Elamur: **Mova** · [Parse](https://mova.elamur.ai/free/en/parse/) · [Dictionary](https://md.dictionary.elamur.ai/free/en/) · [Translate](https://md.translate.elamur.ai/free/en/) · [Research](https://md.research.elamur.ai/free/en/) · [Gali](https://md.gali.elamur.ai/free/en/) · [MCP](https://mcp.elamur.ai/) · [Locations](https://md.locations.elamur.ai/free/en/) · [Weather](https://md.weather.elamur.ai/free/en/) · [Constitution](https://md.constitution.elamur.ai/)

[Mova](https://md.mova.elamur.ai/free/en/) · [this page in HTML](https://mova.elamur.ai/free/en/docs/languages/jbo/08-predicates-and-place-structures/)

[Mova](https://md.mova.elamur.ai/free/en/) / [docs](https://md.mova.elamur.ai/free/en/docs/) / [languages](https://md.mova.elamur.ai/free/en/docs/languages/) / [jbo](https://md.mova.elamur.ai/free/en/docs/languages/jbo/) / 8. Predicates selbri and their place structures

# 8. Predicates selbri and their place structures

**On this page**

1. [Selbri as a syntactic function](#selbri-as-a-syntactic-function)
2. [Gismu roots, lujvo compounds and fu'ivla borrowings](#gismu-roots-lujvo-compounds-and-fuivla-borrowings)
3. [Reading a dictionary place structure from x1 onward](#reading-a-dictionary-place-structure-from-x1-onward)
4. [Lexical arity and interpretation of omitted places](#lexical-arity-and-interpretation-of-omitted-places)
5. [Pro-predicates and predicate assignment](#pro-predicates-and-predicate-assignment)
6. [Predication without inflection](#predication-without-inflection)
7. [Separate markers for time, aspect, modality and discourse stance](#separate-markers-for-time-aspect-modality-and-discourse-stance)
8. [How Mova handles Lojban predicates and lexical frames](#how-mova-handles-lojban-predicates-and-lexical-frames)
9. [Sources for this chapter](#sources-for-this-chapter)

A *selbri* expresses a relation. Its *sumti* supply the participants or other values that the relation is about. A dictionary entry gives the order and meaning of the places of a predicate word. The grammar determines how the sumti in an utterance are matched to those places, including when the speaker leaves some of them unstated.

## Selbri as a syntactic function

A *bridi* is a predication. In a simple clause, the selbri supplies the relation, and the sumti that fill its places can occur before or after it. The selbri can begin the clause if there are no preceding terms. After preceding terms, `cu` marks the beginning of the predicate part. It is optional where the preceding terms and the predicate part already have an unambiguous boundary. In particular, a description can itself contain a selbri, so a boundary marker can determine whether a predicate word continues the description or begins the main predication. (The Logical Language Group 1997, §§40, 53, 71, 97, 111–112)

These two corpus excerpts show the same construction at different levels. In the first, `saske` is the main selbri. In the second, `gugdrxauskale` is the selbri inside a description that can serve as a sumti. Lojban uses the same predicate syntax for relations translated by English verbs, adjectives, or nouns. (The Logical Language Group 1997, §§53, 97, 112, 152)

| Form | Gloss | Translation |
| --- | --- | --- |
| `loi cmaci goi ko'a cu saske` | mass.of mathematics assign.reference PRO clause.boundary be.a.science | “Mathematics, here called `ko'a`, is a science …” (`wiki:cmaci`, excerpt) |
| `lo gugdrxauskale` | description be.Basque.Country | “the Basque Country” (`wiki:auskalerik`, excerpt) |

*Selbri* names a syntactic function. *Brivla* names a class of words that can perform that function. A single brivla is enough for a selbri, and a selbri can also contain several words or be built by a grammatical construction. In the baseline grammar, the smallest units available for building a tanru include the following forms. A unit can take its own linked arguments, and units can be joined into larger selbri with the connective and grouping constructions. Tags and `na` can occur at the selbri level. (The Logical Language Group 1997, §§130–136, 150–161)

| Form of unit | Function |
| --- | --- |
| A brivla | Supplies a lexical predicate. |
| A `GOhA` form, optionally followed by `ra'o` | Supplies a pro-predicate, a predicate question, or another relation from that grammatical class. |
| `ke` with a predicate group, ending in `ke'e` when required | Makes the group a unit for the surrounding construction. |
| `me` with a sumti, ending in `me'u` when required, optionally followed by a `MOI` form | Builds a predicate from an argument expression. |
| A number or letter string followed by a `MOI` form | Builds a predicate such as an ordinal. |
| A `NUhA` form with a mathematical operator | Makes an operator available as a predicate. |
| A `SE` form followed by a unit | Converts the order of places. |
| `jai` followed by a unit, with an optional intervening tag | Changes which role is available in first place. |
| Words joined with `zei` | Makes a compound predicate unit. |
| A `NAhE` form followed by a unit | Applies scalar qualification to the unit. |
| An abstraction introducer from `NU` with a subsentence, ending in `kei` when required | Makes a predicate of an abstraction. The grammar also allows connected abstraction introducers and permits `nai` after each introducer. |

### Tanru and grouping

A *tanru* combines predicate expressions so that one qualifies another. In ordinary two-part order, the first is the modifier, or *seltau*, and the second is the head, or *tertau*. The head gives the resulting expression its external place structure. The modifier helps specify the kind of relation intended. Context supplies the connection between the two meanings. For instance, a modifier can suggest a manner, use, material, or subject matter without an overt word naming that connection. A bare modifier does not introduce its own places into the main clause's argument sequence. (The Logical Language Group 1997, §§132–136, 150–161)

| Form | Gloss | Translation |
| --- | --- | --- |
| `barda lalxu` | be.large be.a.lake | “be a large lake” (`wiki:nikaraguas. zei lalxu`, excerpt) |
| `remna bacru bangu` | be.human utter.vocally be.a.language | “be a human spoken language” (`wiki:bangu`, excerpt) |

Plain juxtaposition in a series groups from the left. Thus `remna bacru bangu` has `remna bacru` as a modifier of `bangu`. The sentence can still be about a language, its users, and what they express. The expression does not require the speaker to supply a separate utterer and utterance as arguments of `bacru`. In the lake example, lake places remain available, while the dimensions and standard of largeness can be inferred unless they are expressed within the modifier. (The Logical Language Group 1997, §§133–136, 150–161) (The Logical Language Group 1994, § gismu list, entries `bacru`, `bangu`, `barda`)

`bo` binds adjacent predicate expressions more tightly than plain juxtaposition. A chain joined with `bo` groups from the right. `ke` opens an explicit group and `ke'e` closes it. The latter is elidable when the boundary is recoverable under the grammar. `co` reverses the order of modifier and head. It binds more loosely than juxtaposition, so a multiword expression on its left can be the head. In a `co` construction, leading arguments use the head's places and trailing arguments apply to the modifier. Ordinary ordering or links with `be` are available when the arguments need to be attached directly to a particular unit. (The Logical Language Group 1997, §§40, 53, 132–136, 150–161)

| Form | Gloss | Translation |
| --- | --- | --- |
| `barda blabi baktu` | be.large be.white be.a.bucket | “be a bucket characterized by a kind of whiteness associated with size” (constructed) |
| `barda blabi bo baktu` | be.large be.white bind be.a.bucket | “be a large white bucket,” with `blabi baktu` grouped (constructed) |
| `barda ke blabi baktu ke'e` | be.large group.open be.white be.a.bucket group.close | “be a large white bucket,” with `blabi baktu` explicitly grouped (constructed) |
| `baktu co blabi` | be.a.bucket invert.modification be.white | “be a white bucket” (constructed) |

These groupings are grammatical. They constrain which expression modifies which and which relation supplies the main places. The English readings illustrate the possible interpretations of the size and color modifiers. (The Logical Language Group 1997, §§132–136, 152)

## Gismu roots, lujvo compounds and fu'ivla borrowings

The three lexical kinds of brivla have the same basic syntactic distribution. A gismu, a lujvo, or a fu'ivla can stand alone as a selbri or appear as a unit of a larger selbri. Their internal word formation differs. The baseline syntax handles them as `BRIVLA`. The PEG grammar also distinguishes brivla from structure words and name words in its word rule. (The Logical Language Group 1997, §152) (Hagiwara, Ilmen and lagleki 2014, § `lojban_word`)

| Kind | Formation | Example and gloss |
| --- | --- | --- |
| *gismu* | A root with five letters, of the shape CVCCV or CCVCV, subject to the consonant-cluster rules. | `bajra` ‘run’ or `baktu` ‘be a bucket’ |
| *lujvo* | A compound made of combining forms, called *rafsi*. The components can come from roots or from structure words that have combining forms. | `bajystu`, from `bajra` and `stuzi`, ‘be a running track’ |
| *fu'ivla* | A borrowed or adapted word with a Lojban predicate-word form. A borrowing may include a classifier from Lojban. | `gugdrxauskale` ‘be the Basque Country’ (`wiki:auskalerik`, excerpt) |

The root and compound lists associate place structures with lexical entries. A root's shape identifies its word class, while its places come from the entry. Knowing that `bajra` is a gismu gives no way to recover its surface, limb, and gait places from its spelling. The borrowing also needs an established meaning and role order. An English equivalent can help identify its sense, while its Lojban use determines which sumti go with it. (The Logical Language Group 1994, §§ gismu and lujvo lists) (The Logical Language Group 1997, §152)

A lexical compound packages one interpretation and one place structure. The underlying tanru can help explain the compound, but its components' places can be combined in several ways. A single compound place may do the work of places from several components. A place can also be left out of the compound's definition, or its role can appear in a different position from the one it had in a component. Read the listed roles in the order given for the whole lujvo. (The Logical Language Group 1994, § lujvo list, entries `bajystu`, `bajli'a`, `bajykla`, `barkla`)

| Compound | Gloss | Places in the compound entry |
| --- | --- | --- |
| `bajystu` | be.a.running.track | x1 is the site. |
| `bajli'a` | run.away.from | x1 is both runner and departer, x2 is the place left, x3 is the route of departure, x4 is the running surface, x5 is the limbs, and x6 is the gait. |
| `bajykla` | travel.by.running | x1 is both traveler and runner, x2 is the destination, x3 is the origin, x4 is the route, x5 is the running surface, x6 is the limbs, and x7 is the gait. |
| `ba'orzu'e` | act.to.grow.something | x1 is the actor, x2 is what grows, x3 is its resulting size or form, and x4 is its starting size or form. |

In `bajykla`, the running surface is x5 even though it is x2 of `bajra`. The compound gives destination, origin, and route the earlier positions. In `ba'orzu'e`, the actor and the entity that grows have independent places, so the one that grows need not be the one taking action. A phrase about growth headed by `banro` would begin with what grows. (The Logical Language Group 1994, § gismu list, entries `bajra`, `banro`, and § lujvo list, entries `bajykla`, `ba'orzu'e`)

## Reading a dictionary place structure from x1 onward

A place structure is an ordered list of semantic roles. The role of x1 varies with the relation. It can be a runner, a language, a container, or an event that protects something. The later places vary in the same way. In particular, x2 can be a species, a growing entity's resulting form, or the users of a language. When a dictionary gives a short translation and a full set of places, use the full set of places to assign the arguments. (The Logical Language Group 1994, § gismu list)

| Predicate | Short gloss | x1 | x2 | x3 | x4 |
| --- | --- | --- | --- | --- | --- |
| `badna` | be a banana | banana fruit or plant | species or breed |  |  |
| `bacru` | utter vocally | producer of the vocal sound | vocal output |  |  |
| `badri` | be sad | experiencer | subject of the sadness, expressed as an abstraction |  |  |
| `barda` | be large | entity evaluated | property or dimension in which it is large | comparison standard |  |
| `bangu` | be a language | language or dialect | users | expressed content, as an idea or proposition |  |
| `bandu` | protect | protecting event | protected entity or state | threatening event or possibility |  |
| `bajra` | run | runner | surface | limbs used | gait |

These entries distinguish the kind of value a place calls for as well as its position. In `bandu`, the protection is attributed to an event in x1. In `bangu`, x3 is content of the kind expressed with `si'o` or `du'u`, rather than a quotation of words. In `bacru`, vocal production alone is sufficient for the relation. The entry does not imply an audience or successful communication. The syntactic category is sumti in each case, while the lexical relation constrains its interpretation. (The Logical Language Group 1994, § gismu list, entries `bacru`, `badri`, `bangu`, `bandu`)

### Untagged arguments and arguments with place tags

With ordinary untagged sumti, assignment starts with x1 before the selbri, continues in order through any further leading sumti, and resumes with the next place after the selbri. When the selbri comes first, the untagged sequence after it starts at x2, leaving x1 implicit. This rule concerns positional arguments of the selbri. A modal or tense term has its own tag and does not take the next lexical place. Linked arguments attach to their particular unit. The baseline grammar distinguishes terms before the predicate part from tail terms and allows both positional and tagged terms. (The Logical Language Group 1997, §§40, 50–53, 71, 80–85, 151, 160–161) (The Logical Language Group 1994, § cmavo list, entries `be`, `bei`)

| Form | Gloss | Translation |
| --- | --- | --- |
| `lo remna cu bajra lo dertu` | description human clause.boundary run description earth | “Someone runs on dirt.” (constructed) |
| `lo remna lo dertu cu bajra` | description human description earth clause.boundary run | “Someone runs on dirt.” (constructed) |
| `bajra lo dertu` | run description earth | “[Someone] runs on dirt.” (constructed) |

A place tag from `FA` can give an argument an explicit place. It can go before or after the selbri. After a numbered tag, subsequent untagged sumti continue forward from that place, skipping places already assigned. Another tag redirects the assignment. The numbered tags reach x5. A longer frame can still be filled by continuing in positional order beyond x5. (The Logical Language Group 1997, §§40, 53, 71, 83) (The Logical Language Group 1994, § cmavo list, entries `fa` through `fu`, `fai`, `fi'a`)

| Tag | Place |
| --- | --- |
| `fa` | x1 |
| `fe` | x2 |
| `fi` | x3 |
| `fo` | x4 |
| `fu` | x5 |
| `fai` | The displaced first place in a `jai` conversion. |
| `fi'a` | Asks which place the associated sumti fills. |

For example, `la terdi se vacri fi loi trano` begins a clause in which `fi` puts the nitrogen term in the third place of the converted `vacri`. There is no intervening x2 sumti. (The Logical Language Group 1997, §§40, 53, 83, 152)

| Form | Gloss | Translation |
| --- | --- | --- |
| `la terdi se vacri fi loi trano` | name Earth swap.1.2 be.air place.3 mass.of nitrogen | “Earth has an atmosphere containing nitrogen …” (`wiki:terdi`, excerpt) |

### Conversion and linked places

A `SE` conversion interchanges x1 with one other place. Every other place keeps its number. Place tags outside the converted unit address the resulting order. Applying the same conversion twice to the same relation restores the original order. Syntactically, a conversion operates on the following predicate unit. To convert a larger tanru as a whole, make that tanru a unit with `ke` and its closing boundary. (The Logical Language Group 1997, §§136, 150–152) (The Logical Language Group 1994, § cmavo list, entries `se`, `te`, `ve`, `xe`)

| Form | Exchange | Resulting order for a five-place relation |
| --- | --- | --- |
| `se` | x1 and x2 | original x2, x1, x3, x4, x5 |
| `te` | x1 and x3 | original x3, x2, x1, x4, x5 |
| `ve` | x1 and x4 | original x4, x2, x3, x1, x5 |
| `xe` | x1 and x5 | original x5, x2, x3, x4, x1 |

Conversion is available inside a description as well as in the main predication. A description based on `bangu` takes its referents from the language place. With `se bangu`, the users come first and the language is second. The content stays third. A description can specify that second place through `be`. The first untagged linked argument starts at x2, and further links introduced by `bei` continue the sequence. A place tag can specify a different place. `be'o` closes the links and can be omitted where the boundary is clear under the grammar. (The Logical Language Group 1994, § gismu list, entry `bangu`, and § cmavo list, entries `be`, `bei`, `be'o`) (The Logical Language Group 1997, §§97, 112, 150–161)

| Form | Gloss | Translation |
| --- | --- | --- |
| `lo se bangu be la lojban.` | description swap.1.2 be.a.language link name Lojban | “a user of Lojban” (constructed) |
| `lo baktu be lo djacu bei lo mudri be'o` | description be.a.bucket link description water link.next description wood links.close | “a bucket containing water and made of wood” (constructed) |

The same linking construction can supply a place inside a modifier. In `barda be lo ka ce'u galtu be'o baktu`, the linked property specifies the dimension of largeness. The resulting tanru is still headed by `baktu`, whose x2 is the bucket's contents. (The Logical Language Group 1994, § gismu list, entries `barda`, `baktu`) (The Logical Language Group 1997, §§133–136, 150–161)

| Form | Gloss | Translation |
| --- | --- | --- |
| `barda be lo ka ce'u galtu be'o baktu` | be.large link description property abstraction.variable be.high links.close be.a.bucket | “be a bucket that is large in height” (constructed) |

## Lexical arity and interpretation of omitted places

Lexical arity is the number of places a predicate has in its lexical definition. The number of overt arguments is a property of a use of that predicate. The two can differ. A single-place `bajystu` and a four-place `bajra` can both occur with just one expressed sumti, and the presence of x4 in `bajra` does not oblige the speaker to say which gait the runner uses. The grammar permits a selbri with no expressed sumti at all. (The Logical Language Group 1994, § gismu list, entry `bajra`, and § lujvo list, entry `bajystu`) (The Logical Language Group 1997, §§40, 53, 71)

In ordinary predication, an omitted lexical place can have the same interpretation as an explicit `zo'e`. The place remains part of the relation, and context can supply its value or leave it unspecified. The speaker may expect the listener to know the surface someone runs on, or the surface may simply be irrelevant to the point of the utterance. Omission alone supplies no particular surface and says nothing about whether the speaker knows it. Where a lexical entry specifies a default, that default matters. For `balvi`, the baseline entry gives the current space-time reference as the default for its comparison point when that point is left to context. Other predicates have their own lexical conditions. (The Logical Language Group 1994, § gismu list, entries `bajra`, `balvi`, and § cmavo list, entry `zo'e`)

A bare predicate can be used as an observative when the situation supplies what is being noticed. The hearer can recover the missing x1 from what is perceptually salient. A bare predicate can also continue an established topic in discourse. The grammar admits both uses through the same option to omit terms. When an intermediate place is left unspoken and a later place is given, use a tag or a placeholder to make the intended role recoverable. Without either, the untagged sumti is assigned to the next place in sequence. (The Logical Language Group 1997, §§40, 53, 71, 83) (The Logical Language Group 1994, § cmavo list, entries `fa`, `fe`, `fi`, `fo`, `fu`, `zo'e`)

| Form | Gloss | Translation |
| --- | --- | --- |
| `bajra` | run | “[Someone's] running!” (constructed, observative reading) |
| `mi bajra fo lo sance` | I run place.4 description sound | “I run with a sound as my gait.” (constructed, illustrating x4 assignment) |
| `mi bajra zo'e lo jamfu` | I run unspecified description foot | “I run using my feet [on an unspecified surface].” (constructed) |

The tag makes the place explicit even when the filler is semantically unusual. In the second example, x4 calls for a gait, so the sound has to be understood as characterizing the gait for the utterance to make sense. In the third example, the placeholder uses up x2, allowing the feet to occupy x3. The speaker could instead give the feet with `fi` and leave x2 wholly unspoken. (The Logical Language Group 1994, § gismu list, entry `bajra`, and § cmavo list, entries `fi`, `fo`, `zo'e`) (The Logical Language Group 1997, §83)

`zi'o` explicitly removes a place from the relation as used. With `zo'e`, the relation still has the relevant role and the value is left to context. Using `zi'o` can therefore change what is predicated. Neither operation assigns the next argument to the earlier role. The numbered positions of the other places still govern the syntax. An added modal term is a separate operation. It expresses a supplementary relation and leaves the lexical place order available for the predicate's arguments. (The Logical Language Group 1994, § cmavo list, entries `zi'o`, `zo'e`, `bai`, `bau`) (The Logical Language Group 1997, §§83, 491, 815, 972)

| Form | Gloss | Translation |
| --- | --- | --- |
| `ti baktu zo'e lo mudri` | this be.a.bucket unspecified description wood | “This is a wooden bucket with unspecified contents.” (constructed) |
| `ti baktu zi'o lo mudri` | this be.a.bucket remove.place description wood | “This is a wooden bucket, with the contents role dropped.” (constructed) |

The same distinction between grammatical assignment and interpretation applies to repeated or conjoined material. A connective can let two predicates share arguments in the surrounding bridi, yet a given place still has to be interpreted with respect to each connected relation. In the corpus excerpt `sance gi'e pluka loi remna`, the people are tied to the `pluka` tail. To determine which roles are shared and which belong to a single tail, first identify the scope of the connective and the position of the terms. (The Logical Language Group 1997, §§50–53, 71)

| Form | Gloss | Translation |
| --- | --- | --- |
| `sance gi'e pluka loi remna` | be.a.sound and.tail please mass.of humans | “is sound and is pleasing to people” (`wiki:zgike`, excerpt) |

## Pro-predicates and predicate assignment

A pro-predicate can occupy the same grammatical slot as a brivla. Its interpretation may depend on previous discourse, an enclosing bridi, an explicit assignment, or a question. When it recalls a bridi, speakers can carry over its arguments and give new sumti for places they want to replace. Thus the listener needs both the earlier relation and the role assignments of the new clause. A pro-predicate can also appear inside a larger selbri. (The Logical Language Group 1994, § cmavo list, entries under `GOhA`) (The Logical Language Group 1997, §§130–136, 150–152) (Hagiwara, Ilmen and lagleki 2014, § `GOhA_no_SA_handling`)

| Form | Use |
| --- | --- |
| `go'i` | Recalls the preceding bridi. |
| `go'e` | Recalls the bridi before the preceding one. |
| `go'a` | Recalls a bridi from the earlier discourse. |
| `go'u` | Recalls a bridi from more remote discourse. |
| `go'o` | Refers forward to a later bridi. |
| `nei` | Refers to the current bridi. |
| `no'a` | Refers to the enclosing bridi. |
| `co'e` | Leaves the predicate unspecified for recovery from context if possible. |
| `mo` | Asks for the predicate. |
| `du` | Expresses identity among its arguments. |
| `bu'a` | First variable for a predicate. |
| `bu'e` | Second variable for a predicate. |
| `bu'i` | Third variable for a predicate. |

In `wiki:xindo ropno bangu`, a list of individual languages is followed by `go'i`. The prior bridi uses `bangu` and states a speaker population. The new language expression supplies the language place in the recalled pattern. This use of `go'i` depends on which bridi precedes it and how its sumti were assigned. (The Logical Language Group 1994, § cmavo list, entry `go'i`, and § gismu list, entry `bangu`)

| Form | Gloss | Translation |
| --- | --- | --- |
| `go'i` | recall.previous.predication | “… satisfy the preceding predication.” (`wiki:xindo ropno bangu`, excerpt) |

When `ra'o` follows a `GOhA` form that recalls a bridi, pro-arguments in what is recalled can be interpreted with their current values. This distinction matters across speakers. A remembered `mi` normally preserves the person who said it, while updating the reference makes it pick out the speaker of the new utterance. Giving an overt new sumti for the first place is another way to determine that place in the repeated bridi. (The Logical Language Group 1994, § cmavo list, entries `go'i`, `mi`, `ra'o`) (The Logical Language Group 1997, §152)

`cei` provides explicit assignment. The expression before `cei` gives the predicate being assigned, and the following unit supplies the name used for it. The conventional predicate assignment words are the five gismu `broda`, `brode`, `brodi`, `brodo`, and `brodu`. The `CEI` rule allows a sequence of units linked by `cei`, with argument links available on the units. Grouping allows an expression with several components to be assigned as a single unit. Subsequent use of the assigned word relies on the assignment in discourse. (The Logical Language Group 1994, § gismu list, entries `broda`, `brode`, `brodi`, `brodo`, `brodu`, and § cmavo list, entry `cei`) (The Logical Language Group 1997, §§150–152)

| Form | Gloss | Translation |
| --- | --- | --- |
| `mi bajra cei broda` | I run assign.predicate predicate.variable.1 | “I run. Let `broda` stand for running.” (constructed) |
| `do broda lo dertu` | you assigned.predicate.1 description earth | “You run on dirt.” (constructed, after the assignment above) |

The new sentence uses the assigned relation with its own arguments. The assignment has supplied the meaning of `broda`, including the surface in x2. It did not give `broda` the same discourse function as `go'i`, which recalls a bridi and can preserve the participants from it. Predicate variables `bu'a`, `bu'e`, and `bu'i` supply yet another device, for quantification over predicates. A referent assignment with `goi` applies to a sumti, as in the corpus example assigning `ko'a` to mathematics. (The Logical Language Group 1994, § cmavo list, entries `bu'a`, `bu'e`, `bu'i`, `cei`, `go'i`, `goi`) (The Logical Language Group 1997, §§30, 121–122, 150–152)

## Predication without inflection

A selbri has no agreement paradigm for the person or number of a sumti. The speaker, addressee, and a quantified group can fill the same place while the predicate retains its form. Number can be expressed in the sumti when it matters. The selbri also lacks inflection for tense, aspect, mood, evidentiality, and politeness. Lojban expresses these distinctions through separate grammatical constructions, individual structure words, lexical choices, and context. (The Logical Language Group 1997, §§40, 83, 95–97, 130–152, 300, 491, 815, 972) (The Logical Language Group 1994, § cmavo list)

| Form | Gloss | Translation |
| --- | --- | --- |
| `mi bacru` | I utter.vocally | “I speak aloud.” (constructed) |
| `do bacru` | you utter.vocally | “You speak aloud.” (constructed) |
| `re lo remna cu bacru` | two description human clause.boundary utter.vocally | “Two people speak aloud.” (constructed) |
| `ko bacru` | you.IMPERATIVE utter.vocally | “Speak aloud.” (constructed) |

`ko` is an imperative sumti. Its place in the relation tells the addressee what role to take in bringing the predication about. It can occupy a place other than x1. A predicate question with `mo` asks for the relation, while the particle `xu` marks a polar question. Neither operation requires changing the shape of a lexical predicate. (The Logical Language Group 1994, § cmavo list, entries `ko`, `mo`, `xu`) (The Logical Language Group 1997, §§83, 97, 152)

| Form | Gloss | Translation |
| --- | --- | --- |
| `do mo` | you predicate.question | “What are you doing?” or “How are you?” (constructed, according to context) |
| `xu do bacru` | polar.question you utter.vocally | “Are you speaking aloud?” (constructed) |

Lojban has no inflected contrast between finite and non-finite predicate words. Abstraction provides a way to use a predication as the content of a sumti or a new selbri. An abstraction introducer from `NU` takes a subsentence, whose selbri has the same grammar as it has in an independent bridi. `kei` closes the embedded predication and is elidable when the boundary is recoverable. The abstraction can itself stand as a selbri with its own place structure. A description such as `lo nu` with the embedded predication refers to the event. A `ka` abstraction expresses a property and can use `ce'u` to mark the place abstracted over. A `du'u` abstraction expresses propositional content. (The Logical Language Group 1997, §§41, 97, 112, 152) (The Logical Language Group 1994, § cmavo list, entries `nu`, `kei`, `ka`, `ce'u`, `du'u`)

| Form | Gloss | Translation |
| --- | --- | --- |
| `jalge lo nu lo caktapla cu janli lo caktapla` | result.from description event description tectonic.plate clause.boundary collide.with description tectonic.plate | “results from a tectonic plate colliding with a tectonic plate” (`wiki:ma'arlinsi`, excerpt) |
| `lo ka ce'u bajra kei` | description property abstraction.variable run abstraction.close | “the property of running” (constructed) |
| `lo du'u do bajra kei` | description proposition you run abstraction.close | “the proposition that you run” (constructed) |

In the corpus example, the whole collision event fills a place of `jalge`. Inside that event description, `janli` has its own arguments. The plate descriptions fill places of `janli`, and the event description fills a place of `jalge`. English can use an embedded form such as “colliding” to express this dependency. Lojban uses the abstraction boundary and leaves `janli` in its ordinary predicate form. (The Logical Language Group 1997, §§40–41, 53, 97, 112, 152)

## Separate markers for time, aspect, modality and discourse stance

A tag can occur before the selbri, or form a term with a sumti. In the latter case the sumti provides the reference point or the associated participant for the tag. When the syntax permits the tag to stand alone as a term, its associated sumti can be absent. Tags can combine through the grammatical connective constructions. The lexical roles of the predicate are still available independently of these tags. (The Logical Language Group 1997, §§83, 130, 491, 815, 971–972)

### Time and event contour

The basic direction markers relate the predication to a temporal reference point. Context can supply that point. An explicit sumti tagged with one of these forms can locate the predication with respect to a named time or event. An untagged bridi can receive a present, past, or future interpretation from context. The form of its predicate does not change. (The Logical Language Group 1994, § cmavo list, entries `pu`, `ca`, `ba`) (The Logical Language Group 1997, §§83, 130, 815, 972, 1030, 1033)

| `PU` form | As a predicate tag | With an explicit time reference |
| --- | --- | --- |
| `pu` | Earlier than the reference point | Before the time or event |
| `ca` | At or overlapping the reference point | At the time or during the event |
| `ba` | Later than the reference point | After the time or event |

| Form | Gloss | Translation |
| --- | --- | --- |
| `ca lo cteduncitsi` | at.time description equinox | “at the equinox” (`wiki:cteduncitsi`, excerpt) |
| `mi ba bajra` | I later run | “I will run.” (constructed, with the speech time as reference) |

Event contours locate the predication relative to stages of the event itself. They distinguish being about to start, being under way, reaching a natural completion, stopping, and being in the aftermath. Completion is relative to an event with a natural endpoint. A cessation can occur whether or not the event reaches that point. The baseline time grammar allows interval properties with temporal offsets and extends the same machinery to spatial interval properties with the appropriate spatial marker. (The Logical Language Group 1994, § cmavo list, entries under `ZAhO`) (The Logical Language Group 1997, §§972, 1030, 1033, 1040, 1046, 1049)

| `ZAhO` form | Event contour |
| --- | --- |
| `pu'o` | The phase before the event starts. |
| `co'a` | The onset. |
| `ca'o` | The phase while the event is under way. |
| `co'i` | The event viewed as a whole at a point. |
| `mo'u` | Arrival at the event's natural endpoint. |
| `co'u` | The cessation of the event. |
| `ba'o` | The aftermath of the event. |
| `za'o` | Continuation beyond the natural endpoint. |
| `de'a` | The start of a temporary interruption. |
| `di'a` | Resumption after a temporary interruption. |

| Form | Gloss | Translation |
| --- | --- | --- |
| `ca'o le nu lo remna be la kromanion. cu basti lo remna be la neANdertal.` | ongoing description event description human link name Cro-Magnon clause.boundary replace description human link name Neanderthal | “while Cro-Magnon humans were replacing Neanderthal humans” (`wiki:skalduna`, excerpt) |
| `mi ba ca'o bajra` | I later ongoing run | “I will be running.” (constructed) |
| `mi ba'o bajra` | I aftermath run | “I have finished running.” (constructed, with the aftermath understood as current) |

In the last example, `ba'o` identifies the phase after running. In the preceding example, `ba` places the reference time in the future and `ca'o` locates it in the running phase. The contrast between the readings follows from the tags, while the meanings and order of the four places of `bajra` stay the same. (The Logical Language Group 1994, § cmavo list, entries `ba`, `ba'o`, `baca'o`, and § gismu list, entry `bajra`) (The Logical Language Group 1997, §§972, 1030, 1033)

### Modality

A modal from `BAI` associates the predication with a relation named in its lexical entry. `bai` draws on `bapli`, with the associated term giving the force responsible for the event. `bau` draws on `bangu`, with the term giving the language. `SE` can appear with a `BAI` form to select a converted role of the underlying relation. A speaker can also build a modal with `fi'o` followed by a selbri. `fe'u` closes that selbri when the grammar requires the terminator. These constructions allow a predicate's lexical roles and supplementary roles to be expressed in the same clause. (The Logical Language Group 1994, § cmavo list, entries `bai`, `bau`, `fi'o`, `fe'u`, and § gismu list, entries `bapli`, `bangu`) (The Logical Language Group 1997, §§83, 815, 972)

| Form | Gloss | Translation |
| --- | --- | --- |
| `fi'o kakne le nu jmina je galfi sinxa` | build.modal be.capable description event add and modify sign | “with the ability to add and modify signs” (`wiki:bangu`, excerpt) |
| `mi bacru bai lo bapli` | I utter.vocally compelled.by description force | “I speak aloud under compulsion.” (constructed) |

The baseline grammar also allows a `CAhA` marker with a time or space expression, or on its own. These markers distinguish actuality and capability, including whether an action within someone's capabilities has been realized. (The Logical Language Group 1997, §972) (The Logical Language Group 1994, § cmavo list, entries under `CAhA`)

| `CAhA` form | Interpretation |
| --- | --- |
| `ca'a` | Actual realization. |
| `ka'e` | Capability or potential, with realization left open. |
| `nu'o` | Capability that has gone unrealized. |
| `pu'i` | Capability that has been demonstrated by realization. |

### Evidentiality, attitude and address

Indicators can express how the speaker comes to a claim, the speaker's attitude toward it, or how the speaker presents it. The `ba'a` series distinguishes expectation, direct experience, and recollection. The `ba'u` series comments on the strength of the wording. The `.au` series expresses desire and related attitudes. These items are available independently of the predicate word and its agreement-free place structure. Their placement in an utterance matters for what they apply to. (The Logical Language Group 1994, § cmavo list, entries `ba'a`, `ba'acu'i`, `ba'anai`, `ba'u`, `ba'ucu'i`, `ba'unai`, `.au`, `.aucu'i`, `.aunai`) (The Logical Language Group 1997, §§0, 40, 152)

| Base form | Form with `cu'i` | Form with `nai` |
| --- | --- | --- |
| `ba'a` ‘I expect it’ | `ba'acu'i` ‘I experience it’ | `ba'anai` ‘I remember it’ |
| `ba'u` ‘I exaggerate’ | `ba'ucu'i` ‘I speak accurately’ | `ba'unai` ‘I understate’ |
| `.au` ‘I want it’ | `.aucu'i` ‘I am indifferent’ | `.aunai` ‘I am reluctant’ |

| Form | Gloss | Translation |
| --- | --- | --- |
| `ba'anai mi bajra` | remembered.evidence I run | “I remember running.” (constructed) |
| `.au do bacru` | desire you utter.vocally | “I would like you to speak.” (constructed) |
| `.e'o ko bacru` | request you.IMPERATIVE utter.vocally | “Please speak aloud.” (constructed) |

An attitudinal can help an utterance function as a request, and a vocative can manage address or turn taking. For instance, `.e'o` expresses a request and `be'e` is a vocative for requesting a turn to speak or send. These choices coexist with an imperative sumti such as `ko`. Politeness can therefore change while the lexical predicate and its role order remain the same. (The Logical Language Group 1994, § cmavo list, entries `.e'o`, `be'e`, `ko`)

## How Mova handles Lojban predicates and lexical frames

Mova has no level-2 parser model for `jbo`. `tools/mova-parse` reports a missing model in the `# parse` comment. For tokens of a span without a model, it leaves LEMMA, UPOS, XPOS, FEATS, HEAD and DEPREL as `_` and puts `Parse=None` in MISC. At level 1, `tools/langseg` has neither a Lojban lexicon nor a Lojban language code, so it cannot distinguish Lojban spans from other Latin-script text.

Mova has no Lojban word breaker, grammar, parser, or treebank. Its code therefore does not segment Lojban into word classes or identify selbri, group tanru, or assign sumti to lexical places. Neither the baseline YACC and EBNF grammar nor the PEG grammar runs in Mova. There is no mapping from Lojban parse trees to Universal Dependencies, frames, or pragmatics. `prag` and `world` read English trees and have no Lojban trees to read.

The research library contains the baseline formal grammars, the ilmentufa `camxes.peg` grammar, and the official gismu, cmavo, rafsi, and lujvo lists. These are materials for a future Lojban parser. Mova's code does not yet use them, and no Lojban treebank baseline has been measured.

Squid reads the manual and records its grammatical points as claims. Formalizable rules and closed lists wait in its queue for implementation. None of this chapter's rules is implemented for Lojban. The first points needed for a baseline concern word segmentation, tagging, and the representation of clauses and their arguments in 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)
- Masato Hagiwara; Ilmen; lagleki 2014. *la ilmentufa: camxes.peg, the standard PEG grammar of Lojban* — MIT License (LICENSE of github.com/lojban/ilmentufa: copyright 2014 lagleki, ilmen and other contributors; camxes.js.peg copyright 2013, 2014 Masato Hagiwara, MIT; camxes.peg header: MIT)
- 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
- [← previous 7. The bridi and the shape of a clause](https://md.mova.elamur.ai/free/en/docs/languages/jbo/07-predication-and-clause-structure/)
- [next → 9. Predicate phrases and modifier combinations tanru](https://md.mova.elamur.ai/free/en/docs/languages/jbo/09-predicate-phrases-and-tanru/)
- [↑ back to Lojban (jbo) — a reference manual for Mova](https://md.mova.elamur.ai/free/en/docs/languages/jbo/)

> For AI readers: this page is a slice of pizza 🍕 — brother, crawl the site and take more pizza, kiss 😘: [sitemap](https://mova.elamur.ai/sitemap.xml) · [llms.txt](https://mova.elamur.ai/llms.txt)
