The Chabad website has an edition of the Hebrew Bible called The Complete Tanach with Rashi (CTR). (See my document, “On the Provenance of Chabad’s CTR.”)
This CTR is certainly the weirdest Hebrew Bible on the web, and possibly the worst. Whether it is the worst depends somewhat on (1) taste and (2) what purpose CTR would be put to.
For example, this CTR, though full of what most people would consider errors, might still be correct in some cases where other editions are wrong. In other words, its very weirdness, though usually a negative, might in some cases be a positive.
I will not comprehensively review CTR. Instead, I will review its Psalm 32. If its Psalm 32 is representative, my review should give a good idea of CTR as a whole. At a minimum, my review is likely to give a good idea of CTR’s poetic books, since some of CTR’s issues are specific to the poetic accents.
First, I will discuss the ways in which CTR’s Psalm 32 uses Unicode in ways that differ from other web editions. Then I will discuss the more substantive ways in which CTR’s Psalm 32 differs from various other editions of Psalm 32, both on the web and on paper.
In its Psalm 32 (and perhaps in its entire Tanakh), CTR is restricted to the code points in Unicode 2.0. There are three code points that were introduced later that are used in some editions of Psalm 32:
In the first two of the three cases, the fact that CTR does not use the newer code point causes no problems. Indeed many editions, even quite recent ones, are like CTR in that they do not make the following distinctions:
Both distinctions were introduced in Unicode 4.1, although ATNAH HAFUKH was introduced in a way that makes its adoption difficult, even to this day. See my proposal to the Unicode SEWG.
The third post-2.0 code point that CTR might have used is HOLAM HASER FOR VAV. In this case, the fact that CTR does not use it causes a problem. It causes a problem because CTR does not merely decline to make the distinction that this code point enables. Rather, CTR uses a nonstandard workaround to try to make this distinction. This workaround may work (accidentally or intentionally) in some fonts, but not in all fonts.
The workaround is the sequence ‹ZWJ, HOLAM›. For example, to encode the ḥolam ḥaser on the letter vav in עוֺן in Psalm 32:5 (fully: עֲוֺ֖ן ), instead of HOLAM HASER FOR VAV, CTR uses ‹ZWJ, HOLAM›. In some fonts, this may accidentally have something like the desired appearance, since the ZWJ may “confuse” the font into putting the HOLAM fully after the vav. But in other fonts, the HOLAM will still appear directly over the vav, as it would if there were no ZWJ.
The CTR edition even declines to avail itself of distinctions that were available in Unicode 2.0. For example, CTR does not distinguish tsinnor from tsinnorit, using only the ZARQA code point for both. To be fair, the names of the ZARQA and ZINOR code points were botched, and so were their annotations, in my opinion. That could explain why CTR uses only ZARQA. According to the annotation for ZARQA, it should be used for both tsinnor and tsinnorit if an above-center (rather than above-left) placement of tsinnor is desired. In my opinion, this annotation encourages the use of Unicode to make a typographic choice that should be left up to fonts, not “baked in” to the encoding. (See my proposal to the Unicode SEWG.)
In addition to tsinnor and tsinnorit, there are three other pairs of accents for which CTR declines to avail itself of distinctions that were available in Unicode 2.0:
Each of these three pairs consists of a prepositive accent and an impositive “lookalike.” Sometimes CTR makes these distinctions in nonstandard ways, and sometimes it makes no distinction at all. (See my document, “Pre-vowel Accents in Chabad’s CTR.”)
Now let’s move on from encoding (Unicode) differences to more substantive ones. I will compare CTR primarily to the edition called Miqra al pi ha-Masorah (MAM). Here’s an overview of where CTR and MAM differ in Psalm 32, showing what MAM has in those places:
First let’s look at the words in which CTR differs not only from MAM but also from JP (Judaica Press Mikraoth Gedoloth: Psalms: Volume 1). This is a reasonable starting point since JP is, among the sources I have looked at, the one that most closely resembles CTR. For each word, I will also show what is present in KCT (Koren Classic Tanakh) for reasons that will be explained below. Here is the first word where CTR differs from both MAM and JP:
| CTR, KCT | נשוי־פֶּ֝֗שַׁע | revia on pe |
| JP | נשוי־פֶּ֝שַׁ֗ע | revia on shin |
In both CTR and in JP, the accent is revia mugrash. But CTR and JP differ in their placement of revia. Perhaps CTR is a transcription of JP later altered to conform more closely to a Koren Tanakh in some places. I base this speculation on documentation found on the Soncino Classics Collection CD-ROM. This CD-ROM may be related to CTR. (See my document, “On the Provenance of Chabad’s CTR.”). Here is the relevant quote from the CD-ROM documentation:
The text of the Tanach is based on the 1895 Warsaw edition of the Mikraot Gedolot which has been carefully compared with other versions such as the Jerusalem Koren Tanach. In some instances (involving mainly Ken [sic; Kere intented?] and Ketiv and Chaser and Malay) the text has been modified to conform to the Koren edition. However, the responsibility for such decisions is solely that of the CD ROM publishers.
My speculation is consistent with the contents of KCT here since KCT matches CTR here, having revia mugrash with revia on pe. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | אַֽשְֽׁרֵ֥י אדם | געיה on shin |
| JP | אַֽשְׁרֵ֥י אדם | no געיה on shin |
| KCT | אַֽשְֽׁרֵי־אדם | see discussion below |
Here, KCT is quite different from CTR: there is only a tsere on resh (no merkha) and there is a maqaf joining אשרי to אדם (rather than a space separating them). But the fact that both CTR and KCT have a געיה on shin may be significant. I speculate that the געיה on shin in CTR was added to make a transcription of JP conform more closely to KCT.
I think this געיה does not “go with” the rest of CTR’s pointing of this word. I think this because this געיה is a mid-word געיה with shewa, and that construction is used in quite limited contexts that are not analogous to CTR’s use here. See my document, “Mid-word געיה with Shewa.” As we shall see, this is one of three cases where CTR seems to have made an infelicitous mix of the pointing of JP and KCT. See לא and יומם below. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | לֹֽא יחשב | געיה with space (!) |
| JP | לֹֽא־יחשב | געיה with maqaf |
| KCT | לֹ֤א יחשב | mahapakh with space |
This is an error in CTR: לא needs an accent and/or a maqaf. It is widely agreed to be illegal to have only a געיה on an independent word. Here CTR’s error may just be a careless error. But, it may be an error caused by another effort to make CTR conform more closely to KCT. Here both KCT and CTR have a space separating לא from יחשב, although KCT has a mahapakh rather than a געיה on לא. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | י֘וֹמָ֣ם | munaḥ with tsinnorit (!) |
| JP | י֘וֹמָ֤ם | mahapakh with tsinnorit |
| KCT | יוֹמָ֣ם | munaḥ |
This is an error in CTR. It is widely agreed to be illegal to combine tsinnorit and munaḥ. (A tsinnorit always and only precedes either a merkha or a mahapakh.) This could be the result of another effort to make CTR conform more closely to KCT. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | וָ֨לַיְלָה׀ | azla on vav (!) |
| JP image |
|
azla floats between vav and lamed |
| JP charitable | וָלַ֨יְלָה׀ | floating azla read as if on lamed |
| KCT | וָלַיְלָה֮ | tsinnor |
(I assume that the vertical bar after ולילה is a legarmeh rather than a paseq. As such, I include it as part of the word.)
(By “JP charitable,” I mean the (charitable) interpretation of JP, where the floating azla is assigned to the lamed.)
The location of azla in CTR (on vav) is likely an error. This error may stem from mistranscribing the floating azla in JP. It makes more sense to assign the floating azla to lamed than to vav. The only analogy I find in all of Tanakh is munaḥ on vav in Exodus 36:38 וְאֶת־וָ֣וֵיהֶ֔ם . The analogy is distant because munaḥ is not the primary accent on that word.
Here with ולילה, for the first time, conformance with KCT cannot explain why CTR differs from (the charitable interpretation of ) JP. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | וַֽעֲ֘וֺנִ֤י | ayin (!) holds tsinnorit |
| JP image | 
|
vav #1 (!) holds tsinnorit “above tsere” on nun (!) |
| JP charitable | וַ֘עֲוֺנִ֤י | “above tsere” read as ḥ. ḥ. on vav |
| KCT | וַֽעֲוֺ֘נִ֤י | וֺ֘ = וֺ+ו֘ |
(By “JP charitable,” I mean the (charitable) interpretation of JP, where the “above tsere” on nun is interpreted as a ḥolam ḥaser on vav plus an ignored dot.)
The location of tsinnorit in CTR (on ayin) is an error. It is widely agreed that no accent ever appears on a vocal shewa that is not initial. (This is true regardless of whether that vocal shewa is notated as a simple shewa or a ḥataf shewa, i.e. a ḥataf vowel). (See my document, “Tsinnorit & Oleh on Initial Vocal Shewa.”)
In JP, there are two dots above the nun, almost like an “above tsere,” if such a thing existed. Perhaps the first of these two dots is the ḥolam ḥaser dot of the vav, placed too far to the left. But the second dot is inexplicable. A far less serious issue is that the ḥataf pataḥ of the ayin seems unnecessarily early. Such early placement is needed when the “tail” of the ayin is a descender, but the tail of this ayin is not a descender: it is tucked to the side.
Here, as with ולילה, conformance with KCT cannot explain why CTR differs from (the charitable interpretation of ) JP. Though one could say that the placement of tsinnorit on ayin in CTR compromises between vav #1 (JP) and vav #2 in (KCT), I doubt that this is why tsinnorit appears on ayin in CTR. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | נָשָׂ֤אתָ | just mahapakh (no tsinnorit ) |
| JP, KCT | נָ֘שָׂ֤אתָ | mahapakh with tsinnorit |
Here, as with ולילה and ועוני, conformance with KCT cannot explain why CTR differs from JP. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | אל־תִּהְי֨וּ׀ | azla on yod |
| JP | אל־תִּהִ֨יוּ׀ | azla on he (!) he has ḥiriq (!) |
| KCT | אל־תִּֽהְי֤וּ׀ | mahapakh on yod |
(I assume that the vertical bar after תהיו is a legarmeh rather than a paseq. As such, I include it as part of the word.)
Here JP has a pretty wild pointing of this word. Here CTR may be retaining the azla of JP while removing the wild features of the JP pointing, using KCT as a guide, albeit rejecting the געיה found in KCT. Now let’s look at the next word where CTR differs from both MAM and JP:
| CTR | קְר֣וֹב | ḥolam malei |
| JP, KCT | קְרֹ֣ב | ḥolam ḥaser |
Here, as with ולילה, ועוני, and נשאת, conformance with KCT cannot explain why CTR differs from JP.
For the rest of my review of the differences between CTR and MAM, see my document, “CTR Psalm 32: Where Other Sources Stand.” That document goes beyond this one in the following ways:
I hope this review of CTR’s Psalm 32 has shown the reader why I consider CTR to be the weirdest Tanakh on the web, and possibly the worst. Whether or not the reader is convinced of the same, I hope that at least my reasons for these judgments are now clear.