Understanding Hebrew vowels - a pandemic project

I have always found the vowels in the Hebrew Bible quite difficult. If I had to begin again, I would learn Biblical Hebrew today beginning with the music of the accents. I would hope to have a teacher who could train my ear without having to read dots and dashes. The music as reconstructed by Haïk-Vantoura is so beautiful that one can escape both the foreignness of a new language and its religious character by entering into beauty first. Actually my first learning of Hebrew was through music, the Chichester Psalms by Bernstein. Psalm 2 was quite challenging to sing with its angular rhythms. (Different music than the music embedded in the text, of course).

I say, escape the religious nature of the text, because there is a tendency, when translating a religious text, to jump to conclusions based on previous training or exposure to cultural norms. The music bypasses that to some extent.

Having set many verses in English to the embedded music in the Hebrew, (a miniscule amount of the total possibilities here), I am reviewing the language by examining the role of the vowels. Hebrew was written without them for most of its early life. They were added in the Masoretic text in the centuries following the destruction, remembered today. Now the pointed text is 'obsolete in Modern Hebrew', and the unpointed maleh text is used. It bears a relationship to the pointed text, but what is that relationship? The comparison of the two text forms is described at the link, and it was from that wiki article that I took my first rule: the hireq is not realized in words that do not contain /i/ in their base form. My understanding depended first on reading and decoding these rules, a process I found quite difficult.

Slave

That first rule I followed was to my mind only partially true until this moment, but it was an adequate starting point. I had a list of many stems (about 75!) that are now reduced to 5 stems: bin, hih, ihvh, kih, and mmi. Dozens of my 'exceptions to this rule' are managed by the rule above it about schwa nah. (See the summary at the end of this post). If I also used the remaining sub-points in that list below, who knows, rules and exceptions may be reduced again. The problem with decoding human language and re-coding it in a computer data language is subtle. What I have done is a second chiseling of the rock, knocking away some things that are not part of the final image. A third chiseling will happen.

These are a rough statement of the rules I have implemented. I have been able to section my program based on these criteria applied in this sequence:

1. some common middle sequences of vowels and consonants involving vv, (needs further analysis), but it does follow a rule, sort of,
2. vav's appear sometimes depending on hatef qamats(1459), hatef patah(1458) and hataf segol(1459),
3. hireq (1460) prior to aleph in third position on a stem for some suffixes, (ti, im, t) becomes i and the aleph disappears,
4. nearly 50 common suffixes apply universally and about 10 have a few exceptions,
5. qamats-h at the end of a word sometimes is dropped,
6. qamats(1464) in several combinations with i and v becomes ii for some stems,
7. patah 
1. with v becomes ii for some stems,
2. with i likewise,
8. holam(1465)
1. 'oh' becomes o for some stems,
2. for a set of stems, holam does not appear,
3. oa and ao have some specific rules,
9. tsere (1461) 
1. vi may become vii,
2. tsere with yod may have the yod removed,
3. tsere with aleph may drop the aleph,
4. i with tsere may become ii,
5. tsere may become i
10. hireq
1. i + hireq + the first character of the stem with dagesh and patah or qamats becomes ii,
2. hireq disappears for a closed syllable,
3. in the hiphil, hi is suppressed,
4. i may become ii,
5. hireq may become i,
11. segol may become i,
12. a missing aleph may be restored,
13. i with dagesh and segol, patah, or qamats may become i
14. i with holam becomes io,
15. v with dagesh becomes u in front of b, m, and p, (that one is really easy),
16. some stems have a yod removed,
17. for the remaining qamats
1. with aleph may be dropped,
2. for some stems, create a final h, or ha,
3. with m, may become mo,
4. with h, may become ho,
5. in most closed syllables, will become o,
6. with several stems in open syllables will become o,
18. Finally, some i's are removed with common single and double prefixes,
19. some stems do not allow double i,
20. and there are a couple of spelling errors in the WLC that are corrected in the maleh text.

That is still too many rules and too many exceptions, and I haven't let you know yet the lists of stems to which these rules apply or not, but it works and it is now a better predictor for the remaining 7/8ths of the Bible. It could be that some classes of stems may have a rule applied.

Several possible refinements suggest themselves to me as I document this sequence. If you have any insights as to general rules, please let me know in a comment. I think this is a useful exercise. It has been my pandemic project.

Now let me read those instructions yet again!

• Every letter that appears in vowelled text also appears in unvowelled text. [Not so, some /h/, /i/, and /a/ letters are dropped]
• After a letter vowelled with a kubuts (the vowel /u/), the letter ו appears: קופסה‎, הופל‎, כולם. [yes]
• After a letter vowelled with a holam haser (the vowel /o/) the letter ו appears: בוקר‎, ישמור. [sometimes. There are many exceptions.]
• After a letter vowelled with a hirik haser (the vowel /i/) the letter י appears: דיבור‎, יישוב‎, תעשייה. The letter י does not appear in the following situations:
• Before a shva nah, for example: הרגיש‎, מנהג‎, דמיון; [yes with exceptions]
• Words whose base forms do not contain the vowel /i/: ליבי ‎(לב)‎, איתך ‎(את)‎, עיתים ‎(עת); [yes with exceptions]
• After affix letters, like in מביתו, מיד, הילד, and also in the words: עם‎, הינה ‎(=הִנֵּה, and inflected: עימי etc., הינו, etc.), אם‎, מן; [yes with exceptions, further study required]
• Before יו (/ju/ or /jo/): דיון‎, קיום‎, בריות‎, נטיות. [haven't tried this rule]
• After a letter vowelled with a tsere (the vowel /e/) the letter י generally does not appear [ממד ‎(=מֵמַד‎), אזור ‎(=אֵזוֹר‎)], but there are situations when י does appear (תיבה‎, הישג) and in words in which tsere replaces hirik because the presence of a guttural letter (אהחע"ר‎): תיאבון ‎(שיגעון‎), תיאבד ‎(תימצא). [yes but I didn't test to see if a guttural is always involved.]
• Consonantal ו (the consonant /v/) is doubled in the middle of a word: תקווה‎, זווית. The letter is not doubled at the beginning or the end of a word: ורוד‎, ותיק‎, צו. Initial ו is doubled when an affix letter is added except for the affix ו- (meaning "and-"). Thus from the word ורוד one has הוורוד but וורוד (that is, וּוֶרד). [If you can read this rule... my implementation works, perhaps I can explain it.]
• Consonantal י (the consonant /j/) is doubled in the middle of a word, for example: בניין‎, הייתה. The letter is not doubled at the beginning of a word or after affix letters: ילד‎, יצא ‎(=יֵצֵא‎), הילד.Still, consonantal י is not doubled in the middle of a word when it is before or after mater lectionis: פרויקט‎, מסוים‎, ראיה ‎(=ראָיָה‎), הפניה‎, בעיה. [I have all the ii's working, but the statement of this rule is unclear to me]

Those are the most basic rules. For every one of them are exceptions, described in the handbook "כללי הכתיב חסר הניקוד" that the Academy publishes in Hebrew. [which I have not read, and probably could not.]

Here is an earlier post on this subject from a few months ago when I first set a chisel to this rock. Next post on this subject, I will see how many rules have been further refined and what exceptions by stem remain. It may then be time to look in detail at the stem lists and constraints.