Why do some sound combinations form words while others do not? Why can some sound combinations only be found at the beginning of words whereas others can only be at the end? The answer rests in the sonority scale. This discussion follows along with Kelly’s post on “What is a Syllable? Rhyme and Reason”. Syllable structure, what makes a string of sounds a syllable, is based on the principle of sonority. Sonority is the openness of the vocal tract that corresponds directly to loudness of a sound. Consonants are divided into obstruents and sonorants based on how they are articulated. They are called obstruents because the airflow is being obstructed. Say a sound like [t] compared to [n]; [t] is an obstruent and stops suddenly whereas [n] is a sonorant. The sound of [n] is more open and can be stretched out. The more like a vowel a sound is the more sonorous it is. Think of a syllable as a mountain: there is a climb up to the top and then a descent. Vowels are the most sonorous so they are at the peak or the nucleus of a syllable. The consonants preceding a vowel increase in sonority; this is the onset. After reaching the nucleus, the consonant sounds decrease in sonority; this is the coda. The sonority scale from lowest to highest is obstruents, sonorant consonants, and then vowels.
For example, in a word like Atlantic, which is written syllabically as at.lan.tic, the first syllable starts with the nucleus [a] and in the coda when sonority decreases it falls to the obstruent [t]. A new syllable is necessary because [l] is more sonorous than [t]. It violates the sonority principle to have [l] follow [t] in the coda. In the second syllable, [l] is in the onset, rising to [a] the nucleus, and falling to [n] in the coda. The third syllable cannot begin with [n] and then be followed by [t] in the onset because this would violate the sonority principle. So the sonority scale is the reason we can have the word ant but not *atn or *nta. If *nta is a word often a vowel is inserted into the pronunciation [nata]. Can you think of any exceptions where sounds in syllables violate the sonority principle?