Primary Sources

[Sean Manning]

So although Polybius and Livy are both writing centuries after the fact, we are obliged to treat them as primary evidence of the army and campaigns of the republican era, as we don't have the sources they were working from?

It's not really a matter of being obliged to treat them as primary evidence. They are primary evidence.

The terms "primary" and "secondary" have no bearing upon the quality of the sources, only their position in the chain of research.

For what its worth, I say that a source is a primary source for something when that source had direct knowledge of it, and a secondary source for it otherwise. Clearly other people use these terms differently, and it seems to me that arguing about definitions is not helpful.

Instead, I will step back and suggest that the principle that sources closer to an event are more reliable is so important that it needs to be called something. If we want to study Gaugamela, using a secondary source for the battle like Arrian faces challenges which using a primary source for the battle like the Astronomical Diary does not, because its hard to tell how the story was changed as it passed from mouth to mouth to produce Arrian's version. A source with direct knowledge of an event is not necessarily "better" than one without, but its usually easier to assess its perspective and limits. On the subject of early Rome, it would obviously be easier to decide whether to believe stories about the early republic if we had versions earlier than Livy's, Diodorus', and Dionysius of Halicarnassus' summaries!

What people call sources with more direct knowledge of something and sources with less is less important than that they think about which is which.