syllable division in transliteration

[bdenckla]

[...] I think that if there would be three the cut-signs that would be nicer.
The regular sign - the mid-point.
The sign that represents a Dagesh forte cut - as ~ (that tells the reader to pronounce the letter only once).
The sign that represents a connection of two word by Maqaph (can be signed by Maqaph).

(I wanted to add to a discussion from another part of the forum, but I thought it might be good to start a new topic.)

For now, I've made the following minimally-invasive extension to Jacobson-style transliteration:

For all 594 maqaf marks in Tanakh surrounded by the same sound-code (one or two letters), I use a slash instead of a dash.

E.g. the example that started this discussion now looks like this:



It turns out that there's another case of the same phenomenon in Exodus 20. That case now looks like this:



I'm not particularly set on the use of slash; at this point, the principle is more important than the details.

[bdenckla]

BTW I have now web-published the current state of my Jacobson-style transliteration of all of Exodus. The transliteration is based on the edition of MAM at the core of the Al-Hatorah Mikraot Gedolot.

[ducky]

Hi,

1. I've noticed that there is no change in the Maqaph sign.
2. Why on verse 3 the word יששכר is written ישכר?
3. I've noticed that there is no Mobile Sheva in cases of M after a definite article, like the word המילדת on verse 15 for example.
(In this case, there are two exceptions in this chapter that even though there is a Meteg, the Sheva is still silent (twice המילדות on verses 17 and 19).

[bdenckla]

1. I've noticed that there is no change in the Maqaph sign.

I only changed dash to slash for maqaf marks that would introduce ambiguity if they were left as a dash. I.e. I only changed dash to slash for maqaf marks that are surrounded by the same sound-code (e.g. b, v, ts, sh, etc.).

2. Why on verse 3 [of Exodus chapter 1] the word יששכר is written ישכר?

Because this is a qere-only edition and ישכר is the (implicit) ("perpetual") qere. E.g. in an edition of MAM I have prepared for a paper publisher, the ketiv/qere is made explicit, like this:

3. I've noticed that there is no Mobile Sheva in cases of M after a definite article, like the word המילדת on verse 15 [of Exodus chapter 1] for example.
(In this case, there are two exceptions in this chapter that even though there is a Meteg, the Sheva is still silent (twice המילדות on verses 17 and 19).

Did you mean למילדת (not המילדת) in verse 15?

Right now, I am just focused on rendering the sheva distinctions as they appear in Al-Hatorah. At some point in the future, I may go back and actually try to understand (and possibly question) those distinction-choices. Toward that (future) goal, I will record (on my "to do" list) this question of the treatment of המ... and similar words, e.g. the ones in Exodus chapter 1:

• verse 15 לַֽמְיַלְּדֹ֖ת lam-yal-l^e-DÓT
• verse 17 הַֽמְיַלְּדֹת֙ ham-yal-l^e-DÓT
• verse 18 לַֽמְיַלְּדֹ֔ת lam-yal-l^e-DÓT
• verse 19a הַֽמְיַלְּדֹת֙ ham-yal-l^e-DÓT
• verse 19b הַמְיַלֶּ֖דֶת ham-yal-LE-det (a different case, really, but useful for contrast)
• verse 20 לַֽמְיַלְּדֹ֑ת lam-yal-l^e-DÓT
• verse 21 הַֽמְיַלְּדֹ֖ת ham-yal-l^e-DÓT

[bdenckla]

For all 594 maqaf marks in Tanakh surrounded by the same sound-code (one or two letters), I use a slash instead of a dash.

Update: I found one more case, making my current total 595. That new case is worth showing because it is a pretty wild one, because it is the only case I've found for which the sound-code surrounding the maqaf is a vowel sound-code: "u" (1 Chr. 23:13):



Now, of course we could debate whether, regardless of the sound-codes used in this transcription, if those are

• The sounds that should be made, ideally (in a given dialect of Hebrew)
• The sounds that would be made, really (in a given dialect of Hebrew)

E.g. I have always wondered whether an א such as this one was meant, by the Tiberian Masoretes, to be truly silent or was meant to call for a closing glottal stop. For comparison, I think it is agreed that a final ע was meant, by the Tiberian Masoretes, to always call for a sound, not silence. Possibly (probably?) a closing glottal stop. Though I think it is also agreed that the Tiberian Masoretes meant ע and א to represent different sounds. So, it is not clear how far analogies between א and ע should go. I should probably consult G. Khan's book on this and many other issues.

And then of course there is the question of whether initial shuruq was in fact more of a "wu" sound... I'm not sure where current scholarship stands on that.

But these questions, though interesting, are not relevant to my task at hand, which is to represent a dialect of Hebrew whose phonetic abstraction/approximation has the same sound ("u") on both sides of the maqaf.

In particular, my task at hand is to provide a notation that shows that these two "u" sounds span a maqaf boundary. In fact it would be impossible for them to span a plain old syllable boundary, but perhaps that's all the more reason to not use plain old dash between them!

As to what a reader does with this information (in particular, whether they interpret this notation as calling for a single, long "u" sound or two separate medium-length ones), I am going to make no normative prescription. Not for this "u/u" case or any other case ("sh/sh", "t/t", etc.). It is a bit odd to decline to make a normative prescription for a supposedly-phonetic notation, I know. But nonetheless I decline.

[bdenckla]

Some weird-looking doublings appear in my transliteration when there's dagesh ḥazaq after maqaf, i.e. the connected (maqaf compound) form of the conjunctive dagesh. We run into this pretty early in Torah, e.g. in Gen 1:12:



This looks particularly weird when we have both "together" and "split" forms of the same doubling in close proximity, as in Gen. 15:2:



And this looks particularly weird (in a different way) when we have two letters representing one sound, as with "sh":



To be fair, that weirdness is not specific to the connected form of the conjunctive dagesh. It looks pretty weird in the disconnected form as well. E.g. ק֥וּמִי שְּׁבִ֖י in Isaiah 52:2 transliterates to KU-mí shsh^e-VÍ, and there's a phrase in Gen 31:13 that looks like this:



It occurs to me that the "dagesh ḥazaq after maqaf" problem could be considered the evil twin of the "same sound surrounds maqaf" problem. The relationship between these two problems becomes clearer if you state them like this:

• Should we avoid transliterations that look like dagesh ḥazaq, but aren't?
  • E.g. should we avoid n-n where "-" is a maqaf?
  • A follow-up question is, if we should avoid such things, how? (My answer for now: yes, we should avoid such things by using slash, e.g. n/n.)
• Should we avoid transliterations that don't look quite like a dagesh ḥazaq, but are in fact a dagesh ḥazaq? I.e. should we avoid transliterations that look like a failure to split the doubled letters of a "normal" dagesh ḥazaq?
  • E.g. should we avoid -pp where "-" is a maqaf?
  • A follow-up question is, if we should avoid such things, how?
    • My answer for now: no, we need not avoid such things.
    • If we were concerned to avoid things like -pp, we could again use slash in those cases, e.g. /pp. (Or, more fully, ‘ó-se/pp^e-RÍ.)

I can't resist the impulse to try to generalize and speculate that perhaps both of these problems reflect a fundamental tension regarding a maqaf compound:

• In what ways is a maqaf compound, as a whole, just like any other chanted word?
• In what ways is a maqaf compound, as a whole, not just like any other chanted word?

Another way of putting that second question is, in what ways do the parts ("atoms") of a maqaf compound function like individual chanted words?

[ducky]

Hi,

Maqaph.
I still don't get it why you use a dot or a slash to sign the Maqaph when you could use... wait for it...
a Maqaph.
What's the point to find a different point to point a point if you already have a good point?
I mean a true Maqaph (־)

יששכר/ישכר
If you, when you'll finish your nice project, intent to keep the Hebrew text next to the English transcription, then you must make sure that the Hebrew text, which is your source, is reliable. Never mind about the conditions of the transcription.
It is like putting a nice shoe on the left leg while cutting off the right one.

המילדת
all of the case you wrote should have the letter M after the definite article with a Mobile Sheva, except for verse 17 & 19.

Letter Aleph is a consonant which had/have a sound that crack in your throat.
like half a cough.
But already in the biblical era, the Aleph became silent in cases of no vowel.

הוא
If you're talking about the Tiberian dialect.
I don't know any claim that they pronounced the Aleph in this case.

Initial Shuruq probably evolved from "wu".
(in Babylonian dialect, the prefixed W before Sheva was "wi")
But I think that the fact that it has a Shuruq, and not Qibbuts, shows that it wasn't pronounced as "wu" by them.

עשה פרי is Dehiq.

Maybe you should use another sign for SH, like $

[bdenckla]

Maqaph.
I still don't get it why you use a dot or a slash to sign the Maqaph when you could use... wait for it...
a Maqaph.
What's the point to find a different point to point a point if you already have a good point?
I mean a true Maqaph (־)

That's also the suggestion Dr Jacobson made when I contacted him about this need for another symbol. It is not a bad idea, and may be worth trying, but I fear it for two reasons:

• Bidi (RTL/LTR) issues: maqaf has Unicode Bidirectional category "R" (meaning, right-to-left) and I would be mixing it with Latin letters which of course have Bidi category "L". Could be fine, could cause headaches. Would need to try it in a variety of Unicode rendering environments (OS, application) to see. (Theoretically all environments should implement the same Bidi rules established by Unicode but I think in practice they vary.)
• Font ugliness: maqaf will often have to come from a different font than the Latin letters, so it may not match well in stroke width, size, etc. Would need to try it in a variety of fonts and font-environments (OS, application) to see.

יששכר/ישכר
If you, when you'll finish your nice project, intent to keep the Hebrew text next to the English transcription, then you must make sure that the Hebrew text, which is your source, is reliable.

I think I am working with the best liberally-licensed Hebrew text available (MAM). Admittedly, I'm biased, since I've spent the last few years of my life working on it and editions derived from it, almost full-time. In fact I believe MAM is probably one of the best Hebrew texts, even ignoring its license advantages. I'd say it ranks up there with, for example Keter Yerushalayim, which you cannot license for love or money (I have tried money). What about the יששכר/ישכר ketiv/qere calls into question the reliability of my base text?

המילדת
all of the case you wrote should have the letter M after the definite article with a Mobile Sheva, except for verse 17 & 19.

Thanks. I've recorded that as an issue in my issue tracker, for later review.

עשה פרי is Dehiq.

Yeah I should have mentioned that the dagesh ḥazaq there on פ is part of a well-studied phenomenon called dehiq or conjunctive dagesh. In this case it is the "connected" or "maqaf" form of conjunctive dagesh, i.e. for these purposes we consider maqaf to be a conjunctive (pseudo)-accent.

Maybe you should use another sign for SH, like $

Yes, certainly it is less awkward if you avoid digraphs (two-letter combinations) like sh. And yes, dollar-sign ($) is certainly a candidate. In fact it is a classic candidate, since $ is the code for "true shin" (i.e. shin with a shin dot) in the classic pre-Unicode Michigan-Claremont encoding of Hebrew. Or, I could get more pretentious (and more phonetic) and use the IPA symbol ʃ. At the same time, there is value in sticking with Jacobson's "sh" notation as a standard. And, Jacobson's notation kind of requires capitalization, which is undefined for $ and ʃ. So that's an issue.

There's currently only 81 words that involve either shsh or SHSH so it is not that widespread an ugliness. The other relevant digraph, ts, appears in repeated form much more rarely than sh does: currently I find only 7 cases. The only other digraph is kh, and that does not (and cannot) appear in repeated form, i.e. khkh is an impossibility.