Monday, January 18, 2010

Heinlen's SpeedTalk and existing similar languages.

Speedtalk is an idea for a new language put forth by Robert A. Heinlein in his novella, Gulf (1949). Speedtalk was defined as an entirely logic-based language, and it was a key plot device. The basic concept was that the conlang would utilize a complex syntax with a minimal vocabulary and a phonemically extensive alphabet (including such letters as œ, ħ, ø, and ʉ), and it was therefore considered extremely efficient. A single phoneme indicated a word, so a "word" indicated a sentence. In the only example given, a single word meant "The far horizons draw no nearer."

The story invokes the notions of the General Semantics of Alfred Korzybski and the work of Samuel Renshaw to explain the nature of thought and how people could be trained to think more rapidly and accurately.

The supermen communicate in an arcane language, a form of English called Speedtalk, which is both unintelligible and unlearnable by outsiders. Speedtalk is founded upon two principles: a reduced lexicon, and an enlarged phonology.
Any English sentence can be composed from a small vocabulary, such as the word set of Basic English. Also, although the human vocal tract can produce hundreds of different sounds, no existing human language normally makes use of more than a few dozen of them. In Speedtalk, each word from the Basic English set is assigned to a different sound. A sentence in Basic English can therefore be pronounced in perhaps one fourth the normal time.


D: Oh. Been done. A designed language that explores this.

Ithkuil’s phonological system of 65 consonants and 17 vowels is based on sounds from a variety of languages, including such as Chechen or Abkhaz. It is often difficult for a monolingual speaker to pronounce, or even to distinguish, some of the sounds. The consonants of Ithkuil are as follows:

No person is hitherto known to be able to speak Ithkuil fluently; Quijada, for one, does not.[1]

D: Hmm. A language so difficult that it cannot be learned as an adult.
Meaning only those raised on it from birth could possibly use it.
That narrows down its utility...

I suppose, ultimately, that the brain's inherent limit, its Universal Grammar, provides a hard cap on how complex a language can be.

The most documented living language with the most tones currently is Ai-Cham (錦話), a member of Kam-Sui languages in the Tai-Kadai language family. It has a total number of 11 tones[citation needed]; Pinghua has 10 tones[citation needed].
However, preliminary linguistic work being done in the Chatino family of languages in southern Mexico suggests that some Chatino dialects may phonologically distinguish as many as 14 tones.

Language with the most words: English, approx. 250,000 distinct words

Language with the largest alphabet: Khmer (74 letters). This Austro-Asiatic language is the official language of Cambodia, where approx.12 million people speak it. Minority speakers live in a handful of other countries.

The language with the most sounds (phonemes): !Xóõ (112 phonemes). Approx. 4200 speak !Xóõ, the vast majority of whom live in the African country of Botswana.

Language with the most consonant sounds: Ubyx (81 consonants). This language of the North Causasian Language family, once spoken in the Haci Osman village near Istanbul, has been extinct since 1992. Among living languages, !Xóõ has the most consonants (77).

Language with the most vowel sounds: !Xóõ (31 vowels)

D: OK, let's look at a fairly basic CVC (consonant-vowel-consonant) form syllable:
1) about 30 vowels
2) about 80 consonants
3) about 10 tones
4) 3 way vowel gemination? Like Thok Reel.
D: so... 80 x 30 x 80. x10 x 3?
That's ignoring diphthongs.
About... 500 0 0 0 0. Five million possible 1 syllable words? Gack.

So if this can be done, why is it not?
Diminishing returns. Design tradeoffs.

Take the simplest of natural languages, Rotokas.

Rotokas possesses one of the world's smallest phoneme inventories and its alphabet is perhaps the smallest in use. (The Pirahã language has been claimed to have fewer speech sounds, but it is not written.) The alphabet consists of twelve letters, representing eleven phonemes. The alphabet characters are A E G I K O P R S T U V

D: one natural language - a pidgin - Taki Taki- has a vocabulary of 340 words.

Sranan (also Sranan Tongo "Surinamean tongue", Surinaams, Surinamese, Suriname Creole, Taki Taki) is a creole language spoken as a lingua franca by approximately 400,000 people in Suriname.[1] It is the mother tongue of the Creoles. Sranan was previously called nengre or negerengels (Dutch, "negroenglish").

D: the benefit of fewer sounds that are used as significant is that they are more clear.
The difference between normal, nasalized and velarized is pretty subtle.
Toss in less than ideal conditions, such as background noise, and I bet Rotokas is still crystal clear.
That would not be true of Xoo.
Or SpeedTalk.

I also have another objection to SpeedTalk as a concept.
It is so busy trying to break the record for per-syllable info density that it ignores per-time.
By this I mean that complex tones (rising/falling) and vowel gemination (1, 2, 3 time duration) extend the time duration of
a syllable.
The result is that a simpler system, with fewer possible meanings per syllable may very well have a HIGHER info density per second.
And then there is the chance of pronunciation and hearing/processing mistakes.

I'm sure Spanish vowels are more distinct than English ones, for example.

bit, bet, bat, but beat, bait, boot, boat, bought, bite, and bout.

Versus AEIOU. Ah, eh, ee, oh, oo.

Aside. Yup, don't think I'll be making a carbon copy of Lojban. Being so unambiguous and logical is, well, downright unnatural.

D: good primer. Too bad langmaker folded.

Logical languages are also unsuitable

The disciples of logical languages (loglangs, such as Loglan or Lojban) often propose using such a language as an international auxiliary language. However, loglangs are poorly equipped for this purpose. The loglangers tend to overlook the simple fact that language and formal logic do not serve the same purposes. Language is not primarily about making propositions that can be mathematically proven or disproven; its purpose lies in communicating ideas and emotions. There are many facets of language use which logical languages do not cover well. Language is language and logic is logic; they are different things.

It is also such that logical languages are very difficult to learn and use. Most people are not acquainted with the intricacies of formal logic; they cannot be expected to learn it in order to learn a language. I once tried to learn Lojban; I quickly abadoned that attempt because I simply could not understand how the language works. The language does not even have the same kind of parts of speech as human languages have. Instead of nouns and verbs, phrases and clauses, Lojban has things bearing such fanciful names as brivla, cmene or selbri. Those are Lojban words which cannot easily be translated into any other language; it tells a lot that Lojbanists use these words rather than English equivalents when talking about Lojban in English. I haven't met the same difficulties in any grammar of a natural language, no matter how exotic. Logical languages may be absolutely neutral - but only because they work in a way completely different from how human languages work, and are thus exceedingly difficult to master.

D: I found vast lists of alphabetical Lojban vocabulary.
Of no use whatsoever in learning it.
Is there a useful software primer?

D: the Decimese logic/ethics vocabulary will be terse and concise.
But I won't be slapping logic in its core. It would cease to be a viable IAL.

Lojban, having been designed to test the Sapir-Whorf hypothesis, was NOT designed as an IAL.
Its fans seem to think they can tack on 'oh yeah- and it could a great IAL!' as an afterthought.
Language design doesn't work that way.
If the principle is not central during design, it certainly won't appear in the final product.

Notably, I strongly part ways with all cultural neutrality premises. There have been 2 major approaches.
1) Euroclones. Esp-o, et al. Europe rocks- suck it up.
2) Neutral. Nobody opposes it... and nobody supports it.
And I suppose, the usual North American refrain that everybody should/does use English.

So I suppose, in order, Decimese is designed to:
1) be easy
2) be powerful
3) suck up to the English (now) and Chinese (later).

A language that is made to be learned by adults better be much more Rotokas than Xoo.
As many overt and redundant indicators of word boundaries and grammatical component should be present as possible.
I default to subject-verb-object for the reason of 3).
Plus there are two ways to bypass word order:
1) Europe rocks- suck it up. Latinate. Heavily infixed. Hard in itself to most of world.
2) same trick, but using isolated word particles. Like Lojban.

As I've said before, there IS a reason to use a different word order: SOV.
Most 'natural', based on intuitive hand gesturing.
Also, most clearly parsed by a computer.

On a related note, I think I can have an optional 'long vowel' system for external word particles.
Not of any immediate use though. Not in Mandarin? Then don't bother.


D: do I feel a bit like a prostitute, sucking up to English and Mandarin?
Yep. But if nothing before has worked, then it's time for a new sales pitch.
IALs remain a "solution without a problem" in the eyes of most of the world.
So I need to implement a game-changing strategy if I expect to not get lost in obscurity like every other effort.


Brian Barker said...

I think that the choice is between English or Esperanto as the future global language rather than an untried project.

Your readers may be interested in

Dr Kvasnak teaches English at Florida Atlantic University.

A glimpse of Esperanto can be seen at

Matthew said...

sranan (aka taki taki) appears to have vocab in the 1000s already. Not suprising that it left the 3 digits quickly, since it has a few languages (English & Dutch) ready for bulk word loaning.

Re: B Barker. The conlanger's first goal is avoid utter obscurity, not to dethrone or replace Esperanto. As long as a new language address an audience and problem different than Esperanto's, it can find an audience and attract a fan or two. [not to say there isn't a role for throw-away conlangs, too]

dino snider said...

And Lojban fans are sure IT will be next big thing.

I like Model T Fords, Brian, I really do.
Just don't expect it to keep up with a 60s muscle car, let alone a modern sedan.
I noticed the # attending the Esp-o Congresses is not growing over the decades.

Yeah, Matthew, I hope to court a niche patron group. Everything from humanists to hard atheists. And Decimese will actually have the tenets it espouses in the very core of the language, rather than being a gimmick to market it after the fact.

Dino Snider said...

I'm writing a sci-fi story (novella?) on how we could manage to obtain a Speedtalk type of language. Since my story involves stolen brain bits, regenerating stem cells and biomimetic computer neural nets, it's safe to say it won't happen any time soon. If somebody tried to teach Ithkuil to a baby, we might finally see what the limit of Chomsky's "Universal Grammar" really is.

Brian Barker said...

Interesting comparison with the Model T Ford.

Why not compare Esperanto with a bicyle :)

Much more sensible !

Bill Chapman said...

I'm all for Esperanto,but I have no objectionm to any other conlanger having a try!

Unknown said...

there is also Gidzic. Not a loglang but it does use one phoneme per English word. presently in developement but already online: