Similarity and attribution in concept combination: Reply to Wisniewski

Wisniewski raises several important points in his comment on our study of the (non)role of similarity in concept combination. He notes that feature diagnosticity is an important factor in property attribution. He also points out that relational interpretations of combined concepts depend on thematic role plausibility, not constituent similarity. We agree with both of these points. Wisniewski further identifies a possible item artifact that might account for our results. However, reexamination of those items fails to support this hypothesis. Wisniewski also claims that an alignment process is necessary to specify how an attributed property should be instantiated in the combined concept. We argue instead that postcomprehension elaboration processes can account for specific property instantiations. Finally, we suggest that the perceived similarity between the constituents of a combined concept is an outcome of the comprehension process, not a prior condition for, or an integral part of, that process.