- Demographic Estimates and the Issue of Staple Food in Early Japan
In Daily Life and Demographics in Ancient Japan, William Wayne Farris revisits the issue of demographic estimates for Nara and Heian Japan. Following his 2006 consideration of medieval population history,1 Farris, one of the few Western historians of Japan to focus on the life of premodern rural populations, has gone back to the earlier periods that he first took up almost two and a half decades ago. The present work brings together and appraises the findings of several scholars who have worked on Nara and Heian demography; it also provides more details regarding calculations of population size than did his 1985 study, Population, Disease, and Land in Early Japan, 645–900.2
Calculating Population Size: Sources and Problems
Estimates of the size of the population prior to the eighteenth century rest in the first instance on inferences about the amount of acreage under cultivation and how many people that acreage could support. For the early eighth century, Japan possesses an extraordinary set of documents for building such correlations. These include 178 individual household census registers (koseki) and population registers used as the basis for assessing taxes (keichō) as well as a number of other records pertaining to taxes preserved in the Shōsōin repository of Tōdaiji temple in Nara.3 The Nihon shoki (720) [End Page 337] records that in 670 Emperor Tenji (r. 661–671) conducted a nationwide census in which household registers were compiled. He also established the underlying administrative framework whereby fifty households (he) were held to constitute one administrative village (sato) and initiated the practice of revising the population registers every six years. This rule was subsequently incorporated in the Taihō administrative code (Taihō ritsuryō) issued by the court of Emperor Monmu (r. 697–707) in 701 and was retained in the subsequent Yōrō code (Yōrō ritsuryō), instituted in 757.4 The implementation of this system is known not only from the reporting of it in the Nihon shoki, but also from excavated wooden-tablet records (mokkan) dating from the seventh and eighth centuries.5
Contemporaneous materials for interpreting the data from the registers can be found in the Taihō code,6 which stipulated that each adult man receive an arable land allotment (kubunden) of 2 tan (1 tan = 113 square meters) and each adult woman 1.33 tan (the code allowed for slight variations in the distribution of plots depending on the availability of land in each village). It also established a rice tax of 1.5 sheaves of rice per tan.7 Eighth-century family registers that give acreage figures confirm the application of this rule. In addition, books 5 to 9 of the encyclopedia Wamyō ruijushō (Categorized Notes on Japanese Words), written around 934 by the court noble Minamoto no Shitagō (911– 983), includes a list of all administrative villages and the rice-field acreages for each province.8
The first scholar to make systematic use of these materials was Sawada Goichi , the pioneer of modern demographic studies in Japan. In his 1927 study, Nara-chō jidai: Minsei keizai no sūteki kenkyū (Numerical Studies of the Nara Period Public Economy), Sawada discussed the household and population registers, but to estimate population size, he also brought in the data to be found in a small number of district rice registers (guntō chō) and in various records of rice loans (suiko). Hypothesizing an average of 1,400 people per village, he counted province by province from the Wamyō ruijushō list of administrative villages a nationwide total of 4,041 villages, [End Page 338] which would add up to an overall population for the early eighth century of six to seven million.9
The sources Sawada utilized have continued to the present to be the natural starting point for studying the population of early Japan, and in the eighty years since he published his findings, successive generations of researchers have scrutinized them, cross-referencing the...