Collections: Who Were ‘the Celts’ and How Did They (Some of Them) Fight?

This week we’re going to take a bit of a detour to talk about how we should imagine the warriors of Gallic/Celtic armies were equipped and fought. I wanted to write about the topic because the YouTube algorithm served me up a video on it,¹ which isn’t ever fully wrong but struck me as importantly incomplete and with which I had a few issues.

I should also note that, despite the massive length of this post, most of what I’m doing here is the briefest overview of material that – God willing – will be covered in much more depth in my book project once it is done. If you want regular, monthly updates on my progress – well, that’s a perk for my Patrons, who are, in effect, funding it.

Now that video is supposed to promote a book, G. Canestrelli, Celtic Warfare (2022) so I went and read that too. It’s published by Pen and Sword, which as you’ve seen me write before, has very unpredictable output because, as a publisher, they have little in the way of standards. Celtic Warfare is not the worst thing I’ve ever seen them publish, but it’s also certainly not the best.² The book was a bit sad, in a way; Canestrelli clearly has read a lot,³ most of the details are right and most of the errors are small (but some are big, including the central conceit that there is a thing called ‘Celtic warfare’ that we can discuss). But it feels like a book that would have benefited from peer review and a stronger editorial process prepared to cry foul at some of the less well-sourced claims and to push some of the uncertainty into the text.⁴ It is very much, as is normal for Pen and Sword books, a book that adopts a tone of certainty both when it is warranted and when it isn’t. It felt, quite frankly, like Canestrelli could have put together a much better book with just a bit more help and training and I wish he had gotten it (or does get it in the future).

In any case, I thought, rather than write a post complaining about Pen and Sword again (though we’re going to do a bit of that), I’d instead set out my own view, especially since – as blog Patrons who can follow my research work – know, I am currently working on the Gallic-stroke-La Tène material culture section of my own book project right now.

Now I’m going to note one thing here right out of the gate: my focus is mostly going to be on Gallic arms and techniques in the third and second century BC (roughly correlated to the ‘Middle La Tène’ period).⁵ That means this blog post is essentially focused on just about a single chapter of the book in question. We do have evidence that runs earlier than this, indeed much earlier than this, but focus is valuable here especially when we are thinking about the where and who of a military system and its equipment. Fortunately as we’re going to see, the one thing we can see clearly is the stuff of a military system (arms and armor). That said, because I think it is important to sketch out the confines of the known and knowable, uncertainty and the inability of evidence to clarify is going to be a theme in this post.

And if you want to help me afford an aristocratic panoply with ample head and body protection while I battle the forces of excessive certainty in the face of uncertain evidence, you can support this project on Patreon and share it with your friends (and/or enemies). 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, assuming there is still a Twitter by the time this post goes live.

(Bibliography Note: The bibliography on what we might term here La Tène material culture weapons and warfare is, unsurprisingly, dominated by works in French (while works on Celtic-language speakers in Iberia are almost entirely in Spanish and works on those in the British Isles are in English). The standard references, though somewhat aged, on Gallic warfare are J.-L. Brunaux, Guerre et Religion en Gaule, Essai D’Anthropologie Celtique (2004) and J.-L. Brunaux and B. Lambot, Armement et Guerre chez les Gaulois (1987). Beyond that, the question goes to archaeology quite quickly, in volumes that are often very hard to get. The new and definitive work on mail, including La Tène mail is M.A. Wijnhoven, European Mail Armour (2022), staggeringly expensive and worth every bit of it. Probably the best single work on weapons is T. Lejars, La Tène: La Collection Schwab (Bienne, Suisse).  La Tène, Un Site, Un Mythe 3 (2013), a detailed study of roughly a third or so of the total finds from La Tène, including some new typologies; there is to my knowledge a single library copy in the entire United States belonging to the Library of Congress (at time of writing my library has borrowed it for me). Easier to get and equally technical is Brunaux, J.-L, and A. Rapin.  Gournay II: Boucliers et Lances Dépôts et Trophées (1988), notable for advancing the initial typologies for shield bosses and spearheads. On the La Tène shield, the essential article is Gassmann, P.  “Nouvelle approche concernant les datations dendrochonologiques du site éponyme de La Tène (Marin-Epagnier, Suisse).” Annual Review of Swiss Archaeology 90 (2007): 75-88, which doesn’t sound like its about shields, but it is. On helmets, note U. Schaaff, “Keltische Helme” in Anike Helme (1988); P. Connolly, Greece and Rome at War (1981) also has a really good diagram of helmet patterns, but Schaaff is the best typological study. Generally on what we know of culture in this period, the Oxford Handbook of the European Iron Age, C. Haselgrove et al. eds. (?LOL?) is incredibly useful, but also still only available as an ebook via Oxford Academic, a state of affairs that has continued since 2018 (it, in theory, isn’t done yet, but many of the chapters are and are already standard citations in the field; you will probably need some kind of library access to get it). That said, the Handbook‘s chapters do a really good job of stressing how much we do not know, how enormous our guesses often are. Finally, on political structures in the La Tène sphere, N. Roymans, Tribal Societies in Northern Gaul: An Anthropological Perspective (1990) is a decent start, but be aware how conjectural much of it is. Finally, for a state-of-the-debate on Celtic identity, the recent article, R. Pope, “Re-approaching the Celts: Origins, Society and Social Change” JAR 30 (2022) is really valuable, both for the argument it presents but also the ‘potted history’ at the beginning which walks through how this idea has evolved over time.)

Who Are We Talking About?

Now already some of you are noting a curious feature here which is that I keep using the word ‘Gauls’ to describe these folks rather than ‘Celts’ and you are probably wondering why. We’ve actually addressed this question before, but we ought to revisit it here, because I think any approach to ‘Celtic Warfare’ is already potentially begging some pretty important questions (assuming it hasn’t stopped to address them) and, alas, begged the wrong answers (unless it has defined ‘Celtic’ very narrowly). The problem, entirely unaddressed in the original video, is that there is a pretty big gap between what the Greeks meant by the word keltoi, what the keltoi may have meant by the word keltoi and most important what people today understand by the word ‘Celts.’ Instead everyone gets smashed together, with all of the Celtic-language speakers mashed in under the label of ‘Celts,’ a practice that hasn’t been acceptable in serious scholarship for at least 30 years. Let’s talk about why.

From antiquity we have two standard terms. On the one hand, the Greeks encountered a people in the Mediterranean and called them keltoi. From Caesar and Strabo we know that at least some peoples called themselves keltoi (or celtae), though as we’re going to see the people who did this are not actually co-terminus with this military system or with all the people folks (including the original video) think of as Celtic or any identifiable polity or political structure. In particular, Caesar reports that the folks living in what is today France (then Gaul) north of the Garonne and south of the Marne and the Seine called themselves celtae, which he takes to be equivalent to the Latin galli (Caes. BGall. 1.1). Strabo, meanwhile, describes peoples in Spain as both keltoi and also keltiberes (which enters English as Celtiberians, Strabo, Geography 3.2.15) as well as those in Gaul (Geography 4.1ff), but doesn’t make the claim that they call themselves that (instead repeatedly noting these groups broken up into smaller tribal units with their own names). Both Caesar (Caes. BGall 1.1) and Strabo (Geography 4.1.1) go out of their way to stress that the folks they’re talking about do not have the same languages, institutions or mode of life, even those who are, to Strabo, galatikos – ‘Gallic’ or more precisely ‘Galatian-like’ (referring to the sub-group of Gallic peoples the Greeks were the most familiar with).

Galli, rendered into modern English as ‘the Gauls’ (though the latter is not a descendant of that word, but a wholly different derivation), is likewise tricky. We’re fairly sure that both keltoi and galli are Celtic-language words, meaning that (contrary to the video) they’re both probably ‘endonyms,’ (a thing people call themselves) but it is really common for peoples in history to take the endonym of the first group of people they meet and apply it to a much larger group of ‘similar’ (or not so similar) people. The example I use with my students is ‘Frank;’ – it was common in both the Eastern Mediterranean and later in East Asia to use some derivative of ‘Frank’ or ‘Frankish’ to mean ‘Western or Central European’ – the term got applied to the Portuguese in China, and to both Germans and Sicilian Normans during the Crusades. It’s possible that galli in Latin is connected to the Galatai (Greek) or Galatae (Latin), the Galatians, a Celtic-language speaking La Tène material culture group who migrated into Anatolia in the 270s, but a number of etymologies have been proposed. It certainly wouldn’t be the first time the Romans named a massive ethnic group after the first people they met; this is how we get the word ‘Greek’ when the Greeks call themselves Hellenes. So assuming off the bat that all of these different tribal groups that Caesar or Strabo treat as a cultural unity thought of themselves that way is most unwise. The most we know is that if you called some of these folks (but not all of them, as we’ll see) keltoi or galli, they’d say, “yeah, I guess that more or less describes me,” perhaps in the same way describe a Swiss person as ‘European’ isn’t wrong, but it also isn’t quite right.⁶

Surely here linguistics will help us out? If we can identify a Celtic language then surely everyone who speaks that language will have that culture? First, this is yet more question begging; English is the official language of South Sudan and yet the South Sudanese are not English, British or American. Linguistic connections do not always imply ethnic or cultural connections extending beyond language. And, in fact, examining the Celtic language family is a brilliant way to illustrate this.

There is, in fact, a family of Celtic languages and indeed it is only in the sense of languages which you will see me use the word Celtic in a formal way precisely to avoid the giant pickle of confusion we are currently working through. Very briefly, it has been shown linguistically that the various surviving Celtic languages are related to each other and also to the extinct languages of pre-Roman continental Europe that were spoken in Gaul, Noricum and parts of Spain. So far so good, right, we have a nice, perfect match between our keltoi and Celtic-language-speakers, right?

Of course not. That would be easy! Because notice there that Irish, Manx, Scottish Gaelic and Welsh are all Celtic languages. But our sources are actually quite clear that at least the Romans and the Greeks did not consider these folks to be galli or keltoi. Indeed, Strabo explicitly defines the people of Britain against the keltoi as two distinct groups, making it clear he doesn’t think the inhabitants of the British Isles were ‘Celts’ (Geography 4.5.2); Caesar doesn’t either (BGall. 4.21ff). Tacitus sees in the britanniae evidence of German, Iberian and Gallic influence, marking them as distinct from all three, but concludes that Gallic settlement is the most likely cause, a point on which we may be quite certain he is wrong, for reasons discussed just below (Tac. Agr. 11). So the groups described as ‘Celts’ don’t entirely overlap with Celtic language speakers.

Well, surely here the archaeologists can help us out, right? Yes and no. On the one hand, we have a collection of object types, artistic motifs and archaeologically visible patterns that we associate with some of the areas settled by people who our sources regard as ‘Celts’ and who were Celtic language speakers. The older of these two material culture groupings we call ‘Halstatt culture’ after the original type-site in Hallstatt, Austria, though we find Hallstatt culture objects (remember, these are objects, not people, a thing to be relevant in a moment) in a territorial range that forms a sort of crescent shape embracing the northern edges of the Alps, from around 1200 BC to around 500 BC. We then shift to a material culture pattern which may have developed out of late Hallstatt culture which we call La Tène culture after its type-site of La Tène in Switzerland; it runs from around 500 BC (very roughly) to around 50 AD, with lots of subdivisions.

And just about all of the folks our sources will identify as ‘Celts’ or ‘Gauls’ tend to live in areas where where we find, by the third century or so, at least some elements of La Tène material culture (and many in places where they have the full package). So do we at last have a way to identify some ‘Celts,’ by matching wherever we find La Tène material culture?

No. Of course not. That would be easy and history is not easy.

First, not all of the people our sources describe as Celts adopt all or even most of the elements of La Tène material culture. Most notably, the folks in Iberia who were keltoi (according to Strabo) or Celtiberians have some elements of La Tène material culture, but are notably missing others. They don’t have, for instance, the whole La Tène military package – mail in particular is absent in Iberia until the Romans arrive, and the La Tène swords they have are local variations of early La Tène I swords by the third and second centuries, not the La Tène II swords we find in most of the rest of the cultural zone.⁷ The artistic style in ‘Celtic’ Spain is also different and unsurprisingly there’s a lot of Iberian borrowing. As a result, archaeologically, the keltoi of south-western Iberia aren’t some sort of carbon-copy of the keltoi of central France. There’s not no connection here, they are Celtic-language speakers and they have some La Tène stuff, but the Iberian Celtici are quite a bit further from the Helvetii (the folks who probably inhabited the La Tène site) than, say, the Senones.

Meanwhile, we find some La Tène material culture objects in southern Britain, but they don’t fully penetrate the Isles (despite the general assumption that all of the people of Britain and Ireland were Celtic language speakers) and many appear to be expensive, high-status imports. Indeed, while it was once supposed that the arrival of La Tène material culture objects signified some invasion or settlement of Britain by people from Gaul, an analysis of burial patterns⁸ demonstrates pretty clearly that this isn’t happening in this period, because burial practices in southern Britain remain distinct from those on the continent. Instead, we’re seeing trade.

Meanwhile, we find tons of La Tène material culture objects in cultural contexts that we know were neither ‘Celtic’ in any cultural sense nor filled with Celtic-language speakers. The clearest instance of these are in Illyria and Thrace, who spoke Indo-European but not Celtic language (so a language as close to Celtic languages as Latin or Greek or German), where it’s clear that folks adopted at least some La Tène material culture, including weapons and armor. Of course by the third century, when it came to militaria, we’d have the same problem with the Romans, who by the end of the Second Punic War, had adopted a La Tène sword (albeit from Spain and with a different suspension system), a variant of the La Tène shield, a La Tène helmet type (domestically manufactured), and La Tène body armor (mail). If we didn’t have any surviving Latin language material, I am almost certain there would be nationalist pseudo-archaeologists claiming the Roman Empire was clearly some ‘pan-Celtic’ imperial construct on that basis.⁹ And of course in the third century, a Greek varient of the La Tène shield, the thureos, begins showing up everywhere in the Hellenistic East, but that doesn’t make them Celts either (they’d be the first to tell you).

Meanwhile, there’s even more complexity than this, because objects of La Tène material culture aren’t the whole of archaeologically visible culture. There are building habits, burial habits, evidence for social organization and on and on. And those vary significantly within the La Tène material culture zone. I put this in the bibliography and I’m afraid it is a (necessarily) difficult and technical read, but if you want to get a sense of just how complex this can get, check out Rachel Pope’s efforts to define the Celts in the Journal of Archaeological Research (2022). To quote some of her conclusions, “In fact, “Celts” as a historical label does not map neatly onto any archaeological tradition; it overlaps with late Hallstatt traditions in northeast France and less ostentatious archaeologies farther west….Nor did the name “Celt” ever equate to all of Gaul, let alone all of Europe.”

So to be clear, we have Celtic-language speakers who aren’t called Celts by our sources and don’t have La Tène material culture (Ireland, N. Britain), Celtic-language speakers who are called Celts by our sources but don’t have the full La Tène material culture package (Spain, Portugal), non-Celtic language speakers who do have some of the La Tène material culture package but who are clearly not Celts to our sources (Thracians, Illyrians, Dacians, etc.), full La Tène material culture-havers who are explicitly not Celts in our sources (Caesar, specifically) and maybe speak a Celtic-language (the Belgae), and partial La Tène material-culture-havers who do speak a Celtic language but are still explicitly not Celts in our sources (S. Britain). Oh, and while we’re here, by the second century we also have La Tène material culture-havers who probably still speak a Celtic-language and are called Celts/galli by our sources but write inscriptions in Greek (the Galatians) and seem to have different religious structures and folks identified as Celts in our sources who are in the process of ditching large parts of La Tène material culture and learning Latin (Cisalpine Gaul), who might, à la Pope (op. cit.), actually be the direct, local descendants of the ‘original’ Celts.

And then of course we have a band across parts of the Alps and central France where everything lines up: Celtic-language speakers with La Tène material culture who our sources call keltoi or galli and live in a place called Gallia by the Romans. But it would be a mistake to assume this is the cultural ‘heartland’ of a ‘Celtic’ people – indeed, La Tène material culture may be more deeply rooted in more Northern parts of France the Danube region, which has a lot of non-Celtic language speakers in it in this period! Because, to be clear, what we actually have are a host of smaller, tribal societies which share come cultural elements and differ in others, who seem to think of themselves primarily as members of a tribe and who lack notable ‘pan-Celtic’ institutions, to which Greeks and Romans, needing a way to label their neighbors, took whatever ethnic signifiers they had and applied them (over)broadly.

Region	La Tène material culture?	Celtic Language Speakers?	Inhabitants called ‘keltoi‘ or ‘galli‘?
Ireland	No	Yes	No
N. Britain	No	Yes	No
S. Britain	Partial	Yes	No
Belgica	Yes	Probably?	No (by Caesar, at least)
Celtica	Yes	Yes	Yes
Aquitania	No	No	No (but lives in Gallia)
Cisalpine Gaul	Yes	Yes	Yes
Thracia	Some	No	No
Illyria	Some	No	No
S. Portugal	Partial	Yes	Yes (keltoi, not galli)
Central Spain	Partial	Yes	Partial (‘Celtiberian,’ not galli)
Galatia	Yes	Yes (but also Greek)	Yes
Danube	Yes	Both Yes and No	Some

Just about the only combination that does not occur here is that there aren’t any non-Celtic language speakers that our sources think are Celts (but there are Celtic-language speakers they think are not Celts, so it’s hardly an unmitigated win for Strabo’s skill at philology) and even then I wonder if we could see language at a more granular level if we’d find exceptions to that too. Remember when we assess if peoples were Celtic-language speakers, since we don’t have any writing we’re mostly working on place-names and sometimes tribal names, which means what we often know is, “at some point, some Celtic-language speakers probably lived here” but not necessarily when. These sorts of maps of Celtic-language have been termed ‘fossil maps’ and I think that’s a fair characterization of what they tell us.

At no point where all of these people united in a single polity (the closest they get is that most of them get conquered by the Romans) and there’s no indication that they ever saw themselves as a cultural or ethnic unity. And of course we haven’t even gotten into the idea that they might all be somehow closely ethnically related but let’s just go ahead and tag that as ‘very unlikely’ and keep moving.

All of that is to make the point that any treatment of ‘Celtic’ warfare is immediately begging an enormous question because ‘who were the Celts?’ is at best an unanswered question and to be frank, probably an unanswerable question. Crucially, ‘the Celts’ do not share a military system. Warfare among Celtic-language speakers in the British Isles isn’t necessarily based around La Tène material culture, nor is warfare in S. Portugal among peoples identified by our sources as keltoi; both areas seem to have very substantial regional variation. By contrast, the galli of central France and Cisalpine Gaul do seem to share at least substantial elements of a military system with the – according to Caesar – non-celtae of broader Gaul and as well as with the Galatians who live, I must repeat, in Anatolia (having migrated there in the third century). There is thus no ‘Celtic’ military system which maps clearly onto either Celtic-language distribution or peoples described as keltoi by our sources.

Now, some scholars will still use the term ‘Celt’ (or its translations in German or French) and so long as one defines the term that’s fair. Often that will mean finding out that when an author says ‘Celtic’ they mean, “La Tène material culture” or perhaps even more narrowly, peoples who speak Celtic-languages and have La Tène material culture. But of course that’s going to be a definition of Celt and Celtic which is going to cause you more than a little bit of trouble if you break it out in a modern social setting in Ireland, Scotland, Wales or Brittany and is going to confuse a whole bunch of other people unless you define those terms. Meanwhile, if you use ‘Celtic’ as an ethnic, cultural, artistic or military signifier (basically anything but language) and include all Celtic-language speakers, that’s just going to be wrong in quite a few cases.

For my own part, I stick to three terms here and in my other writing: Celtic-language speakers (which covers, surprise, all speakers of a Celtic-language), La Tène material culture (which is not co-terminus with Celtic-language speakers) and finally ‘Gauls’ or ‘Gallic.’ That latter term I find more useful because it has not experienced the nationalist-inspired drift of ‘Celtic’ and does not imply a huge range of Celtic language speakers. Instead the Romans apply it to a pretty narrow group of people with lived in what they termed Cisalpine and Transalpine Gaul, as well as the Galatians who moved into Anatolia (they don’t use the term of Iberian ‘Celts’ or anyone in the British Isles). And conveniently, that captures the ‘everything lines up’ groups pretty well: La Tène material culture-having Celtic-language speakers who get called keltoi (or, of course, galli) by our sources. Those are Gauls. There are, admittedly, a few ‘everything lines up’ groups that don’t get captured by this term, most notably the La Tène material culture Celtic language speakers of the Danube region, so it is hardly perfect. But it at least has the benefit of being clear.

But please note even there, if I say something is ‘Gallic’ what I mean is it appears amongst those people, but is not necessarily restricted to them (indeed, it almost certainly isn’t). History is complicated and when you are dealing with cultures and peoples rather than states, just about any general statement is going to be some degree of wrong.

All of that out of the way, how did the Gauls fight and with what weapons in the third and second century?

Gallic Weapons

And now that we’ve drawn our lines far more narrowly than ‘Celtic’ (and thus far more narrowly than the original video to which I am, in theory, responding), we can actually say some useful things because all of our Gauls share an identifiable La Tène material culture military kit – yes, even the Galatians, half the Mediterranean away (they brought it with them).

Now the modern perception of these fellows is of unarmored barbarians swinging great big swords in an undisciplined mess. That modern perception comes, in part, from our sources, which often lean on those sorts of tropes (e.g. Polyb. 2.33.3; Plut. Cam. 41.4; Polyaenus, Strat. 8.72). It is really striking, by the by, that the these tropes are much more common in the Greek literary tradition, but tend to be absent or less extreme in Latin-language sources and one wonders if familiarity is a major factor in that. Livy, after all, grew up in Cisalpine Gaul (he was born in Patavium, modern Padua) and so may have known better and expected his readers to know better too.

Fortunately, we actually have a surprising amount of evidence for Gallic militaria. Representations of people are relatively uncommon motifs in the La Tène material culture zone, but they do occur and so we do have a handful of visual depictions of Gallic warriors crafted by or for Gauls or a Gallic context; these are far more useful than Greek or Roman depictions which tend to lean heavily on the ‘naked Gaul’ as an artistic motif, far beyond its actual prevalence in Gallic armies. But even more in evidence are archaeological survivals of the actual stuff. The video that sparked this notes funerary deposits – warriors buried with their weapons and armor – as a key source and that is correct; what it inexplicably leaves out are ritual deposits, which are by bulk probably a larger source of objects and information.

Simply put, in much of the La Tène material culture sphere (and indeed, beyond it in some Celt-language speaking cultures (but not others)) the deposition of weapons (but only infrequently armor) was a common ritual activity. Weapons might be deposited as trophies in a ritual precinct on land (possibly in the recognizable classical form of the tropaion); this is for instance how we get a huge deposit of spears and shields from a sanctuary at Gournay-sur-Aronde. Alternately they might be deposited in water-courses of various kinds, such as Lake Neuchâtel at La Tène, where the dominant interpretation of the site is that a wooden bridge out into the lake was constructed for the ritual deposition of objects into the water (including some 166 swords, 270 spears, at least 22 shields and so on). Probably what’s going on here is that success in battle is celebrated by giving part of the spoils to the god by permanently placing them in a ritual space, which might be a sanctified plot of land around a shrine or a body of water. The fairly wide chronological range of deposits suggest not one big deposit of stuff, but lots of little offerings.

The picture that emerges from all of that is a bit different from the popular image. For one, the most common weapon seems not to have been the sword but (and this will surprise no one paying attention) the spear. The La Tène spear was a pretty basic affair, typologically very similar to the Roman hasta or the Greek dory or a hundred other one-handed iron-headed thrusting spears. The wooden hafts on these spears almost never survive (being made of wood), but burial patterns suggest normal lengths around 2.5-3m or so (about 8-10 feet); a single spear from La Tène with the haft still intact was 2.55m long, which for those who don’t know their spear lengths is very typical for one-handed thrusting spears. The haft tended to be around 2-2.5cm thick. The spearheads have a range of shapes, but mostly follow a fairly typical shape – sometimes described as ‘leaf-shaped’ but I find that term less than fully useful. Instead, I’ll just offer a picture:

Some La Tène spears also had metal spear-butts, typically quite small. We generally find around twice as many spearheads in mass deposits as spear-butts, which might mean spears were broken before deposition and only the upper part included, but I think – as this pattern recurs in a range of societies – it probably means that a metal spear butt was optional. Some spears had it, others didn’t. Nevertheless, it was the spear, not the sword, which was the mandatory weapon of the Gallic warrior.

That impression is confirmed by artwork from the La Tène material culture sphere (and earlier Halstatt culture artwork too), where when we see infantry in procession they carry spears but swords may or may not be visible. Thus for instance the procession on the Gundestrup Cauldron¹⁰ all have spears and this motif of spear-carrying warriors with the distinctive large La Tène shield is not uncommon in La Tène artwork once one accounts for how rare representations of humans are.

The La Tène sword was the next key weapon and these are quite common in deposits too. They occur in ritual deposits somewhat less than spears, but at similar rates in burial deposits, which suggests, to me at least, that while the sword was more expensive than the spear (it would have been, it uses a lot more metal), it was probably no less common and most warriors carried both. La Tène sword suspensions are particular and different from Roman or Greek sword suspensions, which is a handy typological indicator, but it hangs the sword at the waist just like everyone else.

The development of La Tène swords proceeds in three fairly distinctive phases (Brunaux and Lambot (1978) have the handiest chart of this). Early La Tène swords (very roughly pre-third century, but later in Spain) tend to be shorter and come to sharp points, with a more clearly pronounced mid-ridge or centerline. Over time, those swords start to get longer, and by the Middle La Tène (third and second century) they’ve reached about 60-75cm of blade length, with a typically 14-16cm hilt (so total lengths normally range from around 75-90cm). By classical standards, that’s long, but clever readers will note that those are almost exactly the dimensions of medieval one-handed arming swords. Weight-wise, I suspect the originals were of similar or slightly lighter weight than the typical c. 1kg medieval arming sword; many La Tène swords survive but rust may have reduced their weight from the original (they tend to cluster around 550-600g when complete and not rusted into their sheaths; note that this would be mass with all of the non-metal components long since rotted away).

The shape evolution here is interesting. Early La Tène swords, as noted, come to sharp points; the Iberian variants keep this feature which then passes to the Roman gladius Hispaniensis. But in the broader La Tène cultural sphere those sharp points give way to a more rounded (but still effective) thrusting point in the Middle La Tène (third and early second centuries, roughly) and then to blunter tips and longer cutting blades in the Late La Tène (late second and first centuries). Sword length increases steadily over time as well. So what we see is a design drift from early La Tène swords which seem to owe at least some of their size and shape to bronze forebears (all of these swords are in iron), but get longer as Gallic smiths get more confident with their materials. At the same time, they shift from multi-purpose cut-and-thrust swords to swords that can thrust but are built for the cut.

That cut-emphasis is often presented as something ‘barbaric’ but it makes good battlefield sense in the conditions these would be used. As we’re going to see in just a second, inside the La Tène material culture sphere, the most likely enemy was another warrior with the La Tène material culture kit. And he was probably not very well armored. A cut against an unarmored opponent is far more likely to disable them – to remove them as a threat – far quicker than a thrust, even if both produce lethal wounds. So if you think your opponent is going to be unarmored or lightly armored, going for a weapon that cuts well is a smart move. And these La Tène swords would have cut well.

Quality – and this is going to be a trend – varies wildly in these swords. Now some caveats are necessary: some swords in ritual deposits may never have been intended to be used, and some of them have very poor metallurgy (on this topic, see Pleiner (1993)). But La Tène swords that have been examined run the gamut from some of the lowest quality swords of antiquity all the way to some of the best of the period. The notion – peddled by Polybius and Plutarch – that Gallic swords bend on the first strike is almost certainly nonsense. These swords worked and the Romans adopted them twice (the Roman gladius, as mentioned, is a variant of the early La Tène sword, while the Roman spatha is a variant of the late La Tène sword). But the designs, especially in the Middle and Late La Tène would have been pretty demanding on the metallurgy of these swords – the longer you make a sword, the more strain you are putting on a set amount of metal – and some of these swords are just not very good.

The point here isn’t that Gallic swords sucked – they didn’t – but that here, as with a lot of La Tène material culture military kit, we see a big impact of social stratification, with a huge gap between the haves and have-nots, both of whom were on the battlefield. I think it’s fair to say that the La Tène period seems to see an expansion of who fights in these armies (and the armies get big as a result), but I don’t think – as the referent video suggests – we can say this was an opening of the ‘warrior caste’ so much as it seems to be an increasing willingness to mobilize the still-poor and still-low-status peasantry (these are, to be clear, all agricultural societies; the Gauls were farmers).

That leaves the La Tène shield, a large oval body-shield. We actually have some wooden fragments and a few decently preserved examples from La Tène of this, along with a ton of metal elements from La Tène, Gournay-sur-Aronde and sundry burial and ritual deposits, so we know quite a lot about the La Tène shield.

The La Tène shield sits in the same family as the Roman scutum and the Greek thureos and is probably the progenitor of the other two; this is ‘daddy oval shield.’ It is flat-faced (unlike the curved scutum) and at c. 110cm by c. 53cm,¹¹ making it a bit smaller than the Roman scutum and a bit bigger than the Greek thureos. Unlike the scutum, which was manufactured via laminated wooden strips (‘plywood’ construction) the La Tène shield was constructed out of two wood planks, glued together, with a hide front facing, a leather strip binding the edges and a metal boss in the center. The wooden core had a gap dead-center of mass for the hand; this was then covered by a wooden reinforcing ridge (the Romans call it a spina) that runs down the center of the shield and is nailed into place. It widens to cover the hand-gap at the center and a metal boss (a metal plate) goes over it, and is riveted through the shield to connect to a metal bar on the back side around which is built the handgrip (in wood or leather). Riveting through like that holds everything together. Metal rims to the shield (as Polybius reports about the Roman scutum) have been posited, but none to my knowledge have been recovered in a La Tène context, so I’m inclined to think they’re not standard.

Compared to the scutum, this shield would be a bit less useful at dealing with ranged projectiles because you can’t place your full body into the curve of the shield; that fact was noted by ancient sources (Polyb. 2.30.3; Livy 38.21.4). At the same time, it was probably a lot lighter than the scutum (perhaps 7kg instead of the scutum‘s 10kg), which would have made it handier in more fluid close-combat. And of course we have to note that it sure seems like almost everyone who encountered this shield decided in fairly short order to adopt it for at least some of their troops. The ubiquitous Hellenistic thureophori ‘medium infantry’ were defined by using it¹² rather than the indigenous Greek aspis and pelte and the Romans adapted it whole-hog, plus it shows up in all sorts of non-state contexts in Northern/Central/Western Europe. It was a fantastically successful design.

Gallic warriors seem pretty clearly to also have carried javelins. The Greeks and Romans picked up a Celtic-language word for javelin, gaesum and occasionally note that its technical meaning is a Gallic javelin. However, this is one of those places were caution is necessary because an actual look at the way they use this word shows they it gets used as much (if not more) of clearly non-La Tène javelins (e.g. Livy 8.8.5, 9.36.6, 26.6.5, 28.45.16; Sen. Phaed. 111). This sort of problem, where ancient authors are not even remotely consistent with how they use terminology and avoid technical terms like the plague, persists across basically every kind of weapon and armor (we’ll come back to it in a second with mail), meaning that typically we do not know if there was a technical name for a given type of equipment.

Canestrelli’s book attempts to equate the gaesum with the soliferreum and this will not do.¹³ His evidene is that Julius Pollux, a late second-century (AD!) sophist who wrote what was essentially a giant thesuarus equates them; that is already rendered more than a little weak by the fact that Pollux is writing centuries after both weapons vanished but also given that after putting “a spear of all iron, called a gaisos” in his thesaurus he adds, “and it is Libyan.”¹⁴ The soliferreum, an all-iron javelin, is actually a pretty well known object type which occurs fairly frequently in Spain among both Celtic- and non-Celtic-language speakers.¹⁵ Where it doesn’t occur is in most of the La Tène material culture sphere. The one odd exception, with Canestrelli cites, are a few that occur in a late 6th/early 5th century Iron Age burial complex in Mailhac, Aude, France. French geography afficionados may already note the problem: this is quite close to Spain, where the soliferreum is much more common and might have been with the territorial range of the Aquitani, who we are told by our sources were more of an Iberian-people (and spoke a predecessor language to Basque). This is very thin evidence indeed to treat this weapon as ‘Celtic,’ as Canestrelli does, in any period, though it may well have originated in Southern France.

That’s not to say javelins are rare in the La Tène material culture sphere. They’re not! They are quite common, depending on when you are looking. As Lejars (op. cit.) notes, javelins of the VIa subtype (he has a typology for them) are common in La Tène A (Early La Tène), fade away as we get into the Middle La Tène, and then by the end of the Middle La Tène we see a new javelin-head type, VIb (they have much longer sockets) appear, joined in the Late La Tène by copies of the Roman socketed pilum (but not its heavier, tanged variety). The form of these is pretty tpical: there’s a wooden haft and the only metal element is the tip; for VIa javelins, that’s quite small whereas VIb javelins have a long metal socket (the similarity to the pilum seems not accidental). And it sure seems like there’s a period from the mid-fourth century or through the third century where javelins are fairly rare and strikingly that is when we get reports of battles where Gallic armies lack missile weapons (Polyb. 2.30.1-4, Livy 38.21.4, cf. Caes. BGall 1.26, 5.34, etc. and Diod. Sic. 5.30.4, etc.)

Note however that even the poorest of Gallic warriors seem like they probably had at least a shield and a spear; probably most had a sword too. The quality may have varied a lot, but that variance is nothing compared to what we see in armor.

Gallic Armor

Let’s start with the obvious thing: did the Gauls fight naked? Some did, but it wasn’t the common mode of combat. That said, I think the referenced video errs badly in assigning this practice to just one tribe (the Gaesatae).

This is a case where we need to be pretty careful with our Greek and Roman sources; the ‘naked Gaul’ was both a literary and artistic trope and it seems clear that Greek and Roman artists and writers blew an unusual cultural practice out of all proportion in constructing a Gallic ‘other’ for their audiences. But we have some reliable reports of naked Gallic warriors too. Polybius reports one of the four tribes at the Battle of Telamon (225), the Gaesatae, fought naked (Polyb. 2.28.4-8). Diodorus reports a range of Gallic clothing when fighting, from nude to clothed to armored; the referent video assumes Diodorus is talking about the Gaesatae in the first case, but he makes no such specification (Diod. Sic. 5.29ff). Indeed, Polybius also reports at least some of the Gauls in Hannibal’s army to be naked (Polyb. 3.114.4) but Livy, in a rare instance of breaking with Polybius, instead describes them only as naked to the navel (Livy 22.46.6), so they apparently had trousers (the Gauls wore trousers); this is one of quite a few instances where Latin literary tradition sands down some of the ‘othering’ of the Greek literary tradition when it comes to Gauls. There are sundry other references; of note, Caesar never describes naked Gallic warriors but does describe naked German warriors, among the Suebi (Caes. BGall. 4.1), but in training, not battle.

The ‘naked Gaul’ is a super-duper common visual motif in Greek and Roman artwork, but quite rare in La Tène artwork. Of course the caveat that people in general are rare motifs in La Tène artwork is necessary. That said, it’s not an unknown motif either. A fifth century Hallstatt scabbard depicted in Brunaux, Les Gaulois: Sanctuaires et rites (1986) looks to me like the warriors have uncovered chests (but their bodies are mostly obscured by shields). Far clearer evidence is the Braganza Brooch (British Museum inv. 2001,0501.1), a fantastic and unique piece of artwork which shows a nude warrior with the La Tène material culture kit. Reading this object is complicated; it probably comes from the Iberian peninsula and reflects a blending of La Tène, Iberian and Greek visual elements, but it pretty clearly belongs in many of its visual motifs to a La Tène context. The reason I think that’s useful is that it means this is an object produced in a La Tène material culture environment where folks will have known what a warrior of this sort looks like. If they thought depicting him nude was reasonable, well it was probably because occasionally, rarely, warriors would go nude or nearly so.

But most Gauls didn’t fight nude, so what sort of armor and protection did they wear? Well, if we mean most Gauls, the answer is ‘not much.’ But Gallic aristocrats were some of the best armored fellows on the ancient battlefield; the gap in protection and equipment cost is staggering.

Let’s start with the aristocrat. The wealthiest sort of Gaul – typically the kind that could afford a horse – was pretty well armored with a metal helmet and mail armor. La Tène helmets are, in English-language scholarship generally divided into two types, ‘Montefortino’ and ‘Coolus’ types, the former defined by the presence of a knob at the crest of the helmet and the latter by its absence, both of which get adopted by the Romans but at different times. This typology isn’t used outside of the English language scholarship very much and that’s because it isn’t very informative and in any case is far better suited to the Roman variants of these helmets than their La Tène originals.

But as helmets, both types are pretty serviceable, manufactured in both copper-alloy (bronze) and iron in the La Tène material culture sphere, with the latter steadily replacing the former; La Tène smiths made the switch to iron as a primary helmet material earlier than Greek or Roman ones did. Early La Tène helmets sometimes have quite high crests (think almost traffic-cone shaped) in the fifth and fourth centuries, but by the Middle La Tène these have diminished quite a bit, with just a bit of elongation and a knob attached to the top of the helmet as a separate piece held in place by a pin that was penned down on the inside of the helmet bowl. The average quality of these helmets drops over time, suggesting armies reaching for manpower in the lower classes, but even by the third and second century these tend to be noticeably lighter than their Roman equivalents on the lower end. But on the higher end, well, no surviving Roman republican helmet is anywhere near as fantastically decorated as the Casque d’Agris or the Ciumeşti helmet.

One note in the referent video is the implication that cheek guards come late and this is correct but some clarification is required. Cheek-guards don’t seem to appear on early La Tène material culture helmets, but we’re talking very early, fifth and early fourth century helmets. By the time we’re at La Tène B (c. 400 and following) cheek-guards become increasingly common and then effectively standard. The other tricky aspect here is that early helmets do feature attachments for a chin-strap (to hold the helmet on) and in some cases the difference between a small hinge for a cheek-guard and a spot to attach a strap can be tricky to tell apart, especially because helmets seperated from their cheek-guards aren’t exactly rare (even in cases where we can be quite sure they originally had them). Still, this was a good detail that was correct. The video uses a copper-alloy helmet as a visual example (it’s a Roman-style montefortino, a Roman copy of the La Tène material culture original), but it should be noted that by 300 or so, iron helmets dominate in La Tène contexts.

What is striking though is that we have fair reason to suppose not every Gallic warrior would have had a metal helmet. Notice, for instance, on the Gundestrup Cauldron; the cavalrymen have the distinctive knob-topped and decorated ‘Montefortino’ helmets but the infantrymen do not, instead having a head covering that looks to be the same material as their trousers (perhaps they have wrapped their head in thickened cloth). Thiery Lejars notes, in terms of prevalence, that “the use of the helmet remains exceptional. It is necessary to wait to the Late La Tène in order to find a significant trace of it.”¹⁶ Radomír Pleiner’s study of the La Tène sword leads him to do a pretty sweeping survey of La Tène material culture flat cemetery graves and he broadly concludes that helmets were confined to the burials of ‘chieftains’ in all but the late La Tène. By way of one example, 67 Middle La Tène graves at Bellinzona-Giubiasco contained 11 sword and 11 spear burials, but just one burial with a helmet; the Late La Tène phase at the same site had 97 graves, of which 23 had weapons but only nine had helmets (all but one of the helmet burials also had a sword and all but two had both a sword and a spear).¹⁷ It seems very likely that even something as basic as helmets were not universally available for all shock infantry among Gallic peoples.

If that’s true of helmets, it is profoundly more true of mail armor. I should note, the referent video makes a bit of a deal about the Romans calling mail lorica gallica not lorica hamata and I have to object; this is a question that the definitive treatment on ancient mail – M.A. Wijnhoven’s European Mail Armour (2022) (reviewed by me here; it’s fantastic, but alas, eye-poppingly expensive. Fortunately, if you review a book, you get to keep the review copy) – spends fully ten pages discussing and comes to nothing like so certain a conclusion. For what it is worth, he concludes that the answer is probably unknowable except that Roman mail armor was a sort of lorica and that the Greeks called it a ἁλθσιδωτός θώραξ (armor of chains), which has a Latin mirror in the phase lorica catena. Lorica hamata or variations on that theme occur three times in Vergil, and a few more times in other sources (mostly poets) but never in a way where we can be entirely sure they mean mail Likewise, lorica catena twice in Statius and in a few other places (also mostly poets), but never in a way where we can be entirely sure they mean mail. Meanwhile ἁλθσιδωτός θώραξ (chain armor) is used with some consistency in Greek where we can be sure it means mail. Prose authors tend to just say lorica without specification. To my knowledge, it’s called lorica Gallica once, by Varro.¹⁸ If the Romans had a technical term for this armor, we cannot be sure of it and any Celtic-language term for it is lost to us.

In any case, the La Tène mail armor is easy to describe and really hard to make. Mail is effectively a metal fabric composed of joined rings; in this case (and indeed all along the European-Mediterranean-West Asian mail tradition) alternating rows of solid rings and rings closed by a rivet.¹⁹ The rings are joined in a 4-in1 pattern (each ring intersects four others). Armor rings were exclusively produced in iron (later steel, but in this period, iron). La Tène and Roman mail was constructed (that is, the rings were put together) ‘in the flat’ without much in the way of shaping. Think of a flat sheet joined to make a rough ‘tube’ of fabric rather than a sewn and tailored garment.

The result was a ‘tunic’ of mail (with an opening for the head), which extended to just above the knees, generally without sleeves (but sometimes with ‘false sleeves,’ which is to say a bit of mail that extended out over the shoulders to offer some upper-arm protection). In some cases, the mail was fastened in tube-and-yoke style, in other cases the shoulder elements were an entire second layer. These look really similar in artwork and it is sometimes very hard to tell them apart. Given how little mail survives in the BC, it’s very hard to know which style predominated (the Romans seem to be to have preferred shoulder-double style, but not exclusively by any means).

This armor appears first in the archaeological record in the late fourth or early third centuries BC (dating is hard) and spreads rapidly, probably – but not certainly – from an origin point on the upper Danube somewhere. It reaches southern France by the 220s (if not earlier) and Rome probably around 225.²⁰ Mail is the only preserved body armor associated with La Tène material culture; bronze breastplates were known in the earlier Halstatt period.

Mail is really effective armor,²¹ particularly against cutting weapons. But it’s also really expensive.²² We don’t have any good price data from the ancient world, but the medieval comparanda suggests a good mail shirt might be at least as expensive as a specially bred warhorse (it is in the seventh century Lex Ripuaria). The thing is, good mail is made up of very small links, generally not much larger than a centimeter across and often much smaller. With rings that small, you might need something like 40,000 or so of them to make a complete shirt. Really fancy mail might use even smaller, finer rings and some mail shirts have ring-counts above 100,000. Each of those rings needs to be made individually, by hand, and then assembled, by hand.²³

Needless to say, such armor was out of reach of all but the wealthiest of people in the La Tène material culture sphere. Mail finds are fantastically rare in La Tène material culture contexts. Now part of that is just that mail doesn’t preserve as well as other objects (made up of tiny iron rings, it rusts away more easily), but even then it is damn rare. The sanctuary site at Ribemont-sur-Ancre, for instance, contained 175 spears, 60 shields, 52 sword-scabbards, six swords, 49 belts and…one set of mail fragments. Mail finds are substantially rarer than helmet finds which, as noted, are substantially rarer than weapon finds. A quick page through Wijnhoven’s (op. cit.) catalog – an exhaustive list of all La Tène, Roman and early medieval mail finds, suggests we may have something like 40 finds of pre-Roman mail all told, in all regions.²⁴ Pleiner’s sword study looked at 1616 La Tène culture flat graves and reports 297 burials with weapons, 33 with helmets and one with mail (the famous Ciumești burial).

Meanwhile, while as noted, Gallic warriors were only infrequently reported to be naked in our sources, they were very frequently reported to be unarmored. One interesting note is that Appian describes the Galatian cavalry at Magnesia as armored (κατάφρακτοι), but does not say this about the Galatian infantry (App. Syr. 6.32; cf. Livy 37.40.5). Likewise, looking at the Gundestrup Cauldron, the horsemen wear a tunic which does not cover the legs (like a mail shirt would) while the infantry have visible trousers, whcih might be the artist representing mailed cavalry but unarmored (at least in metal) infantry.

That leaves armor made of something perishable, which is more likely to be textile – multiple, quilted layers – rather than leather (as the popular imagination might have it). The problem is immediate: our sources don’t generally report those sorts of defenses. Diodorus (5.30.1-3) presents the options as mail, shirts or nothing and he clearly excludes ‘shirts’ from the category of ‘armor.’ Strabo somewhat cryptically says that the common armor among the Lusitani in Spain were linothoraces (Strabo 3.4.15) and one wonders if he is observing some sort of more common textile-based protection. Archaeology is generally no help here as textile doesn’t survive. On this point, Canestrelli looks at the balance of the representational evidence and concludes that textile armor for at least some of the poorer warriors was likely and on this point I agree, though I think we need to be cautious with just how weak our evidence is here.

The point here is that Gallic equipment was much more strongly stratified, as best we can tell, than their period-equivalent competitors. Gallic infantry were shock infantry (see below) but apparently often went without mail; indeed the evidence seems to imply they often went without metal helmets. In the settled societies of the Mediterranean, skirmish infantry with bows or javelins might be this lightly armored, but heavy infantry was, well, heavy.

One way to read this is as Gallic societies ‘opening up’ the ‘warrior class’ to poorer Gauls, but I think this reading is probably wrong (though it is hard to know with confidence). First off, we should be careful in assuming there is a ‘warrior class,’ as opposed to, for instance, a landholding aristocracy that exercised leadership and status display in peace and war.²⁵ Though the Romans will cut it short, I think there is evidence that this a period of consolidating power – the halting, first steps of state formation – in the La Tène material culture zone.²⁶ Caesar certainly gives the impression that by the time he is in Gaul in the 50s BC, we are seeing some truly ‘big men’ emerge in these societies who can mobilize armies of clients and supporters; that’s a pretty normal stage in state formation. Eventually one of those big men would consolidate power and become king (something Caesar says other Gallic elites are actively worried about, e.g. Caes. BGall. 1.3-4, 7.4) leading to the formation of a state. In that context, let me suggest that what we are seeing is not the egalitarian opening of the ‘warrior class’ but rather that Gallic elites are becoming strong enough to conscript their peasants en masse into tribal levies.

Now that isn’t to say that anywhere in the La Tène material culture zone was as stratified as Rome or the Hellenistic kingdoms. They weren’t, if for no other reason than no one in the La Tène material culture zone was anywhere near remotely as rich as Hellenistic kings or the sort of Romans who served in the Senate. Rather what we seem to be missing is the sort of broad afluent class – the assidui at Rome or the zeugitai in Athens – who could afford armor and heavier military equipment, but not horses. Scholars differ on precisely how broad that well-to-do-but-not-rich class of freeholding farmer was in both the Roman Republic and the Greek poleis; I tend to envision it as somewhat broader with my own effort to speculatively model the Roman census classes suggesting to me that the ‘first class’ (who by Polybius’ day were required to wear mail) made up perhaps 20-30% of Rome’s assidui and thus perhaps 15-25% of the adult male Roman free citizen population.²⁷ The matching fellows in Gaul appear to have been quite a bit poorer (while certainly grave good assemblages confirm that the elites of the region could be very wealthy indeed).

How Did They Fight?

And that leads neatly into the question of how they fought. The refent video settles on ‘dynamism’ as the framing for warfare in the La Tène material culture zone and perhaps compared to the early and high Roman Empire I might buy that, but more broadly I’m afraid I see few signs that warfare in Gaul from the fifth to the first century was particularly dynamic, though it also wasn’t particularly static. In so far as we can tell, a fighting system emerged around one-handed thrusting spears and large, center-grip shields focused on shock engagements in the fifth century at the latest and was still mostly structured like that in the first century.

In the same time, the Romans ditched nearly all of their indigenous Italic equipment and adopted wholesale the Greek (or perhaps Gallic!) cavalry model (Polyb. 6.25.3), a Spanish sword, a Gallic helmet, Gallic body-armor, a Gallic or Italo-Gallic shield, introduced an new kind of light infantry (the velites), created a legion based on maniples before moving to a legion based on cohorts and also formed and perfected a system for the mass recruitment of citizen soldiers before abandoning that system in favor of a system of semi-professionals serving for pay before transmuting that system right at the end of the period into a system of long-service professionals serving as a career. They also adopted complex oared warships and learned to fight with catapults. Five centuries is a really long time. Absolutely, warfare in Trans- and Cisalpine Gaul is not static, but I’m not sure I’d say it is especially dynamic either.

That said, both the video in question and Canestrelli’s book push back very helpfully against a popular image of Gauls at war as untutored barbarian idiots charging blindly without reason. Gallic warfare may not have been especially static or dynamic but it was not stupid. La Tène weaponry worked and the military system it was attached to worked which is why we see such eagerness to adopt elements of it outside of the La Tène material culture sphere. In favorable circumstances, Gallic armies – that is, armies with La Tène material culture stuff (remember how we’re defining ‘Gaul’ here) – could and did overwhelm and defeat Roman, Greek and Macedonian armies.

Our sources are actually pretty clear on how Gauls fought. There’s a repeated motif of aristocratic display – challenges to one-on-one duels, that kind of thing – which may seem silly but has valuable morale and social cohesion value in these sorts of society (see below on recruitment). But the main action was generally an infanty shock action. Gallic infantry fought in relatively close ranks – Caesar describes the Helvetii formation as a confertissima acies, “a most dense battleline,” before it ‘made a phalanx’ (phalange facta, Caes. BGall 1.24, note also Livy 10.29.6-7, 34.46.9-10; 35.5.7, etc.). So it’s pretty safe to say the main Gallic body of infantry typically fought in close order, though we shouldn’t overstate the level of discipline this implies: Greek hoplites also fought in close order and were, in the classical period, almost entirely untutored amateurs. That said we also shouldn’t, as past scholarship might, unthinkingly overemphasize some difference between what the Greeks and Gauls were doing; though the shields are rather different, a Greek phalanx and a Gallic battleline are doing substantially the same thing, though the latter has a lot less armor doing it.

A repeated motif is that Gallic armies tended to either win in the first rush or quickly come apart (e.g. Livy 10.29.8-11, but this recurs in a ton of places). That is presented, particularly by Greek authors, as a ‘barbarians” lack of courage but to be frank given how little armor these guys had, it meaks a lot of sense. That first onset of a dense-packed, well-shielded battleline often would just win the battle, but if it didn’t and the issue came down to attritional close combat, the guys wearing very little armor were going to be in a bad way. Moreover, ‘win at the first onset or fall apart’ is also probably a pretty good description of how hoplite battles might work, though of course no Greek would descibe them that way.

How were these armies organized? We mostly don’t know. This is a point where I think Canestrelli errs in using weak evidence to avoid saying, “we dont know.” Caesar presents a quick potted description of Gallic social classes (Caes. BGall. 6.13-15) which describes a not-entirely-implausible social structure. There is an established, professional priesthood (the Druids) and a military aristocracy. Nearly the whole of the commons are politically dependent on the aristocrats and Caesar, at least, views the system as extremely heirarchical (even by Roman standards) with the commons having no political voice. He notes that when the aristocrats go to war, they do so with their retainers (Latin: clientes – clients; Caesar is working by analogy to Roman social systems, which should caution us!) arrayed around them. This is a model that makes sense, but attempting to apply it more broadly chronologically or geographically runs into immediate problems. Some folks the Romans call Gauls (specifically, the Galatians) don’t seem to have druids. Our sources tell us that this strong heirarchy is absent among the Germanic peoples (BGall. 6.22; Tac. Ger. 7.11), some areas of which are well within the La Tène material culture zone. And looking earlier, there are substantial reasons to believe that early Gallic society may have been meaningfully matrifocal in at least some respects.²⁸

Still, generally, with all due caution in hand (and there’s a lot of it, see T. Moore in the Oxford Handbook (2018) on this), it does seem like we can imagine a society whose military activities are organized around a handful of economic, social and military elites who raise military force not through formal institutions but through networks of clientage or even potentially something like vassalage (but please note we’re using that term by analogy; carelessly assuming the Gauls looked like medieval Frenchmen was a common mistake in 19th century scholarship). The exact nature of these relationships are lost to us, but they fit fairly well with the extreme degree of stratification we see among the armed populace: aristocrats with their mail, helmets and exquisite, pattern-welded swords next to their retainers, wearing at best textile armor with swords of the same design but far inferior make.

Beyond this description though, we smash into the dark very fast. Canestrelli proposes that the caterva was the basic constituent unit of Gallic armies and I’m afraid that dog won’t hunt. Caterva is not a Celtic-language root word (it’s got a near-peer in Umbrian, an Italic language) and it is a general word meaning crowd/throng/multitude/rabble/mob,²⁹ which also gets applied to ‘barbarian’ armies by Latin writers. Canestrelli thus points to Isidore of Seville (d. 636 AD; e.g. Isidore, Origines 9.3.46, “Properly we speak of a phalanx of Macedonians, a caterva of Gauls and a legio of our forces.” This is not a technical unit description.) and Vegetius (4th cent. AD) to establish this usage, men writing centuries after this military system, which the Romans never much bothered to understand, had ceased to exist.

He follows that up with an effort to reverse engineer unit structure from the loot reported by Valerius Antias as reported by Livy from a battle in 191. I don’t mean to beat up the fellow too much, but I think its instructive to walk through why this approach is ill-advised. First, we have no way to know if the defeated Gauls, in this case the Boii, were in any way typical. Next, we have no way to know what slice of the actual loot the general (Publius Cornelius Scipio Nasica, cos. 191)³⁰ showed up in the triumph as it wouldn’t be unreasonable to suppose a soldier or two might have made off with some of the finer pieces. Then we don’t know what percentage of gear the Boii had that was actually made loot or if that slice was representative; Livy says that he had captured ‘the entire population’ but ancient sources love to say that even when it is demonstrably untrue. Next, we have the problem that this is being reported by Valerius Antias who – well, let me just quote Livy, who relays this story to us and what he says at the beginning of this very story: “Valerius Antias writes that thirty-eight thousand of the enemy were slain…though in the numbers written there is little trust, because in exaggeration no one is less restrained” (Livy 36.38.7). Livy feels the need, in the sentence that follows, to justify calling it a major victory at all – that’s how little trust he has in Valerius Antias.³¹

To me the most likely organization is one made up of irregular units, organized around individual aristocrats based on personal connections, both vertical (lower-class clients follow their upper-class aristocratic patron) and horizontal (clan, family and friendship ties bond aristocrats and their retinues together) forming a relatively cohesive ‘tribal’ army. It is a system of military organization that shows up a lot when looking at non-state societies which lack formal institutions for conscription or mobilization and what we are told is consistent with that. But we need to be really clear just how dark this room is and how little we know. Reading Livy (35.5.10)’s reference to Gallic duces in the army as ‘officers’ (as Canestrelli does) for instance is unwise; duces is about the blandest, least specific word for ‘leaders’ Livy could have used and could easily just mean ‘aristocrats’ or ‘leading men.’

So in conclusion on the one hand I applaud the effort to save the Gauls (and more broadly ancient Celtic-Language speakers) from the ‘mad barbarians’ tropes, but we must be careful in how we do it. Yes, older scholarship made a mistake by filling in the gaps of the ancient evidence with whatever romantic barbarian nonsense that happened to flatter their nationalism. We do no better, however, by filling in the gaps with whatever happens to flatter our preconceptions either. We need to be honest about the existence of gaps and the profoundly unknowable.

And if anything, that’s the lesson I want to offer as a take-away: admit uncertainty. There are many ancient peoples about whom we know little and about whom, sadly, we will always know little, because their writings (if they wrote) do not survive. What we should not do is reach for every scrap of ‘evidence’ no matter how weak or flimsy and pretend that we have built a secure structure out of the evidentiary twigs. Learning new things about Europe’s iron-age Celtic-Language speakers is possible; ever knowing a lot about them – the way we can know things about Greece or Rome or Persia or even late Bronze Age Egypt and Mesopotamia – knowing that kind of ‘a lot’ is almost certainly forever out of reach.

1. I am not a regular watcher of the channel in question. I will say that in some of his other videos I’ve seen there is a trend towards a level of certainty that the evidence doesn’t permit as well as a tendency to draw equivalences between ancient and modern peoples which I think are unwise to draw over such long spans of time. At the same time, he usually isn’t just comically wrong like some other channels. Be wary of confidence, I suppose, on topics about which there can be little confidence.
2. The best award probably goes to either D. Hoyos, Carthage’s Other Wars (2003) or P. Johstono, The Army of Ptolemaic Egypt, 323-204 BC (2020), both very solid, scholarly works well worth your time by authors who know their business. The thing about a press with no standards is that it may still publish very good books by accident.
3. Though there are some gaps in the bibliography which probably would not make it through peer review.
4. And also copy-edit better. There are errors of both the typo and format variety, particularly in the bibliography where, for instance, the titles of articles are italicized instead of places in quotes as is customary in English and years of publication are not uniformly included. Now, I make a ton of typos here on this blog, because this is an informal venue and I don’t have a copy-editor, but an actual press which is printing actual physical books being sold for actual money should have someone whose job it is to check these things, especially if their stock and trade is publishing works by less-academically-credentialed scholars.
5. The La Tène chronologies vary; Brunaux & Lambot (1987) actually has a really handy chart. Generally when I say ‘Middle La Tène’ I mean what you’ll see called La Tène C (1 and 2) or La Tène II running very roughly from around 300 to around 100. In French scholarship, La Tène moyenne is often understood more narrowly, from around 250 to around 150. Periodization of archaeological artifacts when you don’t have things like coin series to get firm dates is hard and experts vary in their views and employ different systems.
6. Especially in the sense that ‘European’ gets used to mean ‘citizen of a country in the European Union,’ which Switzerland is not. Mostly. The EU is complicated.
7. On these differences, see F. Quesada Sanz, “Patterns of Interaction: ‘Celtic’ and ‘Iberian’ weapons in Iron Age Spain’ in Celtic Connections, vol. 2, eds. W. Gillies and D.W. Harding (2005) and in even more detail F. Quesada Sanz, El Armamento Ibérico. Estudio tipológico, geográfico, functional, social y simbólico de las armas en la Cultura ibérica (siglos VI-I a.C.) (1997). Interestingly, the Roman gladius Hispaniensis seems likely to have been a Roman adaptation of the peculiar Iberian La Tène swords, so you have the La Tène I sword making its way to Iberia, becoming distinctive, being adopted by the Romans instead of the more common (to them) La Tène II sword, thus becoming the gladius. On this, see F. Quesada Sanz, “Gladius Hispaniensis: an Archaeological View from Iberia” JRMES 8 (1997).
8. On this, see S. James, The Atlantic Celts: Ancient People or Modern Invention (1999).
9. On this, see M.J. Taylor, “Panoply and Identity During the Roman Republic” PBSR 88 (2020). On the helmet type and its evolution, see U. Schaaff, “Keltische Helme,” in Antike Helme (1988) for a rundown; P. Connolly Greece and Rome At War (1981), 121 also has a fantastic visual chart of the development of the type in the La Tène material culture zone, where you can see quite clearly where in the fourth century the Italic variants of this helmet type are breaking off from, while the La Tène helmets continue their development in other directions, later to be re-adopted by the Romans who thought it was so nice, they borrowed it twice.
10. Immediate caution: the Gundestrup Cauldron does seem to be a La Tène material culture object but the dating and origins of it are exciting. It was found in Gundestrup, Denmark, outside of the typical La Tène material culture sphere (but not far outside; we generally might guess it was a trade good) and its date is deeply uncertain, typically placed between the back half of the second century through the whole of the first century BC, but dating suggestions outside of this range have also been suggested. The imagery has been taken to suggest an origin in the Danube region (there’s a mix of Gallic and Thracian motifs, so the art historians say). So it’s a complex, difficult to work with object. The upshot though is that we can be really sure the warriors on it are using the La Tène material culture military kit; indeed one of the mounted warriors (the one with a bird on his helmet) is a dead ringer for the famous La Tène material culture aristocratic warrior burial at Ciumești, Romania (who also had a helmet with a decorative bird on it.)
11. Measurements of the one best preserved example, but these measurements fit with artistic depictions. See Gassman (2007)
12. Or at least using the worst variant of it. The Greeks looked at this shield which worked really well because it was big and thought, ‘oh, but that’s heavy. What if I make it small and light and then wear no armor and also why do the Romans and Macedonians keep beating me up and taking my lunch money?’
13. Canestrelli, op. cit. 16
14. The full entry in the Dindorf text (I don’t have the Teubners to hand) reads, “Δορύξους, δόρυ ὁλοσίδηρον. καλεῖται δὲ γαισὸς, καὶ ἔστι Λιβυκόν.” “Spear-maker, a spear of all iron. Called a gaisos, and is Libyan.” Poll. Onom. 7.156. This is not the sort of evidence one should balance such an equivalence on. As the Lewis and Short entry notes, Pollux isn’t alone in applying this word to mean a North African javelin, even though we know etymologically it is a Celtic-language word, not a North African one. This is just one of those many cases where our literary sources are not precise enough to be used in this way.
15. The archaeology of weapons in Spain is complex.
16. Full quote: De façon générale, en Gaule du Nord comme dans la plus grande partie du monde celtique l’usage du casque reste exceptionnel. Il faut attendre La Tène finale pour en trouver une trace significative. “As a general matter, in North Gaul, as in the greater part of the Celtic world, the usage of the helmet remains exceptional. It is necessary to wait to the Late La Tène in order to find a significant trace of it.” Doubtlessly bad translation mine. T. Lejars, “L’armement des Celtes en Gaule du Nord à la fin de l’époque gauloise” Revue archéologique de Picardie (1996), 96.
17. Pleiner, op. cit. 43.
18. And not cleanly so, the passage reads, lorica, quod e loris de corio crudo pectoralia faciebant; postea subcidit galli[ca] e ferro sub id vocabulum, ex anulis ferrea tunica (Varro, Ling. 5.24), which translates to, “Lorica, because they made used to make chest-armor from the thongs (lora) of rawhide; Afterwards, the Gallic one of iron was included under this term, a tunic [made] from iron rings.” He doesn’t actually call it lorica gallica as a technical term; gallica is just doing its normal work as an adjective meaning ‘Gallic.’ Varro, in fact, offers no specific categories in De Ling. Lat. for types of body armor.
19. Later medieval mail was increasingly made ‘oops all rivets’ but that is neither here nor there.
20. On this, see, uh, Me, “The Adoption and Impact of Roman Mail Armor in the Third and Second Centuries B.C.” Chiron 52 (2022). It is deeply amusing to be able to self-cite and I do not see this getting old anytime soon.
21. On this, see, uh, Me, “The Adoption and Impact of Roman Mail Armor in the Third and Second Centuries B.C.” Chiron 52 (2022).
22. On this, see, uh, Me, op. cit. and also Me, “The Material and Social Costs of Roman Warfare in the Third and Second Centuries B.C.E.” (2018) and also Wijnhoven, op. cit.
23. See D. Sim and J. Kaminski, Roman Imperial Armour (2012) for a discussion of the process, which is also detailed in Wijnhoven, op. cit. at somewhat more length.
24. The roughness of the estimate here is because a lot of these finds have meaningful question-marks on their dates. It is hard not to notice, for instance, that in the cases where Wijnhoven seems to have looked at the pieces themselves, their dates tend to move to be more recent in time.
25. On this caution, see S. James, “Warriors, war and weapons; or arms, the armed and armed violence” in The Oxford Handbook of the European Iron Age, eds. Colin Haselgrove, et. al. (2018). And yes, Caesar reports an armed warrior elite among the Aedui, but archaeologists today are cautious about extending that structure too carelessly, e.g. T. Moore, “Wealth, status and occupation groups” in in The Oxford Handbook of the European Iron Age, eds. Colin Haselgrove, et. al. (2018)
26. On this see N. Roymans, Tribal Societes in Northern Gaul: An Anthropological Perspective (1990)
27. But please note how many people are stripped out by those riders.
28. On this, see Pope, op. cit.
29. Thank heavens for words early enough in the alphabet that their TLL entries are complete.
30. Not the Scipio Nasica you are thinking of
31. This isn’t even the only time he throws shade on him either! Cf. Livy 33.10.8, where Livy describes Valerius Antias as, “prone to enormous exaggeration in all sorts of numbers.” For a full look at this problem in Livy, see P. Erdkamp, “Late-Annalistic Battle Scenes in Livy (Book 21-44)” Mnemnosyne 59.4 (2006) and P. Erdkamp, “Valerius Antias and Livy’s Casualty Reports” Studies in Latin Literature and Roman History 301 (2006).