Psalm 22: They Have Pierced Or Like a Lion?
Did the Christians simply invent the "pierced my hands and feet" passage in Psalm 22?
Glenn Miller
Draft: April 5/2000; Final version May 6/2000 (containing MAJOR corrections, revision, and some conclusive data) 

    On a couple of occasions over the past few years, I have gotten a question or statement from someone alleging that the first-century Christians literally INVENTED the translation of Psalm 22.16c [English, 17c Hebrew] ("they pierced my hands and feet"), and IGNORED the "real" Hebrew text (e.g., "like a lion, my hands and feet"). Here are two, with obviously different opponent/proponent dynamics (smile)...

     I don't know if this qualifies as a tough question but here goes.  Recently I have been speaking with a Rabbi and we got into a discussion about psalm 22 and specifically verse 16.  The verse, of course reads "...they have pierced my hands and my feet." and has been traditionally held as referring to the Crucifixion.  However, the Rabbi sees the verse as saying " a lion my hands" and as I don't know Hebrew I cannot argue.  Still, I have a suspicion that this is a late Hebrew rendering of the verse that has been a reaction to the christian point of view.  Which rendering is correct?  I have a feeling that the Septuagint might be able to answer this question as it will give a pre-christian Jewish interpretation of the passage in the Greek, which may be less ambiguous than the Hebrew.  Thank you for your time Mr. Miller.

       Dear Glen:  I have read with great interest your work regarding the use of the LXX.  I have an article that I could email to you where a Jewish counter-missionary is claiming that none of the books other than the Pentateuch were translated in pre-Christian times and that the rest of the LXX is a Christian translation and cannot be used to reflect pre-Christian ideas.  Please help!!  This comment came up when the counter-missionary was arguing that Christians deliberately changed Psalm 22:16 from "like a lion" to "they have pierced" and the fact that the LXX supports this reading is useless since the Christians translated the Psalms and put their own spin on it.

    (I had originally intended to address this in a larger piece dealing with the messianic nature of the Righteous Sufferer psalms/passages, but that is too low on the priority list right now. So I decided to go ahead and deal with the much smaller issue of lions-versus-piercers at this time.)

There are a couple of questions embedded here:

a. Were ONLY the books of Moses translated into Greek before the time of Jesus?

b. Was DOES the text of Psalm 22.16c say--"lion" or "pierced"?


a. Were ONLY the books of Moses translated into Greek before the time of Jesus?

    The first one is quite easily shown to be answered "no"...

    Everybody knows (or should know) that the Pentateuch is always translated first in any OT version, and the LXX is no exception. Each version shows a priority of the Five Books of Moses in time, and this is no surprise to anyone. To claim that the translation stopped there--esp. in the case of the LXX--is grossly in error. The hard data of literary finds bear this out clearly.

    In an earlier piece, I documented the existence of the LXX text-type prior to the time of the New Testament, but in light of the question here, I will add additional detail, showing that the pre-Christian/non-Christian references to LXX texts are broader than simply the Pentateuch.

    I had shown references/dependencies on the LXX at Qumran/DSS, Philo, Josephus, the Pseudepigrapa/Apocrypha, other Diaspora Jewish writers (pre-Christian), and the rabbinic writings. Of these, all I would need to do here is to document ONE CASE of non-Pentateuchal LXX usage to demonstrate the problem with the position. To do this, let me point to a number of such situations;

"In the later books of the Bible [than the Hexateuch], however, he has clearly consulted the Septuagint." [HI:IIW:112-113, implying its existence, of course]

"..Hence it is not surprising that where the style of the LXX is more polished, as in the Additions to Esther or in 1 Esdras, he adheres more closely to its text." [ADB, s.v. "Josephus"]

One example: LAB (Pseudo-Philo), in 61.4 is telling the story of David, Saul, and Goliah from 1 Samuel 17. The OT/Taanach has for verse 32:
And David said to Saul, “Let no man’s heart fail on account of him; your servant will go and fight with this Philistine...Your servant has killed both the lion and the bear; and this uncircumcised Philistine will be like one of them, since he has taunted the armies of the living God.
The LXX changes the subject around (from 'no man' to Saul) and adds an interesting concept to the passage:
And David said to Saul, Let not, I pray thee, the heart of my lord be dejected within him: thy servant will go, and fight with this Philistine...Thy servant smote both the lion and the bear, and the uncircumcised Philistine shall be as one of them: shall I not go and smite him, and remove this day a reproach from Israel? For who is this uncircumcised one, who has defied the army of the living God?" [note the addition of the concept of 'removing reproach from Israel' in the LXX, but not in the OT]
Now look at 61.4 of Pseudo-Philo:
And David said, "Do not fear, King, because I will go and fight the Philistine, and God will take away hatred and reproaches from Israel."
[LAB is clearly dependent on the LXX in this passage, and the translator in OTP makes this connection, and also links it to Lucianic mss of the OT. Also, Ps 151 which is not in the MT, but in the LXX and Syriac, shows up at Qumran(!) and the last line of that psalm goes: "I beheaded him, and took away the disgrace of Israel"]
To this we might add 23.1 (where LAB sides with the LXX against the MT on the location of Joshua's covenant renewal), 23.9 (where he adds a phrase "and the Egyptians humbled" from the LXX into Josh 24.4f), 24.3 (where LAB adds the phrase "and they went away" from the LXX into Josh 24.28), 30.5 (LAB uses 'hailstones' from the LXX instead of 'great stones' from the MT, for rendering Josh 10.11), and others from Judges [e.g., 35.3]. These are clearly non-Pentateuchal.
    One last example--Ben Sira  [1st century BCE (Prologue to Ecc., 15 NRSV)] refers to a Greek translation of more than the Pentateuch:
"You are invited therefore to read it with goodwill and attentions, and to be indulgent in cases where, despite our diligent labor in translating, we may seem to have rendered some phrases imperfectly. For what was originally expressed in Hebrew does not have exactly the same sense when translated into another language. No only this book, but even the Law itself, the Prophecies, and the rest of the books differ not a little when read in the original."
    But one of the pre-eminent Jewish scholars of the LXX text--Emanuel Tov--makes a summary statement that should quite settle the matter [Mikra:162-163]:
"Only a few data are known concerning the time of composition of the translations contained in the canon of the 'LXX'. According to the Epistle of Aristeas the Pentateuch was translated in the third century B.C.E.; this seems plausible in the light of the early date of papyri of the Pentateuch (middle or end second century B.C.E.). The books of the Prophets and Hagiographa were translated after the Pentateuch, since in them extensive use is made of its vacabulary [sic] and it is often quoted. As for the terminus ad quem, since the grandson of Ben Sira knew the translation of the books of the Prophets and part of the Hagiographa (132 or 116 B.C.E., according to different computations), these translations were probably finished before the first century B.C.E. Most of the books may have been translated at an early stage (beginning second century B.C.E. or earlier). One may note that the following books are quoted in early sources: Chronicles is quoted by Eupolemus (middle second century B.C.E., and Job by Pseudo-Aristeas (beginning first century B.C.E.). Additionally, Isaiah contains allusions to historical occurrences which indicated that it was translated in the middle of the second century B.C.E.

"Originally the LXX was a Jewish translation, and hence was quoted by Jewish historians (Demetrius, Eupolemus, Artapanus, Josephus), poets (Ezekiel) and philosophers (Philo). Especially the Pentateuch was also used in the synagogue service. However, at the end of the first century C.E. many Jews ceased to use the LXX because the early Christians had adopted it as their own translation, and by then it was considered a Christian translation. This explains the negative attitude of many Rabbis towards the LXX...This negative approach is visible also in the view of the Rabbis who explained the differences between the MT and LXX as alterations of the latter." [Mikra:162f].

    This is a TINY FRACTION of the pre-Christian/non-Christian references/allusions/usage of non-Pentateuchal LXX texts. A list of the ones just noted above would include: Joshua, Judges, Samuel, Kings, Chronicles, Esther, Job, Psalms, Proverbs, Isaiah, Jeremiah and other prophets and even LXX-only sections (e.g., Ps 151).

    [Sometimes I find someone try to make the 'I wish it were the case' position of the rabbis in b. Meg. 9a into 'proof' that only the Pentateuch was translated, but the plentiful data above (confirmed by Jewish scholarship today) demonstrates that the real world was otherwise...]

    And, just to cover the bases, the transmission of the LXX was not affected much by Christian interpretations. In other words, the Christians didn't smuggle in 'convenient' changes:

"Conversely, the NT influenced the transmission of the LXX but little. Allegedly several Christian changes were inserted at one time in LXX manuscripts, but few have survived to date" [Tov, Mikra:163, who cites Ps 96/95, but doesn't mention Ps 22...and this specific accusation doesn't show up in the rabbinics]

    This should pretty well dispel the notion that the LXX was ONLY the Pentateuch before the Christians got a hold of it...

[Additional note: I have glossed over major differences within the Greek translation families, but not to the point of introducing inaccuracies of the representation above. The situation is certainly more complex than the simple data above, and the rabbinic view of the LXX/greek translations I have discussed toward the end of the former piece.]
b. Was DOES the text of Psalm 22.16c say--"lions" or "pierced"?

    To start this discussion off, let me point out that the word in question here in Ps 22.16/17c is a NOTORIOUSLY difficult TEXTUAL (not 'theological') problem. To give you the sense of this, let me cite just two scholars:

"Ps. xxii 17c is an old crux which has never been satisfactorily explained. The MT's ka'ari yaday weraglay, 'like a lion my hands and my feet', makes no sense, and most modern scholars agree the text is corrupt. They also agree in locating the problem in the word ka'ari, 'like a lion'. All the ancient versions with the exception of the Targum read a verb here, and following their lead, most modern scholars emend the consonantal text from k'ry to k'rw or krw in order to obtain a verb in the 3mpl suffix conjugation." [J.J.M. Roberts, Vetus Testamentum, Vol 23, pge 247f]

"MT’s ka'ari (“like a lion”) presents numerous problems and can scarcely be correct. One must suppose that incorrect vocalization of the consonantal text occurred" [Craigie, Peter C. Word Biblical Commentary, Volume 19: Psalms 1-50. Dallas, Texas: Word Books, Publisher, 1998.]

The reasons that 'like a lion' is generally rejected (even though it shows up in the printed Hebrew bibles as MT) are given by scholars as follows:

1. As Roberts pointed out, the verse makes no sense with 'like a lion' there. He points out that attempts to salvage the MT reading fall into two approaches:

a. you either make the verse elliptical and supply a verb from context; or
b. you redivide the verse, so 16/17c STARTS a new verse, instead of ends one.
but points out that neither of these have produced convincing alternatives:
"Attempts to extract some sense from these enigmatic words without resorting to textural emendations generally take one of two forms. either one assumes an ellipsis of a verb in the line, or one redivides the line. Neither approach has produced a credible meaning, however. The Targum, following the first route, supplies a verb not found in the MT: 'they gnaw my hands and my feet like a lion", but such an ellipsis is incredibly hard and totally unexpected in the context." [op.cit., p.247n2]
Rabbinic commentators (medieval to modern) side with the lion, taking the elliptical tactic:
So Rashi:
"As though they are crushed in a lion’s mouth, and so did Hezekiah say (in Isa. 38:13): 'like a lion, so it would break all my bones.'" [tanknote: Isaiah is different than the Ps passage in that the verb "broke" is EXPLICIT in the Hebrew text. There is no verb in 22.16/17c, if the form is translated "like a lion"]
So a modern rabbi (A Rabbi Reads the Psalms), admitting that the MT reading is "rather strange", takes the elliptical approach and supplies a verb from the first half of the verse:
"a company of evildoers has enclosed me; (they have enclosed) like a lion my hands and feet"
Modern commentators would be quick to point out that the notion of a lion 'enclosing hands and feet' borders on the senseless, and certainly makes no sense in context either.

Indeed, biblical images of lions include much more violent actions like:

Professional translators (e.g., UBS) acknowledge the difficulty, and point out that "the majority of translations use a word appropriate to the action of 'a pack of dogs'" (e.g., niping at hands and feet), referring to the image in play at the beginning of the verse [A Handbook on Psalms, Robert G. Bratcher and William D. Reyburn, UBS Handbook Series:1991, p.221].
2. A second reason it is sometimes rejected (and emended to a verb) is that there is Hebrew manuscript evidence against it--even within the MT text-family. BHS lists two variants in the critical apparatus:
  • k'rw ("a few mss", Kennicott gives seven), a 3 personal plural verb form from the biblical Hebrew root kara(h), meaning "they dug/pierced"

  • krw ("2 mss"; Kennicott adds "in the margin of three"), also a verb form.  [Actually, this is the same word as k'rw, less the aleph which has 'intruded']



    It should be noted that these variants occur in MT-family manuscripts.

    [For the uninitiated, there is no such 'thing' as "the MT"..."MT" designates a family of manuscripts, no two of which agree in all details, and there is no hard-and-fast assumption that the MT is the 'original'. Tov, in the standard work on OT/Taanach, tries to get this across to his readers:

    "It has become clear from the preceding paragraphs that one of the postulates of biblical research is that the text preserved in the various representatives (manuscripts, editions) of what is commonly called the Masoretic Text, does not reflect the "original text" of the biblical books in many details. Even though the concept of an "original text" necessarily remains vague, it will always be legitimate to recognize the differences between the Masoretic Text and earlier or different stages of the biblical text. Moreover, even were we to surmise that MT reflects the "original" form of the Bible, we would still have to decide which Masoretic Text reflects this "original text," since the Masoretic Text is not a uniform textual unit, but is itself represented by many witnesses... Similar problems arise when one compares MT with the other textual witnesses, such as the Qumran scrolls and the putative Hebrew source of the individual ancient translations (tanknote: such as the LXX). We do not know which of all these texts reflects the biblical text faithfully. Thus, it should not be postulated in advance that MT reflects the original text of the biblical books better than the other texts." [OT:TCHB:11, italics his-bold mine]
    3. All of the very early translators (except the somewhat later Targumic writers, doing the interpretive-paraphrase thing) obviously had textual variants in the Hebrew text in front of them (or emended the text because of the difficulties mentioned above) as they translated--none of them had "as a lion" in front of them! 4. Even the other incidences of the phrase "like a lion" are different than this form.
    The phrase 'like a lion' (with the kap "as" prefix) appears in 16 unquestioned passages. Of these,
    5. There is recognition even in the rabbinics that this form does NOT mean the same as the "as a lion" form in the Isaiah 38 passage!
    Keil and Delitzsch bring this data to light in their discussion of this passage:
    "Perceiving this [difficulty of the translation 'like a lion' in the context], the Masora on Isa xxxviii. 13 observes, that k'ari in the two passages in which it occurs (Ps. xxii. 17, Isa. xxxviii. 13), occurs in two different meanings, just as the Midrash then also understands k'ri in the Psalm as a verb used of marking with conjuring, magic characters"
    Let's restate this for clarity:
    1. Since the Isaiah passage clearly means "like a lion" and since the Masora says the Psalm passage has a "different meaning", then it cannot mean "like a lion".
    2. The Midrash understands the form in the psalm as (a) a verb [instead of 'like a lion'] and (b) meaning "marking hands and feet with symbols".
        This last point alone should be sufficient to eliminate "like a lion" from consideration as the preferred translation, but when coupled with the other points above, makes the case very, very strong. Whatever the form means, it is LEAST LIKELY to be "like a lion".

        But there is actually a final piece of data that constitutes extremely strong evidence against ka'rai ("like a lion"), and that is from the single occurrence of the verse in the Dead Sea Scrolls (DSS).

        The DSS scrolls pre-date (historically) the Massoretic Text by centuries. Psalm 22.17 occurs in one of these scrolls from Nahal Hever (XHev/Se4, f.11, line 4). The form in this earliest copy of the Psalm in existence is k'ry, with the waw ending y indicating a 3 personal plural verb form ("they"). This is decisive evidence against "like a lion" (although it will not necessarily help us decide between the competing "they X" variants below). And remember from above, this k'ry variant also showed up in the later MT variants. [Data on the DSS manuscript can be found in The Dead Sea Psalms Scrolls and the Book of Psalms, Peter W. Flint, Brill:1997, p.83]

    So, if someone changed the text, it was in the other direction...(smile)

        Okay, so we've eliminated "like a lion" (for numerous reasons, and agreeing with ancient Jewish translators and ancient rabbinic commentators, and being confirmed by the ancient mss from the Dead Sea Scrolls). Is there any evidence to support "pierced", and if so, is that evidence stronger/clearer than the evidence for "tied/bound" (the other main contender, from Aquila/Symmachus/Jerome [A/S/J])?

        Remember, here we are talking about the meaning of the form ka'ry, not the textual form.

        Modern commentators conjecture that there is another root kara(h) IV, that means 'to tie' and that this is the form A/S/J understood as before them. But as Roberts rather brutally pointed out, they are having to "posit an otherwise unattested root" which is "not otherwise attested and be no more than a contextual guess" (op.cit. p.248,249).

        But let's allow the never-seen-before root to exist, for our discussion. This would mean that our form here (although coming from different, but homonymous verbs) could mean either 'they pierced' or 'they bound,' so we would have to decide on other factors than the textual tradition.

    The evidence in favor of "pierced" can be stated as follows:

    a. Its occurrence in the LXX (in the context of a faulty MT reading) is its strongest argument, and would normally be decisive. The LXX  (esp. the kaige, proto-Theodotian that we find in the DSS) would represent earlier readings/understandings than A/S/J, and therefore be stronger evidence.

    b. The underlying word for 'pierced' (kara/h I--"to dig") is a widely attested root, and is used by the Psalmist in the sense of 'piercing' in Psalm 40:7/6, where the form is karitha (2ms ending). On the other hand, the alleged root for "tie" (kara/h IV) does NOT occur in biblical Hebrew anywhere or even in later rabbinic lit! (See Jastrow, DTX:666). If the textual variants listed above (in the MT family and Qumran) meant "to tie," then they would therefore be unique in all of Hebrew literature, and contrary to all the linguistic data we do have! This is a very, very strong piece of evidence in favor of a 'pierced' reading over against a 'tie' reading.

    c. There are 28 passages (in the NAS) in which the English word "pierced" occurs. These are scattered across 15 different Hebrew words, with no single root accounting for more that 17% of the incidences (see unlabeled pie chart of distribution below). This means that kara was as good a choice as any other word to express "pierced".

    d. One of the canons of textual criticism is that the reading to be preferred is the reading which can account for the other readings. In this case, we have an interesting possible link between the Midrash (giving the meaning of our root as "mark with charms, conjuring symbols") and "pierced". In the ancient world of magic, charms were either (1) tied to various body parts, using string and jewelry; or (2) cut into the skin, or tattooed. Ancient curse procedures sometimes entailed attaching/carving  'symbols of doom' to/into the victim, and a midrashic writer in a period of high magic controversy (the rabbis were CONSTANTLY fighting--and sometimes turning their backs on--magical practices among the Jewish people) could easily have understood the pagan enemies of David (e.g., dogs--a common word for Gentiles) as tattooing magical symbols to their Jewish victim. This would certainly make sense of the midrashic comment. [But note how this understanding might also support the 'to tie' reading (e.g., tying charms on), but the use of the word "mark" would be a little stretch for that application]

    This would give a slight edge to the 'pierced' reading, over the 'tie' readings of A/S/J.

    e. But using the same criterion above, we can find a much stronger argument for 'pierced'. To believe that a form of the word 'tie' that would never show up in any Hebrew texts would be in front of A/S/J would be significantly mor doubtful, than to believe that the well-attested form of 'pierced' appeared in front of them (since 'pierced' occurs 15 times in Biblical Hebrew, four times of which it occurs in other Psalms [  Ps 7:16[EB 15]; Ps 57:7[EB 6]; 119:85; 40.7]). So, if we can find a way to explain 'tie' from 'pierced', then we are done--we have come up with a way to account for the various readings from the form 'pierced'.

    There IS a later, attested root for "tie" that looks a lot like our root for "dig/pierce"--the difference is in a tiny stroke of the pen on ONE letter (a het versus a he). This word could have been confused with our word...

    At the level of root, it would be very simple: the he (of kara/h) was misread as het (of kara/ht), and 'presto, chango' you have 'tie' from 'dig' (but this would have only been a fault at the root vocabulary level--neither the he nor the het are in our text or variant list here). This root for 'tie', although it does not show up in the biblical literature, DOES show up in the later rabbinics and could conceivably have been in parliance back in NT times. But the spill-over of root similarities to actual form similarities might have a bearing here.

    In an unpointed text (as would have been in front of A/S/J), the form k'ry ('pierced') could easily have been understood as kry (since the intrusion of the aleph was common, and understood by the ancients). All that is necessary to change kry (they pierced) into krhty (they tied) would be for a het to intrude, or be supplied by the translator (thinking the verbal root had the het instead of the he). There are numerous ways to do this, including an intruding he that is read as a het. These two letters were often confused, as Tov points out in TCHB:244f:

    "In ancient sources, many letters were interchanged because of unclear writing or roughness of the surface which caused misunderstandings in reading. Most of these interchanges were caused by similarities in the form of letters in the paleo-Hebrew and the Assyrian ("square") script...Several Qumran texts (tn: the text type A/S/J would likely have been working with) show a conspicuous similarity between waw/yod, resh/daleth, bet/mem/kap, het/he...Actually, in several texts such as 11QPs(a) [tn: a scroll containing biblical and apocryphal psalms], it is very difficult to distinguish between waw and yod, especially when they are joined to other letters...Examples of interchanges of letters are copious."
    What this means is that A/S/J could easily have looked at 'pierced' (or a form close to ours) and read 'tied' (remember, in A's first edition, he read something else altogether--"disgraced"--on the basis of simple misreading the form). This would explain both the Greek occurrences of 'tied' AND the Hebrew variants of the MT family--all from 'pierced' as we have it in our verse.

    f. Finally, 'pierced' makes more contextual sense than 'tied' since the action is more in tune with the image of dogs/jackels (although the Psalmist does mix these sometimes). This adds additional weight to the 'pierced' reading. [BTW, a 'lion' image would also be a bit out of place with the main 'jackel' image.]

    So, where does this leave us on what the 'original' or 'furthest back' reading was?
    1. "Like a lion" is rejected for a number of reasons by scholars: makes no sense, MT manuscript evidence against it, all the earliest translations (not interpretive paraphrases) reject it, its highly unusual form (for the 'like a lion' expression), the conclusive existence of the verb reading at Qumran, and even ancient rabbinic rejection of the meaning.

    2. The textual witnesses line up historically like this:

    This sequence alone would make a strong case for "they pierced".

    3. Of the remaining two major candidates (i.e., 'pierced' and 'tied'), 'pierced' is to be preferred since:

  • It occurs in the earliest manuscripts we have (LXX)
  • Its root is widely attested, whereas 'tied' does not even occur in all of existent Hebrew writing
  • It is not a 'strange' way to say this--it is not to be rejected for its infrequency
  • It provides a plausible basis from which to reconstruct (a) the midrashic/masoretic comments; (b) the MT textual variants; and (c) the Greek , non-LXX variants
  • It makes more sense in the immediate context.
  • Accordingly, I have to conclude that "pierced" is the better reading of the alternatives--under the praxis of textual criticism.
    ......................................................................... turn to ask a question of priority: why is anyone "arguing" much about this text?

        It is not cited or alluded to in the New Testament anywhere, so there is no real 'theological agenda' here. [Certainly the biblical scholar and  Church Father Jerome saw no problem with this being other than 'pierced'.]

        The 'pierced' quotations in the NT are from the Zechariah passage, NOT from this one. Psalm 22 is cited/alluded to a couple of times as messianic (of course), but not this verse. In fact, the strongest argument I can find against the 'pierced' understanding above is that it is NOT MENTIONED in the NT Passion Narratives (although it might be superfluous after Zechariah, which is a stronger messianic passage for the point of the New Testament).

    This question, accordingly, is a textual problem--not a theological one.

    Just for the record, (smile)...

    hope this helps,
    Glenn Miller
    May 2000

    The Christian ThinkTank...[] (Reference Abbreviations)