Conditional discrimination literature studies the formation of equivalent classes through several different procedures, such as: the Go/No-go procedure with compound stimuli; the Yes/No procedure; the Respondent-type procedure; the Stimulus Pairing-Response procedure; and the Stimulus Pairing-Orientation Response procedure. Regarding the Respondent-type procedure, the conditionality between stimuli was not demonstrated in training, but it was instead observed in the performance of participants in emergent relations test conditions. Conditionality could not be demonstrated by differential responding in training, since R-S relation was not programmed in this study. Even though no responses are required in these conditions, some responses that naturally occur in the programmed environment (i.e. sensorial responses) are relevant for the performance of tasks, which in this case are the tests. In stimuli control conditions, sensorial or orientation responses are needed, first, to put the participants in contact with the relevant properties of the stimulus that form the programmed environment; and second, to demonstrate that such responses also influence the creation of transitive relations. Considering how relevant orientation responses are, this study aims to track measurements for such responses through tracking the eye movements of participants exposed to Respondent-type procedure. Through the exploration of orientation responses in trainings that lack differential reinforcement, and of tests for emerging relationships, this study can contribute to research investigating the variables of orientation responses, and allow for a manipulation of their spatial-temporal properties. Although classes are formed through Respondent-type training (without an R-S response being programmed into it), would their establishment, in testing conditions, be accompanied by the establishment of selective observing? According to the results, nine in ten participants observed the establishment of equivalent classes, which indicate that our replication was successful. Data obtained through eye-movement tracking allowed us to identify selective observation in MTS training even through no R-S relationship was programmed. Lastly, the establishment of selective observation allowed us to formulate hypothesis based on certain aspects of the procedure, which allowed for the establishment of functions of correct and incorrect stimuli in the experimental task (because the training for conditional relationships was not a discriminative training).