[Journal of the Simplified Spelling Society, 1993/1 p29-33 later designated J14]
[See Journal and Newsletter articles, Pamflet 15 and Cut Spelling by Chris Upward.]

Chris Upward revews

AgiliWriting - the readable shorthand of the English language.

Anne Gresham: AgiliWriting - the readable shorthand of the English language, London: Agilityping Ltd, 1990, ISBN 1-872968-00-7, £9.95. This revew is ritn in Cut Spelng.


Th AgiliWriting self-tuition handbook explains, and givs practis in using, a new shorthand systm based entirely on th letrs of th roman alfabet. It is by defnition therfor of direct intrest to spelng reformrs, ofrng as it dos an inovativ way of riting english that has a numbr of practicl advantajs.

Th first advantaj is that, as required for any shorthand systm, its spelng requires far fewr caractrs than th Traditionl Orthografy (TO) of english. Wile Cut Spelng (CS) saves only around 10% of th letrs used in TO, AgiliWriting (henceforth AW) claims a saving of som 40%.

Th secnd advantaj is that, unlike non-alfabetic shorthands such as Pitmans, Gregg or T-Line, AW shud in principl be readbl by anyone litrat in english. How esily it can be red in practis by th uninitiated may be jujd from th foloing sampl from th bak covr of th book:

W hv plzr n sndg u detls v th gds on spzl ofr untl th end v ths mnth;
but even if it is a strugl to read, AW is clearly accesbl, wher non-alfabetic shorthands ar totaly inaccesbl without lengthy trainng.

A third advantaj of AW is that, since it uses only th letrs of th roman alfabet, it can be kebordd to a word-procesr (or even a typriter). This featur is then valubly complmntd by a computer program cald AgiliTyping, wich autmaticly converts text in AW into TO for editng or printng out.

It is perhaps this third advantaj wich may equip AW for th modrn aje, wher non-alfabetic shorthands ar becomng incresingly obslete. Wher ther is a need for spoken languaj to be recordd in riting at th speed of speech (eg taking minuts at meetngs, or notes of talks) for subsequent typng up, then it is a considrbl gain for ther no longr to be any uncertnty over decyfrng, and for th decyfrng and typng up (ie printng out) to be don imediatly, relybly and autmaticly. And if typng up is don from a dictafone recordng, then th gretr speed of AW again givs a notebl gain in eficiency. One may even speculate that a systm like AW cud rendr th efrts of foneticians and computer experts almost redundnt, ho hav for so long been struglng to produce a usebl systm of machine recognition of speech. And in employmnt terms, AW wud hav th advantaj of not rendrng secretris redundnt.

One may ask if ther is not a practicl drawbak, in that wheras the Pitman secretrys only equipmnt was a pencil and notepad, th AW secretry shud idealy hav a kebord to hand at al times if th advantajs of th systm ar to be fuly realized. But now that laptop computers ar widely availbl, even that is no longr th obstacl that it myt hav been up to five years ago.

It is howevr not only traditionl shorthand with wich AgiliWriting needs to be compared. Inventd as long ago as 1906 and used for instnce for recordng th delibrations of parlamentry Select Comitees (as seen on TV), th stenotyp machine has som of th qualitis of AgiliWriting. It is a systm that uses only th letrs of th alfabet, and is intelijbl to othr users of th systm (se David Crystal The Cambridge Encyclopedia of Language, p207 for a brief acount). It wud be instructiv to hav a comparativ evaluation of AgiliWriting and stenotypng.


So far, so good. Howevr, wen th spelng reformr coms to study th AgiliWriting handbook, certn limitations soon com to lyt. Th first is that ther is litl explnation of th rationale behind th spelngs used, som of wich at first apear decidedly countr-intuitiv. Wy shud th initial <s> of sound, for instnce, be ritn <z>? Wy shud th long valu of <a>, as in pail, be ritn <h>? Ther may be good reasns, but in th short time th presnt revewr cud spend studying AW, he was unable to discovr them. Th same frustrating lak of explnation aplys, as wil be seen later, to som sweepng merjrs of hole sets of digrafs and difthongs wich quite take th breth away with ther daring.

Perhaps, we may optmisticly surmise, ther ar exlnt reasns for th abov featurs, but one wondrs how firm a grasp th authr has of th principls of riting systms wen she describes AW, despite such peculiaritis of sound-symbl corespondnce, as 'based on fonetics'. Th most one can clearly say is that AgiliWriting dos away with many of th most outrajusly 'unfonetic' (mor strictly, 'unfonografic') featurs of traditionl orthografy. Furthr douts ar rased wen, in her expositions, th authr fails to distinguish clearly between letrs and th sounds they stand for. Howevr, here again ther may be a good reasn: it may be that she is cunngly blurng such distinctions so as not to confuse her intendd readrship, ho wil not hav an academic intrest in such nicetis.

But lastly, her own spelng dos not inspire unboundd confidnce wen we repeatdly read of dipthongs and apostrophies. Shudnt we, tho, as peple ho apreciate th endless problms TO causes, perhaps rathr take a charitbl vew of such errs?

For th spelng reformr asesng th AW handbook, a furthr slytly iritating featur is th constnt repetition of th rules, ofn using identicl wordng, with a jenrus alocation of space on each paje. Th esential systm cud probbly hav been described comfrtbly over 10 pages of norml print, rathr than th 230+ of this volume. Howevr, such repetitivness and jenrus spacing may wel be virtus in a self-tuition manul such as th AW handbook primarily sets out to be.


Th presnt revewr howevr regards such criticisms as quibls, wen set beside one remarkbl featur of AW. Quite independntly, Anne Gresham has com to most of th same conclusions as CS in anlyzng redundncy in TO. Th introduction to AW states that 'letters which are phonetically weak or silent' ar substituted or elimnated. And so we find a strong eco of al th CS rules thruout AW.

As by CS Rule 1 (omitng letrs unconectd with pronunciation), such forms as breath, debt, evolve, ignore, money, you ar cut in AW to breth, det, evolv, ignor, mony, u (tho curiusly it apears that write keeps its silent <w>).

As by CS Rule 2, Categry 1 (cutng post-accentul shwa befor <l, m, n, r>), abundant, bundle, doctor, filter, under, upward, urban, cultural, tolerant ar cut in AW to abundnt, bundl, doctr, filtr, undr, upwrd, urbn, cultrl, tolrnt. CS Rule 2, Categry 2 is seen in AW in th past tense <‑ed> reduced to <-d> and th <-able, -ible> endngs reduced to <‑bl> (tho TO <‑ing> reduced to CS <‑ng> is furthr cut to AW <‑g>).

And as by CS Rule 3 (simplifyng dubld consnnts) clock, spell ar cut in AW to clok, spel. And as by combnations of those CS rules, answer, battle, cotton, dagger, heighten, hidden, tackle ar cut in AW to ansr, batl, cotn, dagr, hytn, hidn, takl.

Likewise, th CS substitution rules ar observd, as in AW brij, tuf, fyt, aplyd.


To reduce text by as much as 40%, far mor cuts ar necesry than ar alowd by th rules of CS, and it is intrestng to se wher they ar made in AW. As was observd in th CS Handbook, Chaptr 6, furthr econmis can be made by certn substitutions. So, for instnce, letrs can be saved by always spelng long <i> as <y>, rathr than just substituting <-ig> as in CS sy, syn, syt for TO sigh, sign, sight; and AW sezes this oprtunity, respelng for exampl TO bite, guide, knife, lied as byt, gyd, nyf, lyd (howevr this sound-symbl corespondnce is not consistntly aplyd: idle, item becom idl, itm, not ydl, ytm). Similrly, as urjd by Robert Craig for CS, th varying vowl letrs of heard, her, sir, burn (tho not, it apears, word) ar cut so that th <r> alone represents th vowl sound, givng AW hrd, hr, sr, brn. And as sevrl comntators hav urjd for CS, som pre-accentul shwa letrs ar cut too, most notebly in th unstresd prefix <con->. Th combnation of sevrl of th abov cutng rules reduces TO conserving to AW cnsrvg.


A featur of AW wich shud be of particulr intrest to spelng reformrs is its inovativ use of <w>. Th linguistic lojic behind this device requires som explnation, as it may be unfamilir to many spelng reformrs.

Th letrs <w> and <y> can hav sevrl functions in TO: they can be consnnts (somtimes cald semi-consnnts or semi-vowls), as initialy in TO worry, yellow; or they can function as vowls, as <y> in worry or reply, or <w> in th <ow> digraf in yellow, allow; or they can function (like <u> and <i> respectivly) as glides, as in swayed, lanyard (compared with <u, i> havng th same glide function in suède, laniard). But wile <y> can also hav th valu of a long <i> (as in reply), TO dos not use <w> with th long valu of its vowl-letr equivlnt, wich is <u> (exept, argubly in two). Yet <u> sufrs from th heviest functionl overload of al th vowl letrs, as seen in its standrd valus in but, put, truth, music, fur, persuade. As Robert Craig has again sujestd, th letr <w> cud wel be used to reduce th overload on <u>, perhaps by taking over th latrs valu in music, and/or in truth, givng mwsic, trwth. One is reminded of th welsh spelng of ambulance, wich is ambiwlans.

Spelng reformrs somtimes object that to use <y> both as in yes and as in by is ambiguus, but they overlook th clear positionl distinction between th two valus. With its valu as in yes, <y> must initiate a sylabl and precede a vowl, wheras with its valu in by, it dos not normly do so (prevocalic ocurences of long <i, y> as in ion, iodine, dyer ar few). This grafotactic distinction between th two values of <y> efectivly ensures ther wil be no ambiguity between its valu in yes and its valu in, say, Argyll.

Th same lojic cud usefuly be aplyd to a dubl valu for <w>. If it wer used to releve <u> of one or both of its long values (wen it wud truly hav th valu of 'dubl-u'), that wud not normly ocur befor a vowl, wheras th consnnt valu of <w> by defnition must ocur sylabl-initialy befor a vowl. So th conflictng uses of <u> in cucumber, unused etc cud be resolvd by respelng such words as cwcumbr, unwsed. (Ocasionl oditis such as ww for TO woo or swwp for swoop wud howevr arise.)

AW demnstrates this use of <w> to represent thre values of <u>, namely as in put, truth, music, and ofn saves a letr in th process. Thus tube becoms twb, as oposed to tub, wich is unchanjed. Othr exampls ar bwk, bwgl, acwt, rwl for book, bugle, acute, rule. Spelng reformrs may like to explor th potential of such uses of <w>, wich hav not usuly formd part of ther armry.


An intrestng and injenius (tho perhaps questionbl) device in AW is th vowl scale, by wich th five vowl letrs ar aranjed in ordr of precednce in th sequence <u, o, a, i, e>, with short <u> nevr (?) being cut, <o> being protectd unless th word contains a <u>, <a> havng loer status than <o>, and <i> and especialy <e> being cut al over th place. This hierarchy is rationlized in terms of th desendng ordr of 'resnnce' of these vowls, <u> thus being claimd as th most 'resnnt' and therfor most worthy of protection from cutng. Th fonetic basis for this concept is not explaind, indeed th jenrl confusion of sounds and letrs in th AW handbook makes one wondr wethr this 'resnnce' may not be at least as much a matr of th visul promnnce of th letrs as of th acustic promnnce of th sounds.

Howevr, this scale of vowls dos ofr an esily aplicbl rule for removing vowl letrs, altho th avraj spelng reformr wil be unhappy at th outcom, wich ofn results in th loss of stresd vowls and th retention of unstresd vowls. For instnce, one myt expect that in th word renovate, th first <e>, being stresd, wud survive watevr othr letrs wer cut, and that th <a>, wich carris th secndry stress, wud be mor likely to survive than th unstresd <o> hos valu is merely shwa. Howevr by its vowl-scale rule, AW produces th form rnovt. This countr-intuitiv spelng is just one of inumerabl exampls of th efects of th vowl scale.

It wud be good to no wat th justification for such forms is. Perhaps they ar adequat for this shorthand systm, inasmuch as th form rnovt may stil be recognizebl to anyone familir with TO renovate; or perhaps they ar determnd by th needs of th computerized AW-to-TO conversion program. But on th evidnce of th AW handbook those posbilitis must remain mere conjectur. One is left wondrng wethr, by aplyng othr AW rules, th form renvht myt not hav been mor apropriat.

At al events th form AW rnovt for TO renovate clearly demnstrates wy such a systm canot be considrd as a fuly flejd orthografy (tho paradoxicly, th non-alfabetic shorthand systms myt be so considrd). A propr orthografy needs to tel readrs how to pronounce words they ar not familir with, and to tel riters how to spel words hos pronunciation they no. Rnovt for renovate meets neithr of these criteria.


Anothr disturbng featur of AW for th spelng reformr is th way 'majic' <e> is somtimes cut without compnsation, so that vote for instnce becoms just vot, and sage becoms just saj. Then ther is th way in wich th letrs <y> and <w> ar aplyd not just as described abov, but also to represent a wide ranje of <i> and <u> glides. Thus jovial, medium, onion becom jovyl, medym, onyn, wile fluent, poetry, ruin becom flwnt, pwtry, rwn. Ther is a certn lojicl atraction in these uses of <y, w>, but th efect is so daring that th avraj spelng reformr is likely to react to them as wud th most dyd-in-th-wool orthografic conservativ. Suspicion is then likely to turn to outryt rejection of such forms wen we find that day, may, say ar cut to dy, my, sy, and <w> is brot into service for al th vowls in broad, foil, count, draught, givng AW brwd, fwl, cwnt, drwft.

Undrlyng this featur of AW is of corse th major deficiency of TO - that ther ar nowher near enuf letrs in th roman alfabet to represent al th 40+ fonemes of english unambiguusly, th deficiency being particulrly acute for vowls. In jenrl TO resorts to digrafs to deal with th problm, but also makes som positionl distinctions. AW adopts a difrnt solution: it cooly alocates a singl letr, such as <w>, to stand for a wide ranje of difrnt sounds. Dos that matr, especialy if AW works in practis? On th evidnce of th AW handbook, we ar perhaps not entitled to conclude that it dos matr. Posbly this daring device is based on a quite briliant new insyt. It wud be nice to no.


This revew of AW has not don ful justice to th systm by any means. It has not discusd its extensiv patrns of abreviation ('Shorts' and 'Strings'), but at this point a few lines of AW must sufice to giv an impression of them in th context of th ful systm.

Th foloing TO text:

The instructions for the operation of our Ideal instruments appear on pages 6 to 9 of the illustrated leaflet. Kindly indicate your intentions by initialling on the index each item in which you are interested. We will invoice you for any increases or incidentals.
apears as folos in AW:

Th nstrucns fth oprhzn v ur Idyl nstrmnts apr on pjs 6 t 9 v th ilstlrhtd lflt. Kndly ndct y ntnzns bi nshlg on th ndx ech itm n wch u r ntrstd. Wwl nvz u f ny ncrzs or nzdntls.
Th linkng of two seprat short words, as in fth for of the and wwl for we will is remnisnt of Harry Lindgrens radicly streamlined systm, Fonetic B. It is also noticebl how th shortr words alow a typograficly far mor compact text, with many mor words on each line than in TO. But in terms of readng sycolojy one must say that recognizebility for th initiated, rathr than decyfrbility for th uninitiated, is th ke quality of AW spelngs.


This revew has anlyzd AW as a riting systm, indeed to som extent even as a potential reformd orthografy for english, wich is of corse unfair. Th conclusion reachd is that wile AW contains many useful and som hyly orijnl featurs that ar wel worth considrng for a reformd orthografy, it also contains (inevitbly for a shorthand systm perhaps) many featurs wich wud be simply perverse if desynd for norml use. But such criticisms of AW ar unfair, both because we hav not discovrd th rationale behind som of those featurs, and because we hav not realy atemtd to juj AW for th purpos for wich it was desynd, namely as a shorthand systm. Such evaluations must be made by othrs, that is by teachrs, lernrs and practitionrs of shorthand. Spelng reformrs wud howevr no dout be intrestd in th results of a proprly orgnized professionl test.

Anne Gresham is currently workng on an even mor radicly abreviated version of AW, to be cald Agili+Plus. It wil be fasnating to se wat furthr cuts she manajs to make in th spelngs.

Meanwile, we may end by wondrng wethr, just posbly, AW myt hav been a betr systm even by its own lyts (ie it myt hav been mor readbl) if it had taken on bord som of th basic principls of regulr riting systms. Th most notebl: that th sycolojy of readrs and riters is il-servd if th stresd vowl-letr in a polysylabic word is deleted, wile an unstresd vowl-letr is left unscathed. Th reasn is that th success of alfabetic riting systms depends crucialy on a transparent relationship between ritn letrs and spoken sounds, and stresd vowls represent th fonlojicl cor of spoken english. It is th failur of english to observ this principl itself, aftr al, that is th fundmentl cause of its problms today. Can a shorthand systm that uses only th letrs of th alfabet be exemt from such considrations?

Th ansr may wel be, 'yes', as english jenrly is nown to be stil lejbl with al vowl-letrs removed. At al events, it is hoped that this revew has aroused intrest in AW as a powrful systm that probbly has strengths that hav not been adequatly elucidated here. We shal be intrestd to lern of its progress.

Back to the top.