There is more than one kind of earliest
The earliest mention of the I Ching is not a question with one uncomplicated date. We might mean the earliest event described in a text, the earliest composition of its words, or the oldest surviving physical copy. Those are different claims, and an appealing origin story cannot settle all three.
A story can describe an event centuries before the copy in which we read it. Equally, a manuscript can preserve words already old when its scribe wrote them. The date on a historical narrative and the date of a physical object answer different questions.
A consultation placed in 672 BCE
Edward L. Shaughnessy’s study of the Zhou Changes discusses a divination account in the Zuo Tradition, under the twenty-second year of Duke Zhuang, conventionally dated 672 BCE. This is an early narrated setting for a Changes consultation, not a surviving page written in that year.
The Zuo Tradition is a historical narrative associated with the Spring and Autumn Annals. Its stories of divination belong to the book’s literary and historical evidence. Saying that it recounts a consultation under 672 BCE is more precise than saying that an I Ching manuscript from 672 BCE has been found.
A physical witness from around 300 BCE
A different kind of evidence is the bamboo-strip Zhouyi acquired by the Shanghai Museum in 1994. In his first study of it, Shaughnessy describes fifty-eight surviving strips containing roughly a third of the received text. Its writing suggests a copy made in the southern state of Chu around 300 BCE.
The manuscript contains variant written forms, yet Shaughnessy argues that it also shows the text was already stable by the time of copying. The museum acquired the strips through purchase, so we should not describe them as a manuscript recovered from a securely documented museum excavation.
Read dates as evidence, not as a contest
Neither example supplies the birthday of the whole I Ching. One offers a narrative of an early consultation; the other offers surviving written material. A claim that the complete modern book was created on a single ancient date asks the evidence to do too much.
When you see an age claimed online, ask a small question: what exactly is being dated? A symbol, a story, a manuscript or a commentary? Following that question is a good way into the history, and leaves the book more interesting than a simple “oldest ever” label.