11
$\begingroup$

In the story I am developing, there is an intergalactic community with a plethora of different sentient species, each with its own languages that vary depending own its vocal anatomy.

There are universal translator implants that act as a brain-brain interface and allow all kinds of beings to communicate because a lingua franca would be limited to only the phonemes present in every known phonetic table from all species, which is none.

Since I decided that each character will speak their own created language, I decided that I would create a phonetic table for all physically possible phonemes for all vocal anatomies in this universe, that will also work as a universal alphabet for transcribing speech.

I just need to know if this is feasible.

I thought about researching animal sounds, but I didn't find any kind of table for animal sounds, as it's not a thing that sparks scientific interest or is of any utility.

Do you know of any means through which I could achieve this?

$\endgroup$
14
  • 6
    $\begingroup$ This might work as a question here, if not then our sister-site Constructed Languages would surely help. I've checked, they've not had this question before (at least not without it getting deleted for one reason or another). $\endgroup$ Commented Jul 13, 2022 at 20:39
  • 3
    $\begingroup$ I am not sure what the average reader would be supposed to do with something like [əɻəɣɯ]. It means "beauty" in Tamil, a language with some 75 million speakers in India. And the word is normally transliterated aḻagu; people who actually know Tamil will know that a is pronounced /ə/, is pronounced /ɻ/ (that's a voiced retroflex approximant), intervocalic g is pronounced /ɣ/ (voiced velar fricative) and final -u is pronounced /ɯ/ (close back unrounded vowel). The old advice holds: don't proliferate, transliterate. $\endgroup$
    – AlexP
    Commented Jul 13, 2022 at 20:48
  • 4
    $\begingroup$ Something more broad then the "international phonetic alphabet"? Yes. just beware that not all species will be able to produce or hear all phonemes. This would suggest a family of langues that use subsets that galactic alphabet. Eg Fred has business level gal3 (galactic 3), and gal5, but can't hear all of gal2 as spoken by <species>. $\endgroup$ Commented Jul 13, 2022 at 20:50
  • 4
    $\begingroup$ Possibly related: How does one approach phonology notation for a non-human constructed language? Just a note: some birds, here on Earth, can reproduce any sound they hear, because their phonatory apparatus works more like a loudspeaker than like a wind organ. You cannot make a finite table containing all the sounds a lyrebird can make. $\endgroup$
    – AlexP
    Commented Jul 13, 2022 at 20:52
  • 4
    $\begingroup$ What would such an alphabet permit you to do? Do you need a way to transcribe speech using a universal alphabet rather than relying on language-specific alphabets? The human version of this, IPA, is most heavily used in research areas where the ability to write down a transcription to analyze it is useful. Do you plan to build a world around language research? $\endgroup$
    – Cort Ammon
    Commented Jul 13, 2022 at 23:34

10 Answers 10

1
$\begingroup$

Given that your interplanetary "phonetics" may not just include speech, but also scent/pheromones, physical gestures, or patches of skin that light up/change colour, your plan to make a universal set of phonemes is doomed either to failure, or to limit the types of aliens your world contains.

A better suggestion would be to crib an idea from an old Chinese emperor: in short, China had multiple spoken languages, but one official written language. The symbol for a word — such as "food", or "house" — would be the same wherever you were, but how you read or pronounced it would differ depending on what region you lived in. Even today, you see the same thing with some aspects of Mandarin and Cantonese.

So, you will have a single written language (which might be ideographic, or might be phonetic), you will have multiple "spoken" languages, eachc compatible with a different subset of species.

Earth languages have between 13 (Hawaiian) and 48 (Nemi and Norman) phonomes. So, for the sake of neatness, let's assume your written language has 50 "phonome" symbols, which can be arranged to form any and every word.

Some species will use phonomes that sound very similar to Earth languages. Some species will use insect-like chirps and clicks. Other species will produce a deep bass whalesong. Or specific colours. And so on.

So long as you know that the symbol you know as "Sh" in mammellian is the same one called "Krr'tlik" in insectile, "Vrrooon" in rumblese or 🟧 in chameleoid, and 🤏 in gesturoid, you — or your Universal Translator — can easily translate it, via the written form.

Similarly, you will all understand the meaning of the three symbols making a sign on the wall in the gym, whether you say it as "🟧-🟫-🟪", "Vrrooon-Hmmgh-Earp", or "Sh-ow-er".

$\endgroup$
34
$\begingroup$

Frame challenge: you are giving yourself a lot of unnecessary work.

Your implanted translators mean that your readers or players don't need to know the details of your languages' phonetics. That table of phonemes only works for spoken languages, and does not help at all for signalling with gestures, colour changes, scent, or radio. All of those methods could plausibly be used by aliens.

Unskilled attempts at constructed linguistics are fairly pointless. They make you look silly to people who know linguistics, and don't usually impress people who don't know about languages.

$\endgroup$
2
  • 7
    $\begingroup$ +1 As a non-linguist avid sci-fi fan, I skip over all that stuff. That's including Tolkien's work. I've re-read it at least a dozen times and have spent maybe a total of half an hour on runes and pronunciations total. Most of that was on the invisible runes scene at the entrance to Moria. It's fun decoration for book covers, t-shirts and such but it wears think quickly compared to what you could be spending time on. $\endgroup$
    – user53931
    Commented Jul 14, 2022 at 19:22
  • $\begingroup$ Overall, if you write a story such that understanding such an iconography is critical to the plot line, you will lose almost all of your readers. $\endgroup$ Commented Jul 15, 2022 at 17:46
12
$\begingroup$

The International Phonetic Alphabet is used to describe sounds present in human languages.

It is organised by place and manner of articulation in the vocal tract. I.e. the sound [p] is a bilabial voiceless stop, meaning it is made using both lips, the vocal cords are not vibrating, and as a stop the lips fully touch stopping the airflow entirely before releasing it again.

All this organisation is only possible because we have a good understanding on the human vocal tract. Similar tables could be constructed for any given phonetic system, but a system that aims to include all possible vocal tracts would have to be so large as to be completely unusable. The only way to objectively represent the sounds produced by an arbitrary vocal tract in a way that isn't incredibly unwieldy would be to use a spectrogram.

The fact that language need not be based on sound at all makes it even less plausible. Even dealing solely with human sign languages the array of possible handshapes, locations, movements, and other gestures is so vast that no equivalent of the IPA (that can meaningfully distinguish any two signs distinguished in any sign language) has been constructed, and even transcriptions dedicated to one specific sign language are difficult or impossible to interpret without detailed knowledge of the language in advance (making them more like English spelling than a phonetic alphabet). If your aliens have drastically different body plans, encoding their sign languages would make it even more complicated.

Other modes of communication e.g. using chromatophores, radio, scent, etc will fall somewhere between these extremes (one where it's practical to transcribe within a given species but not across all possible species without just giving the raw sensory data received by the "listener", and one where even phonetic transcription of a single language is impractical).

So wanting to also include all modes of communication would make the implausible utterly impossible.

Your best bet is to avoid attempting phonetic transcription entirely.

In general I would recommend simply translating the speech of aliens, with some note as to the language and mode of communication, especially when the language is not based on sound. Where quoting verbatim is absolutely necessary (e.g. for proper nouns, words with no equivalent in human language, and potentially poetry being recited) I would give the text in the alien language, either followed immediately by a translation, or in a context that makes it unambiguously clear what the meaning is. Note that this is the approach Tolkien (the father of constructed languages in fiction) takes and even then some readers find the amount of his languages he includes off-putting so attempting to include more is almost certain to backfire.

When you do quote alien speech verbatim, I would rely on an ad-hoc transcription. You should not attempt to include all the phonetic detail of the aliens' actual speech, but instead give an impressionistic understanding of how your audience would describe the sounds. Your aliens may have very different vocal tracts, but if they have a sound (or multiple sounds) that sound similar to a [p], just write them as "p". I would avoid using diacritics (accent marks) or apostrophes as these tend to make (English-speaking) readers more conscious of the artificiality of what they're reading (even though of course these are used in various languages to denote specific sounds).

Your ad-hoc transcription should probably reflect those used historically by English-speakers to write words in foreign languages. These typically use vowel letters similarly to their use in Spanish or Italian, use consonants in their usual English sense, and generally avoid c in favour of either k or s. Sounds not present in English should generally be transcribed according to the closest sounding English sound (or combination of sounds), although certain foreign sounds do have fairly common transcriptions (e.g. the German "ch" sound, the same as the Spanish "j" sound, or Russian "х" sound is usually transcribed ad-hoc as either "kh" or "ch", although the latter may be confused with the English "ch" sound [t͡ʃ]).

As an example, this is article 1 of the Universal Declaration of Human Rights in Hindi, first written in the native script, then in an ad-hoc transcription as I would write it if it were an alien language in a novel, in a scientific romanisation you'd see in a grammar of Hindi, and lastly in the IPA:

  • अनुच्छेद 1(एक): सभी मनुष्य जन्म से स्वतंत्र और मर्यादा और अधिकारों में समान होते हैं। वे तर्क और विवेक से संपन्न हैं तथा उन्हें भ्रातृत्व की भावना से परस्पर के प्रति कार्य करना चाहिए।
  • Anuched 1 (ek): Sabhi manushya janma se svetantra or meryada or adhikaro me saman hote heng. Ve terk or vivek se sampana heng tatha unhe bhratritva ki bhavena se perasper ke prati karya kerna chahie.
  • Anucchēd 1 (ēk): Sabhī manuṣya janma sē svatantra aur maryādā aur adhikārō̃ mē̃ samān hōtē haĩ. Vē tark aur vivēk sē sampanna haĩ tathā unhē̃ bhrātr̥tva kī bhāvanā sē paraspar kē pratī kārya karnā cāhiē.
  • [ənʊtːʃʰeːd eːk | səbʰiː mənʊʂjə dʒənmə seː sʋət̪ənt̪ɾə ɔːɾ məɾjaːd̪aː ɔːɾ əd̪ʰɪkaːɾõː mẽː səmaːn hoːteː hɛ̃ː‖ ʋeː t̪əɾk ɔːɾ ʋɪʋeːk seː səmpənːə hɛ̃ː t̪ətʰaː ʊnʰẽː bʰɾaːtɾɪt̪ʋə kiː bʰaːʋənaː seː pəɾəspəɾ keː pɾət̪iː kaːɾjə kəɾnaː tʃaːhɪeː‖]

I would be extremely hesitant to attempt to transcribe any language based on a medium other than sound. Rather than describing a sign or pattern of lights etc, you're almost always going to be better off relying on the best translation you can give (together with a note that they are communicating through signs or lights etc), even for proper names and words with no clear equivalent in English, as descriptions of these are likely to get overly cumbersome and draw a reader out of the story.

If you want to show off your conlanging beyond this, I would suggest doing as Tolkien did and leaving it for the appendices.

$\endgroup$
7
  • $\begingroup$ There are some big weaknesses to the international phonetic alphabet. It doesn't take into account differences in pronunciation of the same word. It doesn't work for homophones. It's horrible for Chinese or other tonal languages (even though the phonetic alphabet can indicate tones). There was an attempt in the 1900s to romanize Mandarin Chinese and Zhao Yuanren wrote a short story using entirely homophones which is totally incomprehensible without the original characters: theslittyeye.wordpress.com/2011/11/27/… $\endgroup$
    – mgarey
    Commented Jul 15, 2022 at 16:33
  • $\begingroup$ the IPA can encoded any information that is phonemically relevant in any spoken human language. I don't know what you mean about differences in pronunciation of the same word because it absolutely can. The issue of homophones applies to speech identically to IPA and so is no real criticism. Likewise the Shi story is as incomprehensible in speech as it is in the IPA so again this is no weakness of the IPA, which is performing its intended function perfectly. The claim that it is horrible for tonal languages is entirely subjective, it is perfectly capable of rendering them succinctly & precisely $\endgroup$
    – Tristan
    Commented Jul 17, 2022 at 15:10
  • $\begingroup$ > Likewise the Shi story is as incomprehensible in speech as it is in the IPA so again this is no weakness of the IPA But it is perfectly comprehensible as written in Chinese, but would be totally incomprehensible written in IPA. $\endgroup$
    – mgarey
    Commented Jul 18, 2022 at 17:19
  • $\begingroup$ @mgarey IPA is a phonetic alphabet. It only represents the phonetic parts of the spoken language and that is all it aims to do. That it fails to represent additional distinctions made in writing is not a reasonable criticism $\endgroup$
    – Tristan
    Commented Jul 20, 2022 at 9:07
  • $\begingroup$ That's a fair point, so long as it is only used as a phonetic alphabet. There have been proposals to replace all written language with IPA, and that's just ridiculous. $\endgroup$
    – mgarey
    Commented Jul 20, 2022 at 20:20
4
$\begingroup$

Humans have "International Phonetic Alphabet". So wouldn't be surprising if a larger civilization had one (or several) as well. Likely they would need to modify it with human phonemes.

Create it without detail / handwave

As we were learning their language the <alien people> we noticed their dictionary had additional symbols after each entry word. Upon asking they showed us some copy of "intergalactic community phonetic alphabet".

That is all that is really needed in terms of creation. Just a description that it exits. With at some point some additional story indicating how that affects the characters meaningfully. In this case perhaps how the main characters were able to pass this information on to their linguists to map human phonemes to the alien ones for use in automated translators.

$\endgroup$
4
$\begingroup$

Expanded IPA

Nobody knows all the vocal anatomies in your universe. Someone from a virgin planet might assume there are an infinite variety; but your question implies this is not the case.

Still, someone made the table. So we'll assume it's someone from Earth. They started with the standard IPA characters. Then they added two more characters to each and every IPA symbol, massively increasing the diversity, and shoehorning almost every alien sound into this system.

Result: you start with some goofy IPA symbol, like ʟ̝, and then modify it up into a Hangul-like symbol, like 한. Note that you have simplified and stylized the character on the right a bit, to look spiffier on the side of an interstellar freighter, and easier for pirates to read. Most folks in Korea could probably help out with this. You've also gone from merely designating the sound as a "dorsal velar lateral fricative", to designating how many spiracles and throats are involved, whether they beamshape the sound with phase-controlled interference, and whether there are any associated bodily fluids encoding a simultaneous meaning.

The good news is that there are hundreds of billions of planets in the Galaxy, many of them inhabited. Among all those worlds, there might be a billion individuals who feel comfortable writing in IPA, and at least several dozen capable of handling its full expansion to alien language systems around the galaxy, provided they have suitable reference works to consult.

$\endgroup$
4
$\begingroup$

There can be not be an interspecies phonetic alphabet

The reason for this is that not all alphabets are phonetic. Let's take this alphabet here:

enter image description here

Much like the phonetic alphabet, this one is designed to be able to compose any word you like from these symbols. It can work for a lot of different languages... as long as it is a language spoken with your hands. Does not matter if you speak ASL or LSA, you can represent nearly any hand sign using this alphabet. But because Sign Languages and Phonetic Languages are compiled from different communication mechanisms, there is a 0% overlap between where these alphabets could be used. At most you could have a character set much like UTF where a computer can store a lot of different character sets from different languages into a single text string; so, you could have something like this with hand-sign characters tacked in too. You could even add in characters for different light, electrical or smell patterns that other species may use to communicate.

Also, some aliens may be blind. In the the Voyage of the Space Bubble series, there was a neat alien race that could only see via echo location so they were completely blind to written text and 2-dimentional imagery.

enter image description here

Ultimately though, just using the same character set is not enough because even though UTF may allow me to write a the symbol from the Arvekian Smellabet which stands for the smell of something rancid, it does not mean that an English speaker would ever know what that symbol means, much less have the ability to communicate by stringing together a series of smells.

... the good news is that it does not matter.

Translation and image recognition software is is already good enough to make this a non-issue. Instead of worrying about needing an IPA for all species, you just need an expand your idea of what an interspecies translator can do. I will never need to know the characters of the Unified Interstellar Smellabet because when I do need to read Arvekian signage, I will just need the neural implant to interface with my optic nerves to translate what I am seeing as well as hearing.

enter image description here

$\endgroup$
3
$\begingroup$

The IPA is at best marginally a system for all human languages. English and French work mostly fine, but you start transcribing Irish you'll note that Irish has pˠ, pʲ, bˠ and bʲ. Juǀʼhoan distinguishes between b, p, pʰ and b͡pʰ. With enough symbols, IPA can get all these details, but it's clearly biased towards the major Western European languages.

When you take very careful look at English and French, you'll realize that English t is usually dental and French t is usually denti-alveolar, slightly different sounds subsumed under one letter; some languages differentiate between both. I suspect the Japanese wouldn't have separate r and l letters if they had created IPA.

For a known set of vocal and auditory equipment, IPA still has trouble with the number of slightly different sounds that distinguish words in some languages yet are used indistinguishably in other languages. The idea of making something for all the distinctions among all alien life with whatever range of vocal and auditory equipment strikes me as largely futile. Playing with it could be fun, but actually working all the details out is like providing the detailed schematics of FTL drive; pointless and glaringly unrealistic to some readers.

$\endgroup$
3
$\begingroup$

It is, I believe, impossible in theory to create a phonetic system that is both universal and useful.

We have a universal system for transcribing any audible sounds today! In fact we have a few of them - MP3, .wav, .ogg, and pretty much any other audio recording format can represent any sound that any of your aliens can produce. It's difficult for humans to read those files directly, but it's easy to imagine a system of translating some audio format into symbols that can be read by eye. Somebody clever could work out a system where one kind of symbol lets you know the pitch, another lets you know the volume, and so on.

A system like that probably wouldn't satisfy you, because it seems like those symbol-ized audio recordings would contain way more data than we actually need! Surely there has to be a way to strip out all the extra stuff, all the variations in speed, tone, voice quality, etc that don't actually affect the meaning? That's what phonemes really are - big groups of similar sounds where we have decided the differences don't matter and our language treats anything in the group as being "really the same" as anything else in the group.

But you are looking for one single system that works for every type of hypothetical species out there in your entire universe! An individual language probably doesn't depend on most of those things. It would take a lot of brain power to even say a word in a language where the meaning could totally change based on how loud each letter is, exactly how high-pitched the vowels and voiced consonants are, how fast you're speaking them, exactly where your tongue is in your mouth, and so on. But the only unreasonable part of that is putting it all into the same language - there are human languages where the meaning of words legitimately does change based on each of those characteristics, and many more besides!

The International Phonetic Alphabet already has a difficult time representing all the different aspects of human speech that can affect meaning. Not to mention finding a way to ignore the aspects that some languages care about but others don't! Imagine the difficulty of including all these other aliens!

  • Parrot-like aliens that are perfect at mimicking individual voices - a word said in your voice means something different than a word said in my voice. How do you write down which voice they're using?
  • A planet with multiple species of aliens that communicate through something like Morse code - it doesn't matter what sounds they make at all, the only thing that matters is how long the sound is and how much space there is between sounds. Do you need to write down whether they sound to you like "blah blah blah" or "hee hee hee"?
  • Aliens like Groot, who sound like they always say the exact same thing, but really all the meaning is in very subtle intonation differences. Will you be able to write down anything except "I am Groot"?

At that point, you're probably better off using audio recordings for everything, and letting the speakers of the language decide which are the real phonemes and which are meaningless extras.

$\endgroup$
2
$\begingroup$

Not with iconography, but maybe down-sampled waveforms

Having actually attempted this, I can tell you that it's impossible due to the variations in noisemaker shape.

Phonetics are broken down into two areas: consonants and vowels. Consonants can be characterized by what part of the mouth provides the primary surface for shaping the sound. For instance, M is done by the lips, N is done with the pallet, and NG is formed using the back of the mouth.

Within that range, where would you categorize a cricket's chirp? The rasp of an angry tarantula? They bear no relationship to each other, so there's no way to map them to a single schema. You can invent separate characters for each of them, but if you don't know what the creaking of an Aluvian Thermobeast sounds like, then a symbol will do you no good.

That said, what you could do is generate squiggles that mimic the wave form of the sounds. This is really down-sampling, and humans wouldn't have the ability to read such a scheme, even with practice, because it's too detailed.

For more detail on how that might be accomplished, here is a good reference: https://swphonetics.com/praat/tutorials/understanding-waveforms/speech-waveforms/

$\endgroup$
6
  • 2
    $\begingroup$ The visual representation used by real-life phoneticians is not waveforms but spectrograms, which show the intensity change of multiple pitch components over time. That roughly corresponds to how the human auditory system works, in the same way the IPA layout corresponds to how the human vocal system works. And apparently experts do indeed train to "read" spectrograms. $\endgroup$
    – IMSoP
    Commented Jul 15, 2022 at 15:13
  • $\begingroup$ @IMSoP, You're still thinking in terms of the range of sounds that humans can make, but it does provide a good example of the kind of approach you'd have to take. $\endgroup$ Commented Jul 15, 2022 at 20:06
  • $\begingroup$ Not really - a spectrogram is literally just a representation of sound - any sound - in three dimensions (time on one access, frequency on the other, amplitude by line width or colour). You can learn to "read" bird songs from them, for instance. You might need to adjust the boundaries and band sizes to catch the details of something particularly alien, but there's no fundamental limit. I think you could even apply the principle to light and other radiation. $\endgroup$
    – IMSoP
    Commented Jul 15, 2022 at 21:10
  • $\begingroup$ (Ugh, "axis", not "access", obviously.) $\endgroup$
    – IMSoP
    Commented Jul 15, 2022 at 21:24
  • $\begingroup$ Are you suggesting that a spectrogram can differentiate between a sine wave and a square wave, or a buzzsaw wave? If not, then you're really just agreeing with me that this would be useful as a starting point, but need to be expanded. $\endgroup$ Commented Jul 15, 2022 at 22:18
2
$\begingroup$

I just need to know if this is feasible.

Not in a way that would be useful as a sort of "Universal Alphabet" to be slapped under signage like マクドナルド (məkˈdɑnəld) for nonlocals to parse... though it is possible that there might be some "internal" phonetic coding scheme that the translator devices themselves transcribe inputs and outputs with (though that is likely unnecessary as well).


The main problem is that, once we go universal, the goal of having a "Universal Phonetic Alphabet" - that is having some minimum viable character set that folks learn so that they can "spell out" anything they wish to convey to others - becomes unworkable due to the sheer variance of possible phonemes available to alien anatomies.

The goal of IPA is basically to have a single character set that covers all input languages without having multiple characters overlap the same sounds. So, for example, the name of the Greek letter "Γ" is spelt out in English as "Gamma" using the Latin alphabet, and it is spelt out as "Гамма" in Russian using the Cyrillic alphabet. Rather than adding both the Latin "G" and the Cyrillic "Г" to the IPA character set - IPA decided that both sounds are equivalent and should use the same letter. So, great! IPA readers have one less character to learn. And if we continue that process, then rather than learning an entire 26-letter Latin alphabet and an entire 33-letter Cyrillic alphabet, we can learn a single distilled IPA alphabet that is maybe only 39-letters or so instead of the full 59-letters without overlap.

Unfortunately, it is unlikely that this kind of character-space "savings" will continue significantly as we go universal. Clearly, if some species of squid-people speak using only their chromatophores, then there is going to be absolutely zero overlap between Earth-IPA and the Squid-IPA. So there is no benefit to learning the "Universal Phonetic Alphabet" over learning the two distinct languages outright, and the problem only gets worse as more and more species communications enter the fray. So, an UPA learner maybe only needs to learn 100,000 phoneme letters instead of the full 1,000,000 involved in learning each species individual alphabet! While, yes, that is a much smaller number, it is still completely impractical.


Really, though, that isn't a problem if your future setting already has real-time translation devices capable of synthesizing the correct audio/visual outputs. Having these devices translate directly from Language-A to Language-B without first taking a detour through Language-UPA is perfectly reasonable.

That being said, there are a few things that still may happen even if these technologies aren't crazy expensive. Firstly, there are likely to be a small number of languages that become dominant due to being the "language of business/commerce" - so it becomes very likely that signage will have the local language and then a translation in the most dominant trade language in the area.

Secondly, there might be regions where two languages (and maybe a few more) happen to share a considerable phonetic overlap in which case a regional-UPA might actually make sense.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .