Abstract
The phoneme, an abstract, descriptive unit of language, is crucial for our intellectual grasp of spoken language reality. It is seen as the basic constitutive unit in modelling human linguistic competence.
However, supremacy in our conscious attempt to understand production and perception of speech needs not presuppose primacy in the domain of natural processes of speech behaviour. The syllable is often viewed as the phoneme's closest competitor at the level of unconscious neurolinguistic planning and decoding.
Reaction time (RT) measurements, based on the assumption that longer RTs mean a more complex or higher-order processing, have been used successfully (and frequently) in various psycholinguistic experiments. In our study we used a word-monitoring paradigm to investigate whether the brain processes phonemes or syllables faster when recognizing words.
For that purpose we broke the speech signal with short stretches of silence and, also, used the English language for Czech listeners to add extra cognitive load to the task, which should underscore potential differences in the outcome. The results showed that phone segmentation exerts greater demands on perceptual mechanisms than syllable segmentation.
Syllable-sized units are apparently easier to process and only then broken into smaller elements. The secondary goal of our study was to compare two freely available computer programmes used for behavioural experiments in order to determine their applicability in linguistic research.