The project, design and development of the diachronic corpus of Japanese began in 2009 at the Department of Corpus Study, the National Institute of Japanese Language and Linguistics, Japan (NINJAL), as a collaborative research project by linguists and literature scholars of NINJAL and the University of Oxford. Its focus is on collecting representative Japanese literary works and classical documents from the tenth century to the nineteenth century. We are currently working on the development of the prototype version of the diachronic Japanese corpus: i.e. selection of materials, digitization of texts, addition of alternative texts (containing different orthography) to original texts, compilation of a basic thesaurus that differentiates between different spellings, and word segmentation. This paper addresses the discussion of the basic concepts encountered during our work on the project: synchronic and diachronic analysis, which led us to the design of a serial comparison model which allows us to examine language change between documents or literary works with respect to time.