First, Mr Hynds admits that most readers would not be likely to have the kind of knowledge he is testing “unless the whole population were to be given a specialist course in phonetics”. So his test is really a test of phonetics, which is not quite the same as phonics. It is illogical to give a phonetics test and use the results to prove that people are weak at phonics.

Second, Mr Hynds says that people learn to read “by employing a whole range of strategies and sources of information”. But his test, by asking people to read words out of context, excludes three of the four sources of information (“syntactic”, “semantic” and “bibliographic”) which he mentions in his published work. What does that leave? Oh dear - only “graphophonic” information, according to his list. So his subjects must have used this in order to read the words, either decoding letter by letter or recognising the words as wholes.

My Hynds seems to have come perilously close, then, to proving that good readers do know their phonics rather than that they do not. Here’s a chance for them to clinch the argument. Can they produce plausible pronunciations for the pseudo words “reshif”, “gwins” and “phaims” (made up out of the letters of the first three Hynds test words)?

If so their phonic knowledge is alive and well. They cannot be relying on syntactic, semantic and bibliographic information and are not recognising the words as wholes as they have never seen or heard them before. If good readers do, after all, know their phonics, we cannot rule out the possibility that this knowledge is what has enabled them to become good readers.

Incidentally, two of the test answers given by Mr Hynds are wrong. In received pronunciation, “common” and “zebra” both have five phonemes, not four.

JENNIFER CHEW The Mount Malt Hill Egham Surrey