The study examined in the biscriptal Japanese orthography if phonological processing may accompany accurate and rapid visual recognition of single kanji characters according to their semantic or phonetic constituent elements, and high- and low-frequency katakana words. The subjects consisted of 108 grades 4, 5, and 6 Japanese children dichotomized into skilled and less skilled readers. The concurrent articulation interference paradigm was used while the subjects were making the lexicality decisions. The results suggest that visual-phonetic recoding may be possible in accessing difficult kanji characters with phonetic elements; and that phonological processing may vary according to the frequency of the katakana words. Further, younger children and less skilled readers are less efficient in their maintenance of the phonological code in processing the kanji and kana lexical items.