On the desire to speak about matters grammatical

Part of an ongoing series on objections to communicative methods

In my last post on these three inter-connected questions (those who don’t want to speak), I discussed the issue of those who object to the ‘content’ of communicative approaches – e.g. the desire not to speak about daily trivia. In this I turn to a similar but again distinct question – what is the ‘nature’ or ‘object’ of language instruction more broadly. Is the point of language instruction in historical languages to teach “grammar, and the analysis of language and texts along grammatical lines”? Or is it something else?

I want to say firmly that it’s something else – that the purpose of teaching a historical language is that students acquire that language as a language, with at least the concomitant proficiency in reading texts in that language, but preferably also some competency in listening, speaking, and writing.

If that is true, then we can ask a separate question about ‘grammar’ – or really two questions. Again, (1) what ‘is’ grammar, as it is taught, and (2) what place could, should, or must grammar have in our classes.

These are actually quite large questions. And the answers I give here may not be shared by all, but I’m articulating what I call “my considered and informed opinion”.

Grammar, as it’s typically taught and understood in classical languages, is a systematic description (descriptive!) of the language as it was used by its speakers and authors, as evidenced in a particular corpus of texts. It is language talking about language. And for that reason it’s incredibly useful, because one of the things that we might like to talk about is how language means. Having a vocabulary, and an understanding, of what nouns and verbs are, what endings are and do, how adverbs function, how clauses relate, etc., allows us to have conversations about how a language is working.

That’s why grammar is metalanguage – it involved a technical (domain-specific!) set of vocabulary useful to have if you want to talk about language itself. And most of our grammar is embedded in the history of Latin literature itself, since grammar as a field emerged among Latin and Greek authors.

But it’s very important to note that the analysis and study of language moved on from Latin grammar and philology to become Linguistics. We now call the systemic and scientific study of language(s) “Linguistics”, and the grammar traditions still incumbent in classics departments are often woefully ignorant of modern linguistics, carrying on with fossilised understandings of how languages work (often highly prescriptive), neglectful of how applying linguistics to Latin and Greek would yield, and already has yielded not only new insights, but in some cases overturn traditional categories and ways of talking. Traditional grammar is often quite bad grammar.

So, grammar is language for talking about language, and that’s quite a useful thing to have, especially since we are very interested in language! But what place should grammar have?

The Grammar-Translation method fundamentally operates on the maxim that if you teach people to understand grammar (and practice applying it in translation exercise), then they will learn the language. Several centuries of experience, and the work of modern SLA theory, strongly suggests that grammar is not the mechanism by which acquisition occurs. So we are left asking, “if grammar is not the mechanism by which acquisition takes place, what role could it have?”

I do think grammar can play a positive role in the language learning space, but before I get on to that let me qualify by saying, it is entirely possible to teach for acquisition without utilising or having a place for explicit grammar at all. Some teachers appear to have adopted that position, and it may be that you have learners who do not want to learn grammar. They don’t have to.

But you may also have learners who really do want to learn grammar in one form or another. So here’s a range of things that I think grammar can do, and you might want it to do.

  • Grammar can be a means of making things comprehensible.

Just like quickly glossing things in English can short-cut having a 20minute circumlocution to get across your meaning staying in language, having the tool of metalanguage can quickly and directly make something understood. This only works if you teach some grammar along the way, but it works. I’d rather use grammar to make something comprehensible in a sentence, than not.

  • Grammar allows you to talk about language use

Precisely because grammar is metalanguage, if you want to talk about language as a topic, if you want to talk about how language means, then this is a domain that learners need to acquire. I wrote in yesterday’s post about how I became quite domain-competent in talking about grammar in Mongolian – it’s because I (and my students) wanted to discuss language.

  • You can do grammar in the target language.

There’s no reason you can’t do grammar talk in the target language. Indeed, since students are likely learning a lot of this metalanguage vocabulary for the first time, doing it in the target language might be even more beneficial. So, in my Latin classes, we do talk about nomen, verbum temporale, modus coniunctivus, et cetera. While my Greek classes get subjected to questions like ἐπὶ τίνος πτώσεως ἐστιν αὕτη ἡ λέξις and the like. You can do communication about grammar in the language, thus creating more opportunities for comprehensible input.

  • You can equip students to access technical resources and secondary literature.

I was asked in a separate question about how I equip students to access various resources. If someone goes on to ‘higher’ study (caveats must apply), and to read technical commentaries, grammars, etc., etc., they will need to come to a mastery of traditional terminology. This is one reason I tend to work on grammar in both the target language and in English – if you’ve learnt both παρατατικὸς χρόνος and imperfect tense-form, you aren’t going to have a problem either discussing it in Greek, or reading a commentary talk about why something is imperfect.

  • You can sideline/background grammar.

One of the things I have tended to do now is to leave it to students to read information on grammar in English, or watch videos that I have produced (examples here), outside our main instructional time. Especially if they feel the need to get that kind of handle on the language (which many do), they can do that, it’s all there, but then we are freer in our limited instructional time to focus on operating in the language.

You don’t have to abandon grammar to do communicative methods, but you do need to let go of doing grammar as a primary mechanism for learners to acquire the language. If you want to teach grammar as the content and goal of language instruction, without acquisition, you are better off designing a course that is “Linguistics of Historical Latin” and adopting full-scale a linguistic approach to the language which presumes that none of your learners have, or will acquire, any facility in the language. That is totally fine. Indeed, I wish there were more places that in fact did that. There is a woeful absence of teaching Linguistics applied to historical languages.

But grammar? Yes, teach some grammar. Just teach it appropriately, and for its fit purpose, and without thinking it leads to acquisition directly, but as a means of talking about language for those who want to do so.

On the desire not to speak (about particular things)

This is a follow-on from ‘on the desire not to speak’, and a continuation of the longer ‘objections’ series. See here for parts one, two, three, and four. And today I want to talk about those whose objection to communicative methods is, paraphrasing Jeremy J. Swist, I don’t want to talk about daily trivia, I want to discuss literature. I’ve got four points on this.

You can talk about anything via a communicate approach. That is, the communicative method does not determine the content of your communication. There’s no reason that a communicative class has to spend its time in practicing how to order lattes in Latin, or asking the way to the bathroom in Ancient Greek. That’s a feature of a subset of modern language instructional material designed to give basic conversational competency to beginners, especially those who might travel abroad, and because those are situations those learners might face and want to have the language for. You are never going to be faced with those situations in Latin or Ancient Greek (unless you deliberately sign yourself up for an immersion event).

A class should talk about things its students want to learn. The method and practice of communicative teaching is about working in the language to make sure learners receive comprehensible messages, which should be interesting to the learners, because interest makes us pay attention and engage. So if you have a class full of people interested in reading ancient texts, the content of the communicative classes should orient itself to those texts. If you have a class full of people interested in discussing medieval philosophy, or South American botany, or contemporary geo-politics in South-East Asia, your classes should work to make that possible.

Language competency is domain specific. It is important to realise that while language structures (e.g. syntax) tend to be broad, competency, especially in vocabulary, is domain specific. To give an example, when I learnt Mongolian I had a real need (teaching) and interest is learning to speak about grammar and linguistics in Mongolian. So I became very familiar with that domain, and could talk about grammar in Mongolian. On the other hand, I didn’t do that much shopping for vegetables, and my competency to talk about various foods was very underdeveloped. The same is true in Latin and Ancient Greek – what you make the content of communication will also be the primary areas you develop vocabulary competency in.

A teacher should be broadly competent. I would say, then, that a teacher ought to be working on a broader competency than just talking about one field. When we talk about higher ‘levels’ of language proficiency, it does involve an ability to talk about a variety of topics. And as a teacher you ought to be shaping your communication to the learners you have, not the learners you think you should have. That requires breadth in order to be flexible. So, for teachers, I don’t think learning how to order lattes is as optional as it definitely is for students. I call this the latte test, by the way. Sure, there is no real reason to order a latte in Latin, but you should probably be able to.

On the desire not to speak

Part of my ongoing response to answers to Critiques and Comments to Communicative Approaches to Ancient Languages. See here for parts one, two, three.

What about students who have selected Latin or Greek because they want to take a foreign language that doesn’t require them to speak, or because they don’t want to talk about going to the shops on the bus, or because they want to analyse language in a linguistics-type manner.

I think these objections often come bundled, but really they need to un-bundled into three separate issues with three related but distinct responses.

  1. The desire not to speak
  2. The content of our classes
  3. The nature of our subject

In the rest of this post, I’m primarily going to talk about the first point here. I’ll deal with two and three in the next two posts.

I recognise that, especially in school and college contexts where “foreign” language credits are required, Latin (in particular) can be an attractive option for students who don’t want to be forced to speak.

And, I recognise too that, especially for children and youth, there’s generally an added aversion to speaking. Being forced to speak, being put on the spot to produce L2 content in front of other people, in environments that are performative, evaluative, and often competitive, is a set of affective features that multiply stressors.

Overall I think my response as an advocate of communicative practices, albeit one who does not typically deal with a lot of students uninterested in speaking, is threefold.

Firstly, it’s a structural problem at a larger level that “foreign” language credits are required, and so that students are forced to take a language. Now, I do think there are good reasons to have that problem – i.e. I don’t take issue with the idea that you might have a school/college system that has decided this is a feature of their programs. That said, it certainly creates a problem – students need to take a language, and they don’t want to take (a) certain languages, (b) any language. This is a problem that generally I have avoided by not having this kind of teaching job!

But recognising this is as a problem doesn’t identify a structural solution, nor am I even suggesting there is one. All it means is that students end up taking classes they don’t want to. And, where Latin (or Greek, but usually it’s Latin) is an option, you then particularly get students who take Latin because “it’s a dead language, and so they won’t make me speak it”.

If they rock up to day 1 of Latin, and it’s all spoken, then you have a different problem, but it is a marketing one. That can at least be solved by making it clear that you teach Latin communicatively (easier said than done). But it doesn’t actually get to the problem that I think is worth talking about:

How can communicative language teaching appropriately cater for learners who are not interested in speaking?

To which I want to lay down four foundations:


  • Comprehensible Input is necessary and (probably) sufficient for acquisition;
  • Output is neither necessary nor sufficient.
  • Compelling output doesn’t aid acquisition, and probably hinders it.
  • It’s okay to (partially) privilege reading (and writing).


1. CI drives acquisition. Whatever role you are prepared to grant to explicit instruction of grammar, from some to none, I’m firmly in the position at the primary and overwhelming driver of learners acquiring a mental representation of a language is their exposure to comprehensible messages in the target language (that are interesting and are produced in communicative contexts).

This comes with two corollaries. If INPUT drives acquisition, you don’t need learners to output to drive acquisition. There’s no acquisition based need to force or compel output. Students don’t need to write, let alone talk, to acquire Latin. They do need to read and listen though. They need to read and listen a lot.

This, by the way, is why I don’t mind doing the bulk of talking in my own classes.

Secondly, you can explicitly cater for students with apprehension about speaking by making this explicit and up-front. “In this language class I expect that you will acquire the language by attending to what I write and say and attempting to understand it.” Now, being lost in a sea of immersive incomprehensible language also raises people’s affective filters, so you need to work hard to make sure that it’s comprehensible, but setting the basis as “your job is to pay attention” is a much better foundation.

2. Output takes on a very different role then. Now, I have just said that OUTPUT doesn’t really lead to acquisition, but it can play three other important roles: (i) learners do need to output if they are going to develop output competencies. So, if students want to speak, they need to speak to improve speaking. But that is a distinct, but not discrete, output skill. And it’s still largely driven by the input. (ii) output can lead to input for other learners, (iii) output can also elicit input as other participants output (e.g. conversation!).

If you are specifically interested in catering to people from who output is going to cause stress, then I recommend working on developing output opportunities that can scale for confidence. At the high end is monologue speech with audience observation. Then working down through dialogue options, single Q&A with sentence answers, one word answers, written answers, through to comprehension-only multiple choice responses where the only thing you are checking is that a student comprehends a written text, through written text comprehension of multiple-choice answers. There are probably other ways too, but if you really want to make it possible for learners to acquire without production, it’s possible.

3. As soon as you force learners to output, you are putting them on the spot, and quite frankly most people don’t like that. Especially if it’s (a) public, (b) instantaneous, (c) spoken, (d) people are judging your correctness. So, really, you don’t want to do this to anyone who’s not wanting to do this. It’s why I don’t mind if students tell me before class (or even during class) that they’re just going to listen and observe.

That said, a teacher who is skilled and sensitive should be doing all they can to reduce the affective filter here – they should be encouraging production at a level that allows learners to respond with confidence, celebrating communicative success, facilitating negotiation of meaning when communication doesn’t quite work, and allowing minimal to no space for negative feedback. Explicit error correction doesn’t appear, according to the SLA research I’ve read, to have any concrete and long-term benefits on acquisition or on outputting correct forms. Even recasting does not appear to have strong positive effects.

4. As I’ve said elsewhere now, since most students of Latin and Ancient Greek are primarily interested in reading, that’s our main goal for most classes, even communicative ones. Which means you can still privilege reading in your class, provided reading is being driven by a communicative understanding – we read to understand messages in the target language, we don’t acquire language by analysis of grammar. That means actually a lot more reading, of (usually) a lot easier material, to make sure it’s comprehensible, and to increase the volume of comprehensible input.

It also means, especially for learners who have zero desire to speak, that you could shift your output activities to writing ones. Writing allows two feature that speaking does not – editing and time. You have time to ‘get it right’, and you can edit what you write. So, students who are feeling a strong negative response to speaking for those reasons can be further alleviated.

All this to say, I still think that a communicative language classroom for historical languages is the best way to go, and I would conclude (i) communication done well meliorates many of the factors of “spoken language” that learners find stressful, (ii) it’s possible to design and run communicative learning procedures that minimise, or eliminate, output, if that’s really a pedagogical necessity; (iii) my end goal for a student who came into a Latin class because they wanted to take a language class they didn’t have to speak in, would be to see them internalise Latin well enough to read it without translating, and be able to speak it if they wanted to.

Upper Level classes in communicative approaches

This is third in a series of answers to Critiques and Comments to Communicative Approaches to Ancient Languages. See here for parts one and two.


not a critique, but I’m curious about how proponents of communicative approaches handle the transition to upper level classes, where literary and historical analysis, secondary readings, etc., begin taking up more class time. Most of the conversations center around intro classes.


I’m going to split this up into three sections: what I’ve heard from others, what I do, what I can imagine.

It is true that a lot of the conversations about communicative approaches focus or centre on intro classes – that range of classes designed to bring learners to a point where they can read and understand texts for themselves. Let’s just pause and recognise that the notion of upper level classes (and I have college classics in view here) strongly revolves around the notion of “intro classes teach them grammar, vocab, and enough translation skills, beyond that they can read authentic texts with minimal help” which I consider patently false. Nonetheless, we do want learners to reach the point where they are reading texts and thinking about the content of those texts (among other things), not simply wrestling their way to an understanding of the words and sentences at a basic comprehension level.

So, as I understand it there are some camps of communicative proponents, in high schools, who have basically said, “given the time we have and the rate of acquisition, we cannot meaningfully get students to acquire enough Latin/Greek to read high register literary texts within our 2-6 years. So we aren’t aiming to.” I have reservations about that position, but I don’t think this is the post to go into them.

Among non-institutions, e.g. the various conventus, conventicula, summer schools, etc., at the intermediate and advanced levels you see courses offered that are basically, “we’ll read and discuss this in the target language”. But these tend all to be short-courses, not college-type upper-level classes. So let’s take a step sideways and think about what a ‘traditional’ college-type upper-level course looks like…

I don’t want to profile a particular course and professor, but using a mostly real example, a course in Greek Drama – we’ll read 3 plays in the original, then some others in translation, and you’ll read a range of secondary literature. Lectures will discuss historical contextualisation, literary features, etc. Students will read a great deal outside class both in and out of Greek, and write various types of papers.

At this point I have two thoughts: Firstly, if you’ve really, comprehensively embraced a communicative approach in your learning, as a student, and are at the point where you can read Greek Drama reasonably well, you should be able to handle a course like this on the Greek side, without much difficulty. A student with communicative competency and reading proficiency can handle being put in a non-communicative program.

My second thought, I’m going to delay until further down. Let’s talk about what I do at SeumasU instead. I don’t have that many ‘upper’ classes, partly due to student numbers. I have tried to offer 200 and 300 level courses – at the 200 level I envisage students who have covered most of the introductory language material, and have had some, though sometimes minimal, exposure to communicative approaches. We read simpler texts, and we work through them at a simpler level – I use the same q&a style that I mostly use in my intro classes, and we’re focused on understanding the texts at a comprehension level. But we’re still doing it mostly in the target language. So this is part of training students to read and comprehend *and discuss* while staying in language.

I’ve started to offer 300 level courses, where my expectation is that students (a) will prepare outside class, (b) can read the text with a degree of fluency, especially with pre-preparation, (c) we’ll read and discuss the text at a meaning and content level, not simply a comprehension level, though we’ll pause to unravel anything that’s not so easily understandable. That’s what I’ve done with Boethius last term, and with Ysengrimus this term. It’s still a long way short of the above college type course, but I’m also not teaching 2-3 hours of a course each week and expecting students to do 8-10 hours on their own over 14 weeks. Nor do I pull that kind of salary.

But could you teach an upper-level college course in the target language? Yes. Here let me draw on a parallel set of experiences. I’ve just finished taking a 200 level college course in Scottish Gaelic poetry from 1900 onwards, taught entirely in Scottish Gaelic. So here’s a college level course with target language texts, taught in the target language. Students need to be at a level where they can comprehend the lectures in the language, but naturally for those who are not natives, their output abilities are likely to lag behind. Much of the secondary literature, an overwhelming majority, is written in English. So that’s unavoidable, as it is in classics (not just English too!), but that reading can be done outside class hours, and it can be discussed in the target language. Assessments, both oral presentations, exams, and essay, all done in the target language.

I raise this example because it’s a minority language, very many of the students are non-native speakers, and even those that are native speakers may not have developed advanced literacy and academic skills yet in the language. And yet it’s possible.

Which circles me around to my delayed thought from above – one can envision an upper level course taught communicatively, if it were developed and supported appropriately. Both students and teachers would need to develop, gradually, the linguistic means and tools to discuss the range of critical, literary, and historical topics that you want to discuss in upper level courses. But of course, that is possible! Students, too, particularly need to be helped, very gradually, to write and express themselves in more complex ways and more academic ways than are often encountered in intro classes. Nobody wants to get to 200 level Latin and be asked to write a 2000 word Latin essay unless they’ve been given the chance to write 100, 200, 300, 400, 500, etc etc, word pieces along the way.

As far as I know, Polis Institute is one of the few places that really teaches some upper level courses in the target language. I don’t have any experience of them, but I know a few of their alumni are occasional readers, and I’d love for them to chime in.

Active Latin and Living Latin

This is second in a series of answers to Critiques and Comments to Communicative Approaches to Ancient Languages. The first post has had some very thoughtful and worthwhile responses so do go and read them. I have quite a few posts to come, so today I thought I’d tackle one that heads in a different direction somewhat:

Robert Low asked,

do you think there’s a useful distinction between active Latin (using spoken and written Latin in the learning process) and living Latin (using it as a modern conversational language)?

It’s an interesting question, and I think it opens up an illuminating consideration. I think I, and many, would quibble at exactly those labels of “active” and “living” Latin to describe those two things, but we can all see what Robert is asking with them, and let’s run with them for the sake of argument. First though, let’s take a step back and consider a few prior questions:

Before all, as I say often, we need to consider what sort of thing we mean by ‘knowing a language’, and what purpose we are learning/instructing for.

My starting point here, and I think it’s fair to say that it’s shared by many, is that our purpose is (usually) “to read historical texts”. That’s probably why most students sign up for Greek or Latin studies, no? Their particular interests may vary – Homer, Classical Greek Authors, Biblical Texts, Byzantine Medical Treatises, Medieval Legal Latin  – but the primary interest and motivation is to gain access to a body of literature and to be able to read that for themselves.

Now, that’s not the only purpose one can learn historical languages for, and perhaps we’ll circle back to that later. But let’s take my second question, what sort of ‘knowing a language’ do we mean? I mean “acquiring a language as a mental representation of a linguistic system in a way analogous to other languages that people acquire and speak”. I don’t mean, “having an externalised content-type knowledge of a language’s grammar”.

This matters so much. Because knowing the grammar of a language does not typically convey an ability to read in a language. But acquiring a language + having the specific skill of reading (and literacy + reading is much more readily learnt when picking up an L2 than L1) does carry that ability with it.

To circle back around then, I’d happily affirm without much doubt that most of those who advocate for communicative methods for historical languages do so because they (rightly) believe that doing so is a more effective way to get students to a point of acquiring enough language to read historical texts without translation. We can call that “Active Latin in the Learning Process”.

What about “Living Latin”? There is a contingent of people, smaller no doubt, who have an interest in using Latin for regular communicative purposes. Of course, there is variation within that group! Some enjoy writing in it, some enjoy contemporary fiction or poetry. Others enjoy speaking in it as a medium of everyday communication. (And there is an even smaller group of people worldwide interesting in using Ancient Greek in this way).

I would conjecture (and feel free for readers to comment on this!) that most of us in this category hold two beliefs about this: (1) that it’s an enjoyable hobby, (2) that it is also a very useful means of ongoing improvement in acquiring a broader and deeper grasp of the language.

That is, it’s not that you ever really finish the learning process, and can thus dispense with active Latin – learning Latin is an endless process, if you keep going, and practising living Latin in your life with other speakers is a highly effective way of advancing in Latin ability. So apart from the enjoyable hobby side of it, it remains a part of pedagogical acuity. Once you’ve reached a point where you can hold a conversation in Latin with without great difficulty, regularly conversing in Latin is one of the most useful things you can do to continue to develop your facility in the language (the other being extensive comprehensible reading).

Granted, every now and again you meet an idealist, usually a naïve novice I dare say, who wants to see Latin revitalised as a language of shared-location community usage + intergenerational transmission. That’s not a position or dream I hold or share, and I think that’s outside most people’s prospect for historical language pedagogy.

I do want to circle back to the idea that there are other purposes for learning Latin beyond reading historical texts. Latin has had a long, protracted afterlife. We have far more Latin from post-antiquity than antiquity. Most of those authors acquired Latin as a second, learned (in both senses) language. There’s no intrinsic reason that has to have ceased – Latin invites us to be fellow authors even today. For myself, I delight that some people want to speak Latin, and author new Latin texts, even into this new millennium. That, certainly, is a goal of “Living Latin” in the communicative sense, and long may it prosper.

Communicative Approaches aren’t fast enough?

This is the first in a series of answers to Critiques and Comments to Communicative Approaches to Ancient Languages (which is a mouthful and I won’t repeat it like that. Essentially, I asked people on twitter for some of their best critiques of Communicative-Approaches, and received a number of critiques, but actually more comments and questions. I intend over the coming weeks to provide my own reflective answer on each of them. There’s no particular ordering to my answers, by the way.

@Ludovic47101295 writes:

The one I see the most (merely as a language student fascinated by pedagogy) is that CI isn’t fast enough at getting students to the level needed to pass some external test.  For example preparing for the AP Latin test in the US.
I want to start my answer here just be recognising that many teachers of historical languages (by which I mean Latin, Ancient Greek, and similar ‘classical’ or ‘ancient’ languages) work within systems in which they have little choice about this – they are teaching students who will face those exams, and they necessarily must accomodate the way they teach to those exams. For them, this is a given.
But if we take a step back, we ought to be asking some prior questions: what are we testing? and what are we testing that for? Because if we look at AP Latin, or the GCSE, or other similar end-of-high-school exams, the content of those exams is very revealing. You could boil down the exams to the following types of questions and tasks:
1. Analysing morphology of words
2. Commenting on syntax structures in sentences
3. Translating texts
4. Writing longer comment answers on the contents of ancient texts.
Or, to whit, Grammar, Translation, and Commentary. So, which came first, the testing of Grammar-Translation or the teaching of Grammar-Translation? Students who are going to get tested on G-T need to be prepared for that, but the really prior question should be – what ought to be the outcome of language instruction?
If the answer to that question isn’t “grammar + translation”, then we’re teaching and testing the wrong thing.
And here’s what I would say – that shouldn’t be the primary aimed-at outcome of language instruction. Now, that’s partly a philosophical position about language instruction, but if you think that acquiring a language is meant to be the point, so that students come out of a program with an ability to read, e.g. Latin, as Latin, and understand Latin texts in Latin, without needing to translate, then both the teaching and testing superstructures that exist are misguided.
There’s a fundamental distinction that CI-based approaches adopt, following Krashen, but widely held in a lot of SLA – that there’s a difference between knowledge of a language (an external, content-type knowledge of a language), and acquisition of a language (an internalised, more-skill-like competency in a language). Those two things may be related, or not; they may impact each other, or not, but they are two separate things. Teaching for acquisition but testing for knowledge is bad pedagogy. Teaching and testing for knowledge but expecting acquisition is likewise bad pedagogy.
With all that said, I think the ‘speed’ objection is possibly true in the short term, but false in the long term. That is, if you have a bit of a lead time, and enough instructional time, a communicative approach should produce someone with a better ability in the language, and then you can teach some grammar/knowledge about the language they have already acquired, and the student will end up more than capable of handling the current crop of traditional exams. So I do think that teachers who have students who will face the traditional grammar/translate/comment exams can get away with communicative teaching, if they have a long enough program and are prepared to adapt some of their teaching towards the inevitable gravitational pull of the test. But it doesn’t have to be that way. We made it that way.