How often do you need to encounter a word before you learn it?

Posted: September 23, 2019 in research, vocabulary
Tags: , , , ,

There has been wide agreement for a long time that one of the most important ways of building the mental lexicon is by having extended exposure to language input through reading and listening. Some researchers (e.g. Krashen, 2008) have gone as far as to say that direct vocabulary instruction serves little purpose, as there is no interface between explicit and implicit knowledge. This remains, however, a minority position, with a majority of researchers agreeing with Barcroft (2015) that deliberate learning plays an important role, even if it is only ‘one step towards knowing the word’ (Nation, 2013: 46).

There is even more agreement when it comes to the differences between deliberate study and extended exposure to language input, in terms of the kinds of learning that takes place. Whilst basic knowledge of lexical items (the pairings of meaning and form) may be developed through deliberate learning (e.g. flash cards), it is suggested that ‘the more ‘contextualized’ aspects of vocabulary (e.g. collocation) cannot be easily taught explicitly and are best learned implicitly through extensive exposure to the use of words in context’ (Schmitt, 2008: 333). In other words, deliberate study may develop lexical breadth, but, for lexical depth, reading and listening are the way to go.

This raises the question of how many times a learner would need to encounter a word (in reading or listening) in order to learn its meaning. Learners may well be developing other aspects of word knowledge at the same time, of course, but a precondition for this is probably that the form-meaning relationship is sorted out. Laufer and Nation (2012: 167) report that ‘researchers seem to agree that with ten exposures, there is some chance of recognizing the meaning of a new word later on’. I’ve always found this figure interesting, but strangely unsatisfactory, unsure of what, precisely, it was actually telling me. Now, with the recent publication of a meta-analysis looking at the effects of repetition on incidental vocabulary learning (Uchihara, Webb & Yanagisawa, 2019), things are becoming a little clearer.

First of all, the number ten is a ballpark figure, rather than a scientifically proven statistic. In their literature review, Uchihara et al. report that ‘the number of encounters necessary to learn words rang[es] from 6, 10, 12, to more than 20 times. That is to say, ‘the number of encounters necessary for learning of vocabulary to occur during meaning-focussed input remains unclear’. If you ask a question to which there is a great variety of answers, there is a strong probability that there is something wrong with the question. That, it would appear, is the case here.

Unsurprisingly, there is, at least, a correlation between repeated encounters of a word and learning, described by Uchihara et al as statistically significant (with a medium effect size). More interesting are the findings about the variables in the studies that were looked at. These included ‘learner variables’ (age and the current size of the learner’s lexicon), ‘treatment variables’ (the amount of spacing between the encounters, listening versus reading, the presence or absence of visual aids, the degree to which learners ‘engage’ with the words they encounter) and ‘methodological variables’ in the design of the research (the kinds of words that are being looked at, word characteristics, the use of non-words, the test format and whether or not learners were told that they were going to be tested).

Here is a selection of the findings:

  • Older learners tend to benefit more from repeated encounters than younger learners.
  • Learners with a smaller vocabulary size tend to benefit more from repeated encounters with L2 words, but this correlation was not statistically significant. ‘Beyond a certain point in vocabulary growth, learners may be able to acquire L2 words in fewer encounters and need not receive as many encounters as learners with smaller vocabulary size’.
  • Learners made greater gains when the repeated exposure took place under massed conditions (e.g. on the same day), rather than under ‘spaced conditions’ (spread out over a longer period of time).
  • Repeated exposure during reading and, to a slightly lesser extent, listening resulted in more gains than reading while listening and viewing.
  • ‘Learners presented with visual information during meaning-focused tasks benefited less from repeated encounters than those who had no access to the information’. This does not mean that visual support is counter-productive: only that the positive effect of repeated encounters is not enhanced by visual support.
  • ‘A significantly larger effect was found for treatments involving no engagement compared to treatment involving engagement’. Again, this does not mean that ‘no engagement’ is better than ‘engagement’: only that the positive effect of repeated encounters is not enhanced by ‘engagement’.
  • ‘The frequency-learning correlation does not seem to increase beyond a range of around 20 encounters with a word’.
  • Experiments using non-words may exaggerate the effect of frequent encounters (i.e. in the real world, with real words, the learning potential of repeated encounters may be less than indicated by some research).
  • Forewarning learners of an upcoming comprehension test had a positive impact on gains in vocabulary learning. Again, this does not mean that teachers should systematically test their students’ comprehension of what they have read.

For me, the most interesting finding was that ‘about 11% of the variance in word learning through meaning-focused input was explained by frequency of encounters’. This means, quite simply, that a wide range of other factors, beyond repeated encounters, will determine the likelihood of learners acquiring vocabulary items from extensive reading and listening. The frequency of word encounters is just one factor among many.

I’m still not sure what the takeaways from this meta-analysis should be, besides the fact that it’s all rather complex. The research does not, in any way, undermine the importance of massive exposure to meaning-focussed input in learning a language. But I will be much more circumspect in my teacher training work about making specific claims concerning the number of times that words need to be encountered before they are ‘learnt’. And I will be even more sceptical about claims for the effectiveness of certain online language learning programs which use algorithms to ensure that words reappear a certain number of times in written, audio and video texts that are presented to learners.

References

Barcroft, J. 2015. Lexical Input Processing and Vocabulary Learning. Amsterdam: John Benjamins

Laufer, B. & Nation, I.S.P. 2012. Vocabulary. In Gass, S.M. & Mackey, A. (Eds.) The Routledge Handbook of Second Language Acquisition (pp.163 – 176). Abingdon, Oxon.: Routledge

Nation, I.S.P. 2013. Learning Vocabulary in Another Language 2nd edition. Cambridge: Cambridge University Press

Krashen, S. 2008. The comprehension hypothesis extended. In T. Piske & M. Young-Scholten (Eds.), Input Matters in SLA (pp.81 – 94). Bristol, UK: Multilingual Matters

Schmitt, N. 2008. Review article: instructed second language vocabulary learning. Language Teaching Research 12 (3): 329 – 363

Uchihara, T., Webb, S. & Yanagisawa, A. 2019. The Effects of Repetition on Incidental Vocabulary Learning: A Meta-Analysis of Correlational Studies. Language Learning, 69 (3): 559 – 599) Available online: https://www.researchgate.net/publication/330774796_The_Effects_of_Repetition_on_Incidental_Vocabulary_Learning_A_Meta-Analysis_of_Correlational_Studies

Comments
  1. Otto Weihs says:

    let alone the different capacity of each and every learner

  2. alexcase says:

    Very interesting. Some surprising points such as “Repeated exposure during reading and, to a slightly lesser extent, listening resulted in more gains than reading while listening and viewing”, which would suggest better learning of vocabulary watching a video without English subtitles.

    • philipjkerr says:

      Hi Alex, I wouldn’t read it in that way! The meta-analysis is only looking at frequency effects in repeated exposure, not at modes of input in a more general sense. Since video with subtitles provides more linguistic information than video without, it is presumably the case that fewer exposures are needed for learning. But if repeated exposures are provided anyway, then video without subtitles leads to a more powerful frequency effect. I think the narrow focus of the research is obscuring the bigger picture. At least, that’s my reading.
      Philip

  3. N says:

    Hm, I’m rather curious about the bit mentioning repeated exposure under massed conditions… That would mean that tasking the learners with “noticing” activities like looking up the examples of the target vocabs in texts/videos is a very sensible thing to do… wouldn’t it?

    The point about non-words is interesting as well: even with regular vocabulary, ‘fancy-looking’ words like ‘ostentatious’ or ‘gazillion’ tend to attract more of the learners’ attention as they stand out from the crowd of less impressive-looking vocabulary, with which more attention is paid to the meaning, while the vocab itself may fade into the background…

    • philipjkerr says:

      I’m not sure that that follows, because (as I understand it) the comment about massed conditions is independent of considerations of ‘engagement’. I suppose that the massed conditions could simply mean something like a whole pile of narrow reading crammed into a short space of time (which, I guess, is what typically happens with narrow reading).

      • N says:

        Hm… Fair enough. Although I’m not completely sure how or if massed conditions and engagement were segregated. The original paper states that the only criterion for the conditions to be considered ‘massed’ was for the treatment to take no longer than one day (interestingly enough, irrespective of the number of texts or exposures) – but how does one make a comment on it independent of other factors (say, engagement) if, in the original study, they were combined? Or weren’t they?
        ಠ , ಥ

        However, my initial example of a ‘noticing’ activity is definitely not about incidental vocabulary learning, so I do stand corrected there.

  4. philipjkerr says:

    To be honest, the more I look at the paper, the less satisfactory I find it. Your point about separating out the various factors is the key one, I think. Even if, in experimental conditions, it were possible to do so, it wouldn’t really tell us much about the messy real world. Meaning-focussed reading / listening is such an idealised construct when it comes to a language learner. I really don’t understand how we can even make claims about what is meaning-focussed and code-focussed reading, as they are typically points on a continuum, rather than separate entities.

Leave a Reply to philipjkerr Cancel reply

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

WordPress.com Logo

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

Google photo

You are commenting using your Google 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