0. Lojban and this manual
Lojban is a constructed language in the Loglan lineage. Its grammar builds clauses around predicates with ordered argument places. This manual describes Lojban for the jbo edition, using the reference grammar and baseline word lists for grammatical and lexical analysis, and examples from community writing to illustrate usage. (Cowan 1997, §§1.1–1.2, 2.1)
1. Lojban as a constructed language in the Loglan lineage
Lojban belongs to the Loglan lineage by design descent. Here, lineage describes the relationship between constructed language projects. The design draws on predicate logic. A clause expresses a relation, and the grammar provides ways to identify its arguments, combine it with other relations, quantify over variables, and mark the scope of an operation. Lojban also has grammatical resources for questions, attitudes, and other uses of language in conversation. (Cowan 1997, §§1.1, 2.1, 2.15–2.16; The Logical Language Group 1997, rules 30, 40, 50–54, 411–413)
The design is visible in the dictionary. A root entry gives a place structure alongside its English keywords. For example, the keyword language points to bangu, whose entry gives three places. These are the language or dialect, its users, and what those users express or communicate through it. The third place takes content such as an idea or proposition. The English keyword is a lookup aid. The full entry gives the information needed to assign Lojban arguments to places. (The Logical Language Group 1994, entry bangu)
The grammar distinguishes what an expression does in a clause from the form of the words used to build it. A content word can serve as a predicate on its own or occur inside an argument description. A predicate can also be built from several words or from a construction that contains an embedded clause. The same vocabulary can therefore serve several grammatical purposes. (Cowan 1997, §§2.9–2.11, 5.1, 5.9–5.10; The Logical Language Group 1997, rules 97, 112, 130–152)
2. Speakers, community use and the jbo edition
Lojban is used by a community of speakers, learners, and writers. The Lojban-language Wikipedia provides examples of extended explanatory prose on subjects ranging from music and mathematics to biology and geography. Those texts show one written register. They provide evidence of how contributors use the language, with no basis here for estimating the number of speakers or how frequently a form occurs in conversation. (Cowan 1997, §1.1; wiki:zgike, wiki:cmaci, wiki:mivyske, wiki:auskalerik)
jbo is the language code used for Lojban and the sole edition served by this grammar. The edition covers the shared reference language. Where an example illustrates a choice found in community writing, its description applies to that use. For example, the articles use both lo and loi before descriptive predicates. These are forms with different grammatical and interpretive functions, so their presence in different texts gives no reason to assign the texts to different dialects. The reference grammar discusses descriptions and their interpretation in terms of individuals, masses, and sets. (Cowan 1997, §§6.2–6.4; wiki:mivyske, wiki:zgike)
3. The reference standard and the scope of the manual
The grammatical reference is The Complete Lojban Language, published in 1997 and available as version 1.1 dated 2016. It explains how forms are pronounced, how words are built, and how constructions are used and interpreted. The third-baseline Lojban Machine Grammar, dated 10 January 1997, gives the formal syntax in YACC and EBNF. The baseline lists supply lexical forms and information such as root place structures and combining forms. (Cowan 1997, §§1.2, 3.1, 4.1, 4.6; The Logical Language Group 1997, EBNF preamble; The Logical Language Group 1994, gismu list preamble and entries)
The EBNF distinguishes grammatical constructs from terminals. It writes construct names in lowercase and terminal class names in uppercase, as with sumti and KOhA. KOhA is a selma'o, a grammatical class label. A form such as mi is written in ordinary Lojban spelling. Numbered rules link the EBNF constructs to their counterparts in YACC. A terminator marked as elidable can be omitted when its absence creates no grammatical ambiguity, with the same meaning as when it is present. (Cowan 1997, §§2.3, 7.2; The Logical Language Group 1997, EBNF notation, rules 90–97)
The formal rules establish how the relevant forms can combine. Lexical entries establish the places of individual predicates. Context supplies information such as what a pro-argument refers to or which relation a speaker intends between the components of a predicate combination. The manual draws on all three when an example requires them. It covers the organization of words and clauses as well as how speakers use those constructions in discourse. (Cowan 1997, §§2.1, 5.2, 7.2–7.7; The Logical Language Group 1994, entries bangu, benji; The Logical Language Group 1997, rules 40, 83, 90–97, 130–152)
4. Predications, predicates and arguments in a first example
A bridi is a predication. Its selbri gives the relation, and its sumti supply arguments to that relation. In the following excerpt, loi zgike is the explicitly supplied argument of sance. The particle cu marks the boundary before the clause predicate. The excerpt comes from the start of a longer sentence. (Cowan 1997, §§2.1, 2.5, 2.10, 6.3; The Logical Language Group 1997, rules 40, 53, 97)
Form “loi zgike cu sance”
Gloss MASS be.music PRED be.sound
Translation Music is sound.
Corpuswiki:zgike
The expression contains two words that can each be used as predicates. Within loi zgike, zgike supplies the descriptive content of the argument. At clause level, sance is the selbri. cu makes this division explicit. In the baseline sentence rule, cu is optional within the construction with preceding terms. An overt cu can prevent a preceding description and the following predicate from being read together. The marker itself contributes no argument place. (Cowan 1997, §§2.5–2.6, 2.10; The Logical Language Group 1997, rules 40, 53, 97, 112, 133)
The predicate benji illustrates how the numbered places work in a clause with several arguments. Its baseline entry assigns these roles. (The Logical Language Group 1994, entry benji)
| Place | Role with benji |
|---|---|
x1 | Sender or agent of transfer |
x2 | Item or content transferred |
x3 | Recipient |
x4 | Source or point of origin |
x5 | Means or medium of transfer |
Form
mi benji ti do(constructed)
Gloss 1 send PROX 2
Translation I send this to you.
In this simple order, without place tags or conversion, mi fills x1. The two arguments after benji, ti and do, fill x2 and x3 in sequence. The source and medium are left implicit. The English labels sender and recipient apply to the places of this predicate. With bangu, for example, x1 is a language. A place number identifies a position in the predicate's place structure. (Cowan 1997, §§2.3, 2.5–2.7, 7.7; The Logical Language Group 1994, entries benji, bangu)
5. Transcription, glosses and example format
The examples use the ordinary Latin spelling of Lojban, so they need no separate romanization. The form line retains written word divisions and characters that matter to the reading. In particular, an apostrophe inside a Lojban word represents an audible separation of vowels, a period marks a pause where written, and a comma used in Lojban spelling marks a syllable boundary. Capital letters within a form can mark stress, as in the name below. The chapter on sounds and spelling gives the conditions for those conventions. (Cowan 1997, §§1.3, 3.1–3.3, 3.9)
Form “la neANdertal.”
Gloss NAME Neanderthal
Translation Neanderthal
Corpuswiki:skalduna
The form line gives the Lojban expression. The gloss follows its order and gives a corresponding label or lexical gloss for each displayed word. A period within a lexical gloss joins English words used to gloss one Lojban word, as in be.sound. This period belongs to the gloss convention. A period in the form line belongs to the spelling. The translation gives idiomatic English and can express a relation through an English copula or preposition, as in is in Music is sound and to in I send this to you.
| Gloss | Meaning in these examples |
|---|---|
1 | First-person pro-argument |
2 | Second-person pro-argument |
MASS | Descriptor introducing a mass description |
NAME | Marker introducing a name |
PRED | Separator before the clause predicate |
PROX | Proximal demonstrative pro-argument |
A corpus example carries its sentence ID, including when the displayed quotation is a shorter clause or phrase from the corpus entry. An example drawn from an open or public-domain grammar carries its source citation. An example supplied for the explanation is marked (constructed). References accompanying an analysis specify the relevant section, rule, or dictionary entry. Throughout the manual, Lojban forms are set off typographically, and grammatical class labels retain their reference notation. (Cowan 1997, §1.3; The Logical Language Group 1997, EBNF notation)
6. How Mova handles Lojban in the jbo edition
At level 1, tools/langseg has neither a Lojban lexicon nor a Lojban language code. It does not distinguish Lojban spans from other Latin-script text. At level 2, tools/mova-parse has no model for jbo. The # parse comment reports the missing model, and LEMMA, UPOS, XPOS, FEATS, HEAD, and DEPREL are _ for its tokens, with Parse=None in MISC.
Mova has no Lojban word breaker for gismu, cmavo, lujvo, fu'ivla, and cmevla. Neither the baseline YACC and EBNF grammar nor a PEG grammar such as camxes runs in Mova. The research library holds the baseline grammar, the ilmentufa camxes.peg grammar, and the official gismu, cmavo, rafsi, and lujvo lists. These are available as reference material for a Lojban parser. Mova's code does not use them yet.
Mova has no annotated Lojban corpus or measured treebank baseline. It has no mapping from Lojban parse trees to Universal Dependencies, frames, or pragmatics. Squid reads the grammatical points in this chapter as claims. Formalizable claims wait in its queue for implementation. The chapter's grammatical rules are not implemented, and the first points needed for a Lojban baseline are those required for tokenization, tagging, and a Universal Dependencies treebank.
Sources for this chapter
- The Logical Language Group 1994. Lojban baseline word lists: gismu, cmavo, rafsi and lujvo (1994–2002) — Public domain: LLG declaration (public_domain.txt in the same directory): Lojban language definition materials, placed irrevocably in the public domain (signed Robert LeChevalier, President, LLG)
- The Logical Language Group 1997. Lojban Machine Grammar, 3rd Baseline (YACC and EBNF versions, 10 January 1997) — Public domain: explicitly dedicated to the public domain by its author, the Logical Language Group Inc. (headers of bnf.300 and grammar.300; LLG declaration PD.txt)
- John Woldemar Cowan 1997. The Complete Lojban Language (Logical Language Group, 1997; Version 1.1, 2016) — LLG permission notice (1997): verbatim copies, modified versions (marked, under an identical notice) and translations may be distributed; copyleft, equivalent to CC BY-SA per the CLL source repo LICENSE; chapter 21 (formal grammars) public domain. Read in the PDF, section 1.8 Boring Legalities, and github.com/lojban/cll chapters/01.xml
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/languages/jbo/00-introduction/