What is needed is a task where the three theoretical accounts yield different predictions. Haryu
(1991; see also Haryu & Imai, 1999) developed a task that does so. The task assesses children’s ability to coordinate two indirect linguistic cues in a Disambiguation task. Haryu presented half of a sample of 3-, 4-, and 5-year-old Japanese children with a standard disambiguation paradigm (e.g., presented them with an apple and a lipstick holder and asked to give puppet a heku, a novel Japanese word). The rest of the children were also given a strong pragmatic cue indicating that the intended referent was the familiar object: ‘‘Mary is hungry. I would like to give Mary (the) heku.” The standard disambiguation condition replicated previous findings. All age groups selected the unfamiliar objects as referents for the novel terms. The second condition (Pragmatic Cue task) presented a very different picture, a clear developmental shift in object choice; whereas 3-year-olds continued to pick the novel object and disregard the pragmatic cue, 5-year-olds selected the familiar object as referent for the novel word and, thus, no longer demonstrated a mutual exclusivity bias