Special Coverage

What Happens To Sublexical and Lexical Representations After They Have Been Used To Understand Speech? (30)

Arthur Samuel (Stony Brook University & Basque Center on Cognition, Brain and Language) and Nicolas Dumay (University of Exeter)

Summary by Brett Myers, Digital Content Associate Editor

Can't Get You Out of My Head

Arthur Samuel  presents his recent work on speech perception in collaboration with Nicolas Dumay .

When you hear a word, it lingers in your mind at the lexical and sublexical levels. For instance, the lexical form of beef has the sublexical forms /b/, /i/, and /f/. Once beef enters your mind, it tends to stay there for a while (in one form or another).

via GIPHY

In previous work, Samuel showed that hearing a word affects lexical performance in a task 10-20 minutes later (Sumner & Samuel, 2007). The task was to identify pseudowords (e.g., peef, paff, jop) in noise that listeners heard earlier in the experiment (see image below).

As shown in the figure below, they found that hearing a similar real word first delayed (inhibited) the identification of the pseudoword. Conversely, hearing a similar pseudoword first sped up (facilitated) the identification of the target pseudoword. There was no effect of repeating the same pseudoword.

For the current work, they asked if they could see the same effects 12 hours later. They used a phoneme monitoring task to expose listeners to the prime items in Block 1 (i.e., searching for a given word-initial phoneme in a word list). Listeners identified pseudowords in a listening in noise task in Block 2. They tested participants at 8AM-10AM and 8PM-10PM for Blocks 1 and 2 respectively.

As in their previous work, they found that hearing a real word in Block 1 interferes severely with recognizing a similar pseudoword in Block 2. However, they didn’t see an effect of hearing a pseudoword in Block 1 on task performance in Block 2.

Taken together with results from Sumner & Samuel (2007), we see that lexical information remains active for at least 12 hours, sublexical information lingers more than 15 minutes but less than 12 hours.