Lojban terms are used for most parts of speech, except for quantifiers and descriptors. Needlessly inconsistent?
In particular (this list is not exhaustive),
quantifier->something with {snicne}? (though personally I (zort) disagree with the veljvo of {snicne})
quantified variable->something with {snicne}?
descriptor->gadri
description->gadri sumti (I haven't actually seen this used)
attitudinal->cnima'o
terminator->fa'orma'o
modal->sumti tcita
It would be better if formal definitions were separated from comments, and not going mixed.
Maybe even a separate chapter with formal definitions for each group of cmavo. More detailed than the word list, less detailed than the CLL.
It would be better if numbered were not only chapters and sections, but also sentences like in the Bible.
Exactly what would be the benefit of adding "verse" numbers? We can easily cite specific sentences without the need to do this. For example, citing from the online refgram: 9.3.7.2 (Chapter 9, section 3, paragraph 7, sentence 2): "Therefore, it is perfectly all right to scramble...." Or for examples: 9.3-3.3 (Chapter 9, section 3, example 3.3): "3.3) klama fa mi fi la .atlantas. ...." mi'e.aionys.
There should be more illustration and diagrams. For instance, with termsets, show with lines and/or colour coding how two sentences are equivalent. See The Japanese Wikipedia for some other examples.
Chapter 4 needs to go somewhere else, ideally near the end. It's crazy to give all that crap to newbies.
Minimum 2 sentence explanation for every cmavo (excluding ones that are part of a pattern like {fo'i}, of course). Chapter 13 is particularly guilty of the length of the explanation of a cmavo seemingly being inversely proportional to the number of cmavo explained in the section; some just get two gloss words worth of explanation, and it's rather opaque (I (zort) would be willing to write more about them).
Maybe the fact that denpabu are optional in writing should be reinforced, since people seem to forget that. Right now it's only mentioned once, early in the book, in a little paragraph in chapter 3. It could be said again in chapter 19 and/or chapter 4.
Typographical
The brush-stroke placeholders for logical connectives should be written in "blackboard bold" instead, to make it look more form
I remember this section being confusing to me because of the brushstroke letters. Suddenly the book became much more formal (in the mathematical sense) and I didn't know what that signified. Writing them as of "-a -e -i -o" might be better, since the hyphen indicates they are not themselves words, without introducing a strange new typographical convention. The book already uses hyphens to show that rafsi are not words, so this would be consistent. Or since "a e i o" actually are words, perhaps they could be used directly. mi'e.paldanyli.
Per section
Chapter 3
Section 6
Section 6, restriction no. 2 already forbids 8 of the 12 consonant pairs that are forbidden by restriction no. 3; perhaps it's better to just explicitly forbid the pairs "cs", "jz", "sc" and "zj" like in the last restriction.
I kind of like that the explanation for those pairs being forbidden is given, even if it includes things already forbidden. I would like it slightly better if the word "sibilant" were used. Perhaps for consistency the reason could be given for the last restriction instead of listing seemingly-random pairs. .imi'e .skaryzgik.
Agreed mi'e.aionys.
Section 9
A bit after example 9.5, the explanation of how to stress words is ambiguous. A possible incorrect interpretation:
set stressed_syllable = penultimate_syllable;
while (stressed_syllable contains "l|m|n|r|y"
&& stressed_syllable != first_syllable) {
stressed_syllable = syllable_before(stressed_syllable);
}
None of the examples demonstrate that only syllables containing "l|m|n|r" as syllables on their own are to be skipped, nor that skipped syllables are to be ignored before you start counting as opposed to skipped while you're counting. The following examples would make it clearer: {djanatyn} is pronounced {DJAnatyn} (not {djaNAtyn}), and {patrica} is pronounced {paTRIca} (not {PAtrica}).
Section 12
The tengwar table should use actual tengwar in addition to their names, since we now have Unicode (CSUR) and good fonts.
Chapter 4
Section 7
In the procedure for making a non-Lojban word into a valid Stage 3 fu'ivla, double consonants are to be eliminated before the sounds are to be converted to their closest Lojban equivalents, but it is possible that consecutive consonants have different sounds (like in 'eccentric'). So these actions should swap places. The same goes for section 8, in the procedure for Lojbanizing a name.
Chapter 10
It would be nice if there was a template for compound tenses around the end of the chapter, like the template for compound cnima'o in chapter 13 section 8.
Chapter 16
Section 9
It says "for no x" (noda) is the same as "it is false for some x" (naku su'oda). I (mi'e zort) interpret "it is false for some x" as "there is an x such that it is false" (su'oda naku), not "it is false that for some x it is true" (naku su'oda), so it should be changed to "it is false that for some x".