Esthetics of Language Design:
The Case of Acadon

Each language that is written in the Latin-French alphabet has a different "look" on the page. This look is imparted by the actual letters chosen by the language, including any modified or marked letters, the frequency of each letter, and the common combinations of letters in strings. For example, English has many TH and GHT combinations, Spanish does not. The sequence VWV or YY does not appear in either language. Within the various literate populations of the world there have always been traditions of what "looks" right in terms of writing systems -- and what does not.

The impression given by text is very subjective. But it is nonetheless very real. Some of these impressions may be personal, but most are widely shared within the individual's linguistic culture. Often they may be spread across dozens of linguistic cultures. These distributions have been created through history, and they form part of the "world wide language situation" as it prevails today. Widespread impressions of "what a proper language should look like" are as real as is widespread vocabulary, and they are just as important in language design.

Alphabets are involved too. Wars and revolutions have sometimes involved issues of writing system. Kemal Ataturk, for example, made romanization of Turkish a major plank in his platform. The struggles in the former Yugoslavia during the 1990's often reflected the tensions between two alphabet systems for what was essentially the same language: Croatian in Roman letters and Serbian in the Cyrillic alphabet. Shopkeepers displaying signs in one alphabet often found their shop windows broken -- even their produce looted -- when the other side took control.

Much of the great resistance to spelling reforms in the English language has been that proposed texts in the new system give those familiar with traditional English the impression they were written by "foreigners" or "illiterates." A sentence such as
"Kuhnadiyuhn kulchir ehmfuhsayzuhz dhuh importuhns uv
pruhzerving dhuh kwaluhti uv layf uv dhuh induhvijwuhl"
would never be perceived as a proper replacement for
"Canadian culture emphasizes the importance of
preserving the quality of life of the individual."
The average Canadian would probably not hesitate to describe the sample rewrite as "ugly." Although the difference is totally subjective, language planners cannot overlook such prejudices. They are strong and run deep. They are part of the worldwide language picture.

Many efforts to design international languages have stumbled badly on these grounds. At least one project for an international language has utilized the letter X as an additional vowel, on the ground that the letter was otherwise little used and thus readily available. Nothing is inherently illogical about this approach, except that it ignores deep prejudices and ancient traditions on what letters properly follow others. A sample text follows with the vowels replaced by W, UU, X, Y, and Z:
Thxs suuntzncw ys przvxdyd zs w suumplz fwr cympzrxsyn
zf thx sybjzctxvy lzxk yf z twxt yn uu puugw znd tw
yllzstrxty hzw chxnguus wn lxttuur zsxgys mzy crxwtz
qxytz dyffwruunt xmprwsszxns uun thz rxydzr bxsyd zn
pxst wxpzrxuuncz xnd lzng-stwndzng uuxpzctwtwzns.
What it says, of course, is that :
This sentence is provided as a sample for comparison
of the subjective look of a text on a page and to
illustrate how changes in letter usages may create
quite different impressions on the reader based on
past experience and long-standing expectations.

Marks on letters, or variant letters, create subtle impressions as well. Three sample texts follow with letters replaced by alternative forms.

1. T_is _enten_e iš prvi_ed a_ _ample fo_ _ompri_on of _he sb_e__ive _ook of _ext on pge nd _o _llu__rte _ow _hnge_ in let_er sgeš my _rete qi_e diff_ren_ imp_ešsion_ on _he reder ba_ed on ps_ exp_rien_e nd _ong-_tn_ing expect_ion_.

2. Ths sentene s prov_ded s a smple fr comprson œf the subj_ctve l_ok œf a t_xt n a p_ge nd t_ ll_strte how ch_nges _n ltter us_ges m_y crete qu_te dfferent mpressons œn the r_ader b_sed œn pst exp_rene nd lœng-stndng xpecttons.

3. tH_S sEntece i_ prVidEd S A smplE fr cmPriSn oF tHe suBjEctivE lOok oF teXt o pge d tO iL_usTrte hOw cHnges i letTer usges my cRetE QiTe diFferet impreSsis N thE reder basEd paSt exPerIece anD _onG-stndiG _xpeCtTioS.

Many users of the Latin-French alphabet would probably agree that the first text looks very difficult and foreign; the second is a bit less jarring; and the last looks downright silly.

Although there might be nothing wrong with the logic of any systematically applied symbolism, it is easy to annoy potential users by choice in design. Beginning learners of Esperanto often start to become disenchanted as they come across sequences such as the last words of the forward to a 1995 English-Esperanto dictionary:
"vortoj kaj esprimoj propraj al
siaj diversaj Anglalingvaj landoj."
This phrase refers to "words and expressions appropriate to the various English-speaking countries." (Benson, Peter, Comprehensive English-Esperanto Dictionary, (El Cerrito, Calif.: Esperanto League for North America, 1995), page 16.) The sounds might not have been as jarring to the look, however, if the final diphthongs, etc., were written by the rules of Acadon. It would then read:
"vortoe cae esprimoe proprae al
siae diversae Anglalinguae landoe."
Whether this would be better is, of course, subjective. But it would fall more into accord with the factors noted here.

Latinate spelling

To take a specific example, Acadon writes the [k] sound with a C, as was the case in ancient Latin. The question then rises: Wouldn't it make more sense to write the [k] sound with a K; and perhaps use C for the [ts] combination? Most prior IAL projects have said "yes" to this (with the exception of Peano's Interlingue) -- or introduced irregularities into the spelling system to accommodate for C in some cases K in others. Acadon says no.

It is true that most languages that have been romanized since the nineteenth century have almost automatically used the German (perhaps one could say, Greek) K for the [k] sound rather than the Latin C. This reflects in large part the great influence of German scholars on nineteenth century linguistics. The very fact that the International Phonetic Alphabet (IPA) uses [k], is an expression of this influence. (So is IPA use of [j] for the sound in German "ja.") In general, these usages were common in those parts of Central and Eastern Europe that had accepted the Roman alphabet -- such as, Hungary, Finland and Poland.

In the late nineteenth century, proposed international languages uncritically followed the German usage. Most African and Asian languages that have been romanized have followed the German pattern. A major exception is Vietnamese (which uses both C and K). Vietnamese was, however, romanized before the rise of the German linguistic school. Other exceptions include languages like Zapotec which were romanized under heavy Hispanic influence.

Today, four major regional languages in African and Asia: Afrikaans, Indonesian, Turkish, and Swahili, use the K pattern. All had experienced heavy German (or Dutch) cultural/political influence. Recall that the Swahili area centers on the former colony of "German East Africa," and that Indonesia was the "Dutch East Indies." It is true, however, that many smaller, often one-nation, languages (including: Fijian, Hausa, KiKuyu, Malagasy, Yoruba, Somali, Tagalog,) have followed along. The prestige of the IPA has carried great weight with designers of languages for many decades, including the nineteenth century missionaries who reduced many African and Asian languages and dialects to written form.

The overt success of the Germanic K now seems so accepted that the point could be made for using K for [k] in any designed language. Nevertheless, reality is more complex. When people around the world study a second language for wider regional or international communication, which do they study? Of course, many study languages like Arabic, Chinese, Hindi, Russian, or Japanese which do not use the Latin-French alphabet at all.

However, among students of languages using the Latin-French alphabet, the vast number are in fact involved in learning one of six languages: English, French, German, Italian, Spanish, or Portuguese. Five of these six (all but German itself) begin words like "capital" or "catholic" with the letter C. Only German uses a K. Far more language students, classes, textbooks, and teachers are working with languages that favor the C over the K in such cases. In using C in such cases, Acadon is simply accepting the pattern set by the majority of the most widely-studied and prestigious languages that employ the Latin-French alphabet.

The issue of prestige among languages is closely related to esthetic evaluations.(#13)* Habit and other expectations from past experience tend to lead many persons to find beauty in the familiar and revulsion in the unfamiliar. "Culture shock" is an extreme example of this. So the language designer should wish to introduce as little such shock as possible.

Seven historically important languages have in fact been very major bearers of high culture to large regions and are therefore among the most prestigious of existing languages. These are:
1) Classical Greek,
2) Classical (Literary) Chinese,
3) Sanskrit,

4) Ancient (Biblical) Hebrew,
5) Latin (Classical, Vulgate, and the ISV),
6) Arabic (of the Koran), and
7) French.

English lacks the deep reverence that many of the above languages have inspired, but it certainly has high prestige of a practical sort today -- far more than when the twentieth century began. It could be added to make a list of eight. Within the fields of music and art, Italian has long rated being called "preeminently prestigious." German has had that status in linguistics and certain of the sciences.

The currently recognized "preeminently prestigious cultural languages" set the patterns by which other languages are judged. A newly designed language will look more esthetic or "proper" on the page if it follows the norms of the more prestigious languages. Acadon attempts to do that with its spelling system.

However, only two languages on the list of seven are written in the modern "Roman alphabet:" Latin and French. If the list of widely "prestigious languages" be expanded to eight, then English is the third that uses that alphabet. Additionally, modern science uses many words in its nomenclature or "international science vocabulary" (ISV) that are from Ancient Greek but in a "romanized" form> Thus, there is a prestigious sub-set of Greek terms that are cast in the Latin-French alphabet. As a result, the most widespread and prestigious forms that might be taken as models for "esthetic standards" for spellings and letter combinations in the Latin-French alphabet include:
1) Romanized Greek (ISV),
2) Latin (and the ISV),
3) Modern French,
4) Contemporary English

Acadon applies the consequent design principle: "If the graphemic design of Acadon is to be esthetic and look as appropriate as possible, its written form will have to approximate as closely as it can the norms set by Latin and French, and to some extent by English and romanized Greek.

Among current spelling systems using the Latin-French alphabet, some give the impression on the page of being more "classical" while others seem more recently "improvised" on the basis of nineteenth century phonetic usages. Heavy use of "Germanic letters" such as K and W, are among the quickest indicators of the more recent "phonetic" style.

Generally speaking, however, spelling systems using patterns closer to the "classical system" convey an impression of longer tradition and greater prestige. As remarked above, with the exception of German itself, all of currently most widely-studied languages that use the Latin-French alphabet follow the classic mode more closely than that of the IPA. This is to say that the French, Spanish, English, Portuguese, and Italian languages all prefer a C over the K in such words as appear in the phrase "Canadian Francophone culture."

To pursue its esthetic goals, Acadon spelling does not attempt to follow the norms set by nineteenth century linguists as exemplified by the IPA. If anything, it reflects patterns to be found in the most widely used and studied languages that use the Latin-French alphabet. This is a realistic approach.

Esthetics of Sound:

The above has addressed the issue of the "look" of written languages, but the sounds of languages are also a major esthetic consideration. Spoken language is, after all primary to written language.

Many have discussed languages in terms of music, and it has been typical to praise languages in terms of their use in music. In this, Italian has usually been among the most highly praised. Its many open syllables (ending in vowels) make singing clearer and rhyme somewhat easier to find. Yet Spanish has fewer vowels than Italian; and Swahili has more open syllables.

Russian has many complex consonant combinations, but is often considered to sound quite pleasant to outsiders. Cantonese, on the other hand, with many falling tones, may sound rather angry to Westerners. English, on first being heard by a Danish boy, sounded "like dogs barking." Within Western Europe, spoken German has often been described as guttural and harsh, but this may be as much the result of prejudices in the wake of two world wars than anything real. Is there any standard of beauty in all this?

The principal argument against having syllables or words end in multiple consonants is probably ease of pronunciation rather than esthetics per se. Yet many designed languages have had words with a wide variety of consonantal endings and even with multi-consonant endings. Esperanto allowed final vowels to be dropped in poetry and song. In distinction, French, when sung, often has additional final vowels supplied. These factors relate to meter as well as to the sounds per se.

Another issue has been whether certain sounds of language are inherently more beautiful, and others ugly. Affection for the familiar plays a major role here. Several French contributors to the IAL movement have implied that the H [h] sound (which French lacks) and other "guttural sounds" were inherently ugly. Yet such a statement might well seem offensive to a Arab if applied to the language of the Koran -- which has many such sounds. Cynics might also note that the "Parisian R" is in fact a guttural sound, often transcribed by GH in the languages of Asia and Africa.

Polls taken of speakers of English often seem to find words with L's high among the "more beautiful words" in the language, e.g., "lullaby." Is this likely to be the case with a Japanese, whose language has no L?

Acadon relies in great part on those principles designed to make it clear and understandable to also make it sound well. It has only five vowel sounds (less than Italian's seven); it sets limits on the number and placement of consonants in syllables; it uses its law of avoidance to separate potentially confusable words, and it applies a stress pattern that leaves most words accented on the final syllable -- and all longer words with a secondary (lighter) accent on the first syllable.

Final Observations: Integrity of the Whole:

Language planning is not comparable to engineering. Engineering is done under objective situations where the determining factors are measurable. In such circumstances, the human factor is less relevant; measurable facts are paramount. This is not the case in language planning.

Use of language is a more subjective human activity and deeply tied to subtle perceptions and expectations that have been set by the larger cultural and historical milieu, and which are often left unarticulated. Helge Heimer makes this point in his book on Mondial:
"We here enter upon another aspect of the international language problem, to which hitherto all too little attention has been paid. To create an artificial language is not only a technical problem, but also an artistic one. A language is not to be compared to an algebraic collection of formulas. The primary and fundamental task of a language is to render by speech all the thoughts and feelings that stir in the human soul."
Heimer then goes on to say:
"To fulfil this task satisfactorily, every language that claims to be a language of culture must possess a certain artistic shape and character, which is most evidently manifested in the rhythm of the language. Like every civilized language, the artificial language, too, should possess such an artistic shape and character, but that will be impossible if it does not essentially adhere to a certain language group."

Unfortunately, this final conclusion does not follow from the logic of the prior remarks. In fact, one of the dangers of designing a language to mirror languages of a rather specific group, is that the result will turn out to resemble a parody of that group. If it is simpler (as a language designed to serve as an IAL is likely to be) then may well seem to be a semi-literate or pidginized form of that language. This may create a comical effect, surely not an elegant one.

Of course, by creating a language that closely follows the constraints of an existing language group, one may more easily assure that it will have a perceivable "artistic shape and character." But since it is an artificial language that we are speaking of, there is no reason that its "shape and character" may not be largely formed by artifice as well. This creates a larger task, of course, but one that can also serve other communicative purposes.

Copyright (C) Leo J. Moser 1997, 1999