Working on Fluency: Part 1, the theory

Posted: April 6, 2014 in ELT methodology
Tags: , , , ,

This is the second post in a series of posts on fluency: it’s about key insights from literature into what factors contribute to (or detract from) fluency; I illustrate most points on a number of real life examples that I used with my students in awareness raising tasks.

I wrote about what motivated me to start ‘digging’ in the first post here – I’ll also add there  the links to the remaining (more practical) posts when they appear.


Working on Fluency. Part 1: the theory

Fluency: what are we talking about?

Fluency is ‘the [person’s] capacity to produce language in real time without undue pausing or hesitation.’ (Skehan, 1996, cited in Thornbury, 2000). It can actually be objectively measured with a number of variables, such as speech rate (the number of syllables per second – including pause time) and the mean ‘length of run’ (the number of syllables between pauses).  However, there’s also perceived fluency (that is, whether the listener considers the speaker to be fluent), which is of course subjective and much less clear-cut, but nevertheless there has been some research into what features of speech influence the listener’s perception of fluency.

There’s also another important aspect of speech, suggested by Michael McCarthy, who calls it ‘confluence’. Confluence concerns the way the speaker maintains (or fails to maintain) the flow of conversation, linking what they say to what the previous speaker said.

Understanding fluency through understanding what we do when we speak

In order to understand where fluency can ‘break down’, its useful to look at a ‘break down’ of the skill of speaking. Here is a model of speech production proposed by Levelt (1989), according to which an utterance goes through the following stages:

1. Conceptualisation (aka macroplanning: taking decisions what to say, which communicative intentions are to be realized)
2. Formulation (micro-planning at the level of an utterance) Retrieving vocabulary and grammaring it up.
3. Articulation, which is pronouncing the utterance
4. Monitoring. Checking that what you said conveys what you meant, with a reasonable degree of precision, and repairing the utterance if need be.

There are two fundamentally different ‘resources’ that we draw upon to carry out the stages outlined above. We draw on our knowledge (e.g. the knowledge of the way conversation is structured in L2 culture, or the knowledge of lexical items) and automatic routines(which include ‘grammaring’ the lexis up and using our speech organs to pronounce the sounds). We draw upon different types of memory, located in different parts of the brain (called ‘explicit’ and ‘implicit’ memory) and employ fundamentally different processing (‘explicit’ and ‘implicit’ processing).  This distinction is important because it provides insight into the limitations of the human brain that are directly related to fluency. By way of a grossly oversimplified summary, in order to formulate an utterance, we use working memory for planning as well as storing lexis; we can only store ~7 items retrieved from explicit memory in the working memory at a time, while grammatical processing does not take up working memory unless one consciously recalls and applies a rule.  The distinction is also important because explicit and implicit processing is developed/automatised in different ways, and thus they call for different treatment in class (see this book for a fascinating summary of recent research; there’s more on explicit and implicit memory/processing in the books referenced in this post).

Getting closer to a fluent speaker’s performance

Stage of of speech production: Conceptualisation

  1. A fluent speaker…
    knows about the socio-cultural conventions for how conversations and monologues of different types are structured, e.g. uses scripts  and the structure of genres. For instance, when we walk into a restaurant, we know what we’re likely to say/what the waiter is likely to say and in what order this will be said – and this might be different in different cultures. In Russia we do not make small talk while buying groceries, so a Russian English language learner might be startled and might fail to respond altogether, let alone fluently, to an amiable English groceries seller.
  2. Implications for learners & teaching I haven’t learnt much on macro-planning yet, but clearly it’s useful to raise learners’ awareness of genre structures and give them a chance to refer to the structure as during planning stages of speaking activities.I also recently came across this presentation by Barry Tomalin: What do you say after “Hello”? Successful networking techniques. His ideas sound absolutely brilliant, I really want to try his framework with students to add some ‘meat’ to classes on ‘socializing’, but haven’t yet.

Stage of speech production: Formulation

  1. A fluent speaker…
    for efficient planning: uses spoken structures, which are much ‘looser’ than written grammar, stringing together a sequence or relatively independent clauses, or using ‘frames’ with slots
    for efficient vocabulary processing: stores in memory and retrieves lexis in chunks (formulaic speech units), which allows them to operate more than 7 words at a time (Nattinger and DeCarrico, 1992, cited in Wool, 2012)
    …formulates thoughts with varying degree with precision, sometimes resorting to very vague language
    for efficient grammatical processing
    …uses proceduralized routines, in contrast to consciously applying ‘rules’Here’s an example from a conversation between Oprah Winfrey and JK Rowling in which some of these features are evident: there are lots of short clauses linked together by conjunctions (and, but) and sentence adverbs/fluency markers (well/obviously); there are some sentence frames (conditional clauses, as well as relative clauses); there are lots of collocations (e.g. ‘a flash of clairvoyance‘) and some vague expressions (like)
    Winfrey: So, didn’t you know?
    Rowling: No.
    Winfrey: Wasn’t there part of you –
    Rowling: Part of me –
    Winfrey: Subconsciously, that knew? Yes.
    Rowling: I + I remember once + and it was like + it was likewell, like + I’m going to call it clash + a flash of clairvoyance now. + Obviously if [it hadn’t come true] it would just be some crazy thought + I had . But I do remember one day+writing Philosopher’s StoneI was walking away from the café + where I’d been working on +
    Winfrey: Philosopher’s Stone + which became Sorcerer’s Stone.
    Rowling: Which became Sorcerer’s Stone, exactly.+ So that’s the first novel. + And I had this moment where + I suddenly thought + It was like another voice speaking to me + and the voice said + “the difficult thing is going to get published. + If [it gets published] [it will be huge.]”
    Winfrey: Wow.
    Rowling: And [that is] + exactly what it was.

    The complete 40-minute interview is available here:

    Implications for learners/teaching

  1. Planning. I believe that one important step is for students to become tolerant of spoken grammar and, instead of trying to produce syntactically complex structures characteristic of written language, analyze real speech for typical patterns of reasoning/’frameworks’ and practice using these simple frameworks to structure their own short utterances. For example, when practicing discussions/decision making, a typical argument might be broken down into ‘suggestion + reinforcement (state why suggestion is advantageous) + summary, e.g. I suggest you + (use skype) // which will (save you a lot of money) /make it (easier) (for you) to //so you’ll be able to…).When I was teaching IELTS students, it helped a lot to categorize the typical questions asked in the spoken part of the exam, provide students with a framework for each questions type and train them extensively in using this framework. Below is an example of materials I was using (adapted from Clark (2007)). It felt a bit like cheating at the time, but now I think that this might have mirrored the way skilled speakers micro-plan their utterances, as it allowed the students to plan at the level of a clause, moving ‘from column to column’ concentrating on producing one or two clauses at a time.
    Solutions: introductory sentence Introduce one option Explain what effect this solution would produce Link to another option
    Well I suppose there are a number of actions that could be taken.

    Well, I personally believe that there are [two] possible ways to tackle/address these problems.

    Well I think we could go about this in a number of ways.

    One way to …. would be

    An obvious measure/answer would be to
    One possibility/option is to

    By doing this, we would

    This would enable/help [us] to…

    As a result..

    This would have a huge impact on

    This would have a (profound) effect on

    Another solution would be to..

    A further step to [eliminating poverty sth] would be to


  2. Vocabulary. Noticing formulaic language is tricky because, at some level, it’s ‘obvious’: in contrast to individual ‘new words’, students often understand collocations and thus don’t identify them as ‘new’ language. Moreover, in my own experience this is almost impossible to do on first read/listen/watch: while your brain is totally engaged with meaning, it won’t notice anything that does not pose difficulty in that respect.

    When I asked my B2/C1 students to identify collocations in the following extract from the interview with JK Rowling, they
     tended to notice smaller chunks and ignore the ‘bigger’ picture (e.g. That was – That was easily the most frightening thing I’ve done in my life.) they identified the expression ‘in my life’ as a chunk, but didn’t notice the frame ‘That was [easily] the most [adj] thing I’ve [V3] in my life.);
    • also, their estimate was that they understood around 90% of these expressions and used around 10%.

  3. An awareness raising task
    Underline fixed expressions that contain the words in bold (suggested answers are in the third column):

    1 Rowling: Yeah, hugely. Hugely. But the odd thing is that that’s just life, isn’t it?  the odd thing is that…
    it’s just life
    2 The books wouldn’t be what they are if she hadn’t died. I mean her death is on virtually every other page of the Harry Potter books, you know?  wouldn’t be [] if
    virtually every // on every other page
    3 At least half of Harry’s journey is a journey to deal with death in its many forms, what it does to the living, what it means to die, what survives death – it’s there in every single volume of the books.  … in its many forms
    every single volume
    4 Rowling: A little. That was – That was easily the most frightening thing I’ve done in my life. Easily. It felt very exposing because this wasn’t me reading-out words that had already been approved. Do you know what I mean? I used to be borderline phobic about public speaking. [easily] the most [adj] thing I’ve [done/V3] in my life
    this wasn’t me v-ing
    public speaking
    5 Winfrey: Wow, really?
    6 Rowling: Yeah. Really. Like shaking so badly I couldn’t – I didn’t know what sentence I was on. So I’ve come a long way. I’m still not – public speaking I’ve got better at but there are things like having to give a speech on T.V. still scares me so much I can’t deal with that very easily.  I’ve come a long way
    I’ve got better at
    to give a speech on T.V.
    can’t deal with [that] [very easily]

    I’ve been trying to ensure students focus on collocations ever since I started teaching – more on that on the more ‘practical’ post on fluency – but some of the techniques that I personally obviously undersuse are:

    > Horizontal alternatives to vertical lists Providing collocations to all new vocabulary
    > Consistently training students to use collocations dictionaries and other lexical resources ( is a recent exciting addition to that list)

  4. Grammar It’s been suggested that in order to achieve automatization, students need to speak under time pressure, as only then are they forced to ‘restructure’ the procedures they currently employ by ‘assembling’ them into larger units. I try to use this insight by regularly giving the students the chance to repeat tasks under tighter time constraints, but one most important lesson I’ve learnt here is that it’s important to  take students on board – that is, make them aware of the rationale behind the activity, as otherwise they don’t see the point in making that extra effort,  cutting corners and meeting the tighter constraints simply by omitting part of the message.

Stage of of speech production: Articulation

  1. A fluent speaker..
    (a)  uses prominence to direct the listener to the words carrying meaning
    (b) chunks their utterance into units (pauses at clause boundaries and not inside clauses)
    (c) pronounces highly frequent chunks, including stretches of function words, such as ‘I’m in my’, fluently (This was what struck me most in all research into fluency – the observation that the listener’s perception of fluency hinges on how fast and naturally chunks are pronounced: ‘It can be stressed that it does not matter how slowly and carefully the rest of the utterance is, or needs to be, constructed. Provided the ‘chunk’ is said fast, the utterance will sound natural; the opposite, a fast message with a slow chunk, will sound completely unnatural and non-fluent.’ (O’Keefe et. al,. 2007)

    Awareness raising task
    I useed this video to raise awareness of various features of pronunciation (especially fast pronunciation of chunks)

    Students listened a number of times
     first time to respond to the message/understand the gist and talk about their experience with obsolete gadgets
     second time they were provided the gapped text in the first column below with ‘|’ marks showing pauses deleted and asked to mark the pauses; then they listened again and underlined stressed syllables
     then they listened one more time trying to complete the gaps (almost impossible), and one more time listening out specifically for speed (it’s only now that most students realize that the gaps are so difficult to fill not just because the words are ‘swallowed’ but that they’re specifically pronounced much faster than the rest of the utterance;
     they then practiced reading the passage: first time concentrating on the pauses, then on the pauses and stress and finally, having practiced the pronunciation of ‘fast’ chunks separately, all of it;
     finally, they discussed their favourite gadgets – the task was to talk but also to try stress the important words as they talk

    Louis C.K.: Ya, because | Everything is aMAzing right now | and NObody’s happy. | Like | in MY LIfetime | the CHANges _______ world _________ inCREdible. . Louis C.K.:  Ya, because | Everything is aMAzing right now | and NObody’s happy. | Like | in MY LIfetime | the CHANges in the world have been inCREdible. .  /ɪnðə/    /həvbɪn/
    _________ kid ________ ROtary PHOne. |We had a phone you had to STAND NEXT to | _________ DIal it, (yes) __________. |___________REalize how PRImitive, when i was a kid we had a ROtary PHOne. |We had a phone you had to STAND NEXT to | and you had to DIal it, (yes) you know |you know, you ever REalize how PRImitive, /wənʌwəzə//əntə/
    /jənə/  / jəve/ 
    ______ making SPARks _____ phone | and you actually would HATE people with ZEros ______ NUmbers’cause it was more (right) oh, __________ two zeros, | screw THAT guy, ____________, ugh… you’re making SPARks in a phone | and you actually would HATE people with ZEros in their NUmbers’cause it was more (right) oh, this guy’s got two zeros, | screw THAT guy, why do i wanna ugh… /jə     /inə//inðe//ðɪsɡaɪzgət//wʌɪdəʌwənə/
  2. Implications for learners & teaching(a) Learners need awareness raising activities in which they identify prominent syllables and practice highlighting prominent words in while engaging in a communicative task; I think that, as it often said about pronunciation activities, the best approach here is ‘little and often’.
    (b) I notice that my students often make in-clause pauses when they are searching for vocabulary; what they do is produce the ‘grammatical part of the clause’, which, having been drilled repeatedly, is the ‘easy’ bit,  and then drag out the final function word while formulating the rest of the clause (e.g. /I have tooo [pause] get up early/). So far I haven’t come up with any ways to address this (at least practicing the correct pronunciation of the structure doesn’t produce any visible effect, as the students revert to unnatural pronunciation as soon as they concentrate on the message).
    (c) Regarding chunks of function words, McCarthy suggests the students practice saying them over and over again  (He lists the following chunks as top priority chunks, as they are top 20 3-word chunks in Cambridge and Nottingham Corpus of Discourse in English:


    (McCarthy, 2006)

    Cauldwell(2013:315) lists another list of clusters of 3 and more words commonly squeezed together.

    > To help the students to produce these chunks with natural pronunciation, it might be useful to raise their awareness of features of connected speech and, in particular, the way diphthongs ‘lose’ weak vowels (/aɪ/ in ‘I’ll’ becoming /ʌ/, /əu/ becoming /ə/ in ‘you know’, etc)

    > I think it’s also paramount that students are made aware of the effect non-fluent pronunciation of chunks produces on the listener’s perception (at least if they do aim to sound fluent), and are consistently encouraged to practice fluent pronunciation whenever they study  functional language and formulaic expressions, either in class or on their own – at least my students normally try to shirk at this point of the lesson because they feel extremely self-conscious; in my experience, rationalizing the need for pronunciation work does help> Adrian Underhill repeatedly makes the point of the need for placing a physical demand on top or the cognitive demand  and challenging students to pronounce what they say with maximum fluency and connectedness possible (e.g. Underhill, 2013:215) and for me the way he demonstrates this is an inexhaustible source of inspiration.

Stage of of speech production: Monitoring

  1. What do a fluent speakers do?

     monitor their message after it’s been formulated and uttered
     use spoken structures that are characteristic of spoken language and that mirror the way attention shifts e.g. those olives, how much are they? (the topic is identified first, then the question is added – cf how much are those olives?); my brother, his wife (cf ‘my brother’s wife)  they are nice, those olives (the sentence starts out with the speaker’s perspective, and then gets ‘clarified’ to include the listener). In the interview extract above, JK Rowling freely starts a sentence and then changes it mid-way:  I remember once + and it was like + it was like + welllike + I’m going to call it clash + a flash of clairvoyance now. Language learners, on the other hand, often feel that they have to speak in full, ‘correct’ sentences and attribute their failure to do this not to natural limitations of the human brain, but to their own lack of proficiency – in short, very often they have unrealistic expectations of what they are to produce.
     Uses a range of expressions to signal the need for extra interpretation on the part of the reader (e.g. ‘you know’/’do you know what I mean?’)
  2. Implications for learners / teaching
    For me as a non-native speaker insight into monitoring was the most revealing. I realized that I, in contrast to Levelt’s speech model, was vehemently trying to monitor utterances before uttering them, because I was never sure would be able to finish the sentence I’d started – a fear that, I suspect, was well-grounded, as my speech was lacking all the features of spoken language outlined above.  

Stage of speech production: linking to the previous speaker’s turn

  1. Fluent speakers…
     react to what the previous speaker said (‘yeah’/’no’/’exactly’/’it is!’)
    comment (‘we’ll need to …’/’which won’t be easy’)
    reformulate (‘Was it good?’ ‘Yeah, amazing!’)
    repeat (Winfrey: Philosopher’s Stone which became Sorcerer’s Stone./ Rowling: Which became Sorcerer’s Stone, exactly.);  backchannel to show interest (mmm/really?)
     use discourse markers and sentence adverbs to start turns (‘well, basically’/’actually’/’so’/’right’).Awareness raising task (adapted from one of Michael McCarthy’s tasks in an IATEFL webinar)
    Again, the interview with JK Rowling mentioned above is a striking example of that – here’s a random sample. I asked students to underline look at the starts of turns and compare that to a transcript of A2 students speaking (I couldn’t produce a transcript of the students themselves due to time constraints, but it would’ve been better: record them doing a task, ask them to analyze turn beginnings in the interview, then listen to their recording to see what features are present.
    Winfrey: Are you in a place now where you can accept that you will always be rich?Rowling: No. Are you?Winfrey: Kind of. Getting there.Rowling: Really? I hope – I hope I – that sounds good.Winfrey: Unless I’m a complete fool.Rowling: But that’s it! Unless I’m a fool! And you know what? I’ve never been a fool with money so why worry? But I do. I think ‘God, if I blew this, how could I look everyone in the face?’Winfrey: But, you know psychologically it’s a difficult thing to come to terms with because it’s like saying – not allowing room for never say never.Rowling: Exactly.Winfrey: You know?Rowling: Exactly. And you feel – I feel – I don’t want to get complacent.

    Winfrey: Right.

    Rowling: I don’t want to take things for granted.

    Winfrey: Correct.

    Rowling: I just – I just – and after all. Well, you do know what, I’m talking absolutely rubbish, aren’t I? I’m talking rubbish. I mean really would have to be very stupid but, yeah, I do still worry.

    Winfrey: Really?

    Rowling: Yeah. Not all the time. I mean mostly I feel great.

    Winfrey: What do you actually think money has done for you? What does it do?

    Rowling: It frees you. That’s what it does. It frees you. That’s why it’s like a super power. You don’t – it frees you. I mean we don’t have to – the luxury of literally being able to sit down and say “where should we go for a holiday?” and not be, in any way, limited.

    Winfrey: I hear you don’t drive.

    Rowling: No, I don’t drive. No. Cars terrify me. I am really frightened of cars.

    Winfrey: So do you have a driver?

    Rowling: I – of – lately I have had a driver. Very lately.

    Winfrey: Is it true that you still take the bus? I read that you still take the bus.

    Rowling: Occasionally. Within the last year I have taken the bus. Definitely, yeah.

    Winfrey: Did you ever imagine your life being the way it is now?

    Rowling: No. Never. And I really, really mean never. It overshot the mark so ridiculously that I – I was so unprepared for it. This is a thing I think I’ve never really spoken about. I was a writer. I had no one near me professionally or personally who could in any way help me when I had questions like “what do you do when the press is searching your bins?” You know?

    Winfrey: Mhmm.

    Rowling: Really crazy stuff that happens. The stuff that makes you feel –

    Winfrey: But that doesn’t happen to most writers, you know?

    Rowling: Exactly. Exactly. So it took everyone around me totally by surprise.

    Winfrey: It’s not like if you’re an actress you could have expected that.

    Rowling: Of course! Of course. You know that if I’m wildly successful that stuff will happen. I’m not going to like it but that will happen. But as a writer there’s no way of thinking “if I’m wildly successful they will want long-lens photographs of me on the beach in my bikini. Never occurred to me in a million years.

    Winfrey: So you weren’t prepared for it.

    Rowling: Totally unprepared. And really running scared for a while.

  2. Implications for learners / teaching
    Again, based on my conversations with B2 learners, it is evident that they have skewed perception of this feature of speech: they felt that fluency markers are ‘rubbish words’/signs of inferior speakers; that adverbs like ‘absolutely’ and ‘exactly’ are too expressive; I personally also feel that we non-native speakers might be feeling the need to prove our linguistic ability by saying something new/driving the conversation forward. My point in that conversation was that if you don’t use those ‘small’ words, your speech does not sound ‘clean’ – it sounds non-native.

    An observations that I’ve made since I learnt about confluence is that I myself completely fail to notice these features, as well as expressions like ‘you know’ and ‘do you know what I mean’, unless I’m paying conscious attention. So reading the transcript of the interview with Rowling (or, more specifically, reading only the starts of turns) was incredibly revealing – even though I’d seen the interview, which suggests the need for explicit conscious raising activities.Another interesting fact is that film English is bad for exploring these features: at least in Friends, a well-known sitcom, these features are drastically underrepresented.

Personal conclusion Looking back up the changes my daily teaching underwent in light of the insights outlined above, I see that I’ve (1) started to use task repetition much more, making sure I inform the students of the rationale to persuade them to cooperate; (2) treat pronunciation of formulaic language much more consistently and (I hope) more persuasively, although prominence and appropriate pausing definitely deserve more attention; (3) see the need to increase learner autonomy in working with lexical resources, although I haven’t figured out how to do this in a systematic way (4) see the need to research typical ‘micro-frameworks’ to help learners with micro-planning. In the next couple of posts I’ll write about some other practical activities that target fluency.


Cauldwell, R. (2013) Phonology for Listening. Speech in Action

Clark, M (2007) IELTS Speaking

Levelt, W.J.M (1989) Speaking. Cambridge, MA: Bradford/MIT Press.

O’Keeffe, A., McCarthy, M. and Carter, R. (2007) From Corpus to Classroom: Language Use and Language Teaching. Cambridge University Press.

McCarthy, M. (2006) Explorations in Corpus Linguistics

Quaqlio, P. (2009) Television Dialogue: The sitcom Friends vs. natural conversation. Studies in Corpus Linguistics

Thornbury, S. (2000) Accuracy, fluency and complexity. English Teaching professional, 16 (available for download here).

Skehan, P (1996) ‘Second language acquisition research and task-based instruction’ in Willis, J and Willis, D (Eds) Challenge and change in language teaching  Heinemann

Underhill (2013) ‘The Inner Workbench: learning itself as a meaningful activity’ in Arnold, J and Murphey, T (Eds) Meaningful Action. Earl Stevick’s influence on language teaching. Cambridge University Press

Wray, A. (2005) Formulaic Language and the Lexicon. Cambridge University Press

More resources:
cognitive bases

McCarthy, M. (2009) Rethinking Spoken Fluency


  1. […] interested then take a look at this rather technical but very comprehensive explanation of what fluency in speaking really […]

  2. […] interested then take a look at this rather technical but very comprehensive explanation of what fluency in speaking really […]

  3. Tina Kern says:

    Dear Olya
    This is a great blog – thanks so much for sharing all your ideas and experience! Tina

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s