As children, part of learning English involved uttering the proverbial ah, eh, ee, oh, oo – the vowels, which along with consonants, form the building blocks of the English Language. Overtime, we learn to easily distinguished between different vowels in speech. To make this discrimination we need to process a wealth of continuous information, such as the pitch, the duration, or the loudness of the sound and then make a discrete choice about what vowel we have heard. Researchers have found it difficult to reconcile the fact that listeners make discrete choices based on continuous information.
Graduate student Gabriel Tillman and Professor Scott Brown from the University of Newcastle along with Titia Benders (Macquarie University) and Don van Ravenzwaaij (University of Groningen) developed a cognitive process model that describes how continuous acoustic information leads to discrete phoneme decisions. In a nutshell, the model posits that people sample evidence from the sounds and this evidence accumulates until a decision threshold is crossed, which triggers an overt response.
The model accounted for choice and response time data from an experiment where Dutch listeners discriminated between Dutch vowels. With the model, the researchers could examine unobserved processes involved in the perception of Dutch vowels. They found that sound frequency information contributes more to the perception of vowels than duration information, that frequency was more important for some of the Dutch vowels than others, and that longer durations did not delay when participants started using information from the sound.
Read more about this study here: