6
$\begingroup$

If while exploring the galaxy, we were to encounter an alien script, would we be able to decipher the pronunciation of the language without actually hearing it?

Assumptions:

  • The language is pronounceable by human tongue
  • The text found is very large, equivalent to several long novels
  • The text describes many aspects of culture and history, and includes many names and possibly even descriptions of how they are pronounced (in the alien language of course)
  • The language follows a writing system that works similarly to an alphabet, with individual letters combining to form syllables which combine to form words
  • Advanced tech level, so computational power isn't an issue
  • The world's linguists are, for obvious reasons, supremely interested in this task

Challenges:

  • I've heard references before that this isn't possible but I can't locate them
  • Would names be harder to figure out than words, assuming the names are not based on actual words to begin with?

Would we actually be able to figure it out, at least somewhat closely to the original? Or would any ideas we came up with be purely guesswork?

$\endgroup$
  • 3
    $\begingroup$ We can't do it for human languages, just reconstruct from existing languages (e.g. Sumerian script but the reconstruction is not without doubts). So no, impossible. Except when this works: "and possibly even descriptions of how they are pronounced (in the alien language of course)" Not possibly but it has to exist, be complete, and the anatomy has to be human like to the detail (so: "humans in space"). Then it is "just" a question of cracking the general language and script code itself (which is unlikely but at least possible). $\endgroup$ – his Aug 1 '15 at 12:08
  • $\begingroup$ If it's alphabet-like (symbols, dashes, etc.) with 20 to 50 characters you could do that with just a couple pages - depending on the content. We did a writeup on legal issues with cryptography, so in a course we were taught classical ways of deciphering. As for the sounds they make...? Does it matter? We would translate it and say them out loud in our language, no? $\endgroup$ – Mikey Aug 1 '15 at 14:58
  • $\begingroup$ @Mikey That's the impression I was under, if the language has limited complexity and is similar enough in structure to our own. The pronunciation has more to do with names and words that don't exist in our language already. $\endgroup$ – thanby Aug 1 '15 at 16:31
  • 1
    $\begingroup$ @ArturoTorresSánchez I've definitely heard of that but one of the prevailing theories is that it's not intended to be translatable and is in fact just a work of fantasy art. $\endgroup$ – thanby Aug 2 '15 at 1:12
  • 3
    $\begingroup$ Heck, we can't even do it for modern English :-) $\endgroup$ – jamesqf Aug 2 '15 at 4:11
7
$\begingroup$

Please encode for Maximum Meaning

With modern day compute power, a xenolinguist would be able to find many many things of interest in this alien script such as the distribution and grouping of letters or ideograms, common words, common clustering of words, numbering system, even the writing style of various authors. With a math text, we might even be able to learn how they encode math. However, stats only gets us so far. Mapping one alien symbol into modern English requires recovering an idea from the mind of an alien.

The math text would perhaps be the closest thing to a Rosetta Stone that we get. Remember that the Rosetta Stone provided the critical piece that allowed modern linguistics to finally translate Egyptian hieroglyphics into a modern language by way of ancient Greek. Prior to its discovery, Egyptian was completely unreadable.

Essentially, these very long texts are random strings of numbers. One can find patterns in those numbers and relationships within the document itself but without some way of mapping a number or number cluster in that long string into a real life concept then translation and pronunciation is impossible.

All Tongue tied

But let's try anyway. Pronunciation won't be possible with any kind of fidelity without talking to an alien or some kind of wicked clever inference from an artifact.

One possible approach exploits how poetry is written to derive pronunciation. Take Middle English poetry. The speakers of that language are long long dead and their tongues with them. There are no recordings to listen to so inferences must be made through textual analysis, specifically which words rhyme together in a poem. Linguistics exploit their knowledge of how poetry works in English. However, this would not work for analysis of Hebrew poetry because they don't worry about clever things with the sounds of words but rather clever arrangements of meanings (or double or triple meanings) between lines. Isaiah is an excellent example of this kind of poet though you'll need a strong background in Hebrew to really appreciate it.

We don't even know if they spoke their language. What if they used scents or bioluminescence to signal with each other? There is no sound to pronounce!

Translation is near impossible. Pronunciation without a "form your mouth like this" guide will be equally impossible.

$\endgroup$
  • $\begingroup$ Very creative idea with the poetry, I can see how that could make it possible given the right conditions. I'm also thinking if the author(s) tried to describe a sound made by an animal or instrument using a vocal explanation ("The man hit the ground with a heavy thud") or onomatopoeia which is referenced in other context as well, that might be a start? It would be drawing a parallel to our own sounds through basic physics. This is all pretty "out there" but I'm just trying to figure out if it's even believable at all. $\endgroup$ – thanby Aug 1 '15 at 16:47
  • $\begingroup$ I believe Egyptian language was hard to crack because the characters weren't detailed enough. They were pictograms which represented whole syllables or words, whereas a Roman language breaks even those syllables down into parts essentially tied to single tongue positions, giving more data to work with. But even so, as Josh mentioned, we still don't know for sure how ancient Latin was truly pronounced. $\endgroup$ – thanby Aug 1 '15 at 16:50
  • $\begingroup$ @thanby true, but exploiting rhyming structures like that requires a deep deep understanding of how a language works and it's historic context. Exposure to a completely alien script won't permit that kind of analysis.... Well maybe but you'd be crazy crazy lucky. $\endgroup$ – Green Aug 1 '15 at 17:23
  • $\begingroup$ Ultimately luck may be good enough (though I'd rather not just hand-wave it with no realistic explanation), just as long as it's possible in some remotely believable way. Any believable fictional story is based on the anthropic principle after all, otherwise there would be no story :) $\endgroup$ – thanby Aug 2 '15 at 1:15
  • 1
    $\begingroup$ @thanby, as long as you include some kind of Rosetta Stone type object, even if you need to go through 1 or 5 steps to get there then the audience needs to suspend far less belief than if you hand wave it with "computers! Stats!". $\endgroup$ – Green Aug 2 '15 at 1:25
3
$\begingroup$

Our ability to translate the language will greatly hinge on context.

From your question, it seems you are describing a singular text. Possibly a collection of news articles or even a "welcome to the planet" type guide. If we were to find this isolated from anything indicating its origin and purpose, we would have almost no hope of translating it. (An infinitely powerful computer would be able to generate many possible translations. It may even be able to calculate the most likely one, but we could never be sure.)

If, however, this was not a singular text, but rather the sum of text found about the ruins of a city, we could certainly translate at least most of it. With such a context, we will begin to see patterns such as enter/exit, open/closed, left/right, etc. (I can't find it now, but I saw a documentary once in which the key that allowed us to to translate an ancient text was noticing a pattern where two words were repeated. Someone hypothesized they were "born" and "died" and they were able to crack the rest of it from there. But would she have made that guess if this text hadn't been discovered with a burial ground, as it was?)

Hypothetically, you could contrive a situation where enough context existed to translate a singular isolated text. Numbers are easily translated if enough are present (math being the only known universal language), and if there were corresponding explanations accompanying the numerical notations, this could be enough to crack the code. Getting a bit more out there, if we suspected the text was written by a race we had never encountered but had heard much about, we might come across a translation by trying to make it fit what we know. Or perhaps the material it's written on or the location it was found lead us to believe the text is a specific one we've heard of (e. g. The Holy Book Of Mak'Ra) and we translate it by already having a rough idea what it says.

Any pronunciation will be guess.

We don't even know how Latin was pronounced in ancient times, and we have a lot of information to go on. We are still speaking several languages derived from it. This helps us to guess at the pronunciation, but there is still much uncertainty involved.

In your scenario, we don't have any such clues. Assuming we translate it, the most accurate we could hope to be in pronouncing the language would be knowing which letters are consonants and which are vowels. Any written pronunciation guide will rely on us already knowing how to pronounce the words or symbols used as reference, which we can't possibly know.

Imagine a Japanese speaker learning English solely by learning the translation of each English word he reads. He can now read English fluently but has no knowledge of pronunciation, so he gets a copy of Webster's Dictionary. Even though he understands everything he reads, he won't know how the phonetic spellings are supposed to sound, and will be no closer to knowing English pronunciation. This is the situation we will find ourselves in.

$\endgroup$
  • $\begingroup$ I think that Latin argument may have been the one I've heard before. Very good point about the context being important, otherwise we would have no starting point. I may be stuck at the "read but can't speak" scenario without some other sort of intervention. $\endgroup$ – thanby Aug 1 '15 at 16:36
3
$\begingroup$

Maybe. It depends both on the writing system and on how much material you have — both written material, and additional clues regarding the sounds used by the language.

One thing you'll need is some indication of what sounds a language can use. With humans, we know this. But with aliens, we would need additional information. This could be recordings of speech (even if we have no correspondence between these recordings and the writing, they would be enough to determine the phonemes of the language). This could be remains or an anatomical text with diagrams. It could be a treatise on pronunciation (once the meaning of the text can be understood). Some information on morphology and psychology can also hint at the pronunciation of a few words, such as onomatopoeia. For example, many unrelated human languages have a word for “mother” that sounds like [ma], which is an easy sound for babies to make.

Languages vary greatly in how difficult it is to relate the writing to phonemes. At one extreme, there's pure ideograms, where the writing is unrelated to pronunciation. At another extreme, there are languages where each mark represents a sound. Among humans, pronunciation and writing tend to evolve independently, so even languages where the writing system was designed to be purely phonetic have usually diverged after a few centuries. English is on the far side of divergence, with spelling only loosely connected to pronunciation, and strong phonemic variation between variants. At another extreme, there's Korean, whose writing system (Hangul) (a syllabary) is systematically constructed out of building blocks where the mark representing a phoneme is a stylized representation of the shape of the mouth when pronouncing that phoneme. If the writing system is designed like the one of English, you won't be able to know how it's pronounced with any certainty since even native speakers vary, your only recourse would be to find a treatise on pronunciation. On the other hand, if the writing system is like Hangul, then you might have a chance to guess at least the original pronunciation (not necessarily the modern pronunciation if it's evolved since the writing system was designed).

The nature of the texts matter as well. If there's any ambiguity in the relation between spelling and pronunciation, our main clue for the pronunciation of human languages is rhymes. Rhyming poetry, theater and songs are how we reconstruct the pronunciations of centuries past. A book on puns would also be helpful. A history treatise, markedly less so — though it might include pronunciation clues for old or foreign names.

Your explorers should definitely be on the lookout for multiple languages. Deciphering multiple languages with etymological connections can provide information about the individual languages. For example, if a language imported words from another, the way those words are modified can provide clues about the sounds of both languages. Of course multilingual texts like the Rosetta Stone can be very helpful. Even better, they might find language textbooks.

$\endgroup$
2
$\begingroup$

It is only possible if (a) there is a sound-based key (a recording of some type), or (b) a detailed anatomical description of the alien's 'mouth' and its operation.

For example it is possible to get some idea of how a Roman musical instrument would sound by looking at an old picture or an archaeological find. This is because we have an idea of the physical appearance of the instrument and the fact that it has strings or there is someone blowing through it.

On a slightly different tack but still connected: See just how difficult it would be even to understand it without a key, have a look at this https://en.wikipedia.org/wiki/Code_talker

$\endgroup$
  • $\begingroup$ I understand your point about the audio key. I wonder if perhaps there was a description of the sound a particular instrument makes along with a description or picture of the instrument itself, if that might be a starting point for deciphering it? Or if the readers managed to translate an onomatopoeia? $\endgroup$ – thanby Aug 1 '15 at 16:39
  • $\begingroup$ That seems possible. I like the idea of getting a clue from an onomatopoeia. It would have to be something that we had in common - but what? The word 'wind' doesn't sound like a wind. Do they have kettles - hiss? Doorbells - ding dong? However if we discovered that their language was based on onomatopoeia and we could reproduce the noises then that would make it a lot easier. $\endgroup$ – chasly from UK Aug 2 '15 at 13:00
  • $\begingroup$ It could be similar to Japanese in that they have an onomatopoeia for just about every sound you can think of. Enough of those in a collection of text (assuming consistency) could give linguists some good fuel for a reasonable guess at it, perhaps? But that's assuming the words themselves are recognizable as such and don't dilute the rest of the language too much, else it might end up making it harder to translate... $\endgroup$ – thanby Aug 2 '15 at 13:28
2
$\begingroup$

Given that a spoken language uses a specific collection of sounds, specific letter representations of sounds are essentially arbitrary.

Even assuming the text is perfectly phonetic, as English is not, there is simply no way to determine which phonemes go with which letters.

So, without an audio recording which corresponds with a specific (and specified) text, pronunciation is unknowable.

$\endgroup$
2
$\begingroup$

I mostly agree with the assessment that it will be impossible. However, there might be one slender ray of hope ...

Mistakes, Homophones, and Texting Abbreviations

With really good statistical analysis, you might find out that a "two" in a certain text should have been a "too" instead. Of course first you'd have to be certain that "two" doesn't mean "too", and that it isn't just a typing error. That tells you "two" and "too" have very similar pronounciations.

Then you come across 4u in a text and figure out that four, 4, and for sound similar.

Still, that only helps to divide words into equivalence classes. You have no idea how each is pronounced.

Onomatopoeia

A few words in most languages try to imitate the sound they describe. If you find such a word, and become reasonably sure that it is onomatopoetic, that might be the rosetta stone to unravel the rest.

If there really are alien ruins with alien script, and if there are enough linguists working on it, expect some papers suggesting pronounciation. Some might even be right. After a century of academic infighting, they might settle on a "consensus theory" ...

$\endgroup$
  • $\begingroup$ That is very ingenious. Of course you'd have to eliminate the possibility that 'w' on the alien keyboards is next to 'o', That would also cause problems with 'two' and 'too'. Still, I think you're onto something. $\endgroup$ – chasly from UK Aug 2 '15 at 12:59
  • $\begingroup$ @chaslyfromUK, I mentioned that typo thing. Are there any keyboards among the relics? $\endgroup$ – o.m. Aug 2 '15 at 13:38
  • $\begingroup$ I'd like to stick to just a handwritten text but I'm still open to the idea of other relics if absolutely necessary. The keyboard would be an interesting possibility, though I'd have to figure out if they even used keyboards at the time (we're already in the process of migrating to touchscreens after all, and 100 years ago there was no such thing as a keyboard). $\endgroup$ – thanby Aug 4 '15 at 5:00
  • $\begingroup$ @thanby, the qwerty keyboard comes from early typewriters, and key placement is just one option for common errors. Perhaps statistical analysis could account for it, too. $\endgroup$ – o.m. Aug 4 '15 at 5:09
0
$\begingroup$

This is unlikely in the extreme. Humans could not decipher the ancient Egyptian language until the discovery of the Rosetta Stone (which had the same texts in Egyptian and Ancient Greek), and we still have no clue about deciphering "Linear A"; the script of the ancient Minoans. There are other languages which are not clearly translated either; ancient Sanskrit is thought to be full of hidden meanings which we don't understand because the ancient peoples who wrote in this language seem to have been fond of word games and puns; which don't translate well, if at all.

Since alien beings will have alien concepts, there will be very little context to even begin translation. Perhaps the only thing that will be translatable is mathematics (and even then some of the concepts might be very difficult to translate).

So there may be some elements that can be translated, but most of what will be seen will be a mixture of conjecture and guesswork.

$\endgroup$
  • $\begingroup$ Good example with the "Linear A" script and Sanskrit. The Sanskrit point, though, could actually allow me to add flavor to a "translated" text, assuming I can find a way to translate it to begin with, because there could be colorful phrases involved with no explanation that add mystery to it. For the "Linear A" argument though, the total text we've uncovered would fit on a single page, so I wonder if we had volumes of it instead would that make a difference? It also has hundreds of characters and I'm looking at a much more restricted set. $\endgroup$ – thanby Aug 2 '15 at 1:24
  • $\begingroup$ Egyptologists had tens of thousands of examples of Hieroglyphic writing to study before the discovery of the Rosetta Stone, so the number of examples is only helpful in doing things like statistical analysis, rather than translation. $\endgroup$ – Thucydides Aug 2 '15 at 20:53

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.