With all these variations in the text, how do we decide which of the variant readings is most likely to be original, i.e., what the author wrote? In academic circles, this field of study is called “Textual Criticism,” and it is extremely complex. Most of our textual critics have devoted their entire lives to the discipline. It takes dedication and intelligence to get a handle on 5,600 manuscripts and a deep knowledge of Greek and three or four other ancient languages.

External Evidence

One criterion textual critics use to determine what is original is “external evidence,” which means they look at the manuscripts as a whole, including how old they are. A manuscript that was copied in the 5th century will, by default, be more trustworthy than a manuscript copied in the 11th century. The 5th century manuscript was copied less than 500 years after the writing of the original, and the 11th century manuscript has had an entire millennium for errors to creep into the process of copies being made of copies. Of course, the manuscript (the “exemplar”) copied by the 11th century manuscript may be more accurate than the exemplar copied by the 5th century manuscript, in which case the 11th century manuscript may be more reliable than the 5th. It can get quite complicated. 

To make things even more complicated, we don’t always know how many copies existed between the autograph and each of these manuscripts. The manuscript from AD 300 could be a copy, or a copy of a copy, of the original, but we are virtually guaranteed that an 11th century manuscript is a copy of a copy of a copy. However, an 8th century manuscript could be just a copy of a copy of the original. We now know that a manuscript could survive 150 to 200 years as a norm, so how many iterations of manuscripts there were between the autograph and the manuscript we have is unknown. As I said, textual criticism can get complicated, and textual critics have to make a lot of judgment calls.

Internal evidence

The other criterion is “internal evidence.” The basic rule here is that the reading that best explains the others is more likely to be original. This is called the “harder” reading. The corollary rule is that the shorter reading tends to be preferred. Consider Mark 9:29 as an example. The disciples are unable to exorcise a demon, so Jesus performs the exorcism. In private, the disciples ask him why they were not able to do so, and Jesus responds, “This kind can come out only by prayer” (NIV). But in the KJV we read, “This kind can come forth by nothing, but by prayer and fasting” (italics added). The additional words raise a question: is fasting required for especially difficult exorcisms? 

Looking at the internal evidence, the textual critic would ask if it more likely that “and fasting” was added or omitted? If Jesus had said “prayer and fasting,” what possible motivation would there be to drop “and fasting”? None. We also know that in later centuries there was an increasing emphasis on spiritual practices like fasting, and so it is quite plausible that a scribe would add it into the text. So the reading that only has “prayer” easily explains the other reading, “prayer and fasting,” and we consider “prayer” to be the shorter and the “harder” reading. Rules like these are based on observation and careful reasoning.  We know from looking at thousands of variants that scribes were hesitant to drop words out of the Bible but were willing to add them in, so we are assured that the words “and fasting” were added.

This is especially evident in the scribal practice of harmonization. There is no evidence that the scribes had nefarious motives, wanting to alter the meaning of the biblical texts. Rather, one of the most common intentional changes was the attempt to make parallel passages agree. This would happen when a scribe was aware of a parallel passage to the one he was currently copying. In Matthew 20:22, Jesus asks, “Can you drink the cup I am going to drink?” In Mark 10:38, the same question is, “Can you drink the cup I drink or be baptized with the baptism I am baptized with?” A variant of Matthew 20:22 harmonizes the two by adding to Matthew and saying, “Can you drink the cup I am going to drink or be baptized with the baptism I am baptized with?” The harmonization is so obvious that only the CSB and NET list the variant in a footnote.

Twelve Basic Rules for Textual Criticism

Aland and Aland give these twelve basic rules for how textual critics do their work. (The Text of the New Testament, Kurt Aland and Barbara Aland, trans Erroll F. Rhodes (Eerdmans, 1987) 275–277.

1. Only one reading can be correct, no matter how difficult it is to decide. There is no place to propose a solution not contained in a text.
2. Readings are determined by both external and internal criteria.
3. Textual criticism begins with external evidence.
4. Internal evidence on its own can not be a determiner.
5. The Greek texts are primary, and the versions and citations by the Church Fathers are secondary.
6. Manuscripts are weighed, not counted, but no manuscripts or family of manuscripts should be followed mechanically (local” principle).
7. Original readings are not found in a single manuscript.
8. The stemma — a chart showing the relationships among a group of manuscripts is important (the “genealogical” principle).
9. Variants are not considered in isolation but within the context of the tradition.
10. The more difficult reading is preferred, but not mechanically (lectio difficilior lectio potior).
11. The shorter reading is preferred, but not mechanically (lectio brevior lectio potior).
12. In order to be a textual critic, one should make a collation of one great papyri, one major uncial, and one significant minuscule.

The debate over the “missing” (“added”) verses would cease if only #12 were followed.

Results

Fortunately, we have a gifted group of textual critics who have done an excellent job of looking at all the manuscripts, applying the principles of textual criticism, making decisions on a word-by-word basis as to which of the variants they believe to be the original, and publishing their work in a form that not only shows the text but also the major variants. The text I’m referring to is published as Novum Testamentum Graece by the Deutsche Bibelgesellschaft (current edition is 28th revised), and is often referred to as Nestle-Aland (or NA), named after two of its main authors. The same text is also published as The Greek New Testament by the United Bible Societies (currently in its fifth revised edition), and often referred to as UBS. The only difference between these two is in the apparatus, the section of the Bible (sort of like the footnotes in a book) where they show variant readings and the manuscript evidence behind those readings. Nestle-Aland lists some 30,000 variants with only the major manuscript evidence for each variant. The UBS shows 1,408 variants with a much fuller list of manuscripts behind each of the variants. These are the texts that almost all modern scholars use when working with the Greek New Testament.

Conclusion

There are over 400,000 differences among the 5,600 Greek manuscripts of the New Testament, and yet not even 1% of those variants are both meaningful and viable. Textual critics look at the external and internal evidence, and they have done an excellent job at combing through all the manuscripts and making their decisions, and there is not a single viable variant that calls into question any point of biblical theology, major or minor.