According to Constantin Tischendorf's critical apparatus (p. 103, footnote for. v. 21) of Matt. 17:21,

ϛ Ln Ti τοῦτο δὲ το γένος (aethppl et talis) οὐκ ἐκπορεύεται (אbG εκβαλλετ. , ti2 al15 εξερχετ.) εἰ μὴ ἐν προσευχῇ καὶ νηστείᾳ (aethppl νη. κ. πρ.)

cum אbCDEFGHKLMSUVXΓΔΠ al omnfere itpler vg syrsch etp copwi etschw arm al Or3,579 Chr Hil Amb Aug …

om cum א*B 33. e ff1. syrcu syrhr sah coped (e edd. Mareschalli apud Mill, vide prolegg1402) aethrom etedd Euscanon

Based on Tischendorf's critical apparatus, it seems that the majority of witnesses support the inclusion of Matt. 17:21:

τοῦτο δὲ τὸ γένος οὐκ ἐκπορεύεται εἰ μὴ ἐν προσευχῇ καὶ νηστείᾳ

Therefore, on what basis does the Nestle-Aland 28th edition and various other modern manuscripts omit Matt. 17:21?

3 Answers 3


According to The Greek New Testament (4th Edition), edited by Aland, Metzger, et al. (2012), this verse does not appear in the best Greek manuscripts available. In order to save ourselves the tedious task of comparing and contrasting the pros and cons of the variant readings, the late Bruce Metzger has also published the companion volume, which is A Textual Commentary on the Greek New Testament (2nd Edition, 2012).

On page 35 he writes the following:

17.21 omit verse {A}
Since there is no satisfactory reason why the passage, if originally present in Matthew, should have been omitted in a wide variety of witnesses, and since copyists frequently inserted material derived from another Gospel, it appears that most manuscripts have been assimilated to the parallel in Mk 9.29.

What he is saying, is that the "Editorial Committee" (comprised of the four editors of The Greek New Testament which include Mr. Metzger) had ascribed {A} to their reading, which means that they are "certain" that the verse was not in the autograph.

In other words, please click here to see the text in question: please note the footnote (highlighted in yellow) in reference to the verse in question. The footnote lists the sources for and against. For example, the manuscripts against that omit the verse include the following: א* B Θ 33 579 892* l 253 it e,ff1 syrc,s,pal cop sa,bopt ethms geo1,A, and of course lists many more manuscripts for that include the disputed verse.

The smoking gun, however, which makes the textual criticism easier to accept a reading against, is the evidence from Codex Vaticanus B and Codex Sinaiticus, which are from the Fourth Century and therefore represent the earliest evidence from all available sources for this passage. Codex Sinaiticus is the oldest complete copy of the Greek New Testament, and is best known for its symbol of the Hebrew letter "א." Codex Vaticanus B does not include verse 21, but Codex Sinaiticus does include verse 21, but as a scribal correction. The actual passage in question can be downloaded here. Please note the marginalia at the top of the manuscript where some copyist added the words from Mark 9:29 using the "%" sign as the asterisk of the time, and of course there was no apparent attempt to make the addition to appear as if the original scribe had made the marginalia note in Codex Sinaiticus. According to the referenced footnote highlighted in yellow from our Greek New Testament, this emendation at the top margin of the passage within Codex Sinaiticus was the work of some second copyist, since the superscript "2" appears next to the "א."

In other words, the earliest evidence that includes verse 21 appears to be uncials C, D, and W, which appear from the Fifth Century. Codex Vaticanus B and Codex Sinaiticus are Fourth Century documents, and since the latter document has the marginalia "correction" of verse 21; and since the earliest evidence extent for verse 21 exists in other uncials from the Fifth Century, the conclusion is that verse 21 did not exist in the original autograph from Matthew's gospel.

  • You answered the question that was asked, but you added some of your own insight. So, I must ask, what's the earliest witness of those that do not have the verse, and what's the earliest witness of those that do have the verse? You seem to believe that the earliest of them all do not have the verse, but I'm not so sure that's true.
    – user862
    Commented May 4, 2014 at 4:25
  • @H3br3wHamm3r81 - Good question, and thank you for asking. I am going to edit my response to answer your question.
    – Joseph
    Commented May 4, 2014 at 4:38
  • Did you say that Codex Sinaiticus is the earliest witness of all those that do and do not have Matt. 17:21? If so, did you not notice that Origen includes it, and he lived in late 2nd century to mid 3rd century (184/185 – 253/254)?
    – user862
    Commented May 4, 2014 at 6:36
  • 1
    One should also note that Mark 9:29 has ...ἐν οὐδενὶ δύναται ἐξελθεῖν..., while Matthew 17:21 (in manuscripts which have it; and, there are many) has ...οὐκ ἐκπορεύεται... If indeed Matt. 17:21 was originally missing, and it was later incorporated by a scribe from Mark 9:29, how do you explain the dissimilarity of the Greek text? Certainly a scribe would have simply copied Mark 9:29 verbatim.
    – user862
    Commented May 4, 2014 at 6:45
  • 2
    BTW, if I originally made the question, I'd already have given you best answer for answering the question asked.
    – user862
    Commented May 4, 2014 at 6:47

While your question isn't an exact duplicate, it is answered here: What major discrepancies exist between “versions” of the Bible?

The answer is that our modern translations are based on differing original manuscripts, and that verse is not in the manuscripts that the NIV and ESV are based upon.

Just to head off the "If there are differences, how can we trust any versions?" questions, bear in mind that the differences are minor, and don't affect important doctrine, as explained in What is “Manuscript Evidence” and how is it useful?

  • Or most any other major version for that matter, besides the KJV. Thanks for your answer. Still seems very strange that most of today's major versions (NIV, ESV, NLT, MSG, CEV, etc) don't contain that verse and numerically skip from 20 to 22.
    – KodeKreachor
    Commented May 3, 2014 at 18:38
  • 1
    @KodeKreachor, if you are wondering why the verse number is skipped, that is because the numbers are for reference only (not part of the original) but they are consistent across versions that use them. This allows readers to go directly to the verse as their pastor reads. If they changed the numbers there would be much confusion.
    – Frank Luke
    Commented May 5, 2014 at 19:39

Two Schools of Thought

While it is somewhat of a simplification because the field of textual criticism is very involved, there are basically two main schools of thought on determining the correct text form for manuscripts.

  1. Priority to oldest extant reading
  2. Priority to majority extant reading

Again, each school of thought has other factors it considers when examining texts, but one of the fundamental presuppositions each takes when approaching textual criticism is either of the two options above.

Since the 19th century, the #1 view has ascended to be the majority view of scholars, and the #2 view has become the minority view. Prior to that, the #2 view was in essence the de facto view because of the fact that the majority of manuscripts available to scholars had those readings (many would not even have been aware of the oldest readings, as they were later manuscript discoveries).

One main presupposition behind the two presuppositional views is whether having a very few older manuscripts with a minority text reading is warrant to assume that there were no older manuscripts behind the later manuscript copies that have the majority text reading. Another main presupposition behind the two views is the nature of God's role (if any) in insuring His words were kept accessible to later generations of believers (this, of course, assumes divine inspiration of the text if it is a factor considered at all).

It is important to understand the above in light of the fact that one will often read statements saying a reading is not supported by the "best" manuscripts, which is usually an indication of a #1 position asserting their presupposition of priority to oldest. However, knowing which group a textual critic belongs to will ultimately determine what he or she means by "best."

Disclaimer: Generally speaking, I am of the pre-19th century majority view, the current minority view on text critical matters.

Basis for the Decision

Most modern translations exclude the verse because they are following a text (such as NA28) that was constructed off of an oldest priority position on the text, which when variations occur is usually a minority for extant reading, and thus (as you have noted in your question), they do not follow the majority of manuscripts. The exceptions are those translations that tend to follow the majority textual witness (essentially KJV, NKJV for translations still in popular use).

So what is the "basis" given for this verse. We have the commentary from Bruce Manning Metzger, United Bible Societies, A Textual Commentary on the Greek New Testament, 2nd Edition (London; New York: United Bible Societies, 1994) to inform us. I have emphasized some points to discuss following:

17:21 omit verse {A}

Since there is no satisfactory reason why the passage, if originally present in Matthew, should have been omitted in a wide variety of witnesses, and since copyists frequently inserted material derived from another Gospel, it appears that most manuscripts have been assimilated to the parallel in Mk 9:29.

  • "no satisfactory reason ... omitted in a wide variety of witnesses": This is based on the presupposition of the textual critics. Its omission in a wide variety of witnesses would be the same reasons for its inclusion in the even wider variety of witnesses, there was an early copy made that did not include the verse, and so that version got copied to a wide variety of witnesses. This in and of itself then tells us nothing about which reading is correct.
  • "since copyists frequently inserted material derived from another Gospel": This is using an understood fact (there is evidence such occurs) to presume such must be so here. But it fails to acknowledge that copyists also omitted material for various reasons (one being simply accidentally skipping a line in the hand copying). It would only take one such "error" to be used as a master copy for other manuscripts to perpetuate the omission to "a wide variety" of manuscripts.


So the decision is largely presuppositional. Both views #1 and #2 have "reasons" they hold to the presuppositions that they do, but those presuppositions largely determine which reading one chooses.

A discussion of driving factors in views can be found here (note: that is from a majority text viewpoint). For some contra argument, see here.

  • 1
    Clear and articulate answer; thanks for it! A bit of friendly push-back on your two bullet points. | • "...of itself then tells us nothing...": you've mis-stated the argument, though. The textual commentary states: "... if originally present ...". That is, on the assumption THAT IT WAS THERE, its omission would have no satisfactory explanation. The reasons for omission and inclusion (if absent) are not the same, and to elide them this way misunderstands the process, as well as the textual arguments. ... / cont'd.
    – Dɑvïd
    Commented May 4, 2014 at 14:23
  • 1
    ... / • Your claim ("using a fact...to presume such must be so") misrepresents what Metzger et al argue. They expressly acknowledge that copyists omit material (that's Bullet #1!), but since no good reason for such can be seen here, the other scenario (originally absent, but added) is trialled. Here there is good reason (AMPLE evidence of this in LXX-Exod/Deut!) to see how parallels flatten in transmission. | In sum, it's not largely a matter of "presupositions", but of how the evidence (all of it!) is best explained. FWIW. :)
    – Dɑvïd
    Commented May 4, 2014 at 14:24
  • @Davïd: I disagree that my omission is a misstatement. I omitted for brevity, and it is clear they make that assumption. But the point is that it is not a "satisfactory" answer under that presuppositional framework. I gave a perfectly "satsifactory" answer how it may have been omitted, as any text is omitted, having been done in an earlier copy these earliest we have were based upon. There is no way to "prove" or "disprove" that, since we do not know the textual history--we only know at least these two readings (with and without) exist.
    – ScottS
    Commented May 4, 2014 at 16:59
  • @Davïd: As to the second, there is also ample evidence things get omitted (without disputing the fact there is evidence things can get flattened also). So it is the presuppositional framework one is using about the nature of textual transmission (how one reads the evidence, not the evidence itself) that leads to the two differing conclusions by the two views, as either is a satisfactory "explanation" of the evidence. Why they leave it out is purely based on those presuppositions of how to read the evidence, which is the core difference between the two views.
    – ScottS
    Commented May 4, 2014 at 17:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.