- Oren Watson - My Thoughts On Lojban ===================== Lojban is a constructed language created in 1987 as a successor to Loglan, and for much the same goals. It is meant to be unambiguous, structurally sound, simple and easy to learn, and international. However, while I think its goals are laudable, I don't agree with the choices made in its development, and there's two things in Lojban's design which I think are very large mistakes. The first major mistake in Lojban's design is in its phonology. As noted in reference #1 at the bottom, Lojban has 6 vowels. The first five are the A E I O U common in continental european languages, which are also common elsewhere in the world. The last is schwa, represented by the letter Y. Schwa is common, but hard to make intentionally. However, I could overlook the inclusion of schwa in the vowel system, if it didn't also cause problems with another feature of Lojban, consonant cluster buffering #2. Lojban words can include difficult consonant clusters, but these can be "buffered" with a vowel. In most languages, the obvious vowel for this purpose would be none other than the schwa, or possibly a vowel of a nearby syllable. In Lojban however, both the schwa, and all other vowels in the language, are forbidden from being used to buffer the consonant clusters. Instead some other vowel sound must be used; but this vowel sound isn't specified! This causes problems for human speakers, who must therefore either be able to pronounce all the lojban consonant clusters, or be able to pronounce seven distinct vowel sounds. This also causes problems for human listeners, who might not recognize whichever vowel a speaker chooses to buffer their clusters with; and even for computer processing of speech, which would have to somehow recognize a variety of buffer vowels, including the ones noted in #2. It would be better to have 'y' represent a high central vowel like those, and buffer clusters with a schwa. The second major problem with Lojban's design has to do with the basic grammatical structure of the bridi. In a bridi, the selbri can have up to five sumti. In a natural language, this is akin to a verb that has 5 completely unmarked participants. It is my opinion that anything more than 3 is too many; and only 2 is ideal. In english for example, only a very few, very common words can have 3 participants unmarked with prepositions, and none 4. Further, selbri with similar meanings can differ widely in their place structure, particularly with regards to the last two sumti. For example, consider the selbri: ciska: x1 writes x2 on x3 using x4 cusku: x1 says x2 for audience x3 via medium x4 benji: x1 transmits x2 to receiver x3 from origin x4 via x5 As a result, one has to remember what roles the two to five participants in a bridi have for each selbri, in addition to the raw meaning of the selbri. The participants of a bridi can be rearranged into different orders, and marked with particles #3, #4. However, these particles merely tell you which numbered role a sumti goes into, and nothing about what that role actually is. As a result the use of marked sumti causes more confusion, not less. It would be better if Lojban had a fixed word order with only two unmarked sumti, and marked other participants with semantic roles, rather than marking positions that have arbitrary semantics. These two problems with Lojban's design are not the only ones. There are smaller nitpicks which can be made regarding orthography, internationality, and vocabulary. However, these two reflect choices made which I think are not only poor, but illogical and cause confusion; they conflict with the basic goal of the language, to be logical and unambiguous. #1 https://lojban.github.io/cll/3/2/ #2 https://lojban.github.io/cll/3/8/ #3 https://lojban.github.io/cll/2/7/ #4 https://lojban.github.io/cll/9/3/