Two experiments tested a basic assumption of the differential arbitrarily applicable relational responding effects (DAARRE) model by examining the extent to which functional (Cfunc) and relational (Crel) properties of stimuli affect performances on the implicit relational assessment procedure (IRAP). Experiment 1 required participants to complete IRAPs in both their native and a foreign language. Twenty-one Brazilian participants (fluent in both Portuguese and English) completed two IRAPs, one in which Portuguese-language stimuli were predominantly used and a second in which English-language stimuli were predominantly used. The IRAP trial-type containing Portuguese-language (1) labels, (2) positively valenced targets, and (3) response options produced the largest IRAP effect across all eight trial-types (four within each IRAP). Consistent with the DAARRE model, the Cfunc properties of the native language stimuli appeared to dominate over the Crel properties, relative to the foreign language stimuli. Experiment 2 sought to extend Experiment 1 using a known-groups design involving stimuli that were assumed to possess differential Cfunc properties across two groups of Brazilian soccer fans. Two groups of participants were recruited, differing in their support for specific Brazilian soccer teams. They completed a single IRAP task that included the names of the teams and their corresponding team badges. It is critical to note that responding to the IRAP trial-types did not require any explicit evaluative response in that the IRAP required participants to simply categorize team badges with team names as "true" or "false." The largest IRAP effect observed for each group was for the trial-type that presented the name and badge for that group's supported team. Both experiments provide support for the DAARRE model analysis by indicating that the Cfunc properties of stimuli are critically important when interpreting IRAP performances. A number of caveats to this conclusion are discussed. [ABSTRACT FROM AUTHOR]