Printer-friendly version

Colwell and the New World Translation Revisited

Colwell and the New World Translation Revisited


In my book Understanding Jehovah’s Witnesses, I devoted about two pages to responding to the Watchtower Society’s claim that their New World Translation (NWT) is vindicated as a superior Bible translation because it merits “a perfect score” based on a test set forth by biblical scholar Ernest Cadman Colwell.1 Years later, a Jehovah’s Witness named Anthony Byatt published an essay that included an attempt to show that my critique of the Watchtower’s use of Colwell was without merit.2 Byatt expressed uncertainty as to whether I had “not properly read what Colwell had written” or was trying to “pull the wool over [my] readers’ eyes” on the cynical assumption “that few of them would check up” on what Colwell wrote.3

One can only be glad that a Jehovah’s Witness would make the effort to read Colwell for himself and assess whether my criticism was accurate. Regrettably, Byatt ignored most of what I said on the matter, choosing instead to try to manufacture a problem in my critique that is simply not there. For the sake of the reader who may not be familiar with this debate, I will review the issue from the beginning.

The 1963 Watchtower Article Citing Colwell

Let us first review what the Watchtower said in its 1963 article, a “Questions from Readers” column in its principal magazine The Watchtower:

In 1947 Professor Colwell made a study of a number of translations and put them to the test as to sixty-four citations in the book of John. The book contains what Professor Colwell considers the correct rendering of each of those sixty-four citations.4

The above statement presents the false premise of the whole article, namely, the premise that Colwell’s study was concerned with “the correct rendering” of particular words in John, with the hidden assumption that the Greek words to be rendered into English were the same for all of the translations that he scored. As I explained in my book over two decades ago and will explain again here, Colwell was concerned with differences among English translations originating from differences in the Greek wording of the passages they were translating.

The Watchtower article makes the following assertion on the basis of the premise just mentioned:

However, if any reader will look up what Professor Colwell has to say about these sixty-four citations and will compare these with the New World Translation he will see that the New World Translation merits a score of sixty-four along with Dr. Goodspeed’s translation of the Christian Greek Scriptures, which the book gives a perfect score of sixty-four.5

This sounds very impressive, but what the article fails to explain is that the test is concerned only to demonstrate which of the English translations of the New Testament were based on the best Greek text available at the time. This was the point of my response, which will be reviewed next.

Putting Colwell’s “Test” in Context

In my response to the Watchtower article, I quoted the opening paragraph from the chapter in which Colwell’s “test” appears:

If you care about your New Testament, you will want to know what “make” it is, who translated it, and, above all, how accurate it is. The translator may diminish, but he cannot materially increase, the accuracy of the particular Greek New Testament from which he makes his translation. In the matter of accuracy, therefore, the primary question is: “How accurate a New Testament was it translated from?”6

By “accurate” Colwell says he means “the accuracy of the particular Greek New Testament from which he makes his translation.” Note Colwell’s statement that “the translator may diminish, but he cannot materially increase, the accuracy of the particular Greek New Testament from which he makes his translation.” As I explained in my book, “What this means is that a translation cannot be any more accurate than the Greek text on which it is based, no matter how good a job the translator does; but it can be far less accurate if the translator does not do his job correctly.”7 The “primary question” with regard to the accuracy of a translation, according to Colwell, is “How accurate a New Testament was it translated from?” In other words, it is the accuracy of the Greek text that the translation is rendering that is the concern in Colwell’s treatment. The same Greek words can be translated in many ways, but this fact is not germane to Colwell’s test, which concerns only which Greek words the translation was attempting to render into English. “Thus, he does not address the matter of how good a job the translators did rendering the Greek into English, but rather, of the accuracy of the Greek text they chose on which to base their translation.”8

Next, I quoted Colwell’s explanation for how he ranked the seventeen English versions he chose to compare:

The ranking of these translations was based originally on the translator’s own statements as to their source; but this has been checked and corrected by a test as to the relationship of each one of these translations to two forms of the Greek New Testament—The Westcott and Hort text and the Textus Receptus or Received Test. I carried out this test for the entire Gospel of John. In this gospel the two Greek texts were compared verse by verse; and, from the large list of differences noted, sixty-four passages were selected in which even the freest English translation must show which of the two Greek texts it supports.9

As I then commented:

Colwell even states that he restricted his test to passages where the textual basis of the translation would be apparent even in ‘the freest English translation.’ In other words, no matter how loosely the translation paraphrases the text, and we would add, no matter how inaccurate the translation itself is, the differences in these particular sixty-four texts on which the test focused were designed to screen out such matters and reveal only which Greek text was used.10

I agreed that the NWT does merit a score of 64 out of 64 according to Colwell’s test passages. However, as I pointed out, “this fact only proves that on the basis of Colwell’s test we may say that the NWT translators used the best Greek text as the basis of their translation, not that the translation itself was accurate in its rendition of the Greek text into English.”11

Byatt’s Defense of the Watchtower's Use of Colwell

Byatt’s rebuttal is directed almost entirely at me, although he begins with a quotation from Michael Van Buskirk that made essentially the same point: “The NWT received a perfect score of sixty-four only because it is based on the Westcott and Hort Text. What received the perfect score was the Greek text utilized by the New World Translation Committee, NOT the English translation they produced!”12 After a couple of brief quotations from me, Byatt comments:

However, if Van Buskirk and Bowman had checked Colwell’s book properly, they would see that it is the English translation that counted, for each of the 64 texts is set out in English on pages 100-104, and the comparisons between TR (Textus Receptus, as AV) and WH (Westcott/Hort) are also given in English, not Greek, so that anyone can take a translation right up to the present time to see what that version scores.”13

Although this is the last reference Byatt makes to Van Buskirk in this connection, it would appear that Byatt based his rebuttal almost exclusively on Van Buskirk’s words “What received the perfect score was the Greek text…NOT the English translation.” In his quotation from Van Buskirk, Byatt underlines the words “Greek text,” and his comment indicates that he was finding fault specifically with those words. Technically, what received the perfect score was the English version on the basis that it followed the most accurate Greek text. Thus, Van Buskirk’s statement can be faulted, though only in a most picayune matter of the way he worded his point. Van Buskirk’s point clearly was the same as mine, which was that the test shows only that the NWT utilized the Westcott and Hort Greek text, not that its renderings of that text were particularly or especially accurate or well done.

Byatt goes on to comment on the fact that Colwell tested the seventeen English versions against a list of 64 specific lines in the Gospel of John. He points out that versions that claimed to follow the Westcott-Hort Greek text in some instances did not do so but instead followed the Greek text of the Textus Receptus. For example, Colwell gave the Twentieth Century New Testament a score of 59 out of 64 because in five places it did not agree with the WH text despite its claim to have been based on it. More precisely, Colwell found that the TCNT agreed with the WH text in 59 places, agreed with the TR in four places, and agreed with neither in one place.14

Byatt appears to be correct in his list of the five places where Colwell thought the TCNT departed from the WH (John 1:18; 5:2, 12; 6:47, 51). Evidently, Colwell considered the translation of John 1:18 in the TCNT, “God the Only Son,” to agree with neither the WH nor the TR. On this point Colwell appears to have been mistaken: the rendering “God the Only Son” is a translation of the reading monogenēs theos preferred in the WH (“God” = theos; “the Only Son” = monogenēs). The same rendering has been followed by at least three other English versions in the late twentieth century that were also based on modern critical editions of the Greek NT, which agree with the WH at this point. Specifically, the New Revised Standard Version (1989) and the Common English Bible (2011) both say “God the only Son,” and the New American Bible (2010)15 says “the only Son, God.” In any case, the relevant point in this context is that in at least four places the TCNT followed the Textus Receptus rather than the WH text, which is why Colwell did not give a perfect score of 64 to the TCNT despite its claim to have been based on the WH Greek text.

Back to Byatt’s rebuttal of my analysis, he comments, “There is no question that Colwell was looking to see how correctly each ENGLISH translation was following its chosen Greek text. Here TCNT departed five times from it, although they stated they were translating the WH text, so this is why Colwell scored them down by five points to 59…. What a reader wants to know, and what he may not be able to judge for himself, is how accurately has an English translation followed the Greek text it is claiming to follow?”16 These comments engage in a bit of misdirection. Colwell was not “looking to see how correctly each ENGLISH translation was following its chosen Greek text.” If he had been, he would have given the KJV a score of 64, not zero!17 Rather, Colwell was looking to see whether each English version followed what he stipulated was the most accurate Greek text, namely, the Greek text reflected in the WH critical edition of the Greek New Testament. Colwell did not score the seventeen versions on how well they translated their “chosen Greek text,” but simply on how closely they matched the WH in the 64 places he drew from the Gospel of John.

In the case of the TCNT, for example, the translators may have chosen to depart from the WH Greek text in a few places where they judged another reading to be more accurate. On the other hand, they may have simply decided that the traditional wording expressed the intended meaning more clearly (this is likely the case in the TCNT rendering of John 5:12 and 6:47, for example). It does not matter for the purposes at hand, or for understanding Colwell’s test, since his only concern was how closely the version followed the WH text.

Byatt concludes, “That is why it is significant that the NWT had a score of 64, for it reveals that the Committee had been careful at all times to translate the text before them.”18 But this conclusion does not differ in substance from what I (or Van Buskirk) had written on the matter. The NWT followed the WH consistently in the 64 places, and for that reason and that reason alone would merit a “perfect” score of 64. But this fact has nothing to do with whether the wording of the NWT accurately expresses the meaning of the WH text. To answer that question, one must consider the exegesis of the texts in question, not just the textual-critical question of the underlying Greek text.

Tearing Down the New World Translation?

Byatt ends his rebuttal of my critique of the Watchtower’s handling of Colwell’s test with some general comments on my intentions.

Did he not properly read what Colwell had written, merely giving it a cursory glance to pick out points he thought condemned the NWT? Or did he try to “pull the wool over his readers’ eyes,” thinking that few of them would check up on his use of Colwell’s work? Only he can answer that, and if he is truly an honest man he will admit his error without trying to justify it. But it has to be said that he does try in every way possible to tear down the NWT in his readers’ eyes, even to questioning the motives of the NWT committee.19

As I hope I have adequately explained, the analysis I gave over two decades ago on this matter was quite accurate. Byatt’s criticism misconstrues Colwell’s test as a test of how well the translators of a particular version followed their chosen Greek text, rather than a test of how closely the version agreed with the Greek text that Colwell identified as the most accurate.

Byatt’s parting shot that I “try in every way possible to tear down the NWT” is a blatant misrepresentation. In concluding my analysis of the Colwell test and its relevance to the NWT, I made the following observation:

Another thing to keep in mind is that no one claims the NWT is a bad translation all the way through. Rather, the usual claim is that the translation is biased in a large number of theologically critical passages, that the arguments in defense of the NWT renderings of those critical passages are for the most part without scholarly support, and that as a whole the translation is uneven and lacking in professional scholarly quality.20

My statement denying that “the NWT is a bad translation all the way through” reveals Byatt’s claim that I “try in every way possible to tear down the NWT” to be an attempt at poisoning the well. The NWT is to be commended where it is commendable, including in its following the WH text rather than the TR. But this does not make it an exegetically sound or linguistically competent version of the Bible.

The NWT and the Westcott-Hort Text: Not So Faithful after All

Finally, it bears pointing out that the NWT did not follow the Westcott and Hort Greek text consistently throughout the New Testament. In fact, on just one translation issue it is notoriously well known that the NWT departed from the Westcott and Hort text some 237 times. I refer, of course, to the 237 places in the NWT where it replaced the Greek word kurios or (more rarely) theos with the name Jehovah. Not only did the NWT deviate from the Westcott and Hort Greek text in this regard, it deviated from all of the thousands of extant Greek New Testament manuscripts. To claim a "perfect score" in regards to fidelity to the Westcott-Hort text on the basis of Colwell's 64 texts in the Gospel of John, while ignoring the 237 places where the NWT deviated not only from Westcott-Hort but from every critical edition of the Greek New Testament and every extant Greek manuscript, borders on the absurd. One might say that it is to "strain out the gnat but gulp down the camel" (Matt. 23:24 NWT).



1. Robert M. Bowman, Jr., Understanding Jehovah’s Witnesses: Why They Read the Bible the Way They Do (Grand Rapids: Baker, 1991), 135-37.

2. Anthony Byatt, “Some Principles and Features of the NWT,” in “Your Word Is Truth”: Essays in Celebration of the 50th Anniversary of the New World Translation of the Holy Scriptures (1950,1953), ed. Anthony Byatt and Hal Flemings (Malvern, Worcs., UK: Golden Age Books, 2004), 20-42 (29-31).

3. Ibid., 30, 31.

4. “Questions from Readers,” Watchtower, 1 Jan. 1963, 95.

5. Ibid.

6. Ernest Cadman Colwell, What Is the Best New Testament? (Chicago: University of Chicago Press, 1952), 85, quoted in Bowman, Understanding Jehovah’s Witnesses, 136.

7. Bowman, Understanding Jehovah’s Witnesses, 136.

8. Ibid.

9. Colwell, What Is the Best New Testament, 85, quoted in Bowman, Understanding Jehovah’s Witnesses, 136.

10. Bowman, Understanding Jehovah’s Witnesses, 137.

11. Ibid.

12. Michael Van Buskirk, The Scholastic Dishonesty of the Watchtower (Santa Ana, CA: Christian Apologetics: Research and Information Service [CARIS], 1975), 21, cited in Byatt, “Some Principles and Features of the NWT,” 29.

13. Byatt, “Some Principles and Features of the NWT,” 30.

14. Colwell, What Is the Best New Testament, 87.

15. This is the Catholic NAB, not the evangelical Protestant NASB

16. Byatt, “Some Principles and Features of the NWT,” 31.

17. Colwell, What Is the Best New Testament, 87.

18. Byatt, “Some Principles and Features of the NWT,” 31.

19. Ibid.

20. Bowman, Understanding Jehovah’s Witnesses, 137.