Michael Hoey, faithful observer of language

Michael Hoey’s was a compelling argument in favour of teaching language lexically, because that’s how language is, and that’s also how it’s learnt and mentally stored. On the linguistics angle, he endorsed the work of Michael Lewis and his ´lexical approach´, and on the pedagogic side, he recommended the ´Monitor Model´of Stephen Krashen.

A key concept in Hoey´s talk was the psychological idea of ´priming´, and in particular, lexical priming. The basic idea is that the brain stores words in association with one another, and definitely not as optional slot fillers in a grammatical pattern, as generative grammarians would have us envisage it. Priming experiments are ingeniously contrived to demonstrate things like this: you will process the item ´swan´ faster if you were exposed to the item ´wing´ some time earlier (semantic priming), or if you see ´scarlet onion´ today, then tomorrow ´scarlet´ will prime you for ‘onion´ (repetition priming).  When we encounter words, we subconsciously notice all the neighbouring words, and this collocational awareness speeds up our comprehension. Density of unrecognised collocations is what makes a text difficult to process.

On the linguistic side, Hoey took us through some findings from his corpus investigations. He showed that in any piece of text there is a dense web of connections among the words and phrase patterns. We become familiar with these connections subconsciously, so much so that if you hear a word, we can predict how likely we are to hear it again. For instance, if ´president´ occurs once in a text, it’s very likely to occur again, whereas if you hear ´frankly´, it isn’t.  I found the following micro-nugget especially illuminating: an affirmative is nine times more common than a negative. So a negative is the marked case.  So if a verb occurs 50% of the time in a negative context, this is marked, and hearing that verb will therefore prime the listener for negativity.

It is very welcome once in a while to sit in a plenary which has not only a more general recommendation (teach English more lexically), but also such a wealth of concrete and specific detail about the very stuff of what we teach – the language itself. But this was also a very entertaining experience, as Hoey marched back and forth gesticulating with energy, and constantly cracking jokes, usually at his own expense, and often in reference to his precarious health. For instance, ‘My doctor told me I had a myocardial infarction. I said, ‘Oh thank goodness, I was worried it was a heart attack!’’.

Geoff Jordan provides a good summary of the talk here, with quote marks to indicate which bits he considers contentious. However, I found Jordan's assertion "The Monitor Model and the Lexical Approach are not true" odd. Models and approaches aren't true or false, but rather useful or not useful. Hoey is pointing out that native speaker language use is much denser in lexical connections than is generally appreciated, and that looks like a useful insight. How far we let this insight determine our ELT practices is another question, of course.