I suggest not being dazzled by the spectre of Science in this case. Historical research in the modern age (post-Ranke) has always proceeded along in two basic ways: the reevaluation of evidence traditionally applied to a topic, and the introduction of evidence new to addressing that topic. And historians have always depended on the secondary scholarship of others--with the full understanding that, if an earlier piece of research is questioned or overturned by one of those two basic methods, arguments that had incorporated it or depended on it must likewise be reevaluated. (Well, ought to be, at least--unsurprisingly, one way that scholars deal with new research casting doubt on their pet hypotheses is to shove their fingers in their ears and double down).

Everyone loves debunking "lies my teacher told me" and "myths of ____ history class." There's one myth that never makes the list, but it's the most important one of all: what we know about history isn't a neat textbook narrative. Ever. Evidence is messy. It's scattered, corrupted, missing pieces; it's a piece of art that might demonstrate an original idea but might be centuries-old stock iconography; it's a text attributed to an author who never existed that survives only in one copy dateable to 500 years after its claimed date of composition; it's the vernacular words of a woman visionary claiming to speak from divine inspiration recorded in Latin by a male priest with an ulterior motive. In other words, it's all ambiguous. Where it doesn't seem ambiguous--so many medieval chronicles! Recording events, places, people!--can be the most treacherous of all, since the perceived straightforwardness tempts us to ignore the fact that it's a literary text composed by an author for a purpose that probably wasn't a modern textbook account.

In other words: historical evidence is hard. It acquires meaning only when set next to other historical evidence. And that meaning is not monolithic. That's not to say infinite interpretations are possible. In the mid-14th century, a massive epidemic killed utterly unthinkable numbers of people across Eurasia and North Africa; that happened. But the evidence of the what, the why, the how doesn't tell us "It was bubonic plague like you can get today and cure with tetracycline." The written evidence tells us that Western chronicle accounts describe different sets of symptoms; it tells us that the virulence/speed of mortality doesn't match up with modern plague. Scientific evidence? You've surely read the "conclusive" studies that yes, the Black Death was Y. pestis; we've isolated it in the teeth of skeletons in known plague-era/plague cemeteries. Ah, but what about the plague cemetery where researchers didn't find Y. pestis traces, but did find traces of anthrax--a leading alternate contender for the Great Mortality crown?

No historian crafting a narrative--whether a sweeping, era-spanning "Rise of Western Christendom" or "Africans in the Making of the Atlantic World" story, or an intimate microstory like "the execution of Marguerite, called Porete"--works completely from scratch. We all depend on the work of previous scholars, especially with sets of evidence whose quirks we're not familiar with. Sure, I'm not qualified to judge the different views on the sediment layer research on Horseshoe Lake that disputes whether a massive flood did or didn't contribute to the dissolution of Cahokia in medieval America. But I'm also not qualified to assemble medieval charters of land grants to monasteries and understand what snippets are standard formula and where it matters to have a woman's name instead of a man's listed. And with that use has always come trust: trust that the scholar you're citing, the scholar you're depending on, has evaluated existing knowledge to the best possible standards of the time they were writing.

Sometimes this trust is misplaced. Sometimes the trust is perfectly, wonderfully justified--and still bites you back, when other research questions those earlier conclusions. (A whole lot of history assumes we can take medieval chronicles straightforward; a good chunk of medieval women's history assumes male clerics related women's words and views straightforwardly).

As with scientists doing their research, as with businessmen making decisions in the corporate world, historians aspire to do the best we can with what we have at a given moment in time. Ambiguity is built in to what we do. We contextualize problematic evidence in evidence that seems less problematic (but may not be); we make judgment calls; and ideally, we're open to revisiting those conclusions when our judgment turns out to have been wrong for whatever reason.

I wrote my second master's thesis based on the discovery that the foundational work on scholarship on my topic had made a small but critical error in reading a primary source that every subsequent piece of research had replicated without checking. And I still had to throw away half of my writing three weeks before the deadline when I read a new article that illuminated a piece of the historical context, shifting what I saw as the significance of some of my evidence. It wasn't wrapped up like a gift; it was a paragraph in a primary source by another author that I probably could have twisted to fit into my current argument. But I could make a logical case for it to be telling us something about the past that did not support what I'd been saying in that part of my thesis. I'm not a historian of 16th century military politics, at ALL, but I had to make a call there and I did.

Don't get me wrong, there is some utterly amazing new data emerging out of DNA studies, paleoclimatology, and new attention to archaeology and osteology. My favorite example is, of all things, medieval manuscripts. Before the rise of paper in the West (and during a substantial overlap period), manuscripts after the age of papyrus were written on parchment--animal skin. We can trace the DNA of the animal skin. We can figure out where the animal's DNA suggests its ancestors came from. This doesn't tell us that a particular MS came from Italy or Anatolia, but it does provide one more piece of evidence to evaluate in looking at the geographic relationship among versions of a text.

Historical evidence isn't straightforward. Historical evidence from science fits right in to that well-established paradigm.