Theories on the mechanisms of phoneme identification generally involve only the actual segment and
the transitions to and from neighbouring segments. In a listening experiment we tested the
importance for vowel and consonant identification of the presence of speech segments beyond the
transition parts. The results clearly show that identification continues to improve when speech is
added beyond the boundaries of the transitions to neighbouring phonemes. Adding speech in front of
the target segment improved identification more than adding speech at the back of the target
segment, even if the sound added in front was actually not part of the target phoneme itself. We also
show that the identification of consonantal segments correlates with the correct identification of the
vowel, and vice versa, in CV-type stimuli, but not in VC-type stimuli. From these results, we
conclude that context beyond the CV- and VC-transitions is used for both consonant and vowel
identification.