It’s the first week of classes, so fireside this week. Next week, we’ll dive into a short series looking at the question of the ‘universal warrior,’ the idea – too often repeated – that there is either a single consistent experience or personality true to all combatants to the present day. Also, for those who want a bit more of my ramblings but may not watch my twitter, I should note that I have been slowly (multiple threads: 1, 2, 3) live-tweeting my way through the really quite flawed ‘Warrior Archetype’ videos by author Steven Pressfield, since it espouses one such version of the ‘universal warrior’ ideal.
I also want to briefly address some concerns, voiced by a few people, that the blog seems to be veering into becoming more politics focused. Worry not! This is going to remain a history-and-popular-culture blog. I had intended the October fireside on stasis to be a one-off thing, and while that intent was somewhat foiled by events earlier in January, that won’t be a continued pattern. So no more politics, hopefully for a good long time. In the meantime I have a number of things in the works: the aforementioned series on the ‘universal warrior,’ the long promised series on textile production, along with (further out) hopefully a longer discussion of the concept of doctrine (I’m thinking of anchoring it, pun intended, around Master and Commander (2003) and Greyhound (2020)), a look at the dynamics of trench warfare and some of its common misconceptions (using the ‘trench scene’ from Wonder Woman (2017) as a starting point) and a discussion of the gulf between how diversity in the Roman Empire is popularly understood and how it actually functioned.
So fear not, we’re going back to our regularly scheduled programming!
(Also, if for some absolutely inconceivable reason you actually want my policy hot-takes – which understandably tend towards geopolitics and nat.-sec. rather than domestic, at least in normal times – you can follow me on Twitter (@BretDevereaux).)
If you want to support that regularly intended programming, please share it; if you really like it, you can support me on Patreon. And if you want updates whenever a new post appears, you can click below for email updates or follow me on twitter (@BretDevereaux) for updates as to new posts as well as my occasional ancient history, foreign policy or military history musings.
For this week’s musing, I want to actually talk a bit about the craft of history. In the mess of my above-mentioned twittering about Pressfield’s videos, someone asked, quite reasonably, how a historian knows which sources to put more trust in and which sources to put less trust in. So I thought I’d speak on that briefly. Learning to read primary sources (and secondary scholarship) critically is an important part of the historian’s training. It’s something I get my students to work at in all levels of courses.
Now I should note at the outset that we should understand ‘doing history’ as consisting of functionally three core tasks: selection, evaluation, and interpretation (this is my own labeling of the tasks, but I think any historian will recognize them). In the first step, selection, the historian formulates a topic or question (note: not an argument, not yet) and then gathers relevant evidence. I call this ‘selection’ because, even in fields with very limited evidence, this is still fundamentally an exercise in selection: based on the question I have, what sources should I investigate and consider?
Next, the historian evaluates that body of evidence. We’ll talk about what that looks like in a moment, but we’re generally evaluating for how reliable a source is, what their particular viewpoint is, and based on that, how useful they are for illuminating our topic or question. Finally we move on to interpretation, where we take our pile of now evaluated evidence and form an answer to our question or an interesting conclusion about our topic (what would be, when we write this all out, our argument or thesis). Of course for a complex project, chances are the historian is engaged in all three of these processes at once, on different fronts, but the basic cycle remains the same: select a question, select evidence for that question, evaluate that evidence, then draw interpretive conclusions from that evidence.
(By the by, part of what I think befuddles students is that history is not written this way because the best way to communicate historical research is not how it is conducted. An article, paper or book written by a historian generally leads with its argument and then proceeds to prove that argument, point by point, until all points are proved with evidence, almost the opposite order in which the research was done)
So how does the evaluation step work? When dealing with a source, there are two main questions a historian is thinking about, as mentioned before: how reliable is the source, and what is the source’s viewpoint. Those hoping for some sort of rubric of mathematical proof for either of these will be disappointed; as we’ve discussed before, the humanities are a field of words more than math and many things in them cannot be reduced to numbers.
Now when I say ‘sources’ fundamentally, I mean texts, although as we’ll see considering a text typically means considering the text’s author or authors (for the historian; the literature critic may disagree for their authors are figuratively dead, whereas my authors are only literally dead). Most of the evidence that a historian works with comes in the forms of text; even what we might think of as physical evidence (like archaeological artifacts, or – where relevant – scientific data and studies) comes to the historian mediated by texts (like an archaeologist’s site report, or a scientist’s peer-reviewed paper).
Second, I should note here that while I am speaking of ‘sources’ generally, the historian is always conscious that functionally all of our sources are produced by people and reflect their knowledge, condition and worldview. Even apparently ‘objective’ data, like a census return or a scientific study is, in the end, a text produced by a particular person or group of people to a particular purpose and infused with their worldview and assumptions. Of course there is, at the root of this, a real world in which events really happened and the historian tries frequently to reach that real world, but we have to always be aware that the objective world is entirely mediated to us by subjective things.
Consequently, considering reliability in this case means asking how likely what a source is telling you is to be a true statement. To be clear, even a source that is clearly lying can be useful! One of the common misconceptions beginner history students have is to assume that sources come in either ‘unbiased’ or ‘biased’ types and that the former are to be trusted and the latter discarded in all cases. The world is more complex than that, and all sources are biased, the question is merely how. Nevertheless, having a sense of how likely a source is to be relating reliable information is a crucial first step to interpreting what that source is telling you.
There are a few standard questions for establishing reliability. The first and most obvious is to ask, ‘is this source in a position to know?’ Thus we ask, ‘how close to the events is this source?’ Greater distance, through either the number of witnesses (generally first-hand accounts are more reliable than third-hand accounts) or chronological distance (generally a report written at the time is more reliable than one written ten years later is more reliable than one written a century later) generally reduces reliability. For older periods of history, where good eyewitness reports are rare, we often ask questions about where a source got their information – do they have access to (now lost) records? Access to eyewitnesses?
(For those older sources this can get really complex because while we still call, say, Herodotus or Polybius a ‘primary source,’ they are actually essentially secondary or even tertiary sources. So we have to ask these questions about them and about their sources.)
Of course then we have to ask questions about their viewpoint. What are the incentives and motivations the source has in writing? Might the source be mistaken or simply lying about what they are saying? What are the constraints of the genre the source is working in (for instance, ancient royal Egyptian inscriptions, as a genre, are constrained to say positive things about the Pharaoh who has commissioned them; factual accurate is less valued in poetry, etc). How does that viewpoint influence the testimony the source gives?
Now, here’s the tricky curve-ball. So far, I’ve been talking about sources as in terms of being evidence for things that happen in the real world. But sources are also evidence for their own viewpoint (and indirectly, evidence for the general viewpoints of people who think like the source or have similar experiences). So a source that is mistaken about a fact (you may know this because you have far stronger evidence contradicting them) can still tell you important things about what people believed about a fact or what was commonly thought! A source that is straight up lying to you can tell you things are their state of mind, their viewpoints and agenda.
For instance, note above where I casually note that ancient Egyptian inscriptions are constrained to be positive – to the point of lying – about the Pharaoh? How do we know that? We might easily guess, but in practice there are enough points of fairly obvious falsehood that we can use those data-points to better understand the constraints of the entire genre. Of course that doesn’t make pharaonic inscriptions useless – they may not be good evidence for events (although they still have to have some connection to events; a lie has to be at least a little connected to the truth, typically, to have any power), but they are good evidence for how Egyptian rulers presented themselves and their accomplishments, what things they thought were worth bragging about (which is to say, what things would increase their legitimacy and power).
Looping back to the original question asked in the Twitter, we have two figures recorded in our sources for the size of the Greek coalition army at Thermopylae; which is correct? There is, of course, a true answer here (although the precise correct answer is a ‘dark figure’ that we no longer possess the evidence to find). But Herodotus tells us 5,200 and Diodorus tells us c. 7,700. How do we evaluate these?
Well, on the one hand, Herodotus is writing far closer to the event. The Histories of Herodotus were probably composed c. 430 BC, 50 years after the battle; that puts them (barely) within living memory, meaning that Herodotus may well have some of those numbers from eyewitnesses (or alternately, from records and reliable reports of the events made shortly after them). Diodorus, by contrast, is writing in the first century BC, more than 400 years after the event; he cannot possible have eyewitness access and even many first-hand records (like inscriptions or state documents like muster-rolls) will have been lost by then.
Moreover, there is the concern of genre: Diodorus is writing a ‘universal history,’ running from when he places the age of myths down to his present. For him, then, Thermopylae is merely one event among many, albeit a significant one. It is thus doubtful that he took the time to, for instance, comb Greece for evidence like muster-rolls or inscriptions. Instead, it is nearly certain that Diodorus had to rely on other, intermediate sources (including, quite possibly, Herodotus) and so his reliability can be no greater than these intervening sources.
But that’s not quite the end of it. Herodotus has one clear error – he lists no non-Spartiates from Laconia. But Diodorus does and it is clear from even Herodotus’ own narrative (what we’d call ‘internal evidence,’ in this case Hdt. 8.25) that there were non-Spartiates in the Spartan contingent, alongside the famous 300 Spartiates. Here we get into viewpoint, because we know that Herodotus has his reasons to try to amplify the heroics of the 300 Spartiates, given the particular argument he is making. We also know, from comparing other things in Herodotus to stronger evidence (typically archaeological or geographic) that he is not entirely averse to fudging the details if it would make for a better story.
Herodotus also has reason to exaggerate the small size of the Greek force so as to make its stand more impressive. Openly lying about the size of the contingents might well be detected; a Corinthian upon reading less than the correct number of Corinthians at the battle might well call Herodotus on his deception. But simply leaving out some non-citizen Spartans (along with summing up the Locrians with “all they had” without estimating that number) lets Herodotus shrink the apparent size of the force with less risk of detection (Herodotus does not add his figures up at the end of recounting them, so creates no obvious inconsistency).
Meanwhile, Diodorus lists 1,000 men from the Malians, which Herodotus does not. ANd now we ask questions about Diodorus’ sources. After the defeat of the Persian army at Plataea, Thermopylae became a pan-Hellenic culture touchstone (which it still is, often with less than great justification); Greek communities would want to represent themselves as having been associated with it and so it is not hard to imagine communities which were not there claiming to have been there anyway. One of Diodorus’ intervening sources might well have picked that up and carried it forward; Diodorus is no position to be any the wiser and in any event Diodorus (unlike say, Livy or Polybius) is not so critical of his sources generally.
And so most historians have tended to take Herodotus’ numbers on the whole (for the reasons of proximity noted above), but added back in the c. 1,000 perioikoi and 1,000 Locrians (replacing the ‘all they had’ non-number in Herodotus) as reported by Diodorus. Diodorus’ report of the Malians is generally dropped on the assumption that it represents a later, opportunistic addition made in the decades after Herodotus when the dimming of memory would permit such fabrication. Doing that gives us a figure of 7,200 (conveniently close to Diodorus’ total, but differently comprised).
The final step in assessing that figure would be plausibility – after all, we are trying to get as close as we can to a real number in the real world which would have had to obey the real physical laws of the real world. As it stands, an army of c. 7,000 (for by this method we cannot be meaningfully more precise than this) is plausible (unlike Herodotus’ figure for the Persian army, which is ludicrously too large). It is on the right scale for the known capabilities of the communities in question (smaller, because many were still mustering) and well within the capabilities of being supported in the physical space.
Of course the historian’s work is hardly done in reaching that number. From there the work of interpreting that number – what does it mean that the number is about 7,000 – begins. But that’s a discussion for another time.
On to the recommendations!
For those looking for naval nerdery, mostly focused from the Age of Steam to the Second World War, check out the videos on Drachinifel’s youtube channel. I stumbled across his series (as yet still unfinished) on the Guadalcanal campaign (1, 2, and 3) and as quite impressed by the level of detail and care, along with an amusing dry Commonwealth wit. While I’m on the topic, if your tastes run more towards text-based naval nerdery, there is a lot of good interesting stuff on the blog Naval Gazing.
Meanwhile, over at Tod’s Workshop – because it wouldn’t be a Fireside, I guess, without a Tod experimental archaeology test – Tod has run some really interesting tests on the Late Roman plumbata, a weighted war dart. In the two videos (1, 2), he looks at lethality and the throwing pattern and his tests suggest a range of at least 50m, possibly further. The historian in me is somewhat cautious with Tod’s suggestion, in the second video, that we ought to think a throwing string was sometimes used because the technology was known, even though we have no evidence of it – it is certainly possible, but it would hardly be the first case where an extant technology like that didn’t pass into general use.
Because the plumbata is weighted – with a lead weight in the center which gives the dart it’s name – it can really do some damage if it strikes flesh or armor (though unlike the pilum, it does not have a long shaft to wound someone behind a shield). I tend to think he is probably right that the underhanded throw is how the weapon was used. The big advantage of a weighted dart like this is that it is going to fall point-first every time and the weight of that lead ball is going to drive the weapon in. We know that the plumbata is in use at least as early as c. 400 (as Vegetius refers to them) and probably earlier and they were used at least as late as 600 (because they are mentioned in Maurice’s Strategikon).
Finally, I think the latest Net Assessment Podcast is a great listen on the United States military budget and the hard choices that so far don’t seem to be being made. The fact is there are real tradeoffs both in terms of the scope of military spending and also how that is distributed. As Zack Cooper notes, if the United States really wants to pivot to China, that ought to mean significantly cutting spending on the U.S. Army – which has relatively little role in an East Asian conflict – and instead directing those funds to the Navy (and I suppose the Air Force if we must). Making that budget stretch to cover essential security needs means potentially dialing back other commitments.
I do like that they get into some of the direct tradeoffs: nuclear vs. conventional modernization, readiness vs. designing a ‘future force,’ naval vs. land capabilities, and so on (for my own part, my answers to those questions are ‘conventional,’ ‘future’ and ‘naval’ but I am hardly an expert).
But simply pretending there are no compromises to be made with an endlessly inflating military budget is poor planning and not likely to be sustainable in the long run. In all of this, I am reminded of Dwight D. Eisenhower’s ‘Chance for Peace’ in 1953, “Every gun that is made, every warship launched, every rocket fired signifies, in the final sense, a theft from those who hunger and are not fed, those who are cold and not clothed. This world in arms is not spending money alone. It is spending the sweat of its laborers, the genius of its scientists, the hopes of its children.” We should be more careful in what we spend.
For the book recommendation, I first want to note that I have put together a page compiling all of my book recommendations to date, which will go up on the top bar of the front page.
But for the recommendation itself, I was surprised in putting that list together to find that I never recommended Isabel Hull, Absolute Destruction: Military Culture and the Practices of War in Imperial Germany (2006). Hull takes an institutional or ‘strategic culture’ framework in analyzing the behavior of the German army from 1871 through the end of the First World War (1914-1919), though the most sustained analysis is on the German ‘suppression’ campaigns in Southwest Africa (1904-1907). Hull argues that the military culture that the Imperial German Army developed during its creation, informed by the values of its society, its own doctrine and the political structures it was embedded into, encouraged commanders at all levels to proceed through a spiral of escalating violence that predictably led into atrocity.
As the clever reader will swiftly note, Hull’s book is a book about the Nazi regime and the Holocaust, in which neither appear save for in the conclusion. (her end-date being squarely in 1920); Hull openly frames her argument around the idea that “the victory of annihilation, the tendency to “final” solutions, was so predominant in pre-war [WWI] German military culture.” Consequently, the book fits into the debate about the sonderweg – the idea before WWII that Germany had a ‘special path’ to nationhood which after WWII became a touchstone for arguments as to if Germany was uniquely positioned for authoritarianism and the horrors of Nazism, including the holocaust. Hull quite clearly argues in the affirmative, Germany did have elements that made it more disposed to escalating violence and cruelty of that sort. This naturally makes the book controversial among specialists in German history, where the sonderweg is still very much a hot debate.
Regardless of where one falls in that argument (I find myself largely convinced by Hull; institutional culture matters and while no European – or American – military has anything remotely close to a spotless record, the scale of the systemic dysfunction of the Deutsches Heer is remarkable), Hull is equally valuable as a primer on how to think about institutional culture in militaries and how doctrine, training, political systems and cultural attitudes influence the behavior of an army. And then, of course, should you process half of an entire society through that army, as Imperial Germany did, how those values will replicate throughout the society at large.