In this study, I propose a refinement of Kuroda’s (1972, 1990) claim that the categorical and thetic judgments are realized syntactically in Japanese. Succinctly put, Kuroda argues that sentences with the topic marker wa represent categorical judgments whereas those with the nominative marker ga represent thetic judgments. In the present study, I demonstrate that wa-sentences do not uniformly represent categorical judgments and that ga-sentences do not represent thetic judgments across the board either. In particular, I argue that ga-sentences represent thetic judgments only on the so-called "neutral-description" reading (in the sense of Kuno 1973); on the "exhaustive-listing" reading, they instead involve categorical judgments. In addition, I demonstrate that wa-sentences also require a parallel refinement: thematic wa vs. contrastive wa. I therefore propose a four-way distinction and further demonstrate that the four readings are interrelated through different types of judgments as follows: thematic wa and exhaustive-listing ga involve categorical judgments whereas contrastive wa and neutral-description ga involve thetic judgments.