Managing Mismatch Between Belief and Behavior

Our behavior doesn't always match the beliefs attributed to us, and sometimes the mismatch raises questions about what our beliefs actually are. I compare two approaches to such cases, and argue in favor of the one which allows some belief‐attributions to lack a determinate truth‐value. That approach avoids an inappropriate assumption about cognitive activity: namely, that whenever we fail in performing one cognitive activity, there is a distinct cognitive activity at which we succeed. The indeterminacy‐allowing approach also meshes well with an attractive view of folk‐psychology: that ascriptions can help shape the attitudes they report.