Part 7: The next 4 to 5 years of adaptive learning in ELT – 10 predictions

Posted: February 3, 2014 in A guide to adaptive learning
Tags: , , , , , , , , , , ,

Given what we know, it is possible to make some predictions about what the next generation of adult ELT materials will be like when they emerge a few years from now. Making predictions is always a hazardous game, but there are a number of reasonable certainties that can be identified, based on the statements and claims of the major publishers and software providers.

1 Major publishers will move gradually away from traditional coursebooks (whether in print or ebook format) towards the delivery of learning content on learning platforms. At its most limited, this will be in the form of workbook-style material with an adaptive element. At its most developed, this will be in the form of courses that can be delivered entirely without traditional coursebooks. These will allow teachers or institutions to decide the extent to which they wish to blend online and face-to-face instruction.

2 The adaptive elements of these courses will focus primarily or exclusively on discrete item grammar, vocabulary, functional language and phonology, since these lend themselves most readily to the software. These courses will be targeted mainly at lower level (B1 and below) learners.

3 The methodological approach of these courses will be significantly influenced by the expectations of the markets where they are predicted to be most popular and most profitable: South and Central America, the Arabian Gulf and Asia.

4 These courses will permit multiple modifications to suit local requirements. They will also allow additional content to be uploaded.

5 Assessment will play an important role in the design of all these courses. Things like discrete item grammar, vocabulary, functional language and phonology, which lend themselves most readily to assessment, will be prioritized over language skills, which are harder to assess.

6 The discrete items of language that are presented will be tagged to level descriptors, using scales like the Common European Framework or English Profile.

7 Language skills work will be included, but only in the more sophisticated (and better-funded) projects will these components be closely tied to the adaptive software.

8 Because of technological differences between different parts of the world, adaptive courses will co-exist with closely related, more traditional print (or ebook) courses.

9 Training for teachers (especially concerning blended learning) will become an increasingly important part of the package sold by the major publishers.

10 These courses will be more than ever driven by the publishers’ perceptions of what the market wants. There will be a concomitant decrease in the extent to which individual authors, or author teams, influence the material.


  1. “Major publishers will move gradually away from traditional coursebooks…”

    This is not necessarily a trend we should lose a lot of sleep over, as traditional coursebooks have long since aligned themselves, willy-nilly, to a grammar-macnuggets production line model of language learning: look at the syllabus of any currently popular coursebook series (and compare it to the syllabus of Headway Intermediate 1986). What the snazzy new adaptive learning platforms do is simply add big data analytics to the existing model, although, of course, in doing so, they further disempower the classroom teacher.

    There are other alternatives to traditional coursebooks that don’t have to go down that route, content-based instruction being one. I’ll stop now.

    • Brendan Wightman says:

      Things are indeed getting bad. I now see that the ideals of ‘informal learning’ are being interpreted and expressed in glitzy promotional videos of cute, middle-class kids accessing platforms with their tablets whilst eating their Rice Crispies at breakfast. Even sadder (or more hilarious, depending on your view) is how the teacher is portrayed – ridiculously passive, staring in quiet, metaphysical rapture at data presented on an Interactive Whiteboard.

  2. The language learning experience will continue to move towards an educational philosophy that affords language learners time to engage with others. This will occur through social and connectivist (complex) theories, appropriate asynchronous & synchronous forms of communication, and some form of online & offline means of information delivery. Publishers will be ahead of the game if they find innovative ways to take these three concepts and find ways to assist educators in creating language learning environments that are effective, efficient, and engaging for the learner. Publishers will not continue to make the same mistake in the past in isolating espoused theories, forms of communication (asynchronous & synchronous), and information delivery methods which are inappropriate to local contexts because educational stakeholders see immediately when publishers are able to add value and when they are simply selling something.

    • philipjkerr says:

      I wish I could share your optimism, Benjamin. The main reason why I don’t is that there is now a significant merging of the most significant (in terms of power) educational stakeholders and the publishers themselves. See, for example, some of the posts about Pearson on Diane Ravitch’s blog (, or have a look at Joel Spring’s work (e.g. ‘Education Networks). Educational politics and practices are rarely driven primarily by educational understandings. I’ll get on to this topic in a later post.

      • Brendan Wightman says:

        Philip, the one point I would make is the sheer number of parties that are complicit in these recent developments. The pressure to ‘modernise’ has been top-down for many years, but increasingly, I’ve noted how the pressures have become horizontal, too: the media, parents obsessed with assessment scores, so-called educational technology ‘gurus’ who keep telling us the future is all digital. The list could go on. Interestingly, the people missing from the debate seem to be teachers, and at no point has anyone asked what the school of the future should be like – its role in society, its purpose, how it should work for the benefit of all. Educational philosophy, as far as I can tell, has lost out to the ‘big rhetoric’ that accompanies big data.

      • Educational philosophy lost out to technology long ago, Brendan, at least if we are to believe the neo-Luddites like Neil Postman. In his day, i.e. the 1980s, it was ‘educational television’ – whose programs, he argued, were ‘conceived by someone’s asking the question, “What is television good for?”, not “What is education good for?” (1985: 157). The same thinking, i.e. what is technology good for? would seem to undergird the current mania to dream up ‘100 classroom uses of Twitter’ (Or Glogster, or Flickr, or Wordle, etc).

        As for the teachers, you are right in saying that they have been written out of the plot. Their resistance is branded as Luddism, and their allegiance to educational values dismissed as either reactionary or romantic. As Diane Laurillard wrote, ‘Technological innovation is driven by many factors, but not one of them concerns a pedagogical imperative. There is no dialogue between teachers and technologists about what kind of technological innovation learners need: neither side knows how to begin the dialogue’ (cited in Brett, D. & González-LLoret, M. 2011. ‘Technology-enhanced materials’ in Long, M. and Doughty, C. (eds.) The Handbook of Language Teaching, Oxford: Blackwell).

  3. I am intrigued by the prediction “The discrete items of language that are presented will be tagged to level descriptors, using scales like the Common European Framework or English Profile”. In fact, if the software is fully automated (i.e. independent of human raters), will it really be able to rate CEF descriptors such as ‘Can produce simple connected text on topics which are familiar or of personal interest’ or ‘Can communicate in simple and routine tasks requiring a simple and direct exchange of information on familiar and routine matters’ (And these are both lower band descriptors).

    The Pearson Versant test (done over the phone and scored automatically) manages to test sentence mastery, vocabulary, fluency and pronunciation, through a variety of tasks including reading aloud, repeating sentences, retrieving words from memory, reproducing a short narrative, and answering some personal questions spontaneously. This is achieved through specific (secret) algorithms, fluency being a measure (probably) of speech rate, unfilled pausing, etc, while pronunciation is assessed using voice recognition software. (I have done this test in Spanish and, I have to say, it is uncannily good at measuring what it aims to measure). But the algorithms that can evaluate spontaneous multi-party interaction or text coherence or ‘comfortable’ intelligibility are still mere glints in programmers eyes (aren’t they?). Hence, I don’t see CEF descriptors yielding easily to adaptive software.

    The English Profile project, however, does seem to offer programmers a way in, since it aims to flesh out the CEF descriptors with specific linguistic items (vocab, structures etc). An algorithm that is primed simply to detect the occurrence of these discrete items might, conceivably albeit clumsily, be able to rate a speaker’s ability to ‘communicate in simple and routine tasks requiring a simple and direct exchange of information on familiar and routine matters’.

    Will one of the unintended consequences of the EP be its cack-handed use for automated testing on adaptive learning platforms?

  4. Jill Hadfield says:

    Thanks for this very informative – albeit scary- blog Philip!

  5. antoniaclare says:

    Watch out for the new Pearson Global Scale of English too, which matches both CEFR and English Profile, mapping out skills competences and relevant language items per level (on a scale of 10-90) Their new testing system, adaptive of course, claims to offer highly accurate measurements of progress and level based on tests similar to the one Scott describes. I haven’t seen exactly how it works yet, but my concerns would be similar to Scott’s. Can algorithms really measure communicative competence in any realistic manner? And if, as they claim, the results are so reliable, where does this lead us?

  6. One thing going through my mind as I read through these posts, is how adaptive learning fits in with teacher training, and how modern courses will also need to adapt to cater for the emerging needs of teachers.

    Interestingly, specific, extensive learning outcomes are currently the backbone of assessment on courses like Celta and Delta (and the Trinity equivalents). Tutors, trainees and prospective employers are presented with these at different stages of the course (and they’re referred to on end of course reports), in much the same way as many publishers seem to be moving more towards in language learning. They’re very useful in showing the trainee where they need to improve as well as showing the tutor what they need to give more support on. But, as with level descriptors for language learners, some parts are open to interpretation, and this is where tutors and trainees can come together to negotiate and discuss what their priorities are (in much the same way as the CEFR was intended to be used). I’m not convinced this part can be done electronically.

  7. PHILIP says:

    Hi Scott, I think a reference to Hegel and the labour of the dialetic might work. Let students who want to get this direct delivered course go in that direction. One can hardly blame publishers going this way with such widespread piracy. While I remain positive that other teachers might start developing their own courses, which is incredibly easy to do, when you see all the tools available on the internet.[counter dialectic].

    Personally I would like to mention I’ve just finished teaching for the evening and I’m so pissed offf that I have to work so hard to make a course work book interesting for students. I really don’t know how so many teachers can stand in front of classes and just use a text book. Ok I know they bring in additional stuff but couldn’t they take that next step and write their own course in dialogue with the students?

  8. Paul Ricketts says:

    This blog is superb, Philip, an excellent overview of AL issues. It gets a doubleplusgood Like from me 😉

    I’ll presume to add a few more predictions:

    1. The role of quality adaptive systems in ELT will become much clearly perceived as appropriate for quantifiable outcomes such as vocabulary and grammar – linguistic competence – than as an existential threat to teachers. Competent teachers will use this opportunity to focus classroom time on less quantifiable but more important outcomes of communicative competence.

    2. Publishers able to offer integrated, blended packages comprising adaptive services and skills will prosper (see Philip’s number 7 – but sophistication and money without vision won’t be enough).

    3. Learners will experience and perceive real value in well-crafted adaptive systems (actually, this isn’t even a prediction, the popularity of Duolingo, Memrise, etc already attest to this, despite their limitations)

    4. Learners will begin to understand the value the personal learning data they generate, and there will be a movement towards open standards, making it possible for learners to own and control their own data.

    5. Publishers will remain under increasing pressure to deliver more innovative services more cheaply. Nobody will have an easy ride, except, hopefully, learners.

    6. The role of the author will be to provide creative, high-quality, versatile content in smaller chunks than traditional textbooks. These will only be MacNuggets if good authors throw in the towel and reject AL as intrinsically evil (OK, maybe there’s some devil’s advocacy in this last one, but I’d love to hear Philip’s view here).

  9. […] Given what we know, it is possible to make some predictions about what the next generation of adult ELT materials will be like when they emerge a few years from now. Making predictions is always a …  […]

  10. kimsenglish says:

    “Major publishers will move gradually away from traditional coursebooks…”

    Shouldn’t the adverb precede the verb – or is there a move towards an adaptive grammar, i.e. allowing NNS errors to be absorbed into the language?
    I was part of the recent ‘iPad implementation project’ in the Middle East which had about two years of ‘growth’ followed by an immediate decline in innovative drive when the directors behind the project were asked to move on…
    In that brief time, publishers were asked to produce ‘interactive versions’ of their course books that could be viewed on this device. It was a huge effort for them to provide scanned PDF files as this was the best solution that could be offered immediately, with the quid pro quo that our teachers would help provide material and ideas to support them in developing material further down the line. Only one publisher was bold enough to engage in this development on a weekly basis but feedback from teachers indicated that the publishing staff had very little knowledge about producing interactive content (!) and no experience about providing an engaging classroom experience using the iPad (!!). During a formal debate between the teaching staff and publishers it was established that if the publishers were to work towards providing material for this device (or other tablet-style devices), it would take at least 5 years before a product would be available on the market. In my view, publishers are quite happy to maintain the status quo and are reluctant to address the teacher’s plight of having to constantly make their teaching material relevant to their learners’ needs. The versatility of devices like the iPad are unquestioned – it is how they are used to produce an adaptive pedagogy that requires research. Many teachers still start from the wrong end of the process by finding the App first and shoehorning it into a lesson. Not surprising then, to find someone or other’s taxonomy shoehorned into a table of ‘suitable apps’ to use in order to give credence to the technology.

    Until the revenue stops flowing, you can be sure that ‘no change’ will be the order of the day. As the American South are fond of saying, “if it ‘aint broke, don’t fix it.” Well, I’m broke, but I’m only a teacher. (Let’s see if anyone spots the obvious mistake.)

    Written and edited painstakingly on my iPad

  11. kimsenglish says:

    And thank you Philip for producing a fascinating blog …tried to add this to my last post but the technology ( and LinkedIn Ads) got in the way ;0)

  12. Sandy Millin says:

    Just saw this post 5 years and a few months on…I wonder how many of these have come true?

    • philipjkerr says:

      I really ought to take a proper look at these predictions some time soon. But, at a quick glance, I think that these predictions were not to far off the mark.

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 )

Connecting to %s