Sunday, 24 August 2025

Status report on the Swan Song

 Still alive and kicking, said the Swan.

Here's the current status on all fronts:

All volumes complete verse by verse with the music -- like this. 


This image contains: 

  • 5 syllables (middle of last line)  in the score 
  • over three words (top line Hebrew) of the first verse of 1 Chronicles
  • a transcription (lyrics) based on a simple mapping of consonants and vowels
  • a translation with verse number and harmonic progression in parentheses
  • the Hebrew accents from which the score is derived
  • the SimHebrew equivalent with verse number (bottom right).

There are about 34,400 separate images for the 23,196 verses. Their production and merge with the html was controlled by several procedures from my database. So they are an accurate representation of my translation and the automated music process, but of course open to textual and linguistic criticism.

Here's a more complex example. See if the syllable counts agree with the column of numbers in the bottom centre of the image. There is an occasional discrepancy +/- 1 difference between the syllable count and the notes in the score. I have found it impossible to find this bug. I'm not the only software developer with this problem. I suspect there are other bugs that I have not noticed.


Tasks to do: 

  • for 7 volumes, update the html to include legarmeh (routine process, takes maybe a few hours of interrupted work to do -- probably unimportant).
  • 10 more covers to design -- really fun to do. 10 days (not contiguous) and it will be done.
  • 17 more ISBN's to get. Routine.
  • 10 + brief introductions to the music of the volume. Very challenging. In this area my work has converged with another student of cantillation history. My data formats may be very useful to him and I have to think carefully to get good query design from my database to see the potential patterns.
Future work: to anyone who is interested, the data that is generally available now online in various places in Unicode can be coded for the questions we want to ask. It's a very large undertaking to consider the relationships in the Scripture itself and to reconcile the history of cantillation with the theoretical discovery that has been demonstrated by Suzanne Haïk-Vantoura. By the way, a large number of video recordings of her work done by John Wheeler are now available (again) here. (Very slow start but they are there.)

My computer apparently has only a few months till its expiry date. The folks who make Windows do not consider it worthy to support. I have routinely 30 to 35 windows active on it. It is very fast and has adequate memory. This sin from Microsoft is duly noted for the heavenly assizes. I can certainly buy another computer but whether I can still install Oracle and the last working copy of my complex software and all the other things I need for this unique environment is moot. Maybe I will just maintain it off the web and go back to sneaker net to get any words out. But my expiry date is also approaching -- who knows when!



Tuesday, 12 August 2025

The Mysterious Decalogue

I have examined an alternate solution to the many te-amim in the two decalogues. The data is taken from here. It's been a good exercise not without difficulty. 

בטעם העליון
אָֽנֹכִ֖י יְהֹוָ֣ה אֱלֹהֶ֑יךָ אֲשֶׁ֧ר הוֹצֵאתִ֛יךָ מֵאֶ֥רֶץ מִצְרַ֖יִם מִבֵּ֥ית עֲבָדִֽים׃
לֹ֣א יִהְיֶֽה־לְךָ֩ אֱלֹהִ֨ים אֲחֵרִ֜ים עַל־פָּנַ֗י לֹ֣א תַעֲשֶֽׂה־לְךָ֣ פֶ֣סֶל ׀ וְכׇל־תְּמוּנָ֡ה אֲשֶׁ֣ר בַּשָּׁמַ֣יִם ׀ מִמַּ֡עַל וַֽאֲשֶׁר֩ בָּאָ֨רֶץ מִתַּ֜חַת וַאֲשֶׁ֥ר בַּמַּ֣יִם ׀ מִתַּ֣חַת לָאָ֗רֶץ לֹֽא־תִשְׁתַּחֲוֶ֣ה לָהֶם֮ וְלֹ֣א תׇעׇבְדֵם֒ כִּ֣י אָֽנֹכִ֞י יְהֹוָ֤ה אֱלֹהֶ֙יךָ֙ אֵ֣ל קַנָּ֔א פֹּ֠קֵ֠ד עֲוֺ֨ן אָבֹ֧ת עַל־בָּנִ֛ים עַל־שִׁלֵּשִׁ֥ים וְעַל־רִבֵּעִ֖ים לְשֹׂנְאָ֑י וְעֹ֤שֶׂה חֶ֙סֶד֙ לַאֲלָפִ֔ים לְאֹהֲבַ֖י וּלְשֹׁמְרֵ֥י מִצְוֺתָֽי׃
Music according to the deciphering key for the first separation in Exodus 20

Note the verse structure. The first image above has only two verses. The atnah (on the word coloured brown in the Hebrew text) appears only twice and the music comes to a possible end only twice. The Aleppo verses 3 4 and 5 and 6 have the mid-verse rest on לְשֹׂנְאָ֑י, seven words before the end. The second image and text below combines verses 2 and 3 and preserves 4, 5, and 6 as separate verses.

This raises the question, of course, as to what is a verse. It is clear that the te'amim define verses. 
  1. They begin on several different notes with frequency varying by book but they always (100% in mgketer.org) end with the tonic, unless there has been a copying error and there are a few in WLC. The big colon is a marker, but not a note. In all the cases above where there is a big colon, a silluq appears in a prior word to bring the music to the tonic.
  2. And often, (92.5% of verses) they have an internal cadence on the subdominant (atnah). 

בטעם התחתון
אָֽנֹכִי֙ יְהֹוָ֣ה אֱלֹהֶ֔יךָ אֲשֶׁ֧ר הוֹצֵאתִ֛יךָ מֵאֶ֥רֶץ מִצְרַ֖יִם מִבֵּ֣ית עֲבָדִ֑ים לֹֽא־יִהְיֶ֥ה לְךָ֛ אֱלֹהִ֥ים אֲחֵרִ֖ים עַל־פָּנָֽי׃
לֹֽא־תַעֲשֶׂ֨ה לְךָ֥ פֶ֙סֶל֙ וְכׇל־תְּמוּנָ֔ה אֲשֶׁ֤ר בַּשָּׁמַ֙יִם֙ מִמַּ֔עַל וַֽאֲשֶׁ֥ר בָּאָ֖רֶץ מִתָּ֑חַת וַאֲשֶׁ֥ר בַּמַּ֖יִם מִתַּ֥חַת לָאָֽרֶץ׃
לֹֽא־תִשְׁתַּחֲוֶ֥ה לָהֶ֖ם וְלֹ֣א תׇעׇבְדֵ֑ם כִּ֣י אָֽנֹכִ֞י יְהֹוָ֤ה אֱלֹהֶ֙יךָ֙ אֵ֣ל קַנָּ֔א פֹּ֠קֵ֠ד עֲוֺ֨ן אָבֹ֧ת עַל־בָּנִ֛ים עַל־שִׁלֵּשִׁ֥ים וְעַל־רִבֵּעִ֖ים לְשֹׂנְאָֽי׃
וְעֹ֥שֶׂה חֶ֖סֶד לַאֲלָפִ֑ים לְאֹהֲבַ֖י וּלְשֹׁמְרֵ֥י מִצְוֺתָֽי׃
Music according to the deciphering key for the second separation in Exodus 20

Suzanne Haïk-Vantoura did her work with the Letteris Edition (19th century). Here is one of her manuscripts from the library collected by Jonathan Wheeler, her American Editor. Jonathan sent me the DVD of this library before he disappeared from the Web. He was a stern taskmaster but I did not interact with him very much. You will find this in the pdfs under articles on page 18 of a file called resume_key_english.pdf. You can see how different the accents are in that edition but the piece is recognizable.
Exodus 20 handwritten by Suzanne Haïk-Vantoura

Haïk-Vantoura's deciphering key applies (she says) without difficulty to the existing verse structure and the sometimes multiple accents per syllable. There are questions, however. How would one decide what sequence to apply two accents on one syllable? My program probably used the sequence in which the accents are coded if they are below the text. I don't know if such a decision is possible with a manuscript or a font image. Suzanne may have made a musical decision different from the ones I have let stand in my automated work. You can see her decisions if you read the whole PDF. When she speaks of first and second sign on a single syllable, I would assume she is reading right to left. That does not seem to be the case in the examples she gives above.

Automation would not have been possible for her. She died before MusicXML was invented. (And lucky her, she was not trained as a programmer but first as a musician, composer, and performer.) My solution using her approach is below. It is workable but I do admit, the split solution is also workable. I am grateful to have been introduced to it. There are other minor instances of two sublinear te'amim on a single syllable. Suzanne's solution applies there with good effect. It's as if the accent pair are a different sort of rhythmic ornamentation.
Haïk-Vantoura's solution. The deciphering key allows for all the existing accents without difficulty.
Doing mercy for thousands is more important for our character development than chasing te'amim. Maybe the music will help us read and act more truly.

Friday, 8 August 2025

A few examples of tsinnor and tsinnorit

Tsinnorit is a new name for an accent to me. I have always found the usage of the accent names inconsistent in all the books I have read (or seen and not read). The name is probably in some of the books I have looked at but I may not have noted it. Nonetheless, confusion or not, there is a pattern here, and it may lead to some insight into the mind of the ancient designer. Definitely the sign  ֮seen above as a sideways s often leads to a particular note. (But its not quite that simple).

Here's a collection of tsinnorit resolving to the supertonic.

Job 20:27 - tsinnorit (on the recitation pitch A) to mercha (f# in the poetry mode -- bars 29-30) preceding the final cadence on the tonic e
Psalms 118:25 - tsinnorit to f# twice: first from the reciting note on the third of the scale (bar 148) and before the subdominant A and again from reciting note f# before the final cadence on e

I see 17 additional examples in the poetry and that's all for the sequence tsinnorit-mercha. Mercha occurs independently also. The ornament preceding the f# may come on the reciting pitch A, g, or f# as seen above, and also on e and even from C (once) as in the following image.

Psalms 72:3 - shalshelet (bar 13) lifts up to tsinnorit on the reciting note C
before the subdominant (bar 15)

The following is an example of tsinnorit joined to the high C. This combination occurs 168 times in my data at present. and the sequence is often tsinnorit followed by C at the beginning of the verse. I had noticed these in previous years (even when I was following Unicode and miscalling it zarqa) because this ornament is often a help to the singer to get to the high note.

Proverbs 8:29 - getting to the high note by way of a tsinnorit.

There are 12 other instances in the poetry of tsinnor (not tsinnorit) preceding f#. There are 180 examples of tsinnor preceding d (galgal), 31 examples preceding e, and 22 preceding C, see Psalms 49:15 below. 

Psalms 49:15 containing both tsinnor and tsinnorit prior to C (mahpakh)

These are clearly different in WLC and its font. Not so clear in mgketer. There are an additional 22 instances of tsinnor with mahpakh following. I don't think the music is significantly different to try to verify these details. Here are the two in copy and paste form -- just highlighting the different placements over the letters.
mgketer.org  שַׁתּוּ֮  וַיִּרְדּ֮וּ 

tanach.us   שַׁתּוּ֮  וַיִּרְדּ֘וּ 

Anyway: I can see that this turn wherever it is placed on the letter often immediately precedes mahpakh which Haïk-Vantoura deciphers as the 6th above the tonic, C in her renderings of the pitches. It makes musical sense no matter which turn is sung. And the similar turn called zarqa and tsinnor may also resolve to d or e in the poetry books.

One more example.

Psalms 16:11 tsinnor preceding C (bar 64)
In the prose, I find 34 tsinnor resolving to e, 563 resolving to B, and 2 resolving to g (both in the 10 words -- special cases). Others may be followed by another ornament. I didn't find any leading to C or f.

It's hard to describe music this way, but one does ask why the designers of this system did not find easier ways to represent these hand signals. Maybe that's the clue -- music notation had not been invented, and the notation we have is limited to what the hands could signify. 

All this sort of detail is not my intent but I am glad I have been forced to pay a little more attention to it. My intent is limited to presentation -- so that the history of the Hebrew Bible that we have may be seen and heard in a musical form that is consistent with what could have been known at the time.

Saturday, 2 August 2025

Learning the data - zarqa and legarmeh

The Hebrew Bible is not an easy task to learn for a non-Hebrew speaker who was not raised in the Hebrew tradition. This is not my 'excuse'. But it is important to know. Because there is a tendency that English readers have to think we know what is there in the Bible even in our ignorance, and when you think you know already, it certainly inhibits learning. (The risk is there for everyone.)

I have checked out all the so-called zarqa accents in the prose books in the Leningrad codex. It turns out that there should not be any. The accent called zarqa does not occur in the prose books in mgketer.org. I corrected all of them to tsinnor. It's a minor change in the melody using Haïk-Vantoura's deciphering key but it does reinforce the separation of the two sets of accents a little more. (Alas, if we eliminate zarqa, what becomes of the zarqa table!)

I have also experimented with inserting the legarmeh into the text in Calibre. Hebrew typing has many problems in that environment. I could define three bugs at least -- can't type inline, can't use the onscreen keyboard, and there are some problems with marking the text and pasting when words are connected by a makaf. I have workarounds. But legarmeh is ignored in the SHV deciphering key. I think of it as an editorial suggestion of a pause -- maybe in some cases a pause of a few billion years of imaginary time. I could illustrate this with Genesis.

Legarmeh in Genesis 2.
Let the performer add an editorial tick where there is a dramatic separation.

I may will include the changes in the 2065 verses that contain legarmeh over 700 or more chapters (in the WLC). I am generating the table entries that could be pasted into the html without error. Also I am generating the html files for the chapters affected. (It only takes about an hour to generate them all from the database. Putting them into the EPUB automatically merges them with the SVG images -- so maybe it is worthwhile doing and would not require too much of my spare time.) These are all support files for the EPUB along with the music files. Legarmeh must be considered musical because music includes the space between the notes. That could be the subject of a book-length treatise. I remember learning rhythm, pulse and agogic both as a young performer and as a middle-aged conductor. 

And legarmeh is also seen as contextual for some other accents. I understand that sequences can be contextual, but how much context, how much read-ahead, how many sequences of accents are there that would have such an effect? And why would anyone have developed a musical notation in this way? I don't know. The appeal of the SHV deciphering key is that it is immediately transparent to a musician. It is her thesis that there are no accents that require the reader to change their mind after reading the next accent. Appeals to context are obvious from the musical line. Nonetheless, legarmeh could be a useful sign and I should probably not have omitted it in the section of my tables with consonants and cantillation accents only. [adding it I will as I write a little about each volume -- the more I see it the better I like it -- Legarmeh is like Talmud in a stroke -- I could use some proof readers if anyone is interested... not that I'm expecting much for nothing but love.]