First Steps in Lojban

So wait — Sevan showed me earlier that you can type .o'imuXAGjisofyBAKnicuZVAtilePURdi (all one string, no spaces) and a Lojban parser can still tell what every word is?

Yes. That's not a trick — it's a core design property called audio-visual isomorphism (AVi for short).

Iso-what-now?

"Isomorphism" means "same shape." The idea is:

• Any properly spoken Lojban utterance can be uniquely written down.
• Any properly written Lojban text can be uniquely read aloud.

This is stricter than most languages. In Japanese, for example, a string of hiragana like くるまでまとう could parse as "waiting by car" or "waiting until [he] comes" — the boundary between words is ambiguous. Lojban forbids that.

Recall the three word types (Lesson 3):

• brivla — has a consonant cluster in the first five letters, ends in a vowel.
• cmavo — no internal consonants except possibly the very first, ends in a vowel.
• cmevla — ends in a consonant.

These shapes were engineered so that once you read a stream of sounds (including where the stress falls), there is exactly one way to cut it into words. Every word boundary is recoverable from the sound alone.

That's why you can ditch spaces if you include stress marking?

Exactly. Spaces are a courtesy to the reader, not a grammatical necessity. Lojban's design makes them optional — as long as stress is explicit, the string is still unambiguous.

(That said: please use spaces when writing to humans. Reserve the spaceless form for puzzles and Twitter character limits.)

Now you can fully understand why certain dots are required:

1. Before vowel-initial words — .i, .abu, .iu, etc. Without the dot, a preceding final vowel might run into the opening vowel and create ambiguity about where one word ends and the next begins.
2. After words ending in y — .y. is a cmavo; the trailing dot keeps it from being read as the start of the next word.
3. Around cmevla — la .soran. has a dot at each end, clearly marking the name as a single unit separate from surrounding words.

All three rules are there for the same reason: to keep word boundaries clear.

So "omajinaï" (Koshon's word for the dots back in Lesson 2) was actually "word boundary markers"!

Right. Koshon skipped the explanation to keep Lesson 2 simple. Now you know.

This design has a few practical consequences:

So the word-shape rules I grumbled about learning aren't arbitrary — they're load-bearing architecture.

Exactly. Lojban is less like a collection of rules-for-rules'-sake and more like an engineered system where each constraint earns its keep. Take out one rule and something else breaks.

That's actually… kind of cool? It's more like a proof than a language.

Some people say exactly that. Whether that's a feature or a bug is left as an exercise for the learner.