Latest version of lujvo place structure paper

[John Cowan <cowan@ccil.org>]

I have just completed revising (once again) Nick's lujvo place
structure paper.  This form will be Chapter 12 of the reference
grammar.  (It's less than half as long as the original.)  Comments
are solicited.

-- 
John Cowan						cowan@ccil.org
			e'osai ko sarji la lojban

Chapter 12
Dog House And White House: Determining lujvo Place Structures
$Revision: 1.10 $


1.  Why have lujvo?

The Lojban vocabulary is founded on its list of 1350-plus gismu, made up by
combining wordlists from various sources.  These gismu are not intended to
be either a complete vocabulary for the language nor a minimal list of
semantic primitives.  Instead, the gismu list serves as a basis for the
creation of compound words, or lujvo.  The intention is that (except in
certain semantically broad but shallow fields such as cultures, nations,
foods, plants, and animals) suitable lujvo can be devised to cover the
ten million or so concepts expressible in all the world's languages taken
together.  Grammatically, lujvo behave just like gismu: they have place
structures and function as selbri.

There is a close relationship between lujvo and tanru.  In fact, lujvo are
condensed forms of tanru:

1.1)	ti fagri festi
	That is-fire waste.

contains a tanru which can be reduced to the lujvo in:

1.2)	ti fagyfesti
	That is-fire-waste.
	That is-ashes.

Although the lujvo "fagyfesti" is derived from the tanru "fagri festi", it
is not equivalent in meaning to it.  In particular, "fagyfesti" has a
distinct place structure of its own, not the same as that of "festi".
It needs to take account of the places of "fagri" as well.  When
a tanru is made into a lujvo, there is no equivalent of "be...bei...be'o"
(described in Chapter 5) to incorporate sumti into the middle of the lujvo.

So why have lujvo?  Primarily to reduce semantic ambiguity.  On hearing a
tanru, there is a burden on the listener to figure out what the tanru
might mean.  Adding further terms to the tanru reduces ambiguity in one
sense, by providing more information; but it increases ambiguity in another
sense, because there are more and more tanru joints, each with an ambiguous
significance.  Since lujvo, like other brivla, have a fixed place structure
and a single meaning, encapsulating a commonly-used tanru into a lujvo
relieves the listener of the burden of creative understanding.  In addition,
lujvo are typically shorter than the corresponding tanru.

At present, there are no absolute laws fixing the place structure of a newly
created lujvo.  The maker must consider the place structures of all the
components of the tanru and then decide which are still relevant and which
can be removed.  What is said in this chapter represents guidelines,
presented as one possible standard, not necessarily complete, and not the
only possible standard.  There may well be lujvo that are built without
regard for these guidelines, or in accordance with entirely different
guidelines, should such alternative guidelines someday be developed.  The
reason for presenting any guidelines at all is so that Lojbanists have a
starting point to explore a likely place structure --- one that others seeing
the same word can also arrive at by similar consideration.

If the tanru includes connective cmavo such as "bo", "ke", "ke'e", or "je",
there are ways of incorporating them into the lujvo as well.  Sometimes
this makes the lujvo excessively long; if so, the cmavo may be dropped.
This leads to the possibility that more than one tanru could produce the
same lujvo.  Typically, however, only one of the possible tanru is
useful enough to justify making a lujvo for it.

The exact workings of the lujvo-making algorithm, which takes a tanru
built from gismu (and possibly cmavo) and produces a lujvo from it, are
described in Chapter 4.


2.  The meaning of tanru: a necessary detour

The meaning of a lujvo is controlled by --- but is not the same as --- the
meaning of the tanru from which the lujvo was constructed.  The tanru
corresponding to a lujvo is called its "veljvo" in Lojban, and since there
is no concise English equivalent, that term will be used in this chapter.
Furthermore, the left (modifier) part of a tanru will be called the
"seltau", and the right (modified) part the "tertau", following the
usage of Chapter 5.  For brevity, we will speak of the seltau or tertau
of a lujvo, meaning of course the seltau or tertau of the veljvo
of that lujvo.  (If this terminology is confusing, substituting "modifier"
for "seltau" and "modified" for "tertau" may help.)

The place structure of a tanru is always the same as the place structure
of its tertau.  As a result, the meaning of the tanru is a modified
version of the meaning of the tertau; the tanru will typically, but not
always, refer to a subset of the things referred to by the tertau.

The purpose of a tanru is to join concepts together without necessarily
focusing on the exact meaning of the seltau.  For example, in the
<cite>Iliad</cite>, the poet talks about "the wine-dark sea", in which
"wine" is a seltau relative to "dark", and the pair of words is a seltau
relative to "sea".  We're talking about the sea, not about wine or color.
The other words are there to paint a scene in the listener's mind, in which
the real action will occur, and to evoke relations to other sagas of the
time similarly describing the sea.  Logical inferences about wine or color
will be rejected as irrelevant.

As a simple example, consider the rather non-obvious tanru "klama zdani",
or "goer-house".  The gismu "zdani" has two places:

2.1)	x1 is a nest/house/lair/den for inhabitant x2

(but in this chapter we will use simply "house", for brevity), and the gismu
"klama" has five:

2.2)	x1 goes to destination x2 from origin point x3
		via route x4 using means x5

The tanru "klama zdani" will also have two places, namely those of "zdani".
Since a "klama zdani" is a type of "zdani", we can assume that all
goer-houses --- whatever they may be --- are also houses. 

But is knowing the places of the tertau everything that is needed to
understand the meaning of a tanru?  No.  To see why, let us switch to a
less unlikely tanru: "gerku zdani", literally "dog house".  A tanru
expresses a very loose relation: a "gerku zdani" is a house that has
something to do with some dog.  What the precise relation might be is left
unstated.  Thus, the meaning of "lo gerku zdani" can include all of the
following: houses housing dogs, houses shaped by dogs, dogs which are also
houses (e.g. houses for fleas), houses named after dogs, and so on.  All
that is essential is that the place structure of "zdani" continues to apply.

For something (call it z1) to qualify as a "gerku zdani" in
Lojban, it's got to be a house, first of all. For it to be a house, it's
got to house someone (call that z2).  Furthermore, there's got to be a dog
somewhere (called g1).  For g1 to count as a dog in Lojban, it's got to belong
to some breed as well (called g2).  And finally, for z1 to be in the first
place of "gerku zdani", as opposed to just "zdani", there's got to be some
relationship (called r) between some place of "zdani" and some
place of "gerku".  It doesn't matter which places, because if there's a
relationship between some place of "zdani" and any place of "gerku", then
that relationship can be compounded with the relationship between the places
of "gerku" --- namely, "gerku" itself --- to reach any of the other "gerku"
places.

Doubtless to the relief of the reader, here's an illustration. We want
to find out whether the White House (the one in which the U. S. President
lives, that is) counts as a "gerku zdani".  We go through the five variables.
The White House is the z1.  It houses Bill Clinton as z2, as of this writing,
so it counts as a "zdani". Let's take a dog --- say, Spot (g1). Spot has to
have a breed; let's say it's a Saint Bernard (g2).  Now, the White House 
counts as a "gerku zdani" if there is any relationship (r) at all between
the White House and Spot. (We'll choose the g1 and z1 places to relate by r;
we could have chosen any other pair of places, and simply gotten a different
relationship.)

The sky is the limit for r; it can be as complicated as "The other day,
g1 (Spot) chased Socks, who is owned by Chelsea Clinton, who is the daughter
of Bill Clinton, who lives in z1 (the White House)" or even worse.  If no
such r can be found, well, you take another dog, and keep going until no
more dogs can be found.  Only then can we say that the White House cannot fit
into the first place of "gerku zdani".

As we have seen, no less than five elements are involved in the definition
of "gerku zdani": the house, the house dweller, the dog, the dog breed
(everywhere a dog goes in Lojban, a dog breed follows), and the relationship
between the house and the dog.  Since tanru are explicitly ambiguous in
Lojban, the relationship r cannot be expressed within a tanru (if it
could, it wouldn't be a tanru anymore!)  All the other places, however, can
be expressed --- thus:

2.3)	la blabi zdani cu gerku be fa la spot.
		bei la sankt. berNARD. be'o
		zdani la bil. klinton.
	The White House is-a-dog (namely Spot
		of-breed Saint Bernard)
		type-of-house-for Bill Clinton.

Not the most elegant sentence ever written in either Lojban or English.
Yet if there is any relation at all between Spot and the White House,
Example 2.4 is arguably true. If we concentrate on just one type of relation in
interpreting the tanru "gerku zdani", then the meaning of "gerku
zdani" changes.  So if we understand "gerku zdani" as meaning "doghouse",
the White House would no longer be a "gerku zdani" with respect to Spot,
because as far as we know Spot does not actually live in the White House,
and the White House is not a doghouse (derogatory terms for incumbents
notwithstanding).

3.  The meaning of lujvo

This is a fairly long way to go to try and work out how to say "doghouse"!
The reader can take heart; we're nearly there. Recall that one of the
components involved in fixing the meaning of a tanru --- the one left
deliberately vague --- is the precise relation between the tertau and the
seltau. Indeed, fixing this relation is tantamount to giving an
interpretation to the ambiguous tanru.

A lujvo is defined by a single disambiguated instance of a tanru. That is
to say, when we try to design the place structure of a lujvo, we don't need
to try to discover the relation between the tertau and the seltau.
We already know what kind of relation we're looking for; it's given by
the specific need we wish to express, and it determines the place structure
of the lujvo itself.

Therefore, it is generally not appropriate to simply devise lujvo and decide
on place structures for them without considering one or more specific
usages for the coinage.  If one does not consider specifics, one will
be likely to make erroneous generalizations on the relationship r.

The insight driving the rest of this chapter is this: while the relation
expressed by a tanru can be very distant (e.g. Spot chasing Socks, above),
the relationship singled out for disambiguation in a lujvo should be
quite close.  This is because lujvo-making, paralleling natural language
compounding, picks out the most salient relationship r between a tertau place
and a seltau place to be expressed in a single word.  The relationship
"dog chases cat owned by daughter of person living in house" is too
distant, and too incidental, to be likely to need expression as a single short
word; the relationship "dog lives in house" is not.  From all the various
interpretations of "gerku zdani", the person creating "gerzda" should pick
the most useful value of r. The most useful one is usually going to be the
most obvious one, and the most obvious one is usually the closest one.

In fact, the relationship will almost always be so close that the predicate
expressing r will be either the seltau or the tertau predicate itself.
This should come as no surprise, given that a word like "zdani" in Lojban is
a predicate.  Predicates express relations; so when you're looking for a
relation to tie together "le zdani" and "le gerku", the most obvious
relation to pick is the very relation named by the tertau, "zdani": the
relation between a home and its dweller.  As a result, the object which
fills the first place of "gerku" (the dog) also fills the second place of
"zdani" (the house-dweller).

This leads to a conclusion, and a corollary.  The conclusion is that,
since the relationship between the seltau and the tertau of the
veljvo is expressed by the seltau or tertau itself, at least one of the
places supplied by the seltau is always going to be equivalent to a place
supplied by the tertau --- and is thus redundant, and can be dropped from 
the place structure of the lujvo.  The corollary is that the precise
relationship between the veljvo components can be implicitly determined
by finding one or more places to overlap in this way.

So what is the place structure of "gerzda"?  We're left with three places, 
since the dweller, the "se zdani", turned out to be identical to the dog,
the "gerku". We can proceed as follows:

(The notation introduced casually in Section 2 will be useful in the
rest of this chapter.  Rather than using the regular x1, x2, etc. to
represent places, we'll use the first letter of the relevant gismu
in place of the "x", or more than one letter where necessary to resolve
ambiguities.  Thus, z1 is the first place of "zdani", and g2 is
the second place of "gerku".)

The place structure of "zdani" is given as Example 2.1, but is repeated
here using the new notation:

3.1)	z1 is a nest/house/lair/den of z2

The place structure of "gerku" is:

3.2)	g1 is a dog of breed g2

But z2 is the same as g1; therefore, the tentative place structure for
"gerzda" now becomes:

3.3)	z1 is a house for dweller z2 of breed g2

which can also be written

3.4)	z1 is a house for dog g1 of breed g2

or more concisely

3.5)	z1 is a house for dog z2=g1 of breed g2

Despite the apparently conclusive nature of Example 3.5, our task is not yet
done: we still need to decide whether any of the remaining places should
also be eliminated, and what order the lujvo places should appear in. These
concerns will be addressed in the remainder of the chapter; but we are now
equipped with the terminology needed for those discussions.


4.  Selecting places

The set of places of an ordinary lujvo are selected from the places of its
component gismu.  More exactly, the places of such a lujvo are derived from
the set of places of the component gismu by eliminating unnecessary places,
until just enough places remain to give an appropriate meaning to the lujvo.
In general, leaving in a place makes the concept expressed by a lujvo more
general; leaving out a place makes the concept more specific. 

It would be possible to design the place structure of a lujvo from scratch,
treating it as if it were a gismu, and working out what arguments contribute
to the notion to be expressed by the lujvo. There are two reasons arguing
against doing so and in favor of the procedure detailed in this chapter.

The first is that it might be very difficult for a hearer or reader, who 
has no preconceived idea of what concept the lujvo is intended to convey,
to work out what the place structure actually is.  Instead, he or she would
have to make use of a lujvo dictionary every time a lujvo is encountered in
order to work out what a "se jbopli" or a "te klagau" is.  But this would
mean that, rather than having to learn just the 1300-odd gismu place structures,
a Lojbanist would also have to learn myriads of lujvo place structures
with little or no apparent pattern or regularity to them.  The purpose of
the guidelines documented in this chapter is to apply regularity and to make
it conventional wherever possible.

The second reason is related to the first: if the veljvo of the lujvo
has not been properly selected, and the places for the lujvo are formulated
from scratch, then there is a risk that some of the places formulated
may not correspond to any of the places of the gismu used in the veljvo of
the lujvo.  If that is the case --- that is to say, if the lujvo places are
not a subset of the veljvo gismu places --- then it will be very difficult
for the hearer or reader to understand what a particular place means, and
what it is doing in that particular lujvo. This is a topic that will be
further discussed in Section 14.

However, second-guessing the place structure of the lujvo is useful in
guiding the process of subsequently eliminating places from the veljvo.
If the Lojbanist has an idea of what the final place structure should look
like (as we did for "gerzda" in Section 2), he or she should be able to pick
an appropriate veljvo, to begin with, in order to express the idea, and then
to decide which places are relevant or not relevant to expressing that idea.


5.  Symmetrical and asymmetrical lujvo

A common pattern, perhaps the most common pattern, of lujvo-making creates
what is called a "symmetrical lujvo".  A symmetrical lujvo is one based on
a tanru interpretation such that the first place of the seltau is equivalent
to the first place of the tertau: each component of the tanru characterizes
the same object.  As an illustration of this, consider the lujvo "balsoi":
it is intended to mean "both great and a soldier" --- that is, "great soldier",
which is the interpretation we would tend to give its veljvo, "banli sonci".
The underlying gismu place structures are:

5.1)	"banli": b1 is great in property b2 by standard b3
	"sonci": s1 is a soldier of army s2

In this case the s1 place of "sonci" is redundant, since it is equivalent to
the b1 place of "banli". Therefore the place structure of "balsoi" need not
include places for both s1 and b1, as they refer to the same thing. So the
place structure of "balsoi" is at most

5.2)	b1=s1 is a great soldier of army s2
		in property b2 by standard b3

Some symmetrical veljvo have further equivalent places in addition to the
respective first places.  Consider the lujvo "tinju'i", "to listen" ("to
hear attentively, to hear and pay attention"). The place structures of
the gismu "tirna" and "jundi" are:

5.3)	"tirna": t1 listens to t2 against background noise t3
	"jundi": j1 pays attention to j2

and the place structure of the lujvo is:

5.4)	j1=t1 listens to j2=t2 against background noise t3

Why so?  Because not only is the j1 place (the one who pays attention)
equivalent to the t1 place (the hearer), but the j2 place (the thing
paid attention to) is equivalent to the t2 place (the thing heard).

A substantial minority of lujvo have the property that the first place of
the seltau ("gerku" in this case) is equivalent to a place other than
the first place of the tertau; such lujvo are said to be "asymmetrical".
(There is a deliberate parallel here with the terms "asymmetrical tanru"
and "symmetrical tanru" used in Chapter 5.)

In principle any asymmetrical lujvo could be expressed as a symmetrical
lujvo.  Consider "gerzda", discussed in Section 3, where we learned that the
g1 place was equivalent to the z2 place.  In order to get the places
aligned, we could convert "zdani" to "se zdani" (or "selzda" when
expressed as a lujvo).  The place structure of "selzda" is

5.5)	s1 is housed by nest s2

and so the three-part lujvo "gerselzda" would have the place structure

5.6)	s1=g1 is a dog housed in nest s2 of dog breed g2

However, although "gerselzda" is a valid lujvo, it doesn't translate
"doghouse"; its first place is the dog, not the doghouse.  Furthermore, it
is more complicated than necessary; "gerzda" is simpler than "gerselzda".

>From the reader's or listener's point of view, it may not always be obvious
whether a newly met lujvo is symmetrical or asymmetrical, and if the latter,
what kind of asymmetrical lujvo.  If the place structure of the lujvo isn't
given in a dictionary or elsewhere, then plausibility must be applied, just
as in interpreting tanru.

The lujvo "karcykla", for example, is based on "karce klama", or "car goer".
The place structure of "karce" is:

5.7)	ka1 is a car carrying ka2 propelled by ka3

A asymmetrical interpretation of "karcykla" that is strictly analogous to the
place structure of "gerzda", equating the kl2 (destination) and ka1 (car)
places, would lead to the place structure

5.8)	kl1 goes to car kl2=ka1 which carries ka2
		propelled by ka3 from origin kl3
		via route kl4 by means of kl5

But in general we go about in cars, rather than going to cars, so a far
more likely place structure treats the ka1 place as equivalent to the kl5
place, leading to

5.9)	kl1 goes to destination kl2 from origin kl3
		via route kl4 by means of car kl5=ka1
		carrying ka2 propelled by ka3.

instead.


6.  Eliminating places

In order to understand which places, if any, should be completely
removed from a lujvo place structure, we need to understand the concept of
dependent places.  One place of a brivla is said to be dependent on another
if its value can be predicted from the values of one or more of the other
places.  For example, the g2 place of "gerku" is dependent on the g1 place.
Why?  Because when we know what fits in the g1 place (Spot, let us say,
a well-known dog), then we know what fits in the g2 place ("St. Bernard",
let us say).  In other words, when the value of the g1 place has been
specified, the value of the g2 place is determined by it.  Conversely, since
each dog has only one breed, but each breed contains many dogs, the g1 place
is not dependent on the g2 place; if we know only that some dog is a St.
Bernard, we cannot tell by that fact alone which dog is meant.

For "zdani", on the other hand, there is no dependency between the places.
When we know the identity of a house-dweller, we have not determined the
house, because a dweller may dwell in more than one house.  By the same
token, when we know the identity of a house, we do not know the identity of
its dweller, for a house may contain more than one dweller.

The rule for eliminating places from a lujvo is that dependent places
provided by the seltau are eliminated.  Therefore, in "gerzda" the
dependent g2 place is removed from the tentative place structure given
in Example 3.5, leaving the place structure:

6.1)	z1 is the house dwelt in by dog z2=g1

Informally put, the reason this has happened --- and it happens a lot with
seltau places --- is that the third place was describing not the doghouse,
but the dog. The sentence

6.2)	la mon. rePOS. gerzda la spat.
	Mon Repos is a doghouse of Spot.

really means

6.3)	la mon. rePOS. zdani la spat. noi gerku
	Mon Repos is a house of Spot, who is a dog.

since that is the interpretation we have given "gerzda". But that in turn
means

6.4)	la mon. rePOS. zdani la spat noi ke'a gerku zo'e
	Mon Repos is a house of Spot, who is a dog
		of unspecified breed.

Specifically,

6.5)	la mon. rePOS. zdani la spat.
		noi ke'a gerku la sankt. berNARD.
	Mon Repos is a house of Spot,
		who is a dog of breed St. Bernard.

and in that case, it makes little sense to say

6.6)	la mon. rePOS. gerzda la spat.
		noi ke'a gerku la sankt. berNARD. ku'o
		la sankt. berNARD.
	Mon Repos is a doghouse of Spot,
		who is a dog of breed St. Bernard,
		of breed St. Bernard.

employing the over-ample place structure of Example 3.5.  The dog breed is
redundantly given both in the main selbri and in the relative clause, and
(intuitively speaking) is repeated in the wrong place, since the dog breed is
supplementary information about the dog, and not about the doghouse.

As a further example, take "cakcinki", the lujvo for "beetle", based on the
tanru "calku cinki", or "shell-insect".  The gismu place structures are:

6.7)	"calku": ca1 is a shell/husk around ca2 made of ca3
	"cinki": ci1 is an insect/arthropod of species ci2

This example illustrates a cross-dependency between a place of one gismu
and a place of the other.  The ca3 place is dependent on ci1, because all
insects (which fit into ci1) have shells made of chitin (which fits into ca3).
Furthermore, ca1 is dependent on ci1 as well, because each insect has only a
single shell.  And since ca2 (the thing with the shell) is equivalent to ci1
(the insect), the place structure is

6.8)	ci1=ca2 is a beetle of species ci2

with not a single place of "calku" surviving independently!

(Note that there is nothing in this explanation that tells us just why
"cakcinki" means "beetle" (member of Coleoptera), since all insects in
their adult forms have chitin shells of some sort.  The answer, which is
in no way predictable, is that the shell is a prominent, highly noticeable
feature of beetles in particular.)

What about the dependency of ci2 on ci1?  After all, no beetle belongs to
more than one species, so it would seem that the ci2 place of "cakcinki"
could be eliminated on the same reasoning that allowed us to eliminate the
g2 place of "gerzda" above.  However, it is a rule that dependent places
are not eliminated from a lujvo when they are derived from the tertau of
its veljvo.  This rule is imposed to keep the place structures of lujvo from
drifting too far from the tertau place structure; if a place is necessary
in the tertau, it's treated as necessary in the lujvo as well.

In general, the desire to remove places coming from the tertau is a sign
that the veljvo selected is simply wrong. Different place structures imply
different concepts, and the lujvo maker may be trying to shoehorn the
wrong concept into the place structure of his or her choosing. This is
obvious when someone tries to shoe-horn a "klama" tertau into a "litru"
or "cliva" concept, for example: these gismu differ in their number of
arguments, and suppressing places of "klama" in a lujvo doesn't make any
sense if the resulting modified place structure is that of "litru"
or "cliva".

Sometimes the dependency is between a single place of the tertau and
the whole event described by the seltau.  Such cases are discussed
further in Section 13.

Unfortunately, not all dependent places in the seltau can be safely
removed: some of them are necessary to interpreting the lujvo's meaning
in context.  It doesn't matter much to a doghouse what breed of dog inhabits
it, but it can make quite a lot of difference to the construction of a school
building what kind of school is in it!  Music schools need auditoriums and
recital rooms, elementary schools need playgrounds, and so on: therefore,
the place structure of "kuldi'u" (from "ckule dinju", and meaning "school
building") needs to be

6.9)	d1 is a building housing school c1
		teaching subject c3 to audience c4

even though c3 and c4 are plainly dependent on c1.  The other places of
"ckule", the location (c2) and operators (c5), don't seem to be necessary
to the concept "school building", and are dependent on c1 to boot, so
they are omitted.  Again, the need for case-by-case consideration of
place structures is demonstrated.


7.  Ordering lujvo places.

So far, we have concentrated on selecting the places to go into the place
structure of a lujvo.  However, this is only half the story.  In using selbri
in Lojban, it is important to remember the right order of the sumti.
With lujvo, the need to attend to the order of sumti becomes critical: the
set of places selected should be ordered in such a way that a reader
unfamiliar with the lujvo should be able to tell which place is which.

If we aim to make understandable lujvo, then, we should make the order
of places in the place structure follow some conventions. If this does not
occur, very real ambiguities can turn up. Take for example the lujvo
"jdaselsku", meaning "prayer". In the sentence

7.1)	di'e jdaselsku la dong.
	This-utterance is-a-prayer somehow-related-to-Dong.

we must be able to know if Dong is the person making the prayer, giving the
meaning

7.2)	This is a prayer by Dong

or is the entity being prayed to, resulting in

7.3)	This is a prayer to Dong

We could resolve such problems on a case-by-case basis for each lujvo (see
Section 14 for when this is actually necessary), but case-by-case
resolution for run-of-the-mill lujvo makes the task of learning lujvo place
structures unmanageable.  People need consistent patterns to make sense of
what they learn. Such patterns can be found across gismu place structures
(see Section 16), and are even more necessary in lujvo place structures.
Case-by-case consideration is still necessary; lujvo creation is a subtle art,
after all. But it is helpful to take advantage of any available regularities.

We use two different ordering rules: one for symmetrical lujvo and
one for asymmetrical ones.  A symmetrical lujvo like "balsoi" (from Section 5)
has the places of its tertau followed by whatever places of the seltau
survive the elimination process.  For "balsoi", the surviving places of
"banli" are b2 and b3, leading to the place structure:

7.4)	b1=s1 is a great soldier of army s2
		in property b2 by standard b3

just what appears in Example 5.1.  In fact, all place structures shown
until now have been in the correct order by the conventions of this section,
though the fact has been left tacit.

The motivation for this rule is the parallelism between the lujvo bridi-schema

7.5)	b1 bansoi s2 b2 b3
	b1 is-a-great-soldier of-army-s2
		in-property-b2 by-standard-b3

and the more or less equivalent bridi-schema

7.6)	b1 sonci s2 gi'e banli b2 b3
	b1 is-a-soldier of-army-s2 and
		is-great in-property-b2
		by-standard-b3

where "gi'e" is the Lojban word for "and" when placed between two partial
bridi, as explained in Chapter 14.

Asymmetrical lujvo like "gerzda", on the other hand, employ a different
rule.  The seltau places are inserted not at the end of the place
structure, but rather immediately after the tertau place which is
equivalent to the first place of the seltau.  Consider "dalmikce", meaning
"veterinarian": its veljvo is "danlu mikce", or "animal doctor".  The place
structures for those gismu are:

7.7)	"danlu": d1 is an animal of species d2
	"mikce": m1 is a doctor to patient m2 for ailment m3
		using treatment m4

and the lujvo place structure is:

7.8)	d1 is a doctor for animal m2=d1 of species d2
		for ailment m3 using treatment m4

Since the shared place is m2=d1, the animal patient, the remaining seltau
place d2 is inserted immediately after the shared place; then the remaining
tertau places form the last two places of the lujvo.


8.  lujvo with more than two parts.

The theory we have outlined so far is an account of lujvo with two parts.
But often lujvo are made containing more than two parts.  An example is
"bavlamdei", "tomorrow": it is composed of the rafsi for "future", "adjacent",
and "day".  How does the account we have given apply to lujvo like this?

The best way to approach such lujvo is to continue to classify them as based
on binary tanru, the only difference being that the seltau or the tertau
or both is itself a lujvo. So it is easiest to make sense of "bavlamdei"
as having two components: "bavla'i", "next", and "djedi". If we know
or invent the lujvo place structure for the components, we can
compose the new lujvo place structure in the usual way.

In this case, "bavla'i" is given the place structure

8.1)	b1=l1 is next after b2=l2

making it a symmetrical lujvo.  We combine this with "djedi", which has the
place structure:

8.2)	duration d1 is d2 days long (default 1)
		by standard d3

The d2 place should have its default value of 1 here, and doesn't provide us
with new information; so it is omitted.  Otherwise, "bavlamdei" is an
ordinary symmetrical lujvo with one additional anomaly:  While symmetrical
lujvo normally put any trailing tertau places before any seltau places,
the day standard is a much less important concept than the day the tomorrow
follows, in the definition of "bavlamdei".

(This is an example of how the guidelines presented for selecting and
ordering lujvo places are just that, not laws that must be rigidly adhered
to.  In this case, we choose to rank places in order of relative importance.)
The resulting place structure is:

8.3)	d1=b1=l1 is the tomorrow of/is the day
		after b2=l2 by standard d3

Here is another example of a multi-part lujvo: "cladakyxa'i", meaning
"long-sword", a specific type of medieval weapon.  The gismu place structures
are:

8.4)	"clani": c1 is long in direction c2 by standard c3
	"dakfu": d1 is a knife for cutting d2
		with blade made of d3
	"xarci": xa1 is a weapon for use against xa2
		by wielder xa3

Since "cladakyxa'i" is a symmetrical lujvo based on "cladakfu xarci", and
"cladakfu" is itself a symmetrical lujvo, we can do the necessary analyses
all at once.  Plainly c1 (the long thing), d1 (the knife), and xa1 (the
weapon) are all the same.  Likewise, the d2 place (the thing cut) is the
same as the xa2 place (the victim of the weapon), given that swords are
used to cut people.  Finally, the c2 place (direction of length) is always
along the sword blade, and so is dependent on c1=d1=xa1.  Dumping the
places of the remaining gismu in right-to-left order we get:

8.5)	xa1=d1=c1 is a long-sword for use against xa2=d2
		by wielder xa3, with a blade made of d3,
		long measured by standard c3.

(If the last place sounds unimportant to you, notice that what counts
legally as a "sword", rather than just a "knife", depends on the length
of the blade (the cutoff point varies in different jurisdictions).  This
fifth place of "cladakyxa'i" may not often be explicitly filled, but it is
still useful on occasion.)


9.  Eliding SE rafsi from seltau

It is common to form lujvo that omit the rafsi based on cmavo of selma'o SE,
as well as other cmavo rafsi.  Doing so makes lujvo construction for common
or useful constructions shorter.  Since it puts more strain on the listener
who has not heard the lujvo before, the shortness of the word should not
necessarily outweigh ease in understanding, especially if the lujvo refers
to a rare or unusual concept.

Consider as an example the lujvo "ti'ifla", meaning "bill, proposed law".
The gismu place structures are:

9.1)	"stidi": agent st1 suggests idea/action st2
		to audience st3
	"flalu": f1 is a law specifying f2 for community f3
		under conditions f4 by lawgiver f5

This lujvo does not fit any of our existing molds: it is the second
seltau place, st2, that is equivalent to one of the tertau places,
namely f1.  However, if we understand "ti'ifla" as an abbreviation for
the lujvo "selti'ifla", then we get the first places of seltau and
tertau lined up.  The place structure of "selti'i" is:
	
9.2)	"selti'i": idea/action se1 is suggested
		by agent se2 to audience se3

Here we can see that se1 (what is suggested) is equivalent to f1 (the law),
and we get a normal symmetrical lujvo.  The final place structure is:

9.3)	f1=se1 is a bill specifying f2 for community f3
		under conditions f4 by suggester se2
		to audience/lawgivers f5=se3

or, relabeling the places,

9.4)	f1=st2 is a bill specifying f2 for community f3
		under conditions f4 by suggester st1
		to audience/lawgivers f5=st3

where the last place (st3) is probably some sort of legislature.

Abbreviated lujvo like "ti'ifla" are more intuitive (for the lujvo-maker)
than their more explicit counterparts like "selti'ifla" (as well as
shorter).  They don't require the coiner to sit down and work out the precise
relation between the seltau and the tertau: he or she can just rattle off a
rafsi pair.  But should the lujvo get to the stage where a place structure
needs to be worked out, then the precise relation does need to be specified.
And in that case, such abbreviated lujvo form a trap in lujvo place ordering,
since they obscure the most straightforward relation between the seltau and
tertau.  To give our lujvo-making guidelines as wide an application as
possible, and to encourage analyzing the seltau-tertau relation in lujvo,
lujvo like "ti'ifla" are given the place structure they would have with the
appropriate SE added to the seltau.

Note that, with these lujvo, an interpretation requiring SE insertion
is safe only if the alternatives are either implausible or unlikely to
be needed as a lujvo.  This may not always be the case, and Lojbanists
should be aware of the risk of ambiguity.


10.  Eliding SE rafsi from tertau

Eliding SE rafsi from tertau gets us into much more trouble.  To understand
why, recall that lujvo, following their veljvo, describe some type of whatever
their tertau describe.  Thus, "posydji" describes a type of "djica", "gerzda"
describes a type of "zdani", and so on. What is certain is that "gerzda" does
not describe a "se zdani" --- it is not a word that could be used to describe
a dog, say.

Now consider how we would translate the word  "blue-eyed".  Let's tentatively
translate this word as "blakanla" (from "blanu kanla", meaning "blue eye").
But immediately we are in trouble: we cannot say

10.1)	la djak. cu blakanla
	Jack is-a-blue-eye

because Jack is not an eye, "kanla", but someone with an eye, "se kanla".
At best we can say

10.2)	la djak. cu se blakanla
	Jack is-the-bearer-of-blue-eyes

But look now at the place structure of "blakanla": it is a symmetrical lujvo,
so the place structure is:

10.3)	xe1=s1 is a blue eye of xe2=s2

We end up being most interested in talking about the second place, not the
first (we talk much more of people than of their eyes), so "se" would 
almost always be required.

What is happening here is that we are translating the tertau wrongly,
under the influence of English. The English suffix "-eyed" does not mean
"eye", but someone with an eye, which is "selkanla".

Because we've got the wrong tertau (eliding a "se" that really should be
there), any attempt to accommodate the resulting lujvo into our guidelines
for place structure is fitting a square peg in a round hole. Since they can
be so misleading, lujvo with SE rafsi elided from the tertau should be
avoided in favor of their more explicit counterparts: in this case, 
"xekselkanla".


11.  Eliding KE and KEhE rafsi from lujvo

People constructing lujvo usually want them to be as short as possible.  To
that end, they will discard any cmavo they regard as niceties.  The first
such cmavo to get thrown out are usually "ke" and "ke'e", the cmavo used
to structure and group tanru.  We can usually get away with this, because
the interpretation of the tertau with "ke" and "ke'e" missing is less
plausible than that with the cmavo inserted, or because the distinction
isn't really important.

For example, in "bakrecpa'o", meaning "beefsteak", the veljvo is

11.1)	[ke] bakni rectu [ke'e] panlo
	( bovine meat ) slice

because of the usual Lojban left-grouping rule. But there doesn't seem to be
much difference between that veljvo and

11.2)	bakni ke rectu panlo [ke'e]
	bovine ( meat slice )

On the other hand, the lujvo "zernerkla", meaning "to sneak in", almost
certainly was formed from the veljvo

11.3)	zekri ke nenri klama [ke'e]
	crime ( inside go )
	to go within, criminally

because the alternative,

11.4)	[ke] zekri nenri [ke'e] klama
	(crime inside) go

doesn't make much sense. (To go to the inside of a crime? To go into a place
where it is criminal to be inside --- an interpretation almost identical
with Example 11.3 anyway?)

There are cases, however, where omitting a KE or KEhE rafsi can produce
another lujvo, equally useful.  For example, "xaskemcakcurnu" means
"oceanic shellfish", and has the veljvo

11.5)	xamsi ke calku curnu
	ocean type-of (shell worm)

("worm" in Lojban refers to any invertebrate), but "xascakcurnu" has
the veljvo

11.6)	[ke] xamsi calku [ke'e] curnu
	(ocean shell) type-of worm

and refers to the parasitic worms that infest clamshells.

Such misinterpretation is more likely than not in a lujvo starting
with "sel-" (from "se"),  "nal-" (from "na'e") or "tol-" (from "to'e"):
the scope of the rafsi will likeliest be presumed to be as narrow as
possible, since all of these cmavo bind only to the following selbri
or "ke...ke'e" group.  For that reason, if we want to modify an entire lujvo
by putting "se", "na'e" or "to'e" before it, it's better to leave the result
as two words, or else to insert "ke", than to just stick the SE or NAhE
rafsi on.

It is all right to replace the phrase "se klama" with "selklama", and
the places of "selklama" are exactly those of "se klama".  But consider the
related lujvo "dzukla", meaning "to walk to somewhere".  It is a symmmetrical
lujvo, derived from the veljvo "cadzu klama" as follows:

11.7)	"cadzu": c1 walks on surface c2 using limbs c3
	"klama": k1 goes to k2 from k3 via route k4 using k5
	"dzukla": c1=k1 walks to k2 from k3 via route k4
		using limbs k5=c3 on surface c2

We can swap the k1 and k2 places using "se dzukla", but we cannot
directly make "se dzukla" into "seldzukla", which would represent the
veljvo "selcadzu klama" and mean something like "to go to a walking surface".
Instead, we would need "selkemdzukla", with an explicit rafsi for "ke".
Similarly, "nalbrablo" (from "na'e barda bloti") means "non-big boat",
whereas "na'e brablo" means "other than a big boat".

If the lujvo we want to modify with SE has a seltau already
starting with a SE rafsi, we can take a shortcut. For instance, "gekmau"
means "happier than", while "selgekmau" means "making people happier
than, more enjoyable than, more of a 'se gleki' than".  If something is less
enjoyable than something else, we can say it is "se selgekmau".

But we can also say it is "selselgekmau".  Two "se" cmavo in a row
cancel each other ("se se gleki" means the same as just "gleki"), so there
would be no good reason to have "selsel" in a lujvo with that meaning.
Instead, we can feel free to interpret "selsel-" as "selkemsel-".  The rafsi
combinations "terter-", "velvel-" and "xelxel-" work in the same way.

Other SE combinations like "selter-", although they might conceivably mean
"se te", more than likely should be interpreted in the same way, namely
as "se ke te", since there is no need to re-order places in the way
that "se te" provides.  (See Chapter 9.)


12.  Abstract lujvo

The cmavo of NU can participate in the construction of lujvo of a particularly
simple and well-patterned kind.  Consider that old standard example, "klama":

12.1)	k1 comes/goes to k2 from k3 via route k4 by means k5.

The selbri "nu klama [kei]" has only an x1 place, the event-of-going, but
the full five places exist implicitly between "nu" and "kei", since a full
bridi with all sumti may be placed there.  In a lujvo, there is no such
room, and consequently the lujvo "nunkla" ("nun-" is the rafsi for "nu"),
needs to have six places:

12.2)	nu1 is the event of k1's coming/going to k2 from k3
		via route k4 by means k5.

Here the first place of "nunklama" is the first and only place of "nu", and
the other five places have been pushed down by one to occupy the second
through the sixth places.  Full information on "nu", as well as the other
abstractors mentioned in this section, is given in Chapter 11.

For those abstractors which have a second place as well, the standard
convention is to place this place after, rather than before, the
places of the brivla being abstracted.  The place structure of "nilkla",
the lujvo derived from "ni klama", is the imposing:

12.3)	ni1 is the amount of k1's coming/going to k2 from k3
		via route k4 by means k5, measured on scale ni2.

It is not uncommon for abstractors to participate in the making of more
complex lujvo as well.  For example, "nunsoidji", from the veljvo

12.4)	nu sonci kei djica
	event-of being-a-soldier desirer

has the place structure

12.5)	d1 desires the event of (s1 being a soldier of army s2)
		for purpose d3

where the d2 place has disappeared altogether, being replaced by the
places of the seltau.  As shown in Example 12.5, the ordering follows
this idea of replacement:  the seltau places are inserted at the point
where the omitted abstraction place exists in the tertau.

The lujvo "nunsoidji" is quite different from the ordinary asymmetric lujvo
"soidji", a "soldier desirer", whose place structure is just

12.6)	x1 desires (a soldier of army x2) for purpose x3

A "nunsoidji" might be someone who is about to enlist, whereas a "soidji"
might be a camp-follower.

One use of abstract lujvo is to eliminate the need for explicit "kei" in
tanru:  "nunkalri gasnu" means much the same as "nu kalri kei gasnu", but is
shorter.  In addition, many English words ending in "-hood" are represented
with "nun-" lujvo, and other words ending in "-ness" or "-dom" are
represented with "kam-" lujvo ("kam-" is the rafsi for "ka"); "kambla"
is "blueness".

Even though the cmavo of NU are long-scope in nature, governing the
whole following bridi, the NU rafsi should generally be used as short-scope
modifiers, like the SE and NAhE rafsi discussed in Section 9.

There is also a rafsi for the cmavo "jai", namely "jax", which allows
sentences like

12.7)	mi jai rinka le nu do morsi
	I am-associated-with causing the event-of your death.
	I cause your death.

explained in Chapter 11, to be rendered with lujvo:

12.8)	mi jaxri'a le nu do morsi
	I am-part-of-the-cause-of the event-of your dying.

In making a lujvo that contains "jax-" for a selbri that contains "jai", the
rule is to leave the "fai" place as a "fai" place of the lujvo; it does not
participate in the regular lujvo place structure.  (The use of "fai" is
also explained in Chapter 11.)


13.  Implicit-abstraction lujvo

Eliding NU rafsi involves the same restrictions as eliding SE rafsi,
plus additional ones.  In general, NU rafsi should not be elided
from the tertau, since that changes the kind of thing the lujvo is
talking about from an abstraction to a concrete sumti.  However, they may
be elided from the seltau if no reasonable ambiguity would result.

A major difference, however, between SE elision and NU elision is that the
former is a rather sparse process, providing a few convenient shortenings.
Eliding "nu", however, is extremely important in producing a class
of lujvo called "implicit-abstraction lujvo".

Let us make a detailed analysis of the lujvo "nunctikezgau", meaning "to
feed".  (If you think this lujvo is excessively longwinded, be patient.)
The veljvo of "nunctikezgau" is "nu citka kei gasnu".  The relevant place
structures are:

13.1)	"nu": x1 is an event
	"citka": c1 eats c2
	"gasnu": g1 does action g2 (event)

In accordance with the procedure for analyzing three-part lujvo given
in Section 8, we will first create an intermediate lujvo, "nuncti", whose
veljvo is "nu citka [kei]".  By the rules given in Section 12, "nuncti" has
the place structure

13.2)	n1 is the event of c1 eating c2

Now we can transform the veljvo of "nunctikezgau" into "nuncti gasnu".  The
g2 place (what is brought about by the actor g1) obviously denotes the
same thing as n1 (the event of eating). So we can eliminate g2 as redundant,
leaving us with a tentative place structure of

13.3)	g1 is the actor in the event n1=g2 of c1 eating c2

But it is also possible to omit the n1 place itself!  The n1 place describes
the event brought about; an event in Lojban is described as a bridi,
by a selbri and its sumti; the selbri is already known (it's the seltau),
and the sumti are also already known (they're in the lujvo place structure).
So n1 would not give us any information we didn't already know.
In fact, the n1=g2 place is dependent on c1 and c2 jointly --- it does not
depend on either c1 or c2 by itself.  Being dependent and derived from the
seltau, it is omissible.  So the final place structure of "nunctikezgau" is:

13.4)	g1 is the actor in the event of c1 eating c2

There is one further step that can be taken. As we have already seen with
"balsoi" in Section 5, the interpretation of lujvo is constrained by the
semantics of gismu and of their sumti places. Now, any asymmetrical lujvo
with "gasnu" as its tertau will involve an event abstraction either
implicitly or explicitly, since that is how the g2 place of "gasnu" is
defined.

Therefore, if we assume that "nu" is the type of abstraction one would
expect to be a "se gasnu", then the rafsi "nun" and "kez" in "nunctikezgau"
are only telling us what we would already have guessed --- that the seltau
of a "gasnu" lujvo is an event.  If we drop these rafsi out, and use instead
the shorter lujvo "ctigau", rejecting its symmetrical interpretation
("someone who both does and eats"; "an eating doer"), we can still deduce
that the seltau refers to an event.

(You can't "do an eater"/"gasnu lo citka", with the meaning of "do" as
"bring about an event"; so the seltau must refer to an event, "nu citka".
The English slang meanings of "do someone", namely "socialize with someone"
and "have sex with someone", are not relevant to "gasnu".)

So we can simply use "ctigau" with the same place structure as "nunctikezgau":

13.5)	agent g1 causes c1 to eat c2
	g1 feeds c2 to c1.

This particular kind of asymmetrical lujvo, in which the seltau serves as
the selbri of an abstraction which is a place of the tertau, will be called
an implicit-abstraction lujvo.

To give another example: the gismu "basti", whose place structure is

13.6)	b1 replaces b2 in circumstances b3

can form the lujvo "basygau", with the place structure:

13.7)	g1 (agent) replaces b1 with b2 in circumstances b3

where both "basti" and "basygau" are translated "replace" in English,
but represent different relations:  "basti" may be used with no mention
of any agent doing the replacing.

In addition, "gasnu"-based lujvo can be built from what we would consider
nouns or adjectives in English.  In Lojban, everything is a predicate, so
adjectives, nouns and verbs are all treated in the same way. This is
consistent with the use of similar causative affixes in other languages.
For example, the gismu "litki", meaning "liquid", with the place
structure

13.8)	l1 is a quantity of liquid of composition l2
		under conditions l3

can give "likygau", meaning "to liquefy":

13.9)	g1 (agent) causes l1 to be a quantity of liquid
		of composition l2 under conditions l3.

While "likygau" correctly represents "causes to be a liquid", a different
lujvo based on "galfi" (meaning "modify") may be more appropriate for
"causes to become a liquid". 

Many other Lojban gismu have places for event abstractions, and therefore
are good candidates for the tertau of an implicit-abstraction lujvo.
For example, lujvo based on "rinka", with its place structure

13.10)	event r1 causes event r2 to occur

are closely related to those based on "gasnu".  However, "rinka" is less
generally useful than "gasnu", because its r1 place is another event rather
than a person:  "lo rinka" is a cause, not a causer.  Thus the place
structure of "likyri'a", a lujvo analogous to "likygau", is

13.11)	event r1 causes l1 to be a quantity of liquid
		of composition l2 under conditions l3

and would be useful in translating sentences like "The heat of the sun
liquefied the block of ice."

Implicit-abstraction lujvo are a powerful means in the language of rendering
quite verbose bridi into succinct and manageable concepts, and increasing the
expressive power of the language.


14.  Anomalous lujvo

Some lujvo that have been coined and actually employed in Lojban writing
do not follow the guidelines expressed above, either because
the places that are equivalent in the seltau and the tertau are in an
unusual position, or because the seltau and tertau are related in a
complex way, or both.  An example of the first kind is "jdaselsku", meaning
"prayer", which was mentioned in Section 7.  The gismu places are:

14.1)	"lijda": l1 is a religion with believers l2
		and beliefs l3
	"cusku": c1 expresses text c2 to audience c3
		in medium c4

and "selsku", the tertau of "jdaselsku", has the place structure

14.2)	s1 is a text expressed by s2 to audience s3
		in medium s4
	
Now it is easy to see that the l2 and s2 places are equivalent:  the
believer in the religion (l2) is the one who expresses the prayer (s2).
This is not one of the cases for which a place ordering rule has been
given in Section 7 or Section 13; therefore, for lack of a better rule,
we put the tertau places first and the remaining seltau places after them,
leading to the place structure:

14.3)	s1 is a prayer expressed by s2=l2 to audience s3
		in medium s4 pertaining to religion l1

The l3 place (the beliefs of the religion) is dependent on the l1 place
(the religion) and so is omitted.

We could make this lujvo less messy by replacing it with "se seljdasku",
where "seljdasku" is a normal symmetrical lujvo with place structure:

14.4)	c1=l2 religiously expresses
		prayer c2 to audience c3
		in medium s4 pertaining to religion l1

which, according to the rule expressed in Section 9, can be further
expressed as "selseljdasku".  However, there is no need for the ugly
"selsel-" prefix just to get the rules right: "jdaselsku" is a reasonable,
if anomalous, lujvo.

However, there is a further problem with "jdaselsku", not resolvable by using
"seljdasku".  No veljvo involving just the two gismu "lijda" and "cusku"
can fully express the relationship implicit in prayer.  A prayer is not just
anything said by the adherents of a religion; nor is it even anything said
by them acting as adherents of that religion.  Rather, it is what they say
under the authority of that religion, or using the religion as a medium, or
following the rules associated with the religion, or something of the
kind.  So the veljvo is somewhat elliptical.

As a result, both "seljdasku" and "jdaselsku" belong to the second class
of anomalous lujvo: the veljvo doesn't really supply all that the lujvo
requires.

Another example of this kind of anomalous lujvo, drawn from the tanru lists
in Chapter 5, is "lange'u", meaning "sheepdog".  Clearly a sheepdog is not a
dog which is a sheep (the symmetrical interpretation is wrong), nor a dog of
the sheep breed (the asymmetrical interpretation is wrong).  Indeed, there is
simply no overlap in the places of "lanme" and "gerku" at all.  Rather, the
lujvo refers to a dog which controls sheep flocks, a "terlanme jitro gerku",
the lujvo from which is "terlantroge'u" with place structure:

14.5)	g1=j1 is a dog that controls sheep flock l3=j2
		made up of sheep l1 in activity j3 of dog breed g2

Note that this lujvo is symmetrical between "lantro" (sheep-controller) and
"gerku", but "lantro" is itself a asymmetrical lujvo.  The l2 place,
the breed of sheep, is removed as dependent on l3.  However, the
lujvo "lange'u" is both shorter than "terlantroge'u" and sufficiently
clear to warrant its use:  its place structure, however, should be
the same as that of the longer lujvo, for which "lange'u" can be understood
as an abbreviation.

Another example is "xancyminde", "to command by hand, to beckon".  The
component place structures are:

14.6)	"xance": x1 is the hand of x2
	"minde": x1 gives commands to x2 to cause x3 to happen

The relation between the seltau and tertau is close enough for there to be
an overlap: xa2 (the person with the hand) is the same as m1 (the one
who commands).  But interpreting "xancyminde" as a symmetrical lujvo with an
elided "sel-" in the seltau, as if from "se xance mindu", misses
the point: the real relation expressed by the lujvo is not just "one who
commands and has a hand", but "to command using the hand".  The concept of
"using" suggests in the gismu "pilno", with place structure

14.7)	x1 uses tool x2 for purpose x3

Some possible three-part veljvo are (depending on how strictly you want to
constrain the veljvo)

14.8)	[ke] xance pilno [ke'e] minde
	(hand user) type-of commander

14.9)	[ke] minde xance [ke'e] pilno
	(commander hand) type-of user

or even

14.10)	minde ke xance pilno [ke'e]
	commander type-of (hand user)

which lead to the three different lujvo  "xancypliminde", "mindyxancypli",
and "mindykemxancypli" respectively.

Does this make "xancyminde" wrong?  By no means.  But it does mean that
there is a latent component to the meaning of "xancyminde", the gismu "pilno",
which is not explicit in the veljvo.  And it also means that, for a place
structure derivation that actually makes sense, rather than being ad-hoc,
the Lojbanist should probably go through a derivation for "xancypliminde"
or one of the other possibilities that is analogous to the analysis of
"terlantroge'u" above, even if he or she decides to stick with a shorter,
more convenient form like "xancyminde".  Plus, of course, the possibilities
of elliptical lujvo increase their potential ambiguity enormously --- an
unavoidable fact which should be borne in mind.


15.  Comparatives and superlatives

English has the concepts of "comparative adjectives" and "superlative
adjectives" which can be formed from other adjectives, either by adding
the suffixes "-er" and "-est" or by using the words "more" and "most",
respectively.  The Lojbanic equivalents, which can be made from any
brivla, are lujvo with the tertau "zmadu", "mleca", "zenba", "jdika",
and "traji".  In order to make these lujvo regular and easy to make,
certain special guidelines are imposed.

We will begin with lujvo based on "zmadu" and "mleca", whose place structures
are:

15.1)	"zmadu": x1 is more than x2 in property x3
		in quantity x4
	"mleca": x1 is less than x2 in property x3
		in quantity x4

For example, the concept "young" is expressed by the gismu "citno",
with place structure

15.2)	"citno":  x1 is young

The comparative concept "younger" can be expressed by the lujvo
"citmau" (based on the veljvo "citno zmadu", meaning "young more-than").

15.3)	mi citmau do lo nanca be li xa
	I am-younger-than you by-years the-number six.
	I am six years younger than you.

The place structure for "citmau" is

15.4)	z1 is younger than z2 by amount z4

Similarly, in Lojban you can say:

15.5)	do citme'a mi lo nanca be li xa
	You are-less-young-than me by-years the-number six.
	You are six years less young than me.

In English, "more" comparatives are easier to make and use than "less"
comparatives, but in Lojban the two forms are equally easy.

Because of their much simpler place structure, lujvo ending in "-mau" and
"-me'a" are in fact used much more frequently than "zmadu" and "mleca"
themselves as selbri. It is highly unlikely for such lujvo to be construed
as anything other than implicit-abstraction lujvo. But there is another type o
f ambiguity relevant to these lujvo, and which has to do with what is being
compared.

For example, does "nelcymau" mean "X likes Y more than she does Z", or "X
likes Y more than Z does"? Does "klamau" mean: "X goes to Y more than to Z",
"X goes to Y more than Z does", "X goes to Y from Z more than from W", or
what?

We answer this concern by putting regularity above any considerations of
concept usefulness: by convention, the two things being compared always fit
into the first place of the seltau.  In that way, each of the different
possible interpretations can be expressed by SE-converting the seltau, and
making the required place the new x1 place.  As a result, we get the
following comparative lujvo place structures:

15.6)	"nelcymau": z1, more than z2, likes n2
		by amount z4
	"selnelcymau": z1, more than z2, is liked by n1
		in amount z4
	"klamau": z1, more than z2, goes to k2 from k3
		via k4 by means of k5
	"selklamau": z1, more than z2, is gone to by k1
		from k3 via k4 by means of k5
	"terklamau": z1, more than z2, is an origin point
		from destination k2 for k1's going via k4
		by means of k5

(See Chapter 11 for the way in which this problem is resolved when lujvo
aren't used.)

The ordering rule places the things being compared first, and the other
seltau places following.  Unfortunately the z4 place, which expresses by
how much one entity exceeds the other, is displaced into a lujvo place whose
number is different for each lujvo.  For example, while "nelcymau" has z4 as
its fourth place, "klamau" has it as its sixth place.  In any sentence
where a difficulty arises, this amount-place can be redundantly tagged
with "vemau" (for "zmadu") or "veme'a" (for "mleca") to help make the
speaker's intention clear.

It is important to realize that such lujvo do not presuppose their seltau.
Just as in English, saying someone is younger than someone else doesn't
imply that they're young in the first place: an octogenarian, after all, is
still younger than a nonagenarian. Rather, the 80-year-old has a greater
"ni citno" than the 90-year-old.

There are some comparative concepts which are in which the "se zmadu" is
difficult to specify.  Typically, these involve comparisons implicitly made
with a former state of affairs, where stating a z2 place explicitly would
be problematic.

In such cases, it is best not to use "zmadu" and leave the comparison hanging,
but to use instead the gismu "zenba", meaning "increase" (and "jdika", meaning
"decrease", in place of "mleca"). The gismu "zenba" was included in the
language precisely in order to capture those notions of increase which
"zmadu" can't quite cope with; in addition, we don't have to waste a place
in lujvo or tanru on something that we'd never fill in with a value anyway.
So we can translate "I'm stronger now" not as

15.7)	mi ca tsamau
	I now am-stronger.

which implies that I'm stronger than somebody else (the elided occupant
of the x2 place), but as 

15.8)	mi ca tsaze'a
	I increase in strength.

Finally, lujvo with a tertau of "traji" are used to build superlatives.
The place structure of "traji" is

15.9)	t1 is superlative in property x2, being
		the x3 extremum (largest by default) of set x4

Consider the gismu "xamgu", whose place structure is:

15.10)	xa1 is good for xa2 by standard xa3

The comparative form is "xagmau", corresponding to English "better",
with a place structure (by the rules given above) of

15.11)	z1 is better than z2 for xa2 by standard xa3
		in amount z4

We would expect the place structure of "xagrai", the superlative form, to
somehow mirror that, given that comparatives and superlatives are comparable
concepts, resulting in:

15.12)	xa1=t1 is the best of the set t4 for xa2
		by standard xa3.

The t2 place in "traji" for property is replaced by the property
abstraction formed by the seltau places, and the t3 place specifying the
extremum of "traji" (whether the most or the least, that is) is presumed by
default to be "the most".

But the set against which the t1 place of "traji" is compared is not the t2
place (which would make the place structure of "traji" fully parallel to
that of "zmadu"), but rather the t4 place. Nevertheless, by a special
exception to the rules of place ordering, the t4 place of "traji"-based lujvo
becomes the second place of the lujvo.  Some examples:

15.12)	la djudis. cu citrai lo'i lobypli
	Judy is the youngest of all Lojbanists.

15.13)	la ajnctain. cu balrai lo'i skegunka
	Einstein was the greatest of all scientists.


16.  Notes on gismu place structures

Unlike the place structures of lujvo, the place structures of gismu were
assigned in a far less systematic way through a detailed case-by-case
analysis and repeated reviews with associated changes.  (No further changes
are contemplated, however.)  Nevertheless, there are certain regularities,
both in the choice of places and in the ordering of places, which may be
helpful to the learner, and which are therefore discussed here.

The choice of places results from the varying outcome of four different
pressures: brevity, convenience, metaphysical necessity, and regularity.
The implications of each are roughly as follows:

	Brevity tends to remove places: the fewer places a
	gismu has, the easier it is to learn, and the less
	specific it is.  As mentioned in Section 4, a brivla
	with fewer place structures is less specific, and
	generality is a virtue in gismu, because they must
	thoroughly blanket all of semantic space.

	Convenience tends to increase the number of places:
	if a concept can be expressed as a place of some
	existing gismu, there is no need to make another gismu,
	a lujvo or a fu'ivla for it.

	Metaphysical necessity can either increase or decrease
	places: it is a pressure tending to provide the "right
	number" of places.  If something is part of the
	essential nature of a concept, then a place must be
	made for it; on the other hand, if instances of the
	concept need not have some property, then this
	pressure will tend to remove the place.

	Regularity is a pressure which can also either
	increase or decrease places.  If a gismu has a given
	place, then gismu which are semantically related to
	it are likely to have the place also.

Here are some examples of gismu place structures, with a discussion of
the pressures operating on them:

16.1)	"xekri":  x1 is black

Brevity was the most important goal here, reinforced by one interpretation
of metaphysical necessity.  There is no mention of color standards here,
as many people have pointed out; like all color gismu, "xekri" is explicitly
subjective.  Objective color standards can be brought in by an appropriate
BAI tag such as "ci'u" ("in system"; see Chapter 9) or by making a lujvo.

16.2)	"jbena": x1 is born to x2 at time x3 and location x4

The gismu "jbena" contains places for time and location, which few other
gismu have: normally, the time and place at which something is done is
supplied by a tense tag (see Chapter 10).  However, providing these places
makes "le te jbena" a simple term for "birthday" and "le ve jbena" for
"birthplace", so these places were provided despite their lack of
metaphysical necessity.

16.3)	"rinka": event x1 is the cause of event x2

The place structure of "rinka" does not have a place for the agent, the
one who causes, as a result of the pressure toward metaphysical necessity.
A cause-effect relationship does not have to include an agent: an event
(such as snow melting in the mountains) may cause another event (such as
the flooding of the Nile) without any human intervention or even knowledge.

Indeed, there is a general tendency to omit agent places from most gismu
except for a few such as "gasnu" and "zukte" which are then used as tertau
in order to restore the agent place when needed: see Section 13.

16.4)	"cinfo": x1 is a lion of species/breed x2

The x2 place of "cinfo" is provided as a result of the pressure toward
regularity.  All animal and plant gismu have such an x2 place; although
there is in fact only one species of lion, and breeds of lion, though
they exist, aren't all that important in talking about lions.  The
species/breed place must exist for such diversified species as dogs,
and for general terms like "cinki" (insect), and are provided for all
other animals and plants as a matter of regularity.

Less can be said about gismu place structure ordering, but some regularities
are apparent.  The places tend to appear in decreasing order of psychological
saliency or importance. There is an implication within the place structure
of "klama", for example, that "lo klama" (the one going) will be talked about
more often, and is thus more important, than "lo se klama" (the destination),
which is in turn more important than "lo xe klama" (the means of transport).

Some specific tendencies (not really rules) can also be observed.
For example, when there is an agent place, it tends to be
the x1 place.  Similarly, when a destination and an origin point are
mentioned, the destination is always placed just before the origin
point.  Places such as "under conditions" and "by standard", which are
often omitted, are moved to near the end of the place structure.