lojban calligraphy

Semantic issues

Why were the forms of the gismu (root words) created from several languages, rather than just one, for instance English?

Because of the ideal of cultural neutrality, one wanted to make the learnability of the vocabulary the same, regardless of one's native language.

Also, making the vocabulary of Lojban too similar to a single language (English has been proposed many times, because it is widely used) would make learners who were familiar with that language import semantic and cultural baggage from that language -- what in linguistics terms is called a substrate language.

Why isn't the default base of Lojban numerals base-sixteen?

Because base-ten seems to be the easiest for humans to use.

Why is there a gismu (root word) for X?

Depending on the specific X, the answer could be one or more of the following:

• It covers one of the base categories in Eaton's thesaurus
• It is a common or important concept

• The concept cannot be expressed by combining other gismu into a tanru (phrase compounds) or lujvo (affix compounds)
• It seems to be useful in compounds

Why isn't there a gismu (root word) for X?

Depending on the specific X, the answer could be one or more of the following:

• The concept can be expressed by combining other gismu into a tanru or lujvo
• The concept is specific, or infrequently needed

Phonological issues

If rafsi (combining forms) unambiguously refer to one gismu (root word), and it is shorter than a gismu, why have not CVCCV/CCVCV gismu at all? Why not use rafsi as brivla (predicate words)?

Because the current system of self-segregation would break down. Specifically, tanru (phrase compounds) and lujvo (affix compounds) would be indistinguishable. Tanru and lujvo have widely differing semantic properties. Tanru are vague and made up on the fly; lujvo are specific and re-used over and over again. In linguistics terms, lujvo are lexicalised, while tanru are not. Such a language would either have no lexicalised compounds (and hence the vocabulary would be expandable through loan words, if at all), or it would be like some natural languages, in which, upon seeing some unknown sequence of words, can't be certain whether they are meant as a sequence or as a unit with a specific meaning.

Also, some rafsi have the same form as cmavo (grammatical words), so they would be ambiguous.

An artificial language that does use single-syllable root words is eqli.

Why weren't numeral words selected so that if sorted in alphabetical order, a number would also sort in numerical order?

The idea likely didn't occur to the designers at that time.

Also, the common way of assigning series of cmavo (grammatical words) is to have the initial consonant the same, but vowels in ascending order (a, e, i, o, u; or i, a, u). Doing this with numerals is not possible, because there are more than five of them. Also, an important principle in creating Lojban numerals was that they were to be maximally distinct.

Why does Lojban use the ' (apostrophe) symbol for the /h/ phoneme, instead of h?