Message-Id: <199509281929.PAA19135@locke.ccil.org>
Date:         Thu, 28 Sep 1995 11:23:13 -0700
Reply-To: "Peter L. Schuerman" <plschuerman@UCDAVIS.EDU>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: "Peter L. Schuerman" <plschuerman@UCDAVIS.EDU>
Subject:      Criticisms and Parts of Speech
To: John Cowan <cowan@LOCKE.CCIL.ORG>
Status: OR

(beginning lojban students may wish to skip towards the end of this
message for an article on parts of speech that may be helpful.)

Hello all!

I've been reading the texts that are available on the Lojban WWW page for
some time now, and I have a comment to make about them.  I feel that they
are far too complicated in relation to the actual concepts they try to
convey.

For instance, the basic concept of sentence structure is *incredibly*
simple... sumti [one or more sumti] [cu] selbri [one or more sumti]
where [] enclose optional elements.  Yet this formula is not presented in
early lesson texts in this way, if ever.

I'd like to take the time here to give some criticisms of the way the
material has been presented.  But, before I do that, I just want to say
that I have been working on personal artificial language projects for
around 15 years and I know how hard it can be to undertake this sort of
work.  I am impressed, to say the least, by the approach that has been
taken with Lojban... everything from the word-generation algorithm to the
tense system to the whole "selbri concept" is ingenious.  Also, the amount
of time and effort that has gone into the project to get it this far must
have been enormous, and I commend all of those involved.

The reason I feel so strongly about commenting is because the underlying
material is so good... it should be conveyed in the most practical and
clear way possible. That being said, the biggest problems I had with the
texts were as follows.

1) There was too much emphasis on how English and Lojban differ, rather
than on their similarities.  Focussing on their similarities would help a
student bridge the gap between the two languages.  In the future, texts
for students whose first language is not English could be adapted
similarly, showing the similarities between their languages and Lojban.

2) Over-elaboration of design features is also a problem.  In the initial
chapters of a Lojban textbook, it would be better to explain just enough
of several topics to make a student be able to speak and write, and leave
the elaborations to later chapters.  Lojban is really wonderful, but we
don't need to know *everything* about every topic, right away.  For
example, *all* of the simple sumti (pronouns, description, name, and
perhaps the le nu ___ [kei] [ku] construction) should be introduced right
away, and complex ones much later.  The first chapter (lesson01) spent far
too much time "selling" the language.  The text should decide whether it
is teaching or selling the language, and should do just one of those
(hopefully the former!).

3) The concept of place-structure isn't discussed until lesson 5!!!  This
should be introduced MUCH earlier!  It's fundamental.  The text says "Many
new Lojbanists are confused by place structures."  I think this is because
they had to wait so long before anyone explained it;  frustrated, they've
been puzzling over the gismu list, forming incorrect ideas.  Also, the
x1, x2, etc. terminology is never explicitly explained... why not mention
that these are, in fact, VARIABLES that represent ARGUMENTS that are
being RELATED by the SELBRI??  To be perfectly frank, the first thing I
thought when I saw x1, x2, was that these were referring to repetitions
(x1 = one time, x2 = two times) and that didn't make a whole lot of
sense. :)  I've seen other people ask about this, and I know I'm not the
only one confused by this.  For example:

sakci /SACK-shee/  suck
     x1 sucks/is suction/vacuum/relative low pressure of
     fluid/gas x2 relative to high pressure x3
     [also suck object/fluid (= sakcpu or sakmu'u)]

Reading this makes my head spin.  Why not say that:

 x1 is "the thing applies suction" (e.g. suction cup, vacuum, mouth, etc.)
 x2 is "the thing that is sucked"  (why limit this to fluid or gas?)
 x3 is "the thing that comes out of/away from x2 as a result"
(I mean, you don't give a reference to atmospheric pressure when you say
you suck soda through a straw, or even under most laboratory conditions!)

I think a much better way to format gismu information would be like this:

batci /BAT-shee/  bite (or pinch... squeeze sharply between two
                  surfaces)

     x1 = biter or pincher
     x2 = bitee or pinchee  :)  (actually, "that which is bitten/pinched")
     x3 = location of bite or pinch on x2
     x4 = what x1 used to bite or pinch x2

The variable-substituted quasi-sentences in the gismu definitions just
don't convey the information clearly, and in any event emphasize a
reliance on English sentence structure.

4) The tone of the material swings from over-explained to cryptic.  There
is a long section in lesson 5 that belabors the point that English has
"sentence subjects" and Lojban has "sentence topics."  This idea doesn't
merit the amount of text, and seems to assume that the reader is quite
dull.  On the other hand, look at the following, keeping in mind that the
words and concepts "doi" "selma'o" "COI" "ju'i" "compound cmavo (such as
ju'idoi)" and "vocative reference" have not been defined yet in the text.
Wouldn't you find this a little bit cryptic?

>e. cmene may optionally be preceded by the cmavo "doi" (English "O,
>   ..." indicating direct address or vocative reference; i.e., when
>   talking to that which is named.  They may also be preceded by words
>   in selma'o COI, which also supports direct address.  However, cmene
>   preceded by selma'o COI require a pause or doi between them to
>   ensure resolvability under all conditions.  In informal direct
>   address outside of sentence constructs, cmene also may appear
>   without any preceding cmavo (e.g. calling out "doi djan." or just
>   "djan."). Thus "ju'idoi djan.", or "ju'i. djan." are acceptable
>   ways to get John's attention (ju'i is in selma'o COI).

I know I did.

In summary, I feel that the information should be presented in this way:
give the bare minimum essentials for communication initially (e.g. parts
of speech, simple sentence structure, attitude, event contours, simple
temporal tense and simple spatial tense) and then go more in depth later
(e.g. the first treatment of temporal tense should just include PU and ZI,
with intervals, etc. described in a later chapter).

I've been restructuring the information in this way for my own personal
use and for the use of friends who are also interested in the text but
are frustrated with the format as it stands.  Here's an example; it's a
rough draft, but it should nicely illustrate what I'm saying.

-----------------------------------------------------------------------

Chapter 2.  Parts of Speech in Lojban

by Peter Schuerman (plschuerman@ucdavis.edu, phandaal@well.com)

This document introduces and defines the three main components of
Lojban: bridi, selbri, and sumti.  It is abstracted from THE
BASIC COMPONENTS (SUMTI AND SELBRI) by veion@xiron.pc.helsinki.fi

2.1  The expression of relationship: bridi

Every Lojban sentence expresses the relationship between various
things.  In Lojban, an expression of relationship is called a
bridi.  The concept of bridi is very much like the concept of
"sentence" in English.

(2.1a) I walk to the store.

In this example, "I" and "store" are the "things" in the sentence
while "walk" describes the relationship between these two things.

A simple bridi includes "things" and a description of how those
things are related; thus, example 2.1a is an English example of a
bridi.  The Lojban term for these "things" in a bridi is "sumti"
and the Lojban term for something that describes how these things
are related is "selbri".

To put it another way, a selbri is a part of speech that relates
a number of things (sumti) to each other.

2.2  Types of selbri

Selbri are parts of speech that describe the relationship of
"things" (sumti) in a sentence (bridi).  Selbri can be of
different sorts, but the simplest (and the ones we will deal with
here) are ones which are composed of a single word.  A word which
may by itself express a selbri relation is called a brivla.
There are other sorts of selbri which are more complex, which
will not be discussed until later.

selbri are sometimes set off from the rest of the sentence by
putting a "small word" (in Lojban, "cmavo") in front of the
selbri.  This is used when it may be unclear where the sumti end
and the selbri begins.  The cmavo that is used to do this is "cu"
(pronounced "shoo").

A brivla, then, is a single word that is used as a selbri.  There
are three types of brivla:  gismu (root words), lujvo (compounds)
and le'avla (borrowings from other languages).  We will focus on
gismu here.

2.3  gismu as brivla

gismu are conceptual "root words" which mean different things
depending on how they are used.  For example, they can be used as
brivla.

klama is an example of a gismu, which represents the concept of
"coming" or "going".  When it used as a selbri, it can connect
several "things" with this concept.

What sorts of things are related to the concept of coming?  There
is a traveller, there is a destination, there is a point of
origin, there is a route and there is a means of travel.  The
gismu "klama" can be used to relate these things when it is used
as a selbri.

Since a selbri draws relationships between various sumti, each of the
things described above (the destination, the point of origin, the
traveller, etc.) must be included as sumti in the bridi (sentence).  The
sumti can be said to be the "arguments" of the selbri; the arguments are
the "input" that the selbri takes to give an "output" that results in a
sentence that relates them.

The way that a selbri takes sumti is dependent on the ordering of
the sumti.  The selbri "klama" designates the first sumti it
takes as "the traveller" and the second sumti as "the
destination".  The first sumti can be called "x1" and the second
"x2" and so on.  So, the definition of klama can be written:

   klama  -  come or go
          x1 comes/goes to destination x2 from origin x3 via
          route x4 using means/vehicle x5

or, alternately:

klama - come or go
        x1 = traveller; come-er or go-er
        x2 = destination of travel
        x3 = origin of travel
        x4 = route taken by x1 between x2 and x3
        x5 = means of travel (feet, vehicle)

The ordering of how a selbri takes the sumti in simple sentences
is simply a matter of how the sumti are ordered in the bridi.
Starting at the beginning of the bridi, the first sumti is x1,
the second x2, and so on.  If English used the structure of
Lojban, this is how sentences could be written:

(2.3a)  I   come   home workplace highway car
        x1  selbri x2   x3        x4      x5

Which would mean:

I come home from the workplace by driving my car along the
highway.

For simple sentences in Lojban, the selbri can be anywhere except
at the very beginning of the sentence (e.g. x1 x2 x3 selbri x4 x5
would be ok, as would x1 x2 x3 x4 x5 selbri, etc.).  This will be
discussed at greater length in the section on sentence structure.

2.4  Types of sumti

There are three types of sumti which will be discussed here.
Pronoun-type sumti, name sumti and description sumti.

2.4.1  "PRONOUN" SUMTI:

      mi            I/we
      do            you
      ko            imperative you
      ti            this/these (this here/these here)
      ta            that/those (that there/those there)
      tu            that/those-yonder (that there-yonder/those
                    there-yonder)
      zo'e          something unspecified (it's either obvious or
                    unimportant)

So we could use klama to make the following sentence:

mi cu klama ti tu zo'e zo'e

mi cu      klama ti           tu      zo'e          zo'e
I  (begin  come  this (here)  yonder  unspecified   unspecified
   selbri)       destination  origin  route         means

which means:

I come here from far away, by unspecified route and means.

zo'e at the end of a bridi can simply be left off, and assumed
they are there; the following sentence is equivalent to the one
before:

mi cu klama ti tu

and remember, cu is only necessary when it might be unclear as to
where the sumti leave off and the selbri begins.  So the sentence
can be further reduced to:

mi klama ti tu

And in simple sentences, the placement of the selbri is not
critical as long as there is at least one sumti before it, so the
following sentences are also equivalent:

mi ti klama tu
mi ti tu klama

2.4.2  NAME SUMTI

sumti can be formed from names by using the cmavo (small word)
"la" in front of the name.  Names in Lojban are phonetic
representations of a name imported from another language (the
details of this process of importing names will not be discussed
here).

Examples of name sumti:

la djan.        (John)
la pitr.        (Peter)
la .abigeil.    (Abigail)
la kalifornias. (California)

So we can use names in sentences like so:

la djan. klama ti tu

or

la pitr. la kalifornias. la kolorados. klama

2.4.3  DESCRIPTION SUMTI (gismu as sumti)

These are formed by taking a selbri and putting "le" in front of
it and "ku" after it.  The "ku" can be dropped if it doesn't
create any confusion (just like cu).  The meaning of a
description sumti is identical to the x1 place of the selbri.
For instance, the x1 place of klama is "the traveller" while the
x1 place of "vecnu" is "the seller".  Similarly, the x1 place of
"karce" is "the car."

le klama ku    (the traveller)
le vecnu ku    (the seller)
le karce ku    (the car or truck)

So we can make a sentence like:

le vecnu ku klama ti la kalifornias. zo'e le karce
The seller comes here from California by some route by car.

2.5  Simple sentence structure

In the examples we've gone over, you should have started to get a
feel for how simple sentences are set up in Lojban.  Basically,
simple sentences are composed of selbri and sumti.  The number of
sumti must not exceed the number of sumti which the selbri
"takes"... e.g. klama takes up to 5 sumti (x1 - x5).  At least
one sumti must come before the selbri, and the rest may be placed
anywhere (i.e. on either side of the selbri) as long as their
order in the sentence remains meaningful (e.g. the 3rd sumti must
correspond to the x3 position of the selbri, etc.).  "cu" is used
to set off a selbri from the sumti before it.  There are two
elements which we have not talked about so far.  The first is
"vau".  vau means "end of sentence" and, like cu and ku, may be
dropped if doing so does not create any confusion.  In most
cases, vau can be dropped without a problem.  The other element
is .i (pause,ee).  This is used at the beginning of sentences and
indicates that the sentence is a continuation of the theme or
topic of the previous sentence.  When a bridi (sentence) is not a
continuation of the same topic, .i is not used.

The simple sentence formula can be written:

[.i] sumti [one or more sumti] [cu] selbri [one or more sumti] [vau]

where brackets enclose optional elements, and where the total
number of sumti does not exceed the number of sumti which the
selbri allows.

Thus, the only critical elements of a bridi are

sumti selbri

There are other sentence structures, which will be covered in
another section.

2.6  A note on tense and number

None of the sentences shown in this lesson specify number or
tense.  The pronoun "mi" for instance, means "I" or "we".  The
gismu "klama" does not specify whether the travel occurred in the
past, present or future.  For example, "mi klama la kalifornias."
can mean many things, including "I came to California", "We will
go to California" or "I am going to California".  Lojban does
allow one to specify these things, but if a speaker does not do
so, the audience must either guess as to what number or tense
would make the most sense, or simply ask for clarification.

------------------------------------------------------------------
Peter Schuerman  (plschuerman@ucdavis.edu)