Are we raising a breed of ‘dyslexic’ foreign language learners?

All teachers who have taught dyslexic children at some point in their career will know how challenging it can be to keep them engaged and motivated, how low their self-confidence as language learners usually is, how frustrated they often get as they struggle to make sense of what they read. But hang on, doesn’t this description also apply to quite a few of the language learners we teach?

Well, in fact, one may argue that many of our students do exhibit a form of deficit in their foreign language competences akin to a disorder called ‘Phonological dyslexia’, described by Ellis (1984) as the inability ‘to read unfamiliar words or non-words aloud, suggesting impairment of grapheme-phoneme conversion and/or phonemic assembly’. In other words, ‘Phonological dyslexics’ are individuals who are not able to sound unfamiliar words out. This inability to convert graphemes (letters) into phonemes (sounds) – unsurprisingly – seriously impairs these individuals’ reading comprehension skills.

Erler (2004) is highly indicative of this state of affairs. In her study of 359 year 7 students of French (from two middle-achieving English comprehensive schools) she found that after one year of instruction (with two contact hours a week) their knowledge of spelling-sound rules was poor, both schools obtaining the same mean score at the rhyming-word test she administered, i.e. 2.75 correct choices out of 14. She concluded:

The results seem to indicate that, with a few exceptions, pupils had little idea after one year of learning French about spelling-sound rules for principal vowel sounds in the language and for the general rule of silent final consonants. These are key rules for being able to decode from print to sound, and are essential for reading comprehension (p. 5)

What is equally interesting – and tragic – is that only 7.6% of Erler’s informants reported they felt ‘happy’ when reading French aloud in lesson; the vast majority of the students felt negatively about decoding and spelling written French.

Another interesting finding by Erler was that 75 % of the total sample Year 7s thought that it was indeed useful to know pronunciation and 63% stated that they were aware of subvocalizing to sounds when reading in French, confirming what we know about automatic phonological loop activation in working memory during reading (in simpler words: whether we are aware of it or not, the brain automatically converts letters into sounds as we read, even when we are not reading aloud).

These findings are dispiriting for several reasons. Firstly, because, as much L1 and L2 research clearly indicates (e.g. Stanovich, 1980; Bryant and Bradley, 1983; Sprenger and Casalis, 1995; Gathercole and Baddeley, 2001) decoding skills are crucial to comprehension of written texts and poor readers often exhibit serious deficits in their deployment. Heaps of research indicate that a child’s early knowledge of phonological awareness is a strong predictor of their fluency in later years (Stanovich, 1986). Muter and Diethelm (2001) found that students of French as a foreign language who were able to tell where syllables in a word begin and end were more proficient readers than those who didn’t.

Secondly, as I have argued in many of my posts and will write about more extensively in my forthcoming book ‘Breaking the sound barrier’ (Conti and Smith, 2017), phonological awareness is crucial in listening comprehension in that it helps the brain make sense of the speech flow by identifying word-boundaries, intonation patterns, etc.

Thirdly, and more tragically, as Erler’s concluded, the students she investigated exhibited decoding deficits (in French) comparable to those symptomatic of dyslexic reading impairment. Now, students who are de facto dyslexic are less likely to have high levels of can-do attitude and self-confidence in language learning, both strong predictors of success in language learning (Macaro, 2007). Could this be one of the reasons why many of our students don’t enjoy learning languages? Could addressing this major deficit enhance their motivation? I believe so.

Erler’s findings chime with my own experience as a veteran MFL teacher with over 20 years’ experience at primary, secondary and tertiary level. Decoding skills – also referred in the literature as GPC (grapheme-phoneme correspondences), spelling-sound rules or Phonological Awareness  – are not duly emphasized in British schools and, when they are indeed embedded in the curriculum, instruction is undermined – in my opinion – by the following shortcomings which serious limit its effectiveness:

1.Decoding skills instruction is not given sufficient prominence in the curriculum;

2.It is often an add-on; it is not fully integrated with the curriculum content and goals;

3.As per Dr Rachel Hawkes’ approach, phonics are often taught through gestures and tongue twisters, but the target phonemes are rarely consciously and systematically recycled in the lesson through listening, reading and speaking tasks which aim at their organic acquisition. This approach limits the acquisition of the target phonemes as it is divorced from fluency across all four skills and from real-life-like communication. In my approach the target phonemes are consciously recycled in every single task (both receptive and productive) I stage to teach the topic-at-hand (be it grammar, communicative function or vocabulary);

4. Instruction rarely ventures beyond word-level practice, which is not conducive to acquisition – the input we process and the output we produce usually contains more than one word…;

5.The target graphemes are usually sequenced randomly without considering (a) the level of challenge they pose to the learner; (b) how their teaching contributes to facilitate other aspects of L2 acquisition, such as grammar – for instance: (in French) a focus on letters that are silent earlier on in the instruction process will serve the purpose of assisting the acquisition of present tense forms. Instructional sequences ought to be based on a (possibly evidence-based) rationale;

6.Each target phoneme is not usually focused on for sufficiently long periods of time and recycled consistently and extensively across the curriculum;

7.Phonemic awareness skills (see picture below) are not focused on explicitly in the early years of L2 instruction, yet I have found that primary and year 7 students benefit greatly from practising them and research shows clearly that they prime the connection of sound to print.

Figure 1 : the Phonemic awareness development continuum (Courtesy of University of Oregon)


8.Students with poor phonemic sensitivity are not identified at the beginning of the course, yet I found it extremely useful to have a good idea from day one as to who was less gifted in this area of language aptitude (the natural predisposition to decode letters and repeat and manipulate sounds);

9.Students are rarely – if ever – tested on their decoding ability or phonological awareness (to assess progress in these areas). This is a serious shortcoming considering how pivotal this set of skills is for language learning effectiveness and success; I found that including opportunities for assessment (e.g. old school dictation or short transcription tasks) has increased my students’ focus on decoding skills and their motivation to learn them.

These and more common shortcomings of decoding skill instruction will be dealt with in greater depth in my next post.


Many foreign language students in England appear to have poor decoding skills. This hinders the development of their reading and listening fluency whilst seriously denting their confidence. As I have written in my forthcoming TES article ‘Enhancing MFL learner motivation – the road less travelled’, one of the most important reasons why our students lack confidence and motivation may relate to their inability to make sense of the target language, be its grammar or its decoding/pronunciation.

One of the group of students I have been trialling my decoding-skill training program with reported to me the other day that they were so much happier to be finally able to read out written French following a set of specific spelling-to-sound rules. They felt empowered by the decoding pinciples they had been taught, as the constant -often random – guessing frustrated them. I do believe that in languages like French and English, where spelling-sound correspondence can be challenging, a students’ sense of efficacy as a decoder can substantially enhance their motivation.

In conclusion, much more effort and thought should be put into effective decoding-skill instruction, which should go beyond the teaching of sounds through gestures, a few tongue twisters and listening or singing along to song; a principled framework should be arrived at, which integrates phonics instruction organically and systematically with the teaching of grammar and vocabulary and extensive practice across all four skills to ensure long-lasting retention and automatization.

To find out more about our ideas on decoding skills instruction get hold of our book, The Language Teacher Toolkit,


Spontaneous talk revisited – courtesy of Pearson…



I thought I was in for a quiet and blogless Saturday evening until I came across this very slick and glossy pdf  document by Pearson entitled ‘Approaches to spontaneous speakingwhich represents in my eyes all that is wrong about the way teachers in England are trained to teach spontaneous talk. Hence this post, that I really did not want to write – as it is based on the work of someone I respect (Dr Rachel Hawkes) – but I felt compelled to.

Whilst I do agree with some of the points made – especially the fact that speaking is under-emphasized in most MFL classrooms and that student-to-student talk is grossly neglected – and I did find the activities listed useful (if limited in scope and variety)  I was very disappointed with the advice given in this article, especially the absolute absence of a useful framework for classroom practitioners on how to get from imitation to speech spontaneity. A pity, since the rationale given for laying more emphasis on speaking is excellent; the rest, however, simply felt like a dry list of speaking activities (speaking instruction classics) and facts based on very feeble evidence described in an uninspiring OFSTED- or GCSE-specification-like style.

Hence, the impression one gets from reading this is the overly simplistic persuasion that by merely staging such activities lesson in, lesson out the students would develop spontaneity. I wish it was that easy !

I wasn’t surprised, though, as this is the kind of stuff teachers taking part in much Egland-based CPD on speaking get – a list of activities, some reference to what OFTSED would like to see in lessons and very little reference to neuroscience and the research that really matters.

These are the main problems I have with the Pearson article in question.

  1. Spontaneous talk equates with spontaneous grammar

As I often reiterate in my blogs, there cannot be spontaneous talk without spontaneous grammar. Whereas in immersive or naturalistic input-rich environments (e.g international school ; the L2 country) the L2 learner can acquire grammar subconsciously, in input-poor ones with little teacher-student contact time this is impossible without explicit grammar instruction and work on the automatization of the target structures/morphemes.

Hence, before being able to cope with unstructured communicative tasks (describing pictures, unplanned conversation tasks, simulations, etc.) language students need to be first trained to automatize grammar by applying it in speech in highly controlled tasks, what DeKeyser calls ‘Communicative drills’ such as the following (on negatives), that I produced for my year 11 French lesson last Monday :


This controlled, highly structured stage is paramount as, if we just give students open questions and unstructured tasks, as the article suggests, the students might simply come out in their answers with the same words and phrases all the time and teachers would not have any control on their output.

Moreover, unless we structure the kind of output we want the students to produce, they might produce language which does not contain the target structures we want them to practise and re-practise over and over again. Hence, recycling being the key to grammar automization, spontaneity will not be achieved. We all now that. Open tasks will not guarantee such recycling. Highly structured communicative drills, will.

Open questions and unstructured tasks, contrary to what the article suggests  should occur much later on in spontaneous-speech development than the authors suggest -in what I call the  Autonomous phase. This phase, if you have not read my posts on the topic, refers to the stage in which the students do not need scaffolding or any kind of support any longer and can stand on their own two feet. Only at this stage, once much consolidation and practice has occurred through structured and semi-structured tasks, should the students be asked open questions and involved in unstructured communicative tasks.

What it is often forgotten or ignored  is that spontaneity is the equivalent of automatization of grammar and vocabulary use across a very wide range of contexts. Hence, to scaffold spontaneity, one needs to get the students to produce the target language fast and accurately across a wide range of contexts through tasks which involve systematic recycling and repetition of core language patterns (e.g. the same sentence stems applied to different vocabulary such as: je ne vais jamais au restaurant du coin; Je ne vais plus au restaurant chinois du coin; je ne vais jamais au bar du coin; etc.)

For example, going back to the negatives example, in the first phase of speech spontaneity-development (structured production) a teacher would make sure that the students practise the structure through controlled tasks on the topic in hand many times over.

Subsequently, the teacher will ensure that the students practise the structure across past and future topics (Expansion phase). Whilst in the Structured and Production phases support materials could still be used, in the subsequent phase (Autonomous phase), such materials are phased out. In the final stage of automization, the teacher will ensure that the students routinize the target structure through tasks which aim at developing speed (Routinization stage). In this final phase, communicative drills like the above one can come in handy again and can be used alongside more unstructured tasks eliciting the use of the target structure in real time conditions.

Being able to sequence instructional activities effectively and knowing at which proficiency stage to use them and for what purpose is what enhances one’s teaching, not random lists of tasks with a brief explanation of what they consist in.

In a nutshell, it is not clear from the article how on earth language students would ever be able to develop the all-important spontaneous grammar, as there is absolutely no mention of it. This is why a lot of the spontaneous talk one witnesses in English MFL classrooms is so ungrammatical and often contains many fossilised (automized) mistakes.

  1. Total absence of reference to receptive processing

The article does not mention the importance of listening in bringing about spontaneous speech – a shocking omission. Yet, how can one hope to develop spontaneous speech without listening ? The often unintelligible pronunciation and intonation patterns that language students exhibit in their speech is due exactly to this widespread and deeply engrained bad habit of teaching students to speak without adequate modelling through listening. Such modelling is imperative for spontaneous speech to happen. As long as the listening-speaking connection is not made explicit and emphasized by CPD providers and teacher trainers,speaking and language instruction will stay inadequate.

Another dimension of receptive processing vis-à-vis spontaneous talk which is grossly neglected in the article refers to Listenership, i.e. the ability of being able to respond to an interlocutor in real time (someone talking to us) in order to stay in the conversation. This is a very important component of spontaneous speech – unless we talk to ourselves, that is. No student training in this important skill is mentioned, yet, it is fundamental. That is why we have in England entire cohorts of language learners who do not comprehend impromptu questions in the target  language .

In sum, students need to have speaking modelled to them through aural input day in day out and need to become expert ‘spontaneous comprehenders’ as well spontaneous speakers as one cannot produce an effective response without understanding the question or stimulus that would prompt it in the first place.

  1. No mention of pronunciation and decoding instruction

As most current models of speech production clearly posit, there is no way a language learner can produce fluent speech without developing fluent pronunciation and decoding skills. The reason for this being that for spontaneous speech to occur, cognitive control over the articulators (responsible for speech production) must occupy subsidiary awareness (i.e. must occur subconsiously). Without pronunciation practice how can students develop correct and fluent pronunciation? Emphasizing the importance of this dimension of speaking instruction is imperative as , sadly, nobody teaches pronunciation these days…

  1. No explicit framework provided

As mentioned above and discussed in section 1, the article provides no framework whatsoever on how to take the students from novice to expert speech production. Not only it advances the preposterous notion that we should start teaching spontaneous speech by asking open questions – which goes contrary to how children acquire languages, which is exclusively through masses of caregiver modelling through the aural modality (listening) and closed questions, at the initial stages ; but it does not even remotely show how a teacher should structure and map out the evolution of spontaneous speech. Hence, it is of no use to any classroom practitioner who may want to design a curriculum or even an instructional sequence aiming at developing spontaneous speech.

A random list of tasks without any recommendation as to how should be sequenced in the process of automatisation of speech production is of no use whatsoever.

  1. Spontaneous and accurate and intelligible talk ? Or simply spontaneous ?

Another gross omission is the reference to an important aspect of spontaneous speech – intelligibility. There is no use in forging spontaneous speakers if these cannot produce intelligible speech. Again, in reading this article the impression one gets is that all teachers have to concern themselves with is spontaneity. And how about accuracy, comprehensibility and appropriateness vs ungrammaticality, unintelligibility and inappropriateness ? For spontaneous talk to be fluent as well as accurate and comprehensible, there must be a skilful mix of speaking tasks focusing on accuracy and tasks focusing on fluency. No mention of this is made in the article and again one is left with a sense of ‘randomness ‘ and amateur nonchalance in the approach to spontaneous talk put forward in this article.

  1. Contain the conversation with students through implicit recasts

The article invites teachers to keep the conversation going through implicit recasts when students make mistakes (in teacher-student conversation), ignoring to consider the fact that implicit recasts usually go unnoticed and are of no use in terms of modelling and learning. There is plenty of research evidencing that (Macaro 2007).  No alternative means of providing effective feedback strategy which may scaffold spontaneity are offered – yet, spontaneity can and must be coached (e.g. through critical listening)

    7. Where do they get the answers from?

Another classic of English-based CPD: showing a task without telling the teacher where the students are going to learn the answers from and how – convenient, as this is the hardest bit of all. In the article there is a picture task with lots of questions in different tenses listed next to it. Do I need to read a document by Pearson to learn that I can get my students to ask questions about a picture? I was taught that on the second day of my PGCE and the GCSE photo-card task is basically that. What teachers need to know is: how do I get the students to comprehend and answer those questions fast and reasonably accurately? That is what a teacher needs to be told.

Concluding remarks

As it often happens with articles and CPD workshops attempting to enhance teacher competence in the development of spontaneous speech, this document by Pearson merely provides a random list of speaking activities without suggesting any approach rooted in sound cognitive or even mentalist theory as to how teachers can use those activities to take our students from A to Z. A teacher reading this article is left with the misleading impression that by simply staging the tasks listed in lessons day in day out they will magically develop spontaneous talk. A highly disappointing piece, coming as it does from a publishing giant.

How to teach pronunciation




In this article I will concern myself with the teaching of pronunciation and decoding skills (i.e. the transformation of L2 graphemes into sounds) within a typical secondary school setting where teacher contact time is limited (e.g. 2-3 hours per week). The present post should be considered as a sequel to a previous blog (‘Nine research facts on the teaching of L2 phonology teaching and learning’) in which I discussed the theoretical background to what I will propose below. Hence, the reading of that post is recommended if you want to have a better grasp of some of the points made below.

  1. Approaches to pronunciation instruction

I could not locate much research on the teaching of pronunciation as integrated within a typical primary or secondary Modern Foreign Language curriculum. This is possibly due to the fact that the Communicative Language Teaching approach does not lay much emphasis on pronunciation and, consequently, many teachers see as the least useful of language skills (Elliott, 1995). In fact, the typical French, Spanish, German or Italian textbooks currently in use in the UK or US hardly deal with the teaching of the L2 phonology system; when they do concern themselves with pronunciation or decoding instruction, they do so superficially never going beyond the mere awareness-raising of key features and providing very limited practice -if any at all –  in the oral production of the target phonemes. Recycling of the target phonemes is also a very rare feature.

In deciding on how to integrate the teaching of pronunciation in the curriculum language instructors ought to consider which one of the following approaches best suits their learning context:

2.1 Accuracy vs Intelligibility – This is the most important decision to make at the very outset of a language course. Curriculum designers must decide what degree of pronunciation/decoding accuracy they aim at. Is it just for students to be understood by an empathetic native speaker? Is near-native accent the aspirational goal? Or are we aiming at a level of mastery somewhere in between? The answer to these questions will determine the emphasis L2 instruction will lay on pronunciation.

In most secondary schools, as far as I know, curriculum designers do not often ask themselves the above questions; yet, in view of the effects that bad pronunciation can have on effective oral communication, listening and reading comprehension and language acquisition in general (Walter, 2008), they really should.

2.2 Intensive vs Extensive instruction – teachers may decide to teach the pronunciation of a specific set of L2 phonemes intensively over a period of a few weeks or whether to do it extensively (a little bit every day) over a period of several months. Whereas I am inclined to opt for the latter approach based on what we know about L2-phonology acquisition, one may want to implement the former in the run-up to a high-stake examination as part of a remedial program (e.g. on realizing that the pronunciation of specific phonemes may seriously impair the students’ performance).

2.3 Deductive vs Inductive – Deductive teaching involves the traditional approach whereby the teacher presents and explains the phoneme(s) to the students – often in a PPP instructional sequence. Inductive approaches (e.g. guided discovery), on the other hand, are student-led; the students are in charge of describing and/or analyzing the target L2 phoneme(s) and/or discriminating between them and similar L1 sounds. The teacher acts as a facilitator, guiding the students with open questioning as they ‘work out’ how the pronunciation of the target sound(s) works . The latter approach has the merit of eliciting greater cognitive investment on the part of the learner, but it is also more time-consuming. I personally like to alternate both.

2.4 Spontaneous vs Planned – A planned approach implements pronunciation instruction in a systematic way, considering the way the target phonemes could be recycled across the various units of work. In this approach the teaching of the target phonemes is carefully embedded in the schemes of work, which is a more pain-staking process, but one that allows for more effective integration within each topic covered.

Spontaneous instruction, on the other hand, is less time-consuming and is based on a ‘if ain’t broke don’t fix it approach’; in other words, if a teacher notices that their students are making mistakes with a particular L2 phonemes in a given lesson, they will impart instruction on that phoneme ad-hoc, on the spot. This approach can work quite well with more advanced classes which generally display good pronunciation and may occasionally exhibit minor flaws here and there. In this approach, the teacher must have an in-depth knowledge of the target language phonology so as to be able to improvise. Ideally they would also have a readily available bank of resources to drawn upon impromptu.

I personally prefer a planned approach with a very prompt start from the very early days of instruction. But do not refrain from spontaneous instruction when the need arises.

2.5 Contextualised vs Discrete – Another decision to be made is as to whether the teaching will be carried out through (1) stand-alone pronunciation lessons; (2) lessons in which pronunciation instruction is embedded within the teaching of other linguistic content; (3) a mix of both. In view of the curricular and time constraints of typical secondary school courses, I suggest following the third approach using the framework I will outline below, as it allows for systematic recycling but also with some degree of flexibility.

2.6 Awareness-raising vs Automatisation – A lot of the pronunciation teaching I have seen in 25 years of career did not consciously aim at automatization, but rather at awareness-raising. Typically, the teacher shows presentations through which they model the pronunciation of the target phonemes first in isolation and then within words; the students repeat the sounds and words aloud. In the best-case scenario they practise the words through tongue-twisters or other drills, and that’s it!

The problem is that the acquisition of L2 pronunciation requires the automatization of at least three sets of skills: (1) being able to discriminate between the L2 target phoneme and the similar L1 sounds in the receptive phase; (2) being able to reproduce the target phonemes accurately in isolation and (3) in combination with other phonemes at word/phrase level in the context of spontaneous oral communication. With this in mind, it is clear that the approach I have just described will never result into acquisition.

In order to develop the three sets of skills just alluded to, pronunciation/decoding instruction should include the following phases:

  • A modelling phase in which the teacher models the target phoneme and/or how to transform the target graphemes into sound;
  • (when necessary) A description and analysis phase in which the differences between similar (but not identical) L1 and L2 phonemes and/or the L1 and L2 allophones of specific graphemes are clearly explained (or arrived at inductively). For instance, when teaching the differences between the Spanish and English pronunciation of ‘t’ the teacher will show how the Spanish ‘t’ is not a plosive sound and how the tongue hits the tooth in Spanish and the pre-alveolar area in English, etc.
  • A receptive awareness-raising and discrimination phase in which the learners receive practice in: (1) matching the target L2 phonemes with letters/combination of letters (i.e. that the sound /uah/ matches the letter cluster ‘oi’); (2) discriminating between the L2 target phonemes and the similar L1 sounds (for instance: students listen to the word ‘bonjour’ first uttered by a French native speaker and subsequently by a non-native speaker pronouncing ‘J’ the English way. The task: to identify the difference). The micro-listening-skills enhancing tasks I described in detail in two previous posts of mine can be used here. Transcription tasks, short dictations on MWBs and songs with transcript can be used, too.
  • A productive phase involving controlled practice in which the target phonemes (pronunciation) and/or related graphemes (decoding) are practised in the context of drills designed in such a way as to elicit a narrow focus on the target sounds (e.g. short and easy role plays and simple tongue twisters).
  • A productive communicative practice phase. This is crucial in bringing about automatization as it is all very well to know how to pronounce sound /decode letters in isolation; but, ultimately, it is the ability of doing that under Real Operating Conditions in the context of words, phrases and sentences that matters. Semi-structured communicative tasks such as surveys, interviews, role-plays, ‘find someone who’, ‘find out what’ , ‘fill in the information gaps’ or oral picture tasks are invaluable ways to train students in pronunciation and decoding. Before engaging the students in the communicative task, the teacher will focus her attention/feedback and her students’ on the target phonemes.

As I will point out below, for the acquisition of L2 pronunciation to occur, the teaching of pronunciation cannot stop at modelling sounds through a few minutes of choral repetition and some tongue-twister practice. Whatever the target phoneme is, it must be practised extensively in the context of oral production tasks which are at the highly-controlled end of the spectrum first and become incrementally more unstructured.

  1. Factors to consider

In implementing pronunciation instruction as embedded in a typical secondary curriculum, instructors must consider a number of important contextual factors (e.g. contact time; examination board assessment criteria; methodology espoused by the institution you work at) and individual and affective variables (e.g. age of learners; levels of motivation; relevance of pronunciation accuracy to their personal goals; ability).

  1. What we know about the acquisition of L2 pronunciation and decoding skills

4.1 The brain ‘hears’ sounds using the L1 phonological filter; hence, an L2 learner will match the L2 sound they hear to the closest approximation they find in their Long Term Memory. For instance, a French or German native speaker will hear the English word ‘thirsty as ‘sirsty’) whilst an Italian will hear ‘Tirsty’ (this phenomenon was discussed at length in my previous post on pronunciation).

4.2 L2 graphemes automatically activate the L1 phonological system in the L2-leaner’s Long-term Memory (also discussed in my previous post). So, words should be taught using visuals or gestures before they are presented in their written form.

4.3 It is better to start teaching pronunciation when the articulators are more ‘plastic’, before puberty. Some research would suggest starting learning pronunciation before the age of 7 (Lennenberg, 1967).

4.4 Pronunciation errors are difficult to correct when they are fossilized (Ellis, 1996) – L1 transfer in pronunciation is a major threat to the acquisition of accurate L2 pronunciation for the reasons alluded in 4.1 and 4.2 above. Hence, pronunciation instruction should be particularly intensive an extensive in the first two or three years of language learning. Students should receive feedback on their pronunciation from the very early stages of instruction in order to avoid fossilization.

4.5 As already discussed above, L2 phonemes require masses of distributed practice in order to be automatized. Learners’ acquisition of the target phonemes usually follows a U shaped developmental curve with a backsliding phase half-way through the process (Pech et al., 2011)

4.6 Working memory resources are limited both in terms of capacity and duration of storage. For instance, we know that words are phonologically stored in working memory for only a few seconds (Walter, 2008). Hence, unless we focus students on the importance of accurate pronunciation and place it firmly in their focal awareness, most of them will not be able to consciously invest their attentional resources long and deeply enough to notice and learn L2 phonemes.

4.7 Pronunciation and decoding skills must be automatized if we want our students to acquire effective speaking and reading skills. Moreover, better pronunciation and decoding skills result in better acquisition of grammar and syntax. As already discussed in my previous post, more effective pronunciation and decoding skills enhance reading and listening comprehension. Finally, as it is obvious, an L2 speaker who struggles to pronounce L2 words for lack of knowledge of the L2 phonology system and pronunciation/decoding practice is likely to experience serious processing inefficiency issues in L2 oral production.

4.8 Some L2 students are genetically more predisposed than others to notice and acquire foreign language sounds (Nardo et al, 2009).

4.9 A positive orientation towards the target language and the target language culture(s) seem to correlate positively with better pronunciation.Marinova-Todd et al. (2000) posit that motivation may play an even more important role than age in the acquisition of native-like pronunciation.

  1. My tips for pronunciation teaching

The following tips are based on my review of the existing literature on the subject, Skill-theories of language acquisition and most importantly, on my classroom experience. Here they go.

5.1 Start pronunciation instruction very early – in Primary – not only because the learners are more developmentally receptive to it, but also so that you will not have to deal with it any longer later on, when you need to focus on higher order skills. It is also crucial to sensitize the students to the importance of accurate pronunciation at this stage so as to place it firmly into their focal awareness and forge sound learning habits. Students should be given examples of how inaccurate pronunciation can lead to communication breakdown, embarrassing misunderstandings and stigmatization.

5.3 In order to prevent cognitive overload adopt a narrow focus. Select only one phoneme or maximum two per lesson and dwell on it/them for several lessons, whilst recycling the ones you will have taught before in order to keep them in the students’ peripheral awareness at all times. Before any oral communicative activities do ask and frequently remind the students to pay selective attention to the pronunciation of the target phonemes both in production and peer- feedback.

5.4 Integrate pronunciation and decoding instruction in the topic-at-hand using the five-step framework outlined above (Paragraph 2.6). The adoption of a narrow-focus approach means that whilst the first lesson on a given phoneme will be a bit longer, the reinforcement of the same one/two target phonemes over a period of three or four weeks will be much shorter, five to ten minutes every day. The most important thing will be, as suggested in the previous point, to keep the students’ focus on the target phoneme(s) during any oral communicative activities staged in class during the entire reinforcement/recycling period.

5.5 Plan integration opportunistically but judiciously. By this I mean that whilst planning a unit of work, you may find that the vocabulary or grammar structure you intend to teach lends itself beautifully to the teaching of specific sounds. For instance, in teaching animals in French, a few weeks ago, the words ‘chat’, ‘chien’, ‘cochon d’Inde’ and ‘cheval’ prompted me to decide to teach that ‘ch’ is pronounced ‘sh’ in French. However, opportunism should not be the only criterion to be used in selecting the target phonemes: intelligibility impact (the extent to which the sound may affect understanding); learnability (how ready the students are to pick it up) and frequency (how often they are likely to come across that sound) should also be taken into consideration.

5.6. Prefer extensive distributed practice (a little every day over a longish period of time) over massed practice. Extensive practice which recycles the target sounds for a few minutes every day is crucial to the success of pronunciation instruction as L2 phoneme acquisition requires a lot of time and contextualized practice. I keep a chart, on a google doc, in which I tally all the phonemes I teach so as to have an overview of how many times I have recycled each sound.

5.7 Model through detailed description and analysis. In the modelling phase take the learners, possibly using audio-visuals, through every single step involved in the production of the target L2 phoneme. There are plenty of charts available online showing how the different sounds are produced in various languages. Remember: effective modelling and analysis require good knowledge of the target language phonology system. Heads of Depatment ought to provide staff with adequate specialised training.

5.8 Stage critical listening activities with narrow focus. These can be modelled in class and then flipped. They consist of getting students to listen to a peer reading aloud (if the focus is on decoding) or talking in the L2  whilst one or more peers focus their attention and provide feedback on two or three phonemes. Alternatively, you can get the students to record themselves and each other. Critical listening fosters collaborative learning, social strategies and metacognition whilst bringing about deep cognitive investment.

5.9 Alternate inductive and deductive approaches for the sake of variety, but also to foster healthy inquiry skills.

5.10. Prevent automatic activation of L1 sounds as much as possible. Keep pictures on the walls which refer to words containing the key L2 phonemes and most students know very well (e.g. numbers); it is important that they did not learn to pronounce these words through the written medium in the first place. These pictures will be a valuable aid each time you will want to correct phonetic mistakes made by your students without providing them with a written example (as this would automatically activate the student’s L1 decoding system possibly causing L1 transfer issues). So for instance, when reminding a student of the decoding of the consonant cluster ‘oi’ you will point to the number ‘3’ on the wall, ‘trois’, which contains those letters.

5.11 Teach and practise the target phonemes in accessible linguistic contexts. When teaching new phonemes it is vital that the students have enough cognitive space available to focus on them. If the ‘drills’ or communicative tasks are complex, contain masses of new vocabulary, challenging L2 structures or even other difficult sounds they have not yet mastered, this will impact very negatively on learning. Some teachers give fun but very phonetically complex tongue twisters to novice L2 learners who do enjoy the challenge and often have a  real blast in the process, but frequently end up making a complete mess of it. When selecting tongue twisters or any other material for the initial modelling and controlled practice phases, choose texts which pose very little cognitive and phonetic challenge.

5.12 Correct judiciously. Feedback is very important in the teaching of pronunciation and may need to be more frequent than the correction teachers provide on other aspects of performance. This is because pronunciation mistakes tend to fossilize more easily. Hence, teachers need to monitor oral communicative activities very closely and step in – even interrupting – when errors made with the target phoneme(s) (1) are made publically and consequently may affect several students’ perception of what is correct/incorrect; (2) impede intelligibility; (3) are made frequently. However, do not overdo correction as it may affect motivation.

Concluding remarks

In this post I have attempted to provide some tips on the pronunciation of L2 pronunciation and decoding skills. The most important point is that teachers should sensitize their students to the importance of accurate L2 pronunciation from the very early days of instruction. A principled approach to the planning and classroom delivery of L2 pronunciation instruction should be devised which provides extensive distributed practice through a mix of inductive and deductive learning and adopts a narrow focus, i.e. one or two phonemes are taught each time. Last but not least, this approach, which integrates pronunciation instruction within most lessons (a few minutes per session) cannot lead to acquisition of the target phonemes unless these are practised in the context of structured and unstructured communicative activities.

In this post I have also recommended the following framework that I have used successfully over the years and integrates the teaching of pronunciation with communicative language teaching and serve the goals of the curriculum I am charged with delivering. This framework includes five phases

(1) A modelling phase

(2) A description and analysis phase in which the differences between similar (but not identical) L1 and L2 phonemes and/or the L1 and L2 allophones of specific graphemes are clearly explained;

(3) A receptive awareness-raising and discrimination phase;

(4) A highly structured productive  phase;

(5) A productive semi-structured to unstructured communicative practice phase.

It should be reiterated that in the fifth phase, crucial in bringing about automatization, the role of the teacher in monitoring and providing feedback on the phonological level of student output is crucial. Also pivotal is the level of student focus on pronunciation that the teacher will have generated; unless the students are encouraged and motivated to keep the importance of pronunciation in their focal awareness no pronunciation instruction will ever succeed. Unfortunately, these days, pronunciation is but a peripheral concern in most Modern Language classrooms.

You can find more on this in the book ‘The language teacher toolkit’ that I have co-authored with Steve Smith, available here

Five pronunciation and decoding issues in French-as -a-foreign-language instruction that seriously affect grammar learning and should be targeted as early on as possible

Please note: this post was co-authored with Steve Smith of


As I explained in several previous blogs, our students’ ability to decode the target language sounds can seriously impact acquisition. And I am not simply talking of their ability to acquire vocabulary and pronunciation. I am also alluding to the learners potential to notice and internalize grammar. Why? Because receptive decoding, i.e. the way the human brain ‘deciphers’ the sounds we hear, can cue us to certain grammatical features of words (i.e. endings) we process aurally that lead us to noticing and making assumptions about their gender and/or number (for nouns and adjectives), person, conjugation and tenses (for verbs) and other ‘anomalies’ (e.g. ‘l’’ before nouns, the pronoun ‘y’ before a verb).

For instance, today, a student I help outside school, had still not grasped the phonetic difference between ‘ne’ (as it sounds in ‘ne’) and ‘n’ai’ (as in ‘je n’ai’). This has led her for many years to presume that ‘je n’ai pas vu’ was actually spelt ‘je ne pas vu’. This in turn affected her assumptions as to how the Perfect Tense is formed in French which resulted in the mental representation ‘Je + past participle’ (i.e. Je vu). I could not blame her as, evidently, her previous tutor must have not emphasized adequately the difference between ‘Je’ and ‘J’ai’.

In this post I will focus on five pronunciation/decoding issues in FLE (Français Langue Étrangère) instruction which do usually receive some emphasis but are not in my experience duly emphasized and practised in the typical L2-French classroom.

Issue 1: [ə] vs [e] in receptive decoding

This sound is one of the most important to learn in terms of receptive decoding, not only for the ‘Je’ versus ‘J’ai’ distinction alluded to above, but also because of the potential it has for cueing the students to the presence of a plural noun. Take for instance the sentences ‘ le fils de Marie étudie l’anglais’ et ‘les fils de Marie étudient l’anglais’. In this context the inaccurate perception of the sound [ə] as [e] (as in les) may easily cause confusion (is the subject ‘fils’ plural or singular?) – confusion that might exacerbated by the fact that the ‘s’ in ‘fils’ may lead the beginner French learner to believe that the noun is plural.

Another huge issue that novice teachers often overlook is the pronunciation of ‘é’ as [e] or [ə] in active decoding (when reading a word). This is particularly a problem when it comes to the Perfect tense. When a student pronounces ‘j’ai mangé’ same as ‘je mange’ unless there is a time marker (e.g. hier) the potential for confusion and communication breakdown is great.

The difference between the two sounds must be duly emphasized from the very early stages of instruction. Most often teachers do when it comes to the ‘je’ vs ‘j’ai’ dichotomy. I have often witnessed the very good practice of modelling the two sounds by over-emphasizing lip movement in an attempt to create a muscle memory of the articulatory process and through minimal-pair work. What I have not often seen, however, is the teachers recycling those through many other familiar and unfamiliar contexts until the acquisition of those sounds has occurred. To presume that one or two minimum-pair demonstrations will lead to the automatization of the phonemes is over-optimistic; for most students, acquiring the ability to discern between the two sounds in receptive decoding will take several weeks or even months.

As recommended in previous posts, teachers may want to engage students in micro-listening-skill enhancers reinforcing the distinction between the two sounds for a few minutes (10?) per lessons over a period of 4 to 8 weeks in order to obtain very good results.  For the rationale for this approach read my previous post ‘How to teach pronunciation’.

Issue 2: pronunciation of ‘t’ /’d’/ ‘p’ / ‘s’ at the end of words vs same letters + ‘e’ or ‘es’

Learning the correct (active) decoding of  the above consonants with and without ‘e’ or ‘es’ at the end of words from the very early stages of French instruction is of paramount importance. Why? The reason is two-fold. Firstly, if the students are not aware of the distinction ‘t’ vs ‘te’ / ‘tes’, they will not be able to distinguish between masculine and feminine nouns and adjectives – a distinction that many L1 English learners of French find very challenging. Secondly, the pronunciation of ‘t’ / ‘p’ etc. at the end of many words or nouns can cause serious misunderstandings in terms of meaning. Think about the nouns ‘point’ and ‘pointe’; ‘vent’ and ‘vente’; ‘coup’ et ‘coupe’.

Issue 3 – [z] vs [s] in liaison

This is another issue that it is not always tackled effectively. This can cause quite a few problems, the most frequent and important of which pertains to the famous cross-association elles/ils ont vs elles/ils sont. Since this distinction is not often ‘drilled in’ adequately, many learners of French end up confusing the two all the way to year 11 (when they are 16 yrs old). Often, this leads to ‘elles ont treize ans’ being interpreted as ’elles sont treize ans’ and the assumption that the French, too, like the English, use ‘to be’ when telling somebody’s age.

Confusing [z] and [s] is very common and can cause misunderstanding in many contexts, for example: ‘elles avaient beaucoup de choses’ was interpreted by one of my student as ‘elle savait beaucoup de choses’.

Issue 4 – ‘un’ vs ‘une’

The effects of decoding issues with these words on the acquisition of the French grammar are possibly the most obvious. If a learner is not clear about the differences in pronunciation between these two words, they will make incorrect inference as which nouns are masculine and which are feminine. This issue will affect, if unresolved, other more complex structures such as ‘aucun’ vs ‘aucune’.

After modelling the lip-movements through over emphasis of the sound articulation and contrasting English words such as untouchable and unsolvable (in which the two first letters ‘un’ are highlighted in red) with words like Unesco, tune, (where the key letters ‘une’ are highlighted) one may want to reinforce the differences through micro-listening -skill enhancers such as ‘Broken words’ whereby the teacher utters words such lune, lundi, rune, aucun, aucune etc. and the students have to complete the gaps in l__di, r____, auc___, auc____, etc.

Issue 5 –  voicing ‘ent’ at the end of verbs

This is a very common issue that in my experience receives some attention but not as much as it actually requires and deserves. When not acquired effectively, the wrong decoding of ‘ent’ in the third person plural of  the Present Indicative and Subjunctive of verbs can cause quite a lot of problems. In a class I observed recently, for instance, the belief that ‘ent’ is voiced, led to the students often confusing (in a listening activity centred on modal verbs)  ‘veulent’ and ‘veut’, ‘peuvent’ and ‘peut’, and ‘doit’ and ‘doivent’. In the past I have also witnessed issues in the pronunciation of the third person of the imperfect indicative and of the conditional. Another related issue pertains to adverbs (e.g facilement, lentement) which are occasionally mistaken for verbs.

Concluding remarks

Whereas most published course-books usually converge in the way they sequence the teaching of grammar (especially tenses), when it comes to pronunciation they use quite a random approach which does not appear to be principled in any way. I suggest that, pronunciation/decoding skills instruction should first deal with the easiest-to-acquire phonemes and then gradually concern itself with the more challenging ones, another criterion ought to be considered, too; the extent to which, that is, the ineffective mastery of those sounds can affect the acquisition of the grammar of the target language. The five sets of pronunciation/decoding issues discussed above are only but a few examples of the way in which phonological awareness and the ability to transform graphemes into sound can affect the acquisition of the target language grammar. Teachers ought to pay attention to this very important facet of language acquisition and devote sufficient time and effort to it using the research-based framework I outlined in previous posts.

You will find more on this issue in the book ‘The Language Teacher Toolkit’ that I co-authored with Steve Smith and is available for purchase at


Nine research facts about L2 phonology teaching and learning that every teacher should know


1. Introduction

In the last three weeks I have been researching L2 phonology acquisition as the teaching of pronunciation and decoding skills is one of my performance management targets for this year. This post, written in collaboration with Steve Smith of and Dylan Viñales of Garden International School, is a ‘prequel’ to a longer and more exhaustive article I will publish in a few days in which I will lay out the approach to phonology instruction I undertake in my lessons. Here I will concern myself with nine research facts about the acquisition and teaching of L2 pronunciation and decoding skills that every modern language teacher should know and that should constitute the starting point for any teaching approach to L2-phonology instruction. Here they are:

2. Pronunciation and decoding are the most neglected skills in Modern Language classrooms

As Elliott (1995) points out, Foreign Language Instruction does not concern itself with pronunciation and decoding skills as much as it does with listening, speaking, reading, writing and grammar. Whereas there is some focus on the L2 sound system during the first stages of instruction, especially when the L2 alphabet is introduced, teachers rarely continue to duly emphasize pronunciation for the rest of the course. It has to be added that often teachers think they are teaching pronunciation whilst they are actually mostly focusing on teaching decoding skills. Not the same thing.

Decoding-skill instruction is about teaching students how to convert the written form of the L2 into sound, that is to say how letters, when combined together, should sound in the target language; Pronunciation, on the other hand, is about learning how to accurately master the L2 Phonological system in any oral production. Whereas teachers do occasionally provide some instruction and practice in decoding skills, they rarely give their students information about subtleties in L2 pronunciation, e.g. the differences between the plosive English /t:/ and the non-plosive Spanish /t/ allophones of the letter ‘t’ (i.e. two different phonemes associated with the same letter). When they do, it is usually on a sporadic, ad-hoc and a-systematic basis; recycling of that information is rare Elliott (1995).

As Forman (1993) pointed out, one of the reasons for this neglect is that teachers do not receive sufficient training in pronunciation teaching. We may add that most modern language teachers in the UK were not formally taught the L2 phonology system in their undergraduates years and, although they often do have near-native or even native pronunciation they do not have explicit knowledge of how different sounds are produced and of the relevant metalanguage (e.g. what kind of sounds the labels ‘affricate’, ‘fricative’, ‘occlusive’, ‘plosive’ refer to).

Implications for teaching – (1) If teachers do not emphasize pronunciation from the very early stages of instruction and sustain this emphasis throughout the course, students will not see it as important and consequently may not develop intentionality (the desire to learn) vis-à-vis this aspect of their L2 learning. This is important in view of findings by Suter (1976), Elliott (1995) and other studies that found that if students are more concerned about their pronunciation they tend to have better pronunciation of the target allophones. In particular, Elliott (1995) found that university students’ attitude towards acquiring near-native or near-native pronunciation was the principal variable in relation to target language pronunciation. (2) Pre-service UK teachers should be trained in the teaching of pronunciation and decoding skills and should be provided with a good understanding of the differences between their L1 and L2/L3’s sound systems.

3. L2 learner levels of Integrative motivation can affect the acquisition of pronunciation

As a university lecturer at Reading University – many years ago – I was always surprised by the huge differences in terms of pronunciation between L2-Italian students who came back to England after spending a whole year abroad, in Italy. Some had near-native pronunciation whilst others still retained a strong British accent. After investigating this phenomenon, I found that those who had the best pronunciation had fully embraced the Italian culture and tried hard to integrate both psychologically and socially with the natives – they displayed, in other words, what Gardner and Lambert (1959) call integrative motivation. The less Italian-sounding students, on the other hand, had made much less effort; yes they had enjoyed Italy and liked the language but had been less open and proactive in terms of integration. My ‘findings’ echo those of many theorists and research studies (e.g. Schumann’s,1986, Sparks and Glachow,1991) who posit that a positive orientation towards the target language/culture is an important factor in developing native-like pronunciation

Implications for teaching – getting the students to develop a positive orientation towards the target language and culture is paramount. It is obvious that a student with a dislike for the French civilization will not want to sound French. This is a further reason to aim at heightening cultural empathy for and appreciation of the target language culture(s) in the L2 classroom. Moreover, ways must be found to get students to practise the language orally with target language native speakers outside the classroom. Considering that social media is our teenage students’ most common past time these days this should not be an impossible task.

4. Age as a catalyst or inhibitor of acquisition of pronunciation

It has been posited by some researchers that there is a critical age beyond which it is impossible to acquire native-like pronunciation. A study of Korean children aged between 5 and 10 adopted by French families (Pallier et al., 2003) indicates that at least until the age of ten humans can still acquire 100 % native pronunciation. However, studies by Bialystock (1997), Bongaert et al (1997) and others have demonstrated that this can be achieved with adult L2 learners, too.

It should be pointed out that the commonly held assumption that simply learning a language as a child will lead to the acquisition of perfect L2 pronunciation is only true of naturalistic acquisition, i.e. of acquisition in a second language context in which the child has masses of exposure to the target language (e.g. a child of immigrants/expatriates acquiring the host country’s language or a non-English native speaker in an English medium international school). However, a five-year-old attending two L2-Mandarin lessons a week will not necessarily develop native-like Mandarin pronunciation just because of their age – in fact, in my experience more than often they do not. Frequency of exposure and other factors (e.g. motivation and aptitude) will play an important role, too.

Implications for teaching – language teachers should not be put off by the misconception that beyond puberty L2 learners cannot acquire native-like pronunciation.

5. L2 sounds are interpreted by the brain using the L1 phonological system

For several decades, language instructors were told by theorists working in the Nativist paradigm (e.g. Stephen Krashen) that there was no need to explicitly teach the target language phonology as students would acquire it naturally by simply being exposed to aural input, very much as children learn the mother tongue by listening to caregivers. The proponents of Communicative Language Teaching have a slightly more positive attitude to the teaching of pronunciation. However, because CLT’s main aim is intelligibility of student oral output, not accurate L2 pronunciation, CLT instruction does not greatly emphasize the teaching of pronunciation either. This is why on your teacher training course you were probably not taught how to teach pronunciation and decoding skills.

The assumption that L2 students will ‘pick up’ accurate L2 pronunciation through frequent aural exposure to the language – like children do in the first language – is very intuitively appealing: if you listen to L2 speakers over and over again, day-in day-out you will eventually get a perfect or at least very good pronunciation.  Yet – as much research has shown – this assumption is flawed. Why?

The reason is that the average human brain, unlike what happens in first language acquisition, automatically uses the existing L1 phonological system to interpret any L2 input it hears. In other words, we match any foreign language sound we hear to the most similar L1 sound stored in our Long-term memory. So, for example, an Italian student of English will automatically hear [t] or [f] instead of [θ] whenever s/he hears the first two letters in the word ‘thirsty’; a French native speaker, on the other hand, will hear [s].

What is interesting is that this perceptual mismatch influenced by the native language occurs even though the sound the student hears does actually exist in their mother tongue but is marked by another more frequent similar sound. For instance, the way the ‘n’ in ‘canyon’ is pronounced in English is marked by the more frequent way ‘n’ is pronounced in the same language (e.g. in the word ‘name’). This means that when an English native is taught to pronounce the Italian ‘gn’ sound – very similar to the way the ‘n’ sounds in ‘canyon’ – they will inevitably pronounce it as ‘n’.

Implications for teaching – This automatic response of the brain to foreign language aural input has huge implications:

(1) if we do not raise students’ awareness of the perceptual mismatch which occurs in the Working Memory from the very early days of instruction, they might – as it often happens – end up automatizing a highly L1-influenced L2-pronunciation.

(2) Whether by using a deductive or an inductive approach, it is paramount to raise L2 learner awareness of the differences between the L1 and L2 phonemes the students perceive as identical. I found visual aids very useful in this regard. In teaching the difference between the way ‘T’s are pronounced in French and in English for instance showing through a diagram where the tongue hits the tooth has helped many cohorts of my students to greatly improve their pronunciation of those sounds.

(3) If teachers do know the sound system of their students’ native language, they will be able to anticipate the barriers to accurate L2 pronunciation that L1 transfer erects and plan their teaching accordingly. A perfectly bilingual teacher with native/near-native pronunciation in both their students’ L1 and the target language will have a greater advantage, in this respect, over a teacher with monolingual mastery.

(4) Frequency of exposure and practice in L2 pronunciation and decoding is pivotal. Better a few minutes every day than one or two pronunciation lessons a month.

6. Accurate acquisition of L2 phonology leads not only to more effective listening skills but also to better vocabulary and grammar acquisition

A number of studies have systematically evidenced the fact that L2 learners who have successfully acquired L2 phonology usually have a better mastery of L2 vocabulary, grammar and syntax. Why? Because they are more likely to pick up from aural comprehensible input grammatical features that L2 learners with a less developed grasp of the L2 phonology may not be able to notice. An example: imagine an L2 learner of French who knows that ‘ts’ at the end of French words is usually silent wherear ‘-tes’ is pronounced ‘t’. On hearing the sentence ‘les grenouilles sont marrantes’ (frogs are funny) this learner will understand from the pronunciation that frog is a feminine noun. On the other hand, a student who pronounces ‘marrants (masculine plural) and ‘marrantes’ (feminine plural) the same way, as many L1 English learners of L2 French do, will never pick up on that.

Implications for teaching – In teaching pronunciation and decoding skills teachers ought to prioritize those sounds that may enhance or hinder students’ noticing and understanding of key grammar features (e.g. ‘e’ versus ‘é’ or verb/noun/adjectival endings in French).

7.L2 graphemes automatically activate L1 phonemes in the L2-learner working memory

Whenever a beginner L2 learner who has not as yet mastered the L2 phonological system is asked to pronounce a new L2 word, they will tend to automatically decode it (i.e. transform the letters into sounds) using their native language decoding system – unless other mechanisms (e.g. overcompensation) set in. For instance, an Italian beginner learner of English is very likely to wrongly pronounce the consonant cluster ‘gn’ as an English native speaker would pronounce ‘n’ in the word ‘canyon’.

What is interesting is that many L1 learners, when reading silently, report often repeating the words in their phonemic form ‘in their heads’ (sub-vocalizing), especially when they struggle with the meaning of a text. This entails the risk of L1 learners learning the wrong pronunciation even as they read silently.

Implications for teaching

The above has important implications for teaching. Firstly, this is another important argument in favour of teaching pronunciation and decoding skills explicitly from day one. Secondly, exposure to L2 words in their written form ought to be avoided as much as possible with beginner learners. When new lexical items are indeed presented, they should first be presented through visual aids or gestures; their written form should be provided only after much exposure / practice with their phonemic form.

Another important implication of this phenomenon refers to the frequent use of word-lists and writing mats by many modern language teachers. Unless the students have mastered the L2 decoding system this practice is likely to be very detrimental to their learning as the chance of them mispronouncing the words on those lists/mats will be high. This is particularly the case when the target words have not been selected according to easy-decodability criteria – as it is often the case in textbooks. Hence, teachers should endeavour to use wordlists – with beginners – that are pitched to the right level in terms of ease of pronunciation. When selecting or creating word lists for use, they should model extensively the pronunciation of the words the mats contain through lots of aural activities aimed at raising learner awareness of the pronunciation of the more difficult items.

8. U-shaped developmental curve of phonology acquisition

As it is obvious, frequency of exposure is more likely to result in better acquisition. What several studies have shown, however, is that a U-shaped developmental curve can be observed when students are being taught pronunciation across a range of L2 phonemes. During the first four weeks of instruction there is usually a marked improvement. In the three or four weeks thereafter the L2 learner seems to make more pronunciation errors due to cognitive overload; after this phase, which lasts three or four weeks, accuracy in production appears to be on the rise again.

Implications for teaching – when imparting pronunciation instruction, teachers must be mindful of the transitional phase observed by research. It is a necessary step the human brain takes in which through trial and error the learner refines their grasp of the target language phonology. Hence, teachers should not feel discouraged and give up on pronunciation instruction; instead, they should double their cognitive and affective support to the students and provide masses of constructive feedback through critical listening (whether by themselves or peers) and remedial strategies (e.g. metacognitive listening) which help restructuring.

9. Effective decoding skills and pronunciation play an important role in L2 reading comprehension

A substantial body of research evidence (e.g. Walter, 2008) has demonstrated that poor L2 readers do not often comprehend L2 texts not simply due to lack of grammar or vocabulary knowledge but because of poor decoding skills and issues with the phonological representation of what they read in their Working Memory (in the Articulatory/Phonological Loop to be more precise). The reasons for this are too complicated and beyond the scope of this post. It will suffice to say that they refer to the obstacles to Working Memory processing efficiency that bad decoding skills pose and which, in turn, hinder comprehension. If you do want to know more about this, read Catherine Walter’s fascinating article ‘ Phonology in second language reading: not an optional extra’ (at ).

Implications for teaching – When confronted with poor L2-reader students teachers often provide them with extra reading practice or focus on widening their vocabulary repertoire. This may not be sufficient; they may want to also focus on enhancing their decoding skills and, in particular on their ability to discriminate between L2 sounds that they confuse – due most often to L1 transfer. For instance, Walter (2008) reports findings from Flege and Mckay (2004) that many Italian immigrants who had been residing in Canada for many years still had problems discriminating aurally between the English sounds /ɑ / and /ʌ/ and between /æ/ and /e / (a very common problems amongst L1 Italian learners of English). This hindered their reading comprehension.

10. The temporariness of phonological storage in Working Memory

A fairly recent acquisition of neuroscience is that Working Memory’s phonological storage does not last more than a few seconds (some say two!) unless, that is, we make a conscious effort through rehearsal (repetition) to keep items in there. This limited storage time has important implications given that memory is phonologically mediated (i.e. when we retrieve L2 words from Long-Term memory we do so through their sound); it means that when we hear the words ‘cats’ and ‘cuts’ and we are not clear as to the difference between /ɑ / and /ʌ/ we do not have much time to decide which one we are actually hearing, unless we have automatized the ability to discriminate between those two sounds. Imagine this kind of scenario happening to one of our students during a high stake listening examination… it would cause confusion, slow down the whole process and, should the ambiguous word be crucial to the understanding of the text, it may seriously undermine their performance.

Implications for teaching: Same as per point 9.

11. There is no link between musical ability and pronunciation ability – Researchers have often attempted to evidence a link between the two and have systematically failed to find one. In fact, they have identified a lot of people who have one of these ‘talents’ but not the other.

Implications for teaching – Do not presume that the musical prodigies in your language classes are being failed by you if they do not exhibit excellent pronunciation. In the past, teachers I have worked with were so baffled by the fact that musically-talented students in their classes were not pronouncing words correctly because of this commonly held misconception. No need to be baffled – the two skills are not necessarily related.

12. Concluding remarks

Pronunciation and decoding skills are the most neglected aspect of L2 instruction in secondary school settings nowadays. This is because the trending language teaching methodologies either posit that L2 phonology is acquired subconsciously exactly as happens in L1 acquisition or concern themselves with intelligible communication – hence, accent does not matter as far as a sympathetic native speaker would understand what the student is trying to convey. However, the level of mastery of L2 phonology can seriously impact the acquisition of L2 grammar, syntax and vocabulary and can affect L2 reading and listening comprehension. In particular, the inability to clearly discriminate between similar-sounding L2 phonemes can slow down the processing of aural and written L2 input with potentially disastrous consequences for L2 learning and performance. Hence, it is imperative that Modern Language curricula lay more emphasis on the teaching of pronunciation and decoding skills. At Garden International School, Dylan Vinales and myself are currently experimenting with pronunciation and decoding-skill instruction through various approaches and techniques which I will describe in the sequel to this post (to be published over the next few days).

You can find more on this topic in my book ‘The language teacher toolkit’ , co-authored with Steve Smith and available for purchase on

Another oral-skills-enhancing instructional sequence for beginner to intermediate learners en route to spontaneous talk

image (1)

Please note: this post was written in collaboration with Steve Smith of 


A point that I have often made in my posts is that for foreign language teaching CPD to be effective it has to go beyond simply describing a recommended learning activity, app or website. It also has to provide instructors with a solid rationale for its adoption and how it can be deployed effectively within a teaching sequence. Unfortunately, in my experience, this rarely happens – especially on teacher training courses.  This is the third in a series of posts which Steve Smith and I have written in order to address this perceived deficit in the area of oral proficiency development.

This post proposes a low-effort/high-impact teaching sequence centred on the use of a very versatile learning activity, ‘Find someone who’ (with cards) which, whilst having the development of oral proficiency as its main focus, does also provide practice in listening, reading and writing skills.

Whilst ‘Find some who’ is a fairly straightforward activity to conduct, how to prepare the students effectively for it and to exploit its full learning potential is much less evident. In what follows I suggest ways in which this can be done without too much effort on the part of the teacher.

  1. The task

Each student is given a different card with a number of details in the L1 or in the L2. In my version of this activity the cards usually have five to eight bullet points which look something like this:

  • Name: Jean
  • Date of birthday: 3rd May
  • Siblings: one younger sister
  • Favourite hobby: reading novels and painting
  • Pet hates: cricket and Facebook
  • Favourite singer: Taylor swift

The students are also given a grid with a number of questions in the L1 or L2 (see image above). I personally prefer to put the questions into the L1 so as to avoid spoon-feeding the students. The questions read something like this:

“Find someone who…

  1. …hates Facebook”
  2. …has two siblings”
  3. …is born in September”, etc.

The students are required to find a person for each of the above prompts (e.g. ‘Jean’ for question 1, above) by asking questions to the other students (in the target language). The student who finds them all first, wins. I usually prepare two or three different sets of questions in order to play more rounds.

I uploaded many (free) samples of ‘Find someone who’ on (e.g.

3.Planning / Preparation

1. Decide on the grammar, vocabulary and other linguistic features you intend to focus on;

2.Prepare a set of cards with four of five bullet points;

3. Prepare one or more sets of questions making sure that each question refers only to one card so as to have more movement around the classroom;

4. Prepare a few very short texts in the target language for reading and listening comprehension purposes which you will use in the run-up to the activity implementation. The texts should contain the same sort of details the students will find on the cards. Example:

‘My name is Sean. I am 13 years old and my birthday is on June 20. I have two sisters. My favourite hobby is reading and playing the violin. I hate social networks such as Facebook. My favourite singer is Sia.’

5. (Optional) prepare vocabulary games, worksheets, quizzes recycling the language to be deployed during the to-be-staged activity to give as homework before the lesson

4. The sequence

  1. Drill in vocabulary (15 minutes) – as suggested above, one can ‘flip’  most of this. However, it is beneficial to do some recycling at the beginning of the lesson anyway in order to activate the target vocabulary in Long Term Memory.
  1. Reading and listening comprehension (based on cards) – 2a. Put the short texts containing the target linguistic features up on the screen. Ask reading comprehension questions on the texts of the sort you expect the students to ask each other later on as part of the ‘Find someone who’. Equipped with MWBs the students answer the questions (all in the target language, of course). 2b.Now read out the texts you will have prepared for listening comprehension purposes. Students still answer comprehension questions on MWBs. Since the purpose of this listening activity is not only to recycle the target linguistic features and assess comprehension but also, and more importantly, to model pronunciation, be mindful of the speed at which you utter each text and repeat as often as the students’ request you to.
  1. Questions and answers – Now it is time to further practise the questions that you expect them to produce during the ‘Find someone who’. The easiest option – the one requiring the least preparation – is to ask the students to carry out a survey using the target questions (partner A asks and partner B notes down answers). This should be conducted entirely in the target language. Teachers will go around facilitating and monitoring.
  2. Find someone who – Now carry out one or more rounds of ‘Find someone who’. Make sure that nobody ‘cheats’ by copying what they see on their peers’ grid – the most common offence.
  3. Fluent writing – Now students work in groups. Students take turns in reading out – in the L1 – the details on whichever card they hold and the rest of the group has a set amount of time to put them into French in the form of a paragraph, writing on MWBs – note: this must not necessarily be a word-for-word translation. The purpose of this activity is to prep the students for the next task.
  4. Fluent speaking – Now students go away in pairs with iPads or other recording devices. Each student is given three cards they have not worked with before. The task is to describe the details on the three cards in the target language talking in the third person whilst being recorded impromptu – without studying the cards prior to the recording (e.g. His name is Jean, he is 13 years old, he hates Facebook, etc.).

5. Conclusions

The instructional sequence just outlined is easy to prepare and manage; it allows for practice across all four skills and continuous recycling of the target linguistic features.’Find someone who’ can be implemented without creating cards with fictitious identities and details; however, this allows for less control over the language you want to drill in. I have been using the above sequence several times in my practice and the students usually enjoy and learn a lot from it.

Using picture tasks to develop spontaneous talk – A low effort / high impact teaching sequence


Please note: This post was written in collaboration with Steve Smith of with whom I am co-authoring the ‘MFL teacher’s toolkit’ (to be published in the new year).

  1. Introduction

One of the approaches I undertake in order to promote L2 oral fluency and spontaneity involves the use of picture tasks. This post lays out a low-effort/high-impact teaching sequence based on the following pillars of my instructional approach:

  • Prep students before you start the teaching sequence through as much (flipped?) vocabulary building as possible.
  • Allow for lots of recycling of the target material throughout the sequence.
  • Provide lots of comprehensible written and aural input tbefore involving the students in written or oral production.
  • Start production with highly structured activities which become increasingly less structured. Withdraw support at the end of the sequence.
  • Aim at automatization of speech production as the end-goal (e.g. fast retrieval from long term memory). Prioritize fluency over accuracy in the process; hence tolerate errors unless they impede intelligibility.

The use of picture tasks is advantageous for the following reasons:

  • They require little preparation.
  • Elicit greater creativity with the language.
  • In life we often describe what we see- hence it is a real-life task.
  • The same picture can be used across various tenses.

Please note that the following sequence usually takes more than one lesson and that I supplement it with quizzes and games aimed at reinforcing the vocabulary as well as any grammar needed to execute the tasks in hand (e.g. verb conjugations).

  1. Preparation

2.1 Select the images – Select pictures ensuring that they are not all going to elicit exactly the same kind of vocabulary from your typical student. Some degree of repetition is desirable, though, for the sake of recycling. Ideally the images that you selected would allow the students to answer a range of questions (e.g. When? How? What? Who?). For the sequence that I envisage in this post you will need two sets of pictures which are similar but not entirely identical; so, if Set 1 incudes picture 1a depicting a Ferrari in a city street, Set 2, will contain picture 2a portraying another means of transport in a similar setting with some variation (e.g. different weather, different looking people, different time of the day). The rationale for this will be evident below.


2.2 Decide on the language focus – In planning the activities try to figure out the sort of verbs/nouns the pictures you chose are likely to elicit. If you intend to focus on one or more specific tenses, do provide practice in time markers (e.g. for the present: usually, every day, always, never).


  1. Activities


3.1. Brainstorming  – Give students the pictures (Set 1 only) and ask them to brainstorm as many verbs per picture as they can in groups of two or three. Ideally, before this activity, some vocabulary building activities drilling in as many verbs as possible should be carried out for 10-15 minutes or, even better, ‘flipped’ in the run-up to the actual lesson. I have uploaded worksheets with such activities on and I have created a self-marking module in the grammar section of (see: Verbs monster work-outs).


3.2. Modelling via written and aural input – show on screen sentences (one at the time) in the target language (based on the Set 1 pictures) and ask the students to write on mini-boards (under time constraints) which picture(s) they think they could refer to. I usually do this as a listening activity, too, so as to model pronunciation, as a follow-up.


3.3. Scaffolded written production– Ask students to create one or more sentences for each picture working alone or in pairs. At this stage you can give the students a list of vocabulary as support. I often do this activity on Padlet or Edmodo for students to be able to share their output with others. This activity is carried out without any time constraints, which allows for more careful self-monitoring during production.


3.4. Scaffolded oral production – Ask students to work in pairs. Partner A chooses a picture and ask three questions (one at the time, obviously) in whatever tenses you have been working on. I put a wide range of questions on the board/screen. I get the students to do as many rounds of this with as many students as possible. 100% accuracy is not an issue. Teacher must go around, facilitate, monitor and provide feedback. This activity, too, is carried out in the absence of time constraints and communicative pressure.


3.5. Eliciting fast written response (teacher led) – So far the students have been working with only one set of pictures (e.g. Set 1). Now the teacher stands in front of classroom and shows three (or more) pictures on the board from Set 2 which, being similar to the Set 1 pictures, are likely to elicit language that has already been practised in the previous phases. The students must now write on MWBs as much as they can about each picture under time constraints. The aim of this activity is to recycle the language learnt so far but to also focus on developing fluency (i.e. fast retrieval from long-term memory under time constraint). The teacher can cue the students to the use of specific connectives and one or more tenses. For example, I divide the screen in three sections, ‘past’, ‘present’ and ‘future’ and place a picture in each section; the task is for the students to write something like: yesterday I went to beach, now I am shopping, later on this evening I am going to go clubbing with my friends.


3.6. Eliciting fast written response (student led) – students re-enact what the teacher did with the whole class (in activity 3.5) in small groups of 4-5. Students take turns in showing a picture from set 1 or 2 and asking a question about it whilst their peers answer in writing on MWBs (in the target language). This can be turned into a competition.


3.7. Unstructured picture-based conversation without support– Now students, equipped with iPad or other recording device, do oral pair-work again. This time with no support whatsoever and under time constraints. Partner A/B selects five or more pictures (a mix of set 1 and set2) for partner B/A and asks questions about them – totally impromptu. Recording is sent to the teacher without any editing. If time allows it, several rounds of this can be carried out; I usually do at least two per student.

4. Conclusion

The above teaching sequence is very easy to prepare and allows for tons of recycling. It is mostly learner-centred and lots of language is produced in the process. One of the advantages of the pictures tasks envisaged here is that it forces students to widen their repertoire of verbs, a wordclass that foreign language teachers often neglect.

On the road to autonomous speaking competence – How to use writing mats to effectively support oral communication and proficiency development through a minimal preparation learning sequence


  1. Introduction

Writing mats, like the FFL (French as a Foreign Language) one in the above picture can be extremely useful as a means to support oral communication. In fact, I usually refer to them as ‘talking mats’, as I rarely use them to scaffold writing. In this post I intend to show how writing mats like mine can be effectively used to boost oral proficiency in the context of a student-centred learning sequence which implements the teaching approach I have laid out in many of my previous posts and is firmly rooted in Skill-building Theory.

  1. The sequence

Step 1 – Select or create a writing mat. I personally like to create my own mats, but you can find many excellent writing mats for all languasges, including (English, Spanish, German, etc.) on Mats should be clear, well-structured and possibly contain the L1 translation. Often, writing mats are overambitious as one wants to pack in as much language as possible; my advice is to stick to the items that have the highest surrender value and wait for phase 5 (below) for new phrases to be added.

Step 2 – Pre-teaching of the writing mats vocabulary and getting the students acquainted with the mats. This phase can be ‘flipped’ and has the purpose to prepare the students for the effective use of the writing mats in lesson on two levels: (a) students learn the words/phrases included in the mats; (b) they learn their way around the mats. I usually make up worksheets which ‘teach’ the target language items (e.g. odd one outs; matching; translations; gapped phrases) and require them to construct sentences using the words/phrases listed in the mats. I also input the content of my mats into the (work-outs) vocabulary modules recycling the target items to death. Teachers can do the same on quizlet or memrise.

This phase is very important, yet it is often neglected by practitioners; the students need to be able to know their way around the mats so as to be able to use them effectively and efficiently under R.O.C. (real operating conditions = speaking under real-life like conditions).

Step 3 – Modelling / Listening. This phase is crucial in that it (a) models pronunciation; (b) how sentences can be constructed; (c) practices student listening skills; (d) reinforces vocabulary. The teacher will make up sentences using the mats uttering them at accessible speed and repeating the sentences as much as the students require (remember: you are modelling, not testing). The students, equipped with MWBs (mini white boards) will write the sentences in the L2 or  L1 (ideally in the L1 to show comprehension).

Step 4 – Using mats to create sentences. Now the students work in groups of four or five taking turns in  creating sentences and uttering them, just as the teacher did in STEP 3. The students will have to write down the L1-translation on the mini-boards whilst having access to the mats – this can be turned into a competition. After a few rounds the sheet can be removed and only the sentence-makers will have access to the mats.

Step 5 – Scaffolded Practice. In this phase the students interact with each other using the mats as a scaffold. A typical task at this stage includes giving them short role-plays containing prompts in the  students’ L1 or in the L2 which will elicit L2 language items found in the mats. Example:

Partner 1: ask partner 2 where they went on holiday;

Partner 2: say you went to the South of France

Partner 1: ask where they stayed

Partner 2: answer you stayed in a luxury hotel in Nice


‘Find someone who’ (with cards) is one of my favourite structured practice activities, too, for this phase (see example at: ).

Another activity involves a structured conversational exchange in which each student is given a card with a series of fictitious details (the cards used for ‘find someone who’ can be recycled here). A different card is given at each round. Here is an example in the context of the topic ‘Holidays’:

Name: Jules

Holiday place: La Rochelle

Duration: three weeks

Accomodation: Three stars hotel:

Activities: tennis, diving, swimming, windsurfing

Weather: hot and windy


A list of possible questions to ask is displayed on the classroom screen. The students go around asking three of four questions in whichever order they like and note down their peers’ answers in the L1 or L2.

Step 6 – Transforming and Expanding. This phase aims at ‘stretching’ the students beyond the boundaries of the writing mats and prepares them for the unstructured practice that will follow. The students are now given time and resources to expand their answers using language that may not be found in the mats. The teacher plays an important facilitating role.

In this stage I usually engage students in online ninteractional writing using Edmodo. I ask the students to post questions of their choice and to reply to each other in the target language. Fluency is the focus, I do not really care about mistakes. Interactional writing gives the students the time to use online resources to expand and transform the stock phrases found in the mats.

Another activity I use at this stage is providing students with visual stimuli which force them to use new language. This is done in writing.

As a way to conclude this phase and pave the way for the subsequent unstructured phase I usually ask the students to answer questions on MWBs under time constraints. This forces the students to retrieve language under R.O.C.. Accuracy is still not a concern, unless it impedes communication.

Step 7 – Unstructured / Semi-structured practice. Now the students should be able to have a go at interacting orally in the context of information-gap activities. The easiest to prepare is obviously the conversational exchange on the topic in hand. Questions can be set or student-initiated. The mats and the notes made in the previous phase can still be used for a round or two, then they should be completely removed.

At the end of this phase I usually get the students to record themselves talking in pairs without any support using Voice Recorder Pro (free app); they then convert the recording into MP3 and send it to me.

Step 8Vocabulary consolidation. I usually have an extensive phase at the end of this sequence in which I recycle all the lexis covered. Snappy (low-stake) quizzes under time constraints (students’ answers on MWBs) can be used as a plenary to conclude the cycle. As homework: a vocabulary work-out on where I usually input the lexis included in my talking mats.

Caveats and conclusions

  1. Do not correct unless error impede communication;
  2. Monitor students constantly and reward creativity with the language even if the utterances are grammaticality wrong as long as  they convey the message;
  3. Make a physical note of the errors you hear more frequently reserving to deal with them at a later stage;
  4. The use of writing mats is much more effective when students have been systematically taught decoding skills (i.e. the ability to convert the graphic form of a word into its phonetic rendition).
  5. The above sequence may take more than one lesson. In fact, it is better if it does!

I have been using the teaching sequence just described for years. It is relatively easy to plan and resource and usually yields effective uptake of the target structures/lexis whilst paving the way for autonomous speaking competence (spontaneous talk). It is high pace and, apart from the modelling and feedback phases you will engage in when you feel fit, it is totally learner centered.

To find out about more my approach to fostering spontaneous speech in L2 learning, please refer to the book I co-wrote with Steve Smith, The Language Teacher Toolkit

Spontaneity in the foreign language classroom: how do we forge autonomous L2 speakers?


0. Introduction

In a previous post I concerned myself with the notion of spontaneity in target language production and provided some pedagogical ‘tips’ on how teachers can foster it in the  typical foreign language classroom. Today, during the professional learning time that Garden International School allocates to discussion and reflection every Friday afternoon, I discussed the issue further with my colleague Dylan Vinales, Head of Spanish, and experienced MFL teacher. This post encapsulates the main points of our discussion and expands on it.

1.What do we mean by ‘spontaneous talk’ in target language production ?

What does the notion of ‘spontaneous talk’ actually refer to in the context of L2 learning? In our view it means that when an L2 learner produces speech to initiate a conversation or respond to an external stimulus they do so ‘thinking on their feet’ , so to speak, without any pre-planning and without relying on any sort of support (e.g. vocabulary lists, talking mats, dictionaries, etc.). In other words, spontaneity equates with unplanned autonomous speech production.

Of course, as I advocated in my previous post, a competent L2 speaker is not simply one who can produce language ‘spontaneously’; s/he must first and foremost be intelligible and fluent and will possess a wide-ranging enough repertoire of vocabulary and discourse functions to be able to communicate effectively across various contexts.

Hence, any sensible foreign language teacher would not aim at learner spontaneity as divorced from fluency and intelligibility as the end-goal of their instruction; and whilst reserving to place a greater focus on accuracy and complexity at later stages of development, they will also aim at developing a level of learner mastery of L2 grammar use high enough to allow for generative power – the ability, that is, to effectively manipulate language grammatically/syntactically so as to generate new utterances from phrases /sentences acquired as formulaic items (e.g. from ‘I want you to know’ to ‘I want them to know’; from ‘To go camping’ – they go camping; from ‘ he plays with us’ to ‘they play with me’).

2. Define the teaching and learning focus

Developing spontaneous talk has been a very ‘trendy’ topic in the international teaching community in recent years. These days every L2 language educator posits ‘spontaneity’ in language production as the ultimate desirable goal of L2 teaching. However, I do wonder how much MFL teacher professional development and planning time is invested in figuring out ways to bring about spontaneity; how much effort is put into planning for spontaneity-fostering activities; how much formative assessment is devoted to tracking learner development across this all-important dimension of oral proficiency development.

For spontaneity to be attained teachers must keep the achievement of the ability to produce unplanned fluent intelligible talk in their focal awareness from the very early stages on L2 learning. A big chunk of their short, medium- and long-term planning must regularly focus on this steep goal; and this needs not be detrimental to the development of the other skills since, as I will discuss below, listening, reading and even writing  play an important role in the process.

3. Fostering Intelligible fluency

Fluency, conceived as a measure of time-to-word speech ratio, refers to the automatization of speech production; the speed, that is, at which words are retrieved from long-term memory to match a speech plan and uttered as part of an intelligible speech production unit (e.g. a sentence). Hence, it can only be achieved through masses of practice in retrieving language from long-term memory under R.O.C. (real operating conditions). Consequently, involving the students in oral interaction as much as possible is imperative; this calls for frequent student-to-student and/or student-to-expert speaker interaction.

4. Classroom practice: from controlled to unstructured

Let me reiterate here a point I have often made on this blog: the importance of starting from an imitative, highly scaffolded and controlled practice stage in which the students receive lots of prompts and support. This stage needs to be one of intensive and extensive practice; it can take a whole lesson, or even two, if the attainment of spontaneity is truly a priority. Throughout this stage the oral activities should become increasingly more challenging and should elicit more varied and complex responses. Activities may include: role plays/dialogues with visual or L1/L2 cues; Find someone who with real or fake identities; Oral translations; Surveys; Simulations; etc. Steve Smith, with whom I am currently authoring ‘The MFL teacher’s handbook’, outlines concisely what controlled practice activities may include in his latest blog:

The highly scaffolded stage will be followed by a consolidation and expansion phase in which the language practised and learnt during the first phase is reinforced through activities aiming at strengthening retention in preparation for the next phase, in which communicative practice will occur without scaffolding. This phase, too, should allow for extensive practice. Interactional writing activities can be used during this phase (see my post at  for this). During this phase the students should be encouraged to take risks and expand their vocabulary autonomously.

The final stage is the autonomous stage in which communication occurs without scaffolding and in which accuracy is not a concern unless it impedes communication and errors should go untreated (common errors raising concern can be dealt with at the end of each round of oral interaction or at the end of the lesson). Whilst scaffolding materials are removed, the teacher will play an important role, monitoring, facilitating and providing feedback on student performance. At the end of this phase the students can be asked to video themselves talking in pairs – rigorously without a script.

In a previous post I proposed a framework which integrates the three phases just outlined with emerging digital technologies used successfully in our school by myself and other colleagues (see: )

Please note that it is crucial that throughout each phase teachers ensure that students are exposed to an increasingly wide range of questions, as ‘spontaneous’ speakers must be able to react to as many external stimuli as possible in real time. The complexity of the questions should ideally increase, too. Research in L1 and L2 acquisition clearly indicates that the variety, length and complexity of the questions learners are asked play an important role in language acquisition.

This also calls for a form of flipped learning which is not sufficiently encouraged and scaffolded in the typical L2 classroom: actively seeking opportunities for oral practice outside the classroom. Funny how since the advent of certain emerging technologies educators put enormous effort in promoting flipped learning of the digitally-mediated kind; however, very few language educators seem to focus on this more potentially beneficial form of autonomous learning, which can but does not have to involve the digital medium.

Think of the enormous benefits of having the vast majority of your students practising oral skills regularly outside lesson time. Yet, how many MFL meetings are devoted to try and work out ways to create opportunities for and instill in L2 students the desire to do just that and to scaffold the process in a principled way? The real flipped MFL classroom is the classroom where students WANT to carry on learning by their own free will; where speaking happens after the lesson is over. In international schools like the one I work in, where there are plenty of target language speaker students and parents, not to explicitly and systematically foster this kind of engagement is a massive missed opportunity.

Another dimension of fluency refers to speed of sound production. When our articulators (larynx, mouth organs, etc.) have not automatized effectively the pronunciation of the target language sounds, speech production will slow down and vocabulary retrieval will be hampered too, as memory is phonologically mediated. It follows that pronunciation must be focused on consistently, too. This will call for instruction in micro-listening and decoding skills of the kind I have advocated in my posts on these topics.

Throughout the process, teachers must obviously ensure that their learners’ output is intelligible. Hence, the practice of getting students – when working in TL oral pair-work – to jot down the meaning of what their partners/interlocutors have just told them in response to a question may be useful. Yes, it does slow down the conversation, but (a) it is a real life task (L2 speakers interpret for their peers all the time); (b) it makes the listener pay more attention to their partner’s input; (c) you will not get the students to do it all the time. Say you have organized an oral communicative activity such as a GCSE style interview; you may want to engage each student in four rounds of interviews; the students will do the L1 translating/interpreting for two of the four rounds, whereas for the other two rounds, the listener will focus on providing feedback on their partner’s pronunciation, range of vocabulary, correct use of tenses or any other language feature(s) constituting the focus of the lesson.

Here are some other practical easy-to-set-up activities to promote fluency other than oral learner-to-learner communicative activities. They are very effective as pre-communicative activities as they foster fast retrieval from long-term memory whilst not involving the use of the articulators and the added emotional stress and cognitive load of oral interaction:

  • Fluency is about speed of retrieval of the required L2 items from long-term memory. Hence, getting students to respond to a visual stimulus, a sentence to finish up in their own words or questions under time constraints (possibly setting a minimum required number of words) provides good training. I do this quite often as a starter with my GCSE classes other with MWBs (mini white boards) or on a google doc shown on the classroom screen. It is paramount to vary the type of stimuli / questions as much as possible and avoid merely sticking to the topic-in-hand;
  • Engaging the students in translations of short sentences on MWBs – again under time constraints – accomplishes the same purposes whilst forcing the students to be accurate and allowing the teacher control over student output. Students are shown short sentences on the classroom screen and must translate under time constraints. 100% accuracy is not a must;
  • Where logistically and technologically possible, give the students an iPad or other recording device and ask them to record themselves talking about a specific topic, possibly one that is not too recent but that you know they can talk about. Better if you give them five or six specific sub-topics to focus on in the way of bullet points; including one or two sub-topics which you know they will find particularly challenging in terms of vocabulary will give you an idea of how well they can cope in terms of compensatory strategies deployment.
  • Interactional writing (see my post:

A final point: teachers often ‘compartmentalize’ teaching and learning by topic; what I mean is that whilst dealing with a unit of work (e.g. work and career) the oral activities during the 6-8 week period devoted to that unit only cover the topic-in-hand. That should not be the case. To develop spontaneity, ways need to be found for learners to be engaged ever so often during those 6-8 weeks in conversation/information-gap activities on past topics or even on topics never encountered before to ‘test’ their communicative limits

5.Vocabulary repertoire and communicative functions

Speaking autonomous competence requires knowing a wide enough range of vocabulary to express yourself across a sufficient range of semantic and functional contexts. In other words, for a learner to spontaneously produce language utterances that enable them to meet their communicative goals, they must possess a repertoire of lexical items (words and stock phrases) with high surrender value which allows them (a) to talk about a wide range of topics; (b) perform the most important communicative functions (personal, interpersonal, directive, referential and imaginative; complete list at: ).

Hence, vocabulary teaching must be an important focus of classroom / out-of-the-classroom L2 learning. Whilst it is more time-effective to ‘flip’ vocabulary learning, it may be more beneficial to choose web-tools/apps which do not simply focus on word level and teach words as discrete items (e.g., or ) but which enable students to learn the words in context and across as many linguistic contexts as possible.

It is important for vocabulary teaching which aims at developing fluent spontaneity to aim at (a) fast retrieval and (b) transfer across contexts. Here, too, the MWB translation activities under time constraints of the kind outlined above can come in very handy. Recycling, as I often reiterate in my posts is extremely important for reason I have already explained to death on this blog. Teachers must plan carefully for recycling ensuring that each unit of work provide opportunities for the recycling of old ‘material’. This rarely happens in MFL department or course-book schemes of work, yet is possibly the most important factor in determining learners’ retention of the target vocabulary.

An important` point: very often British L2 textbooks and online resources provide detailed lists of nouns but not verbs. However, without the mastery of a wide repertoire of verbs the generative power of nouns is limited and the ability to talk across a wide range of topics is drastically reduced – affecting the learner potential to talk ‘spontaneously’ across context. Adjectives/Adverbs are often neglected, too.

For a principled approach to teaching vocabulary refer to my post ’13 steps to effective vocabulary teaching’.

6. Aural skill instruction: providing models and teaching listenership vs ‘quizzes’

In order to effectively foster fluent spontaneity teachers need to change their attitude to listening skills instruction. The listening tasks teachers must involve their students in will not be those which aim at testing their inferential strategies (e.g. true and false quizzes), but those that model useful language and teach learners to be effective interlocutors (e.g. be able to function effectively in the context of a conversational exchange by understanding and responding to their interlocutors’ utterances).

6.1 Modelling useful language

Students must be engaged in listening activities involving exposure to useful comprehensible input which aims at (a) reinforcing and expanding their existing repertoire of vocabulary and communicative strategies; (b) enhancing their pronunciation and decoding skills; (c) modelling ‘spontaneous’ talk.

For (a) and (b) to be achieved teachers must involve students in tasks which require them to pay attention to lexis and sound. A very easy-to-set-up activity is obviously translation. The teacher utters useful sentences and students translate them on mini boards. Transcribing short texts can also help. Another  activity involves providing the students with a – as literal as possible – gapped translation of a listening text and play the audio track; the task: to fill in the gaps in the translation (in English of course). Jigsaw listening and L2 gapped-text tasks can be useful, too.

As far as (c) is concerned, videos of native speakers (not actors) engaging in spontaneous interaction / talk can be beneficial as they model useful linguistic and paralinguistic features of native-speaker spontaneous talk.  Such videos can be found on the web or can be created by teachers interviewing native speakers (e.g. language assistants or L2-native-speaker colleagues). Because they don’t have to last more than a few minutes and have to be as spontaneous as possible, they require no planning – apart from deciding on the questions to ask. Thus the process is not very time consuming.

6.2 Listenership

Listenership development can be fostered by involving students in any listening or oral communicative activity which requires them to understand and respond. Some of the MWBs activities outlined above can be recycled in this context, too. Videos of conversational/transactional exchanges between native speakers where students need to demonstrate understanding of the questions being asked. Frequent practice in answering a wide range of questions and responding to statements – for instance, as a starter, make a statement in the L2 and ask the students to respond saying if and why they agree/disagree with it (example: the food in the canteen is unhealthy). If we are aiming at fluency too, we can do this under time constraints. Obviously, any oral communicative activity involving negotiation of meaning will serve this purpose,too.

7.Encouraging risk-taking and modelling and practising compensatory strategies

Students should be encouraged to take risks. For this it is crucial that errors are tolerated in unstructured oral communicative practice. Risk-taking, however, requires some scaffolding, too. By this I mean that students should be equipped with effective strategies to cope with communication breakdown, e.g. how to make up for lack of vocabulary.

We advocate the teaching of the following compensatory strategies:

Coinage – this involves showing the students how you can create an L2 word from and L1 word (this strategy does not obviously apply to all languages). For instance: how to get the French for university, city or proximity by changing the ‘y’ to ‘é’ or how to obtain the Spanish equivalent of verbs ending in ‘-ate’ in English by replacing ‘-ate’ with ‘-ar’ e.g.: exagerar, alternar, enumerar, etc. This also entails encouraging students to create new L2 words that may actually not exist but can be understood, such as ‘ her eyes are watering’ for ‘she is crying’;

Paraphrase – this involves teaching students how to make up for lack of vocabulary by providing a basic definition/description of a word (e.g. for ‘glass’ – you use it to drink water);

Approximation – this involves using a word that is close enough in meaning to the one you need (e.g. ship for sailboat) with or without the use of miming to enhance expressive power;

Teaching the above strategies can be a lot of fun. Some teachers may frown upon the idea of their students learning how to produce erroneous L2 items in order to get the message through. However, these strategies are not simply compensatory strategies; they are ultimately learning strategies in that their use usually results in a correction which provides the accurate L2 form.

Spontaneity does require the kind of risk-taking and creativity that the application of these strategies entails. Compensatory strategies allow a speaker to keep up the spontaneous talk even when they lack vocabulary and grammar; hence, they are important communicative ‘tools’.


Spontaneity in oral production can only be achieved through tons of practice, especially of the productive kind. For learners to be spontaneous they need:

  • Masses of vocabulary. The more vocabulary a learner knows the more they will be likely to communicate;
  • Practice in manipulating stock lexical phrases (formulaic language) to adapt them effectively to various linguistic contexts. This may require some teaching of grammar and syntax;
  • A classroom climate which encourages one-to-one oral interaction and risk-taking and prioritizes fluency and communication over grammatical accuracy;
  • Extensive controlled and highly scaffolded one-to-one oral interaction practice which leads to unstructured practice. Such practice should aim at developing transferrable communicative routines, whose automatization will ultimately lead to spontaneity;
  • autonomous vocabulary learning and seeking oral interaction opportunities outside the classroom;
  • practice in compensatory strategies;
  • listening which models useful language and transferrable communicative routines.

A final point: if teachers do value spontaneity, fluency and intelligibility in speech production as the most important end-goals of L2 learning, then they should ensure that this is reflected in their curriculum design and everyday teaching. The achievement of spontaneity requires relentless practice and systematic formative assessment. Fortunately, as I have attempted to show above, focusing on spontaneity in speech does not harm the development of the other three skills. Even reading, which I have not mentioned thus far, can indirectly play an important role in enhancing spontaneity by widening our students’ vocabulary – provided, that is, that they possess effective decoding skills which enable them to accurately transform L2 graphemes into L2 sounds.

Five tips to encourage spontaneous talk


Today I came across the following tips for encouraging spontaneous talk whilst browsing the web.

Top 5 tips for encouraging spontaneous talk in the MFL classroom

  1. Have keyphrases on the wall so they can use them when they want – e.g. ‘I would like…’
  2. Take all opportunities to encourage students – ‘I need a pen’
  3. Give as much supportas possible – literacy mats/vocab sheets/peer help
  4. Rewardbravery!
  5. Build it into routineslike entering the classroom (hold up a MWB of a key phrase at the door like ‘opinion in French’ and they have to give an example as they come in. That way each child has said something in TL before you’ve even done the register!)


These are useful tips which can help teachers create a culture which may encourage pupils to talk in class, no doubt. But do they really promote spontaneous talk? In order to answer this question let us have a look at what ‘spontaneous’ means. The Oxford advanced learner’s dictionary offers the following definition of the adjective ‘spontaneous’:

Performed or occurring as a result of a sudden impulse or inclination and without premeditation or external stimulus:

Michelle Cairns’ tips reflect the kind of pedagogic advice given by several other MFL educators and are rooted in a misunderstanding of the notion of ‘spontaneous talk’ both in terms of the cognitive processes it involves and the developmental mechanisms that lead to the ability to produce speech autonomously.

Since ‘spontaneous talk’ refers to the ability to produce speech without prompts, Michelle Cairns’ tips, like those dished out by many other MFL educators, refer but to the very embryonal stage of spontaneous talk, what in my model of speaking skill acquisition I refer to as the ‘imitative’ stage. However, in order to bring our learners from the ability to ‘parrot’ phrases on the wall or on writing mats to what applied linguists call ‘autonomous speaking competence’ (i.e. spontaneous speech) it takes way more than those five tips, unless we hold a very simplistic view of oral proficiency acquisition.

For spontaneous talk to be developed effectively in large classes (i.e. classes of the size typically found in secondary schools), the top tip according to much research (e.g. Varonis and Gass, 1985; Pica,1996; Macaro, 1997; Donato and McGlone, 1997; Macaro, 2007) should be to engage students in NNS (non-native speaker)-to-NNS oral interaction in the context of tasks requiring negotiation of meaning (e.g. information gap tasks).

Whilst the ‘parroting’ stage alluded to above is important, especially when it ‘drills in’ carefully selected high frequency phrases useful in the real world, the most important part of the oral-skill acquisition process occurs when students are practising the skill of putting a message across to an interlocutor, regardless of the mistakes they make. Only after much such practice, supported by writing mats, dictionaries and expert TL speakers at the early stages one can develop spontaneous speech.

Many teachers refrain from staging oral tasks involving oral learner-to-learner interaction for the following reasons:

  • Students do not always stay on task and lapse in their L1
  • Teachers are concerned about the negative effective for learning of pairing students of different levels of proficiency; the more able learners might not be ‘stretched’ enough if they work with less able ones;
  • By working with their peers, learners might pick up erroneous utterance that they might end up fossilizing;
  • Not all students enjoy it.

As for (a), studies by Brooks and Donato (1994), Knight (1996) and Brooks et al. (1997), Anton and Di Camilla (1998) found that students do generally stay on task and do tend to use the TL most of the time. What is more interesting, even when they do lapse into their L1, they tend to use it for TL learning enhancing behaviours, i.e. (1) to facilitate the negotiation of meaning; (2) to talk about the task (e.g. how to conduct it; what the expectations were). These studies also produced an interesting finding: learner-to learner interaction tasks promote a whole host of self-regulation strategies which enhance TL acquisition and that I observe every day in my lessons during such activities (e.g. whispering to oneself to repeat a word they have just heard from a peer or teacher in order to commit it to memory) – a further reason to implement such classroom activities.

As for point (b) and (c), above, Iwashita (2001) investigated if pairing students of different levels of proficiency might have adverse effects on the frequency of interaction and the modified output that would result from the interactions. She got her students to work in three proficiency pairs: High-High, Low-Low and Hig-Low. She found that the lower proficiency students gained a lot from working with higher proficiency students and produced lots of modified output, whilst the higher proficiency learners were not seemingly disadvantaged.

Finaly, as far as point (d) is concerned, Macaro (1997) found that oral pair-work made most of the students feel comfortable and they  reported learning and remembering a lot. Very few of the students reported negative attitudes.

Obviously, the process of acquiring spontaneity in TL speech production will have to be supported by the teaching of masses of TL vocabulary (not just nouns – but a wide range of verbs, too), of discourse function markers and by lots of exposure to comprehensible aural input.

Computer/ Tablet-mediated interactional writing (see my previous post on it) can also play a very important role, as it allows the learner to converse through the written medium at a speed high enough to practise fast TL processing but slow enough to allow for more self-monitoring.

The one tip from Michelle Cairns’ post that I would definitely ‘save’ as pivotal in fostering spontaneous speech is to ‘reward bravery’, not only to create an atmosphere conducive to risk-taking and tolerance of error, but also because it encourages the deployment of another important catalyst of spontaneous speech development: communication strategies, the ways, that is, in which MFL learners compensate for their lack of language by coining new words, paraphrasing or explaining unknown vocabulary, resort to gestures or onomatopoeias or overgeneralize TL rules.

These are my top tips for developing spontaneous talk beyond the obvious imitative stage that Michelle Cairns’ pedagogic advice in her very useful blog referred to:

  1. Teach masses of vocabulary, ensuring there is a balanced mix of nouns, adjectives and verbs. As Macaro (2007) and Conti (2015) point out, far too often teachers neglect equipping their learners with a wide enough range of verbs. Extensive reading should be promoted as a way to acquire new vocabulary;
  2. Involve students in lots of oral interaction involving negotiation of meaning and practising a wide repertoire of communicative functions (e.g. comparing and contrasting; persuading; agreeing and disagreein). Oral interaction tasks should be sequenced wisely in terms of the cognitive load they place on the learners; hence, one would start with highly controlled tasks (the imitative stage Michelle Cairnes alluded to) and gradually move to less structured communicative activities (e.g. the ones investigated by Varonis and Gass, 1985, and Macaro (1997);
  3. Expose learners to lots of comprehensible aural input (e.g. through narrow listening tasks). Increase the amount of listening tasks which aim at modelling language use rather than testing students. In other words, use the listening tasks to draw students’ attention to the language items you want them to ‘pick up’ rather than simply ask to guess if statements are true or false or identify details in the text;
  4. Model to students creative ways to put a message across to an interlocutor when they do not know vocabulary; i.e. train them in the deployment of communicative strategies;
  5. Ask them to practise digitally-mediated interactional writing independently with their peers or other target language knowers on the internet – the way I picked up two of my languages.

In conclusion, the acquisition of ‘spontaneity’ in speech production (autonomous speaking competence) is a complex process which goes from an imitative stage in which the learner is highly dependent on models and scaffolding to an autonomous stage in which the learner has a sufficiently wide repertoire of vocabulary, discourse function markers and compensatory strategies. Teachers must plan for it carefully and work towards its attainment through the systematic application of the five principles just outlined whilst creating a non-judgmental learning environment conducive to risk-taking. Ultimately, the extent to which students become effective autonomous TL speakers will largely hinge on the amount of vocabulary they know, the speaking practice they will have received and their willingness to take risks.

A final point: spontaneous speech without the development of fluency intended as the automatization of intelligible speech production is not conducive to effective communication under real operating conditions (e.g. real life communication). Hence, in my view, MFL educators should posit ‘fluent spontaneous speech’ as the desirable goal of speaking proficiency instruction, where fluency refers to time-to-word ratio in intelligible-speech production.

Apologies to Michelle Cairns for the criticism of her tips, which is not specifically directed at her or her blog – which actually usually contains excellent teaching resources and advice for teachers – but to a general attitude towards MFL pedagogy found in many language teaching blogs which may be misleading as it presupposes and divulges an overly simplistic view of language acquisition.