Nature and Art in Language

by Otto Jespersen, 1933

Part 2

Among countries which prefer making their own terms to adopting foreign words must be mentioned Iceland. The visitor to Rekyavik is astonished to find the great number of native words for things which have nearly everywhere in the civilized world the same names: reiðhjól for bicycle, skrifstofa for bureau, fallbyssa for cannon, skammbyssa for pistol, bókasafn for library (French bibliothèque), sími telephone and telegraph, tundurdufl mine, tundurbátur torpedo-boat. The names of sciences are all native: guðfræði theology, lækfræði medicine, grasafræði botany, dyrafræði zoology, efnafræði chemistry, etc., and the same is true of such scientific terms as afleiðsla deduction, aðleiðsla induction, hlutrænn objective, hugrænn subjective. Some of these terms are comparatively recent, and in many cases my Icelandic friends have been able to name to me the originators of terms that are now current there; Jónas Hallgrimsson started in 1842 many astronomical terms and Magnúss Grímsson in 1852 many physical terms, e.g. ljóosvaki ether, sólnánd perhelium, rafurmagn electricity, tvíætting polarity. From the latest time we have víðboð for broadcasting, as it were wide-message.

Similar tendencies are found in another northern country, Finland. On account of the more foreign character of the vocabulary (Finnish belongs to the Finno-Ugric, not to the Indo-European family of languages) I shall give fewer examples. Nature is luonto (from luo create), religion is uskonto (from uskoa believe), electricity is sähkö, telegraph is sananlennatin, an ingenious formation from sana word, (genitive sanan) and lennän fly, lennättää to make something fly, send off rapidly.

In Czech native formations have been extremely successful in keeping learned loan-words out, as I learn from some interesting papers by O. Vocadlo.

In Germany, and similarly in Denmark and Sweden, both those extremes which we found in modern English and Icelandic, are avoided: there are a certain number of perfectly natural native formations, and by the side of them many Greek and Latin loan-words, chiefly for purely scientific notions, but also for such everyday things (based, it is true, on scientific inventions) as telegraphy, photography, etc. But in these countries purists have for a long time been at work to introduce new native formations, and not unfrequently they have been successful, especially where they have been able to induce political and administrative authorities to take an interest in the matter. It is well known that Kaiser Wilhelm at some time favoured these tendencies, and had some influence in the supplanting of telephone by fernsprecher (fernsprechamt telephone office, fernsprechapparat receiver, etc.). Not many years ago the word billet 'ticket' was officially supplanted by fahrkarte, while the German-speaking Swiss still use the word billet. But the number of these official purisms is not very considerable. Rad for cycle, and radfahren, radfahrer etc., are probably popular in their origin.(5) So far as I know, kraftwagen, though supported officially, is much less used than the more convenient foreign word auto; kraftwagenhalle is a native, but unnatural equivalent of garage.(6)

While a spoken language is found wherever human beings live together and must thus be considered part of human nature, the same cannot be said of written language, which is everywhere of much later origin and must really be called an unnatural substitute for spoken words. But in the same way as so many things that were at first "unnatural" inventions - the use of clothes, fire, and in later times matches, electric light, telephones, etc. - have come to be felt as so natural that our children from the very first years come to look upon them as self-evident things which they think must have existed from the very dawn of human life, so it is also with writing, which we now consider a natural way of communicating with fellow-beings. Natural, that is, to some extent only, for it cannot be denied that, as most of our civilized languages are now written, there is a good deal in them that can hardly be called natural. In the first placas regards spelling. As soon as people had invented the art of representing each speech sound, or let us rather say, each of the principal speech sounds or phonemes, by means of a separate symbol, the natural thing was to write down spoken words as faithfully as possible, and that was what people really did, or tried to do. But soon tradition came into play, and people were taught not to depend on their own ears and write down words as they themselves pronounced them, but to spell as their teachers, that is to say roughly as an older generation, pronounced the words; and as this went on continually and the spelling of words was changed much less than their pronunciation, the gap between the two forms of language became greater and greater. The results with regard to English and some other languages are patent enough, especially if we compare them with the beautifully simple spelling of such languages as Finnish, which have not been literary languages long enough for tradition to have had the same effect. In some countries Academies (as in Spain) or ministers of instruction (as in Germany and Denmark) have from time to time interfered with spelling and discarded some of the worst anomalies: but are such regulations "natural" or unnatural?

There are other artificial elements in written language besides spelling. As writing addresses itself to the eye, many of the subtle effects perceivable by the ear (intonation, etc.) are utterly lost when sentences are written down, punctuation marks being at best a poor substitute for much of what makes spoken words expressive. The whole structure of sentences and their combinations has to be changed, and even the simplest familiar letter has to be formed in a different way from the same communication if it had been oral. This rearrangement has to be learnt artificially, though much of it may come unconsiously by instinctive imitation of models of various kinds.

As in spelling, so in grammar, teachers will often insist on forms that really belong to an extinct stratum of their language. In English - to give only one example - hundreds of passages in Elizabethan writers show conclusively that who had supplanted whom in natural spoken language: "Who did you see?" "Who is that letter for?" Schoolmasters, however, insisted on the old form - at any rate in writing, for in ordinary conversation they were not very successful. The result is that people who want to show off their superior education write whom even in cases where their teachers, had they been consistent grammarians, would have understood that who was the correct form, and it is now easy to collect scores of examples, even in the books of very good writers, of such constructions as these: "power to summon whomsoever might throw light upon the events." "Peggotty always volunteered this information to whomsoever would receive it." "She talked nonsense to whomsoever was near to her."

Poetry and religious prose everywhere are "naturally" fond of "artificial" archaic expressions, and in some countries all prose writing, even the most everyday communication, has to be clothed in a linguistic garb that really belongs to a distant past. In Dutch the written language is only beginning to get rid of the old word genders which were given up in the spoken language long ago, and many people have to consult a dictionary very frequently in order not to make blunders in the use of the forms of the definite article and pronouns. As Dr. Kruisinga says: "in older Dutch, nouns had a threefold gender, and were inflected differently accordingly, as well as their attributive adjuncts. Although this has been lost for many centuries, it has been artificially preserved in the spelling, details being settled arbitrarily by grammarians. This artificial system is still used by the majority of Dutch writers in Holland, and is supposed to be taught in schools, although many schoolmasters practically ignore it."

When Dr. Kruisinga says that this complete severance between written and real language is "unique among the languages of Europe," he is forgetting for the moment Modern Greek, where the written language is artificially screwed several centuries back, not only in one point of grammar, as in Dutch, but in every way. The prestige of the old language with its wonderful literature has been so great that children are taught at school to write many forms that have been extinct for many centuries, and it is the ambition of every Greek writer to keep his language as near as possible to the old standard, though it is of course impossible to blot out everything modern. Feelings are very strong in Greece on this subject, and a revolution was even threatened when the attempt was made to introduce the New Testament translated into the modern vernacular: the original text, it was said, was written in Greek and that ought to suffice (even if much of it was not at all understood by ordinary people nowadays).

Similar linguistic conditions with a written language artificially preserved in spite of its distance from the living speech prevail in other parts of the world, notably in Southern India (Telugu), in Tibet, in China and in Japan; but what I have already adduced here must be sufficient to prove my thesis that much in the so-called "natural" languages is very far from deserving that name.(7)

We may now pass to constructed languages. Their number is legion, and they represent many different stages of artificiality, from the purely "philosophical" or a-priori systems, in which all words are created arbitrarily and systematically without the least regard to any national language, down to those recent schemes which boast of being so natural that they can be read at sight by any educated West-European or American.

Languages of the former kind have this advantage that they may be untrammelled by the deficiencies which (it must be admitted) cling to all national languages, many words of which are sadly wanting in that precision which is desired by the strict logician. But as there is no connection between the coined words of a philosophical language and familiar words, everything has to be learnt anew, and the task of memorizing such a language is enormous, much greater than in the case of those languages whose vocabulary is based on national languages. On the first of them, constructed by Bishop Wilkins in 1668, Leibniz said that besides its inventor the only man who learnt it was Robert Boyle: yet it must be called a truly ingenious piece of work.

Let me now try very briefly to indicate what should be natural and what may be artificial in a constructed language meant to be used for international purposes (8). So far as possible no single element of the language should be arbitrarily coined; everything that is already international should be used, and utilized to the utmost extent. Where there is no perfectly international word (or "stem" or "root") the form which approaches that ideal should be taken, the principle being throughout the maximum of intelligibility to the greatest number combined with the maximum of ease in practical handling.

The question whether ready-made words should be adopted from national languages or new compounds or derivatives be formed with the speech-material already incorporated in the language cannot be settled once and for all; in some cases one, and in other cases the other procedure may be preferable; for purely scientific terms the former procedure will generally be the most natural, but as soon as we leave the domain of pure science and have to speak of everyday objects and occurences, we must remember that many a word formed by means of a well-known derivative ending put on to an international word will be perfectly transparent to everybody, even if it has no previous existence in any national language.

The phonetic system must be as simple as possible and contain no sounds or combinations which would present difficulty to many nations. Hence we can admit the five vowels a, e, i, o, u only, but neither nasalized vowels nor rounded front vowels (ü, ö), which are absent from such important languages as English, Spanish, Italian, Russian. As regards consonants we are similarly obliged to exclude palatalized sounds, such as those in French agneau; It. ogni, egli; Spanish año, calle; and the German ch- and the English th-sounds. By the exclusive use of s, where some languages distinguish a voiceless s and a voiced z, an important simplification is gained, not only because some nations are ignorant of that distinction, but also because the distribution of the two letters would necessarily be often arbitrary and consequently would have to be separately remembered for each word. Accentuation (stress and tone) should not be used to discriminate words.

The alphabet must be that which is known to the greatest number, namely the Latin, in spite of its many shortcomings; but it should not be complicated by arbitrary additions to and modifications of its letters, such as accents over or cedilles under them. Nor is it desirable to use letters in a way that is not familiar to the great majority of presumptive users: if combinations like ca, co, cu are to be used at all, they must have the phonetic value given to them in all European and American countries except Polish and Czech, that is to say that c before these vowels as before consonants should be pronounced like k and not like ts. But as c before e and i is pronounced in four or five different ways in those national languages that count, it seems better to do without that difficult letter altogether; as a matter of fact no one will feel any difficulty in spellings like konsert, etc. This is certainly more natural and less artificial than spelling car and meaning tsar.

The spelling, too, must be as easy as possible; we must therefore avail ourselves of all such simplifications as have already been made in some languages, e.g. f instead of ph, t instead of th; single instead of double consonants, as in Spanish. No letter should be allowed to have two distinct pronunciations according to its position; g in gi, ge must sound as ga, go (cp. Engl. give, get, not as in gin, gem). I know very well that many people would prefer c in conclusione, cria, clari, etc., where I prefer k (which is not beautiful!): but the reader must be asked to consider the fact that not only the Germans, the Dutch and the Scandinavians, but also the Slav nations, thus very many millions, write k in Latin loan-words (in Polish, for instance, kleryk, kredyt, klasa, kronika, krystal; correspondingly in Czech, Russian etc.). The new official Turkish spelling with Roman instead of Arabic letters is in perfect agreement with the rules I had adopted for Novial without knowing of the fact: bank, koridor, fabrika, kontrol, kolosal, sigar, sivil, bisiklet, etc. Anyhow, k seems indispensable before e and i, e.g. anke, kelki, kelke; amike friend (epicene), hence naturally amiko male and amika female friend, amikal friendly. I grant, however, that a moderate use of c and z in those words in which they are fully international would present some advantages and would not essentially affect the character of Novial.

In grammar the same principle of the greatest ease should be carried through, wherever possible. No irregularities of the kind found so often in national languages should be tolerated. The grammatical material should be, and can be, taken from existing languages even to a greater extent than is done in some recent constructed languages. For the plural of nouns the ending -s seems to be the best, as it is found in some of the most important languages and can easily be applied to all words, especially if care is taken not to let substantives end in consonants. To distinguish the two sexes the endings -o and -a seem appropriate, and then -e can be used in all substantives denoting either lifeless things or living beings for which it is not necessary specially to indicate sex. Further it seems a very important principle to apply these endings not only to nouns, but also to pronouns. Most interlinguists do not acknowledge this principle and thus set up special pronominal forms for these two categories, alleging that pronouns are irregular in all national languages, and that it is therefore against ordinary linguistic psychology to create regular pronouns. This, however, is only a half-truth, one might even say that it is a fallacy: in their historical development even pronouns tend towards regularity, and if such simplification comes about very slowly in this class of words, the reason is that the extremely frequent use fixes the forms in the memory. Exactly the same thing happens with the most often used verbs, which for the same reason in all our languages are irregular (am, is, was, be; bin, ist, sind, war; suis, est, sont, était, fut, sera; go, went; vais, aller, ira; gehe, ging, gegangen .....); but in spite of this no interlinguist has proposed to give an irregular inflection to the corresponding verbs in constructed languages. An exception is just as indefensible in one as in the other case. An international language can and must be less capricious and less complicated than even the most progressive national language. Hence, in Novial, just as we have kato and kata for a male and female cat, respectively, and kate, when no sex is to be indicated, and correspondingly artisto, artista, artiste, etc., we have in the pronouns lo he, la she, le he or she (e.g. si omne veni kand le deve, nule besona varta, if everybody comes when he or she ("they") should, no one has to wait). Further nule no one, nulo no man, nula no woman; kelke somebody, kelko some man, kelka some woman. In the plural les they (generally), los = Fr. eux, las = Fr. elles, etc.

In the verbs it seems advisable to have an ending to denote past tense, as this occurs so very often, but otherwise it is in accordance with the pronounced tendency of West-European languages to make an extensive use of short auxiliaries, which may be easily combined to express all manner of complicated ideas: me ha veni I have come; me had veni I had come; me ve (better than sal?) veni I shall come; lo ve ha veni he will have come; la vud ha veni she would have come, etc. An indication of person and number is superfluous in the verbal form, as the subject is always there to give the necessary information in that respect.

Fortunately there exist numerous word-building elements (prefixes and suffixes) that are already known internationally and can be adopted without any change. The only thing required is to define their use and to be free to apply the same prefix or suffix to all words, whereas natural languages present all kinds of more or less inexplicable restrictions. Vague and inaccurate definitions of suffixes should have no place in a rational language, and even less acceptable - to mention one example only - is the use of the two Latin prefixes in in two nearly contradictory senses in Occidental: inscrit inscribed and ínscrit unwritten (the accent is an unsatisfactory and ineffective palliative). One of the great advantages of a constructed language is the power it gives every speaker to form a word by means of a recognized suffix without having first to inquire whether it is already in use; but if radicals and suffixes are well chosen, it is possible to form any number of derivatives which will be immediately understood. Take the ending -torie for a place where something is done: observatorie (from the verb observa), lavatorie, dormitorie, laboratorie, auditorie (from audi, to hear), manjatorie dining-room, gajatorie pawnshop (gaja pawn), kontrolatorie, etc. The procedure may be extended in infinitum.

Regularity thus is one of the foremost requisites of a constructed language. But what exactly does regularity mean? It may briefly be defined as expressing the same idea, the same notion or modification of a notion, everywhere by the same means. But this principle does not carry with it the principle "similar things expressed by similar means", for that leads to uncertainty and mistakes. Let me give one example from my own language, Danish: here the two months June and July are called (as in German) Juni and Juli, but the too great similarity occasions many mishearings, obliging you to repeat what you said. Here English June, July, and French juin, juillet are much better. But in the latest philosophical language, Mr. E. P. Foster's "Ro", the names of the months are tamab (December), tamad, tamat - those three together form tama 'winter' - further tameb tamed tamet, etc. Similarly the days of the week are takak, takal, takam, takan. Now the inevitable result of such systematization is that Ro is utterly impracticable: think of the number of mishearings over the telephone, especially as the numerals in Ro are constructed on the same principle, for a man intending to say Monday the third December will easily be thought to speak of Tuesday the fourth January, etc. In Mr. Wilbur M. Beatty's "Qosmiani" the numerals are nul (1), dul (2), mul, bul ful, xul, sul, gul, hul - perfectly systematic, it is true, but just on that account this detail is quite sufficient to condemn the whole language. We have seen above how national languages tend to get rid of too great similarities between names of similar things which it is often important to keep easily distinct. A smaller fault of the same kind is made in Esperanto with the words for 'right' and 'left', dekstre and maldekstre, which in the marine would give rise to the same kind of mishearings as starboard and larboard did. So we see the importance of this principle: the same thing expressed in the same way, but not: similar things expressed by similar means.

If instead of the fantastic numerals and names of the months just mentioned we simply, like most recent constructed languages, take un, du, tri (known in English through unit, duo, trio), januare, februare, etc., and if we base our vocabulary on the lines indicated above, i.e. utilize to the utmost extent such words as nature, natural, universe, universal, natione, national, periode, forme, literature, teatre, komedie, dansa, autore, historie, kanone, pistole, produkte, produktiv, produktione, akte, aktiv, aktione, labora, laborere, laboratorie, dentiste, dental (whence dente tooth), admira, admiratione, splendid, stupid, steril, sterilisa, simpli, simplifikatione (9), etc. etc. - and if we glue these words together by means of a simple grammatical apparatus, we shall be able to build up a rich and expressive language which will shock no one by its unnatural sound or look and which can be very easily acquired and used by men and women of average intelligence.

Just one little specimen to show how such a language looks in connected speech; it will present no difficulties to any educated European or American:

Kulture es ekonomie de energie in omni direktione. Li kultural valore del universal helpelingue es ke le limita li enormi disipatione de energie a kel li homaro ha es til nun submiset. Per liberisa ti energies on pove utilisa les por li kultural taskes kel li homaro non ha ankore solu, e li gano por kulture ve es non-previdablim grandi. (li definite article; homaro mankind, cp. formularo collection of formulas, glosaro, etc.; solu solve, cf. solutione; gano gain).

To sum up: a close study of national languages reveals the truth that everything in them is not "natural" in the strict sense; and a close study of the best constructed languages shows us that nearly their elements are really just as "natural" as most of the elements of English and French. This should make us give up all the ordinary prejudices against "artificial" languages and make us understand that the introduction of a well-constructed language for international purposes will be a very great benefit indeed for the world at large.

The art of the perfect gardener is not to make artificial flowers but to select the finest of those plants with which nature provides him, to arrange them so that they form a harmonious whole, and perhaps to produce new species by means of the same processes (crossing and mutation), that Nature herself employs. This also describes the task of the interlinguist, who may finally quote two profound utterances of two great poets.

Goethe says:

Natur und kunst, sie schienen sich zu fliehen,
Und ahben sich, eh' man es denkt, gefunden.

(Nature and art seemed to shun one another, and look! they have met unexpectedly.)
And Shakespeare:

Yet nature is made better by no mean,
But nature makes that mean: so, over that art
Which you say adds to nature, is an art
That nature makes ..... This is an art
Which does make nature, change it rather, but
The art itself is nature.

Linguistica, 1933.

Back to Part 1 of this article

(5) Cf. Dutch fiets 'bike'.

(6) On recent endeavours among technicians to bring about linguistic norms for the creation of new words and to settle the meanings of technical terms see E. Wüster, Internationale sprachnormung in der technik (Berlin, VDI-verlag, 1931).

(7) I must refrain here from a discussion of conditions in Norway, where the conflict between Dano-Norwegian (which to parts of the population was more or less artificial) and Ivar Aasen's half-artificial, half-natural landsmål had not yet led to a truly national language.

(8) I have tried to carry out these principles in Novial, see my books "An International Language" (G. Allen & Unwin, London, 1928), German edition "Eine internationale sprache" (Winter, Heidelberg, 1928) - and "Novial lexike" (same editors, 1930).

(9) Details in spelling and endings may be open to discussion. I give the words in the form I think the best for international use.

Back to International Auxiliary Languages

James Chandler 03-Jun-01