Skip to content

Latest commit

 

History

History
222 lines (186 loc) · 9.14 KB

README.md

File metadata and controls

222 lines (186 loc) · 9.14 KB
Raw

Generation of Latin hyphenation patterns

The patterns for liturgical Latin hyph.la.liturgical.txt have been written by hand. They are improved continuously.

The patterns for medieval and modern Latin hyph.la.modern.txt have also been written by hand. They are currently unmaintained.

The original version of the patterns for classical Latin hyph.la.classical.txt have also been written by hand. Because of some deficiencies of the old patterns, we have created improved patterns for classical Latin using patgen. The new patterns support marks for long and short vowels (macrons and breves, e. g. lĭnguă Lătīnă) as long and short vowels are important for classical Latin. We use the following workflow.

Generation of hyphenation patterns for classical Latin

  1. We maintain a list index_verborum of about 7500 Latin words without inflected forms and without hyphenations, but containing information about the inflection class and hyphens in compound words and with special orthographic conventions. The exact format of the word list is described below.

  2. Run the script flexura.lua on this list, which creates all possible inflected forms, for example laudō, laudās, laudat, …, laudābō, …, laudāvī, …, laudātus, laudāta, laudātum, … from input laudō. These forms are stored in the file index_formarum.

  3. Run the script divisio.lua on the output of step 2 to hyphenate all the forms according to the basic rules. This is easy as the input list uses i and u only as vowels. This yields lau-dō, lau-dās, …

  4. Create orthographic variants by means of the script variatio.lua: vī-vō → vi-vo, vī-vō, ui-uo, uī-uō; jūs-tus → jus-tus, jūs-tus, jūs-tŭs, ius-tus, iūs-tus, iūs-tŭs; cæ-lum → cæ-lum, cǣ-lum, cæ-lŭm, cǣ-lŭm, cae-lum, ca͞e-lum, cae-lŭm, ca͞e-lŭm. It is crucial to do this after the hyphenation. The script also removes incompatible hyphenation points in homographs.

  5. Create patterns using patgen. The shell script generate-patgen-input.sh generates a list of hyphenated words as needed by patgen by means of the word list and the scripts mentioned above.

  6. Check the patterns using test files in the tests/nonliturgical directory. Every error can be corrected by putting the erroneously hyphenated word in the input list.

The steps 2 to 5 are executed by the generate-patterns.sh script. The publish-patterns.sh script is used to move the generated patterns to the correct directories.

All scripts are documented in a separate file.

Format of the word list index_verborum

Every line of the word list may contain up to four fields divided by commas. White space and following characters are treated as comments.

The first field contains a Latin word as written in a dictionary (normally, the first person present indicative active for verbs, the nominative singular for nouns, the nominative singular masculine for adjectives).

The second field contains the word type as described below. This field is empty for uninflectable words.

The third field contains the first person perfect indicative active for active verbs and the nominative singular masculine of the perfect passive participle for deponent verbs, but only if this form is irregular. For some nouns, it contains the genitive or the accusative and for some adjectives the feminine form or the genitve, as described below.

The fourth field contains the supine for active verbs, but only if this form is irregular, and the future participle for deponent verbs if its stem is different from the stem of the perfect passive participle.

Orthographic conventions

The orthographic conventions for the word list guarantee that all hyphenation points can be found correctly and that all other orthographic variants can be generated automatically.

  • Mark long single vowels (but no digraphs and diphthongs) with macrons: sēditiō, ædificō; do not mark short vowels.
  • Write j for every semivocalic i: jam, jaciō, Gājus, jējūnium. Also use j in the compounds of jacĕre: ab-jiciō (classical spelling: abiciō).
  • Use u and v according to the modern conventions: vērus, avārus, Ūrania.
  • Write æ and œ for the diphthongs ae and oe: cælum, tragœdia.
  • Use hyphens to mark compound words: ab-scindō, ē-jiciō, ob-œdiō, anim-ad-vertō, long-ævus. Do not use hyphens if a prefix is assimilated to the following element of the compound or if a prefix is extended by an epenthesis (d or s): abstrahō, assimils, afferō, difficilis, efflō, occidō, sustulī, complūrēs, redhibeō. However, a hyphen is necessary, when the second element begins with a vowel or when neg is followed by l: com-edō, red-emptiō, neg-ōtium, neg-legō.
  • Use an accented vowel (or a combining acute accent U+301 for vowels with macron) if an uninflectable word with at least two syllables has its accent on the last syllable: ab-hínc, ad-hū́c. Note: In other cases, the accent is generated automatically by the scripts.
  • au and eu are considered as diphthongs. Use a vertical bar, when au or eu is no diphthong (e. g. le|unculus).
  • au and eu are not considered as diphthongs when the u is part of an inflection ending for nouns, adjectives, and pronouns of the second declension (e. g. deus, meus). Use a combining double inverted breve (U+361) when eu is a diphthong in second declension words ending in -eus, e. g. Erechthe͡us.
  • Only use lowercase letters except at the beginning of proper nouns and their derivatives.

Possible word types

Verbs

  • 1 – verb of the first conjugation
  • 1intr – intransitive verb of the first conjugation
  • 2 – verb of the second conjugation
  • 2intr – intransitive verb of the second conjugation
  • 3 – verb of the third conjugation
  • 3intr – intransitive verb of the third conjugation
  • 3M – verb of the mixed third conjugation
  • 3Mintr – intransitive verb of the mixed third conjugation
  • 4 – verb of the fourth conjugation
  • 4intr – verb of the fourth conjugation
  • VP – verb with perfect forms only
  • VI – irregular verb
  • VIintr – intransitive irregular verb (eō and some of its compounds)

Intransitive verbs are treated separately because they only have impersonal passive forms. Deponent verbs, impersonal verbs, and irregular verbs having no passive forms at all are not marked as intransitive.

Examples:

laudō,1
ambulō,1intr
moneō,2,monuī,monitum
mittō,3,mīsī,missum
capiō,3M,cēpī,captum
audiō,4
hortor,1
vereor,2,veritus
ūtor,3,ūsus
patior,3M,passus
orior,3M,ortus,oritūrus
partior,4
ab-scindō,3,ab-scidī,ab-scissus
ex-audiō,4
meminī,VP
volō,VI,voluī
eō,VIintr,iī,itum

Nouns

  • D1 – masculine/feminine noun of the first declension (ending in -a, -æ, -ē, -ēs, or -ās)
  • D2 – masculine/feminine noun of the second declension (ending in -us, -r, -ī, -os, or -e͡us); if the nominative ends in -r, the third field contains the genitive.
  • D2N – neuter noun of the second declension (ending in -um, -us, -a, -on, or -os)
  • D3 – masculine/feminine noun of the third declension; the third field contains the genitive; the genitive is left out for nouns ending in -ēns/-entis, -is/-is, -dō/-dinis, gō/-ginis, -iō/-iōnis, -ō/-ōnis, -or/-ōris, -tās/-tātis, -trīx/-trīcis; the third field contains the accusative if this ends in -im.
  • D3N – neuter noun of the third declension; the third field contains the genitive; the genitive is left out for nouns ending in -men/-minis.
  • D3gr – Greek noun of the third declension (only for plural forms ending in -es); the third field contains the genitive
  • D4 – masculine/feminine noun of the fourth declension (ending in -us or -ūs)
  • D4N – neuter noun of the fourth declension (ending in -ū)
  • D5 – masculine/feminine noun of the fifth declension (ending in -ēs)

Examples:

pēcunia,D1
angustiæ,D1
lectus,D2
ager,D2,agrī
vir,D2,virī
dōnum,D2N
arma,D2N
canis,D3
carō,D3,carnis
turris,D3,turrim
agmen,D3N
animal,D3N,animālis
currus,D4
cornū,D4N
diēs,D5

Adjectives, pronouns, and declinable numerals

Adjectives are either comparable (e. g. longus, longior, longissimus) or incomparable (e. g. ūnicus). Pronouns, declinable numerals, and incomparable adjectives are similar, but pronouns and numerals do not have adverbs.

  • AC3/AI3 – comparable/incomparable adjective with three endings; if the masculine form does not end in -us, the third field contains the feminine form.
  • AC2/AI2 – comparable/incomparable adjective with two endings
  • AC1/AI1 – comparable/incomparable adjective with one ending; if the nominative does not end in -āns or -ēns, the third field contains the genitive.
  • P – pronoun
  • N – declinable numeral (cardinal, ordinal, or distributive)

Examples:

longus,AC3
ācer,AC3,ācris
ūnicus,AI3
brevis,AC2
prior,AC2
prūdēns,AC1
vetus,AC1,veteris
ego,P
ille,P
ūnus,N
prīmus,N
bīnī,N