Thread by @BretDevereaux

It's time for a long twitter thread on the nature and limits of the evidence for the ancient world!

As you may be aware, compared to even something like the European Middle Ages (much less the modern period) the evidence for the ancient world is really very limited!

1/lots

Because the evidence for the ancient world is so limited, it is often necessary when writing narrative histories for regular people to scaffold around known facts to fill in some of the blanks.

Obviously this has risks and good scholars signal when they are doing it. 2/xx

But a lot of times, when you don't know the evidence, the difference between the fact-supported pillar and the guess-work-supported lintel isn't clear, especially if the lintel is the point of the argument and thus directly asserted as the conclusion of the pillars. 3/

That can be a problem, especially when a given event or issue goes through the game of classics-to-public telephone.

See, scholar 1 might write an article on >EVENT< in the text. The sum total of that event might be only a couple of paragraphs in a few authors. 4/

Very limited information! But the scholar builds out an argument, making inferences from other evidence, or similar events in the past (or maybe just irresponsible guesswork, but usually not).

Suddenly a couple of paragraphs in the sources has become a 15 page article. 5/

Now a trained ancient historian is going to check the footnotes and know full well which is which. But often specialists in other fields (esp. non-history, non-classics fields) are not able - or don't care - to do this.

They take the entire 15-page narrative for truth... 6/

...And then build arguments on top of that.

And so you end up with a book aimed at the public with an argument balanced on a 15 page article, balanced on just three paragraphs in the sources with minimal details.

Clearly not a stable argument! Many claims, little evidence! 7/

But it is unstable in ways that are not going to be obvious to someone who isn't familiar with the nature and content of the sources!

To them, there is a statement in a book, it has a footnote to a specialist article, everything looks good. 8/

They then end up either wildly over their skiis when trying to use that information themselves, or are shocked that the book they like is poorly regarded by specialists (often with some variation of "where it's right, it's not new; where it's new its wrong"). 9/

So let's talk about the nature of the ancient sources, because that gives a good impression of the difference of what it is possible to *know* versus what must be *guessed* about the ancient world (particularly, but not exclusively, Greece, Rome, and the Near East), 10/

1) Literary sources, by which we mean all of the long-form written texts. These are the starting point for basically any sort of investigation. But there are very few of them! The entire corpus of Greek *and* Latin fits in just 523 small volumes. en.wikipedia.org/wiki/Loeb_Clas… 11/

And that is to be clear, original+translation!

And that's the *best* it gets in the ancient world. Egypt, Persia, ANE, Phoenicians, etc? Even less - often a LOT less. Sometimes effectively none! 12/

The problem is that for certain kinds of things, it's literary-evidence or bust. Events (wars, plagues, short crises)? Often literary-or-bust - even if there might be archy evidence, without a literary source to understand that evidence, often hopeless. 13/

Biographies? Literary or bust, most cases. complex political systems? mythologies? philosophy? tricky social values? A lot of that relies heavily on literary evidence.

If it isn't there - if it's lost (e.g. a consitution for carthage!) it's lost. 14/

Moreover, the literary sources have interpretation problems, because they're written by people (mostly men) and reflect their perspectives (mostly elites talking to other elites about elite things).

Also they get things wrong and make mistakes! But better than nothing! 15/

But ok, so you have a question about the ancient world which the literary sources don't answer, or answer only incompletely (like a 3-paragraph reference you want to understand more fully). What are your other options? 16/

2) Representational evidence. Basically artwork. Really good at telling you what things looked like (but beware of artistic conventions!), but little help for names-dates-events kind of work. 17/

Bigger problem with representational - 'so that's what it looks like, what is it?' Trying to match thing-you-see with thing-in-texts is crucial and often hard (e.g. the wtf is a kotthybos argument).

Also, what something looks like is actually often just not very important! 18/

Representational evidence gets a lot more useful if you can say 'X depicts Z event from B lit.source' or 'Z object from B lit.source' but obviously you need to have B to make that work and B is generally doing most of the lifting. 19/

3) Epigraphy. Words carved in durable materials like stone. Upside: more texts to read and also unlike the literary texts (which are basically fixed and we don't find anymore), more of these found all the time! 20/

Downsides: types of texts very limited. Mostly laws, decrees, lists. Narratives of events are rare. Very useful source for legal texts, but you need literary sources (again!) to provide a framework.

But also *very* difficult to read and use...
21/

...often very damaged; typically requires specialists (epigraphers) to reconstruct the text into a form (still not english) that a historian can use.

Also very narrow in scope. Very few major historical events recorded in our literary sources can be attested epigraphically! 22/

4) Papyrology (and related branches of paleography): reading texts on papyrus.

Good news: much larger corpus, which includes lots of every-day documents instead of just lists and decrees. Receipts, private letters, census returns, fragments of lit. texts! 23/

Bad news! Almost entirely restricted to Egypt (and if we add wood tablets, one site in N. England).

Unfortunately, Egypt is weird! It is one of the most unusual places in the ancient Med., certainly in the Roman Empire. Not weird bad, just weird different! 24/

So extrapolating from Egyptian evidence to anywhere else in terms of census data, life expectancy, family size, customs dues, etc. is very hard. Lots ?? because Egypt is different and you may not know if it is also different in the way you care about! 24/(to be continued)

Otherwise, papyrus shares epigraphy's problems: often need specialists to read and reconstruct into plain demotic/greek/latin, often damaged, lines missing, text missing, context missing.

Last part is crucial - say you have a tax receipt, is it typical? 25/

Often getting a clear view of that question requires LOTS of examples to get a sense of what is normal. Good news in Egypt is that you have lots of papyrus - bad news is that very little of it has been edited and published.

And outside of Egypt...::sad!crickets:: 26/x

5) Archaeology. Most of what I do is at least 25% archaeological evidence. Often 50%+

Archaeology is wonderful, easily the biggest contributor to the improvement in our knowledge of the ancient world over the last 100 years. 27/52

(yeah, I said it, suck it Aristotle's Ath. Pol., archaeology is cooler than you).

Thing is, Archaeological evidence is really good to answer specific questions, but *only* specific questions. Most research topics are not archaeological visible. 28/52

Is your research question related to what objects where were at a specific time (objects here broadly could mean 'pots' or 'houses' or 'farms' or even 'people' if you are fine with them being dead)?

Good news, archaeology can help you. 29/52

But only if >object< leaves archaeological remains. Come back to that in a second.

But the thing is, that still covers lots of important questions. "When did people Y start using tool X?" "When did people start building here?" "what sort of pots did they use?" All good! 30/52

The best part is that archaeology is like fax machines (remember fax machines?) the more of it you do, the more valuable it becomes. New discoveries help to date and understand old discoveries. 31/52

With a *lot* of archaeological evidence, you can do really neat things, like chart trade networks, or changing land-use patterns.

The problem is that you really need a *lot* because you need a representative sample. 32/52

Only the best excavated regions (read: Italy) right now are at the point where we can confidently talk about changing patterns of land use, for instance and even then we have ???

Still valuable - archaeology is a young discipline! 33/52

But problems! First, what if the object you want is perishable, like textile cloth? Evidence pool collapses to nearly nothing fast; those objects only survive in weird circumstances (and see above for problems with weird circumstances). 34/52

Suddenly hitting 'representative' in the data set is really hard.

What if the thing you are studying won't leave any archaeological evidence? Well then it...doesn't leave any evidence. 35/52

That can be *really big.* Most wars, plagues, famines - not archaeologically visible. But also social values, opinions, beliefs - do not generally leave archaeological evidence. 36/52

For example: Cult of Mithras - leaves evidence in the form of ritual sanctuaries. But it can't tell us (except for disjointed, hard to use snippets) what they believed, or what rituals they did, or sometimes who they were. 37/52

Archaeology works best as a companion to the sources, but that brings us back to the lack of sources - if there's no lit. text, evidence level plummets.

Easy example of this: pre-Roman Gaul. The Gauls are *really* archaeologically visible. 38/52

They leave lots of prestige objects in shrines, lakes, rivers. Rich burial assemblages, identifiable hill-fort-town-centers. Lots of good archaeological evidence.

But zero textual sources until the Romans show up. 39/52

Consequence: almost everything about their values, culture, social organization before the Romans and Greeks start describing it is speculative. Lots of ????s - tons.

What archy can tell us, we know well - we can chart the changes in their objects really well! 40/52

Like 'when did they shift from shorter, pointier swords to longer, slashier swords' - can do a detailed map!

But 'what was Gallic kingship like in 350 BCE? 🤷‍♂️LOL🤷‍♂️
best guess is to reason from Gallic kingship in 50 BCE, when we have Roman/Greek lit. texts. 41/52

Which brings us to:
6) Comparative Evidence - or (favorite Jurassic Park reference), "the frog DNA."

Basically, if you don't know, fill in the blank with a similar, but more modern society which is better attested in the evidence. 42/52

Absolutely, comparative is the weakest form of evidence - it is what you use when you have nothing, but have a gap which has to be filled with *something* - comparative evidence is better than guessing... 43/52

...or in mid-19th century scholar fashion, just assuming ancient elites were just like your fellow British/French/German/American aristocrats.

The good news is that comparative evidence can be brought to bear on any question.
44/52

The bad news is that societies are different! Societies separated by centuries tend to be *really* different.

Comparative evidence thus works better when you have a reason to think things are less different. 45/52

For instance, subsistence patterns. Ancient people needed to eat too and had similar bodies (and thus dietary needs) to modern humans, so you can sometimes reason from early modern subsistence patterns to ancient ones (but beware tech/selective-breeding!) 46/52

Comp. evidence is easily the worst source of bad-pop-history-mistakes (that I talked about above). It goes like this:
Scholar1 estimates in technical lit. something based on little bit of hard evidence and some comparative evidence. Clearly signals that this is guesswork. 47/52

That guess might be the number of mules in use by the Roman army (based on more modern army pack animal usage). It might be the direction of currency flows in the first century CE (based on modern econ. theory). But it's a guess, not firm. 48/52

Writer2 puts together general survey of question, incorporates Scholar1's guess, but text is meant for general audience, so the guess is not signaled.

It looks like we *know* that fact, when we don't! 49/52

BookReviewer3 cries foul, but in a specialist publication. Meanwhile, Scholar4 (specialist in different field) reads Writer2's book and doesn't realize this is a guess, and so bases his argument on it.

And now you have a scholarly argument supported by nothing but air. 50/52

And ancient historians screaming that arguments are 'built on sand' or that we 'can't know that' from the evidence and being ignored by other disciplines and breathy think pieces in the media. 51/52

That's my rant. When in doubt, ask a specialist "what is this based on?" and signal comp. evidence *clearly.*

Also Check out my blog for more ancient history acoup.blog

And don't worry Ath. Pol., I still love you (Narrator: "...he lied unconvincingly.") end/52