Neanderthal practice? On the teaching of grammar and the slaughtering of sacred cows

(Poznan 2013)

My lecture falls into three parts. I shall explain (1) the principle of double comprehension as the prime requirement for language acquisition to take place.

I shall move on to (2) a re-evaluation of the role of the mother tongue in foreign language teaching, which, rather than be ignored, can be made our most powerful ally. I shall explain (3) the generative principle and demonstrate bilingual semi-communicative pattern drills as a way into the grammar of a foreign language.

The time is ripe for slaughtering two sacred cows of our mainstream philosophy:  1. L1 is not a last resort when nothing else will work. It is indispensable for a quick and complete grasp of new constructions, one which leaves no questions unanswered.

2. Grammar practice need not always be situational or contextual so as to lead the learner gently into communication.There is a place for pattern drills with  disconnected sentences. However, drills must be presented as sense variations rather than syntactical manipulations.

 1. Double comprehension

What are the basic conditions for language learning to take place? What are the prime requirements that have to be fulfilled? There are two. Language acquisition begins when people speak to us in ways we can understand. Comprehended input is the single most important factor in language acquisition. It is the essential precondition for our own participation in dialogue.

But learners must understand more than messages. We also have to understand not only what is meant, but how the messages are put together, how they come to mean what they mean, in other words how things are quite literally expressed. So let’s distinguish these two types of understanding. On the one hand understanding what is meant: Let’s call it situational or functional or communicative understanding: We get the full message, we get the idea as precisely as possible so that we can react meaningfully to it. This understanding is something that a good, idiomatic translation can provide. The second type of understanding refers to the way the message is structured. Let’s call it structural, or formal, or operational understanding. This is something that in most cases a word for word translation can provide us with.

As native speakers, once we understand the message, we also understand how the meaning components interlock to create that meaning. But this is not true for infants learning their mother tongue, and it is not true for FL learners. Here, the distinction is important because languages often express the very same idea in different ways which are not immediately transparent to beginners but remain cryptic for a while at least.

I claim that a kind of double comprehension – situational and structural – is needed for all types of language acquisition, including L1 acquisition. Admittedly, to acquire a mother tongue, it seems that all a child needs is to be talked to in understandable ways, i.e. to understand messages. Okay. But as infants receive more and more comprehended input in recurring and often only slightly changing situations, they do not only understand the messages in a global way so that situations and intentions become clear, they also begin to figure out various parts of the messages, and they notice how the situations change along with, more or less in synchrony with, parallel changes in the language: “ball” might change to “balls”, if there is more than one of these bouncing objects around. With more and more speech contacts, they will break the messages down and work out the syntax all by themselves, which, to me, is one of those miracles of evolution, one of those evolutionary feats.

Parents feel that here is a problem for the children.  An intuitive understanding of the French phrase “maman t’aime” (which, when pronounced, could be heard as one word, a three-syllable word) is not enough. Ultimately, the child must not only understand that this is an expression of love (easy), and that it is “maman” who does the loving or where this love comes from (again easy), but the child must also detect where she herself, i.e. the loved person / the person spoken to is hidden in that phrase and must separate it out from the idea of loving. This is difficult because she only hears a continuous flow of language and does not see this phrase in print where the meaning components are clearly separated. It is also difficult because the person talked to is only represented by just one sound, the consonant t. So parents intuitively help the child. Instead of “I love you”, she says “Mama loves you”, or even “Mama loves Mary”, which makes it easier to relate event components to message components, to make a form-function connection and crack the code.

Parents do quite a lot to make the child’s job easier, but it’s the babies themselves who contribute the lion’s share. Little by little researchers begin to understand that babies’ brains make complicated calculations with the conditional probabilities of languages which they are constantly revising to figure out how the language works.  Now in order to do their statistics they need massive amounts of input over many years and they get it. This is the critical mass hypothesis.

But, for a foreign language taught at school there is simply too little input, too little exposure, there are too few meaningful interactions  for learners to make these form-function connections and extract the patterns all by themselves. Teachers must actively and comprehensively support the process of double comprehension and pattern recognition.

Let us look at the French s’il vous plaît. From an English point of view, real meaning and literal meaning differ. When someone hears s’il vous plaît for the first time, they might think that, like its equivalent “please”, it is a single word, only with three syllables instead of one. Only when they see it in writing does it become clear that the word for ‘please’ consists of four parts:

s(i)      il       vous                    plaît

*if       it       to-you                  pleases; or:

*if       it       you (Dative)        pleases

if       it       pleases                 you

Do we need to know this? Yes, we do, if we want to acquire the FL instead of just making ourselves understood for a moment. Knowledge of the structure not only helps us understand why the phrase means what it does, it equips us with a key phrase we can draw on in the future. We can risk new, analogous utterances such as ‘si l’hôtel vous plaît‘, ‘if you like the hotel’; ‘si le vin vous plaît‘, ‘if you like the wine’. And we can understand similar, as yet unfamiliar utterances when we hear them from others.

Let us consider an example from a less known language. In a bookshop on the island of Malta I recognized the image on a book cover – and the author’s name Saint-Exupéry” made it clear: “Il-princep iz-zghir” meant “The Little Prince.” With that we can not only understand the title but we have also half analysed its structure. “princep” must be “prince”. We need a full analysis, though, to move on: “The prince the little” is how the Maltese put it. Now we can also understand phrases like “il-bahar il-mejjet”, “the sea the dead”, i.e. “the Dead Sea” and we might try to construct (or: generate; generative principle) expressions along the same lines:

il-princep  iz-zghir           the prince the little                  The little prince

il bahar     il-mejjet          the sea the dead             The Dead Sea

il bahar     l-ahmar          the sea the red               The Red Sea

A mere situational understanding from the context does not help the language learner very much. He also needs to understand how the foreign language operates, i.e. identify words and recognise patterns in order to be able to make “infinite use of finite means” (Humboldt 1836 / 1963, 477). (Language  … confronts … a truly boundless area, the scope of everything conceivable. It must therefore make infinite use of finite means.”) Without this analysis, learning a formula such as “s’il vous plaît” would remain mere vocabulary drill.

Here is another pattern from the same language, one that a tourist might need:

Fejn hu l-katidral            Where’s the cathedral?

Fejn hi l-knisja?              Where’s the church?

Fejn hu l-ktieb?               Where’s the book?

The printed text seems to give away most of the structure, and the tourist may wonder why there are two words “hu” and “hi” for the one word “is”. Without the literal translation the phrase remains a half-opened package, and we are not yet on the way to grammar. But once we are told that the Maltese actually say

* Where he the cathedral?

* Where she the church?

* Where he the book?

omitting, like many other languages, the copula “is”, it all makes sense in a satisfying rush of insight, and we feel ready to ask our own where-is-questions.

We love it when we get the trick, don’t we, when we get the hang of it.

Put yourself in the place of a beginner. An English tourist looks up his phrasebook. He wants to ask for the time of the day. “Wie spät ist es bitte?” That’s the German phrase for “What’s the time, please?”.  But this knowledge is a linguistic dead-end, unless he can disembed the linguistic elements from the whole. *”How late is it?” – That’s what the Germans say.  Formulas must be unpacked, words must be freed. Only then can we improvise more sentences such as “Wie groß ist es?” or “Wie alt ist er?”

Only after recognising the component parts for what they are, after understanding both what is meant and how it is said, can we analogise on the basis of this insight and acquire a productive sentence pattern. So the generative principle comes into play.

Or take a visitor to China who wants to be polite and picks up the greetings phrase: “Ni hao”.  But only if he can break the phrase down (= *You good), will it become a stepping stone to other phrases such as “wo ai ni” (= I love you), where “ni” reoccurs.

A German teacher, who teaches monolingually, at the end of the lesson often used the phrase “See you tomorrow”, a phrase that the kids didn’t see printed. Most kids got the idea. For them it meant “Till tomorrow” since that’s the natural German equivalent. It’s a time phrase.  But only if the phrase was mirrored in German: *”Seh euch morgen”, countless location phrases such as “see you in the playground”  would also be easily accessible.

Excuse me for multiplying examples for what may seem obvious. But I need to insist here on the generative principle which is largely ignored by the communicative approach or other recent methodologies. It refers to the human capacity to generate an infinite number of utterances from a finite grammatical competence. It reflects what Humboldt thought was the crucial feature of human language, wich is sometimes called compositionality. Meanings are built out of parts and from the way these parts are combined. A finite stock of words or word groups can be recombined again and again to produce innumerable novel sentences ( – and thus, new ideas).

Telling the time of the day provides illustrative examples of how differently languages operate to express the very same ideas. It’s a lesson popular with teachers because the meanings are crystal clear and the lesson can be easily conducted without resorting to the MT.  I remember a French teacher who was practising telling the time in French with a toy clock. The pupils used the time  phrases correctly, but as an observer, one couldn’t help having the impression that some of them had no idea of how the French expressed the phrase “five minutes to eight“  = “huit heures moins cinq”, i.e *eight hours minus five,  which  means five  minutes missing, five minutes to go until it is eight o’clock.  I think they would have welcomed this information of how the French actually construct this idea. Likewise, English learners of Polish would probably welcome a mirrored version of the Polish phrase.

I also observed that learners spontaneously use mirroring all by themselves. In a school in Geneva I overheard a French speaking boy saying: “En allemand on dit ‘le petit bleu poisson’”. The boy was actually trying to make sense of the German phrase, and to make it legitimate, so to speak. Why has this largely been overlooked by our textbooks?

To sum up: Simple understanding from the context of the situation does not lead the language learner far enough. Multi-word utterances must be broken down for us to understand their internal grammar. Only to the extent that we understand both meaning and form can we turn input into intake which can be processed syntactically. Only then can we make utterances of our own which we have never heard before.

2. Re-evaluation of the role the mother tongue

So much for the fundamental principle of double comprehension. Now what have I done? I  have given you FL examples and clarified them via English, a well-known language to you. If I knew Polish, I would have used it instead of English.

Because the perspectival flexibility of a naturally acquired language to elucidate the form-meaning constructions of an unknown language is without equal.

The best window on the logic of a foreign language is a naturally acquired language – usually the mother tongue -  for which we have developed a real intuitive feel.

I‘ve used English here in two ways: idiomatic translations for message comprehension, and for structural comprehension I have used a kind of literal translation which I have called mirroring. The  foreign phrase is imitated in the native language. The native language is bent and twisted and pressed into service to reveal the FL structure.  With your mother tongue, you usually know how far you can go in twisting and bending and adapting it to a foreign structure before it in itself becomes incomprehensible.

These two types of translation don’t figure much in modern textbooks.  Bilingual devices are not mainstream philosophy, not yet.  There is a mismatch between what research tells us and what textbooks do and what  teachers are told  to do. Here’s  just one example. The Internet course Rosetta Stone explains the approach chosen like this: „It essentally means  that you learn German in German, without translations, like you picked up your  mother tongue.“

No, no. no. If this sounds like nonsense, well, it is. We can never duplicate L1 acquisition in the classroom.

Okay, then, Let’s have a fresh look at the role of the mother tongue. Theoretically, it’s all very simple. You don’t learn basketball when all you do is play volleyball. So we can’t learn English by constantly using Polish . We only learn English to the extent that we actually use it. This is blindingly obvious.

But there is another side to the coin. We all must start from where we are. We can only learn a new skill by building upon existing skills, by making the connection. This, too, is common sense, this, too,  is self-evident. It‘ s also what brain research tells us. We‘ ve got to use established neural pathways and then extend and  modify them. So the opposite notion equally carries conviction.

It follows from the first premise that a language has to teach itself. It follows from the second that we must engage the language skills acquired and promoted by the MT.

The allergy to the presence of the MT in the FL classroom undoubtedly comes from its all-too-frequent misuse. „One obvious explanation for the pervasive L1 use in FL classrooms is the low proficiency levels of teachers.” (Krista S. Chambless, in Foreign Language Annals, June  2012). This is not happening way back in Siberia or with underpaid teachers in the African bush. It’s about language teachers in the USA. Well, yes, it’s a real fiasco when less skilled and less proficient teachers simply succumb to the ease of conducting the class in the MT. I can’t go into this here. But the counterproductive, haphazard, inconsiderate and time-consuming  use of the MT in FL classrooms has certainly been a barrier to the true understanding of the issues involved.

Because years of daily encounters with a first language have given the children foundational skills which they need for school. Because years of MT input and interactions have shaped our minds in ways that are overwhelmingly helpful for the acquisition of new languages.

By the time they start with foreign languages at school, children know a lot about language. As they grow into their mother tongue (1) they have learnt to conceptualize their world and have fully grasped the symbolic function of language; (2) they have learnt to communicate, and combine language with body laguage; (3) they have learnt to articulate and use their voice; (4) they have acquired an intuitive understanding of grammar; (5) they have acquired the secondary skills of reading and writing. In acquiring a first language, they have in fact constructed their selves. The MT is therefore the greatest asset any human being brings to the task of FL learning, it is the sharpest tool to cut into the FL and reveal its anatomy. It provides an indispensable Language Acquisition Support System. If learners did not use this support and make the connection all by themselves, in their minds, FL teachers wouldn’t get anywhere. In a nushell: We only learn language once.

In the following, I shall restrict myself   the problem of grammar.

Many grammatical features are directly available for incorporation into the L2 system. However, the path breaking power of L1 grammar is not dependent on the fact that both languages share some grammatical features. MT grammars have paved the way to foreign grammars in ways which are often overlooked. Let us be quite basic here, for the sake of clarity. What could you do with learners who didn’t have the concepts of ‘before’ and ‘after’ in space and time? How can one expect students to understand the essence of the continuous aspect if they didn’t have the notion of duration? Children do have problems with handling time sequences when the second event precedes the one mentioned first: “Before he left, he had another beer”. This is misconstrued as “he left and then had a beer”.  Consider that complex structural areas such as tenses or passives are acquired step by step. In the beginning, children’s strategy of taking the first noun in the sentence to be the agent leads them astray when confronted with passives. In L1 English or German this is gradually sorted out, firstly through cases of semantically irreversible passives such as “The girl was bitten by the spider” which then spreads to semantically reversible passives.  Their  knowledge of the world helps the children  to re-interpret the sentence. However, by the time the child comes to learn a FL, most of these difficulties have been overcome and we need not go the same long way to grammar a second time.

We may, for instance, take it for granted that we can say “my head” and “my father” as well as “my garden”. But we are also mentally prepared to use a different possessive for “my garden”, as some languages in fact do – because it makes sense to distinguish between ‘alienable’ and ‘inalienable possession’. Even if there is no conjunction  “if” in a language, that language can express conditions, and our mother tongue has the means of clarifying these if-less constructions. It is because all languages have evolved means of expressing abstract ideas such as possession, number, agent, instrument, negation, possibility, condition, obligation and a host of others, no matter how they do this, that one natural language is enough to open the door for the grammars of other languages.  It is because “all languages are cut from the same cloth” (Pinker 2002, 37), it is because all languages dance the same dance.

To sum up. It’s all there already. The challenging engineering problems that children have solved as they learn to speak  are coming to light more and more. It has taken children years to obtain the cognitive,   communicative and grammatical  competencies which make instruction possible in the first place, be it maths, geography or another language. Many skills and knowledge sources are available at the FL initial state. They are the base camp from which we all set out to conquer new language territories It makes excellent biological sense for a new language to piggyback on this open channel of communication. All this has been largely ignored by the monolingual teaching philosophy.

3. The generative principle and bilingual pattern drills

Comprehension and explanation are ways of knowing. But speaking is a skill, so knowledge must be turned into a skill. Comprehension must be followed by practice.

I’ll now move on to a particular teaching technique, i.e. to  bilingual pattern drills and show you the generative principle, with MT support,   works in practice. Remember that pattern drills were designed precisely to help the learner make infinite use of finite means, to help him build new forms according to known forms, to use a model sentence as a slot-and-frame pattern for countless other sentences.

In the sixties and seventies pattern drills were all the rage, mostly in the USA, literally filling hundreds of pages with substitution tables and transformation drills. But an error was made. Linguists, i.e. structuralists and their audiolingualist followers claimed that structures were all important and,  in comparison, words were negligible: “Vocabulary comes and goes. It is the least stable and even the least characteristic of the three components of language.“  True, in a way, but there is a misunderstanding.

This becomes clear when we look at how Humboldt conceived this energeia, this quintessential property of language,  its productive potential or creativity. So let me quote Humboldt again, but more fully than before:

‘Language  … confronts … a truly boundless area, the scope of everything which can be thought. It must therefore make infinite use of finite means, and it achieves this through the identity of the power to generate both thoughts and speech.

According to Humboldt then, combinatorial grammar explains the inexhaustible repertoire of language and thought. Sentence variations must be experienced as sense variations, as variations on meaning. Since language has the power of producing “both thoughts and speech”, we do not just generate novel sentences, but new ideas at the same time. After all, this is what language is for: to communicate  thoughts and ideas from one person to another.  Words are not there  just to play around with, but to convey intentions and ideas.   If words are unimportant, pattern drills can easily turn into a self-contained language game, a mere manipulation of forms, with little relation to the world of ideas and events: In traditional pattern drills, meanings were there, but weren’t   focused on.

The problem seems to be that Humboldt’s energeia is usually only familiar in its abbreviated formulation ‘making infinite use of finite means’ and is thus interpreted in syntactical terms only, and not in semantic terms.

This is probably why pattern drills frequently turned out to be mechanical and monotonous. How could practice on disconnected sentences further communicative competence? To transfer or not to transfer – that was the problem. Audiolingual theory was patently wrong in treating words as mere gap-fillers, a convenient device to avoid the monotony of overlearning the all-important sentence patterns. Instead of re-thinking them, pattern drills were dropped like a hot potato. But pattern drlls, after all, tried to capture what for Humboldt and many researchers is the quintessential property of language, which makes it so extraordinarily powerful, namely our ability to use a finite set of symbols which we combine and recombine to create new meanings.  So they should never have been dropped, but re-designed. That’s precisely whatwe have tried to do. Bilingual pattern drills use oral MT cues, and this makes all the difference, because this ways we start with ideas which have to be put into FL words. There is a dual focus, on content and on fluency.  Ideas, not just words,  are played with and structures are manipulated accordingly,  and the semantic potential of a given structure is explored. Unlike conventional pattern practice, lexical substitutions are not regarded as mere fillers.

Without further ado, let’s jump into teaching practice.

This is how I proceed

  • Look for the productive patterns in the basic text
  • Clarification: double comprehension + visual support
  • L1 cues: easy variations / substitutions/ permutations
  • Interesting variations
  • Over to you!   (and pair work)
  • Interludes /creative writing

(Examples of semi-communicative bilingual pattern drills – and further analysis -  are contained in Butzkamm & Caldwell, The Bilingual Reform, pp. 124ff. This lecture draws heavily on this book)

Another objection raised against pattern drills is that they work with isolated sentences, which would be ineffective in principle. But even in the absence of a context of situation, disconnected sentences, if understood, inevitably take on meaning because the listener can immediately imagine fitting communicative contexts. Can’t you imagine those contexts on the spot, spontaneously? Wouldn’t we all know without giving it a moment’s thought when, where and to what intents and purposes we could use these sentences? This is because the sentences express common concepts familiar to us via the MT. All dictionaries operate with illustrative disconnected sentences because we can make those sentences real for ourselves. That’s why our students are not performing language operations in a void. It is high time the myth about ineffective isolated sentences be dispelled. The successful completion of step seven is also evidence for the fact that isolated sentences can lead directly to a context of situation. A bridge is built between pattern practice and creative message delivery.

The bilingual pattern drills which are proposed here meet five different needs:

• The need to spot the pattern or regularity of a sentence, to see the slots into which different words may be plugged.

• The need to find new, analogous items to fit the perceived pattern, thus learning how far one can ride a given pattern and establish its limits.

• The need to entrench and to automatise a given sentence pattern, achieving fluency(these three needs have been met by traditional pattern practice ).

• The need to probe the communicative radius of a structure and explore its communicative potential, i.e. the need for the proper words.

• The need to integrate the new pattern into the existing language repertoire and to employ it freely in conversation.

Here is some more advice:

• Use productive patterns of wide applicability, which give the learner a lot to learn from.

• Give insight into the patterned nature of both the source and target language. Students

recognize the common grammatical thread.

• Begin with simple, easy sentences in rapid succession; move on to interesting sentences.

The MT cues are supposed to trigger abstract concepts which in turn trigger the FL construction. This is the psychological pathway we normally follow when we say anything: from concept to expression. The drills should provide plentiful exposure to the relevant linguistic forms in the shortest possible time. Drills should be followed by creative and more communicative activities as explained and documented by Butzkamm & Caldwell.

Monolingualism: Yesterday’s dogma

I’ve tried to explain and demonstrate to you an activity based on two ideas which were frowned upon, well, condemned as Neanderthal practice, by the mainstream philosophy. The monolingual dogma tried to banish the learners’ native language from the classroom. The communicative dogma rejected pattern drill and disconnected sentences.

Free yourself from these ill-conceived prejudices which have harmed, and not helped, the teaching profession. Try them out yourself, these contrastive, bilingual, semi-communicative drills, experiment with them and find out for yourself in a process of trial & error. As Pawel Scheffler points out in his recent contribution to the journal Applied Linguistics: “Neanderthal teaching often actually works.” 

What has kept the monolingual dogma alive for so long?  To my regret, I have found that a change of opinions is rarely observed in elderly academics. They have committed themselves to the monolingual idea and hate to admit that they’ve been wrong. So far as I know, there is only one well-known expert, Mario Rinvolucri, who has openly professed a change of mind.  The epiphany he describes is revealing. This is what he writes:

“Thirty years ago I was so much part of the Direct Method orthodoxy of the day that I frowned on bilingual dictionaries and one day found myself miming the word ‘although’ in an elementary class! There were brilliant people in the class: one student whispered to another, ‘He mean “but”?’ When I learnt Spanish academically at secondary school, I wore out a couple of bilingual dictionaries in my keenness to launch from the mother tongue into the unclear waters of the target language. In my teenage foreign language work, mother tongue was the semantic bedrock that all my explorations built up from. How had I managed to exclude my real experience as a language learner from my practice as a language teacher for so many years?” Thanks, Mario, for being so frank.

A pragmatic, face-saving compromise, i.e. a relaxed monolingualism with small concessions to some MT support, has only muddled the issue. It is no longer acceptable. The MT, or the language which has become dominant in a learner’s life, is the very bedrock indeed for the learning of FLs. Because it’s the best window into the logic of another language. The MT, the mother of languages, as an 18th century author put it. What is needed is a paradigm shift which, in any academic discipline, almost resembles a change of government or power. The proponents of monolingualism would lose influence. That’s why we can only count on a new generation of scholars to clean up this mess. But we can start now and free teachers from a self-defeating dogma. My hope, my utopia is to see foreign language teaching and learning made much easier than it is now.