Language Design for Sound Designers
September 25, 2007
   
 

In Twin Peaks, David Lynch utilized phonetic reversal to help create the language of The Black Lodge, a timeless plane occupied by a giant, a dwarf, and a doppelganger. In order to achieve this, the actors had to learn to speak in reverse.


Language sound design is a weird audio craft that few sound designers have gotten the chance to explore. Trying to find documentation on it is rather like figuring out how to win a game of cheese-rolling, the hugely disregarded suicidal sport of Goucestershire. You can pick up a few facts here and there, but as it has not really been acknowledged as a skill you will find few books that illuminate how to do it. Nonetheless, some of the most interesting sound design works have incorporated the use of constructed languages (commonly referred to as conlangs). In this article, I would like to present information on how to invent your own languages from the angle of character sound design. Linguistics is a relatively deep subject but as you will find, you need not get buried far in the principles of language construction to bestow its powers in your sound design work.

Conlangs are languages that are invented by creative linguists most often for use in fictional stories. The Klingon language, designed especially for the Star Trek universe, would be an example of a conlang. Language sound design is a term that I am using to represent sound design work that utilizes a conlang. The most famous examples of language sound design encompass the seemingly endless assortment of non-English speaking alien characters in Star Wars. There is also a grey area in-between. For example, English dialog that is manipulated to the point of sounding non-human, like the voices of the Black Lodge in David Lynch's Twin Peaks. Audio specialists may or may not have involvement in creating a conlang as these tasks are normally handled by linguists and writers. Be that as it may, a sound designer will approach language invention from a very different angle than a linguist, using microphones, signal processors and high tech audio mixing techniques. Language design by sound designers is a whole different ballpark that deserves serious creative investigation. Having knowledge about conlang design will undoubtedly help you to create outlandish character voices for your projects.

When approaching a sound project, I make sure that my technique is not driven solely by the best examples in entertainment. So, for instance, if I am going to create the sound of a diving airplane, the last place I would begin my research is by listening to plane crashes in blockbuster films. This would not give me a clear understanding of the physical sounds I am aiming to emulate and it certainly would not inspire me to design something original. Therefore, in the case of language sound design, a good place to look for ideas might be plain old linguistics.

It is generally understood that without language we would not be able to accumulate and synthesize ideas, pass knowledge to our children and progress as sophisticated beings. Language has helped us become, at least in our own opinions, the mightiest creatures. The strange thing, especially to an evolutionist, is that the animals are not coming up with communication systems quite like our own. We have never been able to explain how and when language began but there are two common theories.

   
 
In Mandarin Chinese the end of the page is the bottom. Likewise, their sense of time metaphorically flows downward.

About half of North America believes that our bodies and minds developed over time from physical substance that matured into living organisms, which became increasingly more intelligent and learned to speak. They tap their forefingers at irrefutable facts in science books to provide proof. The other half of the country looks to a different set of books, in confidence that all living creatures, consciousness and language were created over a short period of time by a deity. This group wields "faith" which has the power to automatically cancel out "proof". Arguments from both sides regularly fill web pages and volumes of books.

Interesting is the impact that language has on our sense of time. This becomes obvious when we look at the spatio-temporal metaphors in languages. For example, in Mandarin Chinese the future is down and the past is up, which is evident in the top-to-bottom direction of their written text.

The indigenous Aymara people of South America, actually have a reversed concept of time, meaning that the past is before them and the future behind. In Aymaran, the word "nayra," means "front," or "past" and the word "qhipa," means "back" or "future." When they make gestures about the future, they point in back of themselves and vise versa. Rafael Nunez, associate professor of cognitive science at the University of California, San Diego indicates that this was passed on through language:

"Until now, all the studied cultures and languages of the world – from European and Polynesian to Chinese, Japanese, Bantu and so on – have not only characterized time with properties of space, but also have all mapped the future as if it were in front of ego and the past in back. The Aymara case is the first documented to depart from the standard model...These findings suggest that cognition of such everyday abstractions as time is at least partly a cultural phenomenon."

In his book, "The Spell of the Sensuous", David Abram attributes the development of the phonetic alphabet to a growing disconnection between humans and the natural world. He determined that early cultures of the earth lived directly through legends and songs, connecting them intimately with the animals, landscape and rotation of seasons. Communities learned how to succeed and live happily through stories about their own surroundings and daily events. Eventually, as civilizations became more complex, scripts were used to record these experiences which dramatically altered people's perceptions. When history was recorded, time was no longer cyclic but laid out along a line pointing towards the future. Ultimately, the phonetic alphabet was used to define ones experience, severing mankind's primordial connection with the earth and the rest of it's inhabitants. Human awareness receded into private worlds.

 

For The Fifth Element Director Luc Besson and actress Milla Jovovich designed "The Divine Language" which consisted of about 400 words.

In modern society, experience is nearly impossible outside the realm of words for we have taught ourselves that all things, actions and feelings must have names. Through the mechanism of language, beliefs and emotions are classified before they are identified, with every human condition fitting neatly within an enormous webwork of associated terms. We have words like "physical" and "spiritual" which we think of as opposites because that is the way we have learned to think of them in our culture. Language gives us a codified understanding of reality in which we are all very convinced. We have a special talent of collecting and processing all of these words and meanings which science can not fully explain. Many Scientists are assured that we acquire language instinctively. Dr. Lise Eliot, a Neurobiologist and expert on early intelligence states, "The reason language is instinctive is because it is, to a large extent hard-wired in the brain. Just as we evolve neural circuits for eating and seeing, so has our brain, together with a sophisticated vocal apparatus, evolved a complex neural circuit for rapidly perceiving, analyzing, composing, and producing language."

Cognitive linguistics claims that we are not born with any autonomous language system and that the acquisition of language happens on a conceptual level, as a natural way of forming ideas. This school of thought indicates that perception is observer-independent. Basically this theory states that things do not exist separate from thought, that the thoughts are the reality. This is a hard one for most people to swallow and is generally accepted as fiction.

In the book "Tlön, Uqbar, Tertius Orbis", Jorge Luis Borges describes an idealist society on the planet Tlön that creates its world by imagining it, each citizen creating an individual reality, incongruous to any singular physical truth. The language they speak has no nouns.

"The nations of that planet [Tlön] are congenitally idealist. Their language, with its derivatives -- religion, literature, and metaphysics -- presupposes idealism. For them, the world is not a concurrence of objects in space, but a heterogeneous series of independent acts. It is serial and temporal, but not spatial. There are no nouns in the hypothetical Ursprache of Tlön, which is the source of the living language and the dialects; there are impersonal verbs qualified by monosyllabic suffixes or prefixes which have the force of adverbs. For example, there is no word corresponding to the noun moon, but there is a verb to moon or to moondle. The moon rose over the sea would be written . . . upward, beyond the onstreaming it mooned." (Borges 264)

The inhabitants of Tlön would find our way of thinking just as strange and disorienting. When we grasp a concept and live by it, like time moving forward through space, the idea envelops us like shell that we can not see through. Our concept of reality has been passed down through the centuries to us through language. When we imagine the way beings from other planets may think, we should not assume that they would share these human concepts that we have been inheriting from our ancestors.

 
Weilgart claimed that an alien taught him aUI, the language of space.

In the 1960's a psychiatrist by the name of John W. Weilgart claimed that in his youth he encountered a little green man from outer space. The alien taught him a language called aUI (pronounced "a-OO-ee") which it said is a universal language that all intelligent beings in the cosmos speak. Weilgart customized the language so that earthlings could speak it and published it in a book, promoting its ability to make people think more logically. Whether or not there is any validity behind Weilgarts story, the language is interesting in that it features an oligosynthetic structure. Meaning, primitive sounds are grouped together to form statements so that a single world can contain a synthesized idea as complex as a sentence. The language consists of 42 phonemes, which are basically just upper and lower case letters in our alphabet, each of these phonemes having its own meaning. For example, the three letters "a-U-I" have the meanings ""space-mind-sound" consecutively (meaning "the language of space".) In another example, the word kEged means "up-matter-movement-through" which in English means an airplane. Unlike the English language in which a word can mean multiple things based on definitions written separately in books, in aUI a word can instantly be broken down into its constituents so that there is little question about what is being communicated. Perhaps a civilization of aUI speakers would be very cooperative, believing that anything can be achieved through the mixture of minds and primordial energies.

When we design languages for characters, especially alien ones, we should consider the way that their language might be a reflection of their culture and perceptions. If they write their text in circles or spirals, for example, they might not believe in a finality of death, rather that life and destruction are cyclical and unending. These types of details may at first seem unimportant to language sound design but the fact is that characters will be much more believable and interesting if they are created with cohesive concepts. Languages should be constructed as an aggregate of the character.


Phonetics

Science fiction films usually feature language sounds based on the phonetics that a human voice actor can produce (e.g. Ewok). More revolutionary approaches may require combinations of animal sounds (e.g. Wookie) or perhaps a combination of human sounds and electronics (e.g. R2D2). Whatever the case, it helps to have an understanding of how an intelligent being would physically form speech sounds. The answers are right under our noses.

In order to speak, you push air out of the lungs, up the trachea and through the vocal folds in your throat (often called the voice box). The vocal folds form a muscular slit that can open and close in a highly controlled manner. When the folds are open, air passes through quietly, when they are almost completely closed they vibrate and produce sound waves with a pitch determined by the tension of the folds. (You can replicate this system with a balloon by blowing it up with air then squeezing the end as the air escapes. The tighter you squeeze, the higher the pitch.) Sound waves then travel through your vocal tract which consists of a number of organs that you can articulate to form words. These organs include your lips, teeth, various points along the tongue, and the alveolar ridge which is just behind the front teeth. The vocal organs can completely obstruct the flow of air, forming consonant sounds, or redirect it to the nasal tract. Phoneticians utilize an exhaustive list of terms to characterize vocal gestures. Here are some common descriptors for consonant sounds. If you find this list to be confusing, rest assured that you are not alone:

Alveolar: Tip of the tongue against the alveolar ridge (nest, day, list, zoo, take, say)
Bilabial: Both lips together (ban, mat, pie)
Labiodental: Lower lip against upper front teeth (foe, vat)
Dental: Tip of the tongue against upper front teeth (the, thick)
Paletal: Front of the tongue against the palate (year)
Palato-Alveolar: Front of the tongue against the back of the alveolar ridge (shoe)
Velar: Back of the tongue against the palate (hose)
Retroflex: Tip of the tongue against the back on the alveolar ridge - not used in English
Stop: The articulators have completely closed off the flow of air (put, ten, kill)
Nasal: Air is directed through the nasal cavity (men, nest, gang)
Fricative: Air passes through a small opening between articulators (fist, thin, sit, shell)
Approximant: Air passes through a medium opening between articulators (yes, will)
Trill: The tip of the tongue rolls against the alveolar ridge - Seldom used in English

Vowel sounds are created primarily by the position of the lips and the top of the tongue, the flow of air being unobstructed. There are numerous reflections and resonances happening within any vocal tract which are regulated by the size, shape, and position of all the vocal organs. Sound traveling through the vocal tract, especially the overtones of the sound, are filtered down by the position of these vocal organs creating particular overtone qualities called formants. One way to hear how the vocal cavity does this is by whispering "e a i o u". This allows you to hear the shape of the vocal cavity without the pitched waveform of the voice box. Vowels can be spoken at any pitch but what makes them unique from each other is the shape of the vocal tract.

 

David Hasselhoff went into my nose and out my mouth in the first ever intranasal sound design experiment.

To learn more about the way sound changes as it travels through the vocal tract, I did an experiment and put a piezo speaker in one of my nostrils (do not try this at home, children). I was curious to see if I could send a waveform through my nasal tract and back out my throat. In theory, this would allow me to introduce a formant filter on the waveform. I used perhaps the only piece of music that I was willing to invite into my nose, David Hasselhoff's Hooked On a Feeling. Although my nasal cavity dampened the amplitude quite a bit, It actually worked. When my mouth was closed, the music was entirely silent. When I opened my mouth, David Hasselhoff resounded. I could make the music sound very different and vocal-like depending on how I positioned my tongue and shaped the opening of my mouth.

When an English-only speaker (like myself) imagines all the phonetic sounds that can be used to form a language, he or she can not help but linger within the confines of English. English is comprised of about 25 different gestures and it is the strict adherence to these rules that makes the language sound so unique. It is difficult to describe sounds that do not occur in English without using phonetic transcription, a set of symbols that describe how languages use the mechanisms of speech. As this set is rather large and time consuming to learn I do not recommend using it just for the sake of language sound design. The best way to expose yourself to different language sounds is to listen to them. Peter Ladefoged's book, A Course in Phonetics, features a CD-ROM filled with language sounds from all over the world, an excellent way to get more familiar with ethnic tongues.

Human phonetic capabilities cover a range larger than any single language. Zulu is one of about 30 African languages that includes clicking sounds, one sounding much like our informal expression of disapproval, another made by pressing the lips together flat and making a sharp kissing sound. Sihndi, a language spoken in Pakistan and India, features "implosive" gestures which, unlike most vocal sounds, are made by pushing air down into the larynx. This is not common in English but you can get a feel for the gesture if you say the word "onion" or make the sound "ayn" while pushing some air back into your larynx.

 

A music teaching method called Sonic Sol-fa was featured in Close Encounters of the Third Kind as a means to communicate with Aliens.

The way a word is pitched is called its "tone" and when we speak in sentences all of the words travel continuously up and down in pitch, creating what is called "intonational phrase". The intonation and tone can convey information paralinguistically, revealing such information as the speakers gender, age, emotional state and syntactic information that can, for example, turn a sentence into a question, a kind remark, or a scathing insult. Paralingual cues are usually cross-cultural and do not belong to a specific language system. However, in some languages such as Chinese, the precise tone of the words will change their meaning entirely. For example, the Chinese sound "ma" can mean four different things: mother, hemp, horse or scold, depending on the tone. Languages like these are called "tone languages".

Another characteristic that differentiates languages is the voice onset time. This is the time between the release of a stopped sound and the start of the voicing. For example, Navajo features a very long voice onset time in words such as "wait" which is pronounced like "t - ah", and "arrow" which is pronounced like "k - ah". These little bits of silence that occur mid-word makes the language sound periodic and choppy. French is quite the opposite as the stopped sounds are nearly always voiced, creating a more flowing, romantic sound such as in the word "châteaux".

 

 

Getting a feel for vowels in other languages is a bit more tricky than consonants, partly because our minds tend to categorize unfamiliar vowel sounds into the closest English facsimile. Danish, for example, features nine vowel sounds, but an English speaker trying to speak Danish may not even acknowledge this. Apparently, the Dutch find this rather humorous and enjoy listening to foreigners say such things as "Rødgrød med fløde" (Red porridge with cream), because nobody but the Danes can pronounce all the vowels properly. Give it a try!

There are 28 different vowels listed in the International Phonetic Alphabet (IPA), a standardized phonetic notation system used by people working with foreign languages. Conveniently, Wikipedia allows you to hear samples of them all here. Listening to them should inspire you with ideas when designing the vowel characteristics of your languages.

People usually prefer to hear beings that sound somewhat anthropomorphic and it is best to bring a voice actor in for this, preferably someone that has had experience performing character dialects for animation. The important thing is to try and broaden your notions of how fictional beings could communicate with sound. Perhaps they do not use the same organs they eat and breath with to form language like humans but instead scrape and rattle their claws over exoskeletons. Maybe they use a single elongated appendage that serves the dual function of picking things up and communicating (similar to elephants). If the being does not have a tongue then they might manipulate fingers or tentacles over their lips to help make a wider range of sounds.

If you are taking a more adventurous approach, please visit my article on Creature Sound Design where I cover lots of creative and technical methods for making non-human communication sounds with recordings of animals. That web page should give you some ideas for alien vocal tracts, alternative types of talking and most importantly how to go about designing the sound the you want.

Languages of Fictional Worlds

Language is so integral to culture that a linguist can reconstruct a culture from its language, just as a biologist can reconstruct an animal from a bone. Tolkien's production of the invented language called Quenya compelled him to depict the Elvish cultures of The Silmarillion. (Noel 3)

 
Tolkien invented scripts such as Tengwar to write his languages. Many beautiful Middle Earth fonts are available here.

The author of the Silmarillion, J.R.R. Tolkien, was an Oxford professor of Anglo-Saxon language (Old English) and a philologist. He began inventing languages in his young teens, at first as an amusing linguistic game with his friends, which often amounted to saying words in reverse or swapping word definitions. Then he took it unto himself to design languages using the building blocks of Latin, Gothic, Spanish, Greek and Italian. He was always on the hunt for beautiful words and submerged himself in the languages that caught his attention. So it was that he came upon Welsh, his first encounters just a few phrases on station-signs, coal trucks and date stones. In these fragments he found "a flash of strange spelling and a hint of a language old and yet alive". In 1911 he enrolled at Exeter College, Oxford, where he studied Old English, German and Welsh. In November 25th, 1911 at the Exeter library he discovered "A Finnish Grammar" which he stated was one of the most important steps in the development of his sentiment and taste of language. "It was like discovering a complete wine-cellar filled with bottles of an amazing wine of a kind and flavour never tasted before. It quite intoxicated me." At the time of his graduation in 1915, World War I was enveloping Europe. Tolkien enlisted for duty, was commissioned in the Lancashire Fusiliers and was sent to the Western Front. After four months of the most mechanized and brutal combat the world had ever seen, he caught trench fever and was sent back to England. Upon his return, Tolkien continued to invent languages more seriously than ever, and soon started work on a language that he thought would be spoken by Elves, called Quenya, which was based on Welsh and Finnish. In order to give his languages a home he began writing mythopoeic works about the place the the Elves inhabited, Middle Earth, the realm of The Hobbit and The Lord of the Rings. Within the stories of Middle Earth, he formulated a tree-like evolution of fifteen languages and dialects.

In "The Lost Road and Other Writings", a compilation of posthumously published Tolkien work, Christopher Tolkien described how his father created the languages of Middle Earth:

"He did not, after all, invent new words and names arbitrarily: in principle, he devised them from within the historical structure, proceeding from the bases or primitive stems, adding suffix or prefix or forming compounds, deciding (or, as he would have said, finding out) when the word came into the language, following through the regular changes of form that it would thus have undergone, and observing the possibilities of formal or semantic influence from other words in the course of its history. Such a word would then exist for him, and he would know it. As the whole system evolved and expanded, the possibilities for word and name became greater and greater."

The prologue for Peter Jackson's highly successful film adaptation of The Lord of the Rings opens with total darkness and the whispered voice of the Galadriel, Lady of the Wood, speaking softly in the language of Elves. Following each phrase is heard the English translation.


The world is changed


I can feel it in the water


I can feel it in the earth


I can smell it in the air

 

Tolkien's linguistic work planted the seed that became the Elves. Likewise, the dimensionality of Middle Earth grew from a geneology of languages, the names and voices of its beings, its myths, songs, and forests all reverberating within a cultural landscape. This is an example of the way good movie directors can create worlds larger than the narrative, convincing the the audience that things exist beyond the edges of the photographic scenery.

The Star Wars universe is similarly complex. The combined efforts of book authors, video game developers and a booming fan culture have produced a world infinitely larger in scope than what is represented in the films. Because of this, the history of the alien races, their social structures and interactions have been documented in great detail, subconsciously re-enforcing the believability of the languages.

 

One of the ingredients used to create Poggle the Lessers voice was the sound of Fruit Bats circling around a microphone.

Appearing in Star Wars Episode II: Attack of the Clones are a race of insect-like beings called the Geonosians. Ben Burtt designed the voice of one Geonosian, Poggle the Lesser, using vocal sounds made by Marton Csokas which he mixed and sequenced with strange clicks and honks, the end result being an unintelligible but fantastic sounding insect voice. Tone and texture portray personality and emotional states, much like Chewbacca's voice, and so that his "words" are understood, when he speaks subtitles appear. I came across a Star Wars fansite featuring one enthusiast's analysis of the language. He transcribed the sounds into a dictionary of words by meticulously comparing the sound effects to the subtitles, absolutely assured that the language existed as a functional system. It is obvious to me that there would be no point in creating a phonetically-ordered language for such a overly-designed voice.

Consider all the avenues that you can take as a sound designer to construct languages. A fully devised language having thousands of words like Tolkien's could be a monumental effort and it is fair to say that most sound designers will avoid playing the part of a hard-core linguist. If the constructed language features enough paralingual cues or some character translating them out in English—"What is it girl? Is Timmy stuck in the well?"—or even subtitles then the language does not have to be ordered and functional. In these scenarios sound designers walk the grey area between language and sound design, making linguistic decisions based on aesthetic taste.

When working on any sort of multimedia project, I try to consider all the possible shortcuts I could take to get it done more quickly and efficiently. Some of these clever schemes come at no cost to creative quality, save an enormous amount of time and often make the project come out better. I am talking about the benefits of simplicity. You might want to consider creating languages with an absolute minimum word count. Sonja Elen Kisa, a 28 year old student living in Toronto invented a language called Toki Pona. She designed it in an effort to make a language that encourages people to say nice things (she has my applause). Being a true product of minimalism, Toki Pona has a modest 14 basic sounds and 118 words. It is starting to get quite popular now because it sounds really fun and one can learned it in just a couple of hours.

In order to get more mileage out of the words in Toki Pona you combine words together in creative ways. Below is a list that she has on her website (www.tokipona.org) to demonstrate how you can create words that are not featured as part of the language. The words in the left column are not in the Toki Pona vocabulary, the words in the middle column are combinations of words do exist in the vocabulary and the right column features the resulting phrase according to the official word list:

     
alcohol crazy + water nasa telo
bed sleep + surface ape selo
friend good + person pona jan
geology earth + knowledge ma sona
girl little + woman lili meli
happy feel + good pilin pona
restaurant food + house moku tomo
teach give + knowledge pana sona
when? what + time? seme tenpo

 

For language sound design, it is essential to fine-tune your vocabulary to fit the project. For example, if you are developing a food-chain game that features lots of small helpless creatures retreating from big hungry ones, then perhaps the vocabulary would consist mostly of words about predators and escape tactics.

One of the great things about a minimal language is that even a casual audience will be willing to figure it out and utilize it. The low word count and creative word-combining aspect of Toki Pona invites people to learn it. If people want to learn what your beings are saying and can do so easily then you can even hide virtual Easter eggs in the dialog—people will want to learn the language because it will give them plot details.

The way other people have designed their languages can provide lots of ideas of how to proceed with your own. Be sure to check out www.langmaker.com which features almost two thousand constructed languages. You might even be interested in simply borrowing someone else's language. If so, be sure to check with the author about licensing issues, especially if you intend to make money with their creation.

Language Sound Design

 

R2D2's emotions are communicated to the audience paralinguistically through tonal and textural changes.

Inventing languages of any sort can be fairly easy if you do not get bogged down with too many minor details. From a sound design perspective it is most important to determine the vowel sounds, consonants, pronunciation and intonation. If your language is not going to be illustrated as text, then writing systems can be ignored altogether. Even grammar can be somewhat sidestepped if you are not intending to create a fully functional language. You may be creating a voice so processed and out of this world that there is no need to focus on a language 'system' at all in which case you should just wing it. To be realistic, consider exactly how much time the audience is going to have listening to the language, and spend a proportional amount of time constructing it. If a thirty second scene in your five minute film is the only time that the audience will hear it then create a vocabulary just for that scene. However, if your constructing a language for a race of beings in a video game, your audience might have 48 hours listening to it, which is plenty of time to get familiar with what the characters are saying. In that case you would be wise to flesh out a complete vocabulary before you start generating the sounds.

Working for a video game company or film production house, a sound designer might say "Well I don't design the characters, I'm in the sound department!" That is precisely why it is so difficult to create visual designs that take full advantage of sound. It is our uncelebrated duty to push sound into the visual design phase and this includes working with other departments on character design. Get involved in the process, keep in touch with the artists and bounce ideas off of them by creating sound for their test renders. It is likely that they will be inspired by your ideas and create characters that are designed to make great sounds (and speak interesting languages). Just as importantly, you need to understand the character that is taking shape—what it looks like, how it forms speech sounds, what world it comes from, its cultural history, underlying beliefs and the way it behaves.

There is no single step-by-step guide that could accommodate all language sound design projects but here are a handful of the most important things to consider, roughly in the order they should be approached:

1. Research

Designs of any sort are improved by working from models. For example, Hutese and Greedo's dialog from the Star Wars films were both based on Queecha, the Native American language of South America. Character designers get ideas by looking at photographs of machines, insects, human anatomy books and fashion magazines. In "Designing Movie Creatures and Characters", character designer Miles Teves, who got his break into the industry working on Ridley Scott's Legend, describes how his ideas materialize:

“I’m a reference nut. I have hundreds of books; my house is full of them. And thank God for Google! I often create montages of relevant animals or characters and tape them all around my desk for inspiration...You need to be able to imbue your designs with a sense of realism and nuance that you can only get from studying nature directly. A lot of young people send me their designs, but they are often images of monsters and characters that are clearly two or three times removed from nature and display the heavy stylistic influence of other artists. You can see the artist grew up looking at comic books and not studying living things."

To reiterate this philosophy in a more general sense, using the entertainment industry as your main source of inspiration is redundancy suicide. You may not even realize that you have a choice not to have that unseen ghost speak in a slow reverberating whisper—its a contageous impulse. Sound designers can benefit greatly with interdisciplinary research in such fields as social, physical, and natural sciences. Just as an artist might adorn her drawing room walls with emotive floral photography, so should a sound designer collect comparative models for research. For a language sound design, inspiration could certainly come from the field of linguistics but also anatomy, psychology, philosophy, music, nature and numerous other sources. You can even use imagery like Miles Teeves and (since you can not hang sounds on the wall) keep your sampler loaded with inspirational sounds that you can access at any second.

2. Orthography

It helps to create a written representation of any unusual sounds in your invented language. A linguists approach would be to use the IPA system mentioned above but learning it for the purposes of language sound design would be overkill. Instead, you can make creative use of standard characters like " ' ^ / : or accent characters like î ç æ é Ü. Try not to get too overzealous with these symbols—an interesting looking language could be very distracting to voice actors; if the sound is not too far removed from your native language, adhere to a more standard representation. Hint: To create accent characters, try holding the Option key down while typing characters (Mac) or install a program called AllChars (Windows).

3. Vocabulary and Pronunciation

All the creative fun happens when you build your vocabulary, however, keep the list as small as possible because the longer it gets, the more difficult and time consuming it is to employ (and learn). If you are attempting to create a conlang phenomenon then by all means have at it. Just be aware that the listening audience generally can not tell if the language consists of 50 or 50,000 words. If you simply create the illusion of a full vocabulary, you have done well.

The words should personify the personality of the character so be conscious of the phonetics and how they will affect character sound design. For example, if you want your character to sound large and powerful, consider using phonetics that emote a large being like the English words "huge", "behemoth" and "massive" as opposed to "tiny", "meek" and "little". Compare the crude sounding Klingon language of Star Trek, with the hissing Parcel Tongue (snake) language of Harry Potter. The vocabulary list can make or break the sound design so give it careful attention.

4. Grammar

Grammar can be the most overwhelming part of language design, especially if you hope to create a language that people can actually use. The big advantage in language sound design (such as for science-fiction films) is that the audience is only hearing it, and not trying to figure out how to say or write it themselves. The casual listener will not even know where the divisions are between the words and as this is the first hurdle to understanding the language, the less words the better. 80% of the words we normally use are function words which have no meaning at all! For example:

Prepositions and postpositions: to, from, away, on, of, in, from, at, with, for, inside, etc.
Pronouns: you, I, me, mine, myself, her, herself, she, nobody, anyone, who, each, etc.
Sentence connectors: anyway, then, to, still, as, yet, otherwise, finally, too, first, also, etc.
Conjunctions: but, and, because, since, for, yet, so, or, nor, unless, until, whereas, if, etc.

To simplify your language you can eliminate these words and still get pretty clear meaning. Even plurals, possessives and inflections can be left out so that all you are left with are unmodified nouns and verbs. For instance:

"When Anouk arrived at her outpost she realized she was in danger of being ambushed."

Take out the function words and it still makes sense, although in English it comes off as caveman-like:

"Anouk arrive outpost realize danger ambush."

It might now work well if you try the same approach with a broken sentence like this:

"She realized she was in danger of being ambushed, or even worse—kidnapped, but the place was empty."

Because if you took too much out it would hard to understand, even for a caveman:

"Realize danger ambush worse kidnap place empty."

But if you retain some conjunctions (or, and, but, etc.) it is still quite understandable:

"Realize danger ambush or worse kidnap but place empty."

You might want some modifiers (words for hot, dirty, all, black, big, scary, etc.) and a sprinkling of interjections (like, yup, nope, ok, ha, hi, bye, etc.) which you could even borrow from another language for a down to earth effect (sayonara, hola, danke, hey!) but keep it simple and nobody will know that anything has been left out.

In a language like aUI, you can pretty much sidestep grammar by nature of the way words are created oligosynthetically. So, rather than creating a natural language that begs for function words, consider designing one that does not require them at all.

5. Recording and Mixing

Whether designing this language for a specific character or a race of beings, unless you are doing this alone,you will most likely need to enlist the help of voice actors. If you are designing a character's language with a particular voice actor in mind, book a trial session with them to create some "mock-ups" of the voice. Depending on how things go, you may end up making some phonetic changes (or even deciding to use a different actor). Make recordings that not only demonstrate phonetic sounds and intonation, but the language as expressed by the actor. If you intend to do a lot of signal processing, these recordings will provide you with a starting point for experimentation.

When the final recording session takes place you should be able to hand your voice actor a script written in your own constructed language. It should be written in such a way that it is easy for the voice actor to interpret and perform with minimal coaching. If you have sound mock-ups and character designs/animations the voice actor may find these enlightening, especially if they feel perplexed by impersonating an outlandish being speaking a conlang.

Another approach that you might take is to record the dialog in layers or fragments then synthesize or mix the parts together. For instance, you might opt to record the consonant sounds separately from the voice box sounds by having the voice actor read the script first unvoiced, meaning just with breath sounds, and then in a separate take read the dialog voiced but without consonants. Later you would then mix with these takes together to create a singular voice sound. Your reason for doing this might be because you want the character's voice box to sound very unusual which can best be achieved by processing just the sounds coming into the vocal tract, but not the consonant sounds produced by the character's lips and tongue.

Qapla': [qXAp^h'lA?] n. success

A sound designer has the unique ability to bring types of languages to life that could not occur outside of a technological process. The more you investigate, the more it will become clear how much cultural identity a non-human character can gain by speaking in tongues. Good luck!

References

Abram, David. The Spell of the Sensuous. New York: Pantheon Books, 1996.

Borges, Jorge Luis. Tlön, Uqbar, Orbis Tertius. Ontario: Porcupines Quill , 1983

Grotta, Daniel. J. R. R. Tolkien: Architect of Middle Earth. New York: Running Press Book Publishers, 2001.

Kisa, Sonja Elen. Toki Pona: The Simple Language of Good. Online. Internet. 25 Sep. 2007. Available: www.tokipona.org

Ladefoged, Peter. A Course in Phonetics. Fifth edition. San Diego : Harcourt Brace, 2006.

Noel, Ruth S. The Languages of Tolkien's Middle-Earth. New York: HoughtonMifflin, 1980.

Núñez, Rafael E. , & Eve Sweetser. "With the Future Behind Them: Convergent Evidence From Aymara Language and Gesture in the Crosslinguistic Comparison of Spatial Construals of Time. Cognitive Science: A Multidisciplinary Journal 30 (2006): 401-450.

Rickitt, Richard. Designing Movie Creatures and Characters. Boston: Focal Press, 2006.

Tolkien, J. R. R., & Christopher Tolkien. The Lost Road and Other Writings. Boston: Houghton Mifflin, 1987.

[back]

 

 
Darren Blondin, 2010