Lesson 17
pazemoi seltadni
Relative Clauses
sumti ke galfi bridi
In trying to express your ideas as relationships (bridi) among
conceptual elements (sumti), you will find that the language tools we
have presented so far are capable of expressing almost anything. You
now have ways of expressing tense and number, and can build a
virtually unlimited 'vocabulary' of concepts using tanru.
The language presented so far is adequate; it is not sufficient for a
variety of human expressive needs. This lesson will concentrate on
one area where people need a great deal of expressive power, that of
description. What can you do with sumti to make them effective at
conveying your ideas?
First, we must recognize what the problems are. The most obvious are
shown in the following exchanges:
[I - You may wish to act this out with an assistant in a form of
dialog. To do so, rewrite it on a separate sheet, making choices
where multiple possible answers exist. The person representing the
'Alice' character, who need not be female, should be properly
frustrated and successively more displeased each time 'Fred' confuses
'her' intent. Ensure that you have appropriate props for your
choices; you may have to modify the dialog to fit props at hand..
Perform the dialog as an introduction. Identify the key elements: the
books, the tables, etc., but do not explain the plot. Then, speaking
slowly so that the students can easily follow what you are saying, act
out he dialog 'with feeling'. Then go back over each line; ask
students what each sentence means and explain only if they didn't
understand. You may then choose to repeat the role-play so that the
students can follow with understanding; this is not necessary if the
discussion indicates that most followed it the first time.
A: ko cpacu le xunre cukta le jubme
"Get the red book from the table."
F: le ki'a xunre cukta
"Which red book?" (The which? red book.)
A: ???
Alice may want to answer by specifying a title or author: "The
one written by Mao Zedong." or "The one named 'Tom Sawyer'." Or
perhaps she is referring to the specific book which is hers.
She could answer by making statements about the book:
"ko'a cukta zo'e la maudzedy,n.", "ko'a se cmene lu
tam. soier. li'u", or "ko'a se ponse mi"
presuming that Fred can guess that "ko'a" is the book and not
the table. If Fred doesn't understand this, he will simply ask: "ko'a
ki'a" ("What is "ko'a" referring to?"), and Alice is stuck.
The case of possession is even more difficult; Fred might
interpret "ponse" as indicating that she has the book, and not that
she legally owns the book, but doesn't currently have it. Alice might
have clarified the distinction with a tanru, but this would have
required another exchange and a chance for Fred to misinterpret the
tanru.
F: le ki'a jubme
"Which table?" (There are two things in the room that might
be considered tables.)
A: ???
Alice wants to say "That one.", or "The one which is near (me)."
She can point with : "ta" or state "le jubme cu jibni mi"
("The table I am describing is near me.").
At which point, Fred will note that the one he thinks she is
indicating isn't really a table, but an overturned box. He then asks:
F: le jubme ki'a
"The table???" I don't see a table there.
A: ???
Alice wants to say "The thing that I am describing as a table
is really a box.": "le jubme cu tanxe"
F: le xunre ki'a cukta
"The red??? book". There is a pink one and a maroon
(reddish-brown) one. He either doesn't recognize either as 'red', or
possibly he sees both as 'red', and doesn't know what standard of
redness Alice is using. There is also a white book with a red picture
on the cover which he doesn't even consider as a possibility since he
has interpreted the tanru as describing the color of the book cover.
A: If she wants either of the 'off-red' ones, she can elaborate the
tanru by adding another unit: "le blabi xunre cukta" or "le xunre
bunre cukta". If she wants the white book, she might say: "le blabi
cukta cu se jadni le xunre" ("The white book is adorned with the red
(thing).")
Fred is now totally confused because Alice is now talking about a
white book instead of a red one. Alice says "oi", obtains the book
herself, and decides that she and Fred don't communicate too well.
Human communication cannot always tolerate inaccuracy in a
description, or in interpretation of a description. Several types of
problems were evidenced in this exchange, including:
- the vagueness of tanru units;
- the semantic ambiguity in the relationship between the tanru terms;
- identifying a pro-sumti that has never been defined or assigned;
- difficulties in expressing locations;
- differences in interpretations of descriptions;
- failure to indicate an 'inalienable' possession -
something Alice considered to be hers even
though it wasn't immediately in hand.
In spite of the confusion evident in the example exchange, Lojban has
the capability of dealing with all of these problems clearly, if Alice
and Fred had known how to do so. This lesson will cover many of the
necessary techniques.
17.1
Identifying Descriptions
***
The first problem we will deal with is that of identifying a
description. Remember that the definition of "le" is "the one(s) that
I am describing as ...". The speaker has a certain idea or picture in
his/her mind, and exactly what elements are being described. The
problem is how to convey these description clearly to the listener -
how to identify the elements using words and structures that both
speaker and listener share. Sometimes the selbri that is used in the
description is sufficient to establish this common ground. We have
seen that this isn't always the case.
In almost every natural language, a major tool or structure exists to
provide identifying information for descriptions. This structure is
called a relative clause, and it can take many forms in the various
languages. In English, there are several irregular forms for relative
clauses; the most complete for familiar uses what is called a
'relative pronoun'. The word "what" in the last sentence was a
relative pronoun, and "which", "that", "who" and "where" occasionally
serve as relative pronouns. The sentence you are now reading is one
which uses the word "which" as a relative pronoun. A simpler example:
The woman who carried the box will read the book. (Example 5-1)
The relative pronoun "who" in this example is standing for "the woman"
in a kind of parenthetical sub-sentence that identifies her:
The woman (She carried the box.) will read the book.
Natural languages are often irregular in their construction of
relative clauses; the speaker can omit the relative pronoun, and
rearrange the construction so that it no longer looks like a
parenthetical sentence:
The man with red hair sits in the chair. (Example 5-2)
which is a simplification of:
The man who has red hair sits in the chair.
which implies:
The man (He has red hair.) sits in the chair.
The relative pronoun may not always be the 'subject':
The baby that Jill holds lives next door. (Example 5-3)
or:
The baby (Jill holds it.) lives next door.
In all of these examples, we are presuming that the listener does not
know exactly who or what is being referred to. The relative clause
provides the additional information needed to clarify the referent.
Relative clauses can be nested inside each other. A simple example is:
The baby that the woman named Jill holds lives next door. (Example 5-4)
or:
The baby (The woman (She is named Jill.) holds it.) lives next door.
Another example is the opening phrase of the preamble to the United
States Constitution:
We the people of the United States... (Example 5-5)
or:
We (We are the people (They dwell in the United States.) .)...
where we are making a convenient assumption as to the exact meaning of
the word "of" in this phrase.
It is very rare for relative clauses to nest more than two levels, and
when such nesting is done, it must be very regular or the listener
will completely lose track of what is being said. The extreme English
example of nested relative clauses is the nursery rhyme: "The House
That Jack Built", which is of course extremely regular.
How do relative clauses work in Lojban? You might guess that you
express the relative clause as a bridi, and you would be correct. You
then embed this bridi in the sentence so that it unambiguously
attaches to the sumti that it is clarifying, using a marker and a
terminator (like "nu" and "kei" for abstraction bridi).
17.2
On poi
me zo poi
The word which attaches a relative bridi to a sumti that it is
identifying is "poi", a member of selma'o NOI, which can be roughly
translated as the English "which is ...". (We will learn about "noi"
later in the lesson.) The terminator is the cmavo "ku'o". Using
'relativized' to refer to the sumti being elaborated, the schematic
structure of a sumti that is relativized is:
<(le sumti ku) (poi {...bridi...[vau]} ku'o)>
where the ellipsis represent the sumti of the relative bridi. ***
rewrite per Change 20 ***
Note that the "vau", as usual, can be elided. You will find that, in
almost all situations, you can elide the "ku'o" as well. We will
present the one identified case where you cannot elide it in a later
lesson.
Let us first express Example 5-1 in Lojban:
The woman who carried the box will read the book. (Example 5-1)
<(le ninmu [ku]) (poi {[zo'e] [cu] pu bevri [vau]}
[ku'o]> [cu] ba tcidu [vau] [fa'o]
or, eliminating the elidable terminators, and the elliptical "zo'e"
that is place-holding for the relative pronoun:
<(le ninmu) (poi {pu bevri }> ba tcidu
le ninmu poi pu bevri le tanxe ba tcidu le cukta
and you can truly see the simplifying advantage of elision. This is
one of the places where you might want to use a "cu", even though it
isn't really needed, in order to clearly mark the main selbri.
17.3
On ke'a
me zo ke'a
The "zo'e" in the last example is NOT the relative pro-sumti. (We
will use the term "relative pro- sumti" for Lojban, and "relative
pronoun" for English, to emphasize that Lojban sumti are not really
the same as English nouns.) We were simply explicitly noting that we
had ellipsized the place where the relative pro-sumti should go. We
use the word "ke'a" to specifically represent the relative pro-sumti
when ellipsis isn't appropriate or desired. "ke'a" in a relative
bridi specifically refers to the description that this bridi is
relativizing; "ke'a"s may be found in different parts of a sentence
relativizing different descriptions. You can thus insert "ke'a" in
the translation for Example 5-1:
ba tcidu le cukta
le ninmu poi ke'a pu bevri le tanxe ba tcidu le cukta
Example 5-2 may similarly be translated:
The man with red hair sits in the chair. (Example 5-2)
le nanmu poi ke'a ckaji le xunre kerfa cu zutse le stizu
Often, you can expand a tanru by rephrasing it with a relative bridi
(but not always). Just as the English relative clause in the last
example could have been expressed as the adjectival phrase "the
red-haired man", the Lojban could have been worded with a tanru: "le
xunre kerfa nanmu". This is riskier in Lojban than in English. The
English bears only one reasonable interpretation. After the exercises
in Lesson 11, you almost certainly can think of a situation where the
Lojban tanru would not refer to the adornment on the man's head. An
example of misinterpretation of tanru as a relative clause is the
basis for the humor in the "lei lojbo" comic strip found at the end of
this lesson.
"ke'a" is likely to be ellipsized when it is found in the x1 sumti of
the relative bridi, right after the "poi", just as the first sumti of
a "nu" abstraction is usually ellipsized because it is obvious. You
are allowed to express "ke'a" there, but it will usually be the
'plausible' interpretation if omitted. Example 5-3, however, has a
straight-forward translation using the relative pro-sumti in a
different place of the relative bridi:
The baby that Jill holds lives next door. (Example 5-3)
xabju le lamji zdani
le cifnu poi la djil. jgari ke'a xabju le lamji zdani
Note that in this case you can elide all of the terminators and the
"cu". The relative bridi marked by "poi" prevents the main selbri
"xabju" from being absorbed into the x1 description, while the "ke'a"
effectively keeps the "xabju" from being absorbed into the relative
bridi. Without the "ke'a", either the "cu", the "ku'o", or the "vau"
would be needed to unambiguously separate the selbri of the relative
bridi from the main selbri. Otherwise, the result is grammatical, but
doesn't make much sense:
(*?)
The baby which Jill graspingly inhabits the next door (of?)...
which is an incomplete sentence sumti with an ellipsized relative
pro-sumti; it is hard to guess how that relative pro-sumti would fit
in, since "xabju" has only two defined places in its structure.
(There are possible ways - our English translation is that Jill
graspingly lives next door to the baby, and we will eventually learn
how to explicitly specify a sumti of a description such as "le lamji
zdani".)
Example 5-3 can be manipulated using conversion to put the relative
pro-sumti in the easily ellipsized x1 position of the relative bridi:
xabju le lamji zdani
le cifnu poi se jgari la djil. xabju le lamji zdani
The translation of Example 5-4 shows a nested relative clause with two
"ke'a"s to represent the respective relative pronouns:
The baby that the woman named Jill holds lives next door. (Example 5-4)
xabju
le cifnu poi le ninmu poi ke'a se cmene zo djil. jgari ke'a xabju
le lamji zdani
The innermost "ke'a" could be ellipsized, since it is in the x1
position, making it easier to identify the other "ke'a" which is
widely separated from the description it refers to:
xabju
le cifnu poi le ninmu poi se cmene zo djil. jgari ke'a xabju le lamji zdani
Better still would be to use conversion to put the outer relative
clause in the x1 position:
xabju
le cifnu poi se jgari le ninmu poi se cmene zo djil. xabju le
lamji zdani
and we need no relative pronouns to be plausibly clear. The
rearranged sentence would retranslate back to English as:
"The baby who is held by the woman who is named Jill inhabits the
adjacent-nest."
Read with no phrasing, an English listener will have trouble telling
what each 'who' means. We either need commas (and corresponding
phrasing pauses) in this English sentence to determine the sentence
structure (this is a typical British English approach):
"The baby, who is held by the woman who is named Jill, inhabits the
adjacent-nest."
or we shorten the sentence to:
"The baby who is held by the woman named Jill inhabits the adjacent-nest."
(more common in American English); the latter eliminates the relative
pronoun and obscures the logical structure, even though it makes clear
'who is named Jill'. A third approach is useful only in written
English:
"The baby who is held by the woman-who-is-named-Jill inhabits the
adjacent-nest."
Clearly, the simpler phrasing in English is not the simpler phrasing
in Lojban, and vice versa.
17.4
On du
me zo du
"du" behaves grammatically like a normal brivla, even though it is a
cmavo rather than one of the standard brivla forms, and it expresses
the relationship of "identity". This relationship is similar to that
of "is identical with", but often implies that one of the two sumti
presents new information not previously known. (It is possible that
both sides are known, of course; classically, 'identity' includes
tautology, such as "ko'a du ko'a".) "du" translates as the "=" of
mathematics.
In non-mathematical usage, "du"has a specific denotation. It is used
to link a known referent with a description that is possibly unknown
to the listener (though not necessarily). "du" is not the English
verb 'is', or any of its relatives; it makes a statement about
identities; the identity of the referent of the 'left' side is the
same as the referent of the 'right' side. "du" is symmetric - it does
not indicate which sumti is the one better known. No connotations or
stylistic criteria for choosing a specific order have been proposed,
and any such criteria would be suspect as far as cultural neutrality,
given the strong asymmetry of similar forms in English.) The
sentences:
la bab. du mi
and
mi du la bab.
are identical in meaning . "du" is merely expressing the relationship
of equal identity between the two sumti on either side.
*** bad example ***
Relative bridi are not restricted to descriptions, just as they are
not so restricted in English. Example 5-5 shows a pronoun that is
identified with a relative clause/bridi:
We the people of the United States... (Example 5-5)
mi poi du le prenu poi xabju la .iunaitydsteits.
We have used "du" in its identifying role here; "du" phrases are
considered bridi sentences for the purpose of being used in relative
bridi. In this case it makes perfect sense to omit one side of the
"equation" indicated by "du". You cannot leave the "du" out and just
use "le prenu" because a relative bridi MUST be a bridi:
*mi poi le prenu poi xabju la .iunaitydsteits.
You cannot remove all but the "prenu", since the second "poi" would
have no sumti to relativize:
*mi poi prenu poi xabju la .iunaitydsteits.
When you use a relative bridi to identify a 'pro-sumti', you do not
generally use "ke'a" to represent it in the relative bridi. It is
easier and clearer to repeat the "mi" if you do not ellipsize:
mi poi mi du le prenu poi ke'a xabju la .iunaitydsteits.
or even:
mi poi le prenu du mi poi mi xabju la .iunaitydsteits.
Repeating the "mi" is quite redundant, though given the last phrasing,
the second "mi" cannot be ellipsized, since it would leave nothing for
the second "poi" to attach to:
*mi poi le prenu cu du poi mi xabju la .iunaitydsteits.
Note that the English "we" used in this example is not the "me and
you" translated as "mi'o"; it indicates instead that the "speakers" in
this sentence are plural. As translated, the Lojban "hides" this
plurality, since the expression of number in Lojban is optional. In a
little while, we will see one of the several ways to make the
plurality of "mi" explicit.
17.5
On po'u
me zo po'u
There is a way to simplify the pure identification that was reflected
in Example 5-5. The word "po'u" may be thought of as something like
an abbreviation for "poi ke'a du":
mi poi du le prenu poi xabju la .iunaitydsteits.
becomes:
mi po'u le prenu poi xabju la .iunaitydsteits.
It is not, strictly speaking, accurate to call "po'u" an abbreviation,
though that concept will help you clearly remember the strong
relationship between "du" and "po'u". "po'u" takes two sumti and
equates them, just as "du" does, but it does so in the context of
making the second sumti a 'relative sumti' (which is not to be
confused with "ke'a", the 'relative pro-sumti').
Unlike "du", "po'u" is not generally symmetrical. "du" generally
implies that the referent of one sumti is more clear to the listener
than the other, and the 'equation' is conveying the information that
both refer to the same thing. There is no requirement as to which
comes first, so "du" is symmetrical.
With "po'u", on the other hand, the relativized sumti is assumed to be
too general and ill-defined to the listener, and the relative sumti
provides useful additional information by providing another
designation: something else that labels the same referent. If this
were not the case, you wouldn't need the relative sumti. The relative
sumti is then said to 'restrict' the relativized sumti. ("po'u" sumti
are called 'restrictive relative phrases' or 'appositives' by
linguists and logicians, and "poi" bridi are called 'restrictive
relative clauses'.)
The elidable terminator for "po'u" (as well as for the other members
of selma'o GOI, to be introduced in this lesson and the next one) is
"ge'u"; it turns out to be almost always possible to elide this
terminator, except when very complex sumti are being related.
It should be evident that a "po'u" relative sumti cannot always be
exchanged with its relativized sumti. It is grammatical to say:
"The people who are we of the United States"
or:
le prenu po'u mi poi xabju la .iunaitydsteits.
but it sounds strange, because the use of "we" (or "po'u mi") in this
example conveys no more useful information about who 'the people' are
(although the "poi" does).
Thus, in Lojban, the relative sumti should be another designation that
provides additional useful information about the sumti it relativizes.
However, just like with "du", the sumti are being equated; it will be
natural for people to treat "po'u" as a symmetric operator when it is
useful for them, even though it muddies the waters as to which sumti
restricts which.
An example of this interchangeability may be shown when two "le"
descriptions are related with "po'u". We can, for example, rephrase
Example 5-2 as:
le nanmu po'u le xunre se kerfa cu zutse le stizu
which translates as:
"The man, who is the red-haired one, sits in the chair."
or more completely:
"The one I am describing as a man, who has the same identity as the
one I am describing as the red-haired one, sits in the chair."
Presumably, "the red-haired one" designates an individual who is
obvious to the listener without additional information. It is
supposed to be a additional restrictive description on "the man".
Imagine a room where most are red-heads, and only a few are men, with
exactly one of the men having red-hair. Note that in and of itself,
"the red-haired one" is a description that could apply to anyone in
the room. If this were so, however, using it with "po'u" in the above
sentence would be misleading. The use of "po'u" implies a "du"
relationship - that all of those described as "the red- haired one(s)"
could also be designated by the description "the man(men)". At face
value, the situation would be a misuse of "po'u". However, if the one
man who has red hair has been previously described as "le xunre se
kerfa" or if there is some other reason why this particular red-
haired referent is "the one that I am describing" by using "le", we
can treat this as a restricting designation. IF THE TWO DESCRIPTIONS
ARE NOT BOTH IDENTIFYING EXACTLY THE SAME REFERENT, THE USE OF "po'u"
IS INCORRECT. Use "poi" as in:
le nanmu poi xunre se kerfa cu zutse le stizu
"The one I am describing as a man who is a red-haired one sits in the chair."
or:
le nanmu poi se kerfa le xunre cu zutse le stizu
"The one I am describing as a man who is haired by red ones sits in the chair."
or stick to the original "poi" phraseology using "ckaji".
Alternatively, you can use "pe", which will be defined in the next
lesson.
The 'equation' of two descriptions should be seen to give a different
flavor than the "poi"-based phrasing. Using "poi" caused us to
express the relative information as a bridi, and the relationship of
that information was thus explicit. We know, in effect, that "ckaji"
relates the hair as a 'property' characterizing the man. In the
variation without "ckaji", the man is literally being festooned with
red hairs.
The "po'u" phrasing does not express this colorful relationship
between the two descriptive designations that are joined, nor any
other relationship except "du" - the two sumti are merely two-
mutually-helpful designations for the same thing.
Another example will show how "po'u" is more likely to be used with
descriptions:
Joe the teacher is coming.
la djos. po'u le ctuca cu klama
where the person named is being distinguished from, say, Joe the
doctor. Here, people named Joe are a probably more restricted set
than teachers, but the phrasing works just as well as:
The teacher who is Joe is coming.
le ctuca po'u la djos. klama
In this case, we are presumably distinguishing between the teacher who
is Joe and, say, the teacher who is Rebecca.
"du" expresses a very specific relationship between two sumti that is
seldom hard to interpret; "po'u" when used in the same way indicates
the same relationship, but in a much weaker way. "po'u"s different
purpose, that of restricting, implies only the non-symmetric "du" case
where the left (x1) sumti is unknown and the right (x2) sumti is
known. Furthermore, the fact of the equivalent identities is made
incidental to the main bridi of the sentence, which is after all the
relationship that the speaker considers most important. Otherwise our
revision of Example 5-2 might read:
"The man who is the sitter in the chair is identical to the red-haired one."
or:
le nanmu poi zutse le stizu du le xunre se kerfa
It is not certain where the restricting effect of interpreting "po'u"
will lead in actual usage; it did lead the language designers to add
other operators with the same grammar as "po'u" for the purpose of
conveying specific relationships that might overlap with the
restricting identity function indicated by "po'u". In this way,
"po'u" will more naturally be retained for that identifying function
which is its nominal denotation. We will cover several relatives of
"po'u" that are most useful later in this lesson.
17.6
The ko'a Series Of cmavo
me zo ko'a cmavo
One function that might be considered a valid use for "po'u" is the
definition of pro-sumti like "ko'a". For example, in the "boring
story" of Lesson 10, the narrator suddenly uses "ko'a" for la
.alis. without ever saying that "ko'a" was representing her. This is
a very English-like short-cut. We often just stick in pronouns like
'her' and 'it', and leave it up to the listener to interpret them
correctly.
This is not always easy to do, and many language-specific jokes in
English rely on the ambiguity of pronouns. This is true even though
in the case of English, the listener has the additional clues embedded
in the 'person', 'gender', and 'number' that are associated with each
pronoun.
The Indo-European languages, of which English is one, tend to be
characterized by the additional modes of 'person', 'gender', 'case',
and 'number'. This is because you often wouldn't be able to figure
out what a sentence means without these vital clues that tie pieces of
the sentence together. In other European languages, these structures
are more elaborate: German divides all nouns into one of three
'genders' which have (for the most part) nothing to do with the gender
or association of their referent. Russian uses a complex set of
declensions and inflections for nouns, pronouns, verbs, and adjectives
that involve 'gender', 'number', 'person' and 'case' (like the
declensions for English verbs, especially irregular ones like "to be",
combined with those for the English pronouns). Russian thus has more
clues in speech as to what words go together, and thus has more
freedom in its word order. Latin, which also had such declensions,
had more freedom of word order than English.
This side note has a point, other than to be informative about the
relationships between Lojban, English, and other languages.
Languages that have lots of clues embedded in special forms of words,
are able to use those words more freely. English, which has
relatively few declensional forms, is very sensitive to word order; we
demonstrated this in Lesson 1. Lojban has no 'person', no 'gender',
no 'case', no 'number', absolutely no clue that tells a listener what
a 'pronoun' refers to. Indeed Lojban has a severely constrained word
order in parallel to the lack of these added clues (though not much
more constrained than English).
Since Lojban is designed with the intent of easing the listener's
burden, there has to be a solution. Lojban provides that solution in
two parts:
- Lojban provides a large set of permitted pro-sumti, and
- Lojban provides a method of assigning them either before or upon
first usage ('forethought' and 'afterthought' definition).
We will discuss the first of these two points now, and the second one
in the next section.
English distinguishes a total of seven singular and plural pronouns:
"I", "you", "he", "she", "it", "we", "they". In an optimal sentence,
you can have up to seven pronoun variables floating around
representing specific things, but only if all seven exactly match the
person, gender, and number of the corresponding available pronouns.
If not, you end up with two "it"s in your sentence, and the listener
inevitably gets confused. We occasionally stretch to include the
numbers such as "one" and "many" as gender-less, case-less, 'pronouns'
to solve the problem of insufficient free pronouns, and grammatical
ambiguity reigns supreme.
(Linguistic realities often constrain ideal goals; they certainly
constrained the designers of Lojban, who had to include all kinds of
special features of significant linguistic value in the attempt to
meet the needs of human language use. A similar instance relevant to
current society is the occasional popular attempts to devise a
genderless English 3rd-person pronoun, intended to eliminate sexual
bias. The need for more pronouns, rather than fewer, causes a
linguistic constraint that interferes with acceptance of such a
concept - irrespective of its societal merits. The elimination of
gendered third person pronouns 'he' and 'she' would increase instances
where one pronoun must represent two kinds of referent at once,
already a problem when trying to interpret "it". This would increase
the already severe burden on a listener. It would also be even harder
to teach English to immigrants who are non-English natives; such
speakers are used to rich clues from their native languages - clues
that are often linguistic relics that no longer represent the
distinction that generated them.)
Lojban needs variables at least as badly as English. As a 'logical'
language, people will expect to manipulate objects symbolically as
variables, which is what 'pronouns' allow. Lojban provides 10 basic
variables for such assignment; these are totally 'free' of the baggage
or the boon of person, number, gender, or case, free to be assigned to
any object or concept that can be represented by a sumti. If these
aren't enough, we will learn later how to gain additional free
variables through unlimited subscripting of each variable, as well as
through the use of lerfu, the word/symbols for the letters and special
characters of the alphabet. Your supply of 'free' variables is both
'free' and infinite.
You've met the first two of these free variables: "ko'a" and "ko'e".
The remainder of the series uses the other three vowels in the final
position, and then uses all five vowels in the final position coupled
with the letter 'f' in the beginning:
ko'a ko'e ko'i ko'o ko'u
fo'a fo'e fo'i fo'o fo'u
While people usually start at the beginning of this list, and seldom
get past the third or fourth, you can use any of them, in any order.
In a noisy environment you will want the extra redundancy of changing
the first letter as well as the last one. Be careful, however. A
Lojbanist, especially an inexperienced one, who enters your
conversation in the middle, and hears you talking about "fo'o" and
"fo'u", might presume that the other eight had already been assigned,
thus being hard-pressed to determine which, if any, she/he can use.
17.7
On goi
me zo goi
Having solved half the problem of pro-sumti, we now present the other
half of the solution:
Lojban pro-sumti are specifically tied to their referents AT OR BEFORE
first use, whenever practical. This is done using "goi", a near
relative of "po'u". "goi" serves as the namesake of selma'o GOI,
which includes all relatives of "po'u" that attach two sumti together
within a single sumti position of a bridi.
The specific denotation of "goi" is identity, and specifically the
assignment of identity. Like "po'u", it indicates that the two sumti
that are joined refer to the same thing.
One of the two sumti is generally a name or a pro-sumti that is
presumed to be available for use as a label or symbol. The other is a
description or designation of something presumably known or
sufficiently identified to the listener. "goi" then indicates that
the unassigned name or symbol shall be considered assigned to the
known referent.
Let us look at the two major ways in which "goi" is used, and we will
quickly see how it differs from "po'u".
FORETHOUGHT MODE:
In forethought mode, the speaker (or writer - forethought is likely to
be used often in writing) has determined in advance that he/she will
be talking about something a lot, possibly the topic of a paragraph or
even an essay. She/he then determines to assign an available member
of the ko'a/fo'a series to not distract the user by continued
repetitions of a longer description.
This, after all, is what 'pronouns' are used for. In English, we use
"it" to save from repeating the referent of "it" - a referent that is
usually longer and always more grammatically complex than the very
simple pronoun. In Lojban, there is nothing with simpler grammar than
a pro-sumti of the KOhA selma'o: it needs no article/descriptor like
"le", it has no terminator like "ku", and of course it has no number,
gender, tense, etc. that have to be expressed every time you use it.
Having determined that we want to use a ko'a/fo'a pro-sumti, we can
then assign it with "goi". Thereafter, unless emphasis causes us to
desire to revert to the longer form, we use the pro-sumti instead.
This use of "goi" can therefore be translated as "henceforth to be
referred to as", or "henceforth used to refer to", depending upon
whether the longer description or the pro-sumti is expressed before
the "goi". (There are no conventions as to which comes first; the
authors tend to put the pro-sumti after the "goi" in forethought
assignments.)
If the translation given sounds like 'lawyer talk' in the U.S. legal
system, this is because lawyers who write agreements have to make very
sure that all pronouns and abbreviations are clearly defined in order
to prohibit misunderstandings based on their use. Lojban, as a
logical language, similarly has to make sure that all variables can be
defined, and not be subject to ambiguities due to misinterpretation.
So, in the 'boring story' of Lesson 10, the first line should read "la
.alis. goi ko'a [cu] klama le zarci". Thereafter, "ko'a" is generally
used to stand for "la .alis.". Every once in a while in the story
thereafter, the name "la .alis." is used instead of the pro-sumti
"ko'a". This is a stylistic choice: "ko'a" could have been used to
replace every instance of "la .alis."
AFTERTHOUGHT MODE:
Normally, most people don't think like lawyers - though perhaps
Lojbanists will prove more capable than most after working with the
rigid defining capabilities of Lojban for a while. But in everyday
speech, conversing people often don't have a single 'topic', and they
seldom think far enough ahead to know just which sumti will be used
enough to justify use of a pro-sumti assignment. Thus, in
conversation, we will often decide on the spur of the moment to use
"ko'a" or its relatives. We are then obligated to define our
reference, our risk being misunderstood.
In this case, we usually will be putting the pro-sumti first, and then
the definition. Thus, if the 'boring story' were to be written in a
conversational narrative manner, it might have started:
la .alis. klama le zarci .i ko'a goi la .alis. djica lenu ko'a cpacu
su'o le taxfu zo'e ...
The order is not a requirement. The speaker could have used "la
.alis. goi ko'a" in afterthought as well.
NON-PRO-sumti USES OF "goi":
You can also use "goi" to label a description with a name:
le prenu goi la sam. dansu le zgike or "The person, who I'll call Sam,
dances to the music."
Since "la sam." already means "the one that I am calling Sam", this
usage must imply something more: specifically that the phrase "le
prenu" refers to someone that can be commonly identified, but that the
speaker is choosing to use the name instead of the description. The
name is not presumed to be known to the listener to refer to "le
prenu", or the name would have been used alone. We can contrast this
with parallel usage of "po'u":
le prenu po'u la sam. dansu le zgike or "The person who I call Sam
dances to the music."
With the use of "po'u" the speaker implies that the name is being used
to restrict the domain of reference from "person" or "people" to "the
one or ones named Sam". However, with the assignment operator "goi",
there is no particular need for the person to even called "Sam" by
anyone - including the speaker - in any other context; it merely
indicates what label is going to be used.
An more natural English example occurs when we use letters - often
initials - to stand for people. In advice columns such as 'Dear
Abby', questioners use initials, or names that they specifically
indicate are not true names. They use descriptions to refer to
themselves. A Lojbanist might thus sign a letter of this type:
".i mi goi la cfipu cu ciska
"I, who I'll describe as "Confused", write."
or
"Signed: Confused"
As we will see in a later lesson, it is perfectly reasonable to use
"la" instead of "le" in situations like this, to indicate that a
description normally to be labelled with "le" is intended for use as a
proper name.