Phonemes are the smallest units of sound in a language. Unlike letters which describe how words are written, phonemes describe how words should be pronounced. There around 44 phonemes in the English language, though this varies with different accents and dialects.
In spɛl ænd spik I hand over the composition of these phonemes to a computer program and text-to-speech software. Unlike the process of haphazardly arranging letters, when phonemes are strung together there is less chance of the result being unpronouncable.
When compososed haphazardly by a computer do these new sounds make sense to human listeners? Can they be mistaken for English? Do changes in the voice, speed, pitch and gender of the computerised voice affect how we interpret these nonsensical sounds? Does the use of a human avatar help our understanding of these sounds as English words?