Monday, July 26, 2010

bliss symbols writing system. thoughts.



http://www.blissymbolics.us/

CHARLES was born Karl Kasiel Blitz in the old Austro-Hungarian Empire, a mixture of peoples where, as he wrote in 1965, “20 different nationalities hated each other, mainly because they spoke and thought in different languages.” In the year of Charles’ birth, the Emperor, Franz Joseph I, was in his 49th year on the throne, Johannes Brahms died, and Gustav Mahler became director of the Vienna Opera. Anton Bruckner had died the year before, and the year after saw the pointless fatal stabbing by an Italian anarchist of the Empress Elisabeth while she was sightseeing in Geneva, Switzerland, on 10th September, 1898.
The languages Charles heard all around him were not logical at all, and by the time he was in high school, he says, he refused to learn the unruly rules of languages, and he refused to differentiate between the relative, indefinite, intensive, reflective or reciprocal pronoun. He would not believe that children must be tortured with such illogical matters. By this time Charles had been impressed by two wonderful logical languages expressed in the symbols of mathematics and chemistry which could be read by anyone no matter what their mother-tongue might be.

---
D: not sure how this is supposed to work. The symbols are not linked to any spoken sounds.
But we sound out written material in our heads when we read.
So thinking in Blisssymbols must be problematic.

D: there are 4-5000 symbols. This is the same as a mid-level fluency with Chinese. So immediately, it does not seem to be learned faster than existing ideographic schemes (logographic really).

D: I read over Crockford's PDF. Right from the start, his appraisal of it seems too optimistic.
Simple, yet recognizable and memorable? This is the same red herring ASL deals with. It is not very iconic. The symbols are only clearly linked to a real-world object or event once explained. For example, a circle means 'sun'. This is only clear after the fact, once explained. So the simplified pictograms are not in fact so easily recognized.
The hand sign is sufficiently abstract that I would never have guessed what it is meant to represent.

Thoughts:
- octomatics seems to have a better grasp of ideal shapes. The computer versions are all straight lines. Whereas the cursive writing forms are rounded. I incorporate this into HIOXian. The rounded shapes don't display well on a monitor.
- many of the symbols are very wide for a standard 8x12 rectangular character space
- many of the symbols involve very tiny parts. For example, the set of wheels for 'car'.
- nouns are the default shape, and an extra indicator is needed to turn this into a verb.
- showing subset parts of 'hand' is very clever. Just have a pointer indicating the subset part.
- it is clearly earth and human-centric. Which is fine, since humans are the intended users.
- the man/woman symbols are *almost* interesting. They remind me of Jewish mystical numerology.
I think father plus mother equals son or thereabouts.
The woman symbol contains that most primitive portrayal of the vulva, the triangle (sometimes a card-diamond).
The author shied away from the obvious portrayal for man (tripod joke from Austin Powers here).
The gender-neutral 'person' figure is anything but intuitive, being a flat base with a pole above it.
It is not clearly related to man, woman, or man and woman.
Now, superimposing the man and woman figures with my anatomically-correct suggestion would work for man-or-woman. De-emphasizing gender would result in a basic 2-legs-torso figure.
- that grass and hair should involve the same symbol at different heights is anything but intuitive.

---
D: I realize he was designing a system for mechanical typewriters. But with the advent of computers, certain portrayals become possible. For example, my proposed layered-colour HIOXian bi/trigraphs.
The man/woman symbols, once superimposed would now indicate the following:
1) both figures involve the the basic 2legs-and-torso character. This is indicated by a 3rd colour which is the other 2 blended.
2) clearly, the 'man-part' is derived from a single figure
3) clearly, the 'female-part' is derived from a single figure.
If we wish to include 'all people of any age', we could even add a 3rd figure with a shorter torso with a similar colour shift.
After a certain number of colours, brown or black could be used.
i suppose we could accomplish the same thing in greyscale.

A stray thought was that a liquid ink could in fact physically blend and accomplish the same thing on a mechanical typewriter.

------
Decimese.
All this colour-blending got me thinking.
If we have compound nouns, verbs, or adjectives we could use a final-syllable condensed indication.
Since all vocabulary items must end in a CVC syllable, with the final consonant being M, N, or NG, we have a problem.
If I only denote the final main element (e.g. black board is noun not adjective) then, we may lose too much meaning.
In Decimese, 'black' does not default to adjective. In fact, 'board' does not default to noun. And nothing defaults to verb either.
Without being assigned a grammatical role, the bound-root cannot be used independently.
The root for 'black', for example, becomes undefined. It has yet to be defined as to blacken, to make black, to become black. It is not yet blackness or blackly. It is not the colour black, or any analogy or metaphor thereof.

So what do we do with a concept such 'blackboard', let alone 'blackboard eraser' or 'black board-eraser'?

Well, if we reserve an HVN (H plus vowel plus nasal consonant) final syllable for this purpose, then we could use vowel diphthongs to this end. This only works with linguistic backgrounds (for both speaker and listener) able to use diphthongs.
English and Chinese, for example.
In the case of blackboard, the concept is a noun with board being the core concept.
So we are left with black as an adjective.
Since -N means noun -NG means verb, and -M means descriptor (either adverb or adjective, depending on sentence position), blackboard in Decimese will end in -N. Ergo, the final diphthong would be selected to indicate an adjective, or adjective plus noun redundantly with the final consonant. In theory, if we avoid this redundancy, we may even be able to denote the nuanced difference between 'blackboard eraser' and black board-eraser' and whatnot.

D: why -N for noun? Noun starts in N. -NG is verb cuz English has verbs ending in -ing a lot. Leaves -M for 'modifier'.
Note earlier proposals for Mandarin vs Cantonese speakers that only involve 2 not 3 nasal consonants.

----
Next stop: recreate my top-1000 common English word list and nail down common recurring concepts.
These words occupy a special niche, being neither closed class function words, nor quite open class vocabulary items.
Because of their common usage and complex meaning, they are assigned simple short-hand one-syllable words to denote them. Their construction involves certain consonants only as an indication of their special status.
They are likely to be formed from LRWY and H and not the usual pairs of voiced/voiceless consonants like PB.
Sufficient planning will allow future expansion to include triple consonant clusters such as STR and such. This is not intended for at least the first generation of users.
Common recurring concepts in the 1000-word list include system, study-of (ology), social and organization hierarchy, and social/political/economic/religious distinctions.
This immediately cuts down on vocabulary demands.

Hopefully it will avoid such past nonsense such as Esperanto and books/libraries.
Book -libro. You'd think book place - librujo? would suffice. Maybe a distinction for sell-place.
But nope- out of nowhere - and never to be seen again in other vocabulary items - we get biblioteko!
Haha found the Espists fixed this one on their own - libraro.
libraro book collection
librejo library
D: not sure what the distinction is.
Biblio Bible
bibliografio bibliography
bibliotekisto librarian
biblioteko library
D: so are these related to the bible? Wouldn't bible-place be a church? The -isto a priest?
teko briefcase, file
So now a sloppy translation of biblioteko might be...
bible... briefcase?
Wow.

We see here that one must not skimp on having a sufficient number of core concepts and words, even if this places increased demands on memorizing lists.
The alternative is heavily compounding in a way that is fairly arbitrary and not at all clear.
Not to mention wordy.

Decimese will be pretty wordy. There are just not that many syllables possible in the most basic form. And it MUST have a basic form, for world-wide appeal and utility.
The shorthand options must squeeze as much as possible out of the limited consonant clusters and vowel clusters.
This is an inherent constraint of my design principles.

It's gonna take a long time to figure out right the first time. So initial users do not end up being beta-testers.
That's the way it has to be.
Plus there is no rush.

My self-imposed goal is 2045 - the 100th anniversary of the UN.
Presumably, around 2020- the 75th anniversary, the subject could be brought up, mostly playfully.
A simple thought experiment.
A what-if.
This gives me to about 2020 to nail down the language- another decade for vocabulary.
After that, the goal becomes to encourage as many early adopters as possible, to create online communities.

And not to stagnate - level off at a marginal rate- like Espo.

Gonna be a heckuva trick to pull off.

Friday, July 23, 2010

detect autism with syllable forming delays

http://www.sciencedaily.com/releases/2010/07/100719162645.htm

The autistic sample showed little evidence of development on the parameters as indicated by low correlations between the parameter values and the children's ages (from 1 to 4 years). On the other hand, all 12 parameters showed statistically significant development for both typically developing children and those with language delays.
Warren says that children with autism spectrum disorders can be diagnosed at 18 months but that the median age of diagnosis is 5.7 years in the United States.

Could be cheaper.

Wednesday, July 14, 2010

making a speakable version of Stokoe's notation via Visemese



D: I just finished a book called "The Linguistics of American Sign Language" by Isenhath.
It was an easy read which was not too heavy on theory.
It outlines the premise that Stokoe used, that there are 5 aspects to each ASL sign.
The devil is in the details, of course.
Some unusual signs use full-head movement.
The speed of the sign also carries meaning.

-----

Aside on sci-fi story language project.

The key lesson I learned was that one can remove the 'filler words' in English - really function words- and still carry on a meaningful conversation when context is already provided on the subject.

I applied this to my sci-fi story English creole which uses musical quarter-pitch notes to convey English function words.
ASL typically does not bother to include definite / indefinite articles such as a or the.
By paring away articles, I finally have enough quarter-pitch-note meaning slots to completely supplant common English function words.
The sentence "Dog bite boy" uses quarter-pitch-notes to convey all additional information.

------

Anyway, I thought cross-mapping Stokoe's notation onto syllables in Visemese that can be pronounced could be interesting.
Every sign typically contains a wealth of steps required to form it properly. ASL shies away from forming compound words in the same fashion as English, since 2 signs consecutively take quite a lot of time to do.
Instead, it tends to incorporate details into a single sign, such as replacing a natural hand position with a letter sign to change the meaning. It also heavily truncates 2 signs, resembling spoken language agglutination.

I remain intrigued by the prospect of a truly international spoken-sign system.
Sadly, we have recreated the national and regional boundaries of spoken languages with proprietary sign languages.

I have noted in the past that Decimese can be modified to interface with my lip-readable Visemese scheme.

I suspect Stokoe's notation will take too long to say since there are so few phonemes available to Visemese.
Various 'cheats' such as vowel gemination could be used. However, as always, Speedtalk-esque gemination has the problem that it then takes longer to say. This does not seem to be worthwhile to pursue.

An interesting approach to compressing Stokoe's notation would be to include facial expressions as a meaning-unit.