This study compared the effectiveness of balanced and skewed input at facilitating the acquisition of the transitive construction in Esperanto, characterized by the accusative suffix -n and variable word order (SVO, OVS). Thai university students (N = 98) listened to 24 sentences under skewed (one noun with high token frequency) or balanced (equally-low token frequency) presentation following either inductive (rule not given) or deductive (rule given) instructions. In the testing phase, they heard 20 sentences (10 SVO, 10 OVS) with new nouns and identified the object. Only the group that received balanced input and deductive instructions detected the novel pattern.