— ash nazg durbatulûk —
A Primer in

SF XENOLINGUISTICS

— borag thungg —
1998+ Justin B Rye
(plus 2017 Postscript)


— eep opp ork ah‐ah —
FANTASY EXOTIC TONGUES – An Introduction

If you're here chasing the search‐string “+fantasy +exotic +tongues”, then I'm afraid you've probably come to the wrong place.  This article is a companion to my guides to SF Chronophysics and Exobiology, dealing with another random feature of fictional world design.  It is written in the same spirit as the chronophysics guide: I'm willing to tolerate any level of implausibility in the quasiscience as long as these flaws are irrelevant to the linguistic issues!  (For a definition of this term “quasiscience”, see my Star Trek Rant.)  And I'll try to avoid jargon, since there's no reason a pedantic distinction between (e.g.) inflectional and derivational morphology should be relevant for non‐terrestrial languages.  So for instance the Universal Translators section is not intended as an assessment of how close we are to building natural‐language interpreter machines in real life – it's just a somewhat idiosyncratic survey of the excuses available for SF writers who want to avoid dealing with irritating language barriers.

By the way, when I say “SF” I don't mean to exclude Fantasy.  This is one of the reasons I avoid the word “sci‐fi”: I can claim where convenient that the abbreviation “SF” stands for something nice and inclusive like “Speculative Fabulation” or “Secular Fantasy”!  (But I do now also have a page specifically about Fantasy Elvish.)


— klaatu barada nikto —
LET'S SPEAK ALIEN – In Ten Easy Lessons

Ever wondered how all those traditional space‐opera and epic‐fantasy races – the pig‐faced warriors, the smug bumheads, and all the rest – came up with their wonderfully clichéd alien vocabularies?  It's not difficult; once you've mastered these basic rules, you'll be able to produce names and phrases just as stereotypical as theirs!

LESSON ONE
Languages described as “High”, like High Martian, Old High Vulcan, or indeed High Draconic, aren't from upland regions (as is the case for, e.g., High German) – they're ancient and complicated prestige dialects preserved from the days when the Empire was much bigger and better and more sophisticated.  Speaking them requires considerable effort, dramatic gestures, and often a special capital‐T Talent.
LESSON TWO
Sounds (and sequences of sounds) common in English are still possible in Alienese, but much less common – there are no exotic alien worlds called Stritty or Thudgewundle.  Sounds (and sequences of sounds) entirely unused in English are also very rare in Alienese – no Star Trek character will ever be named Bwäølh or Ngì!  But sounds (and sequences of sounds) uncommon in English are abundant in Alienese; hence the alien races known as the Xeelee, Chirpsithtra, and Githyanki.
LESSON THREE
Initial K is especially popular (Kazon, Klendathu, Krell, K'kree).  Incidentally, there's a good reason for this (and one I'll credit to Steve Mowbray): aliens are obsessed with triangles, a particular shade of green, the number three, and the letter K because they learned everything they know from our TV broadcasts.  To be more specific, from a particular episode of “Sesame Street”.
LESSON FOUR
Aliens enjoy designing their words to look like Latin or Greek, or occasionally Hebrew; they make heavy use of classical sounds spelt in classical ways, such as X, QU, TH, and PH – hence Thranx, Zarquon, Tholian, Cylon, et cetera.  Some, such as the Romulans, Centauri, and Draconians, take it a step further and steal entire words out of Latin dictionaries (or Atlantean TV broadcasts, maybe).
LESSON FIVE
What's more, aliens tend to put classical‐looking endings on their names: ‑ON and ‑OS are particular favourites for planet names (Axos, Auron, Gothos, Krypton), and if there's any sign of females, their names will end in an unstressed ‑A (Thuvia, Belanna, Dua, Ardana).
LESSON SIX
A civilisation of billions of individuals will have no trouble allocating each one a unique, pronounceable name a syllable or two long (e.g. Worf, G'kar, Worsel, Kal‐El, Shevek).  They may even manage to make them all alliterate.  Exceptions to this rule usually have very long names indeed, though there are a few planets where everyone is called Bruce to save time.
LESSON SEVEN
The names of a species, empire, language, homeworld, homestar and so on will all be self‐evidently related; Ogrons come from Ogros, Arisians come from Arisia, Arcturans come from Arcturus, and Humans no doubt come from Humus.
LESSON EIGHT
When the endings aren't pseudoclassical they usually follow the Middle‐Eastern standard: Pakistan‐i, Minbar‐i, Tymbrim‐i, Kimdiss‐i.  Such words often serve both as racial adjective and collective noun, removing the need for a distinct plural; where alien plurals do occur they either end in ‐I (Fyndii) or occasionally ‑N (Thrintun).
LESSON NINE
A name dominated by guttural consonants and sibilants (Cthulhu, Troxxt, Chasch) indicates savagery; one with lots of front vowels and sonorants (Alderaan, Eloi, Emereli) implies a more civilised nature.  Except of course that mysterious gas‐giant races always have names like thunderous farting.
LESSON TEN
If they use apostrophes, ignore them – they're not serious.  Some aliens will try to tell you that “'” stands for an obscure vowel (F'lar, T'pau, Sp'thra), or a silent consonant (Dra'Azon, Ka'a Orto'o), but in reality it's purely decorative.  It's not clear why they choose to use apostrophes rather than, say, umlauts (à la Mötley Crüe) – or peculiar alien squiggles, come to that.  Maybe they just want to keep things convenient for ASCII.

— cthulhu fhtagn —
THE UNSPEAKABLE – And the Unthinkable

If you're the kind of person who read the Silmarillion just for the linguistic appendices (No?  Oh, well, it's only me then), you'd probably prefer your SF languages not to be quite like the ones lampooned above.  So what's the alternative?  Well, I suppose you could use the Simpsons Manoeuvre – to quote Kang: “No, actually I'm speaking Rigellian.  By an astonishing coincidence our two languages are exactly the same!”  But to the best of my knowledge, only Star Trek has ever had the nerve to offer this excuse with a straight face (see e.g. “Bread and Circuses”)… so if that's out, you're left having to imagine a real alien language.  What could that be like?

Well, there are plenty of ways in which alien languages could be extremely unearthly.  The most basic variables are those of FORMAT:

Understandably, writers (including screenwriters) tend to shy away from untransliteratable dialogue – but no such problems arise from variations in GRAMMAR:

And on the traditional third hand, it's commonplace for languages to make some things more and others less convenient to communicate about by means of alternative styles of CODING:

So, is there anything that can't vary?  Well, there's no surplus of hard evidence, but I'd say that all true languages (see footnote) must have in common the following characteristic properties, necessary for any general‐purpose communicative mechanism:

STRUCTURE
Utterances are constructed and interpreted out of modular constituents according to systems of combinatorial rules (capable of creating arbitrarily many novel sentences).  Some proto‐sapient species might in principle do this “primitively”, as some sort of pidgin (cf. Pleistocenese), but a “holistic, impressionistic grammar” with no such structure won't get them far.
DISCRETENESS
Meanings depend on systems of distinctions between finite sets of elements (sounds, words, grammatical forms, and so on), not on the subtle shadings of their individual qualities.  Continuous variables such as stress may be useful for expressing generalised attitudes and overtones, but they're no good for specifics.
CONVENTIONALITY
Words are arbitrary labels, not representations of what they denote.  The word “giraffe” isn't particularly giraffelike, “big” is a small word, and dogs don't really say “bark”.  Even in sign‐languages and pictographic writing systems, few meanings are guessable from their signs – aliens may use sonar onomatopoeia, but it won't make their grammar any more comprehensible.
METAPHOR
The “literal‐minded” languages of many space‐opera Ancients are impossible; all linguistic categories and rules are formed via analogies, explicit or implicit.  The definition of the word “dance” presupposes a resemblance between waltzes and raves – and pluralising “days” is a fossilised metaphor too: when did you last see a stack of them?  On the other extreme, the “allusive language” of “Darmok” (see Star Trek Rant) is unworkable because it's all metaphor and no grammar.
ABSTRACTION
Phenomena can be discussed other than those directly apparent to the senses, including spatially or temporally remote events, hypothetical or generalised situations, counterfactual fantasies, and lies.  The speakers may have trouble with them, but the language will always provide for “saying the thing that is not”.
CONTEXTUALITY
Meanwhile, phenomena that are directly apparent or previously established can be referred back to by means of special shortcut forms – pronouns and point‐of‐view‐dependent expressions as in “you were behind them”.  Aliens with no way of expressing the first person (even as “this person now speaking”) are unlikely – they'd need unique absolute identifiers for every person, place, moment, and event!

Then again, while I'm sure about xenolinguistics being a branch of linguistics, can I entirely rule out the idea of intelligent lifeforms who simply don't have language?  The usual suggestion is that instead they have some sort of (telepathic or biochemical) “hive mind”.  Well, maybe.  But unless they've got some standardised way of encoding concepts for portability from brain to brain, it's not going to be enough to distinguish them from animals – and if they have, that's essentially a language by another name.


— zog —
UNIVERSAL TRANSLATORS – A Buyer's Guide

Okay, so you walk into the spaceport bar and discover that nobody within a kiloparsec speaks English (or whatever it's become by the year 3000 AD).  You may think you'll be able to tell what that Kzin is saying just by the tone of his voice; you may think you can signal your friendly intentions by showing your teeth a lot; you may even hope he'll speak your favourite interlanguage (cf. [the late] Don Harlow's notes on Esperanto and Science‐Fiction).  But no, take my advice: it's time to invest in a Universal Translator system (henceforth “UT”)!  And beware of dodgy characters trying to flog second‐hand 3PO units or fishy‐looking implants; here are some guidelines to help you avoid wasting your credits on something likely to get you lynched or brainwashed.

For the convenience of any alien readers (especially Gubru, Ramans, and the like), all the points made come in sets of three.

“UNIVERSAL”
There is some room for flexibility here – you needn't splash out on a Star‐Trek‐style model that can handle any number of unknown alien tongues at once, as long as it doesn't specialise in, say, Basque‐to‐Tamil.  Does it need to be able to cope with thoroughly alien mindsets, non‐auditory languages, and so forth, or is everyone in the bar tediously humanoid?  And when you encounter a language it doesn't already have on file, how does it learn new ones?  Possibilities include:
  1. Plug‐and‐Play – “canned” languages, which you can upload into the UT or your own brain.  Note the social implications if you can learn Xemahoa on a whim, Gothic as a fashion statement, or a private cryptolang to baffle eavesdroppers.  Just be careful with black‐market language tapes; don't buy any that claim to be “doubleplusgood”.
  2. Exchangers – including “handshaking” computers that swap lexicons on contact (do you really want to give away security‐risk terms such as “hypnosis” to unknown aliens?), and psychic “language chamaeleons” that can reply in any dialect they encounter (be careful not to use “royal we” back to God‐Emperors).
  3. The Hard Way – language learning by prolonged interaction with cooperative native speakers.  Unless you've got magical assistance, don't expect mere minutes of eavesdropping to help.  Even if the locals point at a spaniel and say “That's a dog!” you can't be sure they mean “dog = Canis familiaris”; it could be “dawg = barking” or “tsäd = Fido”!
“TRANSLATOR”
When it comes to the “three levels of translation” it's worth being clear about your requirements:
  1. “Literal” or word‐by‐word translation is less useful than monoglots tend to imagine; the intelligibility of the results is proportional to the relatedness of the source‐language to the target‐language.  This is hopeless for Syrians, let alone Sirians – if you pass that first line through an online translator a couple of times, you get: “literal” or the word for the translation of the word is little use, of that monoglots if inclines to present itself.
  2. “Official” or phrase‐by‐phrase translation is organised by legal conventions about equivalences, and restricted to subjects with narrow, codified jargons.  If you're an interplanetary lawyer or civil engineer, this might be adequate; but don't expect it to convey subtexts.  Actually, unless your UT device is ridiculously good it's always wise to steer clear of fancy figures of speech, such as jokes or irony (“do I look stupid?”) – be literal and tolerant of apparent threats, insults, and the like.
  3. “Psychological” or concept‐by‐concept translation is the ideal, producing exactly the same effect on a speaker of the target‐language as the original would on a speaker of the source‐language.  This objective is next to impossible for anything less expensive than a trained human‐equivalent brain.  When you're talking about your family life to a Sontaran warrior‐clone, syntax, idiom, and social background knowledge blur into one another as things that need to be translated; so how much are you willing to pay for?  A “shallow” UT confuses “Pat owns an orange pelt” with “Pat has red hair”; better models can deal with slang, xenoethnological trivia, and allusions to Oolon Colluphid.
MECHANISM
Never mind the questions “What is it?  How does it work?  Where is it?” (to quote someone else's (ex‑)rant); all that matters is that there are three broad categories of UTs, each of which has its own pros and cons.
  1. Polyglottisers – mechanisms by which one or both participants can come to understand all of the languages involved (i.e. either a cyberpunk plug‐in language‐chip or a psionic/magical Pentecost Effect).  The drawbacks of this are that if you become a monoglot again afterwards, you're left with baffling memories (if only “Why was that funny?”), and if there's any truth whatsoever to Whorfian Relativism, exotic languages may influence the decisions you make thinking in them!  The effects of human tongues such as Hopi are arguable, but UTs capable of making you take as natural the conversational instincts of a radar‐using methane‐shark shaman are another thing entirely.
  2. Psi‐Dubbing – reads what a speaker is thinking and provides a voiceover, itself preferably telepathic.  Alien brains may turn out to be unreadable, or all minds may prove to be readable regardless of native tongue – it all depends on whether there's a universal nonlinguistic “language of thought” for the Psi‐Dubbing to work in (a very Chomskyan thing to imagine).  Unfortunately, if it works it may outflank all efforts at diplomacy – at any rate, it sounds like a monstrous breach of privacy – and it requires improbable technology such as psionics or neurotelemetry.  Farscape's playful suggestion of “translator microbes” is about the most plausible version I've heard of!
  3. Cyberinterpreters – “expert system” translators, with robot bodies.  Audio‐only “Pocket UTs” wouldn't be able to handle situations as simple as a trip to a Spanish grocer's: the correct rendering of “I'll have that one!” depends on whether it's nearby and feminine (¡ésa!), distant and masculine (¡aquél!), or whatever.  The body needn't be humanoid, but any decent Machine Translator would have to be such a flexible and intelligent AI that it deserves civil rights (I pity C‒3PO, kept as a slave translator for biochauvinist rebels in a society where everything understands English anyway).  Nonetheless, you may have to slow down to allow your interpreter to keep up – a good one may start translating your sentences “incrementally” before you finish, but anything approaching a “simultaneous” translation takes an awful lot of processing power.  So if your UTAI starts speaking at the same moment as you do, shut up and let it do the talking.
FEATURES
Check how good the system is at coping with mismatches between languages in the following fields:
  1. Vocabulary – if one language lacks an idiomatic match for the other's expressions, UTs may paraphrase or neologise to fill the gaps (“he contracted blue fever from a jabberwock bite”).  More problematic are partial matches: we say “cousin (parent's sibling's child)”, they say “cousin (relation of the same moiety and age‐grade)”; they say “water (dihydrogen monoxide)”, we say “water (salt or fresh, but always liquid)”.  The classic example of this problem is the inventory of basic colour‐terms: Russian discriminates between “light blue” and “dark blue”, while Hanunóo uses a single term to cover the entire green/blue end of the spectrum.  And as for Jovians…
  2. Agreement – when a Betelgeusian calls you {addressee‐agentive‐adult‐nondistributed}, the UT throws away as irrelevant all the surplus grammatical features and focusses on its function, parallel to English “you”.  Then when you in reply address the Betelgeusian as “you” the UT has to somehow come up with the extra details the alien agreement system involves.  This kind of routine trimming and padding takes a good deal of creative fudging – especially if there's a risk the Betelgeusian might later turn out to have intended you to pay attention to some of the discarded “trivial” details.  The UT could play safe and insist on conveying every last ambiguity and nuance explicitly, but this is excruciatingly difficult, not to mention distracting (exercise: try paraphrasing the precise differences between “he ate the biscuit” and “she has eaten a cookie”).
  3. Implicatures – ordinary communication relies on speakers obeying a set of “conversational maxims”.  They shouldn't say things that are either uninformative or self‐evident; that are inaccurate; that are irrelevant; or that are either otiose or ambiguous.  Of course, people aren't always apposite, reliable, etc., but the interesting point is that infractions are often themselves communicative: if I say something blatantly inappropriate, it usually means there's a subtext to be found.  The trouble is, aliens are likely to have different conventions about such things, and inhuman intuitions about what needs to be pointed out (“You're very tall…”), what level of hyperbole is acceptable (“Nobody ever comes this way”), what's relevant (“Are you hungry?” – “It's daytime!”), and what's concise (Ents and Vorlons never get along).
INTERFACE
Ensure the UT's output is appropriately customisable – assuming it has output; if the UT's skills are seamlessly integrated into your own mind, you may not have these options.
  1. Confidence – if it has to guess at a translation, does it plough on with fingers crossed, flash unintelligible warning lights, or constantly interrupt with questions?  Backchat can be annoying (“Is that and/or or either/or?”, “You realise that's not answering the question?”), but it's the best way of dealing with errors when they do occur.  Internalised UT systems don't pose these problems, but external ones make great scapegoats…
  2. Anthropomorphism – does it credit you with a lot of background expertise (talking of “zitidars”, “foreclaws”, and “spoo”) or recast everything patronisingly into familiar analogies (“elephants”, “thumbs”, “apple pie”)?  And what style of analogy does it use – does 144 baph‐ʾl‐ghab equate to “503 km” or “a hundred leagues”?  Does gobl‐digûk become “many mooncycles” or “several of your Earth years”?  Or does the UT just convert everything into Galactic Standard hexadecimal Planck units?
  3. Diplomacy – should the UT respect linguistic taboos and conversational etiquette, or translate insults as insults?  If your UT can cope with multiple stylistic registers, it can be set to convert between them and filter out (or enhance) the profanities.  Come to that, any UT that can translate an oratorical welcome can be expected to summarise it too.  There's no need to reply with a formal speech of gratitude; simply configure the translator to turn your terse colloquial English into polite long‐winded Vilani.  Etc., blah, waffle.

— ack ack ack, ack ack ack ack —
CETI FOR BEGINNERS – Little Green Manuals

You may or may not be surprised to learn that some extremely serious attempts have been made to specify in detail the best ways of opening up communications with Extra‐Terrestrial Intelligence, either in radio broadcasts or on the White House lawn.  To start with the classic example:

On LINCOS – an interstellar auxiliary language

Cosmic Intercourse – LINCOS v2

And here are some other TEFLETI projects:

Me Human, You Alien – How to Talk to an Extra‑Terrestrial

Solresol, a musical language – remember Close Encounters?

Personally, I say if they can't be bothered to work out in advance how to say “Take me to your leader” then they can't be trusted to drive unlicenced starships around in our atmosphere…  Blast them out of the sky before somebody gets hurt!


2017 POSTSCRIPT – Ack, Nack, Googlewhack

If only we'd had universal translators as good as Google Translate in the mid‐nineties it might have saved us from the tragic interplanetary failure of communication documented in the movie “Mars Attacks”:

[ack ack (x8) = I'm sorry, you're welcome, you're welcome]