“…Therefore, accurate selections of theta or upsilon remained dependent on the presence of pi to be excluded and did not lead to learning of the new relations between the words "theta" and "upsilon" and the corresponding letters (actually, Dixon repeated the procedure with other stimulus sets, and found that learning of new discriminations increased in probability). Similar outcomes were often obtained (Campos, Brino, & Galvão, 2013;Clement & Zentall, 2003;Domeniconi, et al, 2007). However, learning by exclusion was reported in other studies (e.g., Cabral, Assis, & Haydu, 2012;Ferrari, de Rose, & McIlvane, 1993;McIlvane & Stoddard, 1981).…”