Is a nonverbal working memory task necessarily nonverbally encoded?

The objective of the studies reported in this paper was to determine whethe r memory for positions depends on visuo-spatial or/and on verbal encoding a nd whether the type of encoding is subject to developmental change. Several variants of a visuo-spatial working memory task (Mr. Peanut task, in which positions or associations of positions and colors of a clownlike figure ha ve to be remembered) were presented at encoding time with a concurrent verb al or spatial task. In Study 1, children aged 6, 8, and 10 years were compa red using the Purple-Peanut task (recall. of positions), and the Color-Pean ut task (recall of color-position associations). The results showed that en coding was predominantly spatial in children, even in 10-year-olds. In Stud y 2, the same two tasks were administered to young adults, and combined wit h a manipulation of the presentation mode (simultaneous vs, sequential. pre sentation). The adults relied almost exclusively on verbal encoding in both tasks and both presentation modes. Thus, there is a developmental shift in encoding mode in these tasks; this is in line with other studies in the fi eld, but the shift appears to take place later. In Study 3, also conducted with young adults, the original task was modified by eliminating the outlin e (Ghost-Peanut task), in order to reduce the possibility that positions wo uld be encoded verbally. This time, some spatial encoding took place, while verbal encoding was not totally eliminated. It is concluded that, when pos sible, verbal recoding is almost mandatory in adults, even when information is presented visually and there is no incentive for verbal processing. How ever, it is not necessarily more efficient than spatial processing.