Phonaesthetics in Conlanging

Many conlangers (including myself) focus on linguistics when creating our conlangs. But while learning linguistics is very useful, we shouldn’t forget there is a creative side that needs to be considered. One way in which you can express creativity is through how the conlang ‘sounds’, i.e. its phonaesthetics. This is usually defined as the study of beauty in spoken language, and has applications outside of conlanging (such as poetry and creative writing). But in this post, I not will not just explore beauty, but other phonaesthetic qualities as well. I will also discuss how to utilise them in your conlang.


Euphony is the ‘classic’ definition of phonaesthetics, this is the beautiful, flowing, elegant sound that many conlangers want to achieve. The typical examples of euphonic conlangs would be Quenya and Sindarin.

Here is a list of euphonic features based on the findings in David Crystal’s “Phonaesthetically Speaking”:

  • Words with three or more syllables
  • Stress on the first syllable
  • Frequent use of /m/ and /l/
  • Front vowels
  • More manners of articulation
  • Front vowels before back vowels
  • Open vowels before closed vowels
  • Short vowels

Something to remember is that the list above only covers what words are considered beautiful in the English language and assumes British Received Pronunciation. So I tried to think about words I find beautiful in the foreign languages I’m familiar with:

In Icelandic, euphonic words that come to mind include: tölva /ˈtʰœlva/, saga /’saːɣa/, miðvikudagar /ˈmɪðvɪːkʰʏˌtaːɣʏr/, and of course Eyjafjallajökull /ˈeɪaˌfjatlaˌjɜːkətl/. So I think voiced fricatives could be added to the list above. The consonants /ʒ/ and /ð/ were rated low on David Crystal’s list. However, I think this is because those sounds are relatively rare in English (particularly in open category words).

In Chinese Mandarin I like chūzūchē /ʈ͡ʂʰútsúʈ͡ʂʰɤ́/, shénmeshíhou /ʂə̌nmɤ ʂɻ̩̌xou/, pǔtōnghuà /pʰùtʰʊ́ŋxwâ/ and mápódòufu /mǎpʰwǒ tôufu/. Most of these subvert the list above, but I think the main pattern here is the repetition of tones.

Of course what is considered beautiful may vary from culture to culture, and even from person to person. So it’s probably best to use your personal taste when creating euphonic words/languages. This is what conlangers usually do when they create personal conlangs. But, I think it’s also good to consider what most other people find beautiful in order to help engage your audience.

Phonotactics are usually limited in euphonic languages, strictly obeying the sonority hierarchy (see my blog post on phonotactics). They also avoid heavy consonant clusters and have restricted codas. Open syllables are preferred, but if there any consonants in the coda they tend to be continuants: this includes fricatives nasals, approximants, and trills. Although affricates may be considered continuants, but because they’re a stop + fricative sequence which slightly breaks the sonority hierarchy in the coda. Stops seem to be more euphonic if they are in the syllable onset, but in coda they either have a release that would affect the flow of a syllable; or they have no audible release, which stops the air-stream and makes the syllable feel choppy.


The opposite of euphony is cacophony: an unpleasant, awkward, choppy sound. The typical example of this would be Klingon. A cacophonic conlang would in theory have the opposite traits to euphony, so the following list can be deduced:

  • Words with less than three syllables
  • Stress on the final syllable
  • Avoids use of nasals and laterals
  • Back vowels
  • More places of articulation
  • Back vowels before front vowels
  • Closed vowels before open vowels
  • Long vowels

A common trope, so common some might say it’s a cliche, would be to add ‘guttural’ sounds. The term ‘guttural’ usually refers to sounds that are articulated near the back of the mouth. This includes velars, uvulars, pharyngeals and sometimes glottals. I feel some of these sounds, especially /q/, are a bit overused. Even I have used /q/ in Liyashi and wish I did something a bit more original). While I don’t think it’s bad to use ‘guttural’ sounds, there are other ways to make a language sound cacophonic and I would like to see more original approaches to this phonaesthetic style. There is also the possibility that the ‘guttural’ language trope may originate from unfair stereotypes of certain languages, so please be careful when including those kinds of sounds and consider context and worldbuilding.

One thing on the list above that may sound strange is having lots of places of articulation. David Crystal showed the euphonic words usually have many manners of articulation, so logically cacophonic words would have less variety in manners of articulation. So a cacophonic conlang would have relatively few manners of articulation, but for the sake of variety, it will have more places of articulation. This apparently works as Klingon does this, whether or not Marc Okrand did this intentionally I don’t know, but I will look out for this in other conlangs.

Making a language more consonant heavy can make it more cacophonic. This done several ways: having a large consonant inventory, syllabic consonants, and phonotactics that allow heavy consonant clusters. I made use of several syllabic fricatives in Liyashi, but I also allowed them to be preceded with a /q/ resulting in syllables such as qshh /qʂ̩ː/. In a more recent project, Airës, I experimented with the phonotactics by allowing complex consonant clusters in the onset, but I also put a lot of restrictions on the coda. This resulted in syllables being frequently unbalanced, giving it a choppy rhythm.

Sources of Inspiration

Avoiding Stereotypes

A common source of inspiration for many conlangers is natural languages. While this can help with making a conlang seem more naturalistic, it has a few problems if taken too far. I think an example of this would be Mondir, which I have discussed in my review of it.

Modelling a conlang’s phonology too heavily on a particular natural language has a major risk of stereotyping it. As a mentioned before, the cliché of using ‘guttural’ sounds to make a language sound ‘harsh’ may have its origin in stereotypes of various languages such as German, Russian and Arabic. Those are the languages which are typically characterised as harsh. However, this is actually quite a lazy evaluation, as all three of them can sound beautiful given the right context. Similarly, languages that are typically characterised as beautiful, such as French, can sometimes sound harsh, and languages that are characterised as cute, like Japanese, can sometimes sound quite tough. Anyone who knows those languages at all would know that those characterisations are obviously a load of BS.

So is it wrong to use phonaesthetics at all if it’s based on stereotypes? I would say it’s a bit more complicated than that. Associating certain sounds with emotions is not always based on stereotypes, and even though the phonologies of natural languages are based on the roll of a die, I actually think it’s ok to use some artistic licence for the sake of creativity. And just to be clear: I think it’s ok to use natural languages as models for conlanging, but I encourage you to be careful about what languages use and your reasoning behind it.

One more point I want to bring up is that Marc Okrand apparently designed Klingon to not resemble any real-world language, because it was intended to be a language for antagonists and he didn’t want anyone to associate it with their native language. Despite this, people seem to compare Klingon to languages such as Russian and Hebrew. I highly disagree with this, they sound nothing alike and I think this comparison is a really lazy product of ignorance. I bring this up as a reminder that no matter how hard you try to avoid this situation, people are always going to draw connections.

Child-like Language

Some of you might be looking for a ‘cute’ sound for your conlang. When we describe something as cute, we are effectively saying it resembles a child or baby. So the logical source of inspiration for the ‘cute sound’ would be child-like language.

Children learn open and back vowels before front vowels. They also learn labial consonants and stops first. Laterals and rhotics come later. Child-like language is also repetitive, so make good use of reduplication. If you want to know more, Mark Rosenfelder’s book ‘Advanced Language Construction’ has a good section on the topic (under ‘Language Acquisition’ in the ‘Life Cycles’ chapter).

Animal-based Styles

For various reasons, you might want to take inspiration from a particular animal. Whether it’s a language for anthropomorphic animals or you just want to use a certain animal as a motif. The exact details of this will depend on the animal of choice, but generally vocalisations of animals (or at least the human imitations of them) appear to be restricted and harsh (or at least to us humans anyway). Listening to various recordings (this Wikipedia page has a good list), sounds that might be useful would include trills, pharyngeals, and non-pulmonic consonants. Having an animal-based motif is probably a better reason for ‘guttural’ sounds, since those are the sounds that best replicate the vocalisations of animals and you’re not making an intentional comparison to a real-world culture.

Other Potential Motifs

I think there are many other themes/motifs you can express with phonaesthetics. One thing that you could do is design phonaesthetics based on a style of music, unfortunately I don’t know enough about music theory to go into this properly. Another thing you might want to look into is how the environment effects phonetics/phonology. Artifexian did a brilliant video on this subject, which I highly recommend.

Applying This to Your Conlang

In this last section, I will discuss how the previous ideas can be applied to your conlang.

Phonetics Vs. Phonology

An important thing to remember while designing phonaesthetics is the distinction between what sounds are present and what sounds are phonemically distinct. A lot of beginners seem to get obsessed with their conlang’s phonemic inventory (particularly consonants for some reason). I used to do this too, but as I became more experienced, I realised that the phonemic inventory is relatively unimportant in terms of a language’s phonaesthetics. Most phonaesthetic effects come from the details of phonetics, such as allophones, suprasegmentals and phonotactics. The phonemic inventory just shows which sounds are considered distinct from each other, which doesn’t necessary effect how a language sounds. Two languages could have technically the same phonemic inventory but sound completely different, if their allophones, suprasegmentals or phonotactics were significantly different.

Naturalism Vs. Artistic License

Another thing to consider is the effect of phonaesthetics on naturalism. For those who want to create conlangs that resemble natural languages, then I suggest you go for subtlety. Natural languages have evolved and lack a conscious designer, therefore natural languages lack strong traits of the styles of phonaesthetics.

If you want to make a language euphonic for instance, you could try throwing in a few cacophonic features This will help avoid your conlang sounding cliché and make it sound more naturalistic.

On the other hand, artistic licence is a thing. I and others have criticised Klingon for not being naturalistic enough (see my review). However, what I feel that I may have over-looked is that some conlangers may feel they would want to stray from naturalism to give their conlang more character. I think this works well for Klingon; it makes it stand out and I find it quite fun.

Sound Symbolism

In this post, I have been focusing on how a language sounds but individual words can have their own phonaesthetic quality. This means you can use phonaesthetics to help generate new roots. I recommend creating a list of roots of euphonic words and a list of cacophonic words; and then assign them meanings that feel right to you. You could even create a harmony system where one set is more euphonic, and another set is more cacophonic.

Of course, this post was very subjective, but I hope you have learned something from my ramblings and it will give you some things to think about when creating your next conlang.

If you like what I do and want to support me, you can do so on Ko-fi.

One comment

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s