The sonority hierarchy refers the ordering of speech sounds from highest amplitude to lowest amplitude. Vowels are the most sonorous (e.g. /ɑ/, /e/) whilst voiceless plosive consonants are the least sonorous (e.g. /p/, /t/, /k/).

The simplified hierarchy looks like the following with the most sonorous sounds on the left:

Vowels > Glides > Liquids > Nasals > Fricatives > Affricates > Plosives

Glides include /j/ and /w/ as in ‘you’ and ‘want’ respectively.

Liquids include /l/ and /r/.

Nasals include /m/ and /n/.

Fricatives include /s/, /z/, /f/ and /v/.

Affricates include /tʃ/ as in ‘church’.

Also voiced sounds are more sonorous than voiceless ones so, for example, the voiced fricative /z/ is more sonorous than its voiceless counterpart /s/.

So far, there is nothing particularly special about this hierarchy – we’ve chosen sonority as our criterion and made a hierarchy with it. We could arrange speech sounds in numerous other ways, e.g. by place of articulation, or by frequency of occurrence, so is there anything special about arranging speech sounds in terms of sonority? From the point of view of syllable structure, there is.

Syllables are considered to be hierarchical sound structures. Obligatorily there is a nucleus, and optionally there is an onset (before the nucleus) or a coda (after the nucleus) or both. For various reasons, the nucleus and coda are considered to form a sub-unit within the syllable called the rhyme.

_syllable{ Onset – _rhyme(Nucleus – Coda) }

Cross-linguistically the most common type of syllable is CV, that is an onset-nucleus syllable where the onset consists of one consonant and the nucleus consists of one vowel, and some languages only permit this kind of syllable. English is quite free in terms of the number of consonants and vowels permitted in a syllable, for example, ‘strengths’ has the structure CCCVCCC. What is interesting is that cross-linguistically the vast majority of syllables are built around the sonority hierarchy:

The nucleus is most commonly a vowel (it can also be a consonant such as /l/ or /n/ as in ‘bottle’ or ‘button’), and is thus the most sonorous part of the syllable.

In onset position, consonant clusters (multiple consonants together) generally increase in sonority (a notable exception to this tendency is /s/). For example, ‘plum’ (/p/ < /l/) and ‘tree’ (/t/ < /r/). Note that the reversed order /lp/ and /rt/ are not found in syllable onsets.

In coda position, consonants clusters generally decrease in sonority. For example, ‘help’ and ‘cart’ (in rhotic English accents) which do show the clusters /lp/ and /rt/ respectively. Note as well that the reversed orders found in the onsets above are not found in coda position. You might think ‘people’ has /pl/ at the end, but here the /l/ is the nucleus of the second syllable and not part of the coda of the first syllable.

The sonority structure of syllables is a tendency, there are numerous exceptions to the idea that sounds in a syllable go from least sonorous to most sonorous to least sonorous. But the tendency is very strong and seems to hold across the world’s languages. Thus the tendency itself is worth explanation. Why should syllables generally have this structure? Perhaps it is a way of making the syllable more perceptually salient since each syllable would effectively have its own sonorous peak. Perhaps consonants can be considered as ways of beginning and ending a vowel-sound and in clusters they are more vowel-like the nearer they are to the nucleus. This suggests that consonants must latch onto vowels and cannot really exist in isolation (which seems to be true especially of the least sonorous consonants – try saying a /t/ or a /b/ sound without a following vowel-sound). Of course, now we’re getting into questions about what vowels and consonants actually are and whether there is any sharp divide between them. All interesting questions arising from a quick consideration of the sonority hierarchy.