Some notes on the prosodics of Tiberian cantillation marks I
Starting from some of the conclusions Dresher (2013) reached in his research into the Tiberian Masoretic cantillation marks, I will put some of his hypotheses to the test and confront them with the data in the codices.Dresher, Bezalel E. “Biblical Accents: Prosody.” In Encyclopedia of Hebrew Language and Linguistics, edited by Geoffrey Khan, 288–96. Leiden: Brill, 2013.
In Dresher (2013) the author interprets the prosodic word as the lowest level in the prosodic hierarchy. Dresher defines the prosodic word as either an orthographic word surrounded by spaces in the unpointed text, or as two or more orthographic words connected by מקף. The prosodic word is the domain at which stress assignment takes place and this implies the rule of tone lengthening, for which I refer to Prince (1975, §1.2), to be productive at this level.Prince, Alan Sanford. “The Phonology and Morphology of Tiberian Hebrew.” Massachusetts Intitute of Technology, 1975. Dresher (2013, p. 289–290) uses the vocalisation of the object marker את to exemplify this.
According to Dresher the direct object is vocalized with סגול when cliticised to another word, be it with מקף or by pronominal suffix. In its unbound form though, it is vocalised with צרי. The difference in vocalisation is, according to Dresher, due to the fact that in the case of the unbound form, the word is assigned stress, which is why the vowel is lengthened from סגול to צרי.
A search of the database yielded the following numbers.The database I used for the research is the Groves-Wheeler Westminster Hebrew Morphology (version 4.20) incorporated in the program Accordance as the HMT-4 module (version 1.8). Of the total of 10065 attestations of the six forms of את included in my research the by far most common form is the one vocalised with סגול and cliticised by מקף to the following word. When את occurs in its unbound form, it is attested almost exclusively with צרי.
Table 1: Number of attestations of the 6 attested forms of the direct object particle.
As is clear from the table though, against the 8504 attestations of אֶת־ there are 2 occurrences of אֵת־, while there are 7 exceptions to the 1171 occurrences of אֵת, namely אֶת without מקף. Even though the numbers prove Dresher’s statements to be correct, it would be useful to elaborate on the 9 exceptions.
The form אֶת
The form vocalised with סגול but without following מקף is attested in the following 7 verses.
Deut. 8:3 וַיַּֽאֲכִֽלְךָ֤ אֶת הַמָּן֙ אֲשֶׁ֣ר לֹא־יָדַ֔עְתָּ
II Kg. 18:27 לֶאֱכֹ֣ל אֶת חֲרֵיהֶם [צוֹאָתָ֗ם] וְלִשְׁתּ֛וֹת אֶת־שֵׁינֵיהֶם [מימֵ֥י] [רַגְלֵיהֶ֖ם] עִמָּכֶֽם׃
Jer. 32:13 וָֽאֲצַוֶּה֙ אֶת בָּר֔וּךְ לְעֵינֵיהֶ֖ם לֵאמֹֽר׃
Ez. 46:12 וּפָ֣תַֽח ל֗וֹ אֶת הַשַּׁ֙עַר֙ הַפֹּנֶ֣ה קָדִ֔ים
Ps. 47:5 יִבְחַר־לָ֥נוּ אֶת־נַחֲלָתֵ֑נוּ אֶ֥ת גְּא֨וֹן יַעֲקֹ֖ב אֲשֶׁר־אָהֵ֣ב סֶֽלָה׃
Ps. 60:2 בְּהַצּוֹת֨וֹ ׀ אֶ֥ת אֲרַ֣ם נַהֲרַיִם֮ וְאֶת־אֲרַ֪ם צ֫וֹבָ֥ה
Prov. 3:12 כִּ֤י אֶ֥ת אֲשֶׁ֣ר יֶאֱהַ֣ב יְהוָ֣ה יוֹכִ֑יחַ
The Aleppo codex cannot be consulted for Deuteronomy 8 as it is no longer extant in the Aleppo codex, whereas the chapter from Jeremiah is heavily damaged and has become unreadable the codex.
It is remarkable that the first 4 attestations, Deuteronomy, 2 Kings, Jeremiah, and Ezekiel do not have a cantillation mark at all in the Leningradensis. Therefore we can assume these four to be textual corruptions.I refer to the commentary of Deut. 8:3 in the BHQ where two other codices than Aleppo are taken up as proof of the textual corruption of the Leningradensis B19a. This seems to be confirmed for the book of Kings as well as for Ezekiel when compared to the Aleppo codex.
In case of the passage from 2 Kg. 18:27 the Aleppo codex does not agree with the codex Leningradensis on the use of the מקף, as can be seen in Figure 1. The מקף between את and חריהם is clearly visible.
The Aleppo codex also disagrees from the Leningradensis with regard to the form attested in Ezekiel 46:12. Like the previous example from the book of Kings, here too the particle את is clitisized to the following word by means of מקף.
The remaining three attestations from Psalms and from Proverbs do have their own cantillation mark. For these three the Leningradensis and the Aleppo codex agree on the pointing as well as on the use of מקף.
Psalm 47:5 clearly has no מקף after the particle את. It does have it’s own accent, מרכא. Remarkably, the first occurrence of the particle את is the cliticised form with מקף.
In Psalm 60:2 the first occurence of את has no following מקף though it does have the well known רפה to indicate it the ת to be pronounced spirantised. The particle does have its own accent, מרכא. The second occurende of את is clitisized by מקף to the following word.
In the case of Proverbs 3:12, the last of the three attestations, the Aleppo codex, again, is in agreement with the Leningradensis, the מקף is absent here as well. The vertical stroke over the consonant ת like in the previous example, is the רפה.
Conclusion on אֶת
I assume 4 out of 7 attested forms of את without מקף and pointed with סגול to be textual errors. In this case the error implies one of three possibilities: a missing מקף, a missing accent, or an accidental mistake in pointing, namely סגול in stead of צרי.
Of these 4, 2 attestations could be compared to the Aleppo codex. From this comparison it has become clear, that in both cases the Aleppo disagrees with the Leningradensis on the use of the מקף. Therefore I will assume the two other passages from Deuteronomy and from the book of Kings, to be textual error with regard to the use of מקף as well, rather than an error in vocalisation.
The remaining 3 attestations from Psamls and Proverbs most likely are free of textual corruption, given the fact that the Aleppo codex agrees with the Leningradensis, contrary to the cases in 2 Kings 18:27 and Ezekiel 46:12.
The form אֵת־
The form אֵת־ is attested only twice. Once in Job 41:26 and once in Esther 9:31.Of course, we cannot consult the Aleppo codex with regard to the book of Esther.
The form attested in Esther is, most likely a textual error. Already in the BHS the apparatus suggests to read with סגול in stead of צרי.
Esth. 9:31 BHS לְקַיֵּ֡ם אֵת־יְמֵי֩ הַפֻּרִ֨ים
In the BHQ the text is pointed with סגול.
Esth. 9:31 BHQ לְקַיֵּ֡ם אֶת־יְמֵי֩ הַפֻּרִ֨ים
The second, and last, attestation of the form אֵת־ is found in the book of Iob.
Iob 41:26 אֵֽת־כָּל־גָּבֹ֥הַּ יִרְאֶ֑ה
The form attested in the book of Iob is shown in Figure 6. Here, the Aleppo agrees with the Leningradensis. The vowel צרי is easily explained by the געיה, present in both the Leningradensis and the Aleppo codex, which means the syllable has the secondary accent within the prosodic word, hence the lengthening of the vowel.
Conclusion on אֵת־
The form אֵת־ is attested only twice one of which most likely constitutes a textual error, while the second attestation of this form can be interpreted as a result of tone lengthening. This means that these 2 passages contain apparent exceptions to Dresher’s rule, which is why we can ignore them.
Final conclusion on אֶת and אֵת־
After this concise research, it turns out that of the sum of 10065 attestations, only 3 constitute real exceptions to the rule Dresher (2013) formulates with regard to the changes of the vocalisation of the direct object particle את due to the following מקף.
All 3 exceptions are of the form אֶת. I do not have an explanation for the pointing of this form, which is not a problem per se as they constitute only .03 percent of the total.