In the lecture by Daniel Dennett featured in the previous post there’s a section about recognising language.
He discusses the manner in which we understand language by recognising the discrete packages of sound of which it’s composed.
To illustrate this, Dennett asked the audience to repeat after him an expression: “Mundify the epigastrium.” You may not know what it means, but you can make out the words and you can repeat them back (as the audience did, more or less). This is possible because the words are made up of phonemes (the smallest segments of standardised sound that are put together to form words in any particular language). Dennett then asked the audience to repeat after him a random string of odd sounds. It was impossible. This was because the audience didn’t have “the norms for correction of those sounds.” They just couldn’t get a handle on them.
Another illustration of the phenomenon of recognising the discrete packets of sound that make up words is the process that occurs when you learn a foreign language (Dennett didn’t actually mention this in his lecture – I’m adding it on my own initiative).
Some years ago I started to learn French. At the beginning of the endeavour whenever I heard any French being spoken it sounded like nothing more than a meaningless cascade of undifferentiated sound. How could anyone understand it? Over the years and with increasing familiarity with the sound structure of the language those amorphous sounds gradually crystallised into clearly differentiated words (even though I still don’t understand what half of them mean).
Because of the fact that language is composed of discrete packages of sound it’s possible to compensate for minor inaccuracies in the transmission of those sounds by mentally comparing the perceived sound with the expected, or probable, sound. You put this compensation into action every time you hear someone who has an unfamiliar accent.
This is the well known “optical illusion” consisting of the words “THE CAT” with malformed letters H and A.
With its modified H and A this wording looks slightly more like a contrived effect than it actually is. It looks slightly contrived because the letters are printed – and there are no printed letters that look like that H and A.
However, the phenomenon that you are observing – the correction in your head of letters to their proper form – is one that happens every time you read someone’s handwriting. Especially if you’re reading bad handwriting.
Here are the words as handwriting.
The effect is more or less the same. We recognise the intended letters immediately, due to their context, just as we recognise mispronounced or unusually pronounced spoken words due to their context.
But look at the following words. There’s ‘THE CAT’ on the top line – but what’s that word below it?
But can you read it as such? Possibly, but not as easily as you read ‘THE CAT”.
The H and the A suddenly start looking worryingly like made up letters. There just isn’t enough context for the H and the A to be interpreted correctly. You might decide that it’s the word AT with an extra A at the beginning.
Let’s give the word more cues by putting it in the context of a sentence. Here’s the title of a well-known children’s book.
Mmm. It still doesn’t look very convincing. What’s more, I suspect that the confusion over the H and A in HAT is creating a sort of contagion whereby the other Hs and As start to look shaky, but that’s another subject.
So it looks as though there’s a limit to the ratio of properly formed letters and malformed letters beyond which corrective interpretation breaks down.
It’s not simply a matter of the letters having to be framed within words (The H and A in THE CAT each being nicely bracketed by the T & E and C & T). Look at the expression THAT HAT, below.
It’s possibly nonsense. If you can read it as ‘THAT HAT’ there’s the chance that you’re doing so because you’re mentally referring back to the previous examples.
The problem is probably simply the preponderance of malformed letters. You just can’t get a handle on them. You scan backwards and forwards over the words and you can’t make any order out of them.
Scanning and order – important aspects of the process of understanding written words. But possibly in a more complicated way than you initially assume. To illustrate this, try reading the following sentence.
Algtohuh pcraialclty all of the wdors taht cpoosme tihs stecnene are jebumld you can plrbobay udetanrnsd waht it syas.
It reads: Although practically all of the words that compose this sentence are jumbled you can probably understand what it says.
It’s thought that this interesting phenomenon occurs because you don’t read words from left to right, letter by letter in order, but that you scan the words very rapidly, taking in the whole content at once. I assume that the words also have to be in the context of a meaningful sentence which gives clues as to the probable word that’s intended. I say this because individual words that are jumbled up are just anagrams – and I’m useless at anagrams (or amganars as tehy are smeometis claled). In similar vein, that phrase ‘smeometis claled’ doesn’t easily translate into ‘sometimes called’ in my opinion – maybe it’s too close to looking like real words itself, so it throws you off.