Guidelines for educationally responsible grammar-translation

Grammar-translation in the ELT community is a bit like Republicanism in the San Francisco Bay Area.  It’s practiced widely outside our community by what we imagine are people who generally don’t know any better or whose priorities are twisted against the interests of their socio-economic group, but espousing its values among we who (think we) are more enlightened is like opening up a bag of Doritos at the ballet.

For those in need of explanation of what exactly I’m condescending about,

Yakudoku [grammar-translation in Japan] is defined as a technique or a mental process for reading a foreign language in which the target language sentence is first translated word-by-word, and the resulting translation reordered to match Japanese word order as part of the process of reading comprehension.

(The article that this quote comes from is a good one to read if you wonder why people keep saying Japanese students are good at reading in spite of the entire culture seeming to regard printed English like a dog regards catnip.)

Grammar-translation, back and forth into and from the TL, seems to be the default method in state schooling worldwide.  It is unpopular among trained ELT professionals for a few good reasons. As a top-to-bottom lesson plan or curriculum, grammar-translation clearly fails, and there is no justifying its nonetheless overwhelmingly widespread practice (claims of its necessity for standardized tests, even when true, just increase the scale of the problem).  It surrenders all responsibility for language learning to the most mentally taxing tasks – memorizing connections between abstract tokens, applying rules of transformation in ways that place a heavy burden on short-term memory rather than automatization, knowing long lists of exceptions that invalidate most of the rules you just learned.  I have a hunch that part of the reason for its continued practice in public schools worldwide is precisely because it turns language learning from an activity that most of humanity has engaged in successfully for most of the history of our species into a bell-curve producing all-purpose test of general intelligence and dedication to the ritualized study process.  What it doesn’t do is produce functional language users, hence its unpopularity among ELT specialists and incredulity that it could be so common.

Grammar-translation also has a role in the ongoing cartelization of skills of native speaker and non-native speaker teachers in Japan.  Demanding as it does fluency in Japanese as well as long familiarity with the conventions that constitute “correct” translations, grammar-translation is seen as the exclusive domain of Japanese (NNS) teachers.  Whether NS teachers can ever become competent practitioners of it is beside the point; they are never asked to.

There is reason though to believe that grammar-translation can still have a place in a responsible and modern curriculum.  Since the 1990s, grammatical teaching, i.e. teaching the rules explicitly, has made something of a comeback among the ELT elite, albeit usually reactively and among lots of input and interaction.  Approaches seen as forward-thinking in recent years, including Task-based Language Teaching and Dogme, recommend explicit negative feedback and focus on form, i.e. some time to look at the abstract rules that govern grammaticality and what is correct or incorrect according to them.  The latest incarnation of the need for explicit as well as implicit knowledge and teaching seems to be the volume reviewed here, which I hope to read when the price drops. Grammar-translation can be a very helpful type of reactive grammar-focused activity and has strengths that other such activities don’t.  In this post I plan to introduce a few rules of thumb for grammar-translation as part of a modern language class.


First step: Throw away this textbook.

Here are my guidelines for rescuing grammar-translation from the SLA junk heap.

Don’t forget input and interaction

I made this point above, but it bears repeating: Grammar-translation is not a lesson in itself, and what benefits it confers will probably be forgotten unless scaffolded with the normal (for most teachers with recent training) amount of input and/or interaction.

For instance, say you task students with translating various types of requests, and come to the insight that a great many ways of making requests in English come in the form modal + you + VP, a la “Can you pass the salt?”  Students have likely encountered can before in the sense of “have the ability to”, as this is how it is usually taught first (I had a citation saying this, but can’t find it at the moment).  What they are encountering with this other can might be called an idiomatic second meaning, except that this idiomatic second meaning is actually the much more common way that can is used (again, citation somewhere lost in my hard drive).  Without copious practice and real-world use (or something close), this arguably main usage of can is likely to become a mere footnote of the other can and sooner or later be forgotten.

If you’re wondering what the role of grammar-translation would be if you already have input and interaction, well in the above example translation would help students to notice the movement of can (and other modals) to the front of yes/no questions, a feature unusual in world languages and certainly absent in Japanese.  Having students try literal translations in addition to figurative would highlight also that languages use different tools from their grammatical toolkits to achieve similar pragmatic objectives – requests in Japanese don’t use a word that can also be about ability like can (although some principles of politeness, like remoteness, are similar), and it’s likely that any other languages students will learn in their lifetimes will use different means as well.  Grammar-translation offers clear opportunities for comparison of surface structure and meaning.

Democratize the process

Students can benefit from watching the teacher conduct a translation to or from the TL on the blackboard, but in my view much more useful insights can be had by the students doing and sharing translations themselves.  An activity I will introduce in another post is based on students sharing and comparing translations.

Translation is a means to competence; it is not competence itself

The point, as with other types of focus on form, is to facilitate noticing (OK, I have this citation memorized – Schmidt, 1994) when particular grammar or vocabulary comes up in input again in the future.  You can think of it also as providing a little new perspective on something your students have heard but never really connected the dots in their mind about, like the first time you figure out what I-IV-V-vi means for popular music or hear the Wilhelm Scream.  The point is not to build up great speed and skill in translating or to reinforce the idea that you can only understand something when you can translate it quickly into your L1, your most natural means of communication, but to get a fresh set of conceptual eyes on the relationship between the TL and your L1.

Take a short phrase like “It’s supposed to rain tomorrow”.  Practice with translation of the various meanings of be supposed to can help students understand what someone is saying to them the next time they hear this or something like it – they still have to figure that in this case it’s indirect information that’s being talked about (often denoted in a very different way, using a quasi-adjective らしい rashii in Japanese) and not the rules of an establishment (e.g. “you’re supposed to pay at the counter”) or an order from a superior (“you’re supposed to speak at the meeting today”) or unrealized intention (“what is this slop supposed to be?”), but knowing a few possible interpretations gives you fewer ways to be wrong.  Knowing those translations though doesn’t give you much of a hint of what kinds of words are likely to come after a sentence starting “It’s supposed to…” and when you would usually say them.  You get that by saying that or hearing people say that a lot in particular situations, and build connections in your head between the words in the phrase, the sound of it, and the situation.  I had done translations of past tense verbs in Japanese before coming to Japan, but I still remember the sound of my landlady’s voice telling me a package from my mom had “arrived”.  For actually using the language in a natural way, collocations and the other statistically gathered information collectively called “language sense” are more important than quickly mentally substituting らしい when you hear “supposed to”.  That part just helps get your foot in the door.

Know enough of your students’ first language to be able to judge rough equivalency

I am assuming that the teachers thinking of including some translation in their lessons know enough English and Japanese (or the students’ L1, whatever it may be.  Translation gets harder when students have different L1s) to tell whether a given English and Japanese utterance can be called equivalent.  This doesn’t mean the teacher should know every way a given utterance could be phrased while remaining grammatical, just whether that particular utterance is compatible with the state of the world described in the utterance it is translated from or into.

This can be tricky in Japan, as other teachers are notorious sticklers for the “right” translation, for example “The weather will be fine next week” rather than “It’s going to be sunny next week”.  Students pick up on this and may reject “good enough” translations as sanctioned sloppiness.  I recommend you make the “good enough” principle explicit.

Remember that grammar translation is not the method of translation that most professional translators use, and the reasons for that are instructive

I suggest as an exercise conducting grammar-translations of some samples of a movie script, and then comparing them to the subtitled or dubbed version of that movie.  Of course, writers of subtitles and dubbed dialog work under some particular constraints when translating – timing and all that – but even knowing that students will still likely be surprised by how different professionals’ 意訳 (translation for meaning) is from their own 直訳 (direct translation, which to me is badly named).  I use the Japanese terms here but the principle applies for any language, although I suppose finding movies in some languages may be harder.

The point of this is that beyond grammatical equivalence or literal meaning in a very narrow sense, there are questions of frequency, collocation, idiomaticity, and other facets of what many call “native sense” which professional translators incorporate into their translations and which are generally ignored in grammar-translation.

I was watching one of the less popular Die Hard movies and one of the characters said something like, “Loaded”.  The subtitles said 弾が入った tama ga haitta, literally “bullets have entered”.  Now, we can all guess that the Japanese lends itself to a few different “literal” translations in English and vice versa, but the point is that none of the grammatical or narrowly literal translations of either of those phrases yields the other.  In having students do literal translations of lines from movies you are setting them up to fail in a sense as the result of literal translation is almost never what professionals come up with, but the failure of grammar-translation to produce natural language can be instructive.  Ideally, it will prepare students to think of the links between surface structure and meaning a bit more flexibly.

Use grammar-translation to generate real appreciation of and respect for the students’ L1

I’ve been teaching in Japan for almost all my career, and one goal for learning English one frequently hears here is fomenting respect for the Japanese language.  I don’t share the government’s concern with the Japanese language in particular, but grammar-translation is certainly a fine method for examining similarities and differences between languages, with an eye toward psychological commonalities across all humanity (e.g., the well-known tendency to speak of time as if it were 3D space, with future events in front of you and past events behind, a point English and Japanese have in common).  Recognition that “literal” translation does not incorporate the unique aspects of any language is compatible with the goal of respecting the students’ L1.

Grammar-translation as it’s currently practiced, at least in Japan, seems to me to do the opposite, forcing teachers to skyhook in unfounded assertions of their L1’s uniqueness.  The pretense that all sentences are translatable simply by looking at their surface structures and applying transformations implies that no language can have features unique enough to defeat this process.  I’ve known some teachers who maintain the appearance of superiority of their language by only translating into it and never from it, saying to students “we can do their thing but they can’t do ours”, but in my experience most teachers do translation into and out of the TL.  Teachers who do this and still feel the need to instill pride in Japanese sometimes just resort to periodically dropping slogans on their students a la “Japanese don’t usually say this because we and only we communicate non-verbally”.  Statements like this are usually not true, or at least the teacher certainly hasn’t done enough research to be able to say them, but are said anyway because the teacher has painted him/herself into a corner by teaching with the implication that all languages are translatable, and therefore one must resort to magical thinking to find anything unique about any language.

Grammar-translation as part of a complete language course doesn’t have this problem.  All languages differ from each other in how they use the common palette of vocabulary and grammar to describe the reality that we all mostly share, and it is entirely natural for literal translation to produce unnatural TL usage.  The fact that languages have different words and grammar for familiar things and actions is no surprise and is not very satisfying for a student looking for validation of his/her non-English identity.  Acknowledging that the different words and grammar are surrounded by different conventions and cultures which influence how they are used and forces literal meaning and contextualized meaning to be considered separate puts that fact into a context more compatible with a healthy appreciation of the uniqueness of every language.  Thoughtful comparison of how languages illustrate reality makes claiming “all languages are basically the same, but ours is special because woo-woo” unnecessary.

A point I like to make on translation that is very compatible with common notions of the appreciation for Japanese is that the relationship of Chinese loanwords to Japanese is very similar to that of Latin to English.  When translating, for example, “犯人を確保した” hannin wo kakuho shita as opposed to “犯人を捕まえた” hannin wo tsukamaeta, both of which mean something like “we caught the criminal”, the teacher can bring attention to the fact that only the first uses a kanji compound, which are almost always Chinese loanwords, for the verb.  Latinate words occupy roughly the same technical/formal space in the English lexicon as Chinese loanwords do in Japanese due to the similar statuses Chinese and French/Latin once had in Japan and England, i.e. the language of the professions, of religion, and the ruling class.  Therefore a translation that better captures the technical register of the first would use a Latinate word like “apprehended” or “arrested” rather than “catch”.  Actually on looking up those words I see that “catch” is Latin in origin as well – but I think we can agree the other two have more of a formal ring usually associated with Latinate words.  The existence and discursive role of Chinese loanwords has some qualities that are of course unique to Japanese, but their history also gives Japanese something in common with English and numerous other world languages.  This is a point that is easily accessible through the comparison of languages that occurs in translation.

The point

A student, a kikokushijo or returnee to Japan after having been educated abroad, once told me of her English teacher at the cram school she was attending, “He says grammar is the whole point when it comes to English”.  And clearly he had in mind the dry, dead grammar that only exists as a means for translation into Japanese.  When this student told me this I didn’t want simply to contradict her teacher and say that grammar, even the kind he’s talking about, is unimportant while communication is.  The truth is even the type of grammar the cram school teacher considered the be-all-end-all to English can offer some interesting perspective on the English that people should be learning (and often think they are when learning translation).

Grammar-translation should not be the first and last word on English, but it can be one of the words in the middle.

Whew!  Thanks for reading!  Have some Pad Kee Mao!


2 thoughts on “Guidelines for educationally responsible grammar-translation

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s