Observational studies suggest that children as young as 2 years can evaluate some of the arguments people offer them. However, experimental studies of sensitivity to different arguments have not yet targeted children younger than 5 years. The current study aimed at bridging this gap by testing the ability of preschoolers (3-, 4-, and 5-year-olds) to weight arguments. To do so, it focused on a common type of fallacy—circularity—to which 5-year-olds are sensitive. The current experiment asked children—and, as a group control, adults—to choose between the contradictory opinions of two speakers. In the first task, participants of all age groups favored an opinion supported by a strong argument over an opinion supported by a circular argument. In the second task, 4- and 5-year-olds, but not 3-year-olds or adults, favored the opinion supported by a circular argument over an unsupported opinion. We suggest that the results of these tasks in 3- to 5-year-olds are best interpreted as resulting from the combination of two mechanisms: (a) basic skills of argument evaluations that process the content of arguments, allowing children as young as 3 years to favor non-circular arguments over circular arguments, and (b) a heuristic that leads older children (4- and 5-year-olds) to give some weight to circular arguments, possibly by interpreting these arguments as a cue to speaker dominance.