Can pictographic cues make an augmented script such as the Shaw Alphabet easier to learn and remember?

An extended comment on Alice Coleman's article on the Shaw alphabet.

New article: Read said that Shavian was the end of the line - and his quick script largely abandoned the analytical elements in Shavian in favor of shapes that were easier for users to draw, use, and remember.  Monofon - monoline fonetic - aka pictografic monofon - manages to be both more analytic than Shavian - more building block approach, more grouping of similar sounds with similar shapes, more memorable (more pictographic cues).

Shavian - Quick Script - and Monofon
and the ASCII keyboard

An issue with any extended or augmented alphabet is where to position the new characters on the old keyboard.  There has to be a transitional script (usually upper case letters) that can be converted to the extended font.

The following table of 44 phonemes of English shows how these four notations are related.
 
 

By Steve Bett

In a recent article (JSSS, 1-1995, p. 25 ), Alice Coleman argued that while the Shaw Alphabet excelled on several important dimensions such as ease of reading and writing, it was seriously deficient in others. She was particularly disappointed with its ahistorical character, its use of non-traditional shapes, and with the amount of time and effort required to learn it.

Most of the benefits of Shavian stem from its logical structure, its regularity, and its selection of simple single stroke character shapes. Coleman finds these efficient streamlined symbols to be too divergent. Such a drastic change (if adopted) would, in her words, cut us off from the past and from the community of nations who have adopted the Roman Alphabet.

Coleman probably overstates the nature and extent of the divergence. Most of the letter forms used in Shavian have historical roots. What is novel is the way that Read assigned sound values to his symbol set. This disconnect from traditional sound-shape associations probably reflects Shaw's wishes: "The new alphabet must be so different from the old that no one could possibly mistake the new spelling for the old." ¹

Much of what Coleman seems to want would seem to be attainable if there were answers to the following questions:

1. Can a fixed logical set of relationships between sounds and simple character shapes be established that retain enough of a visual relation with familiar alphabetic forms to be immediately readable?

 

 
 
 

2. Can the beneficial features of Shavian be retained in a script that rearranges the assignments of sounds to shapes in such a way to sustain a connection to the past and to a new set of acrophonic letter names?

 

 
 
 

3. Can these relationships and correspondences be extended and enhanced to make the resulting notational system easier to learn and use?

In Shavian, each unvoiced consonant is matched by its voiced equivalent which is usually the same shape rotated 180 degrees. The distinctive shapes were designed for rapid writing and easy reading. According to Coleman, Shavian can be written twice as fast as conventional letter shapes. Writing speeds of up to 100 words per minute can be attained. Because the characters and are more compact and create more distinctive and compact word patterns, Coleman claims that texts printed in Shavian can be read 50% faster than conventional texts. Pitman (1962, p. 13) believes that Shavian text could be read 75% faster.

The narrow condensed shapes used in the Shaw alphabet require less space on the printed page. Narrow letter forms combined with the elimination of digraphs results in a savings of up to 33% in space and printing costs. (Pitman, 1962, p. 12)

The claims for ShawScript are impressive. Compared to TO, Shavian is alleged to be superior on the following dimensions:
 

Improving Ease of Learning

Neither Shavian nor the mostly Roman script used by TO incorporates any mnemonic strategies for simplifying or facilitating learning. Neither script is particularly "learner friendly." Coleman suggests that the Shaw alphabet is more difficult to learn than traditional letter forms. Shavian’s use of similar shapes for similar sounds, its regularity, and its elimination of upper case forms should make the script easier to learn (from scratch) than TO (traditional orthography). The basic problem with Shavian is that it requires English readers to unlearn familiar sound and shape associations.

Coleman (p. 26) argues that a new script should be sufficiently like the old one to (1) make acquisition fairly effortless and (2) ensure that those trained in it would not be cut off from the wealth of earlier written and printed material. Shavian falls short on ease of transition, ease of learning, and connection to the past.

If ease of learning is to be the primary consideration in the evaluation of a constructed script, there are at least three strategies to explore:

1. grouping and relating similar sounds with similar shapes
2. keeping the shape of new phonograms close to the shape of familiar ones
3. using a mnemonic to connect shape and sound (esp. for new phonograms)

There is at least one script based on Shavian that had ease of learning as its primary design consideration. That script is known as pictographic monofon (or PMF). PMF could be used as an initial teaching medium. However, it was designed to provide a strong connection between shape and sound, not to make it easier to learn TO. The design goal for PMF was to create a script that can duplicate the literacy level attained in two years with TO in 2 weeks and to provide a way to teach 18 new letter shapes and sounds in 30 minutes.

For readers of English, PMF would not be the easiest transitional scripts: Cut Speling is clearly the best candidate for this. A script such as PMF would, however, be among the easiest unigraphic (or augmented) phonemic scripts.

Pictographic Monofon (PMF) demonstrates how it is possible to develop a an iconic acrophonic alphabet which uses the same single stroke simplicity of letter form found in Shavian without losing the connection traditional letter shapes. PMF illustrates how it is possible to retain the benefits listed by Coleman for the Shaw alphabet with a script that is more historical, more readable, and easier to learn than notational systems which arbitrarily assign shapes to sounds.

As shown in the Figure 1, Shavian is a monoline phonemic script that has the appearance of a foreign language. While nearly 50% of the new letter forms could be related to traditional Roman letter shapes, only 15% retain a familiar shape-sound connection.² For example, in Shavian, the J shape is associated with the /f/ sound and the 180 degree rotated J, which looks something like an f, is associated with the /v/. There are a number of other instances where simply reversing the sound to shape assignment would result in a better visual relationship to traditional forms. See the complete 40 character chart.

The limitations of establishing closer relationships to familiar alphabetic forms

Even if 26 of the 40+ Shavian characters were similar to traditional letters, there would still be 18 or so new characters and the fact that no phonemic writing systems duplicates the word pattern of TO. As shown in the illustration above, rule based scripts always look a little strange on the printed page because they change the spelling of 30% to 60% of the words.¹

Yule (1982) who studied both regularized rule based spelling and cut spelling, concluded that "readers can adapt to shortened spelling within minutes."

Her research suggested that rule based (or phonemic) spelling such as WES (World English Spelling) is more disruptive to old reading habits than cut spelling. Yule's graphs (1982, p. 657) indicate that cut spelling could be read about 10% faster than TO. Phonemic text was read at a rate that was about 20% slower than TO (traditional orthography).

Based on Beech's (1982) study, one might presume that the reading rate for phonemic spelling would improve with practice. Beech estimates that the time required for adults to adjust to the difference between one form of regularized spelling and traditional spelling to be around 2 hours: "…adults regain a normal reading speed after they have read 6,000 words of regularized text." When familiar digraphs are replaced by new letter shapes, this adjustment period is likely to be a little longer. A reasonable hypothesis might be that augmented alphabets would have an initial reading rate that was at least 30% slower than TO and that 12,000 words of practice would be required to regain ones normal reading rate.

The time required to write with a regularized orthography may be over four times that required to learn to read it. In Beech’s study, 30% of his subjects were writing accurate regularized English by the end of an afternoon session. However, while their speed increased during the practice session, it was still significantly slower than their normal writing speed.²

The power of pictography as a mnemonic device

The mnemonic device used to facilitate the learning of two of the earliest scripts (Egyptian and Semitic/Canaanite/Phoenician) has not been used for some 2000 years. The trick was to use the letter name as the bridge between shape and the sound. There is no reason why this picture letter mnemonic should not be revived to make a contemporary alphabet easier to learn and remember.

PMF (pictographic monoline phonemic aka monofon) is an attempt to make an augmented phonemic script easier to learn and familiar enough to be immediately readable. One way it does this is to make a font change signal a variation in sound.

The key feature of PMF is the use of pictographic shapes. Figure 2. illustrates how naming letters after familiar objects provides a bridge between shape and sound.

PMF presumes that the chief problem in learning any script is maintaining the shape sound relationship. If I forget what sound /ae/ is associated with, I have to consult an IPA code book. With PMF, every letter looks and sounds just like its name. The IPA character "ae" doesn't look like anything. Monofon shapes are iconic, they are supposed to look like or represent something. The letter AX looks something like an ax and from this any native speaker can reconstruct the appropriate sound.

Both Shavian and PMF are monoline phonemic scripts. This means that shapes can be formed with one quick continuous stroke and there is one and only one shape for each significant sound in the language. Both scripts are phonemic: they follow an alphabetic rule and spells words the way they are pronounced.

Most of the PMF characters are simplified versions of Roman characters. This applies as well to the characters for the 18 sounds that are not represented by a single character. Nine of these new characters are vowels. The four A-likecharacters are avian, ax, hawk, and up. Ox is also one of the sounds of the A, as in *father but unlike the other sound signs, it does not look like a traditional A. It does look something like the horns on a Semitic 'alef (or ox), the shape that was rotated to become the original Greek A.

As shown in red in the sidebar, the phrase "A banana" has 4 different A sounds: /Ae bah-na-nuh/. These can be clearly indicated in PMF without resorting to digraphs. PMF is a pronounceable script, TO isn't.

Pictograms are used today for international symbols but have not been consciously used in the design of scripts for nearly 3,000 years. By using a sun sign ¤. to identify the brightness control on a TV, the set can be sold in any country. Manufacturers avoid having to identify the control in any particular language. While an explanation of the symbol can be found in over three languages in the documentation, most people can figure out the connection between the icon and brightness without consulting the instructions. Pictographic scripts are a little different because the sun sign, ¤, which might represent /b/ for brightness in English is probably associated another sound in a different language. Phonographic pictograms tend to be language specific.

Pictograms or picture letters make direct reference to objects, persons, relationships, or events. They re-present by sharing criterial attributes, such as contour, surface structure, number of parts, pattern, color, and order of connection. Pictograms, unlike most logograms, are not arbitrary. The similarity between the pictogram and the reference may be remote, but the resemblance or likeness is easily grasped post linguistically (i.e., after it is pointed out).

Since monofon shapes are abbreviated, the connection to the letter name is not always obvious. The picture on the right, derived from an Egyptian wall painting dated 2800 b.c., provides a more realistic version of the letter "avian". Each letter in the PMF script has a similar realistic picture (and also a gesture or sign language) that would be used to connect each shape with the appropriate imagery. Picture letters have often been used in beginning reading books to provide an external association e.g., "a is for apple," "a is for archer," or "c is for cat." They have not been used as an internal or integrated mnemonic for nearly 3,000 years.

The original D phonogram was a hand (Eg. DR-T) which became a door (DLT) in most Semitic scripts. The acrophonic name in English that is closest to hand is digits. Door would work as a transcription of daleth but the delta shape never looked much like a door. On the other hand, the v shape could easily represent the thumb and index finger on a hand. Rotating this form 180^o looks like the tip of a spear. In some ancient scripts ti does represent the point of a spear or arrow.

The upper case H looks more like a hurdle or saw horse than the lower case h used here. The shape is named hurdle because the ancient heth referred to a barrier. In the ancient Sinai script, the shape appears to be an abbreviation of the Egyptian glyph for hail [hai], a kneeling man with arms upraised.

The following table shows the vowels in cardinal vowel order and the blends or diphthongs. In some cases the shapes of the blends are derived from the shapes of the component sound signs. The letter owl, for instance, is a set of horns (ox) on top of a circle (oat). The letter hook /hoo-uhk/ is a blend of the letter hoop / ooze and the letter up / acute.

Letters are said to be meaningless sounds arbitrarily linked to meaningless shapes. When the pairs that must be associated are meaningless and arbitrary, they are difficult to recall.⁴ Pictographic scripts significantly reduce the amount of effort required to learn a new script. Pictograms provide a meaningful link between shape and sound through the letter name. Anyone who can remember a list of 40 object names can easily recall the associated shape and sound. By having a reference, pictograms provide an effective mnemonic for learning and retaining letter shapes, sounds, and shape-sound correspondences.

Many of the simple monoline shapes found in Shavian can be used in a pictographic script. In most cases, the simpler the form the more distant the association with a particular object. For some people even a complex shape such as the lower case "a" will be too abstract. They will require coaching to see the "a" as a representation of a bird (avian).

A pictographic script with historical roots similar to the one described would seem to overcome most of the objections that Coleman had to Shavian. Such a script would be less of an obstacle to easy reading and easy learning than Shavian.

The chief advantage of a pictographic script is that it does not have to be used all the time to be remembered. One can learn it in a couple of hours, and it will stick. One does not have to consult a code book to reconstruct the appropriate letter form for any sound one can pronounce. To write "ouch" one recalls the imagery associated with "owl + chick" Since chick looks like a ligature of the traditional CH digraph, in addition to a chick's head, all that remains is to recall the shape of an owl's head to regenerate the "horned circle" or Taurus zodiac shape. Pictographic scripts make excellent parallel phonemic scripts.

Clearly there are some trade offs and compromises. It is difficult to make a pictogram as streamlined as the characters in ShawScript. Without abbreviation, they could not be written quite as fast. PMF characters are generally not as familiar as those found in New Spelling or WES. This would result, initially, in a greater loss of reading speed.

It is only when PMF is compared across four or five dimensions that it begins to look noteworthy. It may only win one of the five races, but it is always a contender. When the cumulative scores across four or more dimensions are compared, PMF comes out on top (provided that one of the dimensions is ease of learning).

Summary

Coleman is right, the ShawScript is ingenious and in some respects near perfect but falls short on the ease of learning dimension. Shavian is as difficult to learn as a shorthand system. This flaw has prevented it from attaining its potential as a parallel replacement script for TO.

PMF shows how it is possible have a fixed logical set of relationships between sounds and simple character shapes that retain a visual relation with familiar (or historical) alphabetic forms.

About 30 Shavian characters can be related to traditional Roman forms. Changing the sound assignments might make 75% of the characters easier to recognize and remember but it would not be enough to make the revised script, in Coleman's terms, "a starter." Simply reworking the shape-sound pairings is not likely improve the probability that the new script would be widely adopted.

PMF also shows how the beneficial features of Shavian can be retained in a script that rearranges the assignments of sounds to shapes in such a way to sustain a connection with the past and to a new set of acrophonic letter names. Using letter names that are rich in imagery and which start with the appropriate sound may be one of the best ways to enhance the script and make it easier to learn and use.

Pictographic Monofon (PMF) demonstrates how it is possible to develop a an iconic or pictographic alphabet which uses the same single stroke simplicity of letter form found in Shavian without losing its connection traditional letter shapes.

PMF (Pictographic Monoline Fonetic) borrows 30 Shavian shapes and copies other Shavian features such as using similar shapes for similar sounds. PMF retains most of the benefits of Shavian while using simplified letter shapes that are much closer to traditional Roman forms. This makes PMF immediately readable and ideally suited to be easily remembered phonemic notation for English pronunciation.

Eleven of the 40+ PMF characters are direct copies of shavian shapes and another twenty are close. The PMF characters differ from the ShawScrip in two important respects: they usually resemble traditional Roman letter forms and they incorporate a mnemonic to enhance teaching/learning and retention. As a consequence, PMF retains most of the benefits listed by Coleman for the Shaw alphabet while being more readable and easier to learn.⁵

PMF is easier to learn and use than Shavian but not quite as easy to write. The shapes are not quite as abbreviated as Shavian because they have to retain enough character to resemble particular objects. Once learned, however, abbreviated shorthand forms could be used for personal notetaking (See the column titled AMF [abbreviated monofon] ). For instance, the letter hook has a barb. The barb makes it look more like a fish hook and indicates that the sound is a blend of the hoop [U] and acute [/]. The shorthand form is just a reversed J. The letter chick is ligature of the digraph [ch]. It also looks like the head of a chick. The shorthand form is a J rotated 180^o.

Using traditional letter forms is not the only way to promote "learner friendliness." The earliest Semitic alphabets (e.g., Phoenician) incorporated a mnemonic strategy, borrowed from the Egyptians, that partially accounts for their rapid diffusion. The pictographic strategy was to use the letter name to link shape and sound. The earliest letters looked and sounded like their names. PMF revives this memory aid.

The key to PMFs learner friendliness is its built in mnemonic hooks. Every character resembles a familiar object and the initial sound in the object's name is the sound to be associated with the shape.

Can pictographic cues make an augmented alphabet easier to learn and retain? There are enough studies of imagery and learning to indicate that the cues found in PMF would work. There is enough data to indicate that meaningful pairings of name and shape are retained better than arbitrary pairings. Pictographic alphabets would be easier to learn. However, there is no hard data to indicate how much easier pictographic alphabets would be in practice and if there is a significant difference between integrated mnemonics and add on mnemonics. Is there a significant difference between (1) The letter is avian and (2) The letter A stands for avian?

For more information, consult the web page at the following URL:
http://www.luorc.edu/ecrc/stevehp.htm

Bett, Steve T. (1995) Logical Letters. (unpublished)

Bett, Steve T. (1996) Pictographic Monofon. (unpublished but available on the web))

Bett, Steve T. (1996) The Trouble with Spelling. Louisiana Middle School Journal.

Beech, John R. (1992) Adaptation of Writing to Orthographic Change. Jour. of Gen. Psych. 119 (2), 161-179.

Beech, John R. (1983) The effects of spelling change on the adult reader. Spelling Progress Bulletin, 23, 7-13.

Brown, Bob. (1991) The Shaw Alphabet Competition-Some Background. SSS Publication

Coleman, Alice. (1995) Graphology and Writing Systems: the case of the Shaw Alphabet. JSSS 1, 25-30.

Dewey, Godfrey. (1971) English Spelling: Roadbock to Reading. New York: Teachers College Press.

Drucker, Johanna. (1995) The Alphabetical Labyrinth: the alphabet in imagination. Thames & Hudson, London.

Pitman, James and John St.John. (1969) Alphabets and Reading. London: Pitman Publishing.

Shaw, G. B. (1948) preface to Richard Wilson's The Miraculous Birth of Language, New York: Philosophical Library.

Shaw, G. Bernard. (1962) Androcles and the Lion (1st published in 1913) Reprinted with parallel text in the Shaw Alphabet. Includes notes on spelling by Peter MacCarthy and suggestions for writing in ShawScript from Kingsley Read. London: Penguin Books.

Wilkins, Bishop John. (1668) An Essay Towards a Real Character and Philosophical Language.London.

Wood, George. (1883) A Popular Tretise on the History of Letters. Hartford, CT.
One of the early attempts to link the letters of the alphabet to pictograms. Wood's thesis was the A-shape really was derived from the ax shape: The E was an eel, and the S-shape was a snake. B a bow, and C a crescent. Wood had an excellent imagination and was a good artist, but his theories about the origin of the alphabet was not based on conjecture not fact.

Yule, Valerie (1982) Shorter Words Mean Faster Reading. New Scientist. 9 December, 96, 656-657.

Yule, Valerie. (1986) Design of Spelling. Harvard Educational Review.

1. The reason for the low correspondence may have had something to do with the nature of the 1958 "new alphabet" competition that Read entered. Bernard Shaw had promulgated his ideas for a non-roman alphabet in a letter to the Times on 15 April, 1941and in the preface of Wilson's book (1948). According to Bob Brown (1991, p.1), "Shaw came to the radical view that digraphic schemes like New Spelling, or proposals using diacritics or extra letters, just would not do..." According to Pitman (1960, p. 102f), Shaw became the most trenchant publicist for a complete rejection of the roman alphabet. "The new alphabet must be so different from the old that no one could possibly mistake the new spelling for the old." (1948, p. 39) The proposed British alphabet had to be non-roman and contain at least 40 characters which would enable the English language to be written without indicating single sounds by groups of characters or by diacritical marks. Shaw's preference was for a 42 character script with 18 vowels. (1948, 39).

While changing the sound-shape pairings to bring them closer to TO might satisfy Coleman's desire for a more traditional shapes and correspondences, it may have disqualified the script as a contender in the new alphabet competition sponsored by the executors of the Shaw estate.

2. Most English sounds have four common spelling patterns in TO. Since there are 560 ways to spell the 40 sounds in English, the average is about 14 patterns per sound. If only one spelling pattern is allowed, 60% of the words are changed. If two patterns are accepted for some sounds (or some words), percentage decreases.

Beech's regularized English changed 30% of the words. According to Dewey (1971), World English Spelling (and New Spelling) normally change about 60% of the words in a passage. An analysis of the percentage of regularly spelled words in the Gettysburg address reveals the following:Figure 4 indicates that while only 25% of the phonemes manifest irregular spelling, 60% of the words in the address would have one or more sounds that would not be spelled according to the most common spelling rule. This means that 60% will look "funny" when converted to any phonemic orthography.

3. The author was unable to locate any published research on the time required to adapt to Shavian or PMF. If Shavian had been part of the studies by Yule or Beech, the number of correct words per minute would have started out near zero. If PMF had been included in Yule's study, the reading rate for this script would probably have begun at 30% slower than TO or 20% slower than WES.

4. The literature on imagery, meaningfulness, and learning is substantial. The following books contain references to the many articles that have been published on "paired associate" learning and other experiments demonstrating the role of imagery in recall and retention.

• Richardson, John T.E. Mental Imagery and Human Memory St. Martin's Press, N.Y., 1980
• Yuille, John C. Imagery, Memory and Cognition. Erlbaum, N.J. & London, 1983
• Goss, Albert (1965) The Role of Meaningfulness in Learning
• Broudy, Harry J. (1987) The Role of Imagery in Learning.

The 18 new letters required to eliminate digraphs in TO

The complete character set

The new notation grid.

The PMF short vowels

The PMF long vowels

The Pictographic Monofon Vowels

A poorly scanned notation grid showing the number of ways a given sound can be spelled.

On the Origin of the Alphabet

______

additional notes.

The thesis that PMF is mnemonic and therefore much easier to learn and retain than non- pictographic meaningless letter signs, sounds, and names has some support in the research on paired associate learning, imagery, and meaningfulness. Meaningful material, particularly when it can be related to an image, is learned and retained much more readility than nonsense material.

The alphabetic principle makes literacy easier by allowing readers to pronounce words from their spelling and writers to spell them from their sounds.

The pictographic principle makes learning and retaining the shape - sound relationships by relating both to name of a familiar object which becomes the letter name. It allows readers to reconstruct the sound of individual letters from their shape and writers to reconstruct the appropriate shape of a letter from its sound.

At some point in the future, PMF could be part of a spell check program that would allow people (1) to enter a phonetic spelling and get a TO spelling and (2) allow a computer user to query the pronunciation of a particular word and get it spelled out in PMF.

PMF is only marginally better than other more familiar notational systems. The only dimension where it might be "twice as good" as competitive scripts is on ease of learning or teachability. It is not yet a "starter" except possibly as an auxiliary script that might replace IPA in an English pronunciation dictionaries. It could also be an initial teaching alphabet which would continue to be used throughout life to sound out words.

The basic principle of alphabetical writing is to represent a single sound of a spoken language by a single letter. A phonetic writing system is the best approximation. Even the best writing system requries an orthography because (1) language changes over time (2) there are regional variations, (3) some blends sound the same which means there would be more than one way to spell a particular word.

Phonemic systems are as Yule explains, more like a sketch than a photography of language. The sketch doesn't have to be perfect, just good enough to communicate. The constructed writing system does not have to be accurate enough for technical linguistic analysis.

The subjectives ratings of a small group of adult English readers for Shavian on the first four dimensions, reported by Bett (1995), was slightly different. Rather than potentially readable at a 50% higher rate, they found the script to be unreadable. The perceptions of the rating group probably correspond to the perceptions of those in Coleman's university circles who were indifferent to or unable to see the potential of the new script.

It is not enough for a proposed script to be better than a traditional script in a few areas. To be what Coleman calls "a starter" it probably needs to be nearly twice as good in virtually every important respect. PMF may be an improvement on Shavian on a couple of dimensions, but it has yet to be developed to a stage where it can be said to be twice as good as alternative writing systems on every important dimension. It is an extrapolation to claim that it is twice as good on any dimension.

PMF does require a new font and this used to be a problem. Today, the font can be made available on the Internet free of charge to anyone who wants to take the time to download it.