by Rabbi Dr Jeffrey M Cohen
Friday Night address delivered as Scholar-in-Residence
at Hampstead Garden Suburb Synagogue
Chanukah 5767/2006

Ha-nerot hallalu is found for the first time in one of the minor tractates of the Talmud, Masechet Sofrim, which dates back to the eighth century. Jacob ben Asher (1270-1340), author of the Tur, quotes it in full, and then states, “and Rabbi Meir of Rothenburg and my revered father, the Rosh, were accustomed to recite it” (See, also, end of Avudarham’s quote). The implication is that, although it was clearly recited in Israel in the period of Masechet Sofrim, yet its recitation had not subsequently become more universal. Those distinguished scholars of the thirteenth and fourteenth century, referred to by Tur, were clearly attempting to reintroduce its recitation. And the question is, why?

First, what is the purpose of this composition? This is a question that requires to be asked, since, at first sight, it appears as an unnecessary intrusion into the ritual to provide an explanation of its purpose (Hanerot hallalu anachnu madlikin...“We kindle these lights on account of the miracles, the deliverance and the wonders that You performed ...”) Shabbat is a biblical ritual, also characterised by the lighting of candles, but we have no such prescribed formula elucidating the purpose of the Sabbath lighting. Indeed, Ha-nerot hallalu interrupts the ritual (since according to most authorities it is recited as soon as the first candle has been lit) not for the purpose of explaining the lighting ritual, but to provide a prohibition against using the lights for secular purposes. There are no precedents for such a cautionary formulation being injected into a ritual. So why was this ein lanu reshut-formula introduced here in the form of the Hanerot Hallalu?

Now it is our universal practice for the entire family to sing aloud, and with gusto, the most popular Ma’oz tzur, known almost by heart, by even the very young children, on account of its lively traditional tune. It was composed by an Ashkenazi poet of the thirteenth century by the name of Mordechai ben Isaac, and, according to A. Z. Idelsohn, its popular tune became universal from the sixteenth century.

It is important to bear in mind these dates in order to answer the question we have posed. For a moment let us go back to the period before the 16th century, before Ma’oz tzur was invested with its universal popularity and before its melody made it such a favourite composition. Those who did recite it would certainly have required to hold their text to the light in order to read its words. The instinctive tendency would have been to employ the Chanukah lights, next to which they were standing, in order to illumine the page. Now, bearing in mind that the attempts to reintroduce and popularize the recitation of Ha-nerot hallalu were made (on the evidence of Tur) not long after the composition and dissemination of Ma’oz tzur (13th cent.), we may conjecture that these situations are inter-linked. In other words, the increasingly popular practice of reciting the Ma’oz tzur called for a cautionary prelude, in the form of Ha-nerot hallalu, advising people not to make any use of their Chanukah lights in order to view the still unfamiliar text of the recently-introduced Ma’oz tzur.

Bearing in mind that we are still in the pre-printing age, any text of Ma’oz tzur would have been hand-written, and probably in a small script to utilize as little parchment as possible. Hence the extra probability of all the family members being tempted to pour over it by the light of the Chanukiah in order to read the hymn. And hence the necessity of the cautionary Ha-nerot hallalu composition.

*** *** ***

For the second part of my talk I wish to refer to numerology, which was an important facet of liturgical composition. Now, the Be’er Heitev (676:4 (3) ) quotes R. Solomon Luria (MHRSHL) to the effect that there are 36 words in the Ha-nerot hallalu composition, corresponding to the total number of Chanukah lights kindled during the eight nights. The Sha’ar ha-Tziyyun (sec.13) quotes Eliyahu Rabba to the effect that we should not therefore include the word heim, in the phrase ha-nerot hallalu kodesh heim since the addition of that word gives a total of 37, which destroys its numerical structure and significance. (1)

I’m sure you’ll agree that that sounds most interesting, and adds considerably to the significance of Ha-nerot hallalu. The only problem is that it is not true. For, if we consult the various versions that we possess of this composition, going back to the 8th cent. Masechet Soferim, and including all the printed versions of the Siddur, there isn’t a single version containing so few words.

The version of Soferim has 42 (or, according to the Vilna Gaon, 41) words, that of David Avudarham (illustrious 14th cent. liturgical authority) has 40, as does Singer’s. The Siddur of Ya’akov Emden has 48 (or 50 if we include the bracketed words v’al nifl’otekha), Art Scroll has 49 and Luria’s Siddur HaAri has 51. No version has a mere 36 words. This prompted the Arukh Ha-Shulchan, R. Yechiel Michel Epstein of Novardok to observe: V’zeh she’amar sheyeish b’khan lamed vav teivot, k’minyan nerot Chanukah, lo ukhal lekhavein ha-cheshbon, ‘And that statement that there are 36 words, corresponding to the number of Chanukah lights – I am unable to work out how that is arrived at.’

Now, the task I have set myself – bimchilat kevodo – is to solve this conundrum; and it occurred to me that there was only one textual device that could provide a clue.

You will have noticed that, in our Singer’s and the newly issued Authorised Daily Prayer Book (or Sacks’ Siddur), though not in Art Scroll, some of the words are linked by a hyphen, such as, al yedei and v’khol shmonat, or, al nisekha and v’al yeshu’atekha. The purpose of these linked words is to enable us to construe them as a single conceptual unit. Now, these hyphenated words were arbitrarily created by scribes and printers; and my thesis is that Mishnah Berurah’s tradition of 36 words was arrived at by his having regarded each of the linked words in his Siddur tradition as but one composite expression – one word instead of two.

In Singer’s, even if we regard the linked couplets as one word, we still do not get to a total of 36 words (We actually reach only 35). But, by a process of linking and/or de-linking some of those related phrases in the composition, it might well be possible to reconstruct the version followed by Mishneh Berurah, and recreate the magical sum of 36 words that he regard as of such numerological importance.

*** *** ***

Now there is only one Siddur, that I am aware of, that allows for such a 36-word reconstruction; but before I reveal its identity, I should explain how it is that we have some versions, that I have already highlighted, that have so many more words than 36.

Those who have studied critical editions of classical sources will not be surprised by the many variant readings – shinuyei nuscha’ot – that bound in different manuscripts of the same hand-written text. These variants occur for a variety of reasons, mainly carelessness in scribal copying. The copyist whose concentration lapses for an instant may easily repeat the word he has just written (dittography), or inadvertently omit an adjacent letter (haplography), or even worse, homoeteleuton, where the copyist (who may have just had a break for his lunch, returns and) having noted the phrase with which he ended off, he looks at the master copy, sees the identical phrase and carries on copying from there, not realising that that same phrase occurs again four or five lines below where he left off, and that he had continued after its repeat occurrence, not the first, and had consequently omitted four or five lines of the master text.

Another, quite common, error is that of transposition, whereby the copyist, transcribing part of a phrase with which he is very familiar from another more popular source, inadvertently incorporates the continuation of that familiar source into the text he is copying. I believe that this is precisely what is responsible for the conflated versions of Ha-nerot hallalu that appear in such Siddurim as those of Ya’akov Emden (with 48-50 words), Art Scroll (49), Otzar Ha-Tefillot (50), Siddur Ha’Ari (51), etc. The phrase that they have most obviously transposed is [she’asita la’avoteinu] bayyamim ha-heim bazman ha-zeh. And we know precisely where the author of that version inadvertently copied it from, namely the Al ha-nisim composition: [Al ha-nisim v’al ha-porkan v’al ha-gevurot v’al ha-teshu’ot v’al hanifla’ot] she’asita la’avoteinu bayamim ha-heim bazman ha-zeh. It is obvious that a popular master-copier of Hanerot hallalu was so familiar with the Al Ha-nisim that, after the phrase al ha-teshu’ot v’al ha-nifla’ot sha’asita la’avoteinu, he just used the Al Ha-nisim continuation bayamim ha-heim bazman ha-zeh. From his version the early printers perpetuated and disseminated that error. (That it was extraneous is obvious from the continuation which actually gives no sense, since it now refers to miracles, wonders, etc., that you wrought for our fathers bayamim haheim bazman ha-zeh, -‘ in those days and at this time’ – al yedei kohanekha ha-kedoshim, ‘by your holy priests’. It goes without saying that we are unaware of any miracles bazman ha-zeh - in the post-Chanukah period – that were wrought by priests.) So here are four words that have inflated the number of words in the original Al Ha-nisim.

Again, through the influence of the phrase v’al ha-milchamot in the Al ha-nisim, that identical phrase inadvertently crept into the Ha-nerot Hallalu.

In addition there is the unnecessary word heim, in the phrase, ha-nerot hallalu kodesh heim, which is not found in the 14th century version of Avudarham. Nor does he have the word [leshimkha] ha-gadol, which is also clearly super-numerary.

The word ha-gadol also slipped readily from the copyist’s quill, after the word leshimkha. This is because that identical phrase, leshimkha ha-gadol, occurs in the final line of the Bimei Matityahu composition: V’kav’u shemonat yemei Chanukah eilu l’hodot ul’hallel leshimkha ha-gadol. We can see how an inattentive scribe could so easily borrow from one Chanukah context to another. And so it came about that, with the introduction of printing, all those variant versions, inflated by those several phrases and words, found their way into the printed editions and became the accepted nusach of the communities that used those prayer books.

*** *** ***

In conclusion, I promised to reveal an edition which, albeit unintended, gets closest to Mishnah Berurah and enables us to reconstruct with ease how we arrive at a version containing 36 words, or, more accurately, verbal units.

Believe it or not, it is the new edition of Chief Rabbi Sacks. But not, I hasten to add, if we merely count the words as they stand. No, to arrive at a total of 36, we have to adopt that well-attested approach of counting each hyphenated couplet as but one verbal unit. And this, I submit, is the solution to the mystery that baffled the illustrious Arukh Ha-shulchan. QED! V’Chanukah same’ach!

*** *** ***

(1) MHRSHL actually states that ‘there are 36 words, excluding the first two words, corresponding to the total number of lights.’ It goes without saying that no version of Ha-nerot Hallalu omits those first two words, and it borders on the absurd to read any numerical significance in a text that omits them!