Elgazette Logo newtrans  The magazine for English language teaching and English medium education

Comment: Why ELT fell for the lexical sets myth – hook, line and sinker


Evidence shows learning semantically related vocabulary in groups isn’t the best way to learn lexis, but the idea is entrenched in language classrooms, writes Melanie Butler

Iam now officially ancient. I must be. I was around at the birth of at least one ELT myth and I expect it will outlive me. If there is one thing I have learned in forty years in ELT, it is that a good myth is hard to kill.

The myth in question says that vocabulary is best learned in lexical sets. So, words such as knife, fork, spoon and plate are best learnt at the same time, as a group. Ditto shoes, hat, coat, umbrella.

Doesn’t sound like a myth, right? Sounds like common sense.

But it was declared a myth by Keith Folse in his excellent book Vocabulary Myths, Applying Second Language Research to Classroom Teaching (2004). Folse did his best to bury it in a mountain of evidence against it. In the UK, however, it remains alive and well – and littered all over our coursebooks, as I was reminded by these words about EFL materials I saw on Twitter: ‘Vocabulary should be taught in lexical sets (logical).’

Wearily, I tweeted back: ‘Vocabulary should NOT be taught in lexical sets. It makes it harder to learn. The evidence has shown that for 20 years…’ The lexical set hypothesis started life well. It was not borrowed from pop psychology, nor thought up by a bunch of teacher-trainers looking for something new and sexy to present at Iatefl. It was a scientific hypothesis, put forward by the eminent psycholinguist Professor Jean Aitchison.

As part of her work on adults with dysphasia, a speech deficiency that often results in the wrong choice of word, Professor Aitchison noticed that the errors produced followed a consistent pattern: they all came from the same lexical set. She posited, probably correctly, that that words were stored in the brain in such sets. She then hypothesised that the best way for L2 learners to learn vocabulary was in lexical sets.

I don’t remember the first time I heard the lexical-set language-learning hypothesis, but I have known Jean for nearly forty years. I have met her often in professional and social settings and love her work.

I do remember, however, my reply when she first made the argument to me that we should learn L2 words in sets because we store L1 words in sets. I said, ‘I store all different kinds of pasta in the same part of my kitchen cupboard, but that doesn’t mean I acquired it all at once.’

Jean was convinced she was right, and when it came to vocabulary research she nearly always is.

So, in 1987 on the publication of her seminal work Words in the Mind, Jean began attending ELT conferences and arguing that vocabulary should be taught in lexical sets to adult L2 learners.

This was only one small part of what Jean had to say about the mental lexicon. She argued, for example, that L1 children acquired vocabulary in an entirely different way. They acquired a word in three stages. First comes labelling, where the child makes a link between the sounds and the object.

Second comes packaging, where she learns the range of meanings of the word and its limits.

Finally comes network building, which involves understanding the synonyms, antonyms and collocations of the word, fitting it into its correct lexical set.

The British ELT community ignored Aitchison’s three-stage model of acquisition. Indeed, they pretty much ignored everything she had to say about lexis – except for her lexical-set hypothesis.

They latched on to that like puppies to a bone and have hung on to it doggedly ever since. But evidence against the hypothesis began to build.

In 1990 Paul Nation, a New Zealand professor of applied linguistics, published a seminal paper showing that giving students lexical sets made it harder to learn and retrieve individual words. Nobody in the UK took the blindest bit of notice.

British academics like Meara, Milton and Schmidt began churning out papers on the subject. Nobody cared.

Neurocognitive research from Bangor found that subjects who learned words in lexical sets had more difficulty retrieving the correct item.

Nothing happened.

The lexical set had attained mythical status. Is this, as the tweet suggested, because its logic is unarguable? Let’s unpick Aitchison’s original argument.

Let us take as given that adult native speakers with language processing deficiencies have difficulty retrieving the correct word from a lexical set. They muddle them up. It may therefore follow that words are stored in lexical sets.

But does it follow that L2 learners – and only L2 learners – will learn and retrieve an individual word correctly if they learn it along with lots of other similar words at once, as part of a lexical set? Or will they get muddled too?

I don’t know about you, but I learned the names of the bones in the human body by heart in a lexical set.

I needed them to pass a school biology test. I remember that the ulna, the tibia and the ilium are all bones. But I’m dammed if I know which is which.

I think that Jean Aitchison’s description of the three stages of vocabulary acquisition may well be right, including for L2 learners.

But I believe her lexical set hypothesis has turned into a myth. It’s time it was laid to rest.