Maltese and Hebrew

Maltese and Hebrew

check out with:

 svn co https://apertium.svn.sourceforge.net/svnroot/apertium/incubator/apertium-mt-he

Todo list

1. Make program to generate a full form list for a given Maltese verb stem
2. Add closed categories to the Maltese analyser (prepositions, conjunctions,pronouns, determiners, numerals)
3. Add closed categories to the bidix
4. Add closed categories to the Hebrew dictionary (prepositions, pronouns, determiners, numerals, conjunctions)
5. Generate Hebrew verb entries from hspell output in Apertium format
6. Add existing verbs to the bidix
7. Look at how attached/clitic pronouns can be treated (Spanish,Catalan,Italian have similar requirements)
8. Add high frequency verbs to the full form list / Maltese analyser
9. Add high frequency nouns, adjectives, adverbs to Maltese analyser
10. Align Maltese and Hebrew bibles
11. Add attached articles to all noun entries -- not 100% sure how this works, see for ex. <e lm="הולדת"><par n="ה"/><i>הולדת</i><par n="עת__n_f"/></e>
12. Wiktionary script: 1) Fix URL encoding 2) Output English gloss as a comment
13. Change mt_verbs.py to generate negation circumfixes for all verbs: ma(--)x or m'(--)x generates <neg>
14. mt_verbs.py: Ability to set alternative forms with a direction restriction, e.g. sp['past.p1.sg'] = ('inkun', 'LR')
15. Possessive suffixes on Maltese nouns. p.100 "Teach yourself"
    • This has been started, see: Completed paradigms above this line
    • Find out if all nouns can take these (in theory), or only certain classes of nouns.
16. Write a script that scans the corpus for adjective/noun pairs, looking for gender agreement errors.
17. Many verbs with the j- prefix can also take an i- prefix or a ø- null prefix, this should be included into each of the class files. (E.g. strong_double_mid_radical.py)
18. Rule in verb prefixes for i- and j-
19. Past participles have gender/number -- like adjectives

Midterm sprint

Day	Date	Target	Achieved	Cov. WP	Cov. KPS	Notes
1	26th June	250	154	53.95%		First batch (top 120) of adjectives from Unhammer's adj_suspected.txt
2	27th June	250	411	55.19%	52.15%	Batches 2,3,4 from same list (next 440 items)
3	28th June	250	210	55.38%		Rest of adj_suspected-ku-ka-ci.txt, adj_suspected.txt
4	29th June	250	218	60.59%		Nouns from Wiktionary script, some high frequency words (det, prep)
5	30th June	250	248	63.20%	58.82%	Some nouns, adjectives, adverbs
6	1st July	250		64.38%	60.07%
7	2nd July	250		67.65%	66.23%
8	3rd July	250		70.88%	70.02%
9	4th July	250		—	—
10	5th July	250		72.13%	71.34%
11	6th July	250		72.62%	72.71%
12	7th July	250		73.85%	74.04%
13	8th July	250		74.32%	74.50%
14	9th July	250		75.05%	75.34%
15	10th July	250
16	11th July	250
17	12th July	250
18	13th July	250

Maltese verbs

No infinitive. Stem is third person singular, masculine perfect tense.

Second verb infinitive does not exist, instead both verbs are conjugated. "I want to eat" = "I want I eat"

A verbal stem can consist of:

1. Three consonants (radicals) with the medial radical between one of six vowel combinations. (Triliteral)
   • kiteb
2. Four consonants, some having two repeated biradical bases. (Quadriliteral)
3. Two consonants, or a consonant and a semivowel

In verbs with 'għ' or a + 'j' as the third radical, only have the first two radicals in the stem word which ends in 'a' (open syllable).

1. Verbs that have three non-semivocalic consonants are called sound or strong verbs.
2. Verbs that have three radicals, with the last radical being 'għ' or 'j' are called defective or weak verbs.
3. Triliteral verbs with long 'a' or 'ie' between 1st and 2nd radicals are called hollow verbs.
4. Triliteral verbs with where the second and third radicals are the same are called doubled or geminated verbs.

Examples:

Type	Example	Cons	Vowel config	Translation
Sound (Tri)	ħareġ	ħ·r·ġ	2. a·e	he went out
Sound (Quad)	ħarbex	ħ·rb·x	2. a·e	he scribbled
Defective	qata'	q·t·għ	1. a·a	he cut
Weak	mexa	m·x·j	4. e·a	he walked
Hollow	qal	q·w·l	1. a·a	he said
Hollow	sab	s·j·b	1. a·a	he found
Doubled	habb	h·b·b	1. a·a	he loved

Tenses:

1. Perfect: Action in the past
   • seraq "he robbed"
2. Imperfect: Action in the present/future
   • jisraq "he steals" or "he will steal"
3. Imperative: Order/command
   • israq (sg), isirqu (pl) "steal!"
4. Present participle: Only from intransitive verbs, and some verbs of motion. Has both verbal/adjectival function. Has m/f/pl
   • nieżel (m.sg) "descending"
   • nieżla (f.sg) "descending"
   • neżlin (mf.pl) "descending"
5. Past participle: Has both verbal/adjectival function. Has m/f/pl
   • misruq (m.sg) "stolen"
   • misruqa (f.sg) "stolen"
   • misruqin (mf.pl) "stolen"
6. Verbal noun
   • serq "robbing", "theft"

Vowel patterns:

1. KaTaB
2. KaTeB
3. KeTeB
4. KiTeB
5. KoToB

Pronominal Suffixes

Pronouns, prepositions etc. can be dropped in favor of complex suffixes added to the verb;

For example,

iktbilha = ikteb + il + ha
write(imp,p2,sg) + to + her

See The Verb with Pronominal Suffixes for through documentation.

Resources on verbs

1. Maltese Grammar
2. Maltese verbs on Wiktionary
3. Maltese conjugation tables on Wiktionary
4. Maltese verbs on Verbix
5. The verbal morphology of Maltese, Robert D. Hoberman and Mark Aronoff
6. GF Summer School: Progress in Maltese
7. Dana Dannélls, John J. Camilleri 2010. Verb Morphology of Hebrew and Maltese - Towards an Open Source Type Theoretical Resource Grammar in GF
8. http://www.archive.org/stream/DiziunariuMillMaltiGhallInglis/diziunariu_malti_inglis#page/n101/mode/2up/search/hajja
9. Morphologies of Asia and Africa: Maltese Morphology

Numerals

We have all Hebrew cardinal numbers and number construction defined in he.dix (still pending tests) (relevant definitions).

For Maltese, we have the cardinal numbers defined ([1]) along with some basic construction rules, but complex numbers is not ready yet (smart paradigms required).

Maltese numeral construction examples:

• 1..9, 11..19 - named numbers.
• 21..99 - "one and thirty" for 31 (wieħed u tletin).
• 100 and up - "three hundred and three and seventy" for 273 (tliet mija u tlieta u sebgħin).
• Same for thousands - "thousand nine hundreds and five and sixty") 1965 (elf disa' mija u ħamsa u sittin)

External resources

• w3dictionary, a wordnet?
• MyMemory.translated.net, search TMX'es
• Open-Tran.eu, search FOSS translations
• mt Wiktionary
• Old, scanned mt-en dictionary

See also

• Pending tests
• Regression tests

• Maltese morphological analyser web demo, using the resources of this language pair