As already remarked, a would-be international language must be built up as far as possible on that word material which is already international, i.e. known all over that part of the world where European civilization has penetrated. The form may vary a little from country to country, but there are thousands of words that are already common to practically all nations, thanks to the sameness, or at any rate similarity, of material and spiritual culture in modern life.
There can, therefore, be no doubt that our I.A.L. must adopt such words as the following - I give them in the form they must have according to the principles developed in previous sections: first, names of many of the products of civilization, good and evil: te (tea), kafe (coffee), chokolate, vine, sigare, sigarete, tabake (tobacco), alkahole, opium, telefone, radio, kanone, pistole, paper, lampe, karte (card), bilete, marmore, bensine, diamante, olie ...; further, the following international names of animals and plants which hail ultimately from the most diverse countries and languages, even if some of them have come to us through Latin: elefante, tigre, lemure, gorila, girafe, rose, lilie, banane, tulipane (tulipe?), jasmine (jessamine), palme, begonia.
Further, we have a great many words like the following which show community of ideas: nature, kulture, literature, musike, poete, komedie, tragedie. opera, kore (choir), rime (rhyme), forme, danse, persone, studie, universitate, profesore, akademie, botanike, medisine, kemie, fisike, kongrese, konflikte, diplome, poste, minute, sekunde, grame, septembre ... Adjectives like komplet, ordinar, enormi, ideal, universal; verbs like stenografa, aranja, representa, dekora, examina, improvisa ... All these, and many more, are practically universal in that part of the world with which we are principally concerned.
But there are, of course, numerous notions for which no such completely international name exists. The task then is to find the most international name and use that, even if there may be countries where it is unknown. It is always a question of degree, but how are we to measure the degree of internationality where two or more names are found? The authors of Idiom Neutral were the first consciously to carry out the principle of choosing everywhere the most international word - for, as already remarked, Zamenhof and other had applied it only instinctively and haphazardly - but the procedure of Idiom Neutral was somewhat superficial, for in each particular case they calculated the number of languages to which a given word was common, including in their calculations Latin, which should not be placed on a par with living languages. Now the essential thing is not the number of languages, for languages are not organisms possessing an individual existence apart from those human beings who speak them. The proper rule, therefore, for determining the internationality of a word or a stem is to count the number of people who know it from their mother-tongue - a simple consequence of the fundamental principle of the greatest facility for the greatest number. It is only natural that each person would prefer to find in the I.A.L. as many words as possible that are already familiar to him, and so, to be impartial, we must attach the same importance to the natural preferences of each of the 150 or 160 millions who speak English as to each of the 120 millions who speak German, or the 90 million Russians, the 60 millions who speak Spanish or French, the 40 million Italians, etc. Even the languages spoken by the smaller nations should be taken into account in proportion to their number.
The choice of words for our neutral language is therefore to a large extent a question of mere numbers. And yet all is not settled by the statistics of the people using each word, for on closer inspection several puzzling questions confront those who want a vocabulary that is satisfactory in every respect, and sometimes unsuspected difficulties in form or meaning are discovered when a word that seemed international enough is put to the test of being used in various contexts or having all kinds of regular derivatives formed from it.
Some of these difficulties have already been mentioned in the phonetic chapter: even if the "same word" is found in several languages it is very often under different forms - mostly due to a different phonetic development - so that the choice of form is often a delicate problem. The English and French may spell change alike, but the sounds are very different; and as we can neither take the F nasal vowel nor the diphthong (ei) as found in the most usual E pronunciation, chanj- seems to be the most convenient form for all. In many cases we have to find what might be termed a common denominator between the divergent forms. If E and D did not differ in the form of many etymologically related words to such an extent that a middle form cannot be found, the Germanic (Gothonic) quota would be predominant, on account of the great number of people speaking those two closely connected languages (cf. water wasser; tooth zahn; speak sprechen; soap seife; week woche; enough genug, etc.). Such being the case, it is the Romanic element of English which weighs down the scale in most cases, since it agrees with F I S, or at least with one of these three languages, and the result is that our I.A.L. will necessarily be very much more Romanic in appearance than might be anticipated.
One very important factor is also operating in the same direction, the fact namely that a large number of Latin derivatives have passed into D (and Sc), even when the primary word itself is not found there. Thus we have in D absentieren, abstinenz, artist, dentist, dental, moral, popular, which greatly facilitate for a German the learning of the words absent, absteni, arte, dente, more, popule, although these words are not found in his own language (apart from pöbel = populache).
In the same way, though the numeral for 100 is practically the same in E D Sc, we should not take hundred for that idea, for the L centum is known all over the world, even in Germany, etc., through per cent. (D prozent) centesimal, centimeter, E centennial, century, centenary, D zentner, Dan centner, the coin cent or centime, etc. (N sent).
Still, it should be remembered that it is not enough that a word is found in a dictionary of a language to make it easy for the man in the street to understand it: many words are so special and scientific that they are only known to a minority of the nation. In order to obtain a word for `bird' it is of very little practical use to go to the word ornithology, and my admired teacher Henry Sweet is unusually wide of the mark when he writes (in the Encycl. Brit.): "Thus the Idiom Neutral ornit `bird' and diurn `day,' are almost self-interpreting even apart from any context, while the Esperanto bird and tag are unintelligible except to those who know English and German." Ornit and diurn will be immediately understood by much less than one per cent. of those who at once understand the two Esp words (though I do not say that these are very well chosen). It is the living knowledge only that is of value.
Let me give a small selection of words justified in the way here indicated by statistics; in each case the initials show the languages used as sources: separa D E F I R S Sc, transporta D E F I R S Sc, transparent E F I S Sc, lauda E F I S, rida F I S, sam E R Sc, blu D E F Sc, bora D E R Sc, blind D E Sc, jena D F Sc, no E F I S, yes E.
Now there is one objection to a language constructed on these principles which must be met here: many people want a homogeneous language and dislike seeing words like sam, blind or yes in a language of so Romanic a type as ours. The objection is emotional or æsthetic rather than intellectual, and perhaps it would be sufficient to rule it out of court by saying simply that we want a practical language, no matter how it looks. Still it may bed worth while to argue the point a little.
Not a single one among the 5,000 (?) languages spoken in the world is perfectly homogeneous, why should an artificial language then be so? There is in many countries a natural reaction against those loan-words which are too obviously intrusive either in sounds or inflexion or which are superfluous because the same idea can be expressed naturally and easily enough through native speech material; but such considerations have no weight in an I.A.L. if the words are well chosen and contain the same sounds and sound-combinations as the rest of the vocabulary. No one objects in E to words like tea, religion, panorama, coach, biography, etc., on the score that they have come from other languages, though nearly everybody prefers short handy "Saxon" words to those long and learned words which pedantry has introduced - but that preference is really an outcome of the principle here advocated, of using the word that is most familiar to the ordinary man.
Curiously enough many people, especially in Romanic countries, object to the immixture of Germanic words like blind or nur in a Romanic language, while they have no strong feelings against Romanic words in a Germanic language. Yet, even so fastidious a nation as the French have adopted or are adopting a certain number of words from E and D: bifteck, rosbif, leader, club, ulster, grog, sleeping-car, football, handicap, sandwich, bitter, bock, kirsch; cf. also fiord, ski, saga from Sc.
Most languages have even examples of hybridism, i.e. words composed of elements from different sources, like E eatable, artless, beautiful, mileage, hindrance; F bureaucrate, blackbouler, snobesse or snobinette, smarteux. Examples of hybrid formations in N have been given above under some of the suffixes.
Such a word as blind is immediately recognized by more than 200 millions, while neither F aveugle (even apart from its vowel) nor I cieco, S ciego, P cego attains nearly the same number (F cécité is not a popular word).
Extensions of the vocabulary are made possible through composition and through word-formation by means of suffixes and prefixes; and it is one of the great advantages of an I.A.L. that there are no limits to these procedures, such as are very often met with in national languages, where formations that are perfectly regular and in accordance with several other words are nevertheless proscribed through tyranny of usage.
Compounds are often very convenient; on the other hand it should not be overlooked that they generally presuppose some element, a preposition or even more, which must be left for the hearer to supply. It is also in some cases more or less arbitrary what components are selected; users of a constructed language are often guided here by their own language without taking into consideration that other nations might dissolve the idea in a different way. Convention come to play a rôle here as in national languages: for `railway' Esp, Ido and N use fervoyo, fervie, which is modelled on D eisenbahn, F chemin de fer, but might just as well have chosen rel-vie like E.
With regard to the number of words or roots to be used in an I.A.L. there are two conflicting principles, both of which are legitimate within certain limits, but which should neither of them be carried to excess, the principle of precision and the principle of economy.
According to the former an I.A.L. should be capable of rendering all shades of thought found in the national languages. As each of these expresses several nuances which are incapable of exact rendering in all other languages, it is clear that the ideal as thus formulated is beyond human power. It will also be found that those who set forth this claim generally have in view only certain nice distinctions that are found in their own, or perhaps one more language known to them, and which appears to them indispensable, while they are not considering the burden thus laid on all the rest of mankind who have never felt the need of those particular nuances. Let me translate here a note I inserted in Progreso (February 1911): "I should like very earnestly and emphatically to insist on an important point, before it is too late: do not let us create too many special words to distinguish nuances that are not absolutely necessary. I think there is a danger in the increasing tendency in that direction. It is easy to understand that nuances can often be necessary or desirable, even where many languages are unable to express them; but many of the proposed nuances can only serve to make the language difficult without any real advantage. (I then give as example the proposal to have three verbs for `to tame': domtar leoni, amansar uceli (birds), domestikigar elefanti, and I conclude:) If one goes on to increase the bulk of the dictionary in that way, I humbly ask people to cease printing over and over again the phrase about `easiest for the greatest number' which I wrote unsuspectingly a couple of years ago." Even now some interlinguists look upon it as their principal task to multiply words without regard to the convenience of users.
The opposite principle of economy was practised by Zamenhof more than by anyone else. The number of roots admitted in primitive Esperanto was extremely small, and a good deal of ingenuity was used to express as much as possible by means of compounds and derivatives. Foremost of the means employed must be mentioned mal- to denote the opposite: malbona bad, maldekstra left, maldolcha bitter, malsato, hunger, etc. Other examples: kreskajho plant (properly `thing growing', as if nothing but plants were capable of growing), irilo stilt (`instrument for going'), pafilego cannon (`big tool for popping'), fajrero spark (`unit of fire'), senkulpigi excuse (`make to be without fault'), vagonaro train (`collection of cars'). Each of these has of course to be learnt separately just as well as special independent words, though they have the advantage, when once learnt, to be firmly associated with well-known ordinary words. There is no doubt that the great number of these rebus-words, together with the "masquerading" of several words from the best-known languages, has deterred many intelligent people from Esp, and incidentally from the idea of an I.A.L. in general.
Both principles are thus faulty when carried through one-sidedly: we must steer between them as between Scylla and Charybdis.
As to the form of international words a difficulty arises from the fact that many Latin verbs had two sometimes pretty widely different stems, both of which are found in our modern languages; as it would obviously be too cumbersome always to carry all these irregularities with us in our I.A.L. we must, in most cases at any rate, select one of them, and the question only is, which one? The want of system in our national languages is seen when we compare the E verb conduct with its derived conducive, and, on the other hand, produce with productive, or negligible (F négligeable) with neglect, or
E discuss discussible discussion
F discuter discutable discussion
D diskutieren diskutabel diskussion
(In an official document issued by the Société des Nations, 1927, I found the F spelling discution - but t is here pronounced like s.)
Here follows a list of the most important of these two-stem words - arranged according to the final sound of the stems, and with indication of the form selected in Novial, most often the one found in the Latin passive participle, because so many derivatives (chief among them those in -ive and -ion, also -ere = F -eur, E -or) are formed on that basis; E also often uses it as its verb stem:
c/ct: dukte (dedukte, -in, -intro, kon-, pro-, re-, se-), cf. induktione, deduktiv, etc.
inspekte, suspekte; inspektere.
ejekte, in-, kon-, pro-.
dikte `say,' edikte, inter-, bene-, kontre-, pre-.
By the side of dikte with diktione we may take diktate `dictate' with diktatione.
g/ct: akte, reakte with -tione, tiv. Akto `act,' also in play. Aktere one who acts; aktore as independent word `actor' on the stage.
redakte `draw up, edit' (F rédiger), redaktere, redaktione.
lekte `read' with lektione. (N.B. lege = law, with legal.)
elekte, selekte, colekte.
neglekte. (But neglije `dishabille.')
junkte join, ad-, kon-.
distinkte `distinguish,' -tiv, -tione. Adj E distinct is distinktet, but generally klar, separat or diferant renders the idea.
It is best to take korekti as an adjective, with the derived verb korektisa `to correct, F corriger'; korektiso `correction.'
Similarly we want direkti as adj; and then direktione as an independent sb; but the vb E to direct, F diriger, D dirigieren, I dirigere, S dirigir, is now so far removed in meaning from either of these that we must have a separate form: diriga (i.e. conduct, at a concert; dirigere conductor).
Funktione with vb funktiona. No verb from L fungi (D fungieren).
ng/ct: restrikte, astrikte.
pikte `paint,' pikture. kontakte.
fikte, -iv, -ione; `feign' is generally simula; `fiction' in the special E sense is romanaro.
h/ct: atrakte, abstrakte, dis-, ex-, kon-, re-. (But N has no simple verb trakte: `to draw' is tira; trakta `treat.')
-/ct: konstrukte, de-, in-, ob-.
p/pt: konsepte, exepte, persepte. Resepte give a reception to, resiva get.
mp/pt: rupte, korupte, erupte.
b/pt: skripte write; skripto, skripture; deskripte with deskriptiv, deskriptione; kon-, in-, pre-, su-, transkripte.
r/rt: inserte, aserte, deserte; exerte.
rq/rt: kontorte, exorte, retorte.
r/st: digeste, sugeste, konqueste, inqueste, exhauste.
n/nt: invente, -ione; but veni, eventa.
d/s: ofense, extense, pretense, expanse, suspense, digrese, exkluse, konkluse, inkluse, divise, desise, kolise. - aplaudi, spenda.
seda `cede' (sesa cease), konsese. Sukseda `succeed,' suksese/a `success' (this in spite of succession, which must be rendered suksedo, and successive, which must be rendered suksedant: it does not mean `which can ...'; it is impossible to conciliate the two meanings `succeed' and `success' with the actual forms if we are to keep our rules of word-formation).
responda, -o `answer'; responsa `be responsible for' with responsant, responso.
vida `see'; visione as an independent word (it does not mean `seeing!'); but revise `revise,' revisione. - Til rivido `au revoir!'
nd/s: konfuse, difuse, infuse.
nd/(n)s: prenda, komprenda; surprisa, enterprisa.
nd/ns, nt: tense `stretch,' tensione; tenda tend, incline to (tendentie `tendency; but tente tent, tenta tempt, tentatione); atente `attend,' atentione; intente, -tione.
t/s: admise, e-, ko-, kompro-, o-, per-, pro-, re-, sub-, transmise. - konfese, diskuse, sukuse, subverse. Konverte with konvertione better than konverse, which would easily be confused with konversa, whence konversatione.
m/s: prese, deprese, exprese, imprese (make impression, but printa print, F imprimer), komprese, oprese, represe, suprese.
ct/x: flekte, anekte, konekte.
r/s: adhese, kohese, -iv, -ione.
r/rs: kurse `run' (kurso running, course, e.g. course of time; but kursu course of lectures, rate of exchange, price), kursivi skripto, exkurse, exkursione; sukurse (kura cure; curatione cure, F curation, S curación); but konkura `compete,' konkurante `competitor' (kunkurse `run together').
l/ls: pulse, impulse, expulse, propulse.
n/(n)s: posi, `place': positione, aposi, de-, ex-, pro-, su-, transposi. Disposi arrange in various places; dispona have at one's disposal, dispose of: li pekunie es ye vun dispono.
There is thus a fairly great number of cases in which the existence of necessary derivatives in -ione and -iv obliges us to select verbs ending in -e (akte, aktione, aktiv, etc.); similarly the verbal ending -i is claimed by -itione (expedi, expeditione, etc.) and -u by -utione (evolu, evolutione). But apart from these verbs it is best to have the uniform verbal ending -a, as it would be intolerable in each particular case to remember whether the original verb belonged to the first or second or third Latin conjugation; sometimes living languages have different vowels (F vivre, I vivere, S vivir; F tenir, I tenere, S tener). Our forms viva, tena may also be defended by a reference to F participles vivant, tenant (N the same) and adjectives F vivace, tenace (N vivasi, tenasi) and E F tenable (N tenabli). Only in the "small" verbs pove can, deve ought, and have the ending -e has been preferred to -a, as the verbs would perhaps otherwise be too heavy. (Voli on account of volitione.)