Sunday, 22 July 2012


Imagine an alternative universe in which a large number of teachers, experts and textbooks promoted a model of teaching which was untrue.  Imagine that despite ample evidence of the flaws of this approach,  in the form of books and journal articles, teachers and publishers just carried on teaching it, -all the time taking students money for this "education".   Unfortunately, many EFL teachers live in that world where reading skills are concerned.  I'm hoping to write more about reading skills, but this entry will only deal with prediction.

Prediction is a popular EAP activity.  One of the key principles of reading listed by Harmer is prediction.  He notes that things like the title can help students to form opinions relating to the work before they begin reading (2007).  Grellet (1990: 56)  adds that, “reading is an activity involving constant guesses that are later rejected or confirmed”.  The British council note that " Prediction is a valuable stage in...reading activities. It mirrors L1 skills use, where predictions form an important base for being able to process language in real time."(online)  Prediction is an idea that comes from one model of reading, "this model of how people read is called the "psycholinguist guessing game model"(Grabe 2009:102). 

Grabe makes two important points about this model of reading.  The first is that it is very popular among "applied linguists" and the second is that "it has been proven wrong in its predictions by accumulating evidence for the past 20 years" (Grabe 2009:102).   Grabe hammers the point firmly home noting that this approach "has no empirical validity and is problematic"(Grabe 2009:103)  adding “One needs only to pick up a newspaper in an unknown language to verify that background knowledge and prediction are severely constrained by the need to know vocabulary and structure.” (1991: 380)  

 20 years of being wrong and we still use it?  Still teachers may not be familiar with articles in obscure second language reading journals and Grabes book only came out in 2009.    If only an article had appeared in something more accessible, like the ELTJ a little earlier, say around 1996.......if only!

In 1996, Amos Paran in an article called "reading in EFL facts and fictions" published in the ELTJ bemoans the use of the "psycholinguistic model" of Reading in EAP courses, noting that it "was never accepted as an important model in the first place "(1996:29) and adds:

                As a final point, it is important to stand back and think how the Goodman and Smith view of reading, with all the reservations LI reading researchers expressed towards it, has been able to hold sway over L2 reading models for such a long time. (1995:33)

A point made perhaps more worrying by fact the paper was first presented at IATEFL in 1992.  Despite this, I can still pick up textbooks, such as Oxford's "well read" which include, in Swan's words, "the standard battery of exercises designed to train students in ‘skimming’, ‘scanning’, ‘predicting’, ‘inferring’ and so forth, that one finds in textbook after textbook" (2008:266)

Swan, beating Paran by 8 years, noted another problem with this model, namely the assumption that non-native speakers lack the ability to predict.  With his usual finesse for cutting through bullshit he writes:

               One of the comprehension skills which we now teach foreigners is that of predicting. It has been observed that native listeners/readers make all sorts of predictions about the nature of what they are about to hear or read, based on their knowledge of the subject, their familiarity with the speaker or writer, and other relevant features. Armed with this linguistic insight (and reluctant to believe that foreigners, too, can predict), we 'train' students in 'predictive strategies'. (For instance, we ask them to guess what is coming next and then let them see if they were right or wrong.) But I would suggest that if a foreigner knows something about the subject matter, and something about the speaker or writer, and if he knows enough of the language, then the foreigner is just as likely as the native speaker to predict what will be said. And if he predicts badly in a real-life comprehension task (classroom tasks are different), it can only be for one of two reasons. Either he lacks essential background knowledge (of the subject matter or the interactional context), or his command of the language is not good enough. In the one case he needs information, in the other he needs language lessons. In neither case does it make sense to talk about having to teach some kind of 'strategy'. (Swan 1985: 8)
Maybe it's time to stop wasting our students' time?

British Council (2012) Prediction.  In teachingEnglish. Retrieved July 6 2012, from

Grabe, R (2009) reading in a second language Cambridge: Cambridge University Press

Paran, A. (1996). Reading in EFL: Facts and fictions. ELT Journal, 50(1), 25-34
Grabe, W. (1991) Current Developments in Second Language Reading Research TESOL Quarterly, Vol. 25, no 3, 375-406

Grellet, F. (1990) Developing reading skills Cambridge: Cambridge University Press

Harmer, J. (2007a) How to teach English Essex: Pearson Education Limited

Swan, M. (2008) Talking Sense about Learning Strategies RELC Journal 2008 39: 262 261-273

Swan, M (1985) a critical look at the communicative approach 1 (1), ELT Journal 39/1, pp.2-12


  1. Thanks for this post - I'm looking forward to more posts about other reading 'skills/strategies', particularly to more skewering of our obsession with 'skimming and scanning' :-)

  2. Thanks for the comment! You're the first (^_^)

    The article originally included skimming and scanning but it started to get a bit big and unwieldy. There are two more reading articles in the pipe line. One about "skimming scaanning and gist" and the other about "guessing words from context".

  3. An interesting read and food for thought.

    Although perhaps in defence of 'predicting' as used in textbooks, isn't it just another classroom strategy (rather than intentionally a real-world strategy) to get around the fact that reading in a language classroom is a bit of an artificial task anyway? It's just another type of pre-reading 'warmer' to help students get ready for reading a text that's probably in a subject area they know little about and have no genuine motivation to read (however well we try to choose our material, you can't get it spot-on every time). In a real-world context, you'd probably come to a reading text with your background knowledge of the subject already 'activated' - in an ELT classroom we need a way of getting students to think about the topic and activate their background knowledge a bit so they're not coming to it completely cold. A predicting activity is just one technique for doing that - you can't predict without thinking about what you already know about something (thus activating both 'content' knowledge and vocabulary). It's also a way of providing a motivation for reading - to find out if your predictions were correct.

    So yes, it is artificial and shouldn't be emphasized as an important learning strategy, but I think it has a role as one of the many ways we use to get over the inherent artificiality of classroom reading.

    Oh, that was rather a long, serious comment for this time of the morning ... surprised myself there, you must've really piqued my interest :)

    1. I think you make a good point and thank you for commenting on the article.

      What worries me is why people are doing it in the first place. Take homeopathy for example, it might make people feel better (through placebo 'mind-over-matter' type effects) but the useful side-effects don't make it any more valid.

      I think you're talking about schemata theory, -which is a valid theory- but I ofetn wonder if the amount of time we spend trying to activate it is really worth it. Students must have schemas for concepts already, so are 15 minutes of brainstorming or predicing really worth it? I don't know.

  4. Interesting article. I certainly think we need to be questioning the sometimes rather constraining models of how to 'do' a reading. It can be very formulaic.
    That said, I agree with Julie that prediction is mostly a way of raising interest and giving a reason to read- something which presumably people reading outside the classroom would already have.
    I'm not sure that coursebooks necessarily set out to train learners in predicting, skimming, scanning and so on. I think these task types persist partly because teachers want to be able to ask students to do something specific and measurable while they're reading, and partly as a way of guiding students through the text. Good comprehension questions should aid comprehension rather than testing it.

  5. "I think these task types persist partly because teachers want to be able to ask students to do something specific and measurable while they're reading"

    I think you're absolutely right here! I'mm planning the next article to deal with this point...Things which help teachers Vs things which help students.

  6. That sounds like it will be a good read :)

  7. Your Twitter (and thus blog) were just recommended to me. I like what you're doing here. Great, well-supported posts. This is not a topic that I've thought deeply about and it's good to have post that gets me to do so.

    You paint with a pretty broad brush here regarding prediction strategies. Even Grabe acknowledges benefits in a number of places in his book in regards to strategies for reading comprehension.

    His scorn seems to be directed more towards Goodman's model of reading (and really the whole Whole language movement) and the use of prediction as a stand-alone activity for reading fluency and vocabulary learning. However, as part of integrated strategy instruction, he seems to deem it acceptable (pp. 103 [pre-reading], 64-65 [patterns/L1 interference], 99 [as a component of skimming], and 223 [monitoring comprehension]).

    Of these, though, only patterns/L1 interference are really things that we need to be teaching. I'd guess both he and Swan would likely agree that the others are simply general reading strategies that can likely carry over from the L1 (assuming they are literate [loaded term] in the L1) and have little use for explicit instruction in the L2 classroom.

    With this is mind, do you see any role for prediction strategies in reading instruction?

  8. ps, you convinced me to get Grabe's book :-)

  9. Thanks for your comments Dan, I found them very thought-provoking.

    Yes I think you're right about Swan and Grabe. I can't help thinking that many of these stategies are more useful for teachers than me a cynic. I'm trying to write something about this at the mo. but it's slow going.

    I'm not sure if there is any role for prediction other than to make a nice DELTA lesson plan. I also think any evidence presented on the subject will not stop teachers from using it. "Men do their broken weapons rather use / Than their bare hands"

    I think language learning is a bit msyterious and if a teacher can pull of a good class using this kind of thing, then perhaps it's ok. I personally always manage to screw these lessons up.

    Me: Ok so the text is called "Orphan Timmy's x-mas wish" what do you think it will be about?

    Student A: Timmy wishes for a new car!

    me: hmmmmmmm, right...interesting...

    I'm also not sure how I feel about the kind of "secondary use" argument that "if it does no harm" then who cares? It reminds me of arguments about homoeopathy whereby people say..."well yes it doesn't work but it doesn't hurt anyone and maybe it has a placebo effect which could be useful."

    Perhaps I'm being too hard-headed but it seems like people are buying those drugs because they believe they work. If they don't work, and that is a fact, then shouldn't we stop using them? Also as an EAP teacher I'm a little bit worried by the lack of critical thinking going on. EAP teachers are constantly usrging their students to "think critically" but how many of the practices they subscribe to have they examined critically?

    Sorry, this seems a bit rambling...

    1. Very interesting. I'm late to this, but thanks for posting on this topic. I have one question which is that a major reason for not teaching predicting is that it comes naturally to us, whether we are reading in L1 or L2. However, I have taught many students who certainly do not read habitually in L1, have had very little schooling in L1. I don't know if they would still be able to innately predict or if it is a strategy which they don't possess and may get value from being made aware of.