Several researchers have reported the increased use of o and semantic/pragmatic or syntactic motivations for direct object marking based on texts from between the 8th and 11th centuries. However, very few studies documented direct object marking after that period on until Modern Japanese, and the claims regarding semantic/pragmatic or syntactic motivations were not substantiated. By quantitatively examining an 11th century text and two later translations of it, this study documents evidence that o-marking was pragmatically motivated in the 11th century. However, as o-marking gradually increased over the years to be fully grammaticized as a direct object marker by the 20th century, its correlation with the pragmatic factors weakened.