The goal of this work was to determine whether participants improve their performance on a task of conditional discrimination in which feedback of their responses was not provided. 188 participants were evaluated in an equivalence-equivalence test, which was divided into four blocks of 9 trials. The results show that the participants who made between 5 and 12 mistakes presented an improvement between the first and fourth block of almost 20%, and in some cases, of until 44.44%. These results provide evidence of learning without explicit reinforcement. The possible contradiction with some studies centered on searching for facilitation procedures of the equivalence-equivalence response is discussed.