We now discuss the substructures of the basic components that make up a sentence. Any variety of selbri may be placed in a sentence, or in another substructure below it that contains selbri. Likewise, any variety of sumti may be placed in a sentence, or in another substructure below it that contains sumti. You may see that this can potentially lead to extremely complicated structures nested within one another. Lojban's unambiguous grammar allows even these most complicated structures to be untangled in only one way.
something unspecified (it's either obvious or unimportant)
Some other words in this category include:
it/he/she/they (a specific value)
it/this (the last sentence)
There are many others, each with a particular meaning. For example, there are 9 other words related to ko'a. Each may be used to represent a separate value of it. Since Lojban has no gender or number, these 10 words represent he, she, and they as well, and it becomes more clear why so many are needed to keep track of distinct entities.
zo'e is a place-filler sumti, allowing you to skip over a sumti place in the ordered place structure without specifying a value. The speaker indicates that there is a value, but that it is not important to specify it, or that the speaker thinks it is obvious given the context.
ri is a quick back-reference sumti. It can have a new meaning, depending on the context, every time it occurs. The rules for counting back to 'the last sumti' include some special cases that can't be covered in this summary, but in most simple sentences, the referent will be obvious. There are two other back-referencing sumti of this type.
|I talk about Lojban in Lojban (to someone unspecified.)|
ko is used to express commands. A statement with ko can be interpreted by replacing the ko with do, and then taking the result as a command to the listener to make the sentence true, with himself/herself considered as do:
ko need not be in the first position in the bridi, but rather can occur anywhere a sumti is allowed, leading to possible Lojban commands that are very unlike English commands:
Lojban names always end with a consonant followed by a mandatory pause (which may be very short). No other Lojban word ends with a consonant. Thus names are easily recognized by both their form, and by being marked with a preceding la.
ma indicates a question about the value of a sumti. It is answered by 'filling in the blank', replacing the ma with the intended sumti value. It can be translated as Who? or What? in most cases, but also serves for When?, Where?, and Why? when used in sumti places that express time, location, or cause.
Like ko, ma can occur in any position where a sumti is allowed, not just in the first position:
le specifies a sumti that the speaker has in mind more completely than the pro-sumti we have seen. It does so by introducing a bridi relationship that the sumti in question forms the first (x1) sumti of. This bridi is represented by its corresponding selbri. Description sumti phrases have a terminator at their conclusion, ku, which is omitted when no ambiguity results.
Though Lojban sentences often translate word-for-word into fairly clear English, selbri relations are actually quite unlike English. For example, the selbri bajra expresses a relation of running.
In some sentence positions, bajra might be interpreted as the 'verb' to run; in other positions, as the 'noun' or 'adjective' running. In le sumti [ku], described in the preceding section, it represents the 'noun' interpretation of its x1 sumti place: 'runner'. (Some English words, like cook, have similar properties, but the analogy is weak.)
The simplest form of selbri is an individual word. A word which may by itself express a selbri relation is called a brivla. The three types of brivla are gismu (root words), lujvo (compounds), and fu'ivla (borrowings from other languages). All have identical grammar; they are allowed wherever any selbri appear in these examples.
Some short words may serve as selbri, acting as variables that stand for another selbri. The most commonly used of these is go'i, which represents the main bridi of the previous Lojban sentence, with any new sumti or other features in the sentence replacing those in the previous sentence. Thus:
When the word go'i by itself refers back to a bridi marked as a xu true/false question, it repeats that bridi, thereby claiming it is true. Thus, in this sense only, go'i can mean 'yes'. xu questions can also be answered 'yes' by repeating the entire sentence in full, but go'i is much easier to say:
The negation particle na can occur at the beginning of any selbri. It says that the relation claimed by the selbri does not hold (this is called contradictory negation). It may often be translated as "It is false that [sentence]".(discussed below).
If the contradictory negation particle na precedes go'i, the combination na go'i denies the relation claimed by go'i. Thus, after a xu true/false question, na go'i expresses the answer "False", or "No".
If you were to use go'i after a sentence that contained a na contradictory negation, the negation would carry over to the repeated sentence. Unlike English, na go'i would not form a double negative; it merely replaces the na by another na leaving the sentence unchanged. Instead, you must cancel a negative by using the positive equivalent of na, ja'a, to replace the na in the previous sentence:
na deals primarily with the truth or falsity of a bridi. Lojban also supports a separate form of negation, called contrary or scalar negation. A scalar negation attaches tightly to the next brivla of the selbri, modifying the meaning of the word on some scale. Scalar negation structures may appear anywhere where a brivla or selbri is allowed. Scalar negation words include na'e (other than), to'e (absolute opposite of), and no'e (neutral on the scale); je'a is a strong positive scale assertion, translating roughly as 'certainly' or 'indeed':
mo may be used anywhere a brivla or other selbri might be. Keep this in mind for later examples. Unfortunately, by itself, mo is a very non-specific question. The response to the above question could be:(below).
It is perfectly permissible for the respondent to fill in other unspecified places in responding to a mo question. Thus, the respondent in the last example could have also specified an audience, a topic, and/or a language in the response:
se and others in its word-category modify a brivla used in a selbri by changing the order of the sumti that are attached. This results in a new selbri that expresses the same relation, but with different order of emphasis. se exchanges the first and second sumti places of the unmodified brivla. This reversal is called conversion.
The bridi sentence:
|That is sold by you to me.|
The effect is similar to what in English is called 'passive voice'. In Lojban, however, a conversion is not 'passive': the converted selbri has a place structure that is renumbered to reflect the place reversal, thus having effects when such a conversion is used in combination with other constructs (such as fi and le selbri [ku]).
The other simple relatives of se are: te (switches 1st and 3rd places), ve (switches 1st and 4th places), and xe (switches 1st and 5th places). The effects of using them may be seen on the 5-place gismu selbri, klama:
tanru are compound selbri -- constructions of multiple brivla/selbri components. Each component might be a single word, or it might be a word modified by place structure converters (like se), or scalar negations (like na'e). tanru take selbri components (including other tanru) in pairs, with the first part modifying the second part.
The kind of modification is vague: tanru may act like an English adjective-noun (fast-runner), adverb-verb (quickly-run) or it may restrict a larger set (runner-shoes). Context will generally indicate what is a plausible interpretation of a tanru. You should allow for creative interpretation: 'runner-shoes' might be interpreted in some imaginative instances as 'shoes that run by themselves'. In general, however, the meaning of a tanru is determined by the literal meaning of its components, and not by any connotations or figurative meanings. So sutra tavla 'fast-talker' would not necessarily imply any trickery or deception, and a jikca toldi 'social butterfly' must always be an insect with large brightly-colored wings, of the family lepidoptera.
With the conversion se klama in the final position, the place structure is that of se klama: the x1 place is the destination, and the x2 place is the go-er:
A similar example shows that there is more to conversion than merely switching places, though:
|la tam.||[cu]||melbi tavla||la meris.||[vau]|
|Tom beautifully-talks to Mary (or) Tom is a beautiful-talker to Mary.|
|la meris.||[cu]||melbi se tavla||la tam.||[vau]|
|Mary is beautifully-talked-to by Tom (or) Mary is a beautiful-audience for Tom.|
Note that there is no place structure change if the modifying term is converted, and hence there is less drastic variation in possible meanings:
|le blari'o ku||[cu]||pimu cu'o||[vau]|
|The blue-green (occurrences) are a .5 probability (have a 50% chance).|
Note the interpretation of x1 in the last example, which is a result of the place structure of probability numbers. Each of these special kinds of selbri have other places besides the x1 sumti that appear in these examples.
Each component of a tanru is not merely a single-word brivla, but a representation of an entire bridi relationship. Lojban grammar allows the sumti that complete and define that bridi to be incorporated into the selbri. Combined sumti are called internal sumti. We'll first show the structure of such a complex selbri component:
where the sumti attached with be is normally x2, and other sumti are optionally attached in numerical order (x3, x4, x5), each preceded by the marker bei. be'o is the end-marker for internal sumti, appearing after the last internal sumti for a brivla or other selbri. Let's now look at one way that this construct is used.
brivla/selbri be sumti [bei sumti] [bei sumti] ... [bei sumti] [be'o]
|[cu]||tavla||be||do||bei||le melbi ku||[be'o]|
|That is a talker-to-you-about-the-beautiful-thing(s) salesperson (or, more simply) That's a salesperson who talks to you about beautiful things.|
In compound constructs such as this one, the normally elidable (omissible) right terminators may be mandatory to keep the sentence unambiguous. Thus, in this last example, either the ku or the be'o must not be elided (ku was chosen arbitrarily). Otherwise, vecnu is absorbed into the internal sumti:
|tavla||be||do||bei||le melbi vecnu [ku]||[be'o]|
|That is a talker-to-you-about-the-beautiful-salesperson (or) That one talks about beautiful salespeople to you.|
In the last example, by omitting the elidable terminators ku and be'o, we ended up with sumti attached to the selbri word in final position. The latter sentence is thus identical in meaning to the same sentence expressed without internal sumti:
|ta||[cu]||tavla||do||le melbi vecnu [ku] [vau]|
|That is a talker-to-you-about-the-beautiful-salesperson (or) That one talks about beautiful salespeople to you.|
We rephrased the English translations of the Lojban in the last two examples in order to simplify the English structure and make the sentence more clear. The same type of rearrangement is possible in Lojban. The technique is called tanru inversion. The modifier selbri is placed after the modified one, with a co separating them:
The co can often be translated as 'of type'.
modified-selbri co modifier-selbri
tanru inversion affects the interpretation of sumti that are in the surrounding bridi relationship. The inversion causes a new brivla to be in the final position, and any following sumti are associated with that final brivla of the modifier-selbri. The sumti preceding the selbri are still associated with the final term of the modified-selbri, because that is the primary relation being claimed by the sentence. So, in this case sumti belong to the selbri they are closest to.
One obvious advantage of tanru inversion is to simplify the apparent structure of a selbri. As we have said, the final selbri in a tanru does not need to use internal sumti structures in order to attach its sumti. The first of the examples above that use internal sumti thus becomes the simpler:
|ta [cu] vecnu||co||tavla do le melbi [ku] [vau]|
|That is a seller of-type talker to you about the beautiful thing(s) (or) That's a salesperson who talks to you about beautiful things.|
All of the forms of selbri listed in this section are governed by rules that form a hierarchy. The most complex constructs are those using rules higher in the hierarchy. In general, constructs built from higher rules cannot be used inside lower-rule constructs. This hierarchy of rules is the primary reason why Lojban's grammar is unambiguous.
Inversion of tanru uses rules which are highest in the hierarchy, thus allowing you to invert almost all other selbri constructs. However, this also means that you cannot substitute a tanru inversion into most other constructs within a selbri.
tanru may be composed of more than two components, any of which may be more complex than the simple brivla and/or other selbri structures discussed above. Lojban allows complex tanru structures to be unambiguously expressed, so that any such complex structure can be broken down into a series of modifier-modified pairs. We present two of the variety of ways to express more complex groupings.
In the absence of any grouping indications, components in tanru are presumed grouped from the left::
|This||is-a boys||school||of-type small.|
|This is a boys school which is small.|
tanru inversion can work for many simple grouping problems. However, since it changes the final brivla, it affects the interpretation of any sumti following the selbri. There is a more general solution that does not affect which selbri is in final position, a choice that might be important because of the structural markers required.
Any selbri, or any portion of a tanru that could stand alone as a selbri, may be surrounded with word-brackets ke (left) and ke'e (right) to indicate priority in grouping. Normally, you will only use ke/ke'e grouping around strings of two or more selbri components, since the structure conveys no useful grouping information around a single selbri. At the end of the selbri, and in other places where no ambiguity results, the ke'e terminator becomes optional (elidable):
A selbri structure surrounded by ke and ke'e has the same grammar as a single word brivla. As a result, you can modify such structures with na'e and other scalar negation words, or with se and other conversion words:
The final selbri form we cover here isn't found very often in selbri in their most basic form. However, it turns out to be one of the most important constructs in the language, showing up frequently as a part of more complex structures. This selbri form is called sentence abstraction. The basic form is:
where the bridi sentence inside the two markers can be a Lojban sentence of any type discussed here, no matter how complex. The word nu indicates an event abstraction, the most common kind of abstraction found in Lojban.
nu-word complete-bridi-sentence [kei]
|event-of "I talk to you."|
|This is an event of my talking to you.|
The term event should not be misconstrued. In Lojban, it can refer to a momentary occurrence, or to a situation lasting hours, days, or even an indefinite period of time. nu can stand for any of these kinds and durations of events. Other words may substitute for nu when you need to be specific and indicate particular event contours, such as a point event in time (mu'e) or a steady, unchanging state (za'i) of indefinite duration.
There are other types of abstractions, as well, each indicated by words that substitute for nu. The most common of these are ka for a property/quality abstraction and du'u for a fact/assertion abstraction. These will be exemplified in the next section. Most abstraction selbri have only one place (x1) which is the event, property, fact or other abstract 'thing' being described by the selbri.
Most abstractions that appear often in selbri like this tend to be abbreviated into a single-word selbri (a brivla) which embeds the nu into the concept. Such compounds don't require a kei, since the abstraction encompasses only the idea expressed within the single word. Lojban compound words (lujvo) are composed of combining forms of their component brivla (content words like bajra) and cmavo (short structure words like nu).
The rules for constructing lujvo are not difficult if you have a list of the combining forms (called rafsi); the rules are designed carefully to ensure that the pieces stay attached together and cannot be accidentally interpreted as separate words. This is because (as in the last example) the grammar of separate words may require added markers and terminators to obtain the meaning that you intend. The lujvo for an event of running is formed from nun, the rafsi combining form of nu, and bajra, which serves as its own combining form:
|x1 is an event of x2 running on surface x3 using limbs x4, with gait x5.|
Comparing this place structure with the one at the beginning of this section, you will see that x1 has been assigned to the event, while the remaining places are those of bajra. With this new lujvo, the next to the last example sentence: