Bible translation requires a solid linguistic foundation

that feeds into the work of the translators

Image by NASA (in the public domain)

“For a successful technology, reality must take precedence over public relations, for nature cannot be fooled.” 

This statement by Nobel-price-winning physicist Richard Feynman after the disastrous explosion of the space shuttle Challenger in 1986 could be used as the underlying principle of what I want to explore in this post: the need to come to terms with reality in an undertaking as big as Bible translation.

In a previous post, I used a comparison with a Christian hospital to make it clear that Bible translation is a technical task, where the hospital is less affected in what it is doing by the spiritual motivation of its staff than by the nature of its task, the physical well-being of the beneficiaries. The connection to reality for the hospital workers in this task is medical science. Medical science provides scientific understanding of the human body, the bodily functions, excretions, interactions of organs and the ailments that can disrupt the proper work of the body. The diagnosis of illnesses, the triage of patients, the various steps of therapy, including surgery, medication or physical therapy – all this is entirely determined by the best current understanding of the worldwide community of medical practitioners, as manifested in countless journals and books, peer-reviewed studies and minutely documented experience. Add to this the myriad best practices surrounding a hospital, beginning from disciplined handwashing and instrument scrubbing all the way to the detailed regulations that determine the set-up of an operation theater or the nutrition plans of the hospital kitchen – none of that is in place without a long and documented history of people dying or getting ill as a consequence of sub-standard behavior. All of this is part of the knowledge provided by medical science. Working in a hospital means acquiring as much medical knowledge as is needed for a specific role, so that one can fulfill the established requirements of each situation.

Bible translation is, as I stated then, a task quite similar in nature. Although the motivation of the Bible-translation work force is usually not technical, the very nature of the task requires that the workers are prepared for the reality of the thing they are dealing with: human language. And just as medical science is the lens through which the workers in a hospital view the human body they need to work with, linguistics is the lens that enables Bible translators to successfully engage with the reality of the human language that is supposed to carry God's good message to a new community.

Some decision makers in the Bible translation movement are of the opinion that linguistics isn't all that important for good Bible translation if the translators are translating into their own mother tongue. Since they are expert speakers of their languages, they have inherently all they need to succeed in Bible translation, as long as they have sufficient training in exegetical principles and some translation basics. But this is about as true as claiming that the nurses in a hospital don't need medical training because they are expert users of their own bodies, and that the brain surgeon does not need to scientifically understand the human brain, because his successful use of his own brain for many decades inherently qualifies him to work at least on his own brain, if not on somebody else's. Whereas this idea would be seen as bizarre in a medical environment, it is currently a widely held position among people who fund or plan Bible translation projects.

This is not the time to go into the reasons why this assumption is held – there is most likely a mixture of what Feynman calls "public relations" and a genuine sense of urgency in the face of low resources that becomes more compromising with regard to standards and what is seen as an acceptable result. The point of this blog is to provide some evidence that the assumption that translation can be done without a thorough linguistic basis is far off the truth and leads to very undesirable results – nature can indeed not be fooled.

The orthography

The orthography is the way in which we write a language. For most languages, the orthography is more or less based on the sound system, which differs very much from language to language. Anyone who ever learned another language quickly becomes aware of the fact that this other language comes with sounds that are unknown to the learner, and that need extra practice. Less obvious is the contrary situation, where the other language doesn't make use of some sounds that are important to us. The crucial thing here is that an orthography for a new language needs to be carefully crafted along the lines of the sound system, because any mismatch between sounds and letters will require more effort in teaching, so that really bad orthographies (such as English) need twelve years of intensive teaching, whereas good orthographies can be learned within a school year. Needless to say, the small languages of the world won't ever have twelve years of schooling, so many Bibles in the world nowadays remain unread because there isn't anyone who can read them. Too many Bible translations are written in an orthography that is not aligned with the sound system at all and are therefore impossible to learn.

The lexicon

The Greek New Testament makes use of roughly 5500 different words (not counting any inflectional changes to the same word). About a third of these words only appear once throughout the whole New Testament, because they are used only infrequently. If you translate the NT into another language, you'd expect to have to find and use roughly as many different words of the target language to reach a similar lexical wealth as seen in the Greek original. If you stay significantly below that number, your translation will either be less precise or to the point in its choice of words, or it will come across as somewhat flat or bland. But collecting more than, say, 2000 words using regular language learning and analysis methods takes an extraordinarily long amount of time, because it is very difficult to encounter most of the infrequent words of a language in stories, texts or normal talk. Most mother-tongue speakers also have a limited active vocabulary, but a much larger passive vocabulary, so that even experienced translators have trouble finding good lexical expressions if they can't access a dictionary of their own language. Dictionaries, though, are not very high on the priority list of current Bible translation projects, and for most languages dictionaries that go beyond short word lists of less than 2000 words are non-existent.

By the way, there is a persistent myth that undeveloped languages have only very small vocabularies with just a few hundred words. This is probably related to the abovementioned difficulty of finding many words when there is just a very small corpus of texts, but the truth is that with the correct methodology it is entirely possible to identify 10,000 words, 15,000 words or even 20,000 words in basically any language, because in each language the speakers have the same needs to talk about everyday concepts as we do, and they certainly have the words for them. It is just an entirely different matter to command all these words in the process of translating, which is why so many translations fall short of the lexical depth of the original.

The grammar

A similar myth is propagated about the grammar of undeveloped languages: they are said to be devoid of grammatical rules, and the speakers basically talk in random and chaotic ways, without any kind of describable structure. And this myth is often not only believed by the outsiders, but even the speakers themselves are convinced that it is true and that grammatical order and rules can only be found in "real" languages like English or French, where they have so much trouble learning these rules. I have met a good number of Bible translators who were convinced that grammar was not a thing in their language.

Quite the opposite is true: Not only are there no languages in the world that do not follow grammatical rules to arrange stems and grammatical forms into well-formed words and sentences, but it appears that the smaller and more isolated a language is, the more likely it is that it maintained some grammatical distinctions in everyday use that observers speaking European languages find extremely fascinating and mind-boggling. A small community is apparently better suited to maintain structures that wouldn't survive long in a larger or even global language, and the best-trained linguists of the world have their work cut out for them to describe these amazing languages in terms that really do them justice. Have a look at this website to find (usually) recent examples of grammars that pursue this noble work.

An example of such a structure could be the dual, a grammatical form that was once used in Greek when it was an obscure dialect continuum spoken by a small ethnic federation on a peninsula in Southern Europe. When this language developed into a trade and even world language, this tricky form quickly disappeared. It was used alongside the singular and plural forms in the case when reference is made to exactly two participants. Homer was still using this form when writing the Iliad and the Odyssey, but Luke was no longer encumbered by its use when he was writing his contributions to the New Testament. He just never had to think about it when referring to more than one participant, whether he really meant just two or more than two. This is not the case for many other languages in the world, where the dual is an important grammatical distinction. If you use the dual form, you really mean just two participants, and if you use the plural form, you really mean that you have more than two participants on the stage. The trouble with the dual is that most untrained speakers are not aware of this distinction in their language – nobody ever told them, and it takes an incredible amount of abstraction to realize without prompting that a) you have three forms where other languages have two and b) what governs the use of these three forms. An unaware translator is therefore very likely to just use a plural form where really a dual form is needed, and in this way distorting the accuracy of the translation. Imagine that you must translate the following passages from the story of the disciples on the way to Emmaus:

• “As they talked and discussed, Jesus himself drew near and walked along with them.” Luke 24:15
• “As they came to the village to which they were going, Jesus acted as if he were going farther.” Luke 24:28
  

You need to be very aware of the context to correctly decide that in verse 15 both highlighted pronouns need to be translated into the dual form for an accurate translation of the situation, as we know that indeed two disciples started out on that walk. What about verse 28? Well, the first they refers to the two disciples plus Jesus arriving at the village, so a dual form would not be correct. But the second they refers to the original two persons who started out on the journey, as Jesus did not announce his destination for this walk. So the dual form is again required.

What we can see here is that the grammar of the target language is forcing a decision on the translators that Luke as the author of the story did not face when he wrote it down for the first time in Greek. But clearly, the translators can only translate this successfully when they are aware of this forced choice in their language, and when they know how this grammatical distinction works. Experience has shown that this is normally not the case in situations where nobody has studied the language and feeds this kind of information into the translation team.

There are many other grammatical distinctions that fall into this category where choices are forced on the translators, and practically every language is expected to have some of them. Some may just lead to unnatural language or somewhat awkward expressions, whereas others may lead to serious misunderstandings and even faulty theology. 

On the other hand, it is indeed true that even untrained translators will not create ungrammatical sentences – these really go against the grain of the translators, as they break the rules of grammar in ways that cannot be ignored by them. This is the kernel of truth behind the claim that mother-tongue translators will always translate correctly into their own languages. Each sentence, taken by itself, will normally follow the rules of the grammar on the word-, phrase- and clause level. Still, there is no guarantee (and not even a high likelihood) that a grammatically unaware translator will always choose the correct grammatical categories of the target language to achieve the needed accuracy, or that the individual clauses and sentences will be combined to a readable text where the reader can track the participants, make sense out of the logical structures, or come to the right conclusions as to what is the important and new information. There are currently many glaring examples of where translation did badly go wrong because the mother-tongue translators were left alone without any kind of help to figure out what grammatical features of their language have an impact on the translation.

What is needed is a conscious awareness among the translators of the features that force choices on them in the translation process, beyond just producing grammatical sentences. Two things need to happen for that: Someone needs to study the grammar of the language to find out what these features are, and some way needs to be found to make that information available to the translation team.

In the old days of Bible translation both were accomplished by having a so-called "linguist/translator" on the translation team. This linguistically well-trained person studied the language in the early days of the project, and then in the daily translation work made the translators consciously aware of the features that mattered, so that they could apply them ever better while translating. Nowadays, though, many translation teams work completely cut off from any linguistic input, and their shortage of linguistic training also does not allow them to benefit from the usually highly technical grammar descriptions that rarely do exist about some languages (as seen on the Grammar-Watch website linked to above).

And there aren't even enough of these grammar descriptions to go round: Of the roughly 7000 languages spoken on the planet, only about 1500 can be counted as adequately described in linguistic terms, which means that most of the languages currently under translation do not have a grammar that could be consulted regarding the features that matter for translation.

So, a solid linguistic foundation is normally not a given in today's Bible translation movement. This problem becomes even bigger as many decision makers refuse to acknowledge it as a problem, discarding linguistics as a lens that would help them to connect with the reality of the human language. As the Challenger disaster has been described as a catastrophe that could have been avoided by taking reality more seriously, I think that there are many Bible translations going on that head for failure because of the willingness of the decision makers to discard necessary steps in the process in order to achieve greater speed. Linguistic foundations are the most apparent victim of this stripping-down of the process, but there are others, too. In spite of everything written above, linguistics is not the only thing that needs to be done right in Bible translation. If it is not applied well, though, the project is very likely to go wrong, so it cannot be ignored.