metanohi/site/writings/na.org

#+title: Na
#&summary
A new spoken and written language in development.
#&
#+license: wtfpl
#&toc

* Na

** Introduction

Na is a probalistic language. The more details you use when explaining an
event, the more probable it is that the recipient understands you. This is not
very different from typical languages ("a blue building" is more precise than
"a building"), but the new thing in Na is that this guesswork is everywhere.

** Alphabet

Optimally, Na should have its own alphabet. To aid in its spreading, a subset
of the Latin alphabet has been chosen instead:

: . A E I O B D F G K L N S T V

All of these letters are pronounced as "short sounds", i.e. "A" is not
pronounced as "AAY", but just "AE". This is the same for all wovels. There are
two types of correct syllables: those of the form <consonant><wovel> and those
of the form <wovel>: "LO" is legal, "LOO", "LLA", "LOI", "AV", and "AEV" are
not. Aside from letters, only the period (".")  character and space between
characters are used. No commas, no colons, no semicolons, nothing else. Also,
there is no difference between uppercase and lowercase letters. Sentences are
read left-to-right (numbers are written right-to-left; more about that later).

With only 4 wovels and 10 consonants pronouncations can be very varied and
still correct. According to the rules laid out above there are 40 different
single-syllable words, 40^2 = 1,600 two-syllable words, 40^3 = 64,000
three-syllable words, 40^4 = 2,560,000 four-syllable words, etc. At ten
syllables (an extreme case) there are 10,485,760,000,000,000 different
10-syllable words and a total number of 10,754,625,641,025,640
any-number-of-syllables-less-or-equal-to-10 words.


*** TODO IPA/SAMPA

The author of this document is not very familiar with IPA, SAMPA, or any other
phonetic alphabet.

** Goals

Na wants to be a useful language somewhat usable by people. Na has been
designed to be fairly global, but since the creator of Na is fluent only in
Germanic languages (Danish and English), and because the Latin alphabet is
used, Na is non-global. Nevertheless, it /wants/ to be global.

** Basics

Na has no nouns, no verbs, no adjectives, no adverbiums, no pronouns, etc. In
Na, the core of everything is a word more or less equivalent to the English
word "existence". This word is *BA*. All other words are based on this base
word (in object oriented programming language terminology, this is equivalent
to subclassing classes).

The groups at the first levels of heritance are:

+ Existence :: BA
  + Number :: BE
    + Real number :: BEBA
    + Integer :: BEBE
    + Fraction :: BEBI
    + Imaginary number :: BEBO
    + Complex number :: BEDA
  + Symbol :: BI
    + Letter :: BIBA  
  + Shape :: BO
  + Relation :: DA
  + Identity :: DE
  + Absolute combiner :: DO
  + Identifier :: FA
  + Relative combiner :: FE


TODO: Improve on groups.

These have been chosen not because of perfection, but because Na was designed
for human beings (who are not perfect). Nevertheless, these groups should
suffice for the time being.

*Note:* These groups are few on purpose. Many words fit into several
groups. Groups are only good for basic things; more complex meanings do not
have to be placed in any group.

Na uses a combination of numbers, identifiers and further subclassing to
achieve (some) usability. In Na, what is written as one sentence in
e.g. English, can often become more than one sentence. To make a statement in
Na, one splits one's sentence into Na sentences, the first ones consisting of
definitions and assignments (named variables), and the final one consisting of
everything put together in the right order. Often, the order does not matter,
and the purpose of the final sentence is really just to point out what
definitions you have chosen to use, and that you are done.

For example, to write "I see a table." in Na, you first extract the objects (in
Na, everything is an object) from the sentence. This gives us:

+ I, the subject
+ see, the verb
+ a table, the object

In Na, this is simplified further into:

+ I
+ to see
+ table, 1

The only two possible sense-making combinations using all the words would be to
say "I see a table" or "The table sees me". To stress that the first version is
wanted, we first combine "I" and "to see", which we then combine with a
definition of "table, 1". Finally, we throw the definitions together, so that
people will understand that the sentence is over.

To specify "I", we use one of the members from the identity (*DE*) group. "to
see" is also an abstraction, because the "I" can see in different ways (if it's
a camera, it doesn't use a human brain to see, for example). Other,
non-generalized terms are also available, but they are not part of the
abstraction group.

To create a variable to hold "table, 1", you combine the number 1 with a
generic table defined using the shape (*BO*) group.

A sentence where you combine words into a new word has this structure:

: <word 1> <word 2>[ <word 3>[...]] DO <new word>.

This is an absolute combination. It is also possible to make relative
combinations where word mixes are not saved in new words:

: <word 1> <word 2>[...] FE <word(s)>[ ...].

In these relative cases, the words before an FE mark are evaluated when FE is
reached. If FE is left out from such sentences, everything will be evaluated at
once. Programmers might find it helpful to think of FE as a stack resetter.

When you do not use the *DO* or *FE* combiner, you state something:

: <word 1>[ <word 2>[...]].

In real languages, this is the equivalent of actually expressing something ---
combines do not express anything, they merely aid in preparing for later
expressions.

** Example 1: I see a table

One can write "I see a table." in two ways: the absolute way with *DO* or the
relative way with *FE*.

Words for "I", "to see", and "table" have not yet been made, but we assume they
are BIGUHA, NULASE, and GAVOTI, respectively. The integer 1 is FA BEBE B.

*** Absolute way

+ Step 1 :: Combine "I" and "to see" into a sentence using combinations from
            different groups. Give it a name, e.g. VAVO.
+ Step 2 :: Combine the number 1 from the integer (*BEBE*) group with the table
            from the shape group. Give it a name, e.g. GALO.
+ Step 3 :: a) Write VAVO GALO, or b) combine VAVO and GALO into e.g. VALO
            (could be any word) and go to step four.
+ Step 4 :: Write VALO.

So:

: biguha nulase do vavo. gavoti fa bebe b do galo. vavo galo do valo. valo.

or, shorter:

: biguha nulase do vavo. gavoti fa bebe b do galo. vavo galo.

The advantage of the first variation is that VALO can be reused again and again
until people have forgotten what its temporary meaning is. For example, if you
are telling a story where the main character often sees a table, you could do
this:

: valo. valo. valo. valo. valo.

It is not necessary to define new combinations all the time. Na comes with a
small built-in set of combinations which is required learning for all Na
speakers. These built-in definitions should be used whenever possible,
eventually in subclassed or changed forms, to make it easier to write --- and
especially speak --- Na.

*** Relative way

The relative way is a bit simpler, shorter and easier to say:

: biguha nulase fa bebe b fe gavoti.

In fact, this relative sentence can be made into an absolute sentence:

: biguha nulase fa bebe b fe gavoti do valo. valo.

*** Notes

In speech the relative way is much more useful than the absolute one. Both are
useful, though.


** Identifiers

Some of the groups can be used as identifiers. When an identifier is used one
or more arguments are expected. The integer identifier requires one argument
(the number), the real number identifier requires two arguments (before and
after the point), and the letter identifier requires one argument. To use a
group as an identifier, prefix it with *FA*.

** Numbers

Numbers in Na are duodecimal, i.e. base 12. The alphabet is used as numbers. A
is used for zero, the consonants range from 1 to 10, and O is used for 11. When
specifying a number, it should be prefixed with *FA BEBA* (real numbers) or any
other name from a subclassed number identifier, and it should be written with
the least important bit first (LSB). To pronounce a number, speak the letters
of the number one by one.

Examples:

+ FA BEBE GB = 16
+ FA BEBE D = 2
+ FA BEBA D AT = 2.108

TODO: Improve this strange system


** Example 2: We transported the ship into another dimension

This sentence is a bit tricky. We start by splitting the sentence into
meaningful parts:

+ we transport (in the past)
+ <something specific> ship
+ into
+ another dimension

Again, base words have not been defined yet, so we assume that "we (generic)"
(we do not know if it is an exclusive or inclusive we) is GELO, "transport" is
LOLOTI, the past is BATATO, something specific is SESE, "ship" is MOLOTE,
"into" is NIKO, "another" is KOBO, and "dimension" is TILE.

It appears that it is actually quite easy to create a relative sentence now:

: batato gelo loloti fe sese molote niko fe fe tile kobo.

Notice the two FE in the end. Without the second FE, the sentence could also
mean "We transported another dimension into the ship.". It would not be wrong
to omit the FE, but it would create an ambigious sentence (which can be
useful).

Remember that the following sentence means excactly the same as the previous:

: loloti batato gelo fe niko molote sese fe fe kobo tile.

Word order is only important in the case of identifiers and their arguments.


** Example 3: 33+2.4i

33+2.4i is a complex number. The x+yi notation doesn't fit well in Na. Instead,
the *FA BEDA* identifier is used (33+2.4i = 33.0+2.4i):

: fa beda td a d g.

Again, this is not perfect.


** TODO Foreign words

Foreign words present a problem in Na.


** TODO Built-ins

~1000 words needed to begin with.


** Miscellaneous

Since there are only 16 characters (14 letters, the period, and space), one
character can be stored in only 4 bits, and 2 characters can be stored in one
byte.

Because of the infinitely many relatively different ways to say the same thing,
texts and speech can be varied a great deal. This could be quite useful in
e.g. poems.