Why do translations of the Bible disagree?

Defending himself against the charge of mistranslating a letter, Jerome told Pammachius plainly: "I render sense for sense and not word for word." He argued the alternative produces something unreadable, and cited Cicero as precedent. In the same sentence he exempted the one text this page is about: "except in the case of the holy scriptures where even the order of the words is a mystery". The rest of the letter qualifies even that, arguing that the evangelists and apostles, quoting the Hebrew scriptures, gave the meaning rather than the words.

That is the whole field. Every translation since sits somewhere on the line between those two, and where it sits is a decision made before a single verse is rendered.

The household codes tell douloi to obey their masters. The word is the ordinary Greek one for a slave. Many English Bibles render it servant, which describes a different relationship and a different ethical question.

Neither rendering is a mistake in the narrow sense, and translators who prefer servant or bondservant argue that the word covered a range of relationships and should be rendered by context. Only one of the renderings lets a reader see that the instruction is addressed to people who were owned.

The Greek word covers both, as Liddell, Scott and Jones set out under separate senses, and 1 Corinthians 14 has been printed for centuries as an instruction to women, which is a rule about half a congregation, and is printed in the translation cited throughout this site as an instruction to wives, which is a rule about a marriage.

The Greek word is the same in both cases. The English is not.

Junia, greeted by Paul as notable among the apostles, is a single Greek form, IOYNIAN, that could be a woman’s name or a man’s. A nineteenth-century proposal that it was a shortened masculine name took hold, and Junia left the printed editions and the scholarly conversation for generations before the reading was challenged and reversed. That was an editorial decision as much as a translator’s, and it worked the same way.

And Mark has Jesus say "this generation will not pass away until all these things happen", where the translation quoted here footnotes that the word could also be rendered race, family or people. A deadline or a description, depending on which is chosen.

Translators are not neutral instruments, and students of translation increasingly say so. The committee behind the version quoted below describes itself as drawn from many denominations, and sets out its philosophy in its preface.

What is argued is how far that shapes the result. The strong version, that renderings track the translators’ doctrine, is easy to assert and hard to demonstrate across a whole Bible. The weak version, that it shapes the contested passages where a choice has to be made, is easier to show: a study of gender in New Testament translation finds it in how Phoebe’s title is rendered, and the cases on this page are more of the same.

Some translations say which end of Jerome’s line they are working from, and they say it in the preface. One describes itself as an “essentially literal” translation emphasising “word-for-word” correspondence, so that a reader can see “the structure and exact force of the original”, and names the alternative it is defining itself against: versions that follow “thought-for-thought” rather than word-for-word, “emphasizing dynamic equivalence”.

Both are answering the question "what is this translation for", and they have different answers because they are doing different things with the book. Which traditions line up behind which answer is the part this page does not have a document for, so it does not say.

Every one of these decisions would be easier if anyone could ask a first-century speaker what the word carried. Meaning is not only definition: it is what a word summons, which is exactly what dies with the people who used it.

Lexicons are built from surviving usage, and surviving usage is a fraction of what was said.

None of this shows the text is unreliable. The places where a decision changes the sense are argued over in public by people who mark their working, sometimes in a footnote on the page, which is close to the opposite of a text nobody can pin down.

What it does show is that reading a translation is reading with somebody.