Karen H. Jobes presented a paper at the Fall 2007 Evangelical Theological Society Annual Meeting entitled “Bible Translation as Bilingual Quotation.” It is now being promoted by Zondervan on their blog. They have not just posted it but are actively seeking out bloggers to comment on it by emailing them. I originally read about it on Ancient Hebrew Poetry and Michael Bird was also contacted and posted a link. I was not requested to respond but they are looking for everyone’s input according to Paul J. Caminiti, Vice President and Publisher at Zondervan.
My first impressions were not that favourable. I was immediately uneasy that Zondervan is using this discussion to promote the T/NIV. Jobes was a translator of the TNIV and has been a member of the Committee on Bible Translation since 1995. This certainly isn’t a problem in itself as I have personally studied under Gordon Fee and Bruce Waltke and know them to be top scholars. However, in this paper Jobes is obviously promoting the T/NIV and she singles out the ESV for comment “I do find it more than a little ironic that what is advertised as an ‘essentially literal’ translation is the most verbose of several popular English translations–and that the ESV has about 30,000 more words than the TNIV!” (16). She also singled out the ESV on page 14 even though it had exactly the same verbosity as the NASB. These comments do not inspire confidence that she is unbiased, especially in that Zondervan proudly displays her chart with the NIV and TNIV at the top.
I am also curious as to how this data allows her to make the following conclusion “As the very verbose ESV demonstrates, all good translations must be a mix of both formal and functional equivalence.” I do not see that she has demonstrated that verbosity is related to whether a translation is formal or functionally equivalent. I decided to compare Zondervan’s “Translation Continuum” to the verbosity results. In order to do so Jobes’ list must be expanded. I counted words using Logos Bible Software’s Speed Search feature except with those translations which include the Apocrypha in which case I used the regular search feature in order to limit the count to the 66 canonical books. I am unsure why the results are slightly different from what Jobes came up with using Accordance but the point still stands. I included the following translations: ASV, CEV, ESV2007, GNT, God’s Word, HCSB, KJV, The Message, NAB, NASB, NCV, NiRV, NIV, NJB, NKJV, NLT2ed, NRSV, RSV, TNIV. For the interests of comparing oranges to oranges I am using Jobes’ counts for the MT (474,316) and NA27 (138,167) texts although they differ insignificantly from Logos’ results (475,525 / 138,103). Our base text therefore contains 612,483 words. All English translation counts are taken from Logos.
Before looking at the results I want to make a couple of preliminary points. Jobes is looking at bilingual quotation and, as she mentions in her paper, some languages are more verbose than others. With regard to the Bible the Hebrew text is significantly shorter by word count than the Greek text. The MT contains 474,316 words by Jobes’ count. The LXX of the 39 Hebrew books contains 502,795 words, an increase of 6%. When we look at a completely Hebrew Bible (MT+Hebrew NT) it contains 585,264 words as compared to 640,962 words for a completely Greek Bible (LXX+NA27) a difference of 9.52%. Obviously Greek is a more verbose language than Hebrew. On the other hand Latin seems to be a less verbose language as the Vulgate contains only 532,834 (even with the Apocrypha the Clementine Vulgate contains only 611,994) words and is thus 13% smaller than our base Hebrew/Greek text. This comparison therefore may say more about the English language than about which translation is best.
In the attached picture I have combined the results from the various translations with Zondervan’s continuum. Hopefully they don’t sue me for using their picture, as I am responding to their post at their request. As the picture shows there does not seem to be a direct correlation between Bible translation methodology and verbosity. The CEV is the closest in verbosity to the base text and is also far over toward the “thought for thought” end of the spectrum but the next two, NAB and HCSB, are both middle to left. At the other end of the spectrum is the NIrV with a whopping increase of 240,362 words (39.24%) over our base text but also very far over towards “thought for thought.”
I think the next task is to also complete this exercise in other languages with a multitude of Bible translations, e.g. German, French. The SESB module for Logos should provide enough data to extend the project to the European languages. I think we will discover that verbosity, or lack thereof, has nothing to do with whether or not the meaning has been conveyed adequately. Formal and functional equivalence debates will continue as long as more and more English translations are being produced.