A Statistical Comparison of Editions
of the Greek New Testament


Using a database of various readings, 1 I compared Scrivener's Textus Receptus (representing the readings underlying the King James Version) with four critical texts, and identified 4165 translatable differences 2 adopted by either Tregelles 1857, Tischendorf 1869, Westcott & Hort 1881, or Nestle-Aland 1979.

In these 4165 places, the types of differences are:

omissions 3      1915       46%       
substitutions157738%
additions52913%
transpositions1443%

In these 4165 places, the numbers for each text are:

Tregelles       3095       74%       
Nestle-Aland332380%
Tischendorf349884%
Westcott & Hort361887%

Within this sample of 4165 places the agreement of the four critical texts is exhibited in the following table. For each pairing, the number of agreements is given over the number of places in which either text differs from the TR. 4

 TregellesTischendorfWestcott-HortNestle-Aland
Tregelles3095/30952765/38282907/38062803/3615
Tischendorf2765/38283498/34983094/40223023/3798
Westcott-Hort2907/38063094/40223618/36183195/3746
Nestle-Aland2803/36153023/37983195/37463323/3323

The percentage of agreement for each pair is thus:

 TregellesTischendorfWestcott-HortNestle-Aland
Tregelles100%72%76%78%
Tischendorf72%100%77%80%
Westcott-Hort76%77%100%85%
Nestle-Aland78%80%85%100%



What About the "Majority Text"?

The "Majority Text" of Hodges and Farstad represents the large majority of medieval manuscripts, and it corresponds much more closely to Scrivener's TR than any of the critical texts. I found that the total number of translatable differences of that text from Scrivener's TR is 1005. 5 In 327 places it differs from the TR without agreeing with any of the critical editors listed above. These differences fall outside the sample of 4165 places used in the statistical comparisons above. But in 678 places it agrees with one or more of the critical editors.



Some Observations on the Findings

From these findings it may be seen that where the critical texts diverge from the sixteenth-century Textus Receptus, they largely agree with one another. 72% of the translatable differences from the Textus Receptus were agreed upon by Tregelles and Tischendorf long before the publication of the Westcott-Hort text. The Westcott-Hort text (1881) departs furthest from the TR. The Nestle text (1979), though it largely corresponds with the Westcott-Hort text, differs from it in 551 places. In 295 (54%) of these places it returns to the readings of the TR. The Nestle text also has the highest percentage agreement with each of the others, ranging from 78% with Tregelles to 85% agreement with Westcott & Hort. The text of Hodges and Farstad represents a completely different method, so that comparing it with the critical texts is like comparing an apple to four oranges, but 67% of its differences from Scrivener's TR are supported by one or more of the critical texts.


Michael D. Marlowe
October 2001




Notes

1. The body of data used for this study is given in the collation of critical editions on this site.

2. These "translatable differences" include even the most trivial differences, many of which would show up only in an extremely literal translation. In fact most of the differences are of this nature. The 4165 differences do not include the longer sections, Mark 16:9-20 and John 7:53-8:11, which all the editors regard as secondary, and they do not include other verses that are doubly bracketed in the text of Westcott-Hort or Nestle-Aland.

3. i.e. "omissions" as compared to the text of Scrivener 1881. Looked at from another point of view, Scrivener 1881 has as many "additions" to the critical texts. Nothing is here implied concerning the original text.

4. The number of translatable disagreements between the texts can be obtained by subtracting the first number from the second number in each pairing.

5. In 156 of these places (16%) the text of Hodges and Farstad does however agree with another representative of the TR tradition (Estienne 1550, Beza 1598, or Elzevir 1624), so that it is necessary to enter into fine distinctions and definitions of the term "Textus Receptus" in this case. The numbers given here--it must be borne in mind--have to do with Scrivener 1881, which represents only the readings adopted by the translators of the KJV.



Valid XHTML