Structure and translation

Let me take a few minutes to clear up a mistake that I made in my last post. (I should learn not to post in haste.) I tripped as I was racing through the syntax of Matt. 13:44 because it was background to the point I was making in the post.

Anyway, Matt. 13:44 is a verse of little theological consequence, but one that has some really interesting syntax -- and more than a little relevance to Packer's comment on structure in Suzanne's recent post. Here's the quote.

Packer: I am still saying there is a difference between trying to give people access to the way in which the sentences were put together, it was thus the thought was expressed in the original, and keeping people from the original by only telling people what you think it means.

The depth of misunderstanding of the nature of language implicit in this quote is hard to overstate.

It embodies two very profound errors. First, it is a mistake to think that there is any translation which is free from (the implied sin of) "telling people what [the translator] thinks it means". Second, it is a mistake to think that "access to the way in which the sentences were put together" is a guarantee of translational accuracy. In this post let me focus on the second error, the one that deals with structure.

For the purposes of this post I will gloss over the fact that for 80 years or so linguists have believed, with good reason, that structures were devoid of meaning. Chomskian linguistics, which has dominated the field for 50 years, all but demands such a view. It is only in the 1990's that a minority opinion came to the fore that there are certain limited kinds of meanings, usually fairly abstract, associated with particular linguistic structures. (This view is part of a movement called construction grammar. Some of the thinking is traceable back to Ken Pike's ideas.) However, the meanings particular structures express are, like words, different from language to language. So simply copying the structure from one language into another can be like translating 'dog' when the word means 'cat'.

So let's go back to our particular example.

The structure of Matt. 13:44 is very, very Greek -- and very, very un-English. The whole verse is:

ομοια εστιν η βασιλεια των ουρανων θησαυρω κεκρυμμενω εν τω αγρω ον ευρων ανθρωπος εκρυψεν και απο της χαρας αυτου υπαγει και πωλει παντα οσα εχει και αγοραζει τον αγρον εκεινον

The Kingdom of heaven is like this. A man happens to find a treasure hidden in a field. He covers it up again, and is so happy that he goes and sells everything he has, and then goes back and buys that field. (GNB)

The structural differences are so many it's hard to know where to start to show that cramming the Greek syntax into English does not give English speakers any meaningful help in understanding "the thought [that Matthew] expressed in the original". We could start with the word order.

"Like is the kingdom of the heavens to a treasure hidden in the field which finding a man hides, and ..."

All but the most literalists grant that Greek word order is not to be mimicked in English. Why? Because the function and meaning of word order is very different in the two languages. In English, it's the main way of telling who is doing what to whom.

John loves Mary.

is not the same as

Mary loves John.

In Greek you can tell who is doing what to whom by the endings on the nouns and adjectives coupled with the agreement marking on the verbs, so the word order gets used to do other things, like guiding the listener/reader's attention to those things that are most important in the mind of the speaker/writer. That's why ομοια is the very first word in the sentence. Matthew wants you to attend to the comparison. There's no neat way to do that in English. It's part of the triage that every translator has to perform to figure out what the essence of the communication is and how to get as close to that as the target language allows.

But word order is mooshy stuff. The differences can be subtle, and the arguments are often not emotionally satisfying. So let me cut to the chase -- the really intricate syntax of the crucial part of the verse. The point I want to make is based around the word θησαυρω, and the second of the two clauses that follow and modify it. The comparison Matthew is making is between η βασιλεια των ουρανων "the kingdom of heaven" (leaving out that the Greek is plural, BTW) and the treasure, θησαυρος, in the parable. The treasure is the thing he is telling the story about, not the man. Greek allows Matthew to make that focus of attention explicit by syntactically backgrounding half the story. A more or less literal rendering would be:

θησαυρω [κεκρυμμενω εν τω αγρω] [ον ευρων ανθρωπος εκρυψεν]
'treasure dat. [hidden in the field] [which finding man hid]'

This is really hard to do in English. Maybe something like:

The kingdom of heaven is like the treasure in this story: A man stumbles across a treasure hidden in a field and hides it again. Then with great excitement goes and sells all he has and buys the field.

The problem is in that second clause, ον ευρων ανθρωπος εκρυψεν. In Greek, the ον can function simultaneously as the object of both ευρων and εκρυψεν. This is just not do-able in English. Translate the Greek structure into English and you get something close to word salad. They tried it in the KJV

the which when a man hath found, he hideth,

but it's hopeless. If you say

He hides.

in English, it means 'He hides himself'. The verb is obligatorily transitive. The same for find,

*He found.

Each verb has to have an object to make it even resemble English. The which will do for one, but you need an it for the other:

which, when a man found it, he hid

but still that's awkward. What you sacrifice is that, in Greek, this is garden variety syntax. Part of the meaning of this verse is it's very ordinariness. Insist on stretching English to fit the Greek structure and you totally lose that ordinariness in the translation.

This example, I hope, shows that if you are going to try and translate the structure, then you'll have to make more or less arbitrary decisions about which structures don't interfere with understanding in English and which structures do. That's tantamount to "telling people what you think it means."

The mere fact that the literal versions sound almost comprehensible is more a testimony to our having had our linguistic senses eroded by exposure to the ungrammatical forms of the KJV, than to the Englishness of the translation.

When I was a graduate student in the 70's, studying syntax, we had to make up ungrammatical sentences to test hypotheses about each particular analysis. If you do this for a long time, it turns out, you can lose your sense of what's grammatical and what's not. I think that's exactly what's happened in the church. We have heard God's Word in obsolete English for so long that we've lost our bearings, and we've come to believe that strange English is the way God's Word should sound. And we're suckered in when some otherwise quite sensible authority stands up and says

"it was thus the thought was expressed in the original, and [you're] keeping people from the original by only telling people what you think it means."

Now for the confession of my sin. In my previous post, in my haste to get it up, I mis-parsed ευρων as a subject relative (which must be post-nominal) instead of a clause initial adverbial adjunct clause. (Sorry for all the linguist speak, but those who recognized my mistake will understand. And those who didn't, if they really want me to yack on about esoteric details of Greek syntax, can send comments or email and I'll explain the reason I thought it could be a word order exception.)