For the THEORY OF WORD BUILDING reference is made to the Interlingua-English Dictionary, "Introduction," pp. xlv-xlix.'
The following observations may serve to elucidate the problem. In a discussion of, let us say the English verb, it is customary to take up a variety of formations under the heading of "special finite forms." Thus "interests" for instance is called the third-person present-tense form of the verb "to interest." But also the participles "interesting" and "interested" are treated as forms of the verb although they function quite freely as independent adjectives and might as well be called derivatives. There is nothing wrong in this, but if the adjectives "interesting" and "interested" are regarded as special forms of "to interest," then it is not easy to see why one should not be allowed to speak not only let us say of "teaches, reteach, taught, teaching" but also of "teacher, teachings, teachery, teacherly, teacheress, school-teaching, etc." as special forms of "to teach."
This somewhat unconventional way of looking at "derivatives" is extremely useful in an auxiliary language because it casts a very clear light on the principles of free or autonomous word building. Just as in English no one consults a dictionary to ascertain whether he may use the form "weighs" on the basis of the infinitive "to weigh," so no one - if English were a constructed auxiliary language - should feel the need to consult a dictionary before building words of the type of "roachy" (on "roach" after the model of "lousy") or "deedlessness" (on "deedless").
In English or any other traditional language convention and usage will set a hampering limit to this type of word building. Not so in Interlingua. Here the limit is set by usefulness and clarity.
The first of these points, that of "usefulness," is hardly in need of further comment. Interlingua - like any traditional language - can build large numbers of "special forms" (or derivatives) which no one ever needs. A "rainer" - to use an English example - is "man who rains," but to most people it will come as a surprise that this word is actually listed in dictionaries.
The prerequisite of clarity leads to the formulation that in the auxiliary language only those formative elements can be tolerated in free or autonomous formation which appear throughout the language as a whole with clearly definable values. If the English formation "mountibund" is impossible, the reason is not that "to mount" and "-ibund" fail to occur. They do. But "-ibund" is incapable of active formation because its occurrence in "moribund" and "furibund" does not give it an easily definable value.
Free formations must be so clear that it makes perfect sense to speak of them as "special forms" of the base word from which they are derived.
The BASIC PRINCIPLE OF PRACTICAL WORD BUILDING in Interlingua is analogical. Every new formation must clearly be patterned after a model in the established vocabulary. A description of the most common structural patterns in the established vocabulary coincides with the prescription governing new formations.
In the succeeding paragraphs the most important structural types are described, illustrated, and applied to new formations under the headings listed below.
The term DERIVATION is here used to refer to word building by means of suffixes. In derivation the part of speech of the resulting formation is determined by the suffix. Nouns and adjectives consisting of stem and termination lose the latter in the derivational process. The termination of the derivative is part of the suffix. Terminations in this sense are, in the case of nouns and adjectives the final vowels a, e, o, and foreign endings like Latin -us, -um or Greek -os, -is, etc. For verbs, see §146 below.
In most instances the joining of stem and affix is a matter of simple juxtaposition.
When the stem ends in -i- and the suffix begins with the same vowel, the full derivative is spelled with only one -i-: rubie 'red' plus -ificar > rubificar 'to redden, make red.'
Note that the addition of suffixes with initial vowels to stems ending in -c- may change the sound value of that consonant. When through derivation an originally hard c (as in franc) comes to precede i or e, its pronunciation changes as a rule to soft c (as in Francia). Inversely, when an originally soft c (as in cortice 'bark') comes through derivation to precede an a or o, its pronunciation changes as a rule to hard c (as in cortical). However, a soft c remains soft c before the suffixes -ada, -ata, -age, -alia, -astro, -astra and must be spelled -ci-: nuce 'nut' plus -ada = nuciada. Similarly hard c remains hard c before the suffixes -eria, -ero, -esc, -ese, -essa, -etta, -ette, -etto, -iera, -iero, -issime, -issimo and must be spelled -ch-: porco 'pig' plus -eria = porcheria 'pork shop.'
Note that the seeming irregularity of an example like ricchessa 'riches' = ric 'rich' is due to the fact that cc is spelled c when it appears at the end of a word.
Words ending in -age (whether this group of sounds be a suffix or not) retain the soft pronunciation of -g- (like z in 'azure') in derivatives based on them. When the suffix begins in a or o, the g is replaced by -gi-: orange plus -ada = orangiada; but orange plus -eria = orangeria 'orange greenhouse.'