There are direct historical accounts in China that date back to the late 4th c. CE that describe monks coming to China and reciting whole Agama collections for oral translation. So, even after writing was introduced and written copies were used to preserve the canon, the oral tradition of memorizing collections of sutras continued.
There’s a subtler problem with Pali tradition that makes it difficult to assess whether or how it has evolved over time, which is that it lacks early witnesses to compare the modern version against. Other Buddhist traditions did change over time, which we can see by comparing early versions of a given text to later ones, so it’s hard to think the Pali tradition didn’t. Without actual witnesses of the early canon, though, it’s difficult to make any solid arguments about this or that passage unless there’s clear historical context (like the mention of historical persons or events that are known).
An example is a translation problem I encountered recently with a Chinese version of the Cakkavatti sutta that I wrote about in a previous essay: On Lamps and Islands: The Ambiguity of Dīpa in Indic Languages.