A Plausible Majority Text Argument
Susan's answer has correctly given the direct answer to your question when she states:
This is a textual issue. That is, some manuscripts have the words and
fasting while others don’t.
That is the simple fact. Which manuscript tradition the particular translation in question is following determines the omission or not.
She has also given a good and accurate presentation of a non-majority text tradition. This affords an opportunity here to show how presuppositions tend to drive the textual decision making, while at the same time presenting an argument for the inclusion of the wording.
One presupposition to watch for in any textual discussion is the oft quoted "the earliest and best" witnesses claim. The issue here is the word "best." Such is a subjective statement, based upon one's presupposition as to which manuscripts are "best." Those favoring earlier manuscripts versus those favoring majority manuscripts may and often do differ on what manuscripts are "best" based upon a number of other factors behind their decisions (their text critical criteria). So do not assume "best" is in fact "best."
As was noted, by far the majority of textual witnesses include the wording. NA28 lists but 4 texts omitting it, while giving a large number supporting the reading. Let's re-cite the witness groups, followed by some further discussion:
Super Minority Witness to Omit
א* B 0274 k
These four witnesses range in time from 4th to 5th century, so they are early witnesses. However, despite Metzger's commentary assertion that...
Among the witnesses that resisted such an accretion are important
representatives of the Alexandrian and the Western types of text.
...the three Greek witnesses given (א*, B, 0274) are all of the Alexandrian text type. It is only the one Latin translation (k) that is considered a form of Western text type.
Eclectic Super Majority Witness to Include
P45vid ℵ2 A C D K L N W Γ Δ Θ Ψ ƒ1,13 28. 33. 565. 579. 700. 892. 1241. 1424. 2542. ℓ 2211 𝔐 lat syh co (sys.p boms)
These 32 listed witnesses hide within them hundreds (possibly thousands) more witnesses, since the ƒ1,13 refers to two groups of witnesses (deemed related), the 𝔐 is a single symbol used to represent the Majority text (which is the category where most of the over 5,000 extant manuscripts exists), and the various translation witnesses (listed last in the group, after 𝔐) also hide multiple witnesses within the single symbols.
The oldest potential Greek witness is the P45vid, as it is 3rd century; but its reading is only "apparent" per NA28 because the actual part of the papyrus document at that point is missing (fifth row up from bottom right of image 6). However, the Latin witnesses noted by lat
here range from 2nd - 5th century according to NA28, showing the reading is just as ancient, if not more so, than the manuscripts omitting. So there is early testimony in the versions.
Regarding the Argument from Fasting in the Early Church
I'm not sure if this is Susan's conclusion or that of textual critics (or both), but this quote from her answer...
Because the addition (but not the omission) is easily explained by the
prevalent emphasis on fasting in the early church, it can be concluded
with reasonable certainty that it was not part of the original text.
...best illustrates the presuppositions behind decisions. This is a reasonable explanation for the wording as an addition if it was in fact an addition. What is excluded from consideration is an explanation if it was in fact an omission. An early error of omission might have occurred (for any number of reasons),1 and it simply found its way into a few copies that propagated the error. The fact that the early church emphasized fasting would then either be inconsequential or in fact support why the reading was retained in the majority of copies.2
The point here, however, is that early church emphasis on fasting is being used as a basis to presuppose the addition on that basis, when it can be just as likely that its retention in the text was because of this, and the original omission was a simple copying error.
Conclusion
Not knowing the textual history itself of manuscripts, it all falls back on the presupposed best criteria for evaluating the evidence that then leads one to either a minority or majority text position.
I personally believe their are major flaws in the presuppositions and criteria used to support the non-majority readings, and so hold to majority readings in most cases. But that is just me—each person has to evaluate what makes sense to them about text critical determinations and then come to one's own conclusions on what is likely the "best" manuscripts to hold to, and thus the best textual reading.
The evidence here in Mark 9:29 is certainly scant for supporting the omission, and the explanation unsatisfactory in my mind, and so I see no reason to reject the reading of the majority of texts.
NOTES
1 Davïd's answer has questioned my "reasons" for omission, so they have been added here. There are both intentional and unintentional reasons for variants: see Bruce M. Metzger and Bart D. Ehrman, The Text of the New Testament: Its Transmission, Corruption, and Restoration, 4th ed. (New York: Oxford University Press, 2005), 250-271. What follows is partly based on that.
One thing that is important to remember is that scholars on both sides of the argument generally recognize that most of the variants arose within the first three centuries of transmission (i.e. very early), prior to most of our actual manuscript witnesses, which means any witness (even some of earliest) can be in error from the original text. The major reason for this early variation is because it is also recognized that in this period of time the copying was not being done by professional scribes, but by your "average Joe" literate Christian. This does not mean the texts are evaluated differently (people make the same types of errors whether professional scribe or not), nor that the texts are necessarily prone to error (the textual consistency of witnesses shows that), but it does mean that more errors arose in that period because the practices of a non-professional are, well... less professional (i.e. double checks, standards in place to help reduce error, etc.), and then later texts tend more often to vary because of trying to evaluate and incorporate various versions from this early period of "more" variation.
Now, assuming one eliminates consideration of an intentional omission, which is probably less likely (though not impossible) given the recognized emphasis on fasting in the early eras, one is still left with a number of possible unintentional reasons for omission. There are four categories of unintentional omission given by Metzger and Ehrman, of which one can imagine any number of possible reasons an unintentional omission might occur (a few I can think of are given below); and given that any one of these reasons only needed to occur once, and then get copied to a few other manuscripts. This is critical to understand, since one such incident of any of the following becomes very probable, and thus it is very plausible that an omission could have occurred (at least equally, if not more so, than an addition).
Visual Issues: Anything that physically would have obscured the longer reading, or that the scribe himself suffered from. Since actual visual issues of the scribe are unlikely to have caused it here, the best explanation in this category would be a damaged document (hole, tear, smudge, etc.), and the earlier, untrained scribes would not be familiar enough with the text to know what to "add back," and so simply copied what was discernible (and not being trained, may not have noted the issue with the Vorlage).
Auditory Issues: Since copies were scarce, especially in the earlier stages of transmission (after all, only one original existed to begin with—one cannot get more scarce than that), "copies" were sometimes made by one person reading, and another (or many others to make multiple copies) were writing what was heard. If any distraction may have come in which the one writing did not hear the "and fasting," then it would have been omitted from the end of the text.
Mental Issues: Because copies of texts were scarce, and because some people were illiterate, texts were often read in the assembly and people would memorize the Scripture. "Copies" of texts may not be copies at all, but recordings from scribal memory, which obviously can have some points of failure, though memory retention then was probably far better than now (since but a few people tend to memorize whole chapters and books any more).
Judgment Issues: If the Vorlage had the wording added in the margin because while copying it had been recognized too late that it had been missed (one cannot easily "erase" or "delete" in this era of time), then the next copiest has to make a judgment call on whether or not to believe the text should be included based off the marginal note. If he judges (for whatever reasons) to omit the marginal reading and not even note it, thinking it spurious, then that next copy moves on without any notation of the possible addition. (Personally, in this case, I believe this is probably the least likely of the four broad category reasons why this particular possible omission occurred.)
One might conjecture other reasons that fit into these categories, but to summarize, there are a number of possible and plausible reasons for the initial omission to have occurred, such that yes, as Davïd commented to Jack Douglas, it is rather "axiomatic" to assume "any omission is already plausible" given the state of affairs for textual transmission in the early centuries. But these unintentional issues are not going to have some "history" for one to tie into (as the "addition" theory does), other than mere experience of the fact that such errors can, have, and do occur under such conditions.
Additionally, while I have no doubt "authorities" can be cited that "in copying literary texts tend to grow in transmission, not shrink," those authorities largely base their understanding off either (1) marginal notes being included into the text, or (2) addition based off another textual witness with a longer reading. But either of these begs the question about why the marginal note was there to begin with (as it could have been a correction of what was to be included), or whether the longer reading was correct or not. Certain notations that seem parenthetical at least have the merit of possibly being an explanatory gloss in the margin that was not in the original, but such would not be the case with this text here (as "fasting" is not explaining "prayer").
So while I agree with Davïd that "which scenario can best explain the readings present in the manuscript tradition" is how decisions are made, that word "best" is still subjective. As I argue in the main answer, what is "best" to one group is not to another, because they are evaluating the evidence based on differing essential criteria. To me, contra Davïd (and his quote of E. P. Gould), there are multiple plausible ways "of accounting for how the short reading arose." This, coupled with my reasons for holding to majority reading to begin with, are more than adequate to deem the reading of inclusion as the "best" reading.
2 The comment by kmote pointing to the additional evidence of Davïd's answer about variation regarding fasting in other passages (Mt 17:21, 1 Cor 7:5; to which Act 10:30 can be added as well) does highlight a broader discussion than simply the verse in question here. The broader discussion certainly could (and probably should) influence decisions one makes about this text in Mk 9:29.
First, however, a correction needs to be made to kmote's constructive criticism. He states (emphasis added):
The plausibility of this statement being accidentally omitted in both
passages [referring to Mk 9:29 and Mt 17:21] by the same scribes seems extremely remote to me.
This would not be the position of myself or likely any other who holds to a majority text position. The passages are from two separate books of Scripture, and the variations (whether accidental or not) very likely first occurred independently in two separate documents by two separate scribes, which in a later compiled edition (such as א) the readings were brought together.
So any extant document is not likely where the variant arose, and in such a case the scribe for the extant document probably did purposefully write the two passages as he did. This may be (1) due to those being taken from the two independent source Vorlagen, or (2) a later copy of an earlier compiled document that did the merging of those sources, or (3) having various Vorlage documents, some perhaps omitting, some not, and deciding to leave the omission out, or (4) in fact introducing the variation intentionally. If numbers (1) and (2), the scribe did his job well, whether what he copies was accurate or not was out of his hands (he had no other source[s] to go by); if (3), he had the tough job of the scribe in determining which reading was correct, and how to communicate the variation (if at all), but is still not specifically culpable for intentional change; if (4), then some agenda was at work that was wrong.
In short, I agree that two (or more) such "accidents" arising in a single document regarding the same "subject" would be highly implausible, but such early variants (accidental or not) almost certainly did not arise in a single compiled document of Scripture to begin with, but during the phase of independent documents.
However, regarding the consistency of evidence pointing to an intentional addition or omission of "fasting," that evidence does indicate that at least some of the variations in the four texts were at some point intentionally caused (and perhaps all were intentional). Since all the passages do have the fasting references in the majority text, can a majority text advocate have a plausible answer?
The short answer is "Yes." The long answer would probably have to argue for at least a later purposeful omission (which still may have been based off an original accidental omission) because of the consistency of the omissions. I am not aware, nor am I going to take the time at present to investigate, if a majority text view on this is yet published (much less examine the quality of the argument). Perhaps its an opening for a future PhD Thesis.
However, I can take a few moments and imagine what direction I would further research in to begin to seek an answer. (My reasons for holding to essentially a majority text view as a whole are much broader and stronger than any one particular textual issue, so I tend to assume there is a plausible answer for variation away from the majority text with each issue until proven otherwise.)
Facts
- The Greek manuscript witnesses omitting the wording are primarily of the Alexandrian text-type, with some Western.
- The version manuscript witnesses are primarily Coptic (Egyptian), with some Latin (chiefly 1 Cor). Only the Matthew omission has representation in Syriac, and that is only from the oldest Syriac texts found in Egypt and Sinai.
- With the possible exception of the Mark 9:29 passage (with its slight and only early manuscript era witnesses), the other omissions likely happened early enough to gain a larger ground amidst the copies.
The scholarly community has identified an early emphasis on fasting in the church, and tendencies to mention "prayer and fasting" together in commentary.
The earliest hints in post-New Testament writings indicate a return to
the external, legalistic, ritualistic practice of fasting. ... Almost
all the church fathers encouraged the practice of fasting.
(Curtis C. Mitchell, "The Practice of Fasting in the New Testament," Bibliotheca Sacra 147 n.588 [Oct 1990], 467.)
However, there is also evidence that fasting (especially in a ritualistic sense) was a contested practice in the early church. Mitchell notes this (ibid.). Ambrose of Milan (4th c.) noted a contention against fasting from two monks, Sarmatio and Barbatianus, whom he attributed to their following Epicurean ways (Epistle lxiii, particularly paragraphs 7-9, 15-17). They may have been followers of Jovinianus (4th c.), another who argued against ritual fasting, who was extensively contested against by Jerome (4th c.) in Book II Against Jovinianus. But contention against ritual fasting was even earlier, for Origen of Alexandria (3rd c.) argued against it as well. Veronika Grimm notes in From Feasting to Fasting: The Evolution of a Sin (Routledge, 2002) some points:
Origen understood clearly that both the Egyptian priests and the Pythagoreans abstained from animal flesh for religious and mystical reasons. As opposed to these, he declares ... strict adherence to the Pauline teaching, that for Christians eating or fasting has no religious significance. ... of no concern to God ... Abstaining from meat was one of the practices of which heretics were often accused by their detractors who liked to call it 'feigned temperance'. Origen himself fought against the ideas of Marcion, Saturninus and other Gnostics with all his might. It would not be surprising if he was wary of advocating their food practices. (134-135)
Grimm states later,
In all of Origen's discussion of Christian food practices there is no mention of fasting. (137)
Both Origen and his teacher (or at least influencer) Clement of Alexandria (late 2nd c.) wanted to
distance themselves from the Gnostics and other 'heretics' (ibid., 139)
and so
condemned also their ascetic practices as feigned piety and pretended temperance (ibid., 139-140).
Both men tended to have
interpreted fasting allegorically, as abstaining from covetousness and lasciviousness (ibid., 140).
Proposed Thesis
So given the facts above, the following could be a proposed thesis to explore reconciling a systematic and intentional omission of fasting references from some early manuscripts. References to the "Facts" given above are in parenthesis.
Early (#3) Alexandrian manuscripts (#1) may have had references to "fasting" removed in an
effort to counter growing Gnostic, pagan, and Judaic influences (#5) trying to ritualize
the practice in the church, which ritual became evident in the later church (#4 and 5).
Such removal then influenced translations done in Egypt from those Vorlagen (#2) as well as
the later line of Alexandrian texts extant today, while not influencing the early Byzantine
texts of Asia minor.
Just on the scant evidence examined, the thesis seems plausible to me. If the thesis itself (or some variation of it) could in fact be shown through more extensive research to go beyond just seeming to be be plausible, then a solid counter argument to the opposing view of the text having added the fasting references could be made in support of the majority text reading.
NOTE: Do not assume I am implying either Clement or Origen were responsible for the omissions from the text, but rather that they are representative of Christians in Alexandria Egypt contending against the ritual practices of fasting, and even advocating few if any points of necessity for fasting. If this is a prevalent attitude of those in Egypt (perhaps even earlier than these men) in control of the copies of texts in that area, then it gives plausible warrant for removing of the references to avoid ritualization and any necessity for fasting (which necessity Mark 9:29 seems to teach in this instance).
P.S. For further extended discussion of text critical matters between myself and Davïd regarding my answer, see beginning here in The Library.