10
$\begingroup$

In one of my stories, I have a race that communicates using only two sounds: "O" and "U". This can be strung together like this:

Uuuoouoouooouuuooouuoouoououooououo

Or something like that. There are no pauses between words within a sentence.
It is essentially a binary language, just like what machines use.

Using only two phonemes, how complex can communication be?

The following concerns come to mind:
1. How can each word be differentiated without any pauses between words?
2. Assuming you can get past the above problem, will it simply take too long for them to convey information?

$\endgroup$
  • 5
    $\begingroup$ This may be cheating, but how about they just make two different sounds that humans can hear? The remainder could be sub- or ultrasonic $\endgroup$ – Punintended Sep 19 at 21:10
  • 4
    $\begingroup$ (1) There is no such thing as a "binary language", and the language described may easily have more than two symbols -- you said nothing about stress, pitch, or vocalic length. (2) There are many ways to determine the end of the words. For example, the first wovel in a word may be stressed, or the last, or the last but one. Or there could be other more complicated conventions. (3) Too long or too short depends on context, and on how long the vowels actually are. (4) And all this is should have already been part of your research. (5) The word "sounds" is meaningless in linguistic context. $\endgroup$ – AlexP Sep 19 at 21:21
  • 7
    $\begingroup$ This sounds a lot like Morse Code to me; I can't see how it is not possible given that while I've never used it, I've always wondered how the signallers and receivers of Morse Code know when the first letter is finished and the second one begins. If you take whatever practice is used in Morse Code and then just convert the dots and dashes to U and O sounds as you prefer, I think you have a good answer. $\endgroup$ – Tim B II Sep 19 at 21:48
  • 1
    $\begingroup$ Let us continue this discussion in chat. $\endgroup$ – overlord - Reinstate Monica Sep 19 at 22:24
  • 1
    $\begingroup$ But the OP asked if two "sounds" could make an entire language and morse code is proof that they can. The dot and dash are the only "sounds" and the equal varrying soundless pause-lengths allow communication of complete thoughts using only two "sounds" $\endgroup$ – HA Harvey Sep 20 at 6:20
9
$\begingroup$

This is fairly easy. With tri-state it's even easier, but for now let's consider just using a click or silence (a la pure binary: 00101010). The answer is a lot and as quick as their memory can handle. Let's lay some ground theory:

  • There are a small set of "proto verbs" (Incorrect terminology so I'm having trouble sourcing this), of quantity less than 30, some of which have no existing counterpart in English. Nevertheless, you can define all verbs using that small set in the right combinations with other words.
  • Similarly, you can do the same with many parts of speech (Cat: four-legged furry mammal...etc.)
  • Nouns/Adjectives are your largest collection. Although you can do either-or. StackExchange would put this at somewhere between 19% and 33% utilization, I believe. Looking at English Dictionaries for a maximum for a number of words puts us at 155 thousand nouns/adjectives which can be enumerated with 18 bits.
  • If the universe has roughly 1082 atoms in it then you can enumerate them with 272 bits which will be of such magnitude that it would be laughable to consider filling it completely. You can use that number as a max. I prefer 64 bits. Your minimum enumeration to match English is less than 18 bits using proto-words (probably something like 12 bits), and 19 bits to match exactly.
  • Words consist of definitions, which are combinations of either other words or proto words. It is compression. Each word added adds to the memory requirements, and the bits needed to enumerate all words. But, adding words increases transmission speed. You can trade memory for speed almost as much as you want here.
  • Crucially you can even compress those words by frequency of usage simply by adopting the same trick that saved Unicode.... UTF-8, just pick a "byte" size that encompasses enough of your common words and make the rest extended bytes. Other variable-length encodings are usable as well. What you end up with is as much content, almost as fast as you please (within a log-factor), as specific as you want (this atom), tailored to your language/culture.

What's the downside?

The number of sounds is a multiplier on transmission speed. Any crafted language could do the above-described things. But the more sounds you have the quicker you can do so. 26 sounds? 4 times the content in two sounds (log(26)/log(2)). But as far as "as quick as their memory can handle" there's a limit: ~380 wpm... which probably is present for non-humans but just different rates.

What about adding silence? Tri-state let's you scrap the UTF-8 "extension bits", or gives you another symbol to play with (log(26)/log(3) now). Speed or compression gains either way.

$\endgroup$
  • 1
    $\begingroup$ I just referenced YouTube... At least it was Veritasium.... (=___=) $\endgroup$ – Black Sep 20 at 10:25
  • 3
    $\begingroup$ If you're proposing an 18-bit word length, then a species which could trill (and crucially hear and _interpret) at birdsong speeds then you'll get ~100WPM, which is equivalent to moderately slow use of normal language. Not too bad. $\endgroup$ – Starfish Prime Sep 20 at 13:54
  • $\begingroup$ Probably a more intuitive way to think of that 4x content factor: Imagine you had 129 characters or so in you language. Now you can write each of these down in a second. But if you want to write down the binary representation of that character (ASCII or 8-bit) you have to write 8x as fast to convey the same information in a second. (8x because here we're log(129)/log(2).... rounded up). $\endgroup$ – Black Sep 20 at 21:03
11
$\begingroup$

It is essentially a binary language, just like what machines use.

Unless they never stop vocalising, it isn't really binary as you have O, U and silence. Binary signals just have high and low, or on and off. Tri-state stuff is distinguished from plain old binary.

How can each word be differentiated without any pauses between words?

So, study of spoken language isn't my forte, but I note that many foreign languages that I do not understand do not have word-boundaries that I can recognise... they just seem like more-or-less seamless flows of sound, broken up into sentences. I can understand the word boundaries in languages I'm more familiar with, because I recognise individual word-sounds and my brain does the parsing for me.

Youmightconsidertextwithoutspacingtobeasimilareffectwhereyoucantellthewordboundariesbecauseyouarefamiliarwiththeactualwordsthemselves.

If you really wanted, you could consider having a special sequence to defined End-Of-Word, but that just seems unnecessary and would take up valuable conversation time.

You'd probably still want to pause for end-of-sentence though.

Assuming you can get past the above problem, will it simply take too long for them to convey information?

Depends how fast they can modulate their voices, doesn't it?

For a technological example, consider radioteletype, aka RTTY (sound sample, youtube link) which can give you 60 words per minute, which is less than half the rate of spoken english (faster rates do exist, but they're not reall meat-emulateable). This is done by sending one character at a time, 5 bits per character. The efficiency at which you can send information depends on how you've put together your lexicon. Morse operators used many specialist abbreviations for important common words, questions and answers. Your language probably needs the same. Uncommon words might take much longer to say, which means that unless the race is quite patient they might have issues communicating complex or technological concepts.

For non-technological examples, consider birdsong. I'm having trouble with the precise search terms that will be of most use to you, but trilled portions of birdsong can go as high as 30 distinct pulses per second (1800 bits per minute, 6 times faster than the basic RTTY mode). That needs some fairly sophisticated muscles and brains to do that trick, though.

$\endgroup$
  • $\begingroup$ It is not Tri-state. There are no pauses between words within a sentence. $\endgroup$ – overlord - Reinstate Monica Sep 19 at 21:56
  • 1
    $\begingroup$ @overlord: Many languages exhibit continuous phonation with no pauses between the words in a syntactic group -- French or Italian or Spanish or Romanian or Hungarian etc. And even in English syntactic groups which can be pronounced continuously are pronounced continuously -- inter-word breaks occur only when continuous phonation would generate a forbidden consonant cluster. $\endgroup$ – AlexP Sep 19 at 22:05
  • 8
    $\begingroup$ @overlord I didn't say between words in a sentence. Do your peeps start making noises as soon as they are born, and continue until they die? If so, its binary. If they ever go silent, then clearly it is a tri-state system. $\endgroup$ – Starfish Prime Sep 20 at 7:15
  • 1
    $\begingroup$ @overlord, the question says "Using only two sounds, and slight pauses between words". $\endgroup$ – Ray Butterworth Sep 20 at 13:30
  • $\begingroup$ @RayButterworth I FORGOT TO DELETE THAT PART, no wonder everyone has been so confused. Thank you for pointing that out! $\endgroup$ – overlord - Reinstate Monica Sep 20 at 13:35
4
$\begingroup$

Latin written with scripta continua had words without spaces. I have observed that some modern Italian speakers also string their words together without spaces except when they need to breathe. The key I think it to recognize words as such as soon as they are spoken and mentally file each one as you hear it and get ready for the next.

The other thing about people is that there are many ways to inflect the sounds O and U. Extra or different emphasis can denote the start or end of the word: for example U U U U U U is not the same as U U U U U U. You can tell I am emphasizing my U because it is louder, and also my eyes bug out slightly and I spit.

$\endgroup$
  • 3
    $\begingroup$ (1) That is scriptio continua, obviously. (2) French is a much better example than Italian. In French, individual words do not have dynamic stress; rather, dynamic stress belongs to a phrase / syntactic group / rhythm group, and such rhythm groups are pronounced with no inter-word boundaries. C'est parfaitement vrai is pronounced [sɛpaʁfɛtmɑ̃ˈvʁɛ]. $\endgroup$ – AlexP Sep 19 at 21:52
  • 2
    $\begingroup$ In some forms of written Japanese there are no spaces either, even to this day. $\endgroup$ – Alice Sep 20 at 9:32
3
$\begingroup$

It could function, but how would it evolve?

A two-sound language is basically binary and is perfectly capable of transmitting arbitrary amounts of information. However, unless the species was constructed artificially or they voluntarily adopted this system after evolving a culture, it seems unlikely that such a species would ever evolve the ability to talk in the first place.

Human language evolved from simpler sounds, each conveying a unique concept, long before we began stringing them together into words and sentences. While the exact sounds early humans used are unknown, we can make guesses; suppose that humming was a sound made while pleased or satisfied, hissing when angry or in pain, a grunt when ready to travel, etc. This could serve as a proto-language where the species was already accustomed to communicating intent through sound, which could lead to using these sounds to convey greater meaning through analogy, like using "ma" (a sound conveying comfort) for mother or water.

A single syllable in a binary language, however, conveys too little information to be useful before you start stringing sounds together. Because of this, it seems very unlikely that a species capable of producing only two sounds would even think of using sound as a means of conveying information. It is more likely that they would develop a language based on non-verbal communication.

Moreover, it is very inflexible - if I replace a B with a P in a sentence, 90% of the time you will understand what I mean - but if I replace a zero with a one in a binary language, you will hear a completely different word. A language this precise would be very difficult for living organisms to use and would have a hard time evolving from simpler forms.

$\endgroup$
1
$\begingroup$

So far no one has mentioned a modulated stream. Consider having two states - off and on. Perhaps the off state is "o" and the on state is "u". You can change the length and or volume of the on state and get just about any sound that you want with the caveat that you will hear an overtone squall. In fact the way humans speak is that we produce a carrier tone, rather several tones overlaid, then modulate that sound to produce "o" and "u" among other sounds, essentially shifting the phase of the overtones.

Consider back before computers had sound cards, software authors developed methods for playing polyphonic sound over the binary PC speaker, called pulse-width modulation. The PC speaker is an internal, simple 5 volt speaker which typically comes with the computer case, and whose hardware only provides an off or on state, i.e. switching that voltage on or off. Tones can be produced by utilizing separate timer hardware ( standard on most computers and separate from the CPU ) for generating an audio-range system timer. That timer signal can then be used by software to turn the power to the speaker off and on at regular intervals.

By determining how many intervals to skip between switching the speaker from an on to an off state, this configuration can be used to approximate polyphonic wave forms on an extraordinarily simple device. Doing this essentially pulses the voltage through the speaker, and because of the mechanical nature of it, the modulated signal actually places the speaker cone in the approximate position it would be in if it were playing a smoothed waveform, with the exception that the carefully spaced pulses of current causes the speaker cone to jolt from one position to another, which has the affect of overlaying the wave with the carrier tone. Also this configuration can only play sound up to half the frequency of the available timer. In the case of the timers available to use with the PC speaker, the max available was either 18kHz or 10kHz, causing a buzzing overtone of about 9kHz or 5kHz respectively.

You can hear an example of the PC speaker thing here. It's messy, and it's much easier to hear in person, but it's easily possible to make out any complex sound.

Not sure how this would sound starting with "o" and "u", but I imagine it wouldn't be much different, just that the overtone would have a quality that is perhaps more voice-y and less like the "8-bit sound" of a PC speaker.

$\endgroup$
0
$\begingroup$

Chinese and Japanese (and other such a languages) have an interesting pronunciation structure: each syllable has a certain "timeframe" and those syllables are pronounced in a rhythm. There are "long" ("two-frame") syllables and "short" ("half-frame") ones, but the timing of "frame" should be kept.

These languages have tens of syllables. But the proposed language would have 10. 4 "basic": OO,OU,UO,UU; 4 corresponding long ("long O"-"long U"); and 2 "short" ones: O, U.

That would be enough for a language, but on average words would be 2-3 times longer than in Chinese and Japanese. This is not a problem though, since some existing primitive languages have some very long words (mostly because they have too few "basic words", and they call a boat "hollow-trunk-that-goes-over-water") and it has worked for them.

$\endgroup$
  • 1
    $\begingroup$ (1) The standard word for what you call a "frame" is a mora. (2) "Chinese"? Mandarin (the national language of the People's Republic and Taiwan) is generally considered to be syllable-timed, with no meaningful distinction between the duration of syllables. (Yes, physically some syllables may take a longer time than others, but this is not considered material for speech, not even in poetry.) $\endgroup$ – AlexP Sep 20 at 8:42
0
$\begingroup$

Very feasible indead.

But the written language needn't be binary. Indeed the letter B could stand for the sequence of sounds uuuouu.

(I assume you are including spaces, or pauses? Or you could have s space be a special sequence say ououououououo. )

If the sounds are continuous with no spaces you might need a special sequence to indicate the start of a sentence e.g. ouuuuuuuuuuo

The might be some ambiguity but in context the language could make sense.

$\endgroup$
0
$\begingroup$

Take a look a Huffman encoding (wikipedia link)-- it's a form of compression that uses shorter binary sequences for more common symbols.

Also consider that all humans have huge brain structures just for our language.

I can't see this limitation being a huge problem.

$\endgroup$
0
$\begingroup$

Other have covered the binary issue pretty well. Remember though that vocal communication, as in humans, is supplemented by many other forms: while the old line about only 20% of communication being verbal is a misunderstanding, the aliens may have even more developed use of gesture, tone, expression etc. (Or semaphore: "Cathy!!")

And if their vocal language does limit them, they may be smart enough to develop a sophisticated written language to transmit data faster. As with humans, reading may be much quicker than listening; with them, writing (or texting, if they have the tech) my be quicker than speaking. What looks like a limitation may mean they end up communicating better.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.