Chapter 2. Your First Sentences#

The Central Idea: Relationships#

In Lojban, the sentence is a predication — a claim that certain things stand in a certain relationship. This unit is called a bridi (roughly: "predication" or "claim").

Every bridi has two parts:

• A relation construct called the selbri — this specifies which relationship holds.
• One or more argument terms called sumti — these specify who or what is involved in the relationship.

In English, we distinguish between nouns, verbs, adjectives, and adverbs. In Lojban, there are no such distinctions. The word prami can act as "love", "lover", "beloved", "loving", or "lovingly" depending on context — it always describes the same underlying relationship.

---

Three Word Types#

All Lojban words fall into one of three categories:

Relation words (brivla in Lojban)

Words that express relationships. They always contain at least one consonant cluster (two consecutive consonants) within their first five sounds and always end in a vowel.

Examples: prami (love), klama (go/come), citka (eat), tavla (talk)

Brivla come in three subtypes — all grammatically identical, only different in origin:

A quick example of each:

ta cu klama le zarci That one goes to the market. (gismu selbri)

ta cu blari'o That is blue-green. (lujvo selbri)

ti cu djarspageti This is spaghetti. (fu'ivla selbri)

All three work exactly the same way in a sentence — the grammar does not care which subtype you use.

Particles (cmavo in Lojban)

Short structural words — the grammar glue of Lojban. They begin with a consonant followed by a vowel, optionally with more apostrophe-vowel sequences. Particles are often written together without spaces.

Examples: le (the), cu (selbri-separator), mi (I), pu (past tense)

Name words (cmevla in Lojban)

Names of people, places, etc. They always end in a consonant and are surrounded by pause marks (periods) when written.

Examples: .alis. (Alice), .teris. (Terry), .lojban. (Lojban)

---

Place Structures#

Every relation word has a place structure — a list of numbered slots that define the roles of the participants. These slots are labeled x₁, x₂, x₃, and so on.

tavla

x₁ talks to x₂ about topic x₃ in language x₄

prami

x₁ loves x₂

klama

x₁ goes/comes to x₂ (destination) from x₃ (origin) via route x₄ by means x₅

citka

x₁ eats x₂

gerku

x₁ is a dog of breed x₂

The place structure is part of every word's definition. When you fill in the slots with arguments, you get a sentence.

Swapping Places: se, te, ve, xe#

Sometimes you want to put a different argument in the x₁ slot — for emphasis or because of what is already in context. The particles se, te, ve, xe swap x₁ with another place:

Using klama (x₁ goes to x₂ from x₃ via x₄ by x₅):

mi klama le zarci I go to the market. (x₁=I, x₂=market)

le zarci se klama mi The market is gone-to by me. (x₁=market, x₂=I — places swapped)

le zdani te klama mi le zarci The house is the origin: I go to the market from it. (x₁=house, x₂=I, x₃=market — wait, let's be precise: te swaps x₁↔x₃, so x₁=origin=house, x₂=mi, x₃=destination=zarci)

A shorter example with vecnu (x₁ sells x₂ to x₃ for x₄):

la .teris. vecnu le karce la .alis. Terry sells the car to Alice.

le karce se vecnu la .teris. la .alis. The car is sold by Terry to Alice. (x₁=car, x₂=Terry)

la .alis. te vecnu la .teris. le karce Alice is the buyer: Terry sells the car to her. (x₁=Alice, x₃ becomes x₁)

Conversions are the standard Lojban way to form passive-like constructions and are very common.

---

The Simplest Sentences#

The basic word order is: x₁ selbri x₂ x₃ …

The first argument before the selbri fills x₁; arguments after the selbri fill x₂, x₃, in order.

mi prami do I love you.

mi

I, me (x₁ of prami)

prami

… loves …

do

you (x₂ of prami)

le tirxu cu klama le barda tcadu The tiger goes to the big city.

le tirxu

the tiger (x₁ of klama)

klama

… goes to …

le barda tcadu

the big city (x₂ of klama — destination)

mi tavla do I talk to you.

la .teris. cu citka Terry eats.

la .teris.

the one named Terry

When you leave a slot unfilled at the end of a sentence, it is treated as "unspecified" — the listener infers from context. So la .teris. cu citka just means Terry eats (something unspecified).

---

The cu Separator#

The particle cu separates the last sumti before the selbri from the selbri itself. It has no meaning on its own — it just marks where the relation begins.

mi tavla do I talk to you. (cu optional here — no ambiguity)

le sutra prenu cu tavla The fast person talks.

sutra

… is fast

prenu

… is a person

Without cu in that last example, le sutra prenu tavla would be parsed as "the fast-person-type-of talker" — a description, not a sentence. The cu signals: "what follows is the relation, not more description".

The rule: cu is needed whenever a description sumti (le…) ends right before the selbri and the parser might otherwise absorb the selbri into the description.

In practice, beginners can start most sentences with mi or do or la .name. and omit cu freely — it only becomes essential in more complex sentences. You'll develop a feel for when it's needed.

---

Basic Sumti: Pronouns#

A handful of particles work directly as sumti (arguments) without needing a relation word:

mi and do do not distinguish singular from plural or any gender — that information can be added later if needed, but is not required.

zo'e is Lojban's "placeholder" — it explicitly says "there is some value here but I'm not saying what it is". Trailing zo'e slots at the end of a sentence can simply be omitted.

More than one sumti before the selbri#

You can place more than one sumti before cu + selbri. The place numbers stay fixed — you are only choosing how many arguments sit in the “before the relation” zone, often for emphasis or style (the same bridi can be said several equivalent ways).

vecnu: x₁ is a seller, x₂ is goods sold, x₃ is a buyer, x₄ is a price.

mi cu vecnu ti ta zo'e I sell this to that (price left unspecified).

mi ti cu vecnu ta I — this — sell to that. (same claim: x₁ mi, x₂ ti, x₃ ta)

mi ti ta cu vecnu I — this — to that — sell. (same claim again)

The examples above cover the same sumti-order moves.

---

Names as Sumti#

Personal names and proper nouns are preceded by la:

la .alis. cu tavla Alice talks.

la .teris. cu klama le barda tcadu Terry comes to the big city.

Names end in a consonant and are surrounded by pause marks. The pause before a name with la is usually written as the period that's part of the name itself.

coi .alis. Hello, Alice!

coi la .teris. Hello, Terry!

Names can be Lojbanized versions of names from any language. The Lojban name for a person named "John" might be .djan., for "Mary" .meris., and so on. The rules for forming valid Lojban names are in Chapter 14.

---

The Observative#

If you start a sentence directly with the selbri — no x₁ before it — you get an observative: a quick announcement of something noticed.

klama Someone/something is going! (or: Coming!)

gerku Dog! (upon seeing one)

Observatives are the Lojban equivalent of English shouts like "Fire!" or "Car!" — they announce the relationship without specifying who is involved. They cannot use cu (there's nothing to separate).

---

Sentence Separators#

Multiple sentences are linked by .i (a standalone particle):

mi prami do .i do prami mi I love you. You love me.

.i is like a period/full stop that is also spoken aloud. It prevents the beginning of the next sentence from being mistakenly parsed as a trailing argument of the previous one.

For paragraph breaks (a new topic), use ni'o:

ni'o la .teris. cu blabi tirxu [New topic:] Terry is a white tiger.

---

Compound Relations: tanru#

When two relation words appear side by side as the selbri, the first modifies the second. The combination is called a tanru.

sutra tavla fast talker (x₁ is a fast type-of talker to x₂ about x₃ in language x₄)

barda tcadu big city

blabi gerku white dog

The place structure of a tanru is always that of the rightmost word. The modifying word adds a flavor of meaning — exactly what flavor is deliberately left vague and context-dependent, just as English compounds like "lemon tree" or "stone wall" are understood from context.

Three or more words chain the same way: each new word on the left modifies everything to its right:

sutra tavla gerku fast-talker type of dog (a dog in the way that a fast talker is)

Tanru are very common in Lojban and an expressive and creative tool. Their vagueness is intentional — when you need precision, you use other constructions (covered in Chapter 14).

---

Descriptions: le … (ku)#

To refer to a specific person or thing, wrap a selbri (or tanru) in le … ku:

le gerku the dog (literally: "that which I describe as a dog")

le sutra tavla the fast talker

le barda tcadu the big city

The ku at the end is often omitted when the description comes at the end of a sentence or when the boundary is clear from context.

le is a descriptor — it says "I have a specific thing(s) in mind that I'm describing as …". The description is the speaker's framing, not necessarily a verified fact (the dog might be debatable, but the speaker is thinking of it that way).

You can use any relation or tanru after le:

le prami be mi the one who loves me (be attaches x₂ inside the description — covered in Chapter 11)

---

Describing the Past and Future (Preview)#

Lojban bridi are tenseless by default — the time is inferred from context. But you can add a tense particle before the selbri:

mi pu klama le zarci I went to the store. (pu = past)

mi ca tavla do I am now talking to you. (ca = present/now)

mi ba klama I will go. (ba = future)

pu

(particle) past tense

ca

(particle) present / now

ba

(particle) future tense

Tense is always optional in Lojban — omitting it does not make the sentence ungrammatical, just temporally vague.

---

A Starter Vocabulary#

Here are some useful relation words to get you started. Each entry shows the full place structure.

People & actions

prami

x₁ loves x₂

tavla

x₁ talks to x₂ about x₃ in language x₄

klama

x₁ goes to x₂ from x₃ via x₄ by means x₅

citka

x₁ eats x₂

sipna

x₁ sleeps/is asleep

cliva

x₁ leaves x₂

vecnu

x₁ sells x₂ to x₃ for price x₄

Properties

melbi

x₁ is beautiful/pretty to x₂ by standard x₃

barda

x₁ is big/large in dimension x₂ by standard x₃

cmalu

x₁ is small/little in dimension x₂ by standard x₃

sutra

x₁ is fast/quick at doing x₂

blari'o

x₁ is blue-green (cyan/teal) of shade x₂

pluka

x₁ is pleasant/pleasing to x₂ under conditions x₃

Things

zarci

x₁ is a market/store selling x₂ run by x₃

tcadu

x₁ is a city/town of x₂

gerku

x₁ is a dog of breed x₂

tirxu

x₁ is a tiger of species x₂

---

A Short Dialogue#

Here is a short Lojban exchange using what you now know:

coi .alis. Hello, Alice.

coi .djan. .i do mo Hello, John. What's up? (mo = what relation? — a selbri question)

mi klama le zarci .i xu do klama I'm going to the store. Are you going? (xu = yes/no question marker)

je'u go'i Yes (literally: "truly [that]")

je'u

indeed, truly (an evidential attitudinal)

go'i

[repeat the previous bridi / yes]

---

Terminology at a glance#

Quick definitions you will see in every chapter (full detail unfolds as you read):

For how sumti divide into description / pro-sumti / name / quotation / number, see Chapter 3.

---

Summary#

• Terminology table: bridi, sumti, selbri, brivla, cmavo, cmevla, gadri — see Terminology at a glance above; five sumti kinds in Ch.3
• A Lojban sentence (bridi) = a selbri (relation) + sumti (arguments filling numbered slots x₁ x₂ …)
• Default word order: x₁ first, then selbri, then x₂ x₃ …
• More than one sumti may appear before cu + selbri (same places — style/emphasis) — see vecnu examples above
• cu separates a preceding description from the selbri when needed
• Pronouns: mi (I), do (you), ti/ta/tu (this/that)
• Names: la .name. — always end in a consonant, surrounded by pause marks
• Descriptions: le selbri — "the thing I describe as …"
• Tanru: two relation words side by side — first modifies second; place structure from rightmost
• Tense is optional: pu (past), ca (now), ba (future)
• Multiple sentences: separated by .i
• Brivla subtypes: gismu (root words), lujvo (compounds), fu'ivla (borrowings) — all grammatically identical
• Conversions se/te/ve/xe: swap x₁ with x₂/x₃/x₄/x₅ (Lojban's passive construction)