In recent times incidental teaching has become the preferred method for encouraging children's learning and language development within the early childhood setting. It is argued that this naturalistic, unstructured and predominantly child-led style of teaching is the best way of fostering children's language acquisition and generalisation across settings. Few studies however have compared this method with more traditional, structured, adult-led teaching strategies. The present experiment compared the effectiveness of child-led and adult-led teaching interactions in the acquisition, retention and generalisation of object names in seven, 4-year old children. A comparison of the effects of the two teaching conditions revealed that children learned more object names during the adult-led interactions. However this difference did not prove to be significant. No difference was found between the two groups in either retention or generalisation. These findings have significant implications for the way in which young children should be taught and raises questions about one of the fundamental assumptions underlying developmentally appropriate practice.