Japanese locative expressions such as “X ni iku” (“I'm going to X”) have been analyzed from the viewpoint of the so-called “placeness” of the noun (phrase). Placeness has been considered to be the syntactic constraint of the noun (phrase) that appears in the position of X, or to be the noun (phrase)’s semantic feature. It has also been explained in cognitive linguistics that placeness reflects subjective interpretation. However, neither theory has explained why possessive forms or definites can be used in locative expressions, where no placeness can be seen. Also, the semantic and cognitivist explanations presented so far form circular arguments. Therefore, the constraints of locative expressions are reanalyzed in this paper from the standpoint of ecological semantics, which assumes that the meaning of language is placed in a potential interaction within the human-environment system.

Analyzing the cognitive/perceptual structure of the noun (phrase) in locative expressions, we can find a reference-point structure. In addition to the combination of positional and non-locational nouns, possessive forms, and directives in locative expressions can also be treated as locative expressions because they each have a reference-point structure. It is also shown that locative nouns such as kōen (park) and byōin (hospital) can be analyzed in terms of their reference-point structure within themselves, since the perceptual structure of their meaning includes their affordances (e.g., park: play, hospital: medical care) as reference points. In sum, locative expressions are used to perceptualize the two-step perceptual structure in the human-environment system.

This analysis allows us to explain the grammatical constraints of locative expressions from an ecological perspective, independent of semantic features such as placeness and the interpretation of the subject. Furthermore, based on observation of the corpus, it also provides an objective analysis independent of the analyst's subjectivity.