parshat Chayyei Sarah: Why a Shalshelet?

The Gabbai, over at a new blog, "These Guys Are No GH," reposts a dvar Torah from Rabbi Sacks about the trup of shalshelet, who cites Ibn Caspi, who says it is means to convey psychological state of uncertainty and indecision (this based on its musical notes and its orthographic sign).

The shalshelet is an unusual note. It goes up and down, up and down, as if unable to move forward to the next note. It was the 16th century commentator Rabbi Joseph Ibn Caspi (in his commentary to Bereishith 19:16) who best understood what it was meant to convey, namely a psychological state of uncertainty and indecision. The graphic notation of the shalshelet itself looks like a streak of lightning, a "zigzag movement" (tenuah me'uvetet), a mark that goes repeatedly backwards and forwards. It conveys frozen motion - what Hamlet called "the native hue of resolution sicklied o'er by the pale cast of thought" - in which the agent is torn by inner conflict. The shalshelet is the music of ambivalence.

He then applies this to this week's parsha, to explain Eliezer's uncertainty, weaving it in to the fabric of the midrash which has Eliezer hope to be Avraham's heir, and his own daughter to marry Yitzchak.

Perhaps. I always like to find the connection between the trup and the midrash. However, we should also understand the mechanics of why the shalshelet should appear. I will cite Wickes and then argue a bit on him.

Before proceeding, a plug: Wickes book, Two Treatises On The Accentuation if the Old Testament, is a very important book to read and own. The link above brings you to Amazon, which has three books left. Other sites have this book for a bit cheaper. Google Book Search has this book online, such that you can search and view any page, and also has links to buy the book.
William Wickes, on page 85 of the second of his Two Treatises On The Accentuation if the Old Testament, a book on trup (plug: there are three left at Amazon), we read:


A little background is in order. The (disjunctive) trup subdivides the pasuk in two. Each subdivision is then divided in two again and again and again, as long as there are three words or more left to the phrase. The place where a phrase is subdivided is based on various syntactic features. The specific trup used to mark a subdivision is based on the trup which stands at the end of the phrase as well as the number of words until the end of a subphrase.

Thus, for example, the etnachta is the trup at the end of a pasuk. If we wish to subdivide a phrase that ends in an etnachta, we will use syntactic rules to find the place in the phrase where we should divide, and we will mark that word with a disjunctive trup mark. There are two possible trup marks (or at least, common ones): tipcha and zakef. If the word is immediately before the etnachta (thus, one word away), we use a tipcha. Two words away, either tipcha or zakef, but most often tipcha. Three words away or more, zakef (gadol or katon).

And so, when we see specific trup, it is not just musical, but also serves to mark logical and syntactic divisions. That the shalshelet looks and sounds the way it does may or may not be relevant to the meaning.

So what of shalshelet? First, we must turn to segolta, which looks like an upside-down segol written over the word. The segolta is really a modified zakef. When the division of the subphrase is four words away, we start seeing the segolta, though very rarely. We see it more common on the fifth word, and on from there. Wickes writes (in footnote 7 on page 73) that on the seventh word away from the end of the phrase (marked by etnachta), there is a preponderance of zakef; at the eight words away, the zakef and segolta are equally common; on the ninth, zakef is rare and segolta is expected, and tenth, zekef is even more rare (only five or six times); and more than ten words away, Wickes regards any zakef as a error (!) and corrects to be segolta.

The thing about segolta, as you will see if you read the text in the image I placed above, it that it must have the servus called zarka before it. Thus, when segolta occurs in the beginning of the verse, it transforms into shalshelet.

We can then see why shalshelet is so rare - only occurring 7 times in the prose books of Tanach. To have such long pesukim is rare. To have long pesukim which are divided towards the beginning, such that the division of 8, 9, 10 +, is even rarer. To have that division be at the very beginning of the verse, on the first word, such that we cannot place a zarka there, is even rarer.

Now, Wickes says that a zakef could appear in any of these instances. To a degree, he is correct. Zakef may appear in any of these instances, but the farther one gets from the etnachta at the end of the phrase, the more prone they are to wish to place a segolta. Wickes would like them to have chosen zakef because segolta would not work, because of the lack of a word proceeding. But whatever caused them to wish to choose segolta in many other instances would also cause them to prefer it here, and from that segolta, transform it to a shalshelet.

This decision may also have been motivated by musical concerns, as he suggests, or by some special meaning, a "fanciful Midrashic explanation," which we have lost. Perhaps. But I see no reason to be forced into such an explanation. The distance from etnachta, combined with the word being the first in the verse, is sufficient cause for me. There might also be some other feature that we are not examining, such as number of discrete ideas, or number of syllables, or a major shift in semantic meaning between the first word and the rest of the verse, or a shift in actor that would otherwise not be as apparent, or the complexity of the trup that follows, that could account for a preference of segolta/shalshelet.

As such, perhaps there is no need to find reasons for the shalshelet, as Ibn Caspi and Rabbi Sacks do - though they don't even discuss the syntactic reasons for the shalshelet.

Wickes gives examples of what he considers fanciful midrashic explanations, and Ibn Caspi and Rabbi Sacks could also be considered to have given such fanciful midrashic explanations. It saddens me when people do not give midrashic explanations their due - Wickes here considers it unimportant, and no serious loss, preferring instead the concrete science of syntax. Yet sytax informs sematics, and vice versa, and both are also worthy of study. Of course, I prefer where the midrash is not based on the orthography of the sign of the specific trill, but rather the unique place they chose to divide, but still...

Let us now consider the seven occurrences mentioned by Wickes.

The first is in parshat Vayera, in Bereishit 19:16, on the word וַיִּתְמַהְמָהּ:

טז וַיִּתְמַהְמָהּ--וַיַּחֲזִיקוּ הָאֲנָשִׁים בְּיָדוֹ וּבְיַד-אִשְׁתּוֹ וּבְיַד שְׁתֵּי בְנֹתָיו, בְּחֶמְלַת ה עָלָיו; וַיֹּצִאֻהוּ וַיַּנִּחֻהוּ, מִחוּץ לָעִיר.

16 But he lingered; and the men laid hold upon his hand, and upon the hand of his wife, and upon the hand of his two daughters; the LORD being merciful unto him. And they brought him forth, and set him without the city.

The trup is:
טז וַֽיִּתְמַהְמָ֓הּ ׀ וַיַּֽחֲזִ֨יקוּ הָֽאֲנָשִׁ֜ים בְּיָד֣וֹ וּבְיַד־אִשְׁתּ֗וֹ וּבְיַד֙ שְׁתֵּ֣י בְנֹתָ֔יו בְּחֶמְלַ֥ת יְהוָ֖ה עָלָ֑יו וַיֹּֽצִאֻ֥הוּ וַיַּנִּחֻ֖הוּ מִח֥וּץ לָעִֽיר

The etnachta is on the word עָלָ֑יו. Thus, the division of the first half of the pasuk, on logical grounds, is:
וַֽיִּתְמַהְמָ֓הּ ׀
וַיַּֽחֲזִ֨יקוּ הָֽאֲנָשִׁ֜ים בְּיָד֣וֹ וּבְיַד־אִשְׁתּ֗וֹ וּבְיַד֙ שְׁתֵּ֣י בְנֹתָ֔יו בְּחֶמְלַ֥ת יְהוָ֖ה עָלָ֑יו

But he lingered |
and the men laid hold upon his hand, and upon the hand of his wife, and upon the hand of his two daughters; the LORD being merciful unto him.

The word וַֽיִּתְמַהְמָ֓הּ is therefore 10 words away from the etnachta - 11 if we were to separate וּבְיַד־אִשְׁתּ֗וֹ into two words (though we do not ever separate the words when making this calculation). Wickes himself wrote that zakef as opposed to segolta is extremely rare 10 words away from etnachta, so it would not be an easy matter to substitute a zakef here, rather than transforming it into a shalshelet. (And 11 - he would have labelled it an error!)

The second example is in parshat Chayyei Sarah, in Bereishit 24:12: Eliezer says:

יב וַיֹּאמַר--ה אֱלֹהֵי אֲדֹנִי אַבְרָהָם, הַקְרֵה-נָא לְפָנַי הַיּוֹם; וַעֲשֵׂה-חֶסֶד, עִם אֲדֹנִי אַבְרָהָם.

12 And he said: 'O LORD, the God of my master Abraham, send me, I pray Thee, good speed this day, and show kindness unto my master Abraham.

The trup:

יב וַיֹּאמַ֓ר ׀ יְהוָ֗ה אֱלֹהֵי֙ אֲדֹנִ֣י אַבְרָהָ֔ם הַקְרֵה־נָ֥א לְפָנַ֖י הַיּ֑וֹם וַֽעֲשֵׂה־חֶ֕סֶד עִ֖ם אֲדֹנִ֥י אַבְרָהָֽם׃

The etnachta, which splits the verse in twain, is on הַיּ֑וֹם. The subdivision of this first phrase is thus:
וַיֹּאמַ֓ר ׀
יְהוָ֗ה אֱלֹהֵי֙ אֲדֹנִ֣י אַבְרָהָ֔ם הַקְרֵה־נָ֥א לְפָנַ֖י הַיּ֑וֹם

And he said |
'O LORD, the God of my master Abraham, send me, I pray Thee, good speed this day...'

The first word is thus 7 words away (8 if we would divide הַקְרֵה־נָ֥א). So here, there is a preponderance of zakef, though segolta (and thus shalshelet) is allowed but not forced.

The third example is in Bereishit 39:8, in parshat Vayeshev:

ח וַיְמָאֵן--וַיֹּאמֶר אֶל-אֵשֶׁת אֲדֹנָיו, הֵן אֲדֹנִי לֹא-יָדַע אִתִּי מַה-בַּבָּיִת; וְכֹל אֲשֶׁר-יֶשׁ-לוֹ, נָתַן בְּיָדִי.

8 But he refused, and said unto his master's wife: 'Behold, my master, having me, knoweth not what is in the house, and he hath put all that he hath into my hand;

The trup:

ח וַיְמָאֵ֓ן ׀ וַיֹּ֨אמֶר֙ אֶל־אֵ֣שֶׁת אֲדֹנָ֔יו הֵ֣ן אֲדֹנִ֔י לֹֽא־יָדַ֥ע אִתִּ֖י מַה־בַּבָּ֑יִת וְכֹ֥ל אֲשֶׁר־יֶשׁ־ל֖וֹ נָתַ֥ן בְּיָדִֽי׃

The etnachta is on מַה־בַּבָּ֑יִת. Thus, the division of the subphrase would be:

וַיְמָאֵ֓ן ׀
וַיֹּ֨אמֶר֙ אֶל־אֵ֣שֶׁת אֲדֹנָ֔יו הֵ֣ן אֲדֹנִ֔י לֹֽא־יָדַ֥ע אִתִּ֖י מַה־בַּבָּ֑יִת

But he refused |
and said unto his master's wife: 'Behold, my master, having me, knoweth not what is in the house...

The word וַיְמָאֵ֓ן is eight words away from etnachta, at which point zakef and seglota (--> shalshelet are about equal). So it is certainly no surprise that we would have a shalshelet here, on the first word. (Plus, we could separate 3 paired words, making it 11 words away.)

The fourth example, in Vayikra 8:23:

כג וַיִּשְׁחָט--וַיִּקַּח מֹשֶׁה מִדָּמוֹ, וַיִּתֵּן עַל-תְּנוּךְ אֹזֶן-אַהֲרֹן הַיְמָנִית; וְעַל-בֹּהֶן יָדוֹ הַיְמָנִית, וְעַל-בֹּהֶן רַגְלוֹ הַיְמָנִית.

23 And when it was slain, Moses took of the blood thereof, and put it upon the tip of Aaron's right ear, and upon the thumb of his right hand, and upon the great toe of his right foot.

The trup:
כג וַיִּשְׁחָ֓ט ׀ וַיִּקַּ֤ח מֹשֶׁה֙ מִדָּמ֔וֹ וַיִּתֵּ֛ן עַל־תְּנ֥וּךְ אֹֽזֶן־אַהֲרֹ֖ן הַיְמָנִ֑ית וְעַל־בֹּ֤הֶן יָדוֹ֙ הַיְמָנִ֔ית וְעַל־בֹּ֥הֶן רַגְל֖וֹ הַיְמָנִֽית׃

The etnachta is on הַיְמָנִ֑ית. Thus, the first half of the verse should be subdivided:
וַיִּשְׁחָ֓ט ׀
וַיִּקַּ֤ח מֹשֶׁה֙ מִדָּמ֔וֹ וַיִּתֵּ֛ן עַל־תְּנ֥וּךְ אֹֽזֶן־אַהֲרֹ֖ן הַיְמָנִ֑ית

And it was slain |
and Moses took of the blood thereof, and put it upon the tip of Aaron's right ear ...

The word וַיִּשְׁחָ֓ט ׀ is 7 words away from etnachta, though 9 away if we separate the paired words.

The first example is in Yeshaya 13:8:

ח וְנִבְהָלוּ--צִירִים וַחֲבָלִים יֹאחֵזוּן, כַּיּוֹלֵדָה יְחִילוּן; אִישׁ אֶל-רֵעֵהוּ יִתְמָהוּ, פְּנֵי לְהָבִים פְּנֵיהֶם.

8 And they shall be affrighted; pangs and throes shall take hold of them; they shall be in pain as a woman in travail; they shall look aghast one at another; their faces shall be faces of flame.

The trup:

ח וְֽנִבְהָ֓לוּ ׀ צִירִ֤ים וַֽחֲבָלִים֙ יֹֽאחֵז֔וּן כַּיּֽוֹלֵדָ֖ה יְחִיל֑וּן אִ֤ישׁ אֶל־רֵעֵ֨הוּ֙ יִתְמָ֔הוּ פְּנֵ֥י לְהָבִ֖ים פְּנֵיהֶֽם

The etnachta is on יְחִיל֑וּן; the phrase is then subdivided:
וְֽנִבְהָ֓לוּ ׀
צִירִ֤ים וַֽחֲבָלִים֙ יֹֽאחֵז֔וּן כַּיּֽוֹלֵדָ֖ה יְחִיל֑וּן

The word וְֽנִבְהָ֓לוּ is thus only five words away.

The sixth example is in Amos 1:2:

ב וַיֹּאמַר--ה מִצִּיּוֹן יִשְׁאָג, וּמִירוּשָׁלִַם יִתֵּן קוֹלוֹ; וְאָבְלוּ נְאוֹת הָרֹעִים, וְיָבֵשׁ רֹאשׁ הַכַּרְמֶל. {פ}

2 And he said: the LORD roareth from Zion, and uttereth His voice from Jerusalem; and the pastures of the shepherds shall mourn, and the top of Carmel shall wither. {P}

The trup:
ב וַיֹּאמַ֓ר ׀ יְהוָה֙ מִצִּיּ֣וֹן יִשְׁאָ֔ג וּמִירֽוּשָׁלִַ֖ם יִתֵּ֣ן קוֹל֑וֹ וְאָֽבְלוּ֙ נְא֣וֹת הָֽרֹעִ֔ים וְיָבֵ֖שׁ רֹ֥אשׁ הַכַּרְמֶֽל׃
The etnachta is on קוֹל֑וֹ. The phrase is subdivided:

וַיֹּאמַ֓ר ׀
יְהוָה֙ מִצִּיּ֣וֹן יִשְׁאָ֔ג וּמִירֽוּשָׁלִַ֖ם יִתֵּ֣ן קוֹל֑וֹ

The word וַיֹּאמַ֓ר is thus 6 away from etnachta.

The seventh occurrence, in Ezra 5:15:

טו וַאֲמַר-לֵהּ--אלה (אֵל) מָאנַיָּא, שֵׂא אֵזֶל-אֲחֵת הִמּוֹ בְּהֵיכְלָא דִּי בִירוּשְׁלֶם; וּבֵית אֱלָהָא, יִתְבְּנֵא עַל-אַתְרֵהּ. {ס}

15 and he said unto him: Take these vessels, go, put them in the temple that is in Jerusalem, and let the house of God be builded in its place. {S}

The trup:

טו וַֽאֲמַר־לֵ֓הּ ׀ אלה (אֵ֚ל) מָֽאנַיָּ֔א שֵׂ֚א אֵֽזֶל־אֲחֵ֣ת הִמּ֔וֹ בְּהֵֽיכְלָ֖א דִּ֣י בִירֽוּשְׁלֶ֑ם וּבֵ֥ית אֱלָהָ֖א יִתְבְּנֵ֥א עַל־אַתְרֵֽהּ׃ {ס

The etnachta is on בִירֽוּשְׁלֶ֑ם. The phrase is divided:

וַֽאֲמַר־לֵ֓הּ ׀
אלה (אֵ֚ל) מָֽאנַיָּ֔א שֵׂ֚א אֵֽזֶל־אֲחֵ֣ת הִמּ֔וֹ בְּהֵֽיכְלָ֖א דִּ֣י בִירֽוּשְׁלֶ֑ם

and he said unto him|
Take these vessels, go, put them in the temple that is in Jerusalem...

The "word" וַֽאֲמַר־לֵ֓הּ is thus eight away from the end.