This study investigates a particular behavior in talk-in-interaction that appears to be, at least in its most explicit form, relatively unique to children, that is, the behavior whereby one participant explicitly instructs another participant to say a specific phrase, after which the first participant then supplies a prefashioned response. Rather than simply dismissing such conduct as the product of an immature psychology, close examination of the sequential organization of such talk reveals that even children at a very early age comprehend the execution of any given speech action demands a particular sequential context for that talk to be both sense-making and effective. The data discussed herein reveal an understanding of preference organization and sequential trajectory displayed by children as young as 2 and 3 years old and that this three-part "say [x]" format has versatile application in children's talk-in-interaction.