Because some spatial terms indicates similar locations relative to a reference point, it is important to clarify the difference between such terms from a geometric viewpoint when we construct a computational model of spatial language. This issue has been explored using an acceptability rating task in which participants were asked to judge how acceptable a given term was for a specified location relative to a reference object. The most acceptable location for a spatial term can be defined as its typical location. However, the question remains whether the location deemed most acceptable for a given term corresponds to the term that is usually granted highest priority in identifying that location. This study investigated this issue with regard to Japanese spatial terms using a paired comparison. The results showed that the term deemed most acceptable for a particular location did not always correspond to the most preferred term for that location.