We next proceed to the separate consonants; some of these need not detain us long, because there is such agreement between the principal languages with regard to their pronunciation and spelling that there can be no doubt as to their employment in an IAL. This is the case with regard to the following pairs: b p d t g k (see below on g) v f, the former of which represent the voiced and the latter the corresponding voiceless sound: here again we can disregard such smaller nuances as that between an unaspirated ("sharp") p, t, k as in F, and the aspirated sounds in E D Sc. It is best to give everywhere the same sound to t, thus to make the two t's in intentione identical.
Further we have l m n s r h.
Of these, l is best pronounced as in F D without that velar modification which is often found in E and in an exaggerated degree in Russian and Polish. N naturally takes the sound of ng in E sing before g, k (and qu). On s see below under z.
R should be pronounced very distinctly and best with trilling of the point of the tongue; English learners should be particularly careful to pronounce it also before a consonant and thus to make the distinction between such pairs as arm- and am-, bark- and bak-, farm- and fam-, kart- and kat- clearly audible (I have here purposely omitted the final vowels of these words as unimportant for the question which interests us in this place).
This leaves us with the following letters of the ordinary Roman alphabet on our hands: c j q x y, and we must discuss each of them separately. The easiest to deal with is x, which may be considered a convenient, shorthand-like abberviation known to everybody from a great many international words like example, excellent, etc. Not only is this extremely easy for everybody, but it also saves us the necessity of choosing between ks and gz. F and E pronounce [ks] in some words, [gz] in others, the two languages not always being in agreement on the same words; most other languages have always the sound [ks], which will probably be the pronunciation generally adopted in the final IAL, though it does not matter very much if some people may here and there pronounce [gz]. Z's unphonetic spelling kz (ekzameni, etc.) will probably please no one outside of Russia.
Q means the same thing as k, but is scarcely ever used in any language except in the combination qu, which is familiar to most Europeans from words like quality, quarter, etc. Only doctrinaires could object to our using this combination in such international words: it is true that the same sound is then written in two different ways, but by this means we keep the familiar aspect of these words and get rid of a fair number of k's (an awkward letter to write), with no more danger than to give some nations a temptation to pronounce [k] instead of [ku, kw] or [kv]--for both these pronunciations must be allowable for qu.
As already mentioned, we have no use for y as a sign for a vowel; but this letter may with advantage be used as a consonant with the only consonantal value ever given to it, namely as in E yes; it is used for the same sound in S (yerba herb, yegua mare) and internationally in a few words like yak, yatagan, yacht.
Instead of y for this sound some people would perhaps propose the letter j, used in this way in D I Sc. But as this letter is known practically everywhere in such international words as journal, jaloux, taken from F, it is better to use j in these and similar words, attributing to it either the French sound or the English one in journal, jealous (which is the F sound with d preposed): the latter is more distinctive, but it does not matter greatly if some prefer the F sound of j.
The same letter j with the pronuncation just indicated must also be used in some cases in which the traditional spelling is g. This spelling has its historical reason in the development of the Romanic languages, through which the sound [g], found everywhere in classical Latin where g was written, has undergone changes similar to, though not exactly parallel with, those we shall soon see in the case of the corresponding voiceless sound [k], written c. But it would not be practical to use j in all those words in which I E have the sound [dZ] and F [Z] spelt in this way, for in some words D R Sc still have the sound [g], and it will therefore be best (as being easiest to the greatest number) in such words to keep the spelling g with the normal sound of that letter. This is especially the case with some more or less learned words, which it would be awkward to spell with j: geologia, geografia, and others with geo-, genealogia and others in -logia; further, gigante, gimnastike, tragedie, genie, general, original, geste; to these I count also rege 'king' on account of regal, regalia, though Ido has rejo. We must also have gardene on account of D E (Sc), though F has jardin and I giardino. But j is preferable, where in popular words the meaning is more or less changed from what it was in Latin: jena F gêner, jentil, jendarmo, plaje F plage 'beach'. J is also preferable in jeneros generous, which in D Sc is pronounced with an imitation of the F sound. Imaje is better than image, and there is some justification for the Ido differentiation of this word and imagina vb imagine. The ending -aje is found in other words as well: kuraje courage, voyaje, etc. N of course spells intrige without u; similarly garda to guard, gida to guide (u pronounced in I only).
Occ has the courage here to follow the practice of some languages, giving to the letter g two pronunciations, before e, i, y as in F, elsewhere as in D. But this extreme "naturalness" is unnatural in the artificial language, which should have the advantage of being more regular than the national languages. In the grammatical scheme of N the Occ double pronunciation would be particularly intolerable, because we have regular word-formations like lege law, legal; rego kind, rega queen, reges royalties; vagi vague, vagum something vague, etc.
The letter g creates difficulties in still another way, through the historical development of the combination gn into the palatalized sounds of F signe, I segno, S seño. In most cases it will be best to keep the traditional spelling and give to g its usual sound, which is often found in derivatives like E signal, thus digni with indignatione, signe, asigna, insigne, regna, stagna and others. But in other words it seems more natural to spell ni: lorniete, champanie, onione (E onion), linie, viniete, koniak (or keep cognac as a trade-name or proper name, like bordeaux?). If we take kompane for companion, we have the regularly formed derivative kompania company. Besono is from F besoin, not from besogne.