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!

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.

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.

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.]

Next steps?

The 18 volumes are complete. Some stats: For the 21,535 verses of the Hebrew Bible, I required 34,442 images or about 1.48 images per verse. A few of these are non-musical images and some others are repeated so one could say that the music of the Bible is about 34 thousand lines of music, where each verse no matter how short it is takes at least one line, but may take up to four. A long verse can take up to 4 lines of 30 or so syllables, give or take a few depending on the ornamentation and frequency of changes of reciting note.  Here's an image of the first verses of Chronicles. This is from the PDF format allowing 6 1-line verses per page and requiring 442 pages! 

Sample page from Chronicles

Here's a page with longer verses from 2 Chronicles 25:

Longer verses on page 434 of the Chronicles PDF

It's curious that Calibre has a report that counts the usage of all characters of all types including the details of the cantillation. The details are not viewable with any degree of ease. The accents are there in a column but are unreadable -- too small in the table and lost in translation when converted to Excel. But if you were interested, there are 2,461 mercha accents in Chronicles... and only one zarqa. 2 Chronicles 19:2.

Only one zarqa!

This should not be surprising since Zarqa is uncommon in the prose, but the similarly shaped tsinnor is more frequent. These sign are similar to the modern musical sign for a turn -- a sideways S over the note to be ornamented. This particular verse has two in a row on one word  חֲנָ֘נִי֮. 

Two consecutive turns - rare example 1/2

One would think that such a combination is unique, but there is another word with two such ornaments on it in 2 Samuel 3:8. I suppose that one good turn deserves another. But to get into this historical rarity technically, musically, linguistically, or theologically, would be a distraction at the moment.

Two consecutive turns - rare example 2/2

Surprising what you discover when you have music in front of you. I would not be able to determine these accents from the Aleppo codex. 

A part of rare example 2/2 from Aleppo

You can see I have allowed myself to be distracted when I should be doing the shopping.

Swan song nearly completed

 I started this process in January of getting my data out of the database and into a form where it could provide 

1. a reading environment in epub or pdf form for people learning to read Hebrew text or music. 
2. a musical environment for hearing the words of Scripture in a form that never fails to surprise people who hear about it.
3. the music in a form that provides a starting point for those who want to develop it into songs and stories as illustrated on this blog over the last 15 years.
4. a translation that is open to criticism and changeable in the data manually for anyone who has the software to read the epub.
5. and all of it in a form that can be maintained by anyone with these two programs of publicly available software: Musescore, and Calibre. 

It's nearly done: 15 chapters left in 2 Chronicles to compare with Aleppo. (Perhaps a few hours work). Then about 40 chapters to format the music (a few days work). I hope in the coming months to analyse the types of differences between Aleppo and WLC, but I don't have a schedule or target for this at the moment. There are multiple types of differences. 

• Missing silluqs in WLC (a few, I did not explicitly look for them -- I found maybe 20 in the 7,000 or so verses I looked at).
• Additional silluqs in WLC -- thousands -- I would guess in 50% of the verses I looked at.
• Silluqs that should have been mercha or munach -- maybe 100 or so.
• Silluqs that seemed to move from one word to another.
• Silluqs that seemed to be imitating the same word repeated nearby.
• Silluqs on unstressed or stressed syllables.
• Silluqs on syllables with all sorts of different vowels, a, e, i, o, or u. I can't guess the frequencies.
• Silluqs (frequent) on common words like la (negative), ci (relative) , and mlc (king), Jerusalem and Judah and other less common names.
• Silluqs on prefixes, like l or m, but particularly frequent on h.
• Sometimes in a verse, all will be removed, or one is taken and the other left, or all of them are in both codices.

Maybe I will identify patterns more securely, but it will take a strategy I have not yet figured out.

Also I know I am prone to both error -- old eyes -- and tiredness. I will have missed several silluqs whose presence at the beginning of a verse would not have changed the music, but may be misleading in terms of stress if there is more than one accent on the lexical word. I will also have missed extra silluqs in Aleppo that are not in WLC if they were on verses that I did not look at. It's also possible that I may have deleted where I should not have. 

When I now read through the entire corpus to proof read my epub, I will look and listen for readability, but if I choose a verse to describe in the volume, I must verify that I have not made these blunders of omission.

Here's the percentage of verses with changes I have found - the last 15 chapters of 2 Chronicles will not change the >25% significantly for that book. (Ended up at 28.22%!)

Differences between Aleppo and Leningrad codices --
comparing silluq in WLC on tanach.us with Aleppo on mgketer.org

Next step is for me -- and anyone else who would volunteer -- to take a few examples for each volume and see if we can find distinct musical patterns in the differing volumes. I will probably try these out as blog posts. If you have verses you want, let me know in a comment. I will be back to blogging more regularly.

I would like this final work --that demonstrates the music embedded in the Biblical text-- to be widely available and used -- I don't know that I will give it away, but I hope it will not be lost. Let me know if you are interested. There are now 18 volumes, 929 music scores and supporting files.

Before Jonathan Wheeler disappeared from the web, he gave me -- and likely a lot of other folks -- a DVD of a huge amount of material on Suzanne Haïk-Vantoura (available here and here). I hope I will have done her work a service by making this final work stimulated in me -- partly by him --  since 2010.

If you want to take part in the exploration of the music, please send me an email at drobertmacd at gmail. Please let me know why you are interested and what skills you can bring to bear on the opportunity.

Interim interruptions

What did Haïk-Vantoura ignore or miss in the patterns of the te'amim? 

This negative question -- what did she miss -- is a brief post rather than a chapter in a monograph (I was going to say monologue since public interest may be low). Whether I will write a monograph is moot in any case since I am much more interested in getting the music out to the world in an updatable form. But every question makes me look a little deeper into the mechanisms of the Hebrew Bible and its cantillation tradition.

There are patterns in the copying practice from generation to generation that are causing me some more thinking time. Ten years ago I put out all the music in a raw form, directly mapped from Unicode to MusicXML and published by a Musescore 2 addon that allowed me to make PDF's and MIDI files from the XML without any intervention.

Years later I collected the PDF's into book length PDFs again just available for anyone who wanted to have them. But I can't recommend these copies any more without encouraging the user to check the melodic shape against the Aleppo codex. These copies represent the data that I have from tanach.us but that data has 1000's of additional silluq accents as discussed on this blog in previous posts and  most recently with a scholarly data expert who has some detailed knowledge that questions some aspects of my approach. The questioning is great. 

The questions that have been raised for me are these:

1. What is the usage of silluq, or metheg aka ga'ya? (U-1469). Can it be clarified? Can the changes in the manuscripts over the last 1000 years be decoded? I.e. what is the rationale? I need to start with a documentation of what words and accents are involved in the changes I have made. Are there predictable changes on words where we would not have an explanation to hand? Are there patterns to changes? Are there unnecessary and complex explanations?
2. Are there additional pairs of diacritics that always appear together? It has been suggested to me that legarmeh is sometimes significant in its impact on an earlier accent. I have not included legarmeh in my extracts. I have not paid attention to it. Separation is potentially a function that music would support. Rhythm is fundamental to music. I have never seen a legarmeh that I would have associated exclusively with one or other presumably adjacent accent. I need at least one verse where I can see if there are identifiable clues that predict similar structures in other verses.
3. What is the significance of pre or post positioning of an accent under the text? SHV preserves differences in ornaments over the text for similar accents: e.g. tsinnor and zarqa, qadma and pashta, geresh and azla (a name that SHV does not use -- she calls it geresh as does my Unicode guide). But she makes no distinction between tifha and its slightly displaced pair, d'khi. These being below the text, they each define the same reciting pitch. Possibly,if it were ambiguous, the pre-positioned d'khi would override a presumed continuation of the previous reciting note.
4. What significance is allocated to tifha vs tarha? Tarha is a new name for the accent to me. Tifha and tarha are both the same Unicode value and both the same shape. It makes no sense to give two separate functions to the same accent under a different name. A sign is always a sign of the same thing or who knows what to sing! "the signs always retain the same meaning" (p 30).

There are many names for these signs. SHV used as far as possible the names in the earliest documents, as noted by her American translator from the English book page 496:
 The attentive reader, familiar with synagogue chant or the grammatical rules used by Hebraists, will notice that the list of names given by Haik-Vantoura for the signs (cf. pp. 97-100) lacks several names found in every modern list of names given by Hebraists or cantors: azla, legarmeh, yetiv, debir, gaya or metheg, and so on. This includes the list given by Mayer Lambert in his treatise (cited by Haik-Vantoura as her source). Those graphical signs which have more than one name in the modern lists are precisely those which are found in more than one grammatical placement relative to the verbal text. In effect, modern specialists have chosen certain names (among the 80 or so found in the ancient treatises) which are found in all the early sources; their lists assign one of these names to each "grammatical accent" used in the text. Haik-Vantoura’s key assigns (from the same group of names) one name for each graphical form used; the names she rejects can be accounted for as later, secondary names added to distinguish between the different grammatical placements of certain signs. Since (as is acknowledged) the grammatical rules of the Masoretes are an arbitrary imposition upon the syntax of the verbal text, only a decipherment of the te‘amim independent of the names given to various signs could clearly demonstrate which names are ancestral and which were added by the Masoretes and later scribes. 

It will be clear that I am approaching this text with a very different perspective from traditional cantors and Hebrew students. There are undoubtedly things which I will not see given the music. I am skeptical of some of the modern explanations even when clearly stated now that I have seen the impact on transparency of the text of the music as deciphered. Yet, just as Copernicus and Galileo were talking about the same solar system, it is simpler to hear the music of the spheres thinking heliocentric rather than earth-centric. But perhaps I am just being eccentric.

Still I am curious as to why these changes were made. I will have to develop a form to see the differences and report on them in greater detail. I have a complete 'before image' of my current data. I will figure out a way to isolate the change to the accents and the position of this change in the text and to each affected word.  Many words repeat. 

For instance, לֹא is frequently (but not always) marked with a silluq in the WLC that is not in the Aleppo codex. E.g. in 2 Samuel 2 verse 19: There doesn't seem to me to be any rationale for the insertion of this silluq in the WLC. I can only assume that it was entered for a reason that is unconnected to music. Both are legitimate but the tone of voice changes if the music descends prematurely to the tonic. 

Before: וַיִּרְדֹּ֥ף עֲשָׂהאֵ֖ל אַחֲרֵ֣י אַבְנֵ֑ר וְלֹֽא־נָטָ֣ה לָלֶ֗כֶת עַל־הַיָּמִין֙ וְעַֽל־הַשְּׂמֹ֔אול מֵאַחֲרֵ֖י אַבְנֵֽר

After:   וַיִּרְדֹּ֥ף עֲשָׂהאֵ֖ל אַחֲרֵ֣י אַבְנֵ֑ר וְלֹא־נָטָ֣ה לָלֶ֗כֶת עַל־הַיָּמִין֙ וְעַֽל־הַשְּׂמֹ֔אול מֵאַחֲרֵ֖י אַבְנֵֽר

Similarly כִּי often appears with a silluq in WLC when it is without in Aleppo. Occasionally I have had to add a missing silluq - maybe a half-dozen times. My work has not been exhaustive. I checked for extra silluqs not for missing ones. 

So far I have changed 3,452 verses since 2023, the last time I exported my data for the file in question. Some of these are translation changes. Most of them are simply removal of extra silluqs. All I need to do is figure out how to sort and present this data with word and location context. I'm not set up for this problem at present. While this is a significant number of verses changed, it is a very small percentage of the accents themselves.

 (I feel like I am going through another final exam in my Hebrew journey. I'm nearly finished my project though, 88%. 111 107 chapters to go, 2 Samuel, and Ezra-Nehemiah, and Chronicles. These volumes will supplement courses in Hebrew by giving learners the option of singing the text with underlying harmonic structures defined by the accents under the text. At the same time, they can compare several different ways of seeing and hearing the letters. Reminder, the first volume of 18 is available here.)