Coarticulation is the overlapping of articulatory movements for adjacent speech sounds. When we speak, the tongue, lips, jaw, and velum do not move in discrete steps from one sound to the next; instead, movements for upcoming sounds begin before the current sound is complete, and traces of preceding sounds persist into the next. At any point in the speech signal, acoustic information about two or more phonemes is present simultaneously.
The articulation of a current sound is influenced by a following sound. The articulators begin moving toward the next target before finishing the current one.
Anticipatory coarticulation is thought to reflect high-level articulatory planning.
The articulation of a current sound is influenced by a preceding sound, due to articulatory inertia.
Coarticulation is the physical mechanism underlying many Connected Speech processes:
In isolated, carefully articulated words, coarticulation is reduced. In fluent speech, it is pervasive, which is why Connected Speech sounds so different from citation forms.
Coarticulation is not merely articulatory laziness. It serves communication by spreading phonetic information across time, making the speech signal more redundant and therefore more robust to noise. Listeners use coarticulatory cues to anticipate upcoming sounds, facilitating faster processing.
Understanding coarticulation helps explain to learners why fluent English sounds different from the word-by-word pronunciation they encounter in dictionaries. It provides the rationale for teaching Connected Speech Processes explicitly: these are not "sloppy" speech but natural consequences of how the human articulatory system operates.