A breakthrough in learning to read English
Children can learn to read English, in a fraction of the time it typically has taken, by means of in-text markup
--
When I wrote this article …
… back in the glory days of 2018, I went into excruciating detail about the mappings in English between Phonemes and Graphemes.
The difficulty, it seemed to me, in learning how to read is that although there are only 26 letters in the alphabet, many of these letters are used in combinations to indicate sounds that in some cases aren’t even in the alphabet — like the “oy” sound in “toy”. As it happens, there are something like 49 distinct sounds in spoken English, which get blended together to make all the words we speak — so obviously there cannot be a one-to-one mapping between letters and sounds. For a start, all the vowel sounds have at least one short sound version(for example “a” in apple), and one long sound version (for example “a” in apron) {not to mention, in the case of the letter “a” the r sound version like when people in the south of the UK take a bath}. And some of the vowel sounds have several such — more about this below.
Even so, absence of a one-to-one mapping between letters and sounds would not be too big a deal were it not that many of the sounds in spoken english (the so-called “phonemes”) are coded into words in multiple different ways (the so-called “graphemes”). For example the long “ay” sound you get in “play” can also be made with about 8 other letter combinations; consider how each of these words has an “ay” sound in it that has been coded using different letter patterns: apron, make, rain, eight, straight, vein, day, grey. Straight off we have 8 different ways that an “ay” sound can be coded into a word — ie. 8 different alternative graphemes that can all be used to represent the “ay” sound. And the “ay” sound is not by any means out on its own in this respect. Almost all of the English vowel sounds have the same issue of being represented in English words in multiple alternative ways, and several of the English consonant sounds do as well.
Then on top of this, to make matters even worse, many of the multiple graphemes that can all be used to make one of the word-sounds (“phonemes”) of English, sometimes also make one of the other sounds. The letter “a” which makes an “ay” sound in apron, also makes an “a” sound in apple, and “o” sound like orange in what, an “e” sound like elephant in any and many, and depending on your regional accent it may even make an “ar” sound like car in fast, path and bath.
{ Aside: The letter “a” can also makes the curious and ubiquitous sound that no one ever tells you about at school, even though you use it all the time, which phoneticists call a schwa. An example in English is the short vowel sound of the “a” in the word “about”. If you actually listen to yourself talking, as opposed to already knowing what sounds you were taught that “a” makes, you’ll notice that you probably don’t say “ay-(like in play)-bout”, and nor do you say “a-(like in apple)-bout”. What you actually say is something like “uh-bout”. And it turns out that the schwa sound can be represented in written English by just about any vowel or any of the typical vowel combinations, as well as by vowels in combination with consonants.
Roughly speaking, de-emphasised syllables in polysyllabic words very often become schwa sounds, regardless of how they are spelt. Consider the word “content” which can be pronounced in two alternative ways so as to make it mean either (1) feeling generally happy about things — the first syllable becomes a schwa when it is de-emphasised — or (2) the sub-items contained within something — the first syllable is emphasised, although the second gets to keep its ‘e’ for elephant.}
Asides aside, let’s get back to our main story …
In information management terminology, this two-way multiple mapping between phonemes and graphemes is what’s known as a “many-to-many” relationship:
- Any given sound in english (“phoneme” — such as “ay”) in many cases may be represented by multiple alternative letter combinations (“graphemes” — a, a_e, ai, eigh, aigh, ei, ay, ey).
- and also
- Any given letter combination (“grapheme” — such as “a”) in many cases may be representative of multiple alternative different spoken English sounds (“phonemes” — ay, a, o, e, ar).
In the article linked at the top, I belaboured this point endlessly. In all there may be about 200 of such phoneme->grapheme / grapheme->phoneme mappings (about 50 consonant sound mappings and 150 vowel sound mappings). When a student learns to read English, in order to decode written letters into sounds, they have to spot, for any given sound in any given word, which one of these 200 mappings is being deployed , and then blend those sounds together to decode the word. Even for words with simple CpVpCp (Cp) consonant-phoneme (Vp)vowel-phoneme (Cp)consonant-phoneme structure, that makes the number of permutations you need to cross-correlate in order to read a word something like 50 x 150 x 50 = 375,000 … that is to say something well over a quarter of a million possibilities that need to be searched across in your “mental computer”!
But now some Good News !
But here’s the good news. While the number of mappings between phonemes and graphemes is something in the order of 200, the number of phonemes is less than a ⅓ of that.
There are something in the region of 45–49 sounds (“phonemes”) that go to make up spoken English — perhaps 27 consonant sounds and 22 vowel sounds — something like that. (People argue about the details, but for our purposes, and at the current time these details don’t matter much.)
So in fact, in order to be able to learn to read English easily, when you are at the stage prior to that of just being able to recognise any given word by sight, all you actually need to be able to do is recognise which one of the 49 phonemes is being represented by any given grapheme. Instead of having to learning 150 mappings, if we could find a way of indicating to students learning to read, which phoneme was being indicated, they would only need to recognise 49 sounds.
Even more good news comes from the fact that only about 8 of the consonant sounds have multiple significant grapheme spelling alternatives. (For example, I take “d” as opposed to “dd” to be not very significant.)
For example, if we could manage to do this, we would have cut range of possibilities for CpVpCp words down from 50x150x50 (375,000) to 27x20x27 (16,038) … Possibly we would have immediately made reading English (“roughly speaking”)
1 - 16/375 = 95% easier !
Consequently (admittedly making a fair amount of assumptions about what slows down progress, when learning to read) we might expect to accelerate the process of learning to read by approximately 20 times.
How to achieve this?
So as I said above, if we could find a way of indicating to students learning to read, which phoneme was being indicated, they would only need to recognise 49 sounds, and not 200 mappings.
But how to achieve this?
The obvious answer would seem to be to annotate text in some way. We don’t want to make any significant alteration to text, because doing that simply puts off till later the learning task that students need to traverse.
But if we could find a way of marking-up text to indicate for each of the multiple graphemes representing a given phoneme, what the target phoneme was in any particular text, we could essentially teach students to read while they are reading.
A tiny snippet of Harry Potter, will serve as an example:
Notice how the “ay” sound in each of the words “they”, “made” and “saying” are all marked up with an a Rain Cloud, indicating that “ey”, “a_e” and “ay” all make an “ay” sound in these cases. Similarly the “oa” sounds in “cloaks”, “know” and “doughnut” and all marked with a Boat and the “oo” sounds in “group”, “too” and “to” are all marked up with a Moon.
Using markup symbols in this way, students can have it hinted to them, right there in the text, the correct phoneme to correspond to the multiplicity of different graphemes that can represent it.
Several things happen as a consequence of this:
- Students become familiar with the hinting symbols and the (single phoneme) sounds that they hint at, ie. indicate.
- Students are able to build up a repertoire of all the different graphemes that can be used to represent that phoneme when used in the context of the different words in which that sound is a part.
- Students are prompted with hints about the correct way to decompose a word into the sounds that go to make it up, there-by making it easier to read unfamiliar words by putting those sounds together.
- All of this can be done while actually reading, rather than having to separate out the learning of phonics from the more rewarding process of actually reading something that might be interesting, engaging or even exciting.
So as to fulfil on this break-through, I have started to develop the “Phonicked” project. Please visit www.phonicked.com for more information about this.
