As teachers, we’ve all had that sinking feeling in year 11: students don’t seem to know enough words to excel in the reading and listening papers, and are flummoxed by the grammar and sentence structure. Deep down we know that at least some of the marks they do gain are the fruit of guesswork. If you’ve ever felt anything like this, you’re not alone, and we see it in the marks that students score at GCSE. Take 2019 AQA GCSE French, for example, where average marks in the receptive (listening & reading) papers lagged considerably behind marks in papers testing the productive skills (speaking and writing).
|AQA GCSE French Higher Tier
||Average % score
Researchers have identified that there is indeed a vocab “gap” at GCSE: The official vocab list for AQA is about 1,400 lines, yet Milton (2006) identified that students sit the exam with only around 950 words under their belt; progress in acquiring new words stalls dramatically in years 8, 9 and 10. There is no data anywhere to suggest that this situation has improved or is improving. Milton also finds that as a country, we’re getting worse, not better.
So, where are we going wrong? Why can’t students read?
Perhaps the most dominant of all concepts in how we should teach reading in MFL today is that of ‘comprehensible input’. It dates back to Krashen’s core hypothesis that we can create the best conditions for L2 acquisition if we look at how we learn our mother tongue. Today as teachers, we are frequently told by commentators, trainers and people selling us resources that the texts we use in lessons should be 95-98% comprehensible – i.e. contain maximum 1 in 20 (and ideally fewer) new words. If we teach in this way, we are told, this helps students learn the words and structures we’re teaching them, and helps students be fluent. This is the defence given when people like me say that we should be reading more interesting and ambitious texts in the MFL classroom. No!, they say. Comprehensibility is king! The key is comprehensibility and repetition, and this is more important than giving students high quality texts – such is the received wisdom.
Most frequently, commentators refer to the work of Paul Nation to justify this assertion. The problem is, though, that’s not what Paul Nation said. Below I lay out the differences between Paul Nation’s work, and how it has been (mis-)construed in the UK teaching environment.
||Paul Nation’s work in 2000 and 2006
||The UK context today
|The research finding
||Nation found that texts need to be composed of 95-98% known words for students to be able to understand the text: he wanted to know “what percentage coverage of text is needed for unassisted reading pleasure”
||The message has been distorted to say that Texts need to be composed of 95-98% known words for students to be able to learn new aspects of language. Comprehensible input is sold to us as a method which makes learners fluent: that’s not what Nation/Hu said.
|Type and mode of reading
||Nation was proposing that teachers adopt “extensive reading programmes” within their curricula: i.e. students should read a book every two weeks and read 4,200 words per week. He felt that only highly (95-98%) comprehensible texts would be suitable for this vast volume of reading as students couldn’t afford to be slowed down
||We don’t do anything approaching “extensive reading” in UK classrooms: we read less than one quarter of that amount per week at best, and it is very rare to find MFL courses in schools where students are expected to read widely beyond lesson time
||Nation’s research was based on undergraduate level learners of English who had already what we could consider a very advanced level of English – who were looking to acquire much rarer words and phrases. They already had the grammar under their belts.
||Our students have only very basic knowledge of vocab and generally are still getting on top of core linguistic features such as tenses and negation. They’re looking to learn the first 1000-1500 words of the TL – not the 5000th, 6000th, 7000th words.
||There was, has been and continues to be healthy academic debate about the validity of Nation’s assertions. Cobb (2007, 2008), for example, rejects the whole hypothesis of “extensive reading programmes” as a valuable and realistic component of a language course. Schmitt et al (2011) rejected the idea of a 95% “cut-off” and suggested that a larger range of texts have value, along a continuum, depending on context.
||Nation’s work tends to be presented to us as irrefutable fact: it isn’t, and never was.
There are some other key flaws, though, specific to the UK classroom context which we need to be mindful of when we are considering the relevance of the 95-98% comprehensibility rule to UK classrooms:
- Mixed ability classrooms. We know that, even in schools which set by ability, there is a huge variety of ability and proficiency among our 30 or so students. So we might design a text which is 95-98% comprehensible – but in reality…
- for some it will be far too easy and their vocab won’t progress
- for others it will be far too hard – and barely comprehensible at all (sub 75%, let’s say)
… so what this means is that by thinking too much about the text, we sometimes don’t think enough about individual learners and their experience of it.
- A blunt measurement tool. Up until GCSE, most learners are really beginners, and they will struggle with all sorts of things beyond vocab: tense structures, irregular forms, idiomatic usage (e.g. the verb plaire à), unfamiliar word order and unfamiliar use of tense (e.g. depuis, venir de, absence of present continuous in FR). So measuring a text in terms of the number of unknown words is pretty simplistic – and doesn’t help us get a true sense of what might be difficult about a text. Remember, what’s often hard about a GCSE text is its unexpected twist in meaning, it’s distractors or its atypical ordering of information. None of this is picked up in the 95% comprehensibility test.
- The GCSE itself. While people wanting to sell us their resources or training might want us to believe that comprehensible input is the silver bullet (after all, they need us to think that, otherwise we wouldn’t buy their stuff!), the writers of GCSE simply don’t sign up to this cosy consensus. The blunt truth is this: GCSE texts are nowhere near 95-98% comprehensible. Here’s an example from AQA French 2018 (higher tier):
- Have a look at the vocab here: it’s tricky! It’s way beyond 95-98% comprehensible for a GCSE student and contains many words that aren’t on AQA’s own GCSE vocab list. A student who has been taught using highly comprehensible texts will flounder when they read this – as they need very good lexical inferencing skills (Nassaji and Hu 2012) or “morphological problem solving” skills (Anglin, quoted in Hu 2013) to make sense of what they see. Even if they can hazard a guess at the questions that AQA asks of this text, students will struggle to really understand it, and feel very uncertain of themselves. By the way, if you think this all boils down to GCSEs just getting harder and harder, have a look at this GCSE text from 1991:
- Comprehensible input isn’t sufficient
So now for some maths – bear with me! Let’s assume the following:
a) students receive a weekly average of 1000 words input (including reading & listening) in MFL lessons (based on typical textbook coverage, adding a bit for teacher TL use and homeworks)
b) students are in lessons for 38 weeks per year
c) students spend 4.5 years preparing for GCSE – allowing for absence, assessment weeks, missed lessons, etc
d) students need to encounter a word at least 8 times before they really can learn and know it (this is a very optimistic figure – as Nation (2014) points out, the real number is probably much higherThis allows us to project how many words a student might work, through their exposure to comprehensible input, based on various scenarios.
– with an average of 98% comprehensible input, a student would learn 95 words per year and 438 over the whole course
– with an average of 95% comprehensible input, a student would learn 238 words per year and 1071 over the whole courseHow accurate are my projections? Well, Milton and Meara (1998) and Milton (2006) have done research to work out the average number of words that GCSE students have under their belt by testing students themselves – and their figure was approx 850. So my figures tally with theirs, and reflect a use of texts with comprehensibility of around 96% (or slightly higher, if we think students need fewer encounters with a word to learn it) across English schools.The problem with this? Well, AQA’s own GCSE French vocab list is 1359 items long – not to mention all the irregular verb forms. That’s a gap of significantly more than 20%. Using the same model, I also calculate that to reach the 1400 words needed for coverage of the specified vocab list, students would need to read considerably more than they currently do – which seems unrealistic given the limited lesson time we all have. Specifically, using the same assumptions:
– to learn 1400 words based on average 95% comprehensibility, students need an average 1310 words input per week during KS3 and KS4 (this would be very challenging at KS3!!)
– to learn 1400 words based on average 98% comprehensibility, students need a whopping 3275 words input per week during KS3 and KS4: this is palpably not achievable in the vast majority of settings.
So if comprehensible input is imperfect – perhaps flawed – what does the evidence say we need to do about it?
Comprehensible input is still valuable – but as a guiding principle it is flawed. The key point that research consistently makes is that learning languages is not all about repetition. That’s contrary to a lot of what we read or are told: games, hacks and tools which ‘trick’ students into repeating words or phrases over, and over, and over again are widely praised and shared (and sold). Yet the man who came up with the 95-98% comprehensibility test himself (Paul Nation, 2015) says that it is not the quantity of encounters with words that matters most, but the quality of those encounters. He even says that the cognitive benefit of students stopping to read, looking up a word in the dictionary and finding its meaning outweighs the cost of any slowdown in pace of reading that this might incur (ibid.). MFL teaching in the UK really is behind the curve on this one. So if repetition isn’t the be-all-and-end-all, what else is important, according to the literature?
1.Content matters because it is the driver of motivation. If students are interested in reading what we give them, they will put more effort in, and learn more, as Hulstijn (1996) wrote:
“First and foremost, assign learners reading texts that are interesting and motivating. If the text does not alert their curiosity, learners will not be willing to devote the required mental effort to unfamiliar words”.
More recently than this, Macalister (2015) and Day et al (2002) insist that the natural and genuine purpose of reading must be preserved in the MFL classroom. In other words, we can’t just dish up any old bland nonsense because it fits our scheme of work – the text has to be worthy in and of itself: “the purpose of reading is usually related to pleasure, information and general understanding”. As Williams pithily put it in 1986, as his very first core principle of teaching using texts in MFL, “in the absence of interesting texts, very little is possible”. If the text we’re using in class is bland and boring, then getting them to learn is going to be like pushing a train up a hill. I like Haastrup’s (1989) quote below in Hulstijn’s (1996) work; some of it is a little anachronistic now, but is core assertion is right.
“‘Learners will always find out the meaning of words that are important to them’ (p43). If a desire to know a word’s meaning is the driving force, the mental elaboration required for acquisition will then come of its own accord. However it is the teacher’s and course designer’s task to provide learners with easily accessible glosses and learner-friendly, nontedious review opportunities”.
2.The type of task matters: we need to maximise deep cognitive engagement – in other words, get students thinking and working hard, not just quickly and superficially.
Nassaji and Hu (2012) found that tasks that enquire “deeper level of cognitive effort and attention enhance vocabulary learning”. They encourage teachers to maximise students’ “involvement” in the tasks they are doing, both motivationally and cognitively, and encourage teachers to design tasks with the following characteristics:
– motivation to read the text and understand what it says needs to be high – i.e. use engaging texts!
– effort needs to be involved in searching for the meaning of new words
– tasks need to involve evaluation of the topic and the words
Williams in 1986 indeed asserted that reading should be an interactive process, requiring students to respond and react to a text, not just comprehend it. His work is old now, but I don’t think it’s outdated.
3.Proficiency matters: Many researchers have found that more proficient students can cope with more unknown vocab and remember more of it more successfully (Sheffelbine 1990): so the 95-98% comprehensibility rule should be flexible – and definitely lower for our more able students. Frequent encounters and repetition matter more less able learners.
4.How we deal with new words matters: Hulstijn (1996) and others remind us of obne of the basics: Never leave new words to guesswork, always gloss them or insist on dictionary use always follow up new words with intentional, directed practice of words that students have seen whilst reading. If we do this well, we can cope with more unknown words.
5.We can tackle ‘harder’ texts: We can read things which are less than 95% comprehensible and still learn effectively (MacQuillan 2016, Schmitt et al 2011, Sanchez and Schmitt 2010). My hunch is that this isn’t an option if we want our students to get top grades – it’s an absolute must, and we need to adapt our teaching, differentiation and styles of tasks to make it possible. Should we make texts which are 85%, 90% comprehensible the mainstay of our classroom resources? Probably not, but they have to appear sometimes.
Milton (2011) openly advocates for faster progression in MFL vocab in the formative years of y8, y9 and y10. Finding that progress in these years almost grinds to a halt, he argues that this is a consequence of too little new vocab being introduced, too much thematic repetition and not enough inclusion of the less-frequent words which are actually necessary for any sense of communicative fluency. He argues that while for some, this repetition of topics and core vocab might be presented as a virtue, the other side of the coin is that it limits progress and makes for “demotivated learners”. He makes the case for a richer, more varied seam of themes and topics during the secondary year, a healthier balance of frequent and less frequent vocab.
OK so what does this mean for me as a teacher?
Firstly, despite all this, highly comprehensible texts still have an important role and they will still need to be a major tool in our toolkit. But they’re not the only one, and we need to think about how we deploy them.
- how comprehensible – or not – is the text for each student in my class?
- is everyone stretched and challenged in the right way?
- is everyone making maximum progress?
- if I can’t adapt each text for each student (almost definitely the case!), then do I use a variety of different grades of text so that each student has opportunities to surge ahead during a few weeks?
- am I doing activities with the new words which guarantee a high return-on-investment, i.e. activities which are cognitively demanding?
Secondly, we need to acquaint students with tricky texts, right the way through the course, so that they are at ease with this when they sit their exam. This means teaching them lexical inferencing skills (Nassaji and Hu) like:
- spotting key letter patterns and word families (chaud -> chaleur -> chaleureux etc)
- contextual clues (Zahar 2001: does the text give a clue to the meaning of a word? what kind of clues might exist?)
- good knowledge of the topics being taught
Thirdly, it is imperative to select texts which maximise students’ motivation and investment. Williams (1986) urged teachers to give a bunch of classroom texts to students to rate, and ask them to categorise them as “interesting”, “all right” or “boring”. I think we know what students would make of the vast majority of KS3 and KS4 resources. I try to apply these three questions each time I’m choosing or creating a resource:
1.Is this the best text I could use to teach this?
2.Are they going to want to read this?
3.What are they learning, beyond just the words?
So here’s a summary for you of what I hope this blog has suggested:
- reading is a problem at KS3 and KS4
- the original research into comprehensible input is quite different from how it is presented to us today
- there are some drawbacks of comprehensible input and the volume of reading it requires when dealing with the English GCSE, or indeed mixed ability classrooms
- that repetition of input isn’t the only important thing: the content of the text and the nature of the tasks associated with the text are more important
- comprehensible input will always have a role, but we should make sure that we prepare students for the reality of GCSE by teaching students how to make sense of harder texts – and also find ways to teach which enable us to use harder texts where the potential gains are bigger
Do I say this because I care solely about getting students through the GCSE? Well, unsurprisingly, no. I also recognise that real life doesn’t deliver “comprehensible input” – even in things like restaurant menus, weather forecasts and ticket machines. More importantly, I want students to continue with their studies of MFL at A-level and beyond. Students will look at the A-level course spec and see that they have to read a book. Unless they’ve been shown and taught trickier texts throughout KS3 and KS4, and see themselves as skilled, perceptive readers, they will rightly balk at the idea of reading a book in Y12 or Y13. And they’ll think we’re crackers when we say that they’re going to have to read one.
Select Bibliography – click link below