One of the current issues concerning Japanese given names that has been widely discussed in the media and increasingly in academic literature are the so-called kirakira nēmu („shiny, sparkly names“), characterized by unusual and interesting sound and graphic forms. A typical feature of most of these names, written in Chinese characters, is a discordant sound- -character relationship, which makes them hard or even impossible to read. This type of name started to appear in the 1980s, experiencing a boom at the turn of the millennium. The rather negative consequences of this phenomenon, however, started to appear only recently, when generations with a high concentration of such names started to become socially active (attending schools, entering the workforce, etc.). Through an analysis of a corpus of 8,390 current children’s names, this paper seeks to identify particular types of deviations in the sound-character relationship and thus pinpoint what it is that makes current names hard to read.