[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >
I agree that it is a wise choice to start a conlang with few words, unless one is sure that his/her language will have lots of adherents (and it is not easy to get them in a list where everybody has his/her own conlang). Otherwise, it might be an enormous waste of time. If we had a consensus, we could create it together. I had a very long conversation by e-mail with the already famous anti-Esperantist Justin B. Rye. When I showed him a possible set of rules to make an alternative to Lojban with phonotactics no more complex than those of Japanese or Swahili (but still keeping word-break detection), he answered with a simplified version of it that was much easier to state: 1. Every word can be described by [(C)V(N)] *; 2. every word begins with C; 3. every word ends with CV; 4. every heard sequence CVC must be understood as a word break: -CV C-; 5. a vowel may be alternatively pronounced as a semivowel when it is in a cluster with other vowel. I was astonished to see how many real words of many languages are covered by these rules: koala, fauna, Hindi, Kyoto, kanji, la, le, tu, canto, toada, mango, cauda, pauta, Bantu, Suomi, jambo, bandeira, Simba, mense, Ruanda, huasi, house (if the "e" were still pronounced in modern English), geisha, etc. It's funny to note that a native anglophone, whose language accept very complex syllables, would probably find more difficult to pronounce some of these words than speaker of languages with simpler phonotactics, because native anglophones pronounce final "o" and "e" of foreign languages as /ou/ and /ei/, and this would mess the word-break detection of a conlang with the above mentioned rules. *: key to notation below: C: consonant; V: vowel; N: nasal stop (with articulation dependent of the following consonant); [ ]: repeatable structure; ( ): optional element; -: represents the rest of a word.