Everything you know about history is false.

(Title is clickbait. What I’m saying is: almost everything you know about pre-modern history is false.)

Take a historical fact you think you know – any historical fact – from the pre-modern period. Ask yourself the fundamental question of rationality: why do you believe what you believe?

Fundamentally, your knowledge is based on all basic sources available, and inferences made thereof. By basic sources, I mean: epigraphical and archaeological sources, literary sources, genetic data, extant traditions/beliefs/technologies.

But these inferences you make are subject to incredible uncertainty, and to make things worse, historians usually aren’t even aware of the primary sources that underlie their beliefs. The way that history academia works is based on consensus: if a fact is present in a secondary or better yet tertiary source, it is accepted as fact.

Maybe you might think that this isn’t too bad. Suppose that the consensus view is ~70% in line with the truth, and only 30% is fabrication: this is good enough, right? We might be making mistakes about a third of the time, but at least most of what we know is true.

But it isn’t.

Historians are almost never even aware of the primary sources to make this “70% good” inference from it. Once something becomes consensus, it is accepted as fact. Any research then done by a historian is from these secondary and tertiary sources, and once again the resulting consensus is only “70% good”. As a result, the uncertainty compounds, and 5 nested references later, only one-sixth of history remains true!

OK, maybe you aren’t convinced: where am I getting this “70%” number from?

Here’s a sentence from the Wikipedia article, Maurya Empire. Wikipedia is a good example to use, because it is, by purpose, a reflection of the mainstream view, of the consensus of experts. The sentence is referenced to a textbook by a prominent Indologist, so you know this isn’t even a topic of much controversy among historians.

Chandragupta Maurya’s embrace of Jainism increased socio-religious reform across South Asia, while Ashoka’s embrace of Buddhism and sponsorship of Buddhist missionaries allowed for the expansion of that faith into Sri Lanka, northwest India, Central Asia, Southeast Asia, Egypt, and Hellenistic Europe.

Here are the statements being made, operationally:

• There was a political entity known as the Maurya Empire.
• There was a man named Chandragupta Maurya.
• Chandragupta Maurya was at some point the Maurya emperor.
• Chandragupta Maurya became a Jain.
• This act caused a significant number of people in South Asia to make changes in their practices and beliefs.
• There was a man named Ashoka Maurya.
• Ashoka Maurya was at some point the Maurya emperor.
• Ashoka Maurya became a Buddhist.
• Ashoka Maurya sponsored Buddhist missionaries.
• These acts increased the prevalence of Buddhism in various territories.
• All of this occured c. 300-200 BC.

Here the full list of relevant basic sources, from which we will try to infer these statements:

• Hindu literature: Mudrarakshasa, the Puranas, commentaries thereof, other datable Hindu literature.
• Jain literature: Parishishta-Parvan.
• Buddhist literature: Divyavadana
• Sri Lankan Buddhist literature: Dipavamsa, Mahavamsa, Vinaya, Samanta-pasadika, commentaries thereof.
• Tamil literature: Mamulanar.
• Chinese Buddhist literature: A-yü wang chuan, A-yü wang ching.
• Greco-Roman literature: Megasthenes, Strabo, Herodotus, Plutarch, Pliny the Elder, Justin.
• Buddhist architecture: e.g. Barabar caves, Sanchi, Bharhut.
• Ashoka edicts: Major pillar edicts, Minor pillar edicts, Major rock edicts, Minor rock edicts
• Other artefacts: Sohgaura copper plate inscriptions, Aramaic inscription of Taxila, various Mauryan coins.
• Later epigraphs: Junagadh inscription, various Buddhist monasteries.

Let’s try to infer just two of these statements: Ashoka Maurya was at some point the Maurya emperor, and converted to Buddhism. This is how historians make this inference: 

1. The Brahmi script is deciphered in such-and-such way.
2. The major edicts refer to a king named Piyadasi.
3. The minor edicts refer to a Buddhist king named Ashoka. (see)
4. The Dipavamsa claims Piyadasi to be the grandson of Chandragupta and the son of Bindusara.
5. The Divyavadana claims Ashoka to be the son of Bindusara.
6. The major edicts are scattered across the subcontinent.
7. The minor edicts are mainly centered in North Karnataka, and a few across the Gangetic plain.
8. Megasthenes describes a Sandrocottus.
9. Sandrocottus sounds like Chandragupta.
10. Megasthenes can be dated to c. 300 BC, because reasons.
11. (2 and 6) or (4 and 8 and 9 and 10) imply that Piyadasi was a Mauryan emperor.
12. (4 and 5) imply that Piyadasi is the same as Ashoka.
13. (3 and 11 and 12) imply that Ashoka was a Buddhist Mauryan emperor.

So to make this exceedingly simple claim that is crucial to the standard narrative of Indian history, we need to believe: (2 and 3 and 4 and 5 and 6) or (3 and 4 and 5 and 8 and 9 and 10). For simplicity, let’s say that the only uncertain statements are 4 and 5, and we assign probabilities to them as 80%, 60% (because 4 is corroborated by 8, 9, 10 since the Mauryas are the only known geographically extensive empire, but there could still be one that we just don’t know about), i.e. “there is an 80% chance of Piyadasi being the grandson of Chandragupta, given the reference in the Dipavamsa”, “there is a 60% chance of …” 

Then assuming the claims are independent, that gives us an ~50% chance of even the simple claim that “The third Mauryan emperor was Ashoka, a Buddhist emperor”. And this claim has been used as the basis for much of the historiography of India, multiplying this 50% with other uncorrelated claims with their own uncertainties. 

In other words: only a small minority of expert claims about the religious history of India can be expected to be true. 

But this uncertainty simply isn’t acknowledged in the history academia. Even when a claim is receives “criticism”, such criticism isn’t organized in a fashion that can be tracked, so the uncertainty of claims depending on this claim can be computed. There is no propagation of this uncertainty: the notion of statistical inference is simply lost in the humanities. 

At best, the historical narratives should be seen as a work of fictional literature. Their notion of consensus is hardly better at all, in the big picture, in selecting for truth, than a movie director’s notion of entertainment.