Chapter 17. Text Structure & Quotation#

Sentence Separators: .i#

In written and spoken Lojban, sentences are separated by .i (a pause + the vowel i). Unlike an English period, .i appears between sentences — not after the last one:

mi klama le zarci .i do cadzu le bisli I go to the store. You walk on the ice.

.i implies the sentences are on the same topic or in sequence. The relationship is left vague unless you add a connective directly after .i:

mi klama le zarci .ije do cadzu — I go to the store, and you walk. mi klama le zarci .ini'ibo do cadzu — I go to the store; therefore you walk.

.ibo after .i signals the next sentence is more tightly grouped with the previous one than normal.

---

Paragraphs: ni'o and no'i#

ni'o — new topic / paragraph break

Signals a topic change. Multiple ni'o cmavo in a row indicate a larger-scale shift.

no'i — resume previous topic

Like "getting back to the point…" in English. Resumes the topic in effect before the last ni'o.

ni'o la .teris. cu klama le barda tcadu (New topic:) Terry goes to the big city.

In writing: one ni'o = minor topic shift; ni'oni'o = major shift (also cancels assigned pronouns); ni'oni'oni'o = drastic reset (also resets tenses and indicators).

da'o explicitly cancels all current pronoun assignments (ko'a, ko'e, etc.) without changing the topic.

---

Topic-Comment Sentences: zo'u#

zo'u separates a topic (a sumti) from a comment (a bridi). This is like Chinese topic-comment structure:

le nuzba zo'u mi ba'o djuno The news: I already know. → I already know the news.

le finpe zo'u citka The fish: eat (ambiguous — is the fish eating, or being eaten? Left vague)

zo'u is also used to introduce quantifier prenexes in logic (see Chapter 16):

ro da poi prenu zo'u da morsi For all X that are persons: X dies. → All people die.

---

Quotation#

Lojban has several distinct quotation types, each with different properties:

---

lu … li'u — Direct Quotation of Lojban Text#

lu opens a quotation of grammatical Lojban; li'u closes it. The contents are treated as a sumti (the quoted text itself):

la .alis. cusku lu mi klama le zarci li'u Alice says "I go to the store."

The pronouns inside lu…li'u refer to the speaker of the quotation, not the outer speaker. So mi inside quotes refers to Alice.

li'u is elidable at the end of a bridi but usually kept for clarity.

---

lo'u … le'u — Error/Out-of-Context Quotation#

When quoting something that may not be grammatical or that you want to mark as questionable:

lo'u mi pu le'u cu se cusku la .djan. "mi pu" [fragment] was said by John.

---

zo — Single-Word Quotation#

zo quotes the next single word as a word (not its meaning):

zo klama cu gismu "klama" is a gismu.

zo coi cu cmavo "coi" is a cmavo.

zo works for any single Lojban word, including cmavo. It cannot quote multiple words — use lu…li'u for that.

---

zoi — Foreign-Language Quotation#

zoi quotes text from another language (or any arbitrary string). The syntax uses a delimiter word that appears on both sides and must not appear in the quoted text:

zoi gy. Hello World .gy. "Hello World" (in English)

mi cusku zoi gy. I love Lojban .gy. I say "I love Lojban."

The delimiter (here gy.) can be any Lojban word that doesn't appear in the text being quoted. gy. (the letter G) is conventional for English text.

---

la'o — Foreign Name as Sumti#

la'o uses the same delimiter mechanism as zoi but creates a name sumti from a foreign word:

la'o gy. Aristotle .gy. cu se cusku ro nimcli Aristotle is quoted by every philosopher.

This is the correct way to use actual foreign proper names in Lojban without Lojbanizing them.

---

me and me'u: Sumti-to-Selbri Conversion#

me turns any sumti into a selbri meaning "is a member of the set described by [sumti]":

ta me le'o That is a lion-like thing. (me le'o = is of the type described by le'o)

le karce cu me lo bloti The car is a boat-type thing. (it functions like a boat)

me'u terminates the me construction (elidable at end of selbri).

This lets you use descriptions, names, and other sumti as predicate words, bridging the divide between sumti and selbri.

---

pe'a: Metaphorical Use Marker#

pe'a marks a word as being used metaphorically rather than literally:

mi pe'a klama le skami I metaphorically go to the computer. → I log in / I access the computer.

It signals "I know this isn't literally true — take it as a metaphor or extension."

ji'a (also) and si'a (similarly) are related discourse markers:

ji'a — in addition, also (additive)

mi klama .i ji'a do klama I go. Also, you go.

si'a — similarly

mi klama .i si'a do I go. Similarly, you [go].

---

LAhE: Sumti Qualifiers#

LAhE cmavo wrap a sumti and change what it refers to — they shift reference between a thing, its symbol, and its containing set or mass.

la'e and lu'e are reverse operations — one goes from symbol to referent, the other from referent to symbol:

zo .bab. cmene la'e zo .bab. The word "Bob" is the name of the referent of "Bob". (the actual person Bob)

lu'e la .bab. cmene la .bab. A symbol for Bob is the name of Bob.

tu'a is practically very common — it lets you say "I want something to do with X" without specifying exactly what:

mi djica tu'a le plise I want something involving the apple. (I want to eat it / have it / something)

mi djica tu'a do I desire something involving you. (deliberately vague about what exactly)

Without tu'a, you'd need a full abstraction: mi djica le nu mi citka le plise. With tu'a you can stay economical.

lu'a, lu'i, lu'o, vu'i convert between different ways of grouping referents:

mi tavla lu'a le gunma I talked to a member of the group. (not the whole group)

mi viska lu'i le'i prenu I see the set of people. (as a mathematical set, not just the people)

lu'o mi'a klama We (as a mass) go. (the mass acts together)

All LAhE cmavo terminate with lu'u (elidable in most contexts):

mi nelci la'e lu'u le cukta = mi nelci la'e le cukta

---

BAhE: Emphasis and Nonce Words#

ba'e stresses the next word, like spoken emphasis or written italics. It marks that the following word is the most important part of what's being said:

mi viska ba'e la .djonz. I saw Jones. (not someone else)

ba'e mi viska la .djonz. I saw Jones. (not someone else)

mi ba'e viska la .djonz. I saw Jones. (as opposed to hearing about him)

za'e marks the next word as a nonce coinage — an on-the-spot, ad-hoc word the speaker is inventing for local use, which may not exist in the dictionary:

mi klama la za'e .albeinias. I go to so-called "Albania". (nonce Lojbanization of a foreign name)

le za'e smacu'i cu zvati le zdani The so-called "mouse-neutral-thing" is in the house. (speaker is coining smacu'i on the spot)

za'e is a courtesy to the listener — it says "I know this isn't a standard word, I'm making it up as I go."

---

MAI: Utterance Ordinals#

mai and mo'o turn numbers into ordinal free modifiers — firstly, secondly, lastly, etc. They can appear anywhere in an utterance:

pamai mi klama le zarci Firstly, I go to the store.

remai mi viska le mlatu Secondly, I see the cat.

romai mi sipna Lastly / finally, I sleep. (ro = all/every → "all-thly" = lastly)

Any number can precede mai: cimai = thirdly, vomai = fourthly, etc.

mo'o works the same way but marks higher-level sections (chapters, major divisions) rather than list items:

pamo'o = Section 1 / Chapter 1 remo'o = Section 2

You can combine them: pamo'o pamai = Chapter 1, point 1.

---

FUhE / FUhO: Attitudinal Scope#

Normally an attitudinal applies to one word — the word immediately before it (or the whole utterance if at the start). But fu'e and fu'o let you explicitly control scope over a longer span:

fu'e opens an attitudinal scope — the attitudinal applies to everything until fu'o closes it:

mi viska le fu'e .ia blanu zdani fu'o ponse I see the owner of what-I-believe-to-be a blue house.

Here .ia (belief) applies to blanu zdani (blue house) but not to ponse (owner), because fu'o closes the scope before ponse.

Without fu'e/fu'o, .ia would attach to whatever single word it follows. This is usually sufficient, but for complex descriptions spanning many words, the scope markers give you precise control.

fu'e with an attitudinal at sentence start applies that attitudinal to all following sentences until fu'o:

fu'e .ui mi klama le zarci .i mi facki lo cukta fu'o [Happiness scope:] I go to the store. I find a book. [End happiness scope.]

---

Parentheticals: to/toi, to'i, and sei/se'u#

These three constructs all insert material that is "outside" the main claim — but they work very differently.

to … toi — Spoken Parentheses#

to opens a parenthetical aside; toi closes it. The content inside can be any text and is structurally invisible to the outer parse:

mi klama to la .djan. pu cusku ke'u toi le zarci I go (John had said this before) to the store.

mi to .ui toi klama le zarci I (yay!) go to the store. (the .ui is parenthesized for emphasis)

to'i marks an editorial or quoted parenthetical — the content was not said by the current speaker (it's someone else's words or an inserted comment):

mi prami do to'i se cusku la .djan. toi "I love you" (said John, according to the text).

sei … se'u — Metalinguistic Commentary#

sei opens a sub-bridi that comments on the discourse itself — on the act of speaking, not on external facts. The sub-bridi inside sei is a full grammatical bridi but does not affect the truth conditions of the outer sentence:

mi citka sei mi cusku se'u vau le plise I eat (I say again) the apple.

le se'i bridi conventionally fills x₂ of the sei-bridi's selbri — the sentence being commented on.

Why sei/se'u exists (grammar perspective): The parser needs a way to embed a bridi inside a bridi without the inner bridi's sumti being confused with the outer bridi's sumti. sei acts as a bracket that says "parse this as a free modifier sub-bridi, not as a continuation of the main clause." se'u closes it explicitly; it can be elided before .i or vau.

Common uses:

sei mi jinvi = in my view / according to my opinion (sei-bridi = "I opine") sei se cusku = as is said / as stated (standard formula)

Comparing the three:

---

SI / SA / SU: Metalinguistic Erasers#

These cmavo are used in spoken Lojban (and informal writing) to correct mistakes on the fly:

mi klama si cadzu le zarci I go— I mean walk to the store.

mi pu klama sa mi ca cadzu le zarci I went— [erase] I am now walking to the store.

si is like a backspace key; sa is like erasing a whole phrase; su is "never mind, starting over."

These have no grammatical role in the text — they are metalinguistic operations on the stream of words. They are most useful in real-time speech but can appear in written texts to represent speech authentically.

Note: si erases only the single preceding word, not any free modifiers or indicators attached to it.

---

pau: Question Pre-Marker#

pau before a sentence marks it as a question even before the question word appears:

pau xu do klama Is it the case that you go? (pau = "I'm asking...")

Normally xu already signals a question, but pau makes the questioning intent explicit from the start — useful in spoken Lojban where you want to signal "this is a question" early.

paunai marks a rhetorical question — one that looks like a question but isn't:

paunai ma zmadu le ka certu mi [Rhetorical:] Who is more expert than me? (nobody, obviously)

---

Hesitation: .y.#

.y. (selma'o Y) is Lojban's formal hesitation sound — the equivalent of English "uh" or "er". It requires pauses on both sides and can be held or repeated as long as needed:

mi klama .y. le zarci I go... uh... to the store.

Unlike a pause, .y. explicitly signals you still have the floor. It is not grammatical in the middle of a word but may appear anywhere else — including between sumti, after connectives, or mid-tanru.

---

fa'o: End of Text#

fa'o (selma'o FAhO) explicitly marks the end of a text. It is almost always omitted but exists for contexts where an unambiguous end-signal matters:

• In computer interaction, to signal end-of-input/output
• In conversations, to explicitly yield the floor
• In written text, to mark an unambiguous stop point

fa'o is outside the regular grammar — the parser stops unconditionally when it encounters it (unless quoted with zo or lo'u…le'u). It cannot appear inside lu…li'u quotations or to…toi parentheticals.

mi klama le zarci fa'o I go to the store. [End of text]

---

cmavo Interaction Rules#

Several structural cmavo affect parsing at a level before normal grammar is applied. Key rules:

• zo quotes the immediately following word, no matter what it is — even si, fa'o, another zo, etc.
• si erases the preceding word (but not if that word is zo)
• sa erases the preceding word and the current grammatical construct it belongs to
• su erases the entire current utterance
• lo'u … le'u quotes all words between them literally (including most special cmavo)
• zei combines the preceding and following word into a lujvo, but cannot combine zo, si, sa, su, lo'u, ZOI, fa'o, or another zei
• BAhE (ba'e, za'e) marks the following word — unless that word is si/sa/su or follows zo
• bu converts the preceding word into a letter word (lerfu) — cannot combine with zo, si, sa, su, lo'u, ZOI, fa'o, zei, BAhE, or another bu
• UI/CAI mark the preceding word — except zo, si, sa, su, lo'u, ZOI, fa'o, zei, BAhE, bu; a following nai is absorbed into the UI
• .y., da'o, fu'e, fu'o behave like UI but do not absorb a following nai

---

Elidable Terminators#

Every construction that opens in Lojban has a corresponding terminator cmavo. Most are elidable — omittable when unambiguous. The full list:

When must you keep the terminator?

Elision is legal only when the parser can determine unambiguously where the construction ends. The main cases where you must keep the terminator:

• ke'e in mid-tanru (if more tanru follow the group)
• kei when an abstraction's x₂ follows (e.g., le su'u … kei be lo fasnu)
• ku before a selbri whose first word looks like a sumti-following word
• vau in compound bridi with tail-terms (gi'e…vau do)
• fe'u before a non-logical connective immediately after a fi'o-modal

In practice, most speech and writing elides nearly all terminators. The rule is: if elision causes ambiguity, keep the terminator.

---

MAI: Ordinal Discourse Markers#

MAI cmavo (selma'o MAI) attach to a number to create discourse ordinals — words that signal position in a sequence of points or topics. They are attitudinal-like: they color the sentence they begin without contributing to the bridi's truth conditions.

pamai mi djuno lo du'u tcima ba carvi First, I know it will rain.

remai le pa'o cu spofu Second, the umbrella is broken.

romai mi ba xabju le zdani Finally, I will stay home.

MAI always attaches directly to a preceding number (no space in speech): pamai, remai, cimai. The number and mai form a single cmavo-like unit.

romai specifically means "last in this series" — not "always" or "every time." It marks the final item without specifying how many items there were total.

---

MO'O: Section and Chapter Ordinals#

Where MAI marks items in a list, MO'O (selma'o MOhO) marks larger structural divisions — sections, chapters, stanzas, or other named divisions of a text:

pamo'o le cmene be le lojbo Section 1: The Name of Lojban

remo'o le gismu Section 2: The Root Words

MO'O differs from MAI in that it names a section heading — it typically appears at the start of a major structural block, not in the middle of argument:

• MAI = "this point is first/second in my argument"
• MO'O = "this block of text is section 1/2 of the overall document"

Both MAI and MO'O can be combined with the xi subscript for nested numbering:

pamo'o xi re = Section 1.2 remo'o xi pa = Section 2.1 pamai xi ci = point 1.3 in a nested list

---

xi: Subscript Rules (Text Structure Context)#

xi (selma'o XI) attaches a subscript to the preceding word. It is used throughout Lojban — on pro-sumti, letters, number words, and discourse markers. Key rules for text structure:

1. xi always follows the word it subscripts: ko'a xi re = ko'a₂
2. The subscript can be a number: da xi mu = x₅
3. The subscript can be a letter string: ko'a xi by. = ko'aᵦ
4. Subscripts nest via xi chains: ko'a xi pa xi re = ko'a₁.₂
5. xi can extend any finite series: fa xi xa = the 6th place tag; ni'o xi re = second-level topic shift

For text structure, the most common uses are:

• Nested sections: pamo'o xi re = §1.2
• Nested list items: remai xi ci = item 2.3

Same cmavo elsewhere: xi also tags extra FA places (fa xi xa), logical variables (da xi re in Chapter 21), and math or letter variables (xy. xi pa in Chapter 18). For a survey of “what can be subscripted,” see the xi (subscript) section in Chapter 18 (math variables) and Chapter 21 (logical variables).

---

Questions & answers: one protocol (hub)#

Questions and answers are spread across several chapters of this book. Friendly chapters spread the pedagogy (Chapter 6 — content questions; Chapter 8 — connective questions; Chapter 7 — pei; Chapter 13 — na / ja'a answers). This subsection is the single roadmap: question devices, what you may answer with, and where to read the lessons.

Fill-in-the-blank questions place a question cmavo where the answer would go. The answer often need not be a full bridi — it completes the open claim of the question (see the question/answer table below).

Parallel questions with fa'u: Chapter 8 explains fa'u (“respectively”) for statements. The same pattern pairs ma slots:

ma fa'u ma klama ma fa'u ma Who and who goes where and where, respectively?

la .djan. fa'u la .marcas. le zarci fa'u le briju John and Marsha [go] to the store and to the office, respectively.

Answers may be grammatical without repeating the whole bridi. Mechanical substitution of answer into question can look odd until elided terminators are mentally restored — the elision rules are covered in Chapter 21.

Kinds of non-bridi answer (not an exhaustive list): one or more sumti; a bare connective; a number or parenthesized mekso; na or ja'a; a relative clause hanging on a prior sumti; a prenex; be/bei linked to a prior selbri; at the start of a text, bare names, indicators, or vague nai. See the table above and the chapters on negation, connectives, and logic.

---

Summary#

Sentence / text structure:

• .i = sentence separator (between sentences)
• ni'o = new topic; no'i = resume old topic
• zo'u = topic-comment separator; also introduces quantifier prenex
• da'o = cancel pronoun assignments

Quotation types:

• lu … li'u = direct Lojban quote
• zo = single-word quote
• zoi gy. … .gy. = foreign text quote
• lo'u … le'u = error/fragment quote
• la'o gy. … .gy. = foreign name as sumti

Sumti qualifiers (LAhE):

• la'e = referent of; lu'e = symbol for
• tu'a = something involving (vague event); lu'a = a member of
• lu'i = the set of; lu'o = the mass of; vu'i = the sequence of

Emphasis and nonce:

• ba'e = emphasize next word; za'e = nonce coinage

Ordering:

• pamai/remai/romai = firstly/secondly/lastly; pamo'o/remo'o = section 1/2

Attitudinal scope:

• fu'e = open scope; fu'o = close scope

Parentheticals and metalinguistics:

• to … toi = spoken parentheses (any content, invisible to parser)
• to'i … toi = editorial/quoted parenthesis
• sei … se'u = metalinguistic sub-bridi (grammatical bridi, comments on discourse)

Self-correction (erasers):

• si = erase previous word; sa = erase current construct; su = erase utterance

Other:

• me … me'u = sumti-to-selbri conversion
• pe'a = metaphorical use marker
• pau = question pre-marker (nai = rhetorical question)

Questions & answers (hub):

• One table + ma fa'u ma example + non-bridi list — see Questions & answers: one protocol (hub) above; detail in Ch.6 / Ch.8 / Ch.13

Hesitation and end-of-text:

• .y. = formal hesitation sound (holds the floor; requires pauses on both sides; can be dragged out)
• fa'o = explicit end of text (outside regular grammar; parser stops unconditionally)

cmavo interaction priorities (processed before normal grammar):

• zo quotes the immediately following word unconditionally
• si/sa/su erase; lo'u…le'u quotes literally; zei compounds two words into lujvo
• BAhE marks following word; bu converts preceding word to lerfu; UI/CAI mark preceding word
• .y., da'o, fu'e, fu'o behave like UI but do not absorb a following nai

Elidable terminators (full list):

• be'o (BE), boi (PA/BY), do'u (COI/DOI), fe'u (FIhO), ge'u (GOI)
• kei (NU), ke'e (KE), ku (LE/LA), ku'e (PEhO), ku'o (NOI)
• li'u (LU), lo'o (LI), lu'u (LAhE), me'u (ME), nu'u (NUhI)
• se'u (SEI), te'u (mekso), toi (TO), tu'u (TUhE), vau (bridi), ve'o (VEI)
• Keep terminator when elision is ambiguous; kei be = must keep before abstractor x₂