What sound changes are most common in natural languages, and in what order do they usually occur? Are there any factors to take into account when modelling sound changes for a conlang that is supposed to resemble a natlang?

  • 3
    Index Diachronica (chridd.nfshost.com/diachronica/index-diachronica.pdf) shows every documented sound change in natural languages. This would be a good reference for naturalistic evolution.
    – Ylahris
    Commented Jan 27, 2023 at 16:05
  • 1
    @Ylahris Index Diachronica aims to show every documented sound change, but it is far from complete and it isn't always the best at clarifying which sound changes are generally accepted and which are disputed
    – Tristan
    Commented Jan 30, 2023 at 9:16

3 Answers 3


By far the most common changes are assimilation, one sound becoming more similar to a nearby sound, and lenition, a sound shifting to require less articulatory effort. These are both broad categories that encompass a lot of more specific types of changes.

Assimilation can include:

  • A sound taking on an adjacent sound's features, like md > nd or zt > st
  • A cluster turning into a single sound, like ts > ss (total assimilation)
  • A sound becoming voiced between vowels (which are usually voiced), or voiceless at the edge of a word (next to the voiceless silence)
  • A vowel taking on the features of vowels in other syllables, like uti > yti (vowel harmony)
  • And various others!

Lenition can include:

  • Sounds using less of a closure, like k > x or x > h
  • Sounds disappearing entirely, like h > ∅ (deletion)
  • Vowels moving toward the center of the vowel space, like a > ə (laxing)
  • Long sounds being pronounced for less time, like tt > t (shortening)
  • Multiply-articulated sounds losing one of their articulations, like kw > k (or p)
  • And more!

Since these both reduce the amount of articulatory effort needed, they're very common types of changes. But they also generally involve losing information, so other types of changes will work to counteract this over time (for example, morphological changes to restore some of this information).

  • Really great answer, especially for simulating naturalistic evolution and change of conlangs over time
    – just Paul
    Commented Jan 26, 2023 at 1:09

Things are probably hard to quantify, but some specific sound changes seem to be more frequent than others, most notably:

/h/ -> /∅/ (loss of /h/)

The consonant system often has gaps at /p/ and /g/:

/p/ -> any: is mostly /p/ -> /f/ or /p/ -> /ɸ/, can go further via /h/ to nothing /∅/

/g/ -> any: lots of choices, e.g., /g/ -> /ʤ/, /g/ -> /k/, /g/ -> /ɣ/ (and further to nothing /∅/)


Here are a couple sound changes I use when I'm not sure what to do:

  • Voiceless consonants becoming voiced between two vowels (intervocalic voicing)
  • [u] and [o] becoming [y] and [ø] in the environment _Ci (that is, when the next vowel is [i])
  • Voiceless nonsibilant fricatives becoming [h]
  • [h], [x], [ħ], or other weak consonants being lost and lengthening the previous vowel (compensatory lengthening)
  • Clusters of two voiceless stops becoming geminated forms of the second one (this is sometimes called Italian-style gemination)
  • Weakening voiced stops to liquids (e.g. [b] to [w])
  • this is a really great list. i'm used to weakening voiced stops to fricatives first ([b] to [v] or [β] and [d] to [ð] or [z]) before turning them to liquids and i never considered doing it straight away was an option
    – just Paul
    Commented Apr 13, 2023 at 2:27
  • @justPaul thanks, I remember in Thakina I had the idea to do j to z and w to v, then replace the liquids with the voiced stops. I think that was the first time I used that sound change, and now it's a favorite of mine. Commented Apr 18, 2023 at 0:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.