University Homepage School Homepage School Contacts School Search

Visual Word Recognition

Raednig thsee wrods semes to be esaeir tahn you mgiht hvae tohuhgt; waht colud epxlian tihs?

by Richard Shillcock

Introduction

Recently a piece of text flew around the internet and attracted a lot of attention. Here is one version of it.

Aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it deosn't mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht the frist and lsat ltteer be at the rghit pclae. The rset can be a toatl mses and you can sitll raed it wouthit porbelm. Tihs is bcuseae the huamn mnid deos not raed ervey lteter by istlef, but the wrod as a wlohe.

Most of the words have had their letters jumbled. Why is it still surprisingly easy to read?

Context effects

Context certainly helps. We can often predict the next word in text. If we think "University" is likely to follow "Cambridge", then "Uinervtisy" is a pretty good clue.

However, context effects are not the whole story. The effect happens with isolated words: "mgiht" on its own is still easy to read as "might".

What information is necessary

We can investigate this effect in the laboratory by timing (very accurately) how quickly people can recognize a word. People can recognize "garden" faster if they see a jumbled version of "garden" ? like "graden" ? very, very briefly just before the real word appears.

Recognizing any one word means distinguishing it from all of the rest of the words in the mental lexicon (the set of words a person knows). To recognize "might" we need to know the identity of the letters ? m, g, t, i, h ? and something about their position

Getting just some pairs of letters in the right order may be enough to recognize a word, or at least help to do so. Pairs of words, like "clam" and "calm", seemingly can't help activating each other in the brain.

We can recognize jumbled words exactly because we don't need to know all about the positions of the letters.

End letters

The jumbled words shown here have all had their first and last letters left alone. These "end letters" seem to be particularly important when the brain recognizes words. Why should this be?

One suggestion has been that the end letters suffer less interference from adjacent letters. They have a space on one side. Another suggestion comes from a computer model of reading (which we discuss below)

How does the brain do it

Does the brain give us any more clues about how it identifies words?

One major clue comes from the anatomy of the brain. When we look at some point inside a word, the part of the word to the left of the fixation point is initially projected to the right half of the brain, and the part of the word to the right is initially projected to the left half of the brain.

The two halves of the word are initially in different halves of the brain. Solving the problem of word recognition involves coordinating these two pieces of information.

The two halves of the brain principally communicate through a thick bundle of fibres, the corpus callosum. This starting point in word recognition constrains how the problem is solved. Whether particular letters are on one side or the other is the most important information about letter position, together with knowing the identity of the end letters. In English, only very few confusions like "trial" and "trail" remain.

Computer Modelling

A good way of testing a theory is to implement it on a computer. The computer blindly follows every order given to it. It can discover unexpected things about the theory and can show any weak points in it.

The illustration shows one particular computer model of visual word recognition, as it appears on the computer screen. The model is vertically divided to resemble the two halves of the human brain.

Such a model automatically relies more on the end letters, and is one way of understanding how stored words compete with each other.

Dyslexia

What is dyslexia and does research of this kind tell us anything about it? Dyslexia refers to a range of reading impairments with different causes, in children and in adults.

In Hebrew a specific kind of "letter-switching" dyslexia has been reported, with confusions like that between "clam" and "calm". Hebrew contains a lot of scope for such errors.

Any research that tells us about normal reading can help us to understand what is happening when reading does not happen as easily as it might.

References

For an extensive discussion of this issue, see: http://www.mrc-cbu.cam.ac.uk/personal/matt.davis/Cmabrigde/

Shillcock, R., Ellison, T.M. & Monaghan, P. (2000). Eye-fixation behaviour, lexical storage and visual word recognition in a split processing model. Psych. Rev., 107, 824?851.