Samuel Putman - A Universal Writing System

7/30/2019 Samuel Putman - A Universal Writing System

1/78

PhonA Universal Writing System


2/78

ii iii

Phon: a Universal Writing System, release 0.7

Copyright 2006,2007,2008 Samuel Putman.

All Rights Reserved.

nb: The information contained in this text is meant to be

shared, and future editions will be under less restrictive license.

For now, please ask my permission before distributing copies to

others. I will probably say yes.

Introduction to the Alpha Edition

Tis document is the alpha edition o Phon: a UniversalWriting System. Te system itsel is complete with the exception osome tentative work in the tone group area; the book itsel is expectedto undergo several revisions beore general release.

Te book is in the general structure it will take and the text islargely in, with the exception o the phonemic alphabet chapter. Tatchapter is kind o pulling in two directions at the moment. Basically,in describing Phon in phonemic terms, I can treat it in terms o theIPA or in terms o its own logic. I attempt to do both, and in the beta I

will give both tasks separate justice. In the mean time, I considered thecharts comparing Phon and IPA to be more immediately important,

and have included them.Te bare minimum diagrams have been provided. In particular

the English chapter will be bulked up considerably in this respect, and theorm and morphology chapter could use more explicit stroke diagramsas well. Te appendix will eventually have a table o Phon characters interms o the logic o Phon rather than IPA; this considerable task willhave to wait until Im done traveling, in April. Phon is in this book, butexpect some o it to be conusing or unclear. I welcome any questionsand suggestions that any o you may have. Also any errors you see,please send by page number; I will be grateul.

Te Quenya spellings in the book are inconsistent. I willeventually do a global search and replace to make all the diacritics showup, but ater awhile I just stopped using them so I could write aster.Comprehension is not damaged by this minor irregularity o style.

Tere are little red tags scattered throughout the book. Teseare uncompleted tasks. Tere will be another alpha edition, wrappingthese up, as soon as Im able to get to it. Te layout, in general, is inthe right order, but no serious eort has been made to clean it up, sothere are large whitespace patches beore diagrams and blank pages at


3/78

the ends o some chapters.Te most glaring absence is the suprasegmentals, which are

incompletely described, unnished, and not listed in their table. Illget there, I promise; i you have a specic interest in this topic pleasewrite me and Ill give you what Ive got.

Many o my alpha readers will have some amiliarity with theproject; some ew o you are getting cold-called because I have goodreason to suspect youll nd it interesting. In all cases, I welcome andencourage eedback on the book! Anything at all which occurs to you,dont hesitate to ask. I need the active support o as many o you as

possible to get a good beta edition together and prepare or generalrelease. You will have my lavish thanks, both personally and in theront o the book.

Te plan is to incorporate eedback while expanding the bookin various ways. When this is done, Ill produce a beta edition: thiswill be available as a sliding-scale PDF le (DRM ree) and as a trade

paperback. Te beta edition paperback will only be available until therelease edition (hopeully with a credible publisher), at which time


4/78

Contents

Introduction to the Alpha Edition iiiPhon: An Introduction and Users Manual 1

A Brie History o Writing Systems 4A Brie History o Phonetics 12

Phon in (and or) Plain English 27English Vowels in Phon Sec1:36

Te Form and Morphology o Phon 45Phon and the Nature o Order 71On Stroke Order and Sort Order 77

Phon as a Phonemic Alphabet 81Te Case For Reorm 82


5/78

Phon: An Introduction andUsers Manual

Phon is the preerred name o a universal writing system,

intended to meet the goal o being a universal auxiliary writing system,that is, one that anyone o any language group can eel comortable


6/78

2 3

important, would be equally comprehensible in each direction.Tis proved possible, ater some time spent pushing a pen across

a page and applying knowledge o symmetry theory, ergonomics, anda ew palette-limiting aesthetic criteria. Te result is a stroke pallete othree primary strokes, our primary combinations, and twelve secondary

combinations. Tese primitives are not arbitrary in any sense: they arethe result o applying certain rules o orm and symmetry to the basicstrokes o an angled pen. All primitives except the dot primitive areasymmetrical across the vertical axis, both with respect to themselvesand with respect to any other primitive.

Tus equipped, Phon was designed as a eatural writing systemcovering the same state space as the International Phonetic Alphabet.Tat is to say, there is at least one unique symbol or each grapheme inthe IPA, and every diacritic and suprasegmental has an equivalent ormalso. By standing on the shoulders o this particular Collossus, we mayairly say that Phon can do everything IPA can, and backwards, too.

Phon, however, is designed to be eatural, so that each strokeo every grapheme provides inormation as to the manner o utteranceo the associated phone. Tis is ound in Hangul, but nowhere elseamong native scripts; in the constructed script world, the engwarbehave similarly, but in Phon the system is quite a bit more rigorous.Utterances themselves are not in any sense ully systematic; Phon canclaim only to be more eatural than any writing system in existence, asperect isomorphism is believed to be neither possible nor desirable.

Te aesthetic rules or stroke generation and recombinationallow or tens o thousands o distinct graphemes, ar more than the

hundred-odd which are needed to implement a phonemic writingsystem. Phon is thus being extended to cover mathematical symbolism,and other endeavors such as musical notation are contemplated.Tere are also a large number o characters, called protograms, whichare considered part o Phon but have no assigned meaning, and asmaller set o mesograms, which are used when writing some languagesmesographically, a term which is explained in the section on strata ouse: the concepts involved are similiar to those invoked by the spectrumo broad to narrow transcription. Te ormal nature by which Phon

the phoneme represented. It may be written equally in either horizontaldirection, as well as both vertical modes, and the direction o sense isinherent in the structure o each stroke; urthermore, let and righthanded individuals can each write one o the modes uently, which

was an important design goal, as let-handed writers are approximately

ten percent o the population. Being based only on an invented script,the engwar o J.R.R. olkien, it is without the prejudice o historyand national advantage ound in any script based on a natural writingsystem, and is in principle equally accessible to anyone who knows howto use a pen: the needed technical terms, similarly, are drawn rom thelanguage Quenya, which is not a historical language, and derive theiroriginal meaning rom aspects o the natural world (or a ew ubiquitousand ancient objects) which are common to our heritage as humans.

Phon is designed to be rational, universal, useul andbeautiul. In meeting these design goals, insight was incorporatedrom the existing writing systems o the world, rom other examples

o constructed scripts, as well as rom phonetics, ergonomics, cognitivescience, calligraphy (in the Romantic and Islamic script traditions),and mathematics, particularly combinatorics and symmetry theory.Te result is a writing system similar to an abiguda, incorporatingseveral distinctive eatures.

Abiguda is a contemporary word or what used to be called asyllabic alphabet; Devangiri is the most widely used, with ibetan,Tai and most o the Indian scripts in this amily as well. Phon maybe considered an abiguda, i one takes the liberty o considering theinherent vowel to be a non-utterance. Alternatively one may think o

Phon as an alphabet, but one in which the vowels are written above andbelow the consonants, as in the Taana script o the Maldives.Phon is unique in several respects rom any other writing

system in use. First, Phon is the native script o a let-handed person,who had become rustrated with the inherent righthandedness o allexisting scripts. Rather than merely design a writing system to suit theneeds o let-handed persons, or simply mirror-write an existing script la da Vinci, the decision was made to try and design a writing system

which could be written equally with the let or right hand, and just as


7/78

4 5

Writing is many o these things, but rstly it is an invention,and one o recent acquisition. Many humans cannot read nor writein any script: they dress and talk alike to their peers, work and loveand raise children as humans will do, but without literacy, and no one

would judge them eral or less human or it, though prejudices (and

practical disadvantages) abound.Writing was in no sense invented all at once. ens o thousands

o years ago, earlyHomo Sapienswas drawing in charcoal and clay on thewalls o caves, and carving symbols into sot limestone and sandstone.From what we know o Aboriginal culture in Australia, these symbolsystems were probably quite complex and rich, keys into the landscapeor stories that guide trails, record hunting history in the region, pointto water and shelter, warn o danger rom man or beast, or mark tribalterritory.

We also begin to nd tally sticks, and calendrical constructssuch as the well-known Stonehenge. Marks to count are combined with

symbols to represent the things counted: this i s one o the protoliterateactions, a practical solution to a pressing need that points towardsurther uses.

Another activity our paleo cultures engaged in is divination.Augury is a skill completely uncomprehended by most writers, sadly;the ight o birds is a rich language and can oretell disaster or goodhunting or one skilled in interpretation. I am less convinced o thebenets o extispicy, but you take the hints you can get. We dont knowor sure that these were practiced, because they dont leave a record.

We do know that the ancient Chinese practiced divination by scoring

tortoise shells and putting them into a re, interpreting the resultingcracks; the practice is reerred to in historical texts, and archaeologicaldigs have revealed the shards. Tis is believed to have been an impetusto the Chinese script, as pictures were carved into the shells to representthe subject o divination.

Irrigation created the rst empires o the old world, in theIndus, the Chinese river basins, Mesopotamia and Egypt. Writing

was developed in all these, a lthough that o the Indus is disputed andremains cryptic. Sumeria was rst, and is believed to have inuenced

characters are specied leaves great reedom in rendering, allowingor distinct amilies o type aces, calligraphic hands, and renderingin the grafti style o urban calligraphy. Phon is ree o Eurocentricbias and other cognitive problems that arise in the use o a Latin-basedphonemic alphabet, and being mnemonic and regular is easier to

acquire and use.Tis manual is meant to introduce the Phon project to as

many people as possible, so that we can get a user and developer baseestablished and really get this tool sharpened up and put to use. Wellbegin by situating Phon in its historical context, as a writing systemand as a product o linguistics as an endeavor distinct rom writing perse. Chapter wo will introduce the English language writing mode;this chapter is meant to be directly accessible, so that one can getstarted reading and writing Phon directly. Chapter Tree will denethe underlying structure o Phon and explore the ormal categoriesused in writing it, while Chapter Four will introduce Phon itsel as

a tool or writing, with emphasis on its role as a tool in phonetics.Eventually we will include writing modes or other major languagesas well. Subsequent sections will introduce the mathematical set oPhon symbols and give a guide to their use and urther development,

will discuss the calligraphic and cultural applications o Phon, addressthe issues raised by Phon in computing, and provide guidance or theurther development o the system.

A Brief History of Writing Systems

Speech cannot be called an invention. It reaches back intoancestors, not just o ourselves but o our very species, and its originsare basically mysterious despite all our etymological research. We allspeak; the rare human raised without speech is deprived o culture andis more like the the other mammals than it is like us. Languages shapeour social reality, and color our perceptions; how we speak is how wethink, where we come rom, and what we have made o ourselves.


8/78

6 7

tradition. Again, the Egyptians developed symbols that were used ortheir sound, used or annotating logograms to make their meaningclearer as well as or transcribing oreign names. Te Egyptians alsodeveloped their script in two orms: Heiroglyphic and Heiratic.Heiroglyphic is the amiliar, picture-based system made amous rom

tombstone art, while Heiratic was mostly written with a reed pen onpapyrus and looked more like writing as we know it. Te systems weresimilar but by no means identical. Egypt had a strong caste system, andthe scribes o Egypt were important people who guarded their literacy

jealously; over time, the Egyptian systems actually became harder touse.

It is rom Heiratic that all scripts in current use, which are notnot based on ancient Chinese, are descended. Te Egyptian empire wentthrough multiple periods o expansion and ruled over Semitic peoplein the Levant during these times, with many Semitic peoples comingthrough conquest or migration to live and work in Egypt. It is here

that the alphabet is born, based on imitation o the Egyptian writingsystem. It is thought to be imitation, rather than adoption, becausethe symbols were adopted based on how they would be pronounced inthe proto-Siniatic language, not in ancient Egyptian; thus we have, orexample, a symbol that ultimately becomes R that is abstracted roma head shaped symbol that was sounded t-p in Egyptian, but was usedor , the root or head in Semitic languages.

Tese early alphabets were inuenced by the structure o Semiticlanguages: most o the symbols reer to consonants, with a ew o them(called matres lectionis) used irregularly to show where vowels go on the

occasions where only a vowel, and not context, can show what wordis meant. Called abjads, the Hebrew and Arabic scripts are still in thisamily; although both later developed vowel diacritics to make vowelsexplicit, they are oten neglected in casual use. Abjad is the word or analphabet in Arabic, ater the rst our letters.

Abjads were the rst alphabetic scripts, and a great power wasborn. Logograms are linked to words, and this is no doubt one o thesources o the Chinese civilizations great strength and endurance.Populations that came under the rule o the Chinese, in the pre-

Egypt in developing their scripts, but no evidence linking Sumeria tothe Chinese writing system has been discovered. Te only other well-attested writing event occured among the Olmec and later Maya othe Americas, which have no current survivors; thus it is accurate tosay that all current writing systems descend rom either Sumeria or

China. Cuneiorm was originally pictographic, carved with a roundstylus on clay, and concerned in surviving samples chiey with inventory.Over time, the shapes o the pictograms were abstracted, and the stylus

was standardized as a wedge-shaped device or rapid impressions. Tisis the transition to logograms, unique and abstract shapes that represent

words. Tese shapes were oten hard to distinguish, and so the scribesstart to notate them with other logograms that are used or their soundrather than their meaning, which is the beginning o syllabic orm. Te

Akkadians adopted cuneiorm as a syllabary or their language, prior tothe decline o its use with the rise o the alphabetic scripts.

Chinese writing appear to have undergone a similar transition:From carving pictograms or tally and oracles, to using them to representunique words, through the process o abstraction. Something happens,then, which puts the Chinese onto their own distinctive track. TeChinese have a limited syllable set, and each syllable has a meaning itcarries in and o itsel, what we call an isolating language or analytic.

Ancient Chinese had this quality to a greater degree, to where eachsyllable can be thought o as a word, in and o itsel. Te Chinesereduced their pictograms to logograms through various methods: aew are in act pictograms still, some are ideograms (the character or

down points down, sort o) but the vast majority consist o two ormore radicals, one o which hints at the pronunciation (or used to)and others which suggest the meaning, with no real consistency. TeChinese never developed a syllabary or alphabet until modern times;the closest they got are symbols which are mostly used in transcribingoreign names.

Te Egyptians were in trade with the Sumerians, and appear tohave developed a logographic system out o symbols they were alreadypainting and carving into stone and clay as part o their native crat


9/78


10/78

10 11

will ocus on Japan and Korea because o the developments in writingsystems which took place there.

In both Korea and Japan, writing was Chinese writing rom thebeginning o literacy, and this meant at rst that the literate read and

wrote Chinese. Over time, certain characters came to be used as a gloss

or their sound, similarly to how characters are composed o radicals inthe rst place; this gave a highly irregular, idiosyncratic syllabary, usedinitially to marginally annotate the meanings o Classical Chinese texts,a development which, i much later than that in Egypt and Sumeria, isnonetheless parallel.

In Japan, these logosyllabaries became Hiragana and Katakana,rom separate traditions (women o the court and monks respectively) oshorthand. Te Japanese never abandoned the logographic componento their writing system, and continue to use a rich and heterodox brewo classical Chinese characters, two syllabaries, and the Roman alphabetto represent their language; their writing system is almost a living ossil,

like the coelecanth or the gingko, more like Cuneiorm or Heiraticthan even the Chinese writing systems (o which there are now two, as

well as a dominant Roman transcription and the bo-po-mo-o systemused or teaching in aiwan). In principle, any o Hiragana, Katakanaor Romaji would sufce to write the Japanese language, but there isvery little interest on the part o the Japanese in doing this.

In Korea another unique writing system was developed,arguably the most interesting script in current use rom the perspectiveo Phon. Hangul is traditionally credited as the work o one man, KingSejong; although it is clear rom the record that he had advisors, it is

quite possible that he developed Hangul singlehandedly, as we havethe example o Phags-pa and Sequoyah (who wasnt even literate) topoint to; interestingly, Phags-pa script would appear to have been aninuence in the development o Hangul. Te king released the system in1446, and it was resisted by the educated elite, or classist and classicistreasons both. Hangul was used or about ty years, abolished by a laterking, and came to be used only by women and the uneducated untilKorean nationalist sentiment arose in the late nineteenth century; it

was not until the latter hal o the 20th century that the Hanzi were

Even more remarkably, we nd our writing system organizedin a way that ollows the eatures o the sounds they represent.

Alphabets have a conventional order which is memorized, while mosto the Brahmi scripts are organized by area o articulation, with rankscontaining voiced, unvoiced, aspirated and unaspirated variants. Te

symbols themselves ollow no overarching logic, but there is somedegree o similarity between characters o similar sound, sometimes.Critically, however, the order o the characters is logical, i not reectiveo current thinking on the subject.

We are ollowing a particular trail, and it leads into themountains o Central Asia. First, the Sogdians, and later the Mongols,

were to take a proto-Syriac script and do something interesting to it.Te Chinese were in the habit o writing vertically, although horizontal

writing was known, and in imitation o this custom the Mongolianswrote vertically as well, by turning their script 90 degrees, thus vertical,let to right, rather than the Chinese vertical, right to let. Tus we

have our modes o writing which have been used by decent sizedpopulations; poetically, we might call them the Latin, Aramaic, Chineseand Mongolian modes.

Te Mongolian alphabet itsel was a bit o a poor t or thelanguage, with phonemically relevant distinctions missing rom therepresentation. Ater the Mongolian conquests, they were let with aprousion o scripts and languages under their rule. Kublai Khan hada ibetan lama named Phags-pa come up with a script, named aterhim, or uniying the languages o his empire. Phags-pa came up witha vertical phonetic script useable or writing ibetan, Mongolian, and

Chinese, which is quite an accomplishment. Phags-pa script died withthe Yuan dynasty, but it was inuential past its t ime.Te Chinese script is at the heart o Chinese identity, and it

appears that many ethnic groups in pre-Dynastic China adoptedwriting, and hence the language, becoming Han in the process. On theringes o the Chinese sphere o inuence, Chinese writing tended tobe adopted by the elite in order to gain the benets o literacy, withouttheir cultures becoming Chinese entirely. Te three nations on whomthis eect was most strongly elt were Vietnam, Korea and Japan; we


11/78

12 13

are phonetic, in concept, rom the very beginning. In a happy accidento history, all o the existing members o this amily descent rom ascript called Phoenician; though this word is unrelated to the wordphonetic, it makes thinking o the phonetic scripts as the Phoenicianamily a mnemonic thing to do. We are mostly concerned with that

amily here.Though the syllabary classes of cuneiform and hieroglyphsforeshadowed phonetics, the very invention of the alphabet may be

considered the rst achievement in the eld, making the others possible.The idea of vowels as distinct sounds which can be represented

explicitly developed slower, but in a similar organic, prehistorical way.We have no way of knowing who developed the rst alphabet, or howthe vowel system was rened into the explicit vowels of Greek or themodied-consonant vowels of Ethiopian and Brahmi script. We haveonly the archaeological record showing parts of this development over

time.

The next major achievement takes place during what I considerone of the great owerings of human history, Northern India duringthe Magadha and Maurya empires. The sages of this era are with usstill: Mahavira, founder of the Jain religion, Siddhartha Gautama theBuddha, and Patanjali and the other major gures in Yoga, were allof this age. Equal in stature to these rishis is the great linguist andgrammarian, Pnini.

Pnini adopted a rule-based approach to grammar, one basedon morphology and phonetics. His achievement, a grammar o theSanskrit language, was so thorough and ar-reaching that it became

not merely descriptive, but prescriptive. Beore Pnini, Sanskrit was ahigh dialect o the common tongue (Prakrt), preserved by the priestlyBrahmin caste, which linguists reer to as Vedic Sanskrit. Ater Pnini

we have the era o Classical Sanskrit, where or more than a thousandyears the proper dialect o the language was dened by the Ashtadhyayi,his description o the morphology o Sanskrit.

Te beginning o the Ashtadhyayi is called the Shiva Sutras,and conveys, in a compact orm, the phonetics o the Sanskrit language.Te signicant achievement here is that Pnini divides the sounds o

ully phased out, and relics do survive in place names and the like, atleast in the South.

Hangul has syllable blocks built up out o phonemes, whichin turn are built up rom strokes that represent what kind o phoneticutterance the phoneme is. Tis makes it a eatural writing system, and

the only one in current use to represent a real language. Tis is also oneo the most recent inventions to be in current use by more than onemillion people to represent their native tongue, as the ashion in thelast ew centuries has been Romanization. Tis is a real pity, becausethe Roman script is poor in symbols, with such symbols as it doeshave lacking amily order or eatural qualities. It may be hoped thatthis development was a eature o European imperialism generally, andthat with the decline o the Eurocentric paradigm we may see otherapproaches to writing gain currency.

Tis history is in no sense systematic; it is meant only to providesome background on writing systems generally, and more inormation

where appropriate on particular scripts that were inuential in thedevelopment o Phon. Parallel to the development o writing systemsis the development o linguistics, which we will explore briey in theollowing section.

A Brief History of Phonetics

In order to properly situate Phon in its historical context, weneed to discuss the history o phonetics, and specically phoneticwriting systems. Weve seen that there are basically two living amilieso writing, one through the Semitic (ultimately the Sumerian) and theother through the Chinese. Te Chinese, it should be noted, had a welldeveloped science o phonetics, and cultures on the rim o Chineseinuence (Korea and Japan most notably) developed scripts orrecording the phonetics o their own language rom a Chinese root. Tesprawling amily o scripts descended rom the proto-Siniatic, however,


12/78

14 15

English, on the other hand...Te creole nature o English, crated rom the interactions

between ruling Normans and their Anglo-Saxon subjects, has madespelling a hopeless mess rom the beginning. Tis evident act may be

why English speakers and teachers have been a rich source o writing

reorms since the seventeenth century. Early eorts such as FrancisLodwicks Universal Alphabet bear a charming resemblance to Phon,actually. Tere were a number o Enlightenment Era reorm proposalso this sort, which we will pass lightly over; in the United States,Benjamin Franklins reorm alphabet is perhaps the best known, andis a conservative, Roman-based alphabet with some inuence on themodern IPA.

By the mid-nineteenth century, several actors converged toproduce a urry o work in the eld. One o these was the movementto teach the dea to speak, through instruction in the articulationand placement o sounds, another was the aorementioned spelling

reorm movement. A third was the use o phonetics in teaching oreignlanguages, and a ourth was the rise in popularity o shorthand methodso writing.

Where education o the dea is concerned, the remarkable Bellamily are the standouts. Te patriarch o the clan, one Alexander Bell,

was an actor and orator who wrote several books on elocution. Hisson, Alexander Melville Bell, is the Bell o interest; grandson AlexanderGraham Bell invented the telephone, but not Visible Speech.

Alexander Melville Bell joined the amily trade, becoming alecturer on elocution at the University o Edinburgh. He married a

dea woman, which may have inuenced his lie work, the teachingo spoken language and lip-reading to dea-mute individuals. owardsthis end, he developed a writing system called Visible Speech, rstpublishing around 1867. Visible Speech was an attempt to make iconsrepresenting the shape o sound articulations directly, thus not merelyeatural but iconic as well. Tis system played a role in the developmento the IPA (as well as the Canadian Syllabics used to write some FirstNation languages), and certain structures turn up later in olkiens

work, so it merits a closer look.

his language into vowels and consonants, and within these makesamilies based on shared properties o sound. All nasals, or example,are grouped together, and all stops, in aspirated and unaspirated orm,are grouped together. Tus, due to Pnini, Indian writing systems havealways been organized this way, while the descendents o Greek and

Aramaic in the West remain in an arbitrary order to this day. AltoughPnini may not have even been literate, and i he was he used Brahmi,I have chosen to include a table rom the Devanagiri script, amiliar toanyone who has studied it . It shows the rank and le order othe main group o consonants, and careul observation will show someirregular resemblances between symbols on a eatural basis. It was thisversion o the basic organization introduced by Pnini which was mostinuential in modern phonetic development.

Pninis achievement, basic to high Sanskrit culture, remainedunknown, and unsurpassed, in the West until the late 18th century.Te discovery o the Sanskrit corpus at this time, and the translation

o Pnini into European languages, marks the beginning o modernlinguistics, due to the resemblance between Sanskrt, Persian, Greekand Latin which was thereby recognized, as well as the incorporationo the Indian tradition o phonetics and morphology into the Westernstudy o grammar.

It is an interesting acet o the phonetic science that every halehuman being has the apparatus to study it. Each o us has a mouth,lips, a tongue, a vocal tract and the means to construct statements inat least one language. As a result o this, basic acts about phonetics(such as the distinction between vowels and consonants) were widely

known rom prehistory, and various distinctions were invented multipletimes as thinkers independently came to the same, basically correct,conclusions about how sounds are produced.

Te spread o Latin writing across the whole o Western Europecame at some cost. Italian, which descends closest rom Latin, remainsclosely linked phonetically with its written orm, and Spanish does wellin this also. As we move north, we nd decent adaptations or theGermanic languages, and while French has somewhat drited, it toobegan with a decent t.


13/78

16 17

can be read, without overt strain, by native speakers o English, andrepresent phonetics precisely. While his reorms never caught on orgeneral reading (they never seem to, with the notable exception o

Webster and his ipped res and dropped us) they were inuentialon the IPA, and in this sequence o alphabets, a detailed pictured o

the development o phonetic thinking in the mid-19th century can beseen.Henry Sweet was a noted philologist specializing initially in

ancient Germanic languages such as Icelandic and Anglo-Saxon, and astudent o Bells writing system. He became an advocate o scientic,phonetic alphabets or teaching English abroad, initially working witha modied orm o Visible Speech he called Organic Speech, and laterbecame an advocate o spelling reorms in the vein o Pitmans latereorts. His modication o Visible Speech did make it easier to use,and he retained an interest in its use throughout his career. But theexpense o maintaining entire typesets, and the ease o acquisition o

the Roman based systems, kept it only modestly successul, used inearly publications by what became the IPA.

Sweet was an inuential member o the International PhoneticAssociation, which was ounded as Dhi Fontik cerz Ascicon in1886, in Paris. Originally concerned mainly with the teaching o Englishas a oreign language using phonetic alphabets, the Association cameincreasingly to be concerned with one goal: the standardization andpromulgation o a single, phonetic alphabet, derive rom the Romancharacter set and useul in the phonetic realization o all languages.Tis development is chiey due to Otto Jesperson, who also devised

the notion o an International Phonetic Association proper. It tookonly a ew years or phonetic concerns to become the chie ocus o thejournal, which was published or many decades entirely in phoneticscript. In early years, various phonetic scripts were introduced andcompared, and Sweet maintained a sort o Organic Speech or the useo the publication, but over time the Journal, and the IPA, settled onthe alphabet now considered the standard in linguistic and phoneticscience.

Te International Phonetic Alphabet is considered in detail

Visible Speech divides consonants into our places oarticulation: lip, point, top and back, reerring to the part o the tongueused to articulate each. In modern thinking we are more interested in

which region o the vocal tract works with the tongue or on its ownto constrict the air ow, and would call these categories labial, dental,

palatal and velar. Due at least partly to the need to cast type or eachunique shape, an expensive and labor-intensive process, Melville Bellchose to rotate his symbols in each o our directions, one directioneach or the our major places. Tis has many consequences, once o

which is that the consonants are all the same length and width, makingthem hard to distinguish, like reading something written in Romancapitals.

Manner o articulation is shown by changing the basic shape invarious ways: closed or a stop, closed with a wiggly line or nasal, a barseparating the stroke or voicing, etc. Te vowels are essentially sticks,

with small modications (hooks and cross bars) to indicate degree o

stricture and placement in the mouth. Tere is a lot more to it, butmuch o the phonetics is idiosyncratic or obsolete; in any case, enoughhas been said here to urther the story.

Isaac Pitman is at least as well known in shorthand circles asMelville Bell is in the Dea education community. Moreover, his systemis still in widespread use, unlike Visible Speech, and is the most widely-used system in the UK and Commonwealth communities to this day.His shorthand was the rst phonetic shorthand, and predated VisibleSpeech by some thirty years, being introduced in 1837. He inventedthe word Phonotypy to describe writing systems where the shape o

the character is related to the aspects o its pronunciation, what wewould now call eatural, and his Phonotypic Shorthand was a greatsuccess, employing a great economy o line and exploiting the ability othe pens in use to draw a thick or a thin line quite easily.

From that basis, he turned his attention to spelling reorm,devising many alphabets in the ensuing decades. Te earliest o themresemble the shorthand quite closely, but in later systems he concluded(as Benjamin Frankin did a century beore) that a modied Romancharacter set was required or widespread adoption. His later eorts


14/78

18 19

o what language should sound like. Tey draw visual inspiration romthe Uncial hand, particularly the Insular development seen in the Booko Kells, just as Quenya and Sindarin drew much o their sound qualityand morphology rom Finnish and Welsh. In ormal structure, however,

we see a continuity with the old iconic school o phonotypy.

As a proessor o Anglo-Saxon, olkien must as a matter ocourse have been amiliar with the works o Henry Sweet, who was alsoa noted Anglo-Saxon scholar, and must thereore have had amiliarity

with Visible Speech. It is perhaps not surprising, thereore, that wecan see parallels between Visible Speech and olkiens writing systems.Te our columns o the main consonant grid, called the tmar,correspond closely to the our places o articulation in Visible Speech,and the tyeller, the manner o articulation, have a similar arrangment.Te engwar are conceived o as meta-letters that are assigned valuesdepending on language, but in practice the assignments are amiliarto a student o Visible Speech. Te strokes are doubled to orm voiced

consonants, which is a similar visual eect to the dividing bar used inVisible Speech to represent the same distinction.

engwar ollows Visible Speech in distinguishing vowelsvisually rom consonants, but does so through tehtar, which areessentially diacritics. Tese are placed above or below the consonants,as one pleases, with dierent ways standard or dierent languages.Te consonants that dont t the classication scheme come in non-standard shapes, that continue to be ormally related to the Insularhand. Te engwar lay close to the heart o olkiens mythology: theinventor, Fanor, later went on to orge the Silmarils, which are the

central pivot o the entire legendarium. It was designed only to pleasehis particular sense o the beautiul, and succeeds, I daresay, in theesteem o most.

Luigi Seraphini takes his own sense o the beautiul andarguably goes even urther with it. An Italian graphic artist, Seraphiniis best known or the Codex Seraphinianus, a monumental work in theorm o an encyclopedia. It is richly illustrated, and entirely laid out inlonghand, eaturing a writing system which has never been decipheredand which may not prove decipherable. Te numbering system, it

elsewhere. Here it is enough to note that it was a great success, developedthrough the process o academic consensus in a paragon o the 19thcentury style. It was possible to type it with a little creative overstrikeand modication, and every printer o academic subjects had a sort othe type on hand. Such later oerings as Shavian sank without a visible

trace; Phoneticians and academic linguists used the IPA, and everyoneelse stuck with their native orthographies, and that was that.It is tting, then, that the next development was a hobby, albeit

the now-amous hobby o a distinguished linguist and author. J.R.R.olkien is best known as the author o Te Hobbit and Te Lord o theRings, but in his proessional lie he was a proessor at Oxord, initiallyin Anglo-Saxon. His proession was linguistics, in other words, and hispassion was inventing languages and writing systems; he has claimedmore than once in his writings that the entire world o Middle Earthand the stories written there are to provide background and history orthe languages, writings and lore that were his abiding interest.

olkien developed two Elven languages, Quenya and Sindarin,which reected (among other things) his great love o Finnish andWelsh, respectively. He designed several ways or each to be written:a mode using Roman letters, which is the one he actually used most,a runic orm called Cirth (not urther discussed here), and two scripthands, Sarati and engwar. O the two, Sarati is older, both in thelegendarium and in historical chronology, and Sarati has intriguingtraces o the orm ound in engwar and later Phon. In particular,olkiens elves were ambidextrous, and Sarati could be written inthree (arguably our or ve) directions. Certainly, a vertical and two

horizontal modes were present.It is engwar which is associated in the public imaginationwith the Elvish language. Te inscription on the Ring is written in theengwar, and in the lm adaptation most o the additional Elvish is inengwar also. engwar is also the only script Phon may airly be saidto be descended rom, though its debt to the IPA is also certain. So its

worth exploring in a bit o detail.Te engwar reect olkiens personal aesthetic as to what

writing should look like, just as Quenya and Sindarin reect his sense


15/78

20 21

should be noted, has been worked out with some satisaction, and onecannot read (i thats the word) the Codexwithout getting the sensethat theres at least some meaningul text in there. It could be entirelydecoration, but i so, it does an incredible job o imitating a language,

with recurring words and the like. Te Codexwas completed in 1978,

and I encountered it as a young college student in the late 90s and ellin love. Te sense o the mysterious that browsing the Codexinvokes inme to this day was a major inuence in directing my attention towards

writing as an aesthetic and esoteric pursuit.One week in October o 2006, I developed the core elements o

the Phon writing system. Ill explore this process in detail in the chapteron the orms and morphology o Phon. Our next task, however, is toexplore Phon in the context o English, the language in which this bookis written. We have overviewed the nature o Phon, and established thehistorical context o the script; now lets take some time to understandsomething about phonetics, phonemics, and how to write and read

English using Phon.


16/78

22 23


17/78

24 25


18/78


19/78

Phon in (and or) Plain English

Phon can be approached rom several dierent angles, moreor less useul or dierent backgrounds. For people accustomed tocomputer programming or mathematics, the ormal underpinnings othe writing system itsel are probably o greatest interest, and ChapterIII ld b th b t l t t t F th ith di i


20/78

28 29

claim, o course, but not ar o. It is emphatically not the case orphonemes. f;, or example, is represented by in all, ph in Phon,gh in enough, while t is ound in think, t tall, that, motion, and silent match.

Some alphabets have a very regular correspondence between

letters and the phonemes they represent, with Italian oten cited as anexample. English is not at all a regular language, with many spellingsreecting old pronunciations which have shited, sometimes multipletimes, since spelling was rened into a standard orm in the 17 th and18th centuries.

Phon was designed to provide a symbol or every phoneme, inany language, and it can used to write English easily. In act, i one canimagine or a moment a native English speaker who not only couldntread but didnt know the letters by shape or name, it is easy to see thatthis person would nd Phon much easier to use. Picture trying toexplain the use o the consonant c, which has the same consonant

sound in its name as the consonant s and which is mainly used eitheras an s sound or as a k sound, or why g and j have the sameconsonant sound in their names when g is used or (among otherthings) and d.

Phon is phonemic, which means that the phonemes o in thiscase English are each represented with a single symbol. It is also eatural,and this means that the shapes o the symbols corresponds to parts opronunciation. One part o a consonant shows where in the mouth itis articulated (made by the tongue), and another part shows how: orexample, whether the airstream is completely blocked by the tongue(a plosive or stop), whether the airstream is directed into the sinuses(a nasal), a narrowing producing a turbulent hissing noise (ricatives)or a narrowing producing a consonant sound but without turbulence(approximants). English examples o these phonemic categories are t,n, s and j: the last is pronounced y in yellow, the rst three asound in their names tee en ess.

Phon aims to be perectly phonemic: that is, to provide a uniquesymbol or every meaningully distinct dierence between utterances inevery language. It does not aim to be perectly eatural, but it is highly

thing. Consider how little meaning is contained in the written sentenceI cant believe it!. We have no idea i the speaker is being serious orsarcastic, which would be opposite meanings. Te distinction is carriedin the intonation o the sentence, which is not written out in English,though the exclamation point serves as a guide to intonation just as the

question mark does.Words are even more interesting, because we discover whenwe listen careully to speech that words are not a part o the audiosignal. Sentences and phrases are set aside by rhythm, breath, and tone:a sentence in English will begin aster and higher in pitch, and breathsare taken only between phrases, with larger, deeper breaths at the endo sentences. Syllables are easily recognized by the absence (really amuting in many cases) o sound between them. Words, however, arecognitive: someone listening to English who didnt speak it could markout the syllables but wouldnt know where a word began or ended.someone listening to English has eight syllables, and someone naive

o English could tell you that, but couldnt tell you that it has ourwords. Te act is that stress, which is a property o syllables, has asemi-regular pattern within words in English which makes them easierto pick out, but a language like Japanese lacks this eature, relying onshits in consonant (e.g. rom a phrase-initial k to a g in the middle oa word) to distinguish words.

At the lowest level, we have dierent words or speech andwriting. In English writing, the atomic unit is the letter; in Englishspeech, as in speech generally, the atomic unit is called the phoneme.Te term alphabet, in the broadest sense, is used or writing systems

where the undamental unit (called a grapheme) represents phonemes.I the graphemes represent syllables, we have a syllabary; i they represent

words, a logographic system.Tere are good reasons to not just call phonemes a letter, as

we call text and speech words with the same word word. A sentencelike anyone who knows it knows it has six words and eight syllables,in spoken or written orm. Tis same sentence has twenty-three lettersand seventeen phonemes. Tere is, or practical purposes, a single

written word or each spoken word in English. Tis is an idealized


21/78

30 31

o this type we get a voiced bilabial plosive, , pronounced b,and representing the b in bin. Tese two words orm whats called aminimal pair: they dier only on the basis o a single phoneme. Minimalpairs are important in sorting out what the phonemes are or a givenlanguage, because there are a variety o phones (actual sounds uttered

by a speaker) which are all considered the same sound by particularlisteners. English speakers, or example, cannot typically tell the p inpen rom the p in spin, but the dierence (the rst is aspirated in thesecond is not) is contrastive in Hindi; there are words that dier onlyby whether or not the p is aspirated. Similarly, the distinction betweenl and (the sounds o lit and rit) is not made by Koreans, who

would not detect lit and rit as dierent words.Te next sirp is called the dental: it is assigned, usually, to

sounds which take place against the teeth or just above them. , anunvoiced dental plosive, represents the sound t in tin, while isthe sound d in din. Te palatal stem, which neither ascends nordescends, comes next; there are no palatal plosives in English, and onlyone actual palatal. Ater this is velar, and we have or cut and orgut, which are k and g in IPA terms; the place where these soundsare made is called the velar region.

Other languages, such as Arabic, have stops even lower in thethroat, such as the uvular stop heard in the word Quran, a q in IPAand a in Phon. For English, were done with the stops. Te nextgroup o sounds are the nasals, where the airow through the mouthis stopped but it is continued in the nose; while this is happening, thevocal cords are vibrating, making nasals voiced. Because nasals are all

voiced, we do not need to double strokes to represent them, and nasalsin Phon are symbolized by a spiral shaped lass, called a hwinya, in thesame direction as the stop lass. Te rst nasal, which is bilabial, is ,the sound m in bam; next we have the dental nasal or n in ban,both straightorward enough. Although were used to representing itng, the sound in bang is a distinct phoneme, articulated wherea g is but voiced like an n; we represent this as . Tere is a palatalnasal, , which is used in Spanish loanwords such as pion and hasthe IPA orm ; some accents (Cajuns or example) will render onion

so, and it can justly claim to be the most reguarly eatural writing systemin existence. Tis makes Phon letters (graphemes, really) easier to learn,because the shape tells you something about how to say them.

Consonants are easier or an actual reader o English to learnthan vowels are, because the consonants o our written language are

(comparatively) regular in use. A consonant in Phon consists o twotypes o strokes: a single vertical stroke called a sirp, and one or morerounded strokes attached to it, called lass. Te words mean stem andlea , and are meant to make you think o the shapes, as well as help youremember the order o drawing them: rst the stem, then the leaves.stem has a specic linguistic meaning thats dierent, and or thatand other good reasons well use our particular terms here, which aretaken rom J.R.R. olkiens Elven language Quenya. olkiens writingis ancestral to the Phon system, and this is done in homage to that, as

well as to emphasize the transnational character o the writing system,which is not based on any historically used scripts, but only on the

hobby o a distinguished linguist.Phon is careully designed so that let handed people and right

handed people can write it, in the same way but with opposite resultson the page. Right handed people write in the usual mode or Romancharacters, rom let to right, while let handed people do the opposite.Tis means that, instead o talking about let and right, we usually talkabout senseward and anti-senseward. Te diagrams in this book are,unless noted, in Rightic Phon, but were produced by writing LeticPhon and ipping it on the computer, on the assumption that letiesare used to reversing things and righties are comparatively helpless inthis regard. Tis means that senseward in the diagrams is towards theright margin.

Te sirp are organized so that ascenders point to the ronto the mouth and descenders point down the throat. Tis means thata sirp that is ascended reers to a bilabial, a sound made with bothlips. I we take the rst lass, which looks like a bow, and draw itsenseward on a bilabial sirp, we have the rst consonant in Phon, abilabial plosive, unvoiced: , pronounced pand equivalent to theIPA character p. Tis represents the p in pin. I we draw two lass


22/78

32 33

Rather than tag the dental sirp to mark variants (were savingthat option or later) we are going to use dierent lass to tell thesestrokes apart. O the three unvoiced options, the one that seems mostdirectly contrastive with our dental stop t is the dental ricative ,pronounced th in thigh, so we write it . Te voiced version, or

, is pronounced th in thy.Te next ricative well encounter is , pronounced s in sip,and its voiced counterpart , z in zip. Also, we have , IPA , thesh in ship, and or , the s in pleasure (p, sadly, is just notan English word). Tese are presented in phonetic order; because Phoncharacters are sorted according to the strokes theyre made o, ratherthan the sounds those strokes represent, words starting with thesesounds would be ound ater all other ricatives.

Te remaining ricatives are the rarely used , or a voicelessvelar ricative, a turbulent version o k. Tis is spelled x in IPA,and is ound in the Scottish word loch as the ch sound, and in the

occasional loanword rom other languages; many English speakerswould pronounce loch with a k. Much more commonly ound is theglottal ricative , pronounced h in hip. Neither o these is oundvoiced in English; although the h is sometimes actually voiced as or this is not contrastive, that is, English speakers hear the samephoneme and the h is voiced or not depending on the neighboringphonemes.

Te approximants come next, sounds where the airstream isnarrowed enough to produce a consonant-like sound, rather than avowel, but not so much that it becomes turbulent. Te r sound inEnglish is an approximant, but the symbol has already been given. Wealso nd the only palatal which is requent in English, the character, spelled j in IPA and providing the y sound in yes.

English also has a lateral approximant in the dental region.As mentioned beore, many languages such as Korean and Japanesedo not contrast between the two so-called liquids r and l. In thoselanguages, both phonemes and anything in between are allophones,sounds which carry the same meaning within the language. Tere arerules that govern which is used when, but these rules are not ones that

with this consonant.English has no true trills, taps or aps. Tese sounds are r like,

and are made by the tongue tapping once or several times against thetop o the mouth: or example, the alveolar trill r, ound in the Spanish

word perro meaning dog. Te r sound in English is produced in the

same spot, but in a dierent ashion, and is called an approximant. InPhon, however, it comes beore most other approximants in order; orthe reason see the discussion o sort order in Chapter III. So the nextcharacter used or writing English is the rhotic approximant , written in IPAese and pronounced like the r in rat.

Fricatives are next. Tese are sounds in which the airstreamis narrowed enough to make it turbulent and noisy, but not stoppedentirely. Fricatives, like nasals, are described as sonorants, which meansthey can be sounded or as long as one likes, such as the extended

written as Shhhh! which English speakers use to obtain silence ina crowd. Tis is because the ricatives do not disrupt the airstream

through the mouth entirely. Because the airstream is making a noise atthe point o articulation, ricatives can be both voiced and unvoiced.

Tere are a lot o dierent ricatives, more than any other typeo consonant. Tere is a bilabial ricative, or instance, but it is notound in English and would sound cartoonish i used. Te closest wehave, and much more common in languages generally, is a labiodentalricative. Tis is ormed by the bottom lip nearly touching the top teeth,and we have a variant o the bilabial sirp to show this articulation.Te unvoiced labiodental ricative is , pronounced f in at, and thevoiced is , v in vat.

Te dental sirp reers to a region o the mouth where no lessthan six distinct phonemes o the ricative type are ound, and all othem mean something in English. Tere are a ew reasons or this: oneis that the dental sirp also reers to the alveolar region, ound abovethe teeth along a ridge on the palate , another is that the tongue, bychanging shape, can change the turbulent qualities o the airstreamquite a bit. Our ears are so sensitive to this quality that we have aspecial word, sibilance, to describe it. A third reason is that English hasmore dental-type ricatives than many other languages.


23/78

34 35

long orm, which is proper but seldom to be encountered, is ort and or d. Tese are just the individual phonemes with astroke, called a alma, connecting them. Te short orms are and

respectively; the place o articulation o the ricative, as well as thevoicedness, are understood. I one wished, the short orm o ts can be

written , to show that it is an aricate, and also to make the word itsor its a single consonant long.o emphasize the phonemic nature o these aricates, they are

commonly written with entirely dierent characters. Phon oers anextension set o characters which do not have globally dened meanings:these can be associated. In English, t is ordinarily rendered , withd being rendered . You will note that these both take palatal stems;there are historic reasons or that, as well as the good reason that thereis only one genuine palatal in English, so using these mesograms oersvisual contrast. Te lass used in each are ound only in mesograms,making them easy to spot as such.

English VowelsFigure 2.1

Beat Boot

Bit Put

Beat Boat

Bet But

Bought Bot

Bat Batted

Batter Bird

a native speaker would need to know; indeed, one would only knowsuch rules i taught them, wheras one learns a native language withoutinstruction. Similarly, the rule as to when to aspirate a p in English isnot one most people who speak English can explain, and Id guess thevast majority dont even know theyre doing it.

We represent the dental lateral approximant , the l sound inlet. Te sirp show that it is dental, while the hook (which is a type o

stroke called a tuima) shows that the sound is lateralized, that is, releasedrom the side o the tongue. Approximants dont have a denite sideo the sirp the way stops and ricatives do, and this particular lass isound on the stop side; one good reason is that there is a lateral ricativealso, and ricatives are all to be ound on the ricative side o the sirp,or consistency.

We are let with a couple obeat consonants, which are banishedto the other symbols part o the IPA table or being so irregular. Tecommon one is , which IPA renders w or weather; this is a

voiced labial-velar approximant, made by narrowing both the lips andthe velar region. In some types o English, we have a second sound, rendered or wh in whether. I you pronounce both w, thatsairly common these days; make the distinction in writing or not, asyou please, but it does provide contrast between common words so itsuseul to include whether you say it or not.

Tose are the simple consonants used in English. Tere are acouple more sounds that we tend to think o as simple consonants,namely the ch in chin and the g in gin (or j in june). Tese soundsare aricates: they begin as stops, but release into a ricative at the sameposition. Te phoneme represented as ch is written t by the IPA,

with a tie bar added sometimes or clarity, but these are considered onephoneme in English rather than two. Consider: i one were to exchangechin or shin, one would not hear it as though a phoneme had beendeleted; and yet we would hear the names ynn and lynn as dieringonly by the absence o f in the latter. Te sound ts as in its is also anaricate, but is treated as two phonemes in English, because we treatit and its as diering by the addition and deletion o an s.

Tus there are several ways o writing aricates in English. Te


24/78

36 37

example, we have the approximant , ound in a word like red. Butwhat o the r sound in the word ower? Tis is assuming that onespeaks a sort o English in which there is an r sound at the end o that

word: much o the Anglophone world does not, including most oBritain and all o Australia and New Zealand. For other places, notably

North America, Scotland, and India, there is a so-called r-coloredvowel, represented in IPA as and in Phon as . Tis is considered avowel, but rhotacized, that is, the tongue is bunched in the way usedto pronounce the approximant but the air passage is not narrowedas much as or the actual approximant.

Another example o this border can be seen by pronouncing theword yellow while stretching the rst sound out, then pronouncingthe word evil. Te ormer uses the approximant j or the rst sound,

while the latter begins with the vowel i, however the sounds are eithernearly the same or identical. Te initial sound in yellow becomes anapproximant because the next sound is the vowel , and it becomes

a vowel in evil because the next sound is the ricative v. Saying thewords only yellow at speed will make this clear, as well as illustratinghow it is that the letter y came to be used or both purposes in Englishorthography.

In Phon, the consonants are written where an English speakerexpects to see letters, that is, along the main line and extended aboveand below it. Vowels, on the other hand, are written above and belowthe main line, either over/under a consonant or on a sirp o lass size,that is, a palatal sirpe. A vowel above a consonant is pronounced beorethe consonant, while a vowel below a consonant is pronounced aterit. Tis sort o arrangement will be amiliar to students o Hebrew,

Arabic, Farsi, Urdu, and to a lesser degree o Hindi, Tai, or languageswhich use related scripts. A student o engwar, o course, will be rightat home.

Te vowel i is as narrow in the mouth as you can get andstill have a vowel, as saying the word easy will illustrate: the tonguenarrows rom i to z and back to i, with no phoneme in between.Te vowel is also as ar orward in the mouth as it can be. Tereore we

write this vowel as , where the bow-like stroke shows orwardness

English Vowels in Phon

Te vowels o English are not as straightorward as ourconsonants are. Tis is partly because our spelling has only minimalbearing on which vowels are actually used. Te vowel structure oEnglish also changes considerably depending on which variety oEnglish is being spoken: or the purposes o this document General

American as spoken by actors and news anchors is the standard.Also, we are taught that there are short and long vowels,

which is true in some languages (Hindi and Finnish or examples) butnot so in English. Tere is a relationship between stress and what arecalled long vowels, but length is not a actor. While we have ve lettersto represent vowels, there are between ten and twelve (varying withregional accent) vowels in English, and they dont correspond neatly to

the vowel markers.What is a vowel? In school, we learn a, e, i, o, u, and sometimes

y and w, but these are o course letters, not phonemes, and ough isalso sometimes a vowel (with a s ilent ugh, which is not an uncommonreaction to English spelling). We are interested in what a vowel is as anutterance, because when we understand how vowels are sounded we canthen understand how they are written in Phon, and more importantly,

why.Earlier, weve talked about dierent kinds o consonants:

stops, where the airow is entirely interrupted, ricatives, where theairow is narrowed enough to cause turbulence, and approximants,

where the narrowing is not enough to cause turbulence. Each o thesecorresponds to a more open air passage than the one beore it: when weopen the air passage urther, we produce a vowel. All vowels are voiced,by denition, and all vowels are also sonorants: the vocal cords vibrateand the sound can be maintained or as long as theres breath in thelungs.

Te approximants lie on the border between consonants andvowels, and it is not always possible to distinguish between them. For


25/78

38 39

in Phon, we would draw the vowel above the approximant . wo,however, are considered rhotasized vowels. Tat is to say that the wordbird is typically rendered bd; the middle character is a distinctvowel with an r quality, not a change rom a vowel to a dierenttongue position as the orthographic spelling suggests. Te other vowel

is used to render butter as bt

, and is a reduced vowel. Tese arerendered in Phon as and , where the little tail on the shows that itmodies the vowel above.

Te shwa , a simple dot by itsel: , is used to render thereduced vowel ound in e.g. the end o the word roses. In mesographicuse, the shwa can also be used or the reduced vowels at the end obutton and bottle. It should be noted though, that these wordsinvolve an unusual airstream mechanism, where instead o a vowel thestop ends when the air either goes into the nasal cavity (button) or isexpressed laterally (bottle). I these special airstream release charactersare drawn instead o the vowel, they must be placed under the stop

consonant, not above the lateral or nasal. Te lateral release character isand the nasal release character is .

Writing consonants is perectly straightorward: they arewritten one at a time, in the senseward direction (let to right or righthanded writers), and they come in only one orm. Vowels, in Phon, are

written as diacritic marks, above and below consonants; in Phon, thesemarks are called cerm, which means grain or harvest. When a vowelis ound alone, it is put onto an unextended sirp, what we call a barepalatal sirp: dipthongs, as shown above, are written above and belowthis character.

Te vowels as written here are p-type, that is they are shownas they would be drawn above and below a plosive consonant, orany consonant with the lass in the senseward direction. I you tryand write, say, the vowel on an -type consonant, you will run intoa problem: the thick stroke indicating mid-closed will get in the wayo the bottom extension o the sirp, assuming the consonant has one.For that reason, and others discussed in the ormal chapter, the vowelsare written dierently or -type and p-type consonants, as well as tomake it easier to read Phon in both directions. Te rule is that the

and the upright stroke indicates a closed vowel, one that is as narrow asit can be without becoming a consonant o some sort.

In English, there is no contrast between rounded and unroundedvowels, that is, there are no vowels which dier only by how roundedthe lips are. Some English vowels are rounded, some, like , are not.

An example o a rounded vowel is , which is u to the IPA andpronounced like the oo in boot. Te swirly looking shape symbolizesa back rounded vowel; i it was back unrounded, it would be like anupside down version o the bow stroke in .

Most o the vowels can be remembered by putting thembetween b and t, as beat and boot show. Tis pair have two similarvowels that are slightly more open: , written and said bit, and

, written which must be remembered with put. More open thanthese are , written e and pronounced bait, and , written oand pronounced boat: note that the latter is a common use or o inEnglish, but that e is seldom used that way (or should i say we which

is certainly not pronounced wi; IPA can get conusing when appliedto a language that uses Roman characters already).

Next we come to the mid-opens, , spelled and pronouncedbet, and , spelled and pronounced but.

Some accents o American English distinguish between ,spelled and pronounced bought, and , spelled and pronouncedbot; i you dont distinguish these sounds, just use . Teres alsoa slightly less open, orward vowel, , spelled and pronouncedbat.

Tese are the main vowels o General American English; thereare also the three dipthongs: , spelled a and pronounced eye,,spelled a and pronounced like the exclamation ow!, and , spelledo and pronounced like the interjection oi!: or, i you preer, theresbite, bout and boy.

English accents are split between what are called rhotic andnon-rhotic accents. Te English, Australians and New Zealanders,

would not pronounce the r at the end o New Zealanders, whereasAmericans, Scots and Indians would. Most o these r sounds renderedin IPA by ollowing the vowel with , which is what we might expect;


26/78

40 41

more like the end o the word hung, but as -ing is so common inEnglish and is practically unknown (hung is properly renderedh), the switch saves eort in writing and is essentially a contraction.Te mesograph or and, , is a contraction o the n and d othe long-orm. Tere is also a symbol that represents the words othe, which is a contraction o those mesograms and rendered ; thisis perhaps somewhat anciul. but ollows clearly rom the existingmesograms (which are useul) and is included as an homage to olkienamong other reasons.

Tats how its done; the key is to practice, particularly thevowels. A sample text in English is given in the appendices, or readingpractice. We hope to publish a journal and blog in Phon, as soon as

we can get any sort o ont together. Tis chapter explores the howo writing English using Phon, but not why one might want to doso, or how Phon came into existence. Our next task is to explore, inconsiderable detail, how Phon got to be the way it is and what this

reveals about the underlying structure o the writing system; in otherwords, the orm and morphology o Phon.

closed mark and the mid-closed mark are both drawn away rom thesirp, with the other part o the vowel next to it; while the mid-openmark is always drawn next to the s irp. Vowels by themselves are alwaysrendered in the p-type, above a bare palatal sirp.

Phon aims to achieve several goals: one is to provide a tool orphonetic and phonemic notation, and another is to provide a practicalscript that can be used by speakers o the worlds languages in order to

write them. Tese are complementary goals: certainly i one can renderan acceptable amount o phonemic distinction, one can use it to writelanguages. Spoken utterances have more inormation in them than istypically written down. Some languages, such as English and Spanish,have stress: in Spanish this is indicated only when it varies rom therule, while in English it is never indicated at all.

In Phon, we distinguish between arcograpic and mesographicuse o the writing system. Arcographic use encompasses both broadand narrow transcription in the phonetic sense: it is concerned with

accuracy, use o a single invariant and international standard, andcompleteness o rendering. Tere is a continuum between arcographicand mesographic use, and between meso- and basigraphic use, whichis using Phon in a way that isnt recognized as valid by the Foundation,or which is simply incorrect in some way.

Te idea is that Phon will be modied in small, systematicways, to make it easier to write and read various real languages. Tesemodications will be part o the ofcial reerence standard o the Phonlanguage. An example weve already encountered is the aricates: theycan be written as two consonants with a alma connecting (the mostarcographic choice), in the short orm (still arcographic), or as the twomesograms and . . Tese characters show that t and d aresingle phonemes, while adding visual contrast to the script particularlyby inhabiting the sparsely-populated palatal region instead o thedensely packed dental.

Tere are several cases where English is contracted by implyingthe vowel with a shwa, such as or the, or o and or theending -ing. It should be noted that, while the proper pronunciationo the and o is close to that implied by the spelling, would sound


27/78

42 43


28/78


29/78

Te Form and Morphology oPhon

Letters are strangely invisible. We see them every day, thousands

o them, o course, but the act o seeing the letters o a word invokesthe word, the sentence, the meaning. We gaze at letters but we see


30/78

46 47

Phon, and or the same reason. But its heart is the idea that there are 15principal transormations which are ound at all levels o organizationin nature, and that these and only these create good structure, livingstructure. By identiying these transormations, and nding exampleso them in good structure (both natural in the limited sense and inthe artiacts o humanity), Alexander provides a system or buildinganything at all, and having it come out beautiul, healthy, and vibrant.Tough I hadnt read Te Nature o Order when I designed Phon, theprinciples are ound in nascent orm throughout the earlier works, andI will reer to them here.

Phon, as stated, began as a simple calligraphy project: i wastired o smearing ink, and generally elt constrained by Latin-basedcalligraphy. I had already tried writing English backwards, using Hebrewor Arabic script to write English, and writing engwar backwards inone o the modes or English writing I ound on the Internet. None othem were what I was looking or: it became clear that I was inventing

something new.Te rst, undamental insight was in Appendix E o

the Lord o the Rings, where olkien explained the logic behind engwar.engwar has a main body o regular characters, all consonants, wherethe vertical stroke (called elco) indicates the method o articulationand the horizontal strokes (the Lva) indicates the place o articulation,and are doubled to indicated a voiced consonant.

I remember eeling the hair on the back o my neck bristleas I thought about this. It echoed a similar electric eeling rom yearsago, when I briey studied Sanskrit, and rst saw the neat ranks whichorganize the Devanagiri script. Except in this ranking you could seethe shape o the rank when you took the grid away! In Sanskrit, andthe IPA, there is an ordering on the basis o eatures o sound, but inengwar, you could see that organization in each o the rank and lecharacters. Tis was exciting.

Note that this germ, this kernel, is still at the heart o Phon. Terst insight was the correct one to generate the whole structure, whichis transormed rom there. I next asked mysel why engwar was notitsel what I was looking or. I came up with a ew things. First, I didnt

we created. Te great antiquity o the Latin alphabet contributes tothis eeling; the invention was lost in time, and custom has given oursymbols a weight, a gravity, that is so pervasive as to be hard to see. Itis as though one pointed out that we wear clothing, or use utensils toeat ood.

Phon is as it is out o a ascination with the orm o writing.It has a deep ormal structure whose shape arises rom a process oelaborating and rening shapes, the shapes a pen makes on paper. .Each stroke in Phon takes a particular shape, and that shape has adenite relation both to the other shapes o Phon and to the meaningconveyed by the stroke. Tis happened because Phon was designed ina particular way, which allowed a simple, supple pattern to elaborateinto something which can encompass the entire phonemic state space.

Phon can best be understood by understanding how it cameto be. Phon began as a simple project: I wanted to be able to writebeautiully, with a ountain pen, without smearing ink on the page, and

without having to orient the page sideways or upside down. Troughsuccessive evolution it became something at once simple and intricate,beautiul and powerul. I dont believe Phon can really be grasped

without understanding this process.Phon is a work o synthesis, which builds on the work o

many. In orm, o course, the greatest debt is to J.R.R. olkien, whosework pioneers many o the concepts and orms ound in Phon. Teprocess, however, owes the most to Christopher Alexander, a Berkeleyproessor o architecture. He has devoted his lie to understandingsomething simple and proound: how lie is ormed and nurtured inspace and time. He is best known or two books, A Pattern Languageand Te imeless Way o Building, which between them describe asweeping vision o how to organize human spaces to serve the greater

wholeness o humanity, society and nature. Tirty years later (in 2003)he published a magnum opus, Te Nature o Order, which generalizesthese insights into a ull understanding o how patterns are elaboratedby transormation into all living systems, be they sand dunes, coralrees, or mammals.

Alexanders work is resistant to summary in a similar way to


31/78

48 49

Place of

Articulation

Doubled

for Voice

Manner of

Articulation

First shape o the Phon patternFigure 3.1

We arent quite done with gleaning wisdom rom olkien,however. His vowels, the ehtar, are ound above and below thehorizontal (lva) stroke, and I liked that. Among other things, vowels

dont clearly t the place/manner scheme, which is accurate only orconsonants. I took a cue rom olkien, as well as rom Devanagiri,Hebrew and Arabic, and put the vowels above and below the consonants.Te new pattern is consonants as described earlier, with vowels aboveand below. Te shape o these vowels is completely undeterminedat this time, and ended up resembling engwar not at all, other thanplacement.

One more thing struck me about the engwar; there wassomething about their shape that was subtly o to me, not good insome way I couldnt quite place. I thought about it or awhile, andeventually I realized I didnt like the characters where the lva is closedo by a bottom stroke. I resolved that my new script would have nospace-enclosing strokes, such as the letter o or the oending engwar.Tis adds no enclosed space to the earlier pattern, without disturbing

what comes beore.Id like to ocus a little on what prompted this decision. It is

on the ace o it aesthetic, and we tend to think o aesthetic choices asbasically arbitrary. It is tempting to think that I didnt have to makethis choice at this point, or that I could have made another choice and

like the irregular consonants; although I thought they were beautiul,they didnt t the crystalline sense o beauty I was developing. Teydidnt resonate with the core insight about symbolizing articulationand place. Every consonant has an articulation and a place, so whysymbolize them in some and not in others? Devanagiri, which is highlyreadable, has an invariant stroke, the top bar, or each consonant, anda common vowel stroke or most instances also.

Next, I realized that I needed to switch the basic axis o thesystem. olkien puts method o articulation along a vertical axis andplace along the horizontal one; he then complicates matters by puttingvoicing, clearly a matter o articulation, along the horizontal axisalso. So I switched them: the vertical stroke would symbolize place,and the horizontal strokes method, o articulation. Tis also accords

with certain eelings about the voice, where the h sound is below thek sound which is below the p sound in physical space. Also, thearticulation o a sound involves a pu o breath along the horizontal

axis. I resolved to maintain the doubling or voicing, because it wassimple and beautiul.

Tese two decisions alone radically simpliy the pattern whilemaking it much, much sharper. In the rst iteration we have no shapeat all, just the idea: Something like engwar where the strokes indicatethe sounds, with place along one axis and type along another. Now

we have a writing system where a vertical stroke denes the place, andhorizontal strokes the manner, o articulation o a consonant, wherea symbol is doubled or voicing.. Tis is a denite shape, and again,this shape is ound intact in Phon as it exists, just as the very rstspecication is accurate.


32/78

50 51

At this point I was done with engwar, but note: just these ewextracted kernels give a denite amily resemblance to the two systems.Tis is the power o doing structure-preserving transormations ona pattern. Note that looks like engwar as a guiding principle wasstripped rom the pattern between the rst and second iteration, butthe result still contains the realization o that desire. I I had merely setout to make something that looks like engwar, drawing characters oneat a time in an eort to make something at once dierent and the same,the result would be a derivative mess. But by isolating a ew patternsrom the engwar, and adding them to the earliest iterations o thePhon pattern, the resemblance is obvious to the untrained eye.

I just said I was done with engwar, but this is not quite true. Iwas down to a small number o symbols, 12 as it happens: those whichhad a telco and a lva, with no closing stroke. In a real sense, thesetwelve (with dierent meanings) are still ound in Phon. It was at thispoint that I had an insight which sent me deeper into the structure o

the evolving system. I realized that all strokes on the let o the telcoopened down, and all those on the right opened up. Tis made it clearthat one o the reasons I didnt like the closing strokes is that theyobscure this asymmetry! Why did this matter?

Eventually (and this is still within the rst week o workingon the system) I got it. We have two eyes, arranged on the horizontalaxis, and optimized or objects with a vertical axis o symmetry. Wealso have two hands and the natural way to write with each o them isaway rom the writing, so that the hand doesnt drag across the ink andso that what was just written is visible. By eliminating any symmetryacross the vertical axis, Phon could be written in either direction, withidentical motions; urthermore the shape o the characters would show

which direction the sentence should be read. For the rst time, I wasdesigning something useul to all o the population instead o roughlyten percent.

More than that, though: I had applied enough constraints togenerate a basic palette o strokes. When the project began, the letterscould be literally any shape; now, we have a pretty denite shape. A Phonconsonant is made o strokes along the vertical axis, corresponding to

still produced something like Phon. Tis would be a mistake; thisdecision was made rom a deep sense that the unenclosed charactershad a Quality, a living nature, that the enclosed characters couldnttouch.

Which is a better picture o the sel?Figure 3.2

Alexander gives an exercise which is useul here. Look at thesetwo shapes, a circle, and something like a circle that is nonethelessopen, and ask: which o these is a better model o my own sel? I amconvinced that the open circle is the answer, and that this answer isnot arbitrary. In a moon celebration circle I used to participate in, we

would say the circle is open and not unbroken, and this same wisdomis expressed in choosing not to enclose space in Phon. Lastly considerthis enso, a Zen painting o a circle, how simple and right it is.

EnsoFigure 3.3


33/78

52 53

Strokes must begin and end on a thin lineFigure 3.5

At some point I realized I had another palette-limiting rule onmy hands, which is that all strokes must begin and end on a thin line.I was using an angled pen, which made lovely thick strokes, and I likedthese strokes to taper o at both ends. It passes the mirror o the seltest, as these examples show. Tere is, at a basic level, a sense thatthese sorts o sinuous shapes taper: it is the shape o a vibrating string,o the body o a worm or snake. As I rened it, the strokes becamebolder, more at right angles to each other and more denitely at 45degrees to the page orientation.

I was on the verge o urther insight here, and trying to reworkthe vowels, I hit on it. I wanted the vowels to be in a particular order,and to represent the numerals as a result. I realized this was a poorapproach to the vowels, but that the basic strokes could be ordered,and that in so doing the numerals could be generated. Tis was thedeepest, undamental, ormal work; this done, the system emerged,combinatoric, powerul, and ready to go.

We begin (and this is important) with the simplest symbolyou can make, a simple dot on the page, moved only enough to orma diamond shape. Tis is symmetrical, in principle, in all directions,

and is the only exception to our rule o asymmetry across the verticalaxis. Tis dot is made with either hand, by a relaxed human, sitting orstanding in ront o a surace, with an angled pen, that is, one where thetip is wider in one direction than another, rather than a stylus (round)or a brush. Tis is called elen, the Quenya word or star. It is our rstprimitive.

place, and the horizontal axis corresponding to manner o articulation,with a doubled stroke or voicing. A vowel is a mark above or below aconsonant. None o these strokes may enclose space, and they must allbe asymmetrical across the vertical axis. is the pattern used to generateeverything that ollows.

VowelsPlace

Shows VoiceNo Symmetry

Manner

Symmetry OK

No EnclosingSpace

Te Phon pattern ater more iterationsFigure 3.4

At this stage I had enough to design several characters and I didso; I also hacked together a vowel system which didnt satisy me, but

which did let me move orward. In studying what it would take to writeEnglish in a rational way with the new system, I became reacquainted

with the IPA and resolved that the system (I was calling it Scriptic atthe time) would implement the IPA scheme o distinctions. Ater all,

this was the consensus verdict o the eld o linguistics, and had beenhammered on or more than a hundred years. In assigning symbols,I drew the telco/vertical symbol like in the version o engwar thatappears modeled on Italic rather than Uncial, which is more like an sthan like an l, and pulled the lva shape o that. I developed spiralsand wiggles as variant lva without any particular plan in mind, they

just looked right. Te vowels werent working though...


34/78

54 55

o the pen and two right turns, and thus symbolized LL and RR. Tis isdescriptive, not proscriptive: it denes the symbols in the specic caseo a let-handed generator o symbols, rather than trying to set rules orstroke ormation in Phon itsel. Te ormal system is dened in relativeterms rom the Letic mode simply because I am let handed, and ndit easier to think this way without headaches and their attendant errors.

At some point a more neutral, mathematical description will be needed.In the meantime, please realize that, as the diagrams shown are rightic,they actual penstrokes go in the other direction rom that indicated bythe words right and let.

Figure 3.7

Our sixth is like a at telco, because we start with our thinupstroke, bring it down as a thick stroke, and bring it back up thinrather than curling it back on itsel. I we start with a downstroke

we get the amiliar vertical telco. We call these alma, or wave, andnr, ame. Tese our symbols are the primary combinations, ormedo three segments with two turns, alma is LR and nr is RL. Teremaining basic palette are combinations o these, consisting o vesegments with our turns.

Figure 3.8

I you combine two lva type strokes, you get a spiral sort oshape. Tere are our possible spirals which embody the combinations

Figure 3.6

Te next stroke is a thin one, made with the thin edge o thepen. Tis can be done either upwards or downwards, in principle. Wedecree upward as basic, because o the stroke that ollows: a thick one,made with the thick edge o the pen. It is more natural to pull a penalong the thick edge, less natural to push it. So to give these two strokesmaximum contrast, the thin one is considered an upstroke and thethick one a downstroke. Because either stroke can in act be written asan upstroke or a downstroke, we name them amban (up) and pend(down), words rom Quenya which mean similar things. Both are at/4 radians (45 degrees) to the page, with the pen, thereore, held at

/4 radians also.Pend, we notice, breaks the rule about symbols beginning and

ending with a thin stroke. It is the only symbol which will do so. It iselt that a simple straight line has an honesty about it which needs notaper, but this simple beauty is easily marred, and I was later to discoverrestrictions in the use o pend that are needed to keep the overallstructure intact. Also, one might see amban and pend as reections oeach other across the vertical axis, but one is thick and the other thin.Because we dont always use angled pens, however, this act is kept inmind as the system is evolved.

Te next most simple shape is the type called lva in engwar.I you bend a thin stroke into a thick one, and then curve around toend it thin, you get a

Documents

Samuel Putman - A Universal Writing System