Why grammar-translation is wrong and doesn't work

Just putting it out there.

I will never cease to be amazed at how committed many people are to the Grammar-Translation method. Here I want to succinctly (re-)state the case for why this method is fundamentally flawed. To do so, let's look briefly at two things: what GTM does, and then the fundamental nature of language acquisition.

Grammar-Translation

Grammar-Translation is a historic method of language teaching which I think can be traced back to Johann Valentin Meidinger and his Praktische französische Grammatik (1793). It formed the first popular instance of what became known as the Prussian method, and was later retro-actively termed 'Grammar-Translation' by Reform movement of the late 19th century.

What GTM did, and still does, was create an approach that sought to distill language into its smallest bite-sized rules, feed them to a student one at a time through explicit explanation, and then have the student do translation exercises (in both directions) directly practising the just-taught grammatical tidbit. That method did evolve through late pedagogues, to something more like:

Explicit Rule, new vocabulary, example sentences, drilled controlled translation exercises, translation of controlled but more authentic texts.

This is basically what is still happening in most historical language classrooms today. Most teachers of ancient Greek and Latin are still using this methodology. Now, compare the following:

The basics of SLA.

I'm going to make heavy use of Florencia Henshaw's book "Common Ground: Second Language Acquisition Theory Goes to the Classroom" here, and let's start with this:

"The simplest definition of language acquisition is: the (mostly) implicit process of building a linguistic system by making form-meaning connections from the input. Basically, acquisition is what happens to you while you’re busy understanding messages." p3.

Not everything that you are exposed to becomes 'input', and not everybody quite agrees about what might be explicit about this process, but there is (as the title suggests) common ground that the process is mostly implicit.

Henshaw makes a couple of further important points in chapter one:

"Output does not build the system, and neither does learning about the language. We don’t acquire a language by learning its rules and applying them." p.4

"that no evidence suggests that only some languages can be acquired through input while others must be acquired by studying grammar rules." p4.

Let's be very clear then. There's nothing about historical languages that means they are learned in a different way to contemporary ones, and learning rules and applying them doesn't lead to acquisition. This is not controversial in the field.

Later in her book Henshaw discusses the role of input, what counts as input, and what role explicit instruction might play. Here's an important quote:

"First, explicit instruction is not necessary for acquisition. There is zero evidence indicating that without explicit grammar instruction, people cannot acquire a language." p.77

But is it helpful? It's much more difficult to say, partly because when research is done with an intervention of explicit teaching on grammar, and then students are tested on explicit grammar, they tend to do better... on explicit grammar. This is not the same as influencing acquisition.

Why GT never gets you to your destination

Grammar-translation doesn't get you to acquisition, because it's like a tram that never goes to the part of the city you want to visit. You're always just stuck on the tramline. Explicit learning + drills do not count as input for language acquisition.

If the goal of learning an historical language is acquisition, even if you just mean "reading knowledge ", i.e. ability to read ancient texts in the language and understand them as such, GT is a method that fights against that goal at every step of the way. I wouldn't go so far to say it never works, or cannot work, but I wouldn't stop far short. If GT works, it works in spite of itself not because of its methods. GT trains you to translate into your L1, understand L2 texts through the lens of your L1, and it only produces "input" for acquisition is infinitesimally small doses, vastly disproportionate to the time and effort needed. To acquire a language through GT is to learn to run an marathon by practising walking ultramarathons everyday: futile and wrong-headed.

(Between writing this article and posting it, this article popped up on my feed, which I suppose offers a third way: maybe your goal is to teach students to translate via decoding. In that case, GT is the fastest way to do that.)

If, on the other hand, you argue that the point of GT is to understand and analyse the language itself, you are also doing yourself and students a terrible disservice. The GTM for historical languages rests on a philological tradition that itself gave birth to the discipline of linguistics. Modern linguistics is the best we have for a scientific study of human languages, and all the best work in the linguistics of historical languages comes from those grounded in that discipline. Why, then, teach an outdated and ill-equipped pre-modern version of the 'grammatical' tradition when instead you could teach students linguistics and use the historical language(s) as your primary case studies and reference points? That is how to teach linguistics, not GT.

I concede that most adult learners feel a very strong desire, a felt "need" to understand how the language they are learning works. They want to understand why. They are actually wrong, in that knowing why, knowing how a piece of language does its job, doesn't contribute much, if anything, to their acquiring the language. But the fact that they are wrong isn't itself a good reason not to explain! Adult learners will feel better about themselves, and more confident, if they feel they know the why. And there is some place for explicit understanding of the language. It's precisely when it's possible and desirable to use the "monitor", as Krashen would put it. It's the real reason why schools put so much time in teaching native speakers the rules of their own language. It makes them better writers and editors. Not necessarily better speakers or readers though.

If you understand even the basics of modern SLA theory, and your aim is anything in the ball park of language acquisition, then it makes no sense to continue to teach based on the Grammar-Translation method.

Next steps

Having written all this, I think I now need to write two further entire series of posts.

Firstly, it would be good for me to put together here history of GTM. Not because it's an area of my specialty, but having a clearer idea about the history of GTM really helps to have clarity about its relationship to teaching historical languages. I heard a presentation just recently that off-handedly claimed that GTM emerged at the Renaissance when Latin stopped being a spoken language. That's wrong in at least 3 ways. So, a reasonable historical grasp of the history of GTM, and then language teaching since then, wouldn't be a bad thing for me to write up here.

Secondly, I have long dreamt of an introduction to linguistics for beginners that focused on ancient Greek. Most students have no linguistics background, and get shoehorned in to a grammatical tradition that is largely unchanged by modern linguistics, and so they are done a double disservice. Being introduced to linguistics with ancient Greek as the example language would do wonders to equip them for a proper analysis of the language qua language.

That is, to the extent that the object of study is the language itself, it should be done via the lens of linguistics, not the tradition of grammar inherited within classics.

I have tossed this idea around before many years back, with both Michael Aubrey and James Tauber. There kind of exists something like this for Latin with Wolfgang David Cirilo de Melo's book Latin Linguistics: An Introduction. I mean, he opens with this:

There is an abundance of introductions to linguistics; there is also an abundance of Latin grammars. However, what my students often complain about is the absence of an introduction to linguistics geared towards Latinists. p1

I think Aubrey and Aubrey's proposed grammar would solve at least some of my problems, but I'm not holding my breath for that. So, let's make a start here.

Okay, feel free to get upset with me in the comments.