my proposals for a philosophical language

The message below is a proposal regarding philosophical languages that I
made in a conversation with Andrew Nowicki, the proponent of Ygyde who we
have recently become familiar with.

In it, I examine some of the problems Ygyde currently faces, much of which
it shares in common with other philosophical languages.  I then propose the
sorts of changes that seem to solve these problems to some degree.  As many
of you know, I'm strictly an artlanger, and relish features that make
languages opaque in their workings.  (Though not to the point of absolute
maggellity.)

First, I propose some sounds that are common to English, Spanish, Arabic,
and Cantonese (more or less), as these languages encompass a substantial
portion of the world's population.  (And because I found some phonologies
for them online.)

Then, I suggest a revision of the method of categorizing words.  The basic
roots are expanded in number and made more specific in meaning.

So take a look and tell me what you think.
1) Would such a language be easier for people to learn than some of the
other philosophical languages out there?

2) Does it make sense?

3) Are the sounds chosen easy for you to pronounce?

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

From: "Andrew Nowicki" <andrew@...>
Subject: Ygyde


> You are welcome to contribute any ideas to
> the Ygyde language. I work mostly on the
> dictionary while Patrick Hassel-Zein
> works on the grammar. We are very open
> minded.
Hey, that's great!  I happen to think the whole concept of a philosophical
language is pretty neat.  A few months ago, I started work on a language
comprised entirely of nouns, adjectives, but no verbs.  Now, the purpose of
that language was quite different than yours, but I discovered a few things.
First, that the system of the language made compound words really useful.
The majority of the language is made of compounds.  Second, I found which
compounds were the most useful.  But most importantly, I found that the
meaning of the word could not be _known_ from the roots, but could be
_guessed_.  After all, as anyone can see, noun-animal-long-loud could refer
to a great many things.  But once you know that that's a rattlesnake, it
makes a lot of sense.

Ygyde as it stands has a number of fundamental problems, as I see it.  Since
I'd like to see it improve to the point where it could easily be used for
communication, I also propose some changes to resolve these problems.

1) Words that need to be distinct in sound are too close, due to their
meanings being similar.  For example, the names of the letters of the
English alphabet are far too similar sounding.  So in situations where
clarity is required, we always pronounce them differently.  Such as when I'm
giving my name on the phone, "F as in Frank".  Or on the radio, "alpha bravo
charlie delta echo...".

2) The same roots are used far too often, again making words seem similar or
repetitive.  For example, the "noun" morphemes are extremely common.  Much
of the impression a word makes on the memory is through the first sounds of
the word.  If a third of the lexicon consists of words starting in the same
sound, it becomes very hard to memorize.

3) The basic sounds being used are hard to differentiate, and even
pronounce, for most people in the world.  This means that the similarity
problems are amplified enormously.


So here's what I'd propose to solve these.

--- SOUND PROBLEMS ---
First, revise the basic phonology.  I would recommend the following
principles:

- Use sounds that are maximally distinct from each other, i.e., spread the
sounds out as far as possible.  Imagine if you could only have three vowels,
and you picked the ones in "put, putt, pot".  A maximally distinct set would
be more like "peat, poot, pot".

- Use sounds that the majority of the world's people have in their own
languages.  I wouldn't recommend using ones that are found in _every_
language as your criterion, as that would limit you to, hmm...  n, t, w, k,
and s?  No, actually, I think it'd be even worse than that.  Anyway, if
_most_ people can speak it, that's as good as we'll get.

- Allow enough clustering of vowels and/or consonants that single and double
syllable morphemes number in the thousands or tens of thousands.  This way
we'll have enough to make up the root words.

Now, I'll pull out some of the books, look around the web, and do some
phonology research.

---

After a bit of poking around, I found some interesting charts of sounds.
Below are all the sounds common to English, Spanish, Arabic, and Chinese
(Cantonese).  Between these four languages, we have a substantial portion of
the people in the world.  All who speak any of these languages have these
sounds:

p t k
m n
f s h
w l y
a i u

That's 11 consonants and 3 vowels.  Allowing CV syllables, that gives us 36
possible syllables.  But with only these three vowels, we have a great deal
of distinctiveness between them.  So let's expand that to:

a i u
ai au

That gives us 5 vowels, including diphthongs.  Or 60 syllables.  If we allow
CVC syllables, that increases the number to 720.  But it gets even better.
Given the rules of these languages, we can allow some other diphthongs:

a i u
ai au
ia iu
ua ui

Now we're at 11 consonants and 9 vowels & diphthongs.  Yielding 1,089 CVC
syllables.  If we were to add even a single additional consonant, though,
that would increase to 1,296.  Looking at the charts, it seems Arabic is the
only language that doesn't have a /ts/ or /tsh/ sound.  But they have a
/dzh/.  So let's add one more sound, spelled |ch|, indicating the consonant
at the beginning of "church" or "juice".

p t k
m n
f s h
ch
w l y
a i u
ai au
ia iu
ua ui

With this system, we could have words like:

chat
mum
mai
chuk
yus
tin
su
hach
yap
chin

And these would all be relatively easy to pronounce by the majority of the
people in the world.  More importantly, none of the sounds would be readily
confusable with others.


--- ROOT WORD PROBLEMS ---
As mentioned before, the roots of Ygyde are too common, and the compounds
too similar.  For example, all the body part words sound somewhat similar,
which is a problem when trying to refer to parts of the body.  And some
roots are used so frequently that they make their compounds look similar to
one another.

So to solve this problem, I suggest two things.  One, that the roots be
restructured on a more horizontal hierarchy than a vertical one.  What I
mean is this:  That words in a specific field, such as parts of the body, be
made from different roots, and that things of similar function or form be
made from similar roots across fields.

For example, in the current Ygyde system, the words for wrist and arm might
be something like "noun-bodypart-middle-joint" and
"noun-bodypart-middle-long".  These are far too similar.  Doctors speaking
this language would be constantly repeating themselves and overarticulating
words to compensate.

I propose a twofold change.  One, that the number of basic roots be
enlarged.  Two, that the naming system be restructured.  That way, the words
for wrist and arm might be "hand-joint" and "top-appendage", with prefixed
roots modifying the root like an adjective would.  I would put part of
speech roots at the end, making words look less similar due to their initial
parts.  And we ought to make the most common part of speech (nouns) the
unmarked one, so that if a word has no part of speech marker, it is a noun
by default.

fan  hand
su  joint, pivoting like the wrist, neck, shoulder
laut  hinge, pivoting like the elbow, knee
chuk  appendage, arm/leg
li  upper
pa  lower
tian  head, chief
yak  center, main body
nal  outer
kis  inner
mai  finger
nun  foot

Also remember that a term can have different meanings, so long as they are
in different fields.  Consider the English word "head".  You can speak of
the head of an animal, of a bed, of a table, of an organization, and it's
always clear.  "Head" means the top and most important part, specifically
that of the body, but generally that of the specified thing.  Each of the
words here has a specific meaning that can be more generally applied.

tian  head
tiansu  neck  (head-joint)
lisu  shoulder  (upper-joint)
lichuk  arm  (upper-appendage)
kis lichuk  upper arm  (inner-upper-appendage)
lilaut  elbow  (upper-hinge)
nal lichuk  lower arm  (outer-upper-appendage)
fansu  wrist  (hand-joint)
fan  hand
mai  finger
yak  torso
yaksu  waist  (torso-joint)
pasu  hip  (lower-joint)
pachuk  leg  (lower-appendage)
kis pachuk  upper leg  (inner-lower-appendage)
palaut  knee  (lower-hinge)
nal pachuk  lower leg  (outer-lower-appendage)
nunsu  ankle  (foot-joint)
nun  foot
pamai  toe  (lower-finger)

Keeping in mind the principle of having specific meanings for words that can
be used generally, we could add the following:

tif  hair, anything shaped like hair - long, thin, flexible, plentiful
lin  leaf, anything shaped like a leaf - flat, somewhat small, thin,
plentiful
han  land
chun  horse
tak  sheep
fai  bird
mau  house

hantif  grass  (land-hair)
chuntif  horsehair  (horse-hair)
taktif  wool  (sheep-hair)
failin  feather  (bird-leaf)
maulin  roof-shingle  (house-leaf)

Now, the single syllable roots we're defining are for words commonly used.
Less common roots would have two syllables.  But with 1,679,616 possible
two-root compounds, using only one-syllable roots, I'd say we've got plenty
for most any word we'll need.  After all, who cares if a rarely used
technical term is a real mouthful, so long as the words needed most often
are simple and short.

Let's try making some more roots, and listing all the ones so far.

chun  horse
chuk  appendage, arm/leg
fai  bird
fan  hand
han  land
hak  six
hip  seven
ich  one
is  ten
kan  tree
kin  five
kis  inner
kuat  four
laut  hinge, pivoting like the elbow, knee
li  upper
lin  leaf, anything shaped like a leaf - flat, somewhat small, thin,
plentiful
mai  finger
mau  house
mit  stream
ni  two
nal  outer
nuan  nine
nun  foot
pa  lower
su  joint, pivoting like the wrist, neck, shoulder
tak  sheep
tas  three
tian  head, chief
tif  hair, anything shaped like hair - long, thin, flexible, plentiful
tus  substance
ul  something strong, large, great
wan  hill
wuk  eight
yak  center, main body

And some compounds:

taktus  mutton
kantus  wood
mittus  water
ulmit  river

Now for a bit of grammar.  As you've noticed, adjectives would obviously go
before their nouns.  Nouns used as adjectives in compounds go this way, so
would everything else.  Numbers, being adjectives of a sort, would come
first.  A nice feature, something like in Russian, would be to allow numbers
to go after nouns, meaning around that many or more.  That would allow a
simple, yet efficient way of expressing plurals in varied and easy to use
forms.  People whose languages don't have plurals would find this simple, as
it is never required, and is always transparently derived from a number.

So we could have the following:

- tak  sheep
- ich tak  one sheep
- ni tak  two sheep
- tas tak  three sheep
- tak ni  a few sheep, two or more
- tak is  many sheep, probably at least ten


--- Some last comments ---

So if you're reading over all this and thinking that we'll use up all the
one-syllable roots too quickly, don't despair.  As any of these can be used
in a more general fashion, many words won't need a one-syllable root, as
they can easily be derived from two.

- wantak  goat
- ulwan  mountain
- hantus  soil
- mitchuk  tributary
- ultus  strength

Anyway, this whole long list is just what I think might make a more widely
accepted language made on a philosophical basis.

If a fisherman is talking about different kinds of fish, they can't be
called by similar names.  When counting, the numbers must sound different.
The more different, the better.  Colors shouldn't sound like each other when
possible.

To sum up:  Words that sound alike cannot be distinguished.


And that's my two cents.

Joe Fatula

Replies