This paper examined the characteristics and patterns of uncommon names in present-day Japan. Uncommon names have increasingly attracted a remarkable amount of attention, both in the academic field and in society at large. In order to capture the underlying nature of the phenomenon of giving uncommon names to babies, it is important as a first step to describe the characteristics of uncommon names and to systematically categorize them within a structured framework. However, past research mostly focused on names that were too unique and unclear about how they were to be read (kirakira names), which reflected partial and potentially misleadingly extreme aspects of the phenomenon. Moreover, previous research has used unique names that were possibly invented and hypothetical, which is not productive to understanding the actual phenomenon of giving uncommon names and might produce/reproduce "anecdotal names" or "urban legend names." Therefore, in this article, names that were uncommon (not too unique) and real (not hypothetical) were examined. It is suggested that there are two ways of giving uncommon names: (1) giving an uncommon reading/pronunciation to Chinese characters and (2) giving uncommon Chinese characters. There are three typical ways of providing uncommon readings: (1-1) abbreviating the common reading of Chinese characters, (1-2) reading Chinese characters with the pronunciation of a foreign word that corresponds to its semantic meaning and (1-3) giving readings based on the semantic meaning of Chinese characters. In contrast, there are two typical ways of giving uncommon Chinese characters: (2-1) giving Chinese characters that are not encountered frequently in daily life and (2-2) including silent Chinese characters that add to the semantic meaning without contributing to the pronunciation. The characteristics of uncommon names and future directions in research investigating uncommon names in Japan were discussed.