='Gospels', top_n=10)) print('διά + Genitive (Pauline):') display(greek_prep_collocates('διά', corpus='nt', case='Genitive', book_group='Pauline', top_n=10))

8. Distribution of a Single Preposition Across All Books¶

In [ ]:

# ἐν distribution across NT
print('ἐν distribution across NT:')
display(greek_prep_by_book('ἐν', corpus='nt'))

# ἐν distribution across LXX
print('ἐν distribution across LXX:')
display(greek_prep_by_book('ἐν', corpus='lxx'))

# ἐν distribution across NT print('ἐν distribution across NT:') display(greek_prep_by_book('ἐν', corpus='nt')) # ἐν distribution across LXX print('ἐν distribution across LXX:') display(greek_prep_by_book('ἐν', corpus='lxx'))