All chapters →

3. Vocabulary and word formation

The vocabulary is completely newly invented (a priori), but sensibly connected: related ideas share a root, and new ideas are formed by compounding.

All roots are in the dictionary.

Roots

In the lessons: Lesson 2: The first sentence

Roots begin and end with a consonant (e.g. nam). On their own they are not words; only an ending makes them pronounceable.

Syllable block: no root begins with a short-word syllable (list in chapter 6). Because the stress falls on the second-to-last syllable, a short word in front of a word would otherwise sound like a single longer word (mi name like one word “miname”). That leaves 33 syllables for the beginning of roots (75 minus 40 short words, minus vu- and “ji”).

Basic type: every root has exactly one basic type: action, quality, thing or quantity. An idea has exactly one root: “food” and “eating” are one idea (nama), and so are “language” and “speaking” (sena).

Basic type of formed words: a derived or compound word has the basic type of its idea, that is, of the part of speech of its basic form in the dictionary: noun = thing, verb = action, adjective = quality, quantity word = quantity. All other endings follow from this: namalime means “to be a cucumber” (cucumber is a noun), ravenare “to be a teacher”, vesagefa “reading” (to read is a verb). Hunger is a noun, so duraname means “to be a hunger”; “to be hungry” = bure duranama (to have hunger).

Sound shapes of roots

Length by rank: the most frequent ideas and those that form the most words have short roots of consonant + vowel + consonant, that is, two-syllable words (nima water, vese to see). Rarer ideas (level 3) have long roots of consonant + vowel + consonant + vowel + consonant, that is, three-syllable words (dorala elephant, sereme to exchange). About 400 short roots are possible; 278 are taken (15 of them number words). Because roots must be told apart, only about 25 to 55 free forms remain in each domain. This reserve is for very frequent ideas and for hub ideas from which many words are formed.

Distinctness: two roots that differ in only one sound are forbidden

• if they belong to the same domain (they come up in the same conversations), or
• if they have the same basic type and the difference is a close sound pair: b/v, p/b, t/d, k/g, f/v, m/n, l/r. Many speakers confuse these pairs (b/v in Spanish, l/r in Japanese).

That is why the former “tal” (big) was replaced by bir: “tal” and tol (clear) are both qualities in the same domain and differ only in the vowel.

For short words the sound-pair rule is only a guideline. They stand in fixed places in the sentence, and the position separates similar pairs: ka/ga (question at the start of the sentence or after a word, “with” before a noun), te/de (“than” after ro, “that” at the start of a clause). le/re are also distinct enough. For short words only this is strict: opposite pairs differ in the consonant.

Long roots: a long root never begins with a short root and never ends in a suffix (-ar, -ev, -ot, -is, -en). Otherwise long root + suffix would be the same as a compound: “nimal” + -ar- + -a would give nimalara, and that already means rain (nim water + lar to fall). For the same reason a new short root must not be the beginning of a long one.

Word length: a word has at most 5 syllables; for frequent words the aim is 4. If a word gets too long, there are three ways out: a root of its own for an idea that is part of many words (tree gera instead of “plant + big”, so that forest is nisagera instead of “nisabimabira”), a simpler formation (to bring = to hold + to come) or a phrase of two words. A word of 5 syllables takes no further building block. Ideas from which more words are formed therefore get a root of their own early: telephone gonela instead of “senalipota”, so that mobile phone gonelisa is possible.

Selection rules when inventing. They only concern the choice of new forms, not the grammar; older roots such as lah (five) remain valid.

• No j or h at the end: with the ending -i this would give “ji”, and h between two vowels is easily lost in fast speech.
• Not the same consonant twice (“nan”), so that the root stays clear.
• No form that is a well-known word with a different meaning in a widely spoken language or sounds vulgar.
• Three different consonants in long roots.

Endings show the part of speech

In the lessons: Lesson 2: The first sentence · Lesson 3: Qualities · Lesson 4: Being without “to be” · Lesson 12: Want, can, must

Each ending has exactly one job, whatever the root:

• -e on thing roots means “to be an X”: dova hi dove. = This is a house. It is the same rule as bire = to be big. A thing word never becomes an action (“to house”), because that would have to be learnt separately for every root (water + -e: to drink? to pour? to swim?).
• -i on thing roots means “like an X”. “Belonging to the house” or “made of stone” is said with noun + noun.
• -u exists only on quantity roots. A word like “namu” is not formed (just as vu- exists only where there is an opposite).

With -u the last free ending is taken. Future parts of speech have to be handled with short words.

Derivation elements

In the lessons: Lesson 3: Qualities · Lesson 11: Building words

Prefixes consist of consonant + vowel, suffixes of vowel + consonant. This way the syllable pattern is always kept.

Like the endings, each building block has exactly one job that applies to every basic type:

-en- replaces the earlier formation with “make” as a separate part (“binelage” for to open) and saves a syllable. It is not the same as -ot-: lumota is the lamp (the object that makes light), lumene means to make bright.

There is no building block for “done” (participle). “The bought house” is said with a relative clause (dova je mi nide po); frequent states have a quality root of their own (open, broken, finished, known).

Compounds

In the lessons: Lesson 11: Building words

The head word comes first, the specifying word follows. The ending at the end of the word always shows the part of speech of the whole word.

Linking vowel: between two parts there is always -a-; only the last ending shows the part of speech. To read, from “see” + “write”, is vesagefe, hunger, from “want” + “eat”, is duranama, grey, from colour + stone, is nusaboti. A built-in idea keeps its own stem: university = school (raveva) + big → ravevabira. This way action and thing stay one idea, as with roots: senaduke to sing, senaduka singing, song. The linking vowel used to depend on the part of speech; that had to be known separately for every word.

• nama (food) + lim (green) → namalima = cucumber
• written separately: nama limi = “green food”; written together: a fixed term (like “glove”)
• In speech, a compound and a chain of nouns differ only in stress: nisagera (forest, ni-sa-ge-ra) versus nisa gera (“place of the tree”, ni-sa ge-ra). A fixed term should therefore never mean something completely different from its chain.
• Proper names are never compounded; they always stand as a separate word: rava Senatola (see chapter 4, noun + noun).

Example: the word family sen

In the lessons: Lesson 11: Building words

All chapters →