Question: What makes a translation accurate?
A translation is accurate when it faithfully renders the intended meaning of the biblical author into a receptor language. Biblical scholars differ over what approach to translation best achieves this goal. Those who favor a dynamic or functional equivalence approach argue for thought-for-thought translation. Those who favor a formal equivalence approach argue for word-for-word translation. In truth, no translation consistently follows either of these approaches. Rather, what we have is a spectrum with formal equivalence at one end and dynamic equivalence at the other. Translations like the NASB, ESV, and KJV belong somewhere on the formal equivalence half of the spectrum, while the GNB, NLT, and LB belong on the dynamic equivalence half. The NIV is probably somewhere in the middle.
Which of these two approaches produces the most accurate translation? Neither of them does if followed in their purest form. A pure formal equivalence translation would look something like the English glosses in an interlinear version. In such a version, there is a wooden substitution of English equivalents for the individual words of the original biblical text. No translator worth his salt would accept an undiluted formal equivalence approach, because it produces English sentences that make little to no sense. In fact, one of the complaints against formal equivalence translations is their lack of readability (though this complaint is sometimes overwrought). Likewise, hardly anyone would favor a pure dynamic equivalence approach because while readable, it can produce renderings that sit too loose to the linguistic structures and content of the donor language.
I would argue that the most accurate approach to translation is one that retains formal equivalence as far as is compatible with good English. There are nuances and implicatures of language that are retained in such an approach but that can be lost in dynamic equivalence renderings. Here are a couple of examples where nuances are lost in dynamic equivalence renderings.
Holman Christian Standard Bible
1 Peter 3:1 “Wives, in the same way, submit yourselves to your own husbands so that, even if some disobey the Christian message, they may be won over without a message by the way their wives live, when they observe your pure reverent lives.”
The issue in this text is the rendering of two occurrences of the Greek word logos. In the first half of the verse, logos is translated as “Christian message” and in the second half of the verse as “message.” The strength of this rendering is its clarity of meaning in the first occurrence of logos. Clearly logos is a reference to the Christian message (i.e., the gospel). In the second instance, however, such is not the case. One can only assume that the translators attempted to be consistent in the second instance of logos and also translated it as “message.” But this introduces a potential distortion into the text because Peter is not arguing that husbands will be won to the gospel without a “message.” The difficulty is alleviated if the dynamic equivalent “Christian message” is replaced with the more formal equivalent “word.” Word in this case has a similar semantic range to logos and allows the translator to preserve the play on the word logos. Logos in the first instance does mean “Christian message,” but in the second instance it means “speech” or “speaking.” The play on words is lost in the dynamic equivalent rendering, but is retained in the formal equivalent rendering. This is the kind of subtlety that is often missed in dynamic equivalence approaches.
Today’s New International Version
Hebrews 2:6 “What are mere mortals that you are mindful of them, human beings that you care for them?”
There was a great deal of controversy surrounding this verse when the TNIV was first released because it reflects the anthropological interpretation of the Greek phrase huios anthropou (lit., “son of man”). The formal equivalent translation “Son of Man” allows for both the anthropological and the messianic interpretations. “Human beings” only allows for the anthropological one. The upside of the TNIV rendering is its clarity. The downside is that it mutes the intercanonical connections with “son of man” language found in the Old and New Testaments—language that in some Old Testament contexts is unmistakably messianic (Psalm 8:4; Daniel 7:13). Thus the TNIV rendering obscures an interpretation that is widely held both in today’s Hebrews scholarship as well as in the early church. As such, this translation does a disservice to readers and should be revised (see Barry Joslin, “‘Son of Man’ or ‘Human Beings’? Hebrews 2:5–9 and a Response to Craig Blomberg” JBMW 14.2 : 41-50).
The issues here are complex, but more examples could be adduced to show that accuracy is best achieved by retaining formal equivalence as far as is compatible with good English. When there is no suitable formal equivalent in the receptor language, a dynamic equivalent may be used.
Denny Burk is associate professor of New Testament and dean of Boyce College, the undergraduate arm of the Southern Baptist Theological Seminary in Louisville, Kentucky.