When did the use of fabricated speeches in historical works go out of fashion?

Question

In his History of the Peloponnesian War, Thucydides includes a number of made up speeches based upon "what was called for in each situation". While this is frowned upon in modern histories, it seems to have been perfectly acceptable practice then and their inclusion had no bearing on the historicity of the rest of the work. Every work by an ancient historian that I have read contains such speeches, although the quality and relevance to the situation varies.

What was the first documented case of a historian expressing opposition to the use of fabricated speeches? When did the practice begin to lose acceptance amongst historians in general? What were the reasons for this change of view?

Answer 1

To an extent it still isn't entirely 'out of fashion'. You can find people who do it, or something very similar, such as implying that "Person X will have thought Y" or "Person A must have said B", but they'll tend to be purveyors of 'popular history' rather than academic historians.

Many people claimed over the century only to write history based only on 'facts' - in other words, 'what really happened'. In reality, they invariably fell short of that mark. Important and respected figures in the development of history such as David Hume still used invented speech in the 18th century.

The turning point when it became professionally untenable to deal in fabricated speech can be dated from the 19th century and the career of Leopold von Ranke. Ranke stressed the need for history to be based entirely on verifiable sources, and brought about the transition of history to a professional academic discipline. Previously it was in essence a literary pursuit of amateurs and chroniclers who were permitted (or permitted themselves) great poetic license with what was said. This was considered acceptable in the pursuit of what they would have deemed a 'larger truth'.

Answer 2

This answer is partial and speculative. I think that one crucial idea in the development of modern history is, (ironically, given his status as a leading "fabricator",) first expressed by Herodotus:

I am not able to say for certain, nor do I declare any question other than that which [is reported].^a1

Gradually you find more and more general and refined statements of this basic idea, getting over the centuries to something like:

Let the reader be aware that whatever I mention in my book is relied on the news that were narrated by some men. I had attributed these stories to their narrators, without inferring anything from their incidents [...] If a certain man gets horrified by a certain incident that we reported in our book, then let him know that it did not come from us, but we only wrote down what we received from the narrators.
_{— al-Tabari, Muhammad ibn Jarir (c.915, Arabic). History of the Prophets and Kings. Baghdad.}

I am speculating that this principle had to become firmly established first in order for speech-fabrication to be viewed as "off-limits."

A: actually, he named a specific source – "what the Argives themselves report" – but it's the general principle that interests us here.