I just read an interesting new article from the journal Scientific Studies of Reading called "The Categorical Perception Deficit in Dyslexia: A Meta-Analysis", by the wonderfully-named Mark W. Noordenbos and Willy Serniclaes.
There's been some research showing that dyslexics (notwithstanding the debate about this term) have trouble noticing the differences that matter between speech sounds, and not noticing the differences that don't matter.
This article explored the reliability and significance of this finding by analysing data from lots of relevant research.
When we're babies, we're capable of discriminating all the sounds of all the languages of the world, but we quickly discard the ones that we don't hear in the language(s) around us, and focus on the ones we do (there's a great Ted Talk called The Linguistic Genius Of Babies about this, if you want to know more).
Noordenbos and Serniclaes trawled through the scientific literature and found 632 studies related to dyslexic students' speech sound perception. They distilled this down to 36 studies of a total of 754 individuals with dyslexia which were relevant to their research question, properly-conducted, complete and comparable enough to put through their statistical mill. The studies were of speakers of English, Dutch, French and Chinese.
They found that dyslexics were indeed reliably and significantly worse at discriminating between speech sounds (phonemes), and more sensitive to sound variations within speech sounds (allophones), than peers matched by age, as well as peers with the same reading age (though the differences were greater between the dyslexics and age peers). Sound discrimination tasks showed up their problem more than sound identification tasks.
To use an analogy, if speech sounds were pets and dyslexic kids were asked to sort cats from dogs, they would:
- Put a few large, friendly cats and a hyperactive flop-eared bunny in the dog category, and a couple of the smaller, quieter, fluffier dogs and an iguana in the cat category.
- Be quite distracted by the differences between labradors, Great Danes, Jack Russells and chihuahuas, and between Siamese, Angora, tabby and Grumpy cats, even though these differences were irrelevant to the task at hand.
If the filing system for speech sounds in your head isn't clearly defined, and you tend to slice sounds up too finely, it makes sense that you'll have a hard time working out how they're represented by letters.
Confusing similar sounds is a part of normal speech development, and only becomes a problem if it affects quite a few sounds or goes on for longer than usual.
Tiny tots replace "k/g" with "t/d", and thus propose to doe outside and play with the doddy, or help toot dinner. Many seven or eight year olds are still inclined to replace "th" with "f/v" and say vey fink vere might be a funderstorm on Fursday. One little chap I know often says "n" instead of "l", so he tells me what he nikes and doesn't nike, and not to nook at his cards when we play TRUGs (he never nets me win).
Try saying the word "let" and really concentrating on what your mouth is doing. Now do the same while saying "net". Not much difference, eh? Tongue, teeth and lips are all in much the same position, the voice is used, but the sound comes out the sides of the tongue for "l" and via the nose for "n".
Lots of sounds we think of as quite different are acoustically very similar. Try ringing someone up and telling them you've just been on a picnic, and had some lovely trees under the cheese in the park, and a nice bottle of wine. I bet they don't notice anything odd, because "ch" and "tr" sound very similar (as "ch" is really "t" plus "sh"). Even if they do notice the switch, they're more likely to think the problem was with their ears or the phone than you playing Rhyming Spoonerisms.
The boundaries between vowel sounds are especially mushy, because vowels are produced by adjustments to an open mouth shape. Try describing how your mouth produces the sound "er", and how it differs from the sound "or".
Try saying the following vowels one after another, running them all together: ee (as in me), i (as in pin), e (as in red), er (as in her), a (as in cat), u (as in cup), ar (as in arm), o (as in hot), or (as in for), oo (as in good), ew (as in chew). Good luck deciding perzackly where the boundaries between these sounds lie.
Sounds also get pulled around in the mouth by neighbouring sounds, so that the "i" in "pip" (between consonants produced at the front of the mouth) has a fairly different sound from the "i" in "kick" (between consonants produced at the back), but most of us just file them both under "i" and don't notice the difference.
As adults we think the sound "l" sound in "tell" is the same as the "l" sound in "let", but acoustically, they're quite different. However, English-speakers tend to ignore the difference because it doesn't matter for meaning.
Sound differences across languages
I've learnt from my Somali-speaking students, and discussions with taxi drivers, that Somali has no "p" sound.
If you say "p" in Somali, it is perceived by other Somali speakers as "b", which is the closest thing to "p" in Somali.
When I start work on English spelling with my Somali-speaking students, I say "pet", and they write "bet". When they speak English, a lot of the "p" sounds are pronounced as "b" (with voice).
We have to create a "p" in their English phonemic inventories, or much of what they but on baber in English brobably won't be sbelt broberly.
Years ago I had a delightful classmate from Hong Kong called Man Tuk, who used to laugh at my attempts to say his name with correct Cantonese sounds, which involved tonal vowels and a kind of swallowing sound at the end instead of a "k".
Of course I always got it wrong and ended up calling him things like "flea umbrella" or "peanut elbow", much to his amusement. I couldn't hear the difference between what he was saying and what I was saying, because my sound system didn't include the relevant phonemic distinctions/categories.
I have a Scottish friend who rhymes "put" with "boot", and an Aboriginal friend who rhymes "been" with "tin". In most US accents, the first vowel in the words "father" and "body" is the same.
The main audible differences between accents of the same language are differences in vowel sounds.
There can also be differences in consonants – Americans and lots of other Englishes put an "r" sound after vowels, the Queen and people with pretentions to the peerage say "whale" differently from "wail", and when teaching English in Mexico I found that Mexicans tend to pronounce "yellow" as "jello".
These differences are subtle enough for us to usually understand each other, though I remember sitting in the back of a bus in Jamaica wondering what language my fellow passengers were speaking, then realising it was English.
The multiple spellings of English sounds seem a big nuisance if you only consider them from the perspective of a single accent, but maybe multiple spellings are an advantage for a language with multiple accents. I guess Samuel Johnson, the person who wrote the first important English dictionary in 1755 and thus nailed down our spelling system, was aware of many accents and the need to devise a spelling system that could accommodate them all.
My Scottish friend can file the "u" in "put" with "flu", "truth", "gluten" and "super", and the "oo" in "boot" with "good" and "book".
Americans can file the "a" in "father" with "want", "quality", "salt" and "wallet", while I put it with the "a" in "grass", "past" and "task".
Mexicans could file "yellow" with "Yogyakarta", at a stretch, or just pay more attention to their English teachers, and stop mucking around and having so much fun in class.
My Aboriginal friend can only file "been" with "breeches" and "threepence", but don't worry, she can spell, we had a great Grade 2 teacher. However, I do wonder whether materials to teach Aboriginal students about sounds and letters are designed with enough regard for their accents.
Whatever your speech sound categories are, you need them to be clear and well-defined, and not to pay too much attention to differences inside them, to make sense of the complex way sounds are represented by letters in English.
Noordenbos and Serniclaes say that with reading instruction and experience, perception of speech sound categories improves, and that dyslexic students start to behave more like their peers, but more sensitive neurophysiological studies suggest they still have subtle differences in their speech sound perception.