Cream or ice cream or custard?

 We all know that the human brain can learn languages. We learn our first language. And if we live immersed in another language we can learn that language. There are arguments about whether learning a language as a child and learning a language as an adult are different. People even talk about a childhood age limit for acquiring fluency in a language. Even though most people who studied languages at university owe their fluency to a year abroad in their early 20s. There's the arguments around whether classroom learning gives enough time and real immersion for an acquisition model to work. And arguments around whether adults have an advantage in deliberately learning (rather than acquiring) a language. Unlike an infant*, adults and adolescents already have knowledge of the world, knowledge of a first language, perhaps knowledge of a second language, ability to read, self awareness as a learner...

*I love that this word literally means unable to talk

As there are no clear cut answers to any of these issues, the pragmatic view is that there are aspects of language learning which resemble acquisition and there are aspects which are more conscious learning.

On the acquisition side, we learn through comprehensible input, real communication, the music and rhythm and earworms of language. Meaning is paramount, error is natural, and patterns emerge as our grasp of the language expands.

On the learning side, we deliberately learn vocabulary and grammar and then practise using it in ways which increasingly approximate to real communication and expression.

Pragmatic teaching uses both approaches in a balance between input/communication and instruction/practice.

We are aware of the potential and the pitfalls of both. There are plenty of things that we explicitly teach, without it guaranteeing success. Let's say the perfect tense in French. We know you can explain it and practise it as much as you like. But it doesn't mean it will be learned*. Likewise you can expose pupils to it as much as you like. But how much exposure to j'ai mangé would it take for a pupil to acquire the rule that they could say j'ai joué or ils ont joué without thinking about it? So we are prepared for a process of plenty of both approaches, lasting as long as it takes, trying to incentivise learning (through testing) and acquisition (through invitation to self expression) until something starts to stick.

*I am teaching my dog to whistle. Really? What tunes can he do? I said I am teaching him. I didn't say he was learning.

What if there were an area we could put our finger on that we don't explicitly teach, but which learners successfully learn?

I think I have found one.

Stress patterns in spoken Spanish. I nearly wrote "stress rules". But I put patterns. Because I don't teach the rules. But pupils do learn the patterns.

In fact it is generally successful and pain free. The stress patterns on Spanish words is an important aspect, and pupils master it without me teaching it. At A Level, I do teach the rules, but I teach the rules so they can understand when and why a written accent is required. Usually on words they are already spelling correctly with/without accents, but would like to know why. And when I explain, it's drawing on knowledge of language they already have, in order to arrive at the rule. Rather than giving a rule in order to say or write correctly.

I was never taught the rules. I worked them out for myself from my fluent knowledge of the language. In fact there are rules around diphthongs that I still don't know properly. For example when students ask about continúa and continua or the accent on leído, I just tell them there are rules around when a diphthong counts as one syllable or two and leave it at that.

It's an interesting one for the new GCSE. Because the premise of the new GCSE is that we don't test pupils on things that they are not taught. So in the reading aloud, are we just testing the sound-spelling correspondences in the syllabus? Or are pupils also marked for correct word stress?

Well, imagine my surprise when I checked. It IS to be taught.

The Subject Content for the new GCSE - DfE

For a start, I won't be teaching the rule like this. It seems backwards. It starts by saying to stress the last syllable of a Spanish word (unless it ends in a vowel or an s or an n).

When I eventually talk to A Level students about the rules, I would do it the other way round. We start with words ending in vowels. And they instantly realise they have been stressing the penultimate syllable: casa, escucha. Then it makes sense that if you change the ending, it shouldn't alter the stress: casas, casan, escuchas, escuchan. Then you look at words ending in other consonants: hotel, profesor, escuchar. And it all makes sense.

But I still won't be teaching it for GCSE. Even with the read aloud and the dictation.

How important is this example? How many other things are there like this that are best not explained? It seems significant that there are aspects of the language that are easily and unconsciously acquired. With the explicit rule too unwieldy to be of any practical use, even if it helps make sense after the pattern has been acquired.

I know that I have more idea of German articles from memorised snippets such as Alle Kinder schauen auf das brennende Haus* than from the table of cases and genders.

*From an inappropriate 1980s joke trend

Can it be that all learning is actually acquisition? You can't will learning to happen. We don't memorise or understand things by just willing it to be learned. It has to be through an extended process of input, making sense, making connections, tentative use and articulation with other knowledge.

I don't think it means we should swing away from deliberate instruction based on selected and sequenced items of phonics, grammar and vocabulary to be learned. But I do think it means we should guard against swinging away from learning centred around meaning, comprehension and exposure to authentic language. We need to keep our pragmatic approach, designed to keep a balance and a richness to our teaching and pupils' learning. I have many great memories of my friend Liz the bus driver. But the first that comes to mind is always her answer to the question, "With cream or ice cream or custard?" Liz always asked for, and got, all three. So we can definitely have learning and acquisition. And plenty of both.

The new GCSE was meant to stop rote learning in the Conversation...

 My position is: I teach pupils to speak spontaneously, developing their answers. So keeping the Conversation part of the Speaking Exam in the new GCSE was a huge relief. But how will it work out?

From 2011 to 2017 we had the controlled assessment GCSE which totally destroyed language teaching. If you wanted to teach spontaneous speaking, you were letting your pupils down. The interpretation of the markscheme, with its emphasis on accuracy, variety and amount of information, meant that the answers that scored best were long rote-learned scripts. If your pupils were speaking more spontaneously, they would deliver less information, make more mistakes, and have fewer fancy expressions.

So we breathed a sigh of relief when the current (now legacy) GCSE came in, in 2018. The conversation was to be unscripted, with teachers able to prompt pupils for further information and to develop their answers. And it was no longer on one specified topic that could be memorised. It could be on any two of the three Themes, and could not be limited in advance to just some of the topics. There were to be marks for interaction, and less focus on delivering pre-prepared fancy expressions. The examiner's reports have commented on the successful transition from rote-learning to more spontaneous speaking.

Of course there is some use of rote-learning still, as I noted in this post looking at research published by Rachel Hawkes and Emma Marsden. But the exam didn't incentivise this memorised answers approach. So we had gone from a GCSE that actively penalised spontaneous speaking, to one which incentivised it. It was easier to teach pupils a core of language and how to deploy it across topics, than to ask them to memorise answers to questions across such a wide range of topics.

At least one person (very close to the powers-that-were) and a member of the panel responsible for creating the new GCSE, didn't realise this. I wrote in this post how he told me that the reason they decided so bizarrely to almost immediately replace the then brand new GCSE, was the clamour from language teachers on social media. The clamour of teachers finding that they couldn't teach rote-learned answers to such a wide range of topics. Which, if he had been in a position to understand, he would have seen was directly linked to the declared desire of the new GCSE to do away with rote-learned answers.

My current Year 11 are taking the "old" GCSE. When it came to starting revision for the speaking exam, they were gobsmacked that I didn't let them get out their books and look at the "answers" they had written on each topic back in Year 10. Instead I went directly to speaking. Tell your partner in Spanish what you remember from your stories in Year 10, using Spanish you know.

This works fantastically for students aiming for a high grade. They are able to speak Spanish, using their repertoire of language, spontaneously and with good interaction with the examiner. This also works fantastically with students who would struggle to memorise answers or pupils who are not necessarily motivated to memorise answers. The important thing is they get good at delivering answers using Spanish they know, without having to memorise word by word.

So I said to the pupils, Here's a blank piece of paper. Write me your answer. Then you can look at your Year 10 answers afterwards and see if there's anything wrong or anything missing.

And this is the sort of thing I got:

Some of it recreated a story they remembered (especially if it was true) from Year 10. Some of them made up new answers completely from the routines we have embedded of opinions, reasons, stories.

But remember, this is for the legacy GCSE with Year 11. These pupils are taking a GCSE which incentivises having narration, interaction, improvisation. And most of all, being able to construct an answer on any of the topics, rather than rote-learning an answer to each of the possible questions across the full range of topics.

The question is, Will I keep this approach with the current Year 10?

I certainly want to. And I am teaching that way for now. But when it comes to the exam, which approach will win out? Will rote-learning once again out-score spontaneous speaking?

The guidance for the conversation still includes this important stipulation:

So teachers can't narrow down the Themes to favourite topics to be rote-learned. But on the other hand, the content of the exam has already been streamlined. This exam has cut down on the "overload" of topics following "clamour" from teachers. Has it shifted the balance and incentives back towards rote-learning?

Have we let testing Listening destroy teaching Listening?

 I once found the perfect YouTube video for doing actual Listening with my Year 10 class. It was someone complaining about how noisy slurpers spoil the experience of going to the cinema. It was in angry full speed Spanish, with deliberate slurping and crisp crunching thrown in. I described how my Year 10s coped in a previous post:

They understood it was a rant. They understood that she used to like to go to the cinema but that it's ruined by noise from the audience. That it can be the best film or the worst film, and she would love to watch it on the big screen and with great sound, but now she would prefer to watch it at home. They told me enthusiastically what they understood.

They did not understand every word. I did not understand every word. You couldn't in fact hear every word over the crunching and slurping. But that only added to the message rather than detracting.

This is going to be a post about GCSE. But meanwhile three more stories.

I wrote a post about listening to the Extra TV series with Year 7 learners. I stop every couple of minutes and we do a quiz. My questions guide them through the programme and make sure they are keeping up:

She dumps her boyfriend by email. What was his name?

She says "Yes, I got the present." What was the present?

They are watching the action, the interaction, the body language, the tone of voice. And picking up on language, but not language in isolation. The next time when we come to carry on watching, I can start the episode again and ask about language:

What does "C'est fini" mean?

What does "Oui, j'ai bien reçu le coussin" mean?

And they can tell me word for word. The language falls into place because you understand what is going on.

It makes me think of when I was learning a musical instrument. The best bit of the lesson was at the beginning when the teacher was doing some "warming up". (Otherwise known as showing off.) Getting to hear someone playing fluently and brilliantly, making music. How about when it came to French lessons? Our teachers were English and had studied French from books at university in the 1940s. Hardly anyone in my class had been abroad. We didn't do school trips abroad. No-one in my class spoke another language or knew anyone who spoke another language apart from English. When we learned, Monsieur Marsaud est grand mais Claudette est petite, was that the extent of what French was and could be? When did we ever get to hear it in full flow and be amazed at the real thing? We owe it to our pupils to show them what a language is!

And the third story. I was in the Czech Republic on tour with the orchestra of a school where I was teaching. I spoke not one word of Czech and in Karlovy Vary no-one spoke a word of English. We had a tour guide/translator. I witnessed a conversation in the Reception of the hotel which I followed very carefully through a lot of back and forward, as the two speakers went through various stages of consternation, disinterest, insistence, reassurance: The hotel had our booking but they had the teachers in rooms of two. This meant that I would be sharing with a female teacher. They couldn't do anything at the moment but OK, I could sleep in the duty manager's office for one night until it got sorted. I followed the back and forth of the conversation without understanding a word of Czech. But being there in the situation, following looks, tone of voice, actions and gesture, I knew exactly what was going on. Is this related to language-learning? Well, firstly I imagine that 5 years of living abroad must have given me confidence in my ability to tune in to what was going on. And secondly, language-learning emerges from exactly this kind of witnessing interaction. I didn't have time to pick up much Czech from one conversation, but much of my learning of Spanish and French would have happened in exactly this way.

So what we've been looking at is the question of whether we arrive at understanding of meaning by parsing known words and language and working out what the meaning is. And the answer clearly is that we also do the opposite. We grasp overall meaning, and so can get some grasp on the words.

The problem is that our language teaching in schools now, is so governed by testing, that this approach to language as the actual language itself and the ability to cope with it, has almost completely gone. We test pupils' knowledge of known words and known grammar. We teach a bottom up approach of putting together known words and known grammar, to test how well pupils can recall and put together specific known words and known grammar. And we want the texts they see and hear to model how to put together specific known words and known grammar.

The idea that Listening is in some way different to Reading or Speaking or Writing, and is a skill to be developed, is being denied. The word "modality" is deliberately being used to replace the word "skill". The testing of Listening is declared as a process of transcribing known sound-spelling links into known words and known inflections, which are parsed so that you arrive at meaning. The idea that the skill of Listening is to make sense of something when you don't necessarily understand all the individual words, is now lost. Even though when I watch the news in French, I'm not sure I do understand all the words. Or when I watch a film in Spanish, if I catch myself subtitling every word (whether that be in Spanish or into English) I give myself a stern talking to and stop doing it, in order to just be in the film. (This is also the mistake that people make when reading books in a language they are learning.)

You can really see this in GCSE listening. All the things that make it an actual listening are removed. It's read aloud, deliberately keeping tone of voice clues to a minimum. It's slow pace, with no natural interaction or relationship between the speakers. The content is often slightly off-beat, to stop pupils from using assumptions or deductions. The lengths that they have to go to, in order to strip out all the listening cues and use of actual listening skills, is what convinces me that these skills must actually be real. If something was imaginary or illusory, you wouldn't have to remove it.

Then there's the markschemes. Remember my Year 10s who understood a full speed angry Spanish YouTube video in great detail? That it was a rant, that it could be the best film in the world, that you wanted to see it on a big screen with great sound, but that in the end you prefer to watch it at home because of the slurping and crunching? Well that is not what is wanted. Because the markschemes are constructed to make sure that's NOT what gets the marks.

We have so many examples of how what appear to be comprehension questions are not. Because a pupil who gives a correct answer to the question gets 0 marks. What you have to do is show your knowledge of known words and known grammar.

There's these, from memory. Some from Listening, some from Reading:

What impressed her about one school?

She was impressed that one school grew fruit and veg on the school field. Nul points.

She was impressed that they grew fruit and veg on PART OF the school field. Correct answer.

As if that's the part that impressed her. The part that impressed her was that it was PART of the school field. See what I mean? That's not a comprehension question. That is a directly transcribe and translate word by word what she said otherwise I am not giving you the mark question.

He doesn't get on with his teachers. Nul points.

He gets on badly with his teachers. Correct answer.

Comprehension gets you no marks. Direct translation of all the words is what they think listening is.

It's good for your skin. Nul points.

It looks after your skin. Who even says that? AQA. That's who.

She gives talks to pupils about energy saving measures. Nul points.

She gives talks to pupils about HOW to save energy. Correct answer. According to AQA.

This isn't an accident or a quirky markscheme. This is how they see Listening.

I would actually prefer if they got rid of fake Comprehension questions. And just made every question a translation question. And sometimes I feel as if in my lessons every activity is really a translation activity. And I have to make an effort to bring in language that is not directly modelling known words and known grammar, and let them see and hear real French and real Spanish. Or hear me talking to someone in full flow, just like in my musical instrument lessons.

To be generous, we could say it's how they see testing Listening. And it's our fault if we drag that into our lessons and let it deform our teaching. But I'm not being generous when we've been told that Listening isn't a skill anymore. When we're told that Listening has to be made up of known words and known grammar. When we're told that you arrive at meaning by parsing the words in a one way street from decoding to meaning. In my opinion, it's not true, and it's not language-learning. You may not agree with my conclusion, but please at least don't ignore the question!

The New GCSE Photo Card

 This half term we started working on the Photo Card for the new GCSE. The textbook included it right from the first unit, at least in the assessment pages, but we have delayed it because it requires such different language to the rest of the GCSE.

In an earlier post on Describing, I wondered if "Describe" in the new GCSE was going to replace the emphasis on "Narrate" in the old GCSE. I didn't use to teach description as part of the old GCSE. We concentrated on building narration around opinions, reasons, if sentences, conflict of opinion, argument, decisions, what was happening, what happened, disappointment and hopes for next time. Description just wasn't an important aspect of the old GCSE, either for tasks or for the markscheme. As an aside, I will say that although narration is no longer mentioned in the markscheme for the new GCSE, I am still teaching it because pupils will have to talk for 5 minutes on one topic in the Conversation part of the exam.

But the new exam, with its long list of adjectives, its "describe your favourite celebrity/teacher/friend" questions in the Role Play, and the much longer photocard description... seemed to push teaching more towards describing.

So we concentrated in the first term on getting very good at the core of language for developing answers using opinions. And decided to delay working on description and on the photocard until term 2.

What have we discovered?

I was correct in thinking that describing has a lot to unpack. The verb to be is highly irregular in different persons and tenses. There's ser / estar to be dealt with. And adjectival agreement. Plus the verb to have.

The thing is, is this really what's required for the photocard? How comfortable are we with an obsession with looks, focusing on eye colour, hair colour, skin colour? Is that what we want from the photo description in the exam: There are three people in the picture. One is black, one is white and one looks as if they have Asian heritage. One is quite slim but the other two are not. One has dark curly hair, one is blond with blue eyes... Where is this going?

So although we have done physical description, and the pupils enjoyed doing and writing detective puzzle scenarios, when we have come to do photo descriptions, not one pupil has gone for physical description.

So what are the things that we have done?

Firstly a top tip. For AQA, there are 2 photos and the pupil has to talk about both. Even if it's only a quick mention of one of them. But what if they never get round to talking about the second photo? They lose track of time, or they get into a muddle on the first one and give up in a panic, or do so well on the first one they think they've done everything, or just forget to talk about the second one? So here's what we've decided as a class: Mention both photos straight away. There are two photos. In one photo... In the other photo... You always say this before getting stuck into the detail of one or other or the photos. At least that way you know you've mentioned it.

Then we come up against the language needed for the photocard. In the old GCSE, it was very much a stand alone task, using language that was only needed for the one question "What is in the photo?" This was a pain, having to deal with a totally different repertoire of language just for one question. Generally this meant learning There is / there are X people, they are in X location, they are doing X and they are doing X. This language wasn't useful anywhere else in the exam, which was almost entirely based on first person opinions, reasons and examples. With the odd incursion into we. And only the highest achieving candidates venturing much into he/she/they.

One way round it was to do the photocard in the first person. After all, why are we showing this photo and talking about it? I am in the park with my friend. Look: I like to play tennis. It is sunny so we are going to have a picnic. My friend said, "I love cheese." 

We experimented with this first person approach with the old GCSE photocard and it never felt comfortable. Just as pupils don't go for "She is short and a bit fat with long blond hair and blue eyes", neither did they go for talking about the photo in the first person. It felt as if the scenario wasn't one of showing a picture to a friend. It felt like an exam where you were being tested on your ability to give some formulaic sentences about a picture chosen to put you under pressure. Not surprising, given this was indeed exactly the situation.

Another way round it was to use the imperfect. Talk about what was happening in the picture. In Spanish, pupils love the imperfect with its aba/ía endings which work just as well for 3rd person as for 1st. In French it's a bit more tricky, because if you say il jouait, it's difficult not to sound as if you are making a mess of saying "il joué" or something that the examiner won't like at all. And while we are on the topic of French, let's not even start on the lack of a present continuous and how it all goes wrong as soon as a pupil tries to say They are... working. But that's what I meant right at the top of this post when I said that the photocard requires very different language from the rest of the GCSE.

The problem with the old GCSE was that you had to do this different language just for one question where the pupil had to give a shortish formulaic answer.

With the new GCSE, they have to talk a lot more on the photos. Around a minute, which is a long time. I think this counts as another important insight from this half term. Try it yourself. Pick a photo and see how long you can talk about it using the sort of language pupils have at their disposal. What on earth is there to say that can keep you going that long. Especially if you are steering away from physical description.

I actually think this is a good thing. Learning to use 3rd person singular and plural and how to say they are ---ing, and there is... for just one short answer in the whole exam seemed unwieldy. Now it is for a much longer answer and has to be done properly. It is a big feature of the course. And what we need to do is to find ways to use the language learned for the photocard in other areas of the exam.

Time for some examples.

I created some model answers for 2 photos and mixed them in together. Pupils had to separate them out.

Then next lesson, pupils had to talk about the photos using the structures that they remembered. And do a written version. 

You can see by the time they come to do the written version, it's changed from the language I gave them in the first lesson. Including mixing using the present continuous and the imperfect.

Then they had to recycle as many of the structures as possible to the next set of photos we used. Again, you can see that there's some formulaic aspects as to what to include and what language to use, but also the pupils are using their knowledge of the language to create their own descriptions, rather than relying on set structures:

Some things to note:

• You can see the initial In one picture... In the other... tactic.
• I worked a lot on pupils doing this in spoken Spanish and spontaneously. But in the exam, they do have time to write down answers in the preparation time.
• "They seem" is useful. It avoids the ser / estar problem, and also allows you to use personality adjectives. They seem kind.
• They are using a mixture of present tense llevan uniforme, present continuous están comiendo and imperfect cocinaban. Different pupils have their personal favourite.
• Using the present continuous can get you into trouble: están llevando uniforme doesn't sound right.
• I am still not sure this is enough for a minute.
• We will have to keep an eye on the exam board guidance on marking this task. Marks are for the amount of information conveyed clearly (code for accurately even where accuracy marks are not allocated). So may be just "There is" and a list of nouns would be enough.

The secret of this is going to be to keep using the language features we've worked on for the photocard in the language we use for other aspects of the exam. The imperfect is definitely part of our core repertoire. Using 3rd person is something I'm keen to do much more than with the old GCSE. Description is something we can build in to all topics. That leaves the present continuous. Which just seems to be used for this one task. At least with this GCSE it's for a more substantial task than the old GCSE. Introducing a whole tense just for one utterance didn't feel worthwhile. I think we were right to delay introducing this until pupils had a strong core repertoire under their belt. I think you can really see this in how quickly they were able to pick up the new language for describing pictures and run with it. And what I can now do, is make sure describing a photo is part of something we do regularly across topics, so none of this language gets forgotten.

There is an alternative version of this which I don't want to contemplate because it is so awful.

The AQA markscheme for the photocard is entirely based on "clarity" and "amount of information". Clarity sounds a lot like a way of marking for accuracy even though accuracy marks aren't supposed to apply here. And lewizrs.bsky.social has tipped me off on bluesky, that in AQA training, "a lot of information" has been quantified as saying 15 things. So for full marks, our pupils need to say 7-8 things about each photo, keeping absolute "clarity".

As we get nearer to the exam, this could mean simplistic formulaic responses are what works best.

There is a man. He is big. He is fat. He is blond. There is a woman. She is big. She is fat. She is not blond.

Repeat for the other picture. Full marks every time.

This reminds me of going back to the disastrous Controlled Assessment GCSE where AQA's interpretation of "amount of information", along with "variety of language" meant that rote learned answers scored higher than teaching pupils to be able to speak. 

You can see in this post, the approach I am trying to take. Where pupils develop a coherent body of language that they can deploy across topics and across the exam tasks. Avoiding silos and rote learned single use language. What if AQA's interpretation of the markscheme means this approach actually penalises pupils in the exam?

Here we go again?

One thing that costs nothing which would revitalise languages in England.

A sure-fire way, costing no money, to revitalise and reinvigorate take up of MFL at GCSE.

Do you see what this is?

It's anonymous targets for a random bunch of pupils in different GCSE subjects.

These targets may be given to pupils as they start their GCSE course. Or perhaps when they are considering their options to give them an idea of what they could be aiming for in each subject. Some schools may even give them to pupils in year 7. And they may be used throughout KS3 and KS4 to assess whether pupils are "on track" or not.

They are based on the pupils' KS2 SATs results in English and maths, and how this correlates to how pupils with similar scores go on to perform at GCSE. And by the law of averages, for the thousands of pupils in the year group, there is an overall correlation on average between the predictions and the GCSE grades awarded. Although you might want to look more carefully into just how many (how few) individual pupils get their "target" grade. Because the words "on average" includes... surprise surprise... some getting above and some getting below target. And of course there's another reason why the correlation is a good one: the same data is used to allocate the overall number of grades given. So it's not a coincidence, or a function of the correlation or averages. It's deliberately calculated that way.

What these targets do NOT show, is any reflection of an individual pupil's performance, past, present or future, in a language. The targets are not based on any knowledge of the pupil's study of languages. They are no reflection of the teaching they receive. No reflection of any government initiatives or new GCSE syllabuses or Ofsted Research Reviews. None of that will have any impact. They are set before they start their GCSE course; in fact they are determined before they even start KS3.

But they are a true reflection of something. They very clearly show one important thing.

Look down the column for the target for French. And compare it to the other subjects.

Look at the French column. It's striking. This is not a marginal difference.

It is lower for French than for the other subjects in the range of grades 5 and above. For a pupil targeted a grade 3, it is more in line. But the higher the target grade, the more likely it is to be a grade lower than in other subjects. A grade lower. Not a statistical tendency somewhere in the scale of things. A grade lower compared to other subjects.

This is not a reflection of the pupil's performance in languages. It's nothing to do with the content of the course or the difficulty of the exam. So what is it a reflection of?

It is a true and accurate reflection of the fact that lower grades are given out in languages. It's right there, correctly presented to pupils to help them inform their options choices. And confirmed to them by the reports of exam results from older pupils, siblings and friends.

We can tell the pupils it is no reflection on them. And that the target grade doesn't limit what they can get. Although by then the damage is already done. The targets speak for themselves. 

But if we tell them the truth, we should explain to them that if they pick a language at GCSE, although no reflection on them, the targets do accurately indicate that they are likely to be given a lower grade than in other subjects.

FFT have reported on the statistics of severe grading in MFL many times over the years. But to see it so starkly in the targets given to pupils drives home the fact that this is not some marginal statistical issue to be debated and tweaked by experts in ivory towers. It's right there in black and white in the targets given to pupils: "You are likely to get a lower grade in languages than in other subjects because that's how many high grades get given out."

So before we do ANYTHING else. Before we have a subject review, or national initiatives, or hubs, or an Ofsted Research Review, or a new GCSE. Before we spend another single pound of tax payers' money. Before we publish any more articles lamenting take up of MFL, or have a debate in the House or a committee. There's one thing we can do. Stop downgrading the performance of our pupils compared to the other subjects we compete with.

If grades in Modern Languages were given out in line with other subjects, no-one would be clamouring that they ought to be downgraded because standards of teaching are lower. No-one. It's a historical anomaly that we got stuck with.

If an exam one year happens to be easy, and too many pupils get high marks, they use the grade boundaries to bring it into line. If one year an exam happens to be a bit difficult and not enough pupils get high marks, they use the grade boundaries to bring it into line. Except for modern languages. Where every year they use the grade boundaries to make sure the grades are out of line with other subjects.

And then wring their hands over woe is me, what is to be done.

Not one penny would it cost to put this right.

NB There is an argument that it's hard to get MFL teachers, therefore the ones you get aren't always going to be the best. So therefore that is why French grades are lower.

Firstly. This is untrue. This is not the reason grades are lower. Grades have been lower for decades. It's a historical anomaly, nothing more.

Secondly, it is also hard to recruit specialist maths teachers. People are not crying out for maths grading to be less generous. Or physics.

Which leads to one conclusion. This is sexism. The subject singled out for lower grading is the one most associated with female teachers and female pupils. What is the statistic? One state school boy in 10 gets a "pass" in GCSE French?

They know the grades are unfair but they keep them that way. Would they do that for maths? There's no "would" about it. They DON'T do that for maths. This blights the careers of teachers and the futures of our pupils alike. There really needs to be a class action against the government if they persist with this clear and pernicious discrimination.

The view from the classroom

 I don't know if it is the time to be going over the old arguments about the proposals for the new GCSE. Especially as we are busy working out the ins and outs of the new exam, which is in some important ways different from what was originally proposed. Before Christmas, a paper was published by Emma Marsden and Rachel Hawkes, looking back at what was proposed and how the proposals were received. Balancing evidence-informed language policy and pragmatic considerations: Lessons from the MFL GCSE reforms in England. This link takes you to the abstract. To open the full document, use the links that say "Access to Document." It's very readable and I would encourage you to read it.

It puts the case for the proposals very clearly and tries to address what it sees as some misconceptions. It is also clear from the introduction that the new GCSE was seen as a way to reform not just assessment, but teaching. It stresses that the GCSE assessment  (whether one likes it or not) heavily shapes curricula, materials, and pedagogy. So it is important that we continue to engage with the thinking and research behind these proposals.

You can see the arguments for this from both sides. But what it actually comes down to is what happens in the classroom when you try it out for real. Whether it has an impact, pushes us towards changes, and actually works out as intended. Which is the reason why I am writing this post, although I am somewhat conflicted as to whether I want to. So here goes...

The paper makes the case that the vocabulary list in the legacy GCSE was not central to the exam or to teaching. Teachers did not build their curriculum around the GCSE list, and the exam boards did not deliberately design the exams to test knowledge of all the words in rotation over a few years of the lifetime of the exam.

While this is true, it wasn't necessarily a problem. Or rather, the exam boards and teachers didn't turn it into a problem. Teachers taught the words that pupils needed to perform the tasks and topics required by the exam. Without consulting the vocabulary list. In the speaking and writing exam, what mattered was the pupils' ability to use their language to express themselves, rather than testing knowledge of items from a list. Apart from in the translation questions, where knowledge of the specific word in the text was required, and this isn't something that has changed. In the reading and listening exams, certainly for AQA if not for the other exam boards, the key to the texts and questions wasn't a mass of low frequency topic vocabulary. It was words like often, some, always, other, ago. And the focus on exact rendering of word by word language, rather than an answer to a comprehension question was the biggest criticism of these exams. Which seems likely to be exacerbated not alleviated by the proposed "new" approach.

So the paper is right in that the vocabulary list wasn't central to teaching or to the exam. I'm not sure that this was ever a problem. But if you do think that it was absurd that the list of vocabulary wasn't central, and thought it was reasonable to specify all the vocabulary to be learned, then you can see the case for reform. Does this mean that the new vocabulary list should therefore be determined by frequency?

I can see the arguments. The words that get used all the time are worth learning because they get used all the time. Emma and Rachel base their work on research that shows that the very highest frequency words are common to whichever corpus you use. And the words are a good fit for continuing to A Level and for reading authentic resources. 

I think this last example is something that we really ought to be exploring. In the 1990s and 2000s when reading for information and pleasure and "adapt the task, not the text" was in vogue, we encouraged pupils to spot cognates and topic words from context. And maybe told them to gloss over the little words in between. Or thought that pupils would pick them up from context eventually. Whereas now, if we concentrated on the little words in between, we would find that the topic words and cognates are just as obvious as they always were, but equipped with the high frequency words, pupils would be much more able to access authentic texts. This is something I'd love to come back to. Although in the debate about reforming the curriculum, the Ofsted Research Review warned us off authentic texts, because of its focus on language as modelling known words and grammar, rather than for comprehension or interest.

But in the classroom, teachers are finding it very hard to adapt to the high frequency vocabulary approach. We are not used to following a list of vocabulary and constructing a syllabus out of it or constructing texts out of it. It's not that we have swapped the old vocabulary list for the new one. It's that we have moved from basing our teaching on the tasks and topics that pupils are required to deal with, to basing our teaching on a list of words to be tested.

We were quite happy teaching the words pupils needed for a topic. We were quite happy teaching pupils the words and structures they needed for the exam tasks. And the thing is, we are still doing that. But with the extra worry of whether we are teaching the right words.

For many, it means no change. Apart from the creeping worry. For others, it means we are incentivised to try to gain an advantage by doing some of the things the new approach favours: paying greater attention to recycling words; rehearsing what pupils can do with words on a list, rather than saying what they want to say; making sure we cover words from a list in any context we can shoehorn them into; making sure we work across topics. Some of these incentives are positive and some are negative. But recycling vocabulary across topics was already a feature of strong teaching under the legacy GCSE.

It also gives rise to absurdities. If we are still teaching topics based around pupils' personal likes and dislikes, do we discourage them from learning words like skating, chess, chicken which may not be on the list? Or do we go against the whole point of the reforms and add more words which aren't on the list? There seems to be a serious mismatch between the tasks and topics and the language we are given. And I still haven't got to the bottom of the mismatch with the grammar. If we have to teach jouer à and jouer de, are there actually enough sports and instruments on the list?

And it raises interesting questions about the notion of difficulty in language learning. Some of the words on the Higher list are cognates. The difference between Higher and Foundation isn't about difficulty. It's about the number of words. It's interesting to think about this and whether we can control and segregate our pupils' exposure to words according to a list.

Even if you were to insist it makes sense for the GCSE to be built around a vocabulary list, wouldn't it make more sense to make sure the list matches the tasks, topics and grammar that pupils will be required to demonstrate? The paper seeks to argue that this question is a misunderstanding of the power of the high frequency vocabulary. But in practice we are finding that we are very much teaching words which are not on the list, in order for pupils to perform in the topics and tasks.

One final spectre on vocabulary that this paper raises, is the idea that the exam boards should be held to the stipulation that they should test the full range of vocabulary on the list over the lifetime of the exam. This risks meaning that the exams quickly exhaust the more obvious topic words, and that the exams look very different from year to year, and start to look very different from the texts in the coursebooks. When I've mentioned this before, representatives of the exam boards have been very puzzled by the idea and told me that it's not something they are contemplating. So I think we can stop worrying about that.

Another issue the paper raises is the Conversation part of the exam and the use of memorised answers.

This is the part of the debate that leaves me most bewildered. I am someone who prides himself on teaching pupils how to develop extended spontaneous answers in speaking and writing. The old Controlled Assessment GCSE which forced us to abandon this in favour of pre learned fancy answers was an absolute disaster for language teaching. The legacy GCSE started to see an end to this. Some teachers were quicker than others to start to make the switch, and rote learned answers had been embedded from earlier in the curriculum. All of this needed to be undone. The breadth of topics and the requirement for interaction of the examiner, responding to the pupils' initial answer, mean that pupils' ability to use their language flexibly and across topics is winning out over the idea of pre learning answers to every possible question on every possible topic.

The paper quotes research saying that a large proportion of pupils still do some pre learning of answers. This isn't the same as saying that pupils learn and regurgitate a fancy answer word for word. It may well be that faced with a high stakes exam, many pupils will want to be well prepared. It may well be that one of the steps towards spontaneity and fluency, is to have pre learned answers to some questions. It may well be that the pre learned answers are adaptable, flexible and not actually even used in the exam.

I have had pupils who thought they were doing the right thing by trying to memorise answers for the speaking exam. And it doesn't mean it goes well! They are focused on word by word regurgitation and recall. The first thing I do is interrupt them and push them away from these answers, and then they can start to talk from their repertoire and routines and make a much better job. The examiner's reports for the legacy GCSE comment on how pupils who can do this are much better equipped than pupils who have attempted to memorise answers. Keeping the Conversation, where the teacher can ask open questions and follow them up with further prompts, was the most important thing we have managed to do in engaging with these new proposals.

The proposed alternative version of "spontaneous speaking" meaning responding to unexpected questions, is one of the trickier parts of the new GCSE. We know from the Role Play in the legacy exam that the unexpected question is tricky. In an exam situation, pupils pick up on a key word from the question and give an answer which may or may not fit the question. It's a short answer, often panicky.

The new exam has four of these questions. And the expectation is for an answer with some development. We are working hard on these. My pupils have a greater familiarity with question words than previously. But we shouldn't underestimate the processing load required to accurately understand the question, formulate a sentence response, attempt to give some extended detail. Some of the questions don't help. "Where is your school?" could induce bafflement in a pupil in an exam, where both interlocutors are sitting in said school. And how to develop that answer on the spot is not clear to me at all. These questions seem designed to catch pupils out, rather than offer a platform for them to develop spontaneous speaking. This vision of "spontaneity" is from the point of view of testing pupils' knowledge, rather than their ability to use the language.

And I think that's the key. The research paper finds plenty of evidence to back up its own point of view. But it is firmly rooted in one vision of language learning. It wants to be a fair test of pupils' knowledge of what they have learned. To specify what should be learned, and to test that knowledge and to some extent the application of that knowledge. A perfectly reasonable point of view. But one that isn't so simple in the classroom when you are dealing with real pupils and real language learning. Is this the problem? That acknowledging that the GCSE (whether one likes it or not) heavily shapes curricula, materials, and pedagogy led to an attempt to deliberately shape what is learned and how it is taught. And for us in the classroom, the experience of being reshaped feels awkward, confusing and sometimes painful.