This paper is about the classification and systematization of Japanese spatial nouns. The three types of frame of reference proposed by Levinson (1996), classifies whole static spatial arrays in the following way: (i) Intrinsic frames of reference (not discussed here), (ii) Absolute frames of reference (not discussed here), and (iii) Relative frames of reference (discussed here). We could classify spatial nouns into two major classes; territory (dimension 2, 3)-oriented and direction (dimension 1)-oriented nouns. The territory-oriented nouns concern a territory of the reference object. The direction-oriented nouns concern the direction of a location with respect to the reference object, such as “X no migi (the right of X)”, “X no temae (in front of X)” and so on. And territoryoriented spatial nouns are further classified into three subclasses: designating internal location of the reference object (internal), designating a location apart from the reference object (external), and designating a location on shape of the reference object (boundary).