Wednesday, November 19, 2014

Calling all English teachers!

It's been brought to my attention that English teachers are not immediately receptive to Unspell. When presented with a copy of Unspeller, they ask the rather obvious question: Why do my students need to learn this? And so here is the answer. If you are an English teacher, please read this. And if you know any English teachers, please send the link to them, and tell them that this is something they need to read. Thank you.

[Download printable PDF] [Japanese version]

The Unspell Teaching Method

1. Problem Statement

Students in English-speaking countries do significantly worse in learning to read and write than students in most other countries with comparable educational systems. Whereas in most countries it takes just 2-3 years to learn to read arbitrary texts with good diction and to take dictation accurately (although comprehension may lag), in English-speaking countries this process takes on the order of ten years. What's more, often it never completes: the failure rate is unacceptably high, resulting in functional illiteracy rates that approach 50% in some countries. The effects of such systemic failure are wide-reaching. There is the opportunity cost: students waste years attempting to acquire rudimentary skills instead of learning something interesting. There is the hit to economic productivity from so many people incapable of retraining themselves on their own but requiring oral instruction. There is an adverse effect on health and public safety from so many people unable to read safety instructions and brochures. Functional illiteracy is especially widespread among the prison population and hampers the efforts to rehabilitate prisoners upon release.

In spite of the vast resources and effort directed at achieving basic literacy in English-speaking countries, and in spite of the excessive failure rate of these efforts, few people have dared to ask the simple question: Why is this? Yet all you have do is look, to find both the source of the problem and its solution. It is curious how a culture that embraces radical change in some ways chooses to remain tradition-bound in other ways, even where these old ways inflict great harm.

English spelling presents a unique set of challenges to any child learning to read, because written English is an opaque code. Unlike most other languages, it is not a rendering of speech that is based on orthographic rules but a hodgepodge of orthographic styles collected over the centuries from an assortment of languages, most of them extinct. Some 40-50% of English spellings displays some degree of irregularity; as for the rest, the student has to explicitly memorize the fact that they are unexceptional. For instance, having learned the words “over,” “open,” “only” and “okra” as unexceptional, the learner then has to discover by trial and error that “oven,” “other,” and “osprey” do not follow the same pattern. In essence, the only way to learn to read English is to memorize both the spelling and the pronunciation many thousands of words—a task that calls for more rote memorization than just about any other task in which humans regularly engage.

What makes this task even harder is that the learner isn't being offered any way to directly translate English spellings into sequences of phonemes, for ease of memorization. The human mind is a thirsty sponge for spoken words, which are sequences of phonemes. It is neurally wired for the two very complex, distinct tasks of speech perception and speech production, and phonemic memory is the vital link between the two, for which human mind is wired for it as well. In essence, every child comes equipped for building a mental dictionary, and the symbols that comprise this dictionary are not letters but phonemes. In languages where letters map directly to phonemes this distinction is largely irrelevant, but an opaque code such as written English is a major impediment to learning. This is because the human mind, and especially a child's mind, is not especially good at memorizing sequences of abstract symbols, such as phone numbers, lists of random pictures or the spellings of English words. Thus, the task of learning English spelling relies on something that is essentially a talent worthy of a savant, which much of the population does not possess.

A second challenge posed by English is that there is no easy bootstrapping mechanism for learning to read it. The typical sequence of events in learning to read an alphabetic language is as follows:

1. learn what sounds the letters make
2. learn to form syllables out of these sounds
3. learn to form words out of the syllables

Instead, the student has to memorize the spelling of each word as a whole and then look up its sound in non-verbal memory. Any unfamiliar word becomes an indigestible blob, because the student is afraid to sound it out for fear of making a mistake and remembering it incorrectly.

The only work-around, or fallback, that is currently made available is for the student to “spell out” words. This sort of “spelled-out” English is, in essence, a language that consists of just 26 words. All human languages share some important commonalities—they all have both vowels and consonants, and they all have words that consist of syllables. Beyond these commonalities there is a wide variety of linguistic forms, but to date no human language has been found to consist of just 26 different words. The reason for this is simple: such a language would be far too long-winded and far too lacking in variety to be easily learnable.

“Spelling out” explodes a monosyllabic word like “strengths” into “Es, tee, ar, en, gee, tee, aitch, es”—a perfect example of the weakness of this system. “Spelling out” is not a mnemonic technique but a bizarre parlor trick for those who have already memorized words as sequences of abstract symbols. It is like speaking in Morse Code: another savant-type skill of which few people are naturally capable. No other language has anything similar; the usual way to convey how a word is written in an alphabetic language is simply to pronounce it carefully, placing equal stress on each syllable.

In summary, the problem with teaching written English is this: the student's mind is naturally adapted to memorizing words as sequences of phonemes; instead, it is being forced to memorize words as sequences as abstract symbols that have no direct and unambiguous relationship to phonemes. The student is not being provided with something vital: a way of converting between sight and sound, and back, that can quickly become effortless and automatic. This is the main cause of trouble with basic education in English-speaking countries—adequately accounting for both its inefficiency and its unacceptable failure rate.

2. The Unspell method

The Unspell teaching method offers a way to cleanly circumvents all of these difficulties. Unspell consists of a minimal set of symbols which directly represent generalized speech sounds of the English language in a way that is maximally independent of accent or dialect. These symbols are largely disjoint with the Latin alphabet, eliminating interference effects with spelled English. Each symbol directly represents a specific sound. The student learns to sound out each symbol, then group these sounds into syllables, syllables into words, and words into phrases in what is essentially a self-governed, self-motivated process. The role of the teacher is to guide the student through this process, and does not involve imparting any specialized knowledge. Virtually all that needs to be memorized is presented on the following wallet-sized card:

It is important that the symbols used by Unspell are not referred to as “letters” because that causes confusion between spelled and unspelled English. Rather, they should be referred to as “shapes,” and it should always be stressed that “shapes” make “sounds.” Each shape makes a different sound, and to read unspelled English all you have to do is drag your finger along, making the sound of each shape. The sounds should not be syllables. This is easier with vowels, liquids such as L, N, M and fricatives such as S and F, which can be sustained. It is harder with affricates such as P, T, and K. But it is important that they be taught as pure sounds, not as sounds embedded in nonsense syllables, because the nonsense syllables can mask the distinction between unvoiced and voiced consonants (P, T, K vs. B, D, G).

Classwork can proceed as follows:

1. The students are introduced to the 13 basic shapes of Unspell, and learn to draw them in chalk, crayon, ink brush or by finger-painting. (The term “draw” is helpful, to differentiate Unspell from “writing” English, since most children enjoy drawing but don't enjoy writing.) If they learn to appreciate the shapes of Unspell as a sort of abstract design they can embellish, so much the better. The emphasis need not be on correct form or penmanship, but on what distinguishes each shape from the rest: this is the only kind of “mistake” that a student can make. However, it should be stressed that all the shapes are formed along a horizontal line, hanging down from it. How far they hang down distinguishes consonants from vowels; how wide they are along the line distinguishes stressed vowels form unstressed ones.

2. The students work through the Unspeller booklet, learning all of the sounds made by all of the shapes. There are 12 lessons, but some of them are quite short, and so anywhere between 5 and 7 sessions may be sufficient. The exercises at the end of each lesson are cumulative, incorporating the knowledge gained in the previous lessons. Repeating the last few exercise at the beginning of each session may be helpful.

3. The students start reading unspelled texts; first, by chanting out the sound of each shape, rhythmically, first in monotone, then giving stressed vowels a higher pitch. Second, by going over the same text, but now grouping the sounds into syllables, stressing each syllable. Third, by grouping the syllables into words. Fourth, by grouping the words into phrases. This sets up the basic pattern for conquering the unfamiliar without the teacher's assistance: whenever a student encounters an unfamiliar word, the approach to it is to drop down as many levels as needed to learn to say it.

3. Learning spelled English

Once the student is proficient in reading Unspell, the process of learning English spelling becomes greatly simplified. This because now the student cannot help but sound out spelled English one letter at a time. Of course, it sounds all wrong, because spelled English represents how English sounded centuries ago, if at all, and is essentially a dead language. (Since children like dinosaurs, it may be useful to explain to them that written English is like the dinosaurs—long extinct, but we still admire their majestic skeletons, and try to imagine what their mighty bellows might have sounded like.) Just as with dead languages like Latin and Classical Greek, the fact that we don't really know how they sounded does not make the trick of sounding out Greek and Latin words any less effective in helping to remember them. To help the student get the general knack of this entirely informal, subjective and, in the end, private skill, it is helpful to introduce the student to the Pseudolatin Alphabet presented below, in which each letter or letter combination used in written English has been assigned a unique sound.

Note that although the “names” of the Latin letters and letter combinations are presented above as syllables, this is only so because English spelling cannot express sounds except as syllables; the unspelled versions of them are simply sounds.

Also note that although this mapping is largely arbitrary, there is a principle involved: it cannot destroy information. For instance, C cannot be conveyed to either a “keh” sound or a “seh” sound because remembering it that way would preclude going back from sound to symbol; would it be K, C, or S? Similarly, although Ph is phonetically equivalent to F, it is mapped to puheh, because otherwise arbitrary spelling distinctions (“fantasy” vs. “phantasm”) would be impossible to convey in sound.

Lastly, it is helpful to stress that double-consonants, which abound in English, should be pronounced with a sort of stutter. They are largely meaningless, having no phonetic reality in contemporary spoken English. They used to be related to a short/long vowel distinction which no longer exists, were blindly carried over from Latin and French, or were thrown in haphazardly by Dutch typesetters who used them to pad the bill for the job because they were paid by the letter. The stutter makes them easier to remember.

4. Conclusion

The Unspell Teaching Method reduces the problem of teaching students to read and write English to the problem of teaching students to read and write a largely phonetic form of modern English, then using it as a basis for teaching them an archaic dialect of English, in which most words have one-to-one translation in contemporary spoken English but sound somewhat different. This is by no means an insurmountable task.

5. Computer teaching tools

 The task of teaching Unspell, then teaching English spelling, can be enhanced through the use of the following online tools:

1. A tool that presents a table of Unspell shapes, and sounds out each one when touched or clicked on.

2. A tool that pronounces arbitrary sentences written out in Unspell, by sounding out each shape while highlighting each Unspell symbol as its sound is being played.

3. A tool that pronounces arbitrary spelled English sentences using the Pseudolatin Alphabet shown above, highlighting each letter as its sound is being played.


sjoh said...

I will send this to some teachers I know in Korea. An earlier commenter on another post mentioned King Se Jong's similar project with hangeul in the 15th century. Unspell has a few functional and aesthetic parallels with hangeul, which i sense would make it particularly well received in Korea. So, I suggest you particularly explore possibilities for unspell there. Some Korean english learners transliterate english words into hangeul to help them remember the pronunciation. But, of course, this is not ideal due to differing language structures/sounds. So, a tailor-made system for english was necessary to invent. I learned Korean "spelling" in about 4 hours, and my initial dabbling with unspell suggests it is just as easy to learn (and just as fun). I must emphasise the sense of relief Korean learners would have, just as I sensed relief when I first learned hangeul, and amazingly there were NO exceptions (to spelling rules)! But, for a sensational hit it would probably be good to develop a smart phone app (to translate english text or isloated words at least to unspell), then you'd have a few million more users in your Korean market.

Dmitry Orlov said...

Hangeul was part of my inspiration in creating Unspell. Yes, Koreans, and other CJK-language English learners, could benefit greatly from Unspell. There are some difficulties with English beyond spelling, having to do with vowel reduction and the unbounded set of possible syllables, which Korean does not have, so, yes, Unspell is necessary, and I hope it finds a receptive audience. And the smartphone app you mention is definitely on the list.

rl said...

I find this YouTube video a good primer as to why English got so messed up in the first place ->

And Sir Orlov, by now you should know that common literacy isn't always an objective when a country tries to control its populace ;)

Anonymous said...

I have many questions, but I'll ask only one to begin.

As I understand it, Unspell is a copyrighted proprietary system for which people will have to pay the owner to get books and materials. I couldn't publish and sell books using the Unspell orthography without paying the owners a licensing fee. (Imagine having to pay the Chinese government every time you printed a book using Chinese characters.)

How is such a bottleneck conducive to making this system widespread? (If my assessment is wrong, please correct my mistake.)

Dmitry Orlov said...

James -

The purpose of Project Unspell is to create a public good: a simple, easy-to-learn way to read and write English that removes barriers to literacy and learning by eliminating the need for formal education and rote memorization. But since no public entity is about to step in to create this public good, it is being created through private means. This, along the way, creates some opportunities for profit.

Unspell is copyrighted and will be made available under a specific license. The intent is to make Unspell free for public, non-profit uses in areas such as public education, religion, publicly-funded social services and the like. It will also be free for personal, non-commercial use, as well as for artistic uses, such as designing and selling new fonts or incorporating Unspell in artistic installations and productions, whether for profit or not.

Opportunities for profit will exist wherever Unspell is used to extend or replace existing commercial offerings: selling unspelled versions of books and other publications, selling software, teaching tools and other materials. In all cases where a product is sold to the public, a royalty-based fee structure will apply. A separate fee structure will apply for uses of Unspell in advertising and political speech, along with certain restrictions on content, to prevent Unspell from being turned into a tool of economic or political exploitation.

The initial revenue stream will come from reprints of books via print-on-demand, subscriptions for premium access to Unspell-related Internet resources, and from app downloads.

ArtS said...

Perhaps you should promote unspell for use in the teaching of English as a second language.

I've seen statements that there are as many as a billion people studying ESL around the globe (hard to believe).

Maybe some brave instructor somewhere would try a randomized comparison of two classrooms, one taught traditionally and the other taught using unspell.

Individuals being taught primarily to be able to provide phone support might be able to avoid standard written English altogether!


V. Arnold said...

Frankly, as an English teacher in S.E. Asia; I think unspell is, how can I put it? Unworkable to put it kindly.
My synapses fire and scramble logic to incomprehensible levels.
English comprehension in S.E. Asia is very low and I do not see how unspell can help.
It would further incomprehension, IMO.

Dmitry Orlov said...

Arnold -

Not sure what's wrong with your synapses or how they relate to Unspell. How can a reasonably phonetic alphabetic writing system NOT help learn a language? Perhaps you can explain.

Chris said...

This is a very interesting solution to a serious problem, though I'm not sure learners of English who already employ phonetic alphabets in their native language would necessarily benefit that much. The way I personally learned spelling complicated English words was by sounding them out in phonetic Hungarian. For instance, to this day, every time I write down "beautiful", I have to recall the Hungarian phonetic pronunciation of the English spelling (beh-ah-oo-tee-ful) and this is how my students memorize the spelling of unfamiliar words as well. What English really needs in my opinion is a reformed latin alphabet based spelling, which is a lot simpler and entirely consistent with standard American pronunciation. I agree with Dmitry that RP (received pronunciation) isn't really suitable as such, mostly because too few people actually speak it natively and the R-dropping is extremely unhelpful. German spelling has been simplified quite a few times over the last hundred years, it is time English followed suit. America should take the lead and abandon British English spelling altogether.

Dmitry Orlov said...

Chris -

Yes, most ESL students learn to decode written English quite separately from learning to speak and understand spoken English.

As far as English spelling reform, all you need to understand about it is contained in these two words: NOT HAPPENING. Therefore, we won't even discuss it here. It's a waste of everyone's time.

Unknown said...

My son is a very bright but easily distracted 8 year old who really struggles with reading. He applies the "sound it out" technique with the results that you describe... The sounded out 'words' often do NOT sound out like real words, he has to use his memory to convert the sounded out letters into a recognisable English word, and he is easily discouraged when he applies the technique and cannot recognise a word, because of the problems of English that you have identified.
So I will have a go at teaching him with your Unspeller. Expect my order soon, and I'll let you know how it goes.

I learnt spoken and written Spanish as an adult a few years ago, taught myself, I am fluent, and certainly noticed the difference in the logic and ease of learning that language. I feel very sorry for any adults trying to learn written English!

flyingcardealer said...

after purchasing and working my way through unspeller, these are my initial thoughts...

one, very easy to learn. i spaced the lessons out over a few days, but i'm confident i could've worked my way through the whole book in one sitting. i should probably point out that i have, thanks to my english teaching days in japan, some experience with learning syllabaries, such as hiragana and katakana. i'm not sure if my past experience transferred over to making the unspell learning process easier, but, either way, i didn't find it difficult.

two, the shorter vowels is a clever idea--and helpful. i surprised myself with how quickly i did exercise three in lesson three.

i've also discovered, surprisingly enough to me, that i can transcribe english to unspell quicker than i can read it. i'm not sure why that is, and i'm curious if others have had a similar experience. then again, i've only had the book for a week now and my practice has been rather sporadic. i'm sure my reading speed will improve over time with enough exposure.

also, i noticed you have a symbol for the "'a' as in snake" diphthong, but not one for "oi" as in "oil" or "boy". there is an example of how you would transcribe it in "the wind in the willows" text where you transcribe "joy" using the vowel combination of "'o' as in ostrich" and "'y' as in yak". however, that combination translates more to an "eye" pronunciation for me. i tried switching out the "ostrich" for "'o' as in goat," but that really didn't work either. (i pronounce that "joh-y" which rhymes with "buoy".) so you might want to consider creating a symbol for the "oi" sound.

on a more general note...

i'd recommend a close proofreading of the unspeller text for transcribing errors. i did find one in "the wind in the willows" portion on page 78 of my copy. the second sentence down, starting "how he scraped and scratched...," you transcribe "how" with the "'o' as in goat" symbol instead of the "'ow' as in owl" one. nitpicky i know, but it did stop me dead in my tracks as i tried to read it. anyways, i haven't tried working my way though either of the other text sections yet, so i couldn't tell you if there are any other errors. i'd just recommend another proofreading before your next printing.

as far as use as a teaching tool goes...

one of the most frustrating experiences i had with teaching english in japan was overcoming pronunciation errors created by katakana. katakana is the japanese syllabary that is used for loan words and, unfortunately, it doesn't always help. for example the word "bridal" translates to "bu-rye-da-ru" in katakana. having the students repeat my pronunciation wasn't always successful. i think unspell could work as bridge-gap in this instance. if they can hear the pronunciation and see how it looks on paper (unspell is pictorially fairly similar to katakana), then unspell might offer some needed scaffolding. i'd recommend trying to get in touch with any of the english conversation school chains in japan and pitching your ideas to them. you might have some luck there.

but yeah, i enjoyed working my way through unspeller. good luck :)

Dmitry Orlov said...

flyingcardealer -

Yes, your experience with finding Unspell initially easier (although perhaps not faster) to write than to read is quite typical.

Which vowel/diphthong is which will shake out. Unspell uses a vowel/diphthong inventory that works well across most accents or dialects, and to do that it has to use generalized phonemes, too many for a few accents, too few for a few more. I kept oi as a digraph because it is the only diphthong that is almost everywhere some sort of o + some sort of i, so that's how it will stay.

The typos you caught have already been fixed; I have made two rounds of updates to Unspeller, mostly to fix typos such as the one you mention.

I agree that Unspell (ung-spe-ru) would make a good replacement for katakana, which does sound badly mangled and probably impedes communication as much as it helps it.

As far as contacting English-teaching schools in Japan, I am not sure how to go about it, and would welcome any help. One thing I can do is come up with a Japanese version of Unspeller. For credibility, it would be good to have an introduction (in Japanese) written by someone who has taught English in Japan. The way the Japanese pursue fads, if Unsperu catches on, it could be a very good thing for someone.

Leigh Blackall said...

I learned Mandarin following Bopomofo Unfortunately Pinyin is now the dominant form of teaching it
I wonder if Unspell would have a better chance at success if it used the existing characters rather than these new ones. It would be a shame though, a body of text in these Unspell character sets would look graphically wonderful.

Esker Coffey said...

This is an interesting concept. I am an engineer by trade. My brother and some of my friends and other relatives are dyslexic. It is a difficult mental condition to understand if you've never seen it. There might be value in 'Unspell' in teaching these people to read. I have a rough working knowledge of Russian and I've never heard of a native Russian speaker being dyslexic. Do these people exist? I suspect that this is due to the different and logical structure of Russian in contrast to English.

Unknown said...

@Esker Coffey if you want assistance, or anyone for that matter, I have been working on language processing and this sounds like a unique project.