In terms of maximum amount of information conveyed per amount of language used, that is. I’m interested in two answers, one for the oral form and one for the written one. And I’m not looking for types of shorthand, either - just languages that are used in ordinary conversation or writing.
The oral form would have to be sign language.
Written? Ummmmm…Newspeak?
Answerquery lang doubleplus ungood rewrite fullwise antesubmiting.
Written-Chinese. One character = one word.
-
-
- I don’t quite understand by what you mean by “efficient”, but one MS developer reference I have somewhere notes that you should provide extra long/large text areas in programs intended to include multilanguage support because compared to English, the same text in most other languages is longer. - MC
-
-
-
- Well nuts. It may have been the Borland C+Builder 5 book instead (this is the book that comes with the full version of C+Builder 5 and it doesn’t have an ISBN number, it’s not available separately AFAIK):
-
- Chapter 11: Internationalizing Applications-
-Page 11-8 “…English text is almost always shorter than its translations… …Avoid abbreviations, they do not exist in languages that use ideographic characters…”
Table 11.2: Estimating String Lengths
Length of English String Expected Increase
(in characters)
1-5 100%
6-12 80%
13-20 60%
21-30 40%
31-50 20%
over 50 10%
~
-I am fairly certain I have seen this general point mentioned in either MS’s “Programming Windows 5th Edition” or MS’s “Programming Visual C++ 5th edition” but I can’t seem to locate it now. - MC
-
-
- It is true that one character=one word is concise, but as I whined about in another thread, it is horribly inefficient to learn, and it’s impractical to make typewriters for it. Also mentioned in that other thread, the same word is not pronounced the same way everywhere Chinese is spoken, so I’d say that prety much blows it out of first place.
-
- IF everyone everywhere pronounced the same word the same way, and IF the ideogram was a graphic representation of the sound of the word, I’d agree with you. - MC
Many languages use a single word to say the same thing as an English phrase. But sometimes the word is longer than the English phrase because it has more letters. Or sometimes because it has more syllables.
So, what do you mean by “amount of language”? Are you counting the number of words, number of letters, or number of syllables? Or the length of time an average speaker takes to utter a statement? Or the amount of time an average reader takes to read it?
In measuring the amount of oral language used, are you counting non-verbal content? How many “word equivalents” do you add for hand motions, for example?
Kind of true, but their are a whole dang lot of compound words that in english would be represented by a single word, for instance wine is putaojiu (3 characters) socialism is shehuizhuyi (4) even eye needs two characters in yenjing, plus about a million or so more examples.
The effect that the Borland manual talks about is really an effect of translation and not specific to English. As a general rule, if you write some text in one language and translate it into others, the original text will be shorter than the translations. This isn’t always true, since you can find some exceptions to this, especially for very short texts.
The rule of one character = one word is more true for Classical Chinese (wenyanwen) than it is for modern Chinese (which is called baihua, or “white language” for some reason). You whiners, you think ordinary modern Chinese is hard to learn? Try Classical, where a single five-or-six character sentence must be interpreted by an entire essay in the modern language (in, say, the works of Confucious). Classical Chinese can also be thoroughly unintelligible when spoken aloud, but makes perfect sense when read on the page. (I once saw a short story written with characters that were all pronounced “yi”, though with different pronunciations.)
I’d say classical Chinese is the most efficient traditional mode of communication, but to understand it requires not only an understanding of the characters themselves, but also a thorough grounding in the culture whence it arose. Moreover, AFAIK it’s a semi-dead “language” and would probably be useless for trying to convey modern ideas like democracy or microchips.
BTW, it’s true that modern written Chinese can be understood by people who speak different dialects. But why does MC argue that this is a disadvantage, or that it somehow disqualifies written Chinese vis-a-vis the OP? MC, you sound embittered by the language.
Esperanto is in fact very inefficient, since the prefix-suffix system provides very little vocabulary to learn, but pretty long words. But I think the efficiency award must go to an artificial idiom. Maybe the hard-to-learn but very abstract and mathmatic languages Lojban and Loglan? I don’t know too much about them, but as far as I know, they can express very complicated things in very short phrases. Check the web for information.
A little off-topic, but I remember reading a Heinlein story about an elite group of super-intelligent people (it may have been called “the great divide”).
Anyway, the language they used was super-efficient. As I recall, the language had many many different sounds. Each sound corresponded roughly to an english word. Thus a “word” in their language was like a whole sentence.
Realistically, I don’t think anything like this could actually work. Languages are full of redundancy because it’s impossible to listen with 100% accuracy. When you chat with someone, you may think you hear every word, but usually your mind is filling in some gaps.
Sign language isn’t oral. In fact, the whole point to sign language is that it isn’t oral. 
It’s symbolic, which puts it in a separate category from oral and written (Along with semaphore, Morse and smoke signals, I guess.)
To answer the OP, maybe it’s English. English certainly has the largest vocabulary of any language, which gives you more opportunities to express a concept in one word rather than multiple words.
But what I don’t know is how English compares to all languages in terms of the percentage of words that are syntax helpers - articles, conjunctions, etc. It seems to me that the ideal language would minimize non-contents words while maximizing the effectiveness of sentence structure and syntax. The only other language I speak is French and it’s comparatively inefficient, though it’s a pretty language, and I know some sign language, which has basically no syntax words at all, but has a much more limited selection of terms.
I agree with schnitte, though, that idiom is often the best language. Having defined, rigid rules creates its own inefficiency because you can’t possibly predefine all the different ways humans might want to communicate. Look at the various permutations and meanings you attach to the now-commonly-used term cool, which used to just mean “not warm” but can now express many different levels of subjective approval depending on context or inflection.
[hijack]
The name of the story was “Gulf.” I agree that the language would not be practical. I work in two way radio communications, and you would be amazed what kind of problems with intelligibility that you can get into through relatively harmless changes in the audio signal. A restricted frequency response and a little noise, and familiar words turn into gibberish.
The written form of the language Heinlein proposed would be more interesting. He proposed something like one hundred base words and the numbers up to , I think, sixty. All one hundred and sixty words would be represented by a single symbol each (each, of course, with its own pronunciation arranged to give one syllable.) More words were to be made by prefacing one of the base words with one of the number words. This gives you six thousand, two syllable words of two characters each. Further words were to be made by adding another number word. You then had 36000 words of three syllables/characters. Worst case was to have been four character words. This would be a very compact written language. Like I said, though, you lose the redundancy of normal languages. I would also imagine that some of the combinations would be a genuine bitch to pronounce.
[/hijack]
For the sake of discussion, let’s ignore non-verbal content, since the language can also be used in a non-visual context (on the radio or phone, for example). And as far as efficiency, for oral language, I meant minimum amount of time to convey a particular idea. For written, smallest amount of text used. I realize that this can become a bit vague when you look at Chinese characters, where a word can consist of a single character, but each character is much more intricate than a language with a more “familiar” (familar, that is, to a Romance-language-speaking person) alphabet, so, unfortunately, I don’t really have a good answer for that. If we accept traditional Chinese as the most efficient written language, what would most efficient if we only looked at languages with a more “familar” type alphabet (i.e. few characters).
I’ve always thought that one’s “efficiency” in communication (in any language) is more a function of true literacy than anything else. Some people might understand when I speak of someone else as being a bit of a “Hamlet”–one word. If they don’t get the reference, then I’ll have to explain that “he seems immobilized and indecisive in the face of his own rage and guilt, and there’s also this thing with his mother . . .”
How about loglan/lojban?
It probably compresses complex concepts and abstractions better than most because you can apply various short “operator” words to predicates, rather than having to construct ancillary phrases as modifiers. For simple statements and ideas, it’s probably no shorter.
Witness the fact that most translation of something in loglan usually requires some hyphenated thing to attempt to convey the degree of abstraction / generalization / time sense being conveyed.
Loglan / lojban are interesting beasts, if you are not familiar with them:
What I find interesting is that it is a language that could concievably be used for human communication, but is ammenable to machine parsing because of the totally unambigous grammar.
(loglan / lojban are a major departure from what most people consider a “language” - it throws out the normal “parts of speech” concepts and uses a basic vocabulary of predicates with with functional slots for operands, and operator words to handle modifications like tense. It was originally intended in the 1950’s as a means of testing the Sapir-Whorf hypothesis by constructing an extremely logical language to see if using it facilitated logical thought. It has evolved far beyond that purpose.)
I’d like to take a different approach to your question. Mathematics is the the most efficient human language for expressing complex ideas.
Before you rip me to shreds, strictly speaking, mathematics qualifies as a language - it uses written symbols to represent thoughts and ideas. Sure you can’t express common everyday ideas like “My hovercraft is full of eels” as an algebraic expression, but you can’t accurately describe the full import of the fundamental theorem of calculus in any social language either.
The amount of information expressed per symbol is extraordinary. There is no simpler or more meaningful way to express to following ideas (not even in Welsh ):
E = mc^2 , or
F = (G * m1 * m2)/d^2, or even
a^2 + b^2 = c^2
I’m skeptical of the efficiency of a language that can’t even settle on a name.