Lesson 5
mumoi seltadni

Place Structures
tergismu

You may have noticed that many bridi end up looking much like English
sentences in structure, if not in words.  After all, doesn't "mi cu
klama" look a lot like "I come."?  "klama" seems in this case to
closely resemble an English verb, with "mi" as the 'subject'.

The first problem with this shows up in, for example, the color
words. "mi cu blanu" still conveys a sense of "mi" as a 'subject', but
English speakers find it hard to think of "blanu" as a verb if the
bridi is translated as "I am blue".

Even harder to look at in this manner is "mi cu pritu ta ti" - "I am
to the right of that, based-on-facing this."  But "ko cu pritu ta ti"
can be easily understood as similar to an imperative verb: "Be to the
right of that, based-on-facing this."  The English translation inserts
the helping verb "to be" to turn an 'adjective'-like structure into a
verb.  This use of a helping verb is called a 'copula' by linguists.
Lojban, like many other languages besides English, has no copula.
("cu" is not acting as a copula, by the way - it is actually more akin
to a right parenthesis mark or comma, marking the end of that which
comes before and separating it from the selbri.)

To understand bridi, it is necessary to accept that the selbri
specifies nothing more than the relationship between the sumti, and
that this relationship need not be classified as noun, verb, or
adjective.  Those classifications are possibly of some interest if you
want to compare Lojban word meanings (semantics) with English
meanings, but they are simply irrelevant to Lojban.


5.1
All Places Are Created Equal
ro tergismu cu dunli selfinti

In English, the equivalents of sumti are known as 'subject' and
'objects'.  There are two kinds of objects in English: a 'direct
object' and 'indirect objects'.  The 'indirect objects' are usually
attached as part of a 'prepositional phrase'.  (This terminology isn't
important if you aren't already familiar with it.)  Quite often, the
form or usage of a word changes depending on whether it is part of the
subject, object, or indirect object.

Lojban, on the other hand, does not really distinguish between the
various sumti.  We can arbitrarily label the x1 sumti as being a
'subject'.  A better term, however, is based on the fact that the
first sumti is the one we are most interested in - it is the 'topic'
of the sentence or bridi.  But the other sumti are of identical form,
and it is possible to rearrange the values of a bridi without changing
any of the forms or usages.  Thus:

mi cu klama la bastn. la .atlantas. ti la ford.

is a valid Lojban sentence meaning:

"I go to Boston from Atlanta via this-here using the Ford."

We could also say the unlikely:

la bastn. cu klama ti la ford. mi la .atlantas.

by changing no word forms and rearranging the sumti, the result meaning:

"Boston comes to this-here from the Ford via me using Atlanta(?!)"

or in colloquial English:

"Boston comes here on Atlanta from Ford through me."

If we heard this English version, we would be totally confused, even
though it closely correlates with the Lojban place structure.  English
does not cope with randomized place data.  In Lojban, we know that
someone is expressing relative nonsense, and we know what specifically
is nonsensical about it.


5.2
What Is The Place Structure Of ...?
ma tergismu

Many new Lojbanists are confused by place structures.  Each word has a
different structure; often two words that are similar in English have
different place structures.  How can you possibly memorize them?  Is
there no system for figuring them out?

The answer is yes, but also no.  Place structures are certainly not
assigned randomly.  The place structures of the gismu list that you
have looked at, or of the vocabulary words in these lessons, are
merely a result of judgement on the part of those who developed the
word list.  Each place structure represents a decision as to what
concept a specific word should represent.

Since a brivla is indicating a relationship among the places of the
structure, the place structure is in some ways the definition of the
brivla.  The example of "klama", and the related gismu: "muvdu",
"bevri", and "cliva" will be used to show this.

The place structure of "klama" is:

x1 klama to x2 from x3 via x4 by means x5

In assigning this place structure, we are defining "klama".  If there
isn't someone or something moving such that all 5 places are
specifiable, then "klama" is not the word, since it would be relating
any non-existent places to the rest.  A brivla with a different place
structure, or a different relationship between the places, represents
a different concept; it is a different word.

Thus "cliva" has been defined to have a different place structure from
"klama", and thus to represent a different concept.  The two brivla
differ only in that "cliva" omits the "to..." place:

x1 cliva from x2 via x3 using means x4

It was determined that many uses of the English words "come" and "go" are totally interchangeable.  
If you substitute one for the other in many sentences, it doesn't change the meaning of the sentence:

"I go home from the office."	"Go here."
"I come home from the office."	"Come here.

There may be some subtle emphasis in "go" on the departing, and in
"come" on the arriving, but both concepts clearly have an origin and a
destination, and the distinction between them is probably too subtle
to define in terms of brivla.  In English of course, we have helping
words that CAN make a difference between the two words.  Thus "go
away" equates to "leave" and cannot be substituted for "come" in the
above examples without changing the meaning drastically.  The concept
"leave" doesn't even imply that a destination exists; it indicates
only motion away from an origin.

muvdu offers a slightly different contrast from "klama".  In the place
structure:

x1 muvdu x2 to x3 from x4 via x5

it is evident that the same places exist as in "klama".  Is it the same concept?

"klama" and "muvdu" are indeed quite close.  Rearranging the order of
the place structure does not change the essence of the relationship
between the places.  There must be something else different.  It was
judged in the case of "muvdu" that the first place x1 was not quite
the same as x5 of "klama".  Can you think of some mode of transport
that we might use to go somewhere, that would not be said to 'move
us'?

To explain how the two places differ requires a foray into an area of
semantic (word meaning) analysis called 'case theory'.  We will delve
into case theory a bit more in later lessons, but for now the essence
is to note a 'means of transport', the x5 of "klama" is a tool or
machine that does the moving.  There is nothing of import about x5
other than that of enabling the 'coming'; there is no hint of volition
or purpose.  The x1 place of "muvdu", however is commonly seen to have
a more active role in the process of transport.  Case theory makes a
distinction between an 'actor' and a 'means', and this is carried
across into a difference in gismu.  To be more concrete, we might use
'by foot' to fill x5 of "klama".  ko'a would be most unlikely to say
that "My feet moved me to the store" because this statement implies
something about feet other than simply being a means to go somewhere.
Specifically, it implies that the feet are 'actors', rather than
'means'.

"bevri" has an almost identical place structure to "muvdu", but it
differs from both "muvdu" and "klama".  The x1 place of "bevri" is an
actor, just as in "muvdu"; "bevri", however, implies a different
relationship between x1 and x2.  If something is dragged across the
ground, we might say that it is 'moved', but it is not 'carried'.
Since there is a different relationship between the places of the two
concepts, a different bridi is needed to express each kind of
relationship.

Place structures are thus quite intricately bound into the meaning of
a brivla, or of any bridi.  It is therefore quite reasonable to
suspect that you need to know a brivla's place structure in order to
use it effectively in all circumstances.  But you do not need to
memorize them, any more than you need to memorize every dictionary
definition of every English word in order to speak English.


5.3
Color Words
skari valsi

Colors seem very straightforward.  At first thought, we think
"Everyone knows what 'red' is."  This is, unfortunately quite wrong.
Every language divides up the spectrum in its own way.  There are
languages with only a couple of color words.  There are languages like
English that assign dozens of names to fine gradations or color.
Lojban is more like English, but there are significant differences
from English as well.

Lojban's color word gismu are shown below.  Look at the words that are
given; look for those that are missing.  Lojban has no gismu for
"pink", for example.  What English-speakers describe as "pink", a
Lojban speaker describes as either a kind of "red", "gray", "white",
or "orange", or (as we'll learn in later lessons) a mixture of these.
Russian speakers face a similar problem with "blue".  They have two
words for different kinds of "blue".  They must decide how to
interpret the colors they learned in Russian in terms of the colors of
Lojban.

Since there are no native Lojban speakers, there are very loose
boundaries between the colors.  This isn't as nebulous or unworkable
as it sounds.  It is part of the fun of learning a brand new language
- you get to help decide what the words mean!  You also will note that
we inherently have the same problem in English - I'm sure everyone can
recall a time where they thought something was one color, and someone
else thought it a different color.  Not to mention the color-blind.

When you first learned English, you had to learn the colors by being
told: "This is 'blue'.  That is 'red'." and we learned to categorize
things.  We will do the same in Lojban, except the categorization will
be by a slowly evolving consensus.

To begin with, look at the following list of colors.  See what is
there, and picture in your own mind what each color means.  Picture a
color half-way between two colors on the list, and try to figure out
how much it has to change in order to call it one or the other.

	blabi	white	x1 is white; adj.
	blanu	blue	x1 is blue; adj.
	bunre	brown	x1 is brown; adj.
	crino	green	x1 is green; adj.
	grusi	gray	x1 is gray; adj.
	narju	orange	x1 is orange; adj.
	pelxu	yellow	x1 is yellow; adj.
	xekri	black	x1 is black; adj.
	xunre	red	x1 is red; adj.
	zirpu	purple	x1 is purple; adj.
	cicna	cyan	x1 is cyan/turquoise/greenish-blue; adj.
	nukni	magenta	x1 is magenta/fuchsia/purplish-red; adj.

The last two colors may not be familiar to you; they are better known
to artists and color physiologists than to most other people.  cicna
is the color midway between green and blue, and nukni is the color
midway between blue and red.

Here are a few more gismu related to colors:
	
	carmi	intense	x1 is intense/bright in property x2 to observer x3
	kandi	dim	x1 is dim/dull/less colored, of color x2
	
	manku	dark	x1 is dark; adj.
	
	skari	color	x1 is of color x2 as perceived by x3 under conditions x4

To make things a bit easier, note at the bottom of the list that there
are gismu corresponding to "bright" and "dim".  These two words, in
the case of colors, refer to the richness or intensity of the color.

There are other components that vary within color.  One of variations
is what we call in English "light" and "dark".  We have included
"dark" in our vocabulary list here to point out that:

IT DOESN'T MEAN WHAT IT DOES IN ENGLISH.

At least not in terms of colors.  All Lojban gismu are intended to
cover a single concept, though it may be narrow or wide.  "manku" (and
its counterpart "gusni", which is not included in this lesson) refer
to amounts of illumination.  A 'dark' color in Lojban is one that
isn't illuminated; it does not change the color.  So to obtain a "dark
white room" in Lojban, turn off the light switch.

The color variation associated with the English "dark" and "light" is
fully covered by "xekri" ("black") and "blabi" ("white").  In Lojban,
a "light blue" will be covered by the metaphor ("tanru") "blabi
blanu", or "white-blue", or possible its reverse "blue-white"; you can
decide where the boundary between those two color concepts lie.  We
will discuss in later lessons how to use tanru to express all kinds of
small variations in concepts like colors.

Now that you've seen the list of colors, note that their place
structure has only one place defined.  Does this mean that colors
relate only one concept?  (to what?)  No.  In each brivla there is
always a potential involvement of other concepts.  In the case of
colors, other possible components have been much debated over the
years of Lojban development.  It was decided to use the smallest
subset of places that relate to a color - specifically the x1 place of
an object evidencing that color.  Using mechanisms we'll discuss in
later lessons, you can explicitly bring in sumti components such as:

"perceived by ..."
"under conditions ..."
"more than ..."
"according to standard ...

You can also throw in references to background and illumination, which
can change the way color is perceived.  In short, you can cover the
entire range of scientific and artistic definitions of color.  But if
you don't want to specify them, you can leave them out.  These
potential sumti are still there as potential sumti, and a listener
(ko'a) can ask about them if ko'a feels it is important to ko'as
understanding of your statement, ko'a can ask.

You may have noticed already that a lot of Lojban works like this:

- Singular/Plural in KOhA
- Gender in KOhA
- Singular/Plural in Names
- Additional implicit sumti in bridi

There is all kinds of implicit information involved in statements,
information that does not need to be made explicit.  A Lojban speaker
(ko'e) always recognizes that ko'e is leaving out information about a
statement.

The final gismu in the list, "skari", exhibits one way to bring in
this additional information.  "skari" explicitly brings in an observer
and the conditions of observation (which could include illumination
and background among other things) and associates them with an x1
which is a color or colored object.

In English, all the color words are adjectives.  This isn't
necessarily true in Lojban.  Thus:

ti cu blanu

says that	"This/These is/are blue." (adjective), but also
	"This/These is/are blue object(s)." (noun) and
	"This/These is/are bluing." (verb!).

As a verb, "blanu" is the equivalent of an intransitive English verb
(usually abbreviated "v.i." in dictionaries) which has no 'object'; in
the example, "ti" isn't doing anything to anything else.  Lojban,
having no 'verbs' nor 'objects' per se, does not directly distinguish
between active, passive, or intransitive 'verbs'.  Thus, for the
scientifically minded, you can think of "ti cu blanu" as conveying the
sense of an object giving off photons of blue wavelength, and/or
absorbing photons of other colors.  This active perception of "blue"
is not easily communicated in English.  It is implicit in Lojban.


5.4
Assumptions About Place Structures
sruma fi lo tergismu

Sometimes the inclusion of one or more places may seem arbitrary.
This is because one particular concept or relationship was chosen for
the brivla out of several possible ones, and one of these sumti was
needed to convey this choice.  The individual colors are defined
without standards or observers, because they are most often used
without consciously referring to the differences among observers or
standards.  But color is inherently based on perception, and it was
felt that at least one gismu should implicitly carry an observer place
as part of the relationship.  Thus, the observer is part of "skari";
anything said about this Lojban concept must be said in reference to
an observer.

Lest the learning of place structures seem too confusing or arbitrary,
we will mention a few conventions and consistencies that you may have
noticed:

- among the individual colors, as mentioned, the places have been left
  as minimal;

- parts of the body implicitly have the body they are part of as one
  of their places;

- all motions have some or all of the following places, depending on
  the nature of the concept and whether or not the place is
  applicable; if a given place is present, it will occur in the order:
  'actor', 'thing moved', destination, origin, route, means; if the
  'thing moved' is the 'actor', the two place are combined into one.

When people talk, they make assumptions about the meanings of words.
When you speak in Lojban, you must do the same.  You have the great
advantage with Lojban of knowing that there is a single correct
answer.  There is only one place structure for "klama", and only one
meaning.

If a speaker does not know the correct place structure/meaning and
misuses a brivla, there are two possibilities.  The listener (ko'e)
may recognize the error, or ko'e may not.  If ko'e recognizes a
possible error, the two can discuss it and clarify the intent.
Communication occurs, in spite of the knowledge gap.  This will be the
normal case, because in most communication, ko'e checks what is heard
to make sure that it is plausible (this is called, therefore a
plausibility test).  If something seems implausible, ko'e suspects an
error.

In English, when something we hear fails a plausibility test, we tend
to try guessing what the speaker 'really meant'.  Miscommunication
seldom occurs because we guess pretty well.  But it is difficult in
English to speak about implausible things because this correction
mechanism steps in too often.

Lojban, without grammatical ambiguity, and with only one meaning per
word, tends to invoke the plausibility test a bit less than English
does.  However, the interchangeability of places tends to make
guessing correct interpretations harder.  As a result, Lojban was
designed to instead make it easier to ask the speaker about
uncertainties in the intended place structure.

Thus, in Lojban, if you were to hear about the nonsensical journey of
Boston in the example given a couple of pages ago, you would
immediately ask the speaker whether indeed you heard correctly and
interpreted the places correctly.

Of course, it is possible that the plausibility test will not rule out
miscommunication.  If you have forgotten that motion bridi always give
the destination before the origin, you might misinterpret:

mi cu klama la bastn. la .atlantas. ti la ford.

and think that I ended up at Atlanta instead of Boston.

Lojban enables better communication.  It does not guarantee it.  Yes,
you are better off knowing the defined place structures.  But you can
rely on slow assimilation of those place structures over time as you
use the language with others, for there is no teacher better than the
miscommunication that results from errors in semantic meaning.

This slow assimilation of place structures provides an added benefit.
It removes a potential flaw in the language.  After all, the defined
place structures are only the best judgement of a few people.  Actual
usage may determine that some of the places defined are not always
relevant to the concept.  Or perhaps, other places will be found
needed that are not specified.  Those who do not accurately know all
the place structures serve as a valuable continuing source of
reevaluation of the true essence of meaning as you attempt to
determine 'on the fly', what the places of a brivla 'have to be' in
order to have a sentence make sense.

This doesn't mean that you are free to change any place structure you
don't like.  First you need to try to figure out why the place
structure is the way it is defined.  This will generally take
comparison with several other related brivla, and will usually result
in your having a clearer idea of what concepts each of those brivla
are intended to cover.  Until you know most of the gismu, therefore,
you will be unlikely to identify any real problems.

When you speak the language reasonably well, however, and believe that
a place structure is not optimal, you will be able to propose changes
to a language standards committee that will review and resolve such
proposals.  The standards committee will also watch for problems that
arise among large numbers of users, and may adjust place structures
accordingly.  The change process will be slow, however, and will be
comparable to the 'linguistic drift' that causes changes in meanings
of words in every other language.  (Twenty years ago, the English word
'gay' was synonymous with 'happy' - it is now seldom used for that
meaning, because plausibility tests cannot distinguish between that
meaning and the totally unrelated interpretation involving
homosexuality.)

Place structures will also be closely watched for evidence of cultural
bias.  Since the current place structures were devised by native
English speakers, when Lojban is spoken by non-English speakers, it
may be found that certain places are not innately part of a brivla's
meaning.  Proposals for changes in place structures will similarly be
studied to avoid introduction of new English biases.


5.5
So What Do I Do?
mi gasnu ma

Make your best guess as to the places of a brivla when you do not know
them.  If communication occurs, your knowledge was adequate.  If not,
you will learn, or possibly improve upon, what is correct.  Use
plausibility tests, and ask the speaker when something seems
implausible.  You may be the one who is correct, and end up aiding the
speaker in using the language better in the future.

There are a few patterns and conventions that we will point out as we
go through the vocabulary, like the ones mentioned above.  We will
also try to point out specific comparisons between related brivla so
that their meanings are clearly differentiated.  In most cases,
however, you will learn to use a place structure by simply trying to
guess what types of information are relevant to the concept and in
what order of importance, and you will probably be close to correct.

In a later lesson, we will learn how to specifically request the
speaker to identify the intended place for a sumti.  Until then, you
can ask the speaker by using the phrase "{sumti in question} cu mo",
or "{sumti in question} ki'a", both of which imply some confusion as
to the sumti, thus prompting the speaker to clarify how the sumti
relates to the bridi.


5.6
Places That Aren't Expressed
nalsku tergismu

In most of the examples so far, you may have noticed that we haven't
filled in all the sumti.  We have, for example, given the short
statement: "mi cu klama".  Does this bridi have only one place?

The answer is no.  One major reason is that there would be no
difference then between "mi cu klama" and "mi cu cliva"; these two
differ only in the existence in "klama" of a destination sumti.  We
cannot allow two Lojban gismu to have the same meaning, or a lot of
the advantage to the language would be lost.

The more important reason is based on the converse of this.  "klama"
and "cliva" do have different meanings.  "mi cu klama" and "mi cu
cliva" therefore mean two different things.  "klama" relates the x1
sumti "mi" to four other sumti, whether specified or not.  "cliva"
relates the same sumti to three other sumti.  The relationship remains
different even when the sumti are unspecified; the bridi also remain
different.

When you leave sumti out, they are really still there.  The speaker
has simply left them unspecified.  In English, we would just put a
nondescript 'something' or "somewhere", as in:

mi cu klama thus translates as "I go somewhere from somewhere via some
route using some means."

Leaving out implicit information is called 'ellipsis' in English, or
sumti nalnunsku in Lojban.

Most speakers of English and other languages use sumti nalnunsku a
great deal.  We just don't notice it, and there is no clear cut way to
tell when the speaker has omitted data.  In Lojban, we can tell.
There is sumti nalnunsku:

ANY TIME THERE IS A bridi WITHOUT ALL PLACES FILLED IN.

If the speaker (ko'a) uses sumti nalnunsku to leave information out,
it is still implicitly there for the listener to ask about.  One way
to ask is to repeat the bridi using "ma" for each of the places you
want filled in (e.g., "do cu klama ma ma" - remember that each ma
might stand for something different); there are other ways as well
that we will learn in later lessons.  In general, if sumti nalnunsku
is used, either ko'a is indicating that the information is irrelevant,
that it is presumably obvious and goes without saying, or that ko'a
hasn't bothered to figure out what goes in the elliptical place.  (If
ko'a is not familiar with the brivla being used, the apparent sumti
nalnunsku could also mean that ko'a has forgotten that the elliptical
place is part of the bridi place structure.)


5.7
Places That Aren't In The Place Structure
lo tergismu na te gismu

Sometimes, when you attempt to figure out what the sumti places of a
brivla are likely to be (as opposed to memorizing them), you will come
up with places that aren't going to be there.  These places may seem
like they should be part of the definition, or they may be places that
you want to specify as part of a bridi.  Some of these are:

- at time ...
- at location ...
- according to frame of reference ...
- by standard ...
- according to observer ...
- more than ... (or less than ...)
- because of ...
- resulting in ...

There are many such possible additional sumti that could be included
in any brivla, and some of the above would seem to be a part of almost
any brivla.

In fact, these sumti places have been omitted from nearly all brivla
definitions, even though they could be said to be a part of nearly
all.  Why?  to make it easier to memorize the places that are specific
to individual brivla.  There are ways (again we're saving them for
later lessons) to add these sumti to a bridi when they are relevant to
your statement.  They are omitted in most other cases.

Thus, every brivla implicitly carries some potential incompleteness.
You should only expect to find the above sumti types included in a
place structure if they are integral to the basic concept.  Thus
"pritu" requires a frame of reference because it is truly impossible
to say something is on the right of another without indicating an
orientations standard.

These added sumti are not 'equal' to the other sumti.  Each must be
marked, and we don't assign numbers to them.  Because these added
sumti aren't implicitly there as part of the definition of the brivla,
we do not call their omission 'sumti nalnunsku'.


5.8
Standard Form of bridi Sentences
tadni ke bridi tarmi


We have written all examples of Lojban sentences with exactly one
sumti to the left of the selbri.  This is called the 'standard form'
of the bridi, and it is the simplest form of Lojban sentence.

Other forms are permitted, however.  You can put two, three, or all of
the sumti on the left side of the selbri.  This order is very common
among the world's languages, but occurs rarely in English, usually in
forms such as poetry.  Our example from above with the selbri last is:

mi la bastn. la .atlantas. ti la ford. cu klama

We also use other forms in English (usually in poetry) that fit
neither of these orders: for example, 'To the store go I.'

You might wish to express the sumti of a bridi other than in standard
order for the same poetic reasons you do in English.  You might also
wish to change the relative emphasis of the selbri as opposed to the
sumti by changing the order from the canonical order.  Simply by doing
something unusual, you call attention to that which is different.  A
third reason for using a different order is to communicate with a
Lojbanist whose native tongue uses a different order.

This variation is possible because the Lojban grammar truly treats all
of the numbered sumti as equal.  There is no grammatical difference
between any of these possible orderings in Lojban.  The assignment of
numbering to the various sumti has been made a question of
interpretation or semantics.  This is useful for a lot of reasons.
Most important, it means that a change to the place structure of a
brivla is not a grammatical change.  The wall between structure
(grammar) and interpretation (meaning, or semantics) is retained in
Lojban wherever it is possible.  It is this trait that allows Lojban
to have an unambiguous grammar, while still allowing for the ambiguity
that cannot be eliminated from the semantics of a human language.

In general, as long as you keep the places in order x1, x2, x3 ..., it
doesn't matter where in the sequence you put the selbri, especially
when you clearly mark that it is coming with "cu".  The sumti will be
numbered from the left as x1, x2, x3 ..., etc.  If you want to change
the order, though, things get more complicated.  We will deal with one
way to change the sumti place ordering in the section on 'conversion'
below.  In Lesson ***, we will learn a more general approach to the
problem of place ordering.

There is one exception, based on the fact that the standard form has
exactly one numbered sumti to the left of the bridi.  If all of the
sumti are to the RIGHT of the selbri, then - following the convention
just mentioned - the first would still be x1.  This could lead to a
sentence like

klama mi la bastn. la .atlantas. ti la ford.

This is not ungrammatical.  However, it doesn't mean the same thing as
the previous example, which is what it would appear to mean under the
usual sumti numbering convention.

When ALL of the sumti appear to the right of the selbri (or when there
are no sumti at all), then the canonical ordering rule takes
precedence over the numbering rule, and we assume that the x1 place
was elliptically omitted, and the first place to the right is numbered
x2.  In fact, nalnunsku of the first sumti is so common and useful in
Lojban that the freedom to move sumti was deemed less important than
the capability to easily omit the x1 sumti.  (It turns out that you
still can put all of the sumti on the right side using the rearranging
techniques of Lesson 22.)

We will discuss this special aspect of sumti nalnunsku, and some other
practical considerations in using bridi place structures, in Lesson 7.