Upvote:7
Firstly, the author's intentions are important. If this book/movie/article/game is intended to be historical fiction (think Downton Abbey or Apocalypse Now), then the author/creator would have full license to create their own dialogue and events. Within reason, of course - such fictional works typically have to retain a certain level of authenticity in the dialogue and costumes.
If the book is intended as an academic study of an event or period in time, many authors these days choose to refrain from using apocryphal quotes. In cases where there is no verbatim (or reliable) record of a conversation or meeting, authors will often choose to discuss the intentions or outcomes of the meeting, rather than speculate on what each participant said.
In cases where someone has given exact quotes, you could always check their footnotes or endnotes (which a reliable book should have). There will often be references to the author's sources for these quotes, which will also help to give you a better idea of the reliability of the quote.
Upvote:11
This is why it is important to always keep the writer's intention in mind when studying some historical source. There are some different scenarios how that could happen:
In the end, you are right with your scepticism. Never assume that a conversation happened the exact way it was written down. We only have a chance to approximate truth by studying and comparing many different sources .. if several people with independent backgrounds (e.g. from different neighbouring cultures) agree on a thing, it might actually have happened that way. Maybe.
Upvote:29
The simple answer is that often the people involved write down (or verbally pass on information that is later written down by other people) these 'private communications' after the event, in memoirs, journals, letters, etc. Also 'official' secrets don't have to permanently stay secret, especially if the original reason for keeping them secret no longer applies, so 'private' correspondence and documents can be released to the public (and therefore to historians) once the need for secrecy is gone.
The issue of fabrication is a different one and opens up a wider discussion of intended and accidental bias in historical reporting. There's also the issue of (mis)interpretation if the original language of the parties in the conversation is different to the language used by the historian.