Skip to content

Tag: phonemes

Worldbuilding 102 – How To Make a Constructed Language Sound Natural

I’ve been looking forward to writing an article about constructed languages for a while. But when it comes down to it, I find it difficult. I love languages and I have studied them for over forty years. I want to write everything there is to know about how to construct a language.

The difficulty for me is to rein it in. The problem is that, when writing fiction, less is more. You want a constructed language to add a fantasy flavor to a fictional world. That is, create words, names, and places unique to your world that sets the scene and creates a feel for the people living there. You don’t need a fully fleshed out grammar or writing system. You don’t want to overdo it and make your story hard to read.

Contents:
Consonants
Vowels
Pitch and Tones
Choosing Sounds
Orthography
Phonotactics

Language as Cultural Identity

It is a common belief that different cultures have different languages. For the most part, that’s true, though there are some cultures that differ in some ways but share a language that is almost exactly the same. For example, Serbian and Croatian are very similar, but the two cultures are divided by religion. The languages might be mutually intelligible if spoken, but one is written in the Latin script and the other in Cyrillic.

Language is often used as an indicator of group identity. Subtle differences in speech can mark a person as being part of a group or not. Often, it can be used to distinguish class as well. With your constructed languages, think about how different groups within your world use those languages. Do street people have a jargon they use to confuse the elites and avoid authority?

The problem for the fantasy writer is how to convey those differences. Often, you need characters to interact using dialog. If they don’t speak the same language, how do you write dialogue? One solution is to have everybody speak the same language. But if you do that, you lose the opportunity to breed misunderstanding and conflict.

Creating Consistency in a Constructed Language

In addition to showing differences, a language has to sound consistent within the group of speakers. Often, writers make up names and words on the spot, but they sound phony and ad hoc. In order for your world to to ring true to the reader, it helps for the invented words and names to have some consistency. That is, the words and names should sound like they are coming from the same language.

One of the easiest ways to create consistency is to simply use a real-world language as a template. Of course, then you are implying your fantasy world is Earth or some analog. You can copy a real-world language and make a constructed language that is similar, like J.K. Rowling did with spells in Harry Potter. All the words were pseudo-Latin, making the spells sound scholarly.

Whether you copy a real-world language or create one from scratch, it helps to do some basic work in constructing a language in order to get that consistency. One of the first things to think about is what it sounds like.

Sounds and Phonemes

Consonants

Every language uses a limited number of phonemes, or sounds. Some use more than others. For example, Lingit, the language of indigenous peoples of Southeast Alaska, has 47 different consonants, many of which sound similar to a native English speaker’s ears. On the opposite range, native Hawai’ian has only eight consonants. By comparison, English has 24 consonants represented by 21 letters.

The various sounds created by human speech are recorded in the International Phonetic Alphabet (IPA). No one language uses all of the sounds represented in the IPA.

A portion of the official chart of the IPA. International Phonetic Association, CC BY-SA 3.0, via Wikimedia Commons

English speakers are familiar with a fraction of these sounds. Arabic uses retroflex and uvular consonants. The glottal stop has no letter in English, but is common. Think of the Cockney pronunciation of bottle: /bo(‘)l/.

Lingit uses ejectives, which are a class of non-pulmonic consonants. That is, they are created by a push of air from the throat, not from the lungs. The Khoisan languages of south and east African use clicks, which are created by pushes of air in the mouth.

/R/ is pronounced differently in most languages. As a alveolar trill in Finnish, uvular trill in French, German, and Arabic, approximant in American English. English dialects use both postalveolar and retroflex as well as some dialects with the aveolar flap or trill to pronounce /r/.

You will need to choose which sounds you will use to build your constructed language.

Vowels

There are generally far fewer vowels in a language than consonants. The difference between consonants and vowels is like the difference between digital and analog. Consonants are distinct sounds, while vowels exist along a range.

The range of vowel sounds expands along two axes, from the front of the mouth to the back, and from closed to open. Also, vowels are often long or short (tense or lax).

An example of front of the mouth and closed is /i/ as in English ‘bit’. The long version is Spanish /i/ or English ‘ee’ as in ‘beet’. An example of open back of the mouth is /o/ as in English ‘hot’.

File:IPA Vowels with all vowels.svg
IPA vowels. Babbage via Wikimedia Commons

Vowels might be different within different dialects of a single language, though they are very similar. Compare an American saying ‘pin’ or ‘pen’ against a New Zealander saying it. The two words are switched in pronunciation. Even within the United States, vowels shifted as people moved from New England to the Midwest in the 19th century.

Nearly all languages have a minimum of three vowels. Few have more than ten, but some have as many as thirteen.

Often, they are combined into diphthongs, such as /ai/, similar to the English long /i/ as in ‘bite’, or /au/ as in the English ‘ow’, ‘au’, or ‘ough’ as in ‘bow’, ‘bauhaus’, or ‘bough’.

Pitch and Tones

In some languages, speakers change the pitch of their voice to convey meaning. That is, the speaker will add a rising or falling pitch. English speakers use this to add a questioning tone to the end of a question, though English is not a tonal language. Sometimes, this lift at the end of a sentence is used in statements when the speaker is looking for the listener to agree with the statement.

Speakers of some languages have multiple pitch changes throughout a sentence, not just at the end. This variation occurs even in various dialects of English. Compare an American speaker of English to someone from India. Americans speak in very flat tones while Indians speak with a variety of rising and falling pitches like ripples on a pond. These pitch changes might be conveying meaning, shifting the mood of the speaker’s meaning. Since I am an American, I speak in flat tones and any possible meaning from these pitch changes are lost on me. ¯\_(ツ)_/¯

In tonal languages, tonal changes can be very complex, changing the meaning of words. This is common in east Asian languages such as Vietnamese, Cantonese, or Mandarin. For example, the word ‘ma’ in Mandarin has five different meanings depending on the nature of the tone:
mā (媽/妈) ‘mother’
má (麻/麻) ‘hemp’
mǎ (馬/马) ‘horse’
mà (罵/骂) ‘scold’
ma (嗎/吗) (an interrogative particle)

File:Pinyin Tone Chart.svg
The tone contours of Standard Chinese. Wereon, CC BY-SA 3.0, via Wikimedia Commons

I don’t use tonal changes in my constructed languages because I am not very familiar with them.

Choosing Sounds for your Constructed Language

When designing your constructed language, think about what it will sound like in the reader’s head. Is it beautiful? Is it harsh? Does it sound like Latin or Greek? Chinese or Japanese? This will have an impact on how your readers approach the culture and your characters.

For example, Tolkien chose many unvoiced sibilants (f, s, l, h, hy) and often used the vowel ‘a’ when constructing elvish because he liked the sounds. Conversely, in the Black Speech and Orcish, he chose many fricatives such as th, s, z, sh, zh, kh, gh. The vowel ‘u’ is more common than in elvish.

It helps to sound out the words and phrases when you create them. You’ll get a better idea of how your your constructed language will sound to your readers.

Example: Savannah Speech

For Savannah culture, I want it to sound more throaty and harsh. The reason is that I intend this culture to be harsh. They are slave traders and the bad guys are the rulers, nobles, and elites of their area. Like Tolkien, I find throaty sounds more harsh and not as pretty as sounds produced in the front of the mouth and glides such as l, y, and wh. Of course, there are some front of mouth sounds that are basic in just about any human language.

The twenty-five consonants (written in IPA) I decided on using for Savannah Speech are: p, ɸ, m, b, β, t, s, n, d, ts, z, t̠ʃ, ʃ, d̠ʒ, ʒ, k, x, ŋ, ɡ, ɣ, q, ɢ, ʀ, ʔ, h, ɦ.

Because the Savannah people use a lot of consonants in the back of the mouth, the vowels follow the same pattern. I chose these four: a, œ, u, and ɒ. We will use the short version of all of them and the long version of a, u, and ɒ. We’ll also use the diphthongs au and ei.

Cultural and Language areas within Pancirclea. Image by Michael Tedin

Constructed Language Orthography

Orthography is how a language is written. There are three basic types of writing systems: logographic – symbols represent words. E.g Chinese
syllabic –   symbols represent syllables. E.g. Japanese or Korean, and 
alphabetic – symbols represent phonemes. E.g. Latin, Cyrllic, or Arabic alphabets.

We have been discussing sounds using IPA lettering. In order for an English reader of fiction to understand them, we’ll need to transcribe them. For the most part, this means using letters of the Latin alphabet.

You could design a fully functional script to write your language, but if you are doing it for written fiction, you likely won’t be using it. Even if you do come up with a script, good luck getting a publisher to typeset anything using it. For more information on types of writing systems, check out this Conlang wiki or the Language Construction Kit.

Tengwar.svg
“Tengwar” in the Tengwar script created by J.R.R. Tolkein

For the most part, it make sense to simply write the letters out using their English equivalents. The sounds might not match exactly to the sounds English speakers use for those same letters. This might cause some confusion as to how the words are actually pronounced, but that’s what book appendices and speaker’s forums are for, right?

Accent Marks

Accent marks can indicate differences between pronunciation of English and that of your constructed language. Be careful how you use these. Accent marks mean certain things in different languages. Too often, I see writers use accent marks to make words look foreign with no sense of what those accent marks mean. My own pet peeve is Mötley Crüe. It’s supposed to sound like Mawtley Crew with some “cool” accent marks. I think it reads like Mootly Cryu.

Some examples of accent marks used in different languages are apostrophes (‘), accent grave (è), accent acute (é), tilde (ñ), umlaut or diareses (ä), among others. These accents might mean different things in different languages. For example, the accent acute and accent grave indicates a tonal change in Vietnamese, but a change in pronunciation in French.

See the source image

Many languages use apostrophes for consonants. The one I see most often is the apostrophe for a glottal stop, as in Hawaiian. For example, Kauai and Hawaii are pronounced with a glottal stop as Kaua’i and Hawai’i, not Kauayi or Hawaiyi.

You can find a list of codes to type out letters with accents at alt-codes.net. If you just want a visual chart without pop-ups, you can find one here.

Phonotactics – Letter Combinations

Accents don’t work for all sounds. Often, we use letter combinations to indicate different sounds. For example, in English, we use a following ‘h’ to indicate a fricative. English doesn’t use þ (thorn) or ð (eth) for the dental fricative like Icelandic. Instead, we add an ‘h’ after a ‘t’, using ‘th’ for both sounds.

Often, English speakers will transcribe sounds in foreign languages the same way. For example, we don’t have letters for IPA x or ɣ. Standard English doesn’t use them, but Russian or Arabic do. We often transcribe them with ‘kh’ or ‘gh’.

One of my favorite letter combinations is the German transcription of . In English we simply use ‘ch’. In German, they need four letters to write the sound: ‘tsch’.

Another method of transcribing phonemes is by doubling the letter. For example, ‘ll’ in Welsh makes the ɬ lateral fricative sound (like ‘l’ but with air moving on either side of the tongue). The same combination in Spanish makes the ʎ lateral approximant (similar to English ‘y’). The ‘tl’ combination is used in Tlingit. Even IPA has a hard time with this, combining two symbols to represent one sound: ‘

Example: Savannah Speech

With twenty-six consonants, seven short and long vowels, and two diphthongs, it would stand to reason that my constructed language for the Savannah People would need thirty-five letters, but most languages use the same letters for the long and short vowels, so that makes it fewer.

The sounds ɸ, β, ts, t̠ʃ, ʃ, d̠ʒ, ʒ, x, ŋ, ɣ, ɢ, ɦ can be written using letter combinations, such as ‘pf’, ‘bh’, ‘zh’, ‘kh’, ‘ng’, and ‘gh’ for ‘ɸ’, ‘β’, ‘ʒ’, ‘x’, ‘ŋ’, and ‘ɣ’. I’ll add diacritical marks to change the sound of others, like in Czech. Also, ‘č’ and ‘š’ can stand for ‘t̠ʃ’ and ‘ʃ’.

I’ll use ‘dzh’ for ‘d̠ʒ’. I could use ‘j’, but I like the harsher look of ‘dzh’. The ‘ɢ’, and ‘ɦ’ are the voiced versions of ‘q’ and ‘h’, we can use ‘g̲’ and ‘ẖ’. They will only be between vowels or at the beginning of a word. Finally, I’ll use ‘r’ and ‘ for ‘ʀ’ and ‘ʔ’.

For the vowels, we have ‘a’, ‘œ’, ‘u’, and ‘ɒ’. The first three can be spelled with ‘a’, ‘ö’, and ‘u’. I’ll use ‘o’ for ‘ɒ’. The diphthongs ‘au’ and ‘eɪ’ are straightfoward enough, but I’ll use ‘e’ for ‘eɪ ‘. For the long version of ‘a’, I’ll simply double it: ‘aa’. I’ll use accents to distinguish long ‘u’ and ‘ɒ’: ‘ū’ and ‘ō’.

This gives me a set of sounds written with b, bh, p, pf, m, t, s, n, d, ts, z, č , š , dzh, zh, k, kh, ng, g, gh, r, q, q, h, h, a, ö, u, o, aa, ū and ō. The alphabet is much shorter: a, b, č , d, e, f, g, g̲, h, ẖ, k, m, n, o, ö, p, q, r, s, t, u, z, and ‘.

Next, we can start building words and giving them some meaning. Also, we’ll finally get around to naming some places in Pancirclea in their indigenous languages.

1 Comment