Fall of France, 9: Manpower, Part 1

This is a posting about demographics, so I could be taking this opportunity to talk about the British Army and the late  Roman Empire while complaining about the fate of Generation X. I'm not going to, though, because politics blah.


So I put it elsewhere.


Besides, this is tricky demographics. As I keep saying, one way of assessing the catastrophe of 5/40 is that the Commonwealth did not have enough men at the front. Yet there were plenty of men who were not at the front. So it must be that there were not enough men qualified to be there. I find it implausible that countries that could draw on limitless supplies of (admittedly aging) veterans, a significantly  larger population in the case of the Canadian junior partner; and a cohort subject to conscription beginning three months before the outbreak of the war 8 months ago could field fewer fusiliers than in 03/15. So we are talking about men with a set of qualifications that could not be produced in eightish months.

That's a deduction a priori, or, to put it another way, I pulled it out of my ass. So let's look at evidence. It ought to be more fun than proctology!

History of Now

Edge of the American West is on hiatus. And I'm sad. EotAW was the first blog I followed online, beginning three years ago. And having been a nodding acquaintance of blog proprietor Eric Rauchway when I was in graduate school, I felt as though I had some deranged-stalker-type personal connection to it.

Blogs come and go. Whether they're worth the effort in the abstract or not, you can't argue with a blogger who decides that it is no longer worth it for them. It happens, and I'll leave the sic gloria transit mundi of it all to someone who actually knows Latin.

Except this: I'm a hopelessly optimistic person, but the hiatising of a great University of California blog is still a depressing sign that the demographic trends first spotted by David Foot are not magically going to release our grip on us any time soon. And by "us," I mean the professionally over-educated. (But mainly you guys, who, all my pessimism aside, have a far better chance of getting gainful academic employment than I do.) We're just doomed, something I shall now dilate upon.

Malthus, Trajan, Alexander, Adrianople: Why Not Think About Them Together, With Math?

Chris Wickham, one of the best historians working today, spent most of the last two decades working on two books to cover the period alternately called "Late Antiquity," "the Dark Ages," or "The Early Middle Ages." (Amongst others that could as easily include, "the period when barrels took over from amphorae.")

Of all the things that Wickham stresses in these books, I am singling out the terminological problem because it suggests two visual aids to me, which I will inflict on you if you like on the way to making this tenuous connection between the first of the Five Good Emperors, and the great prophet of catastrophe.

Malthus More

There was one perfect day in college when it all worked out as planned. We learned something in calculus class (I think it was the natural log function), walked out of that lecture and into chemistry class, and saw what we just learned applied, I suppose to calculate concentrations. Then we went from there to physics class, and saw it used again --at this late date, I'm not sure on just what.

Needless to say, it's not common to learn something in math class and then see it applied in history! I mean, given that some historians use pretty extensive databases, there is a place for statistics, but beyond that? The one example I can think of is the day we were introduced to some old dude name Thomas Malthus, who told us that "population expands exponentially, food supply expands arithmetically." This was something we did, indeed, learn in math class, where we plotted x2=y and x=y and watched the rising curve of the exponential function ascend, intercept the straight diagonal of the linear function, and head for infinity. Math and social studies came together to tell us of of the three ages of Man: the age of abundance, represented by the corner below the interception point, the imminent point of catastrophe, and beyond that the post-apocalyptic age of horror. Though how it could be horrible with V-8 Interceptors and mutants and zombies everywhere I do not know. But that's beside the point, because at the interception point, we'd all be living in cars in the street and eating Soylent Green, and that's terrible.

It's also debatable in many, many ways. But when you get to the core point, you start with P. J. O'Rourke's take: "Way Too Many of Them, Way Too Few of Us." There's any number of things you can take away from O'Rourke's comparison of the hellish overpopulation of Bangladesh with the spacious elegance of life in a California county with the exact same population density, including some pretty effective criticisms, and also the updating observation that the rate of population increase in Bangladesh has fallen from 2.02%/year to 1.55%/year in the last three years alone. Certainly there's no getting around the fact that he shows the whole thing to be a little undertheorised. Which is putting it mildly. Historians and mathematicians have a heck of a lot to say about this.

On Humans Not Being Bunnies: Malthus and Stuff

If anyone cared about my opinions, I would be rich and famous now. (Yes, that is an invitation to you to make me rich and famous.) So instead of inflicting them on you (for now), I offer the  links promised elsewhere.

Here's Archatlas. Have fun and come back soon! http://www.archatlas.dept.shef.ac.uk/Home.php.
Here's Ken Pomeranz dissecting Clark at length (probably limited access):  http://www.humanities.uci.edu/history/pomeranz/AHRreviewofFarewelltoAlms/ahr.pdf
If you haven't had enough Pomeranz bashing, check out Jan Luiten van Zanden, “Simply a Better Class of People? The ClarkThesis Assessed,” Agricultural History Review 57, 1 (2009), 124—29, pimping a book that's probably better. 

The mere microhistorical world of the clock on the wall is remarkably unsympathetic to my need for more time before work to extend this post. As a result, this will be all for now, although I do promise a selection of monographs for a "round table"-type seminar Real Soon Now.


PS: Who is the real Malthus? Check out  James Bonar, Malthus and His Works, conveniently reissued by F. Cass's "Library of Economic Classics" in 1966. Remember: overpopulation isn't human nature. It's something that Hindus, Catholics, High Church Anglicans, and other barbarians do. 

Fall of France, 8: Money

Last time I suggested that the real problem for the BEF on 10/05/40 was that not enough money had been spent on it. Or there would have been more of this.



(Check out the divisional cavalry moving through the outskirts of Bruxelles at the end of the clip.)

Sure, you say. It's not for nothing that we talk about Neville Chamberlain and the Guilty Men.
Let's understand the absolute, rock bottom point here. British (and Canadian) governments had been supporting enough infantry, cavalry and support services to equip more troops than were in France on that date. In particular, two armoured divisions, one entire infantry corps, and 6 armoured regiments (1 Army, 5 divisional) existed, had existed, had been paid and quartered and fed year in and year out for 70 years, and were in France at the equivalent date after mobilisation in 1915. These units existed in 1940. Some, notably the first armoured division and the infantry corps were to come over within the next two months. So the gap in mobilisation is very, very small. And the reason for that is ....money?

So I Caught An Anglo-Saxon, and He Was This Long. Me on Fleming, Britain After Rome

So I've been reading Robin Fleming's Britain After Rome: The Fall and Rise, 400 to 1070 (London: Allen Lane, 2010).

This isn't a review. If I finish a book these days, it's either because I think it's awesome on its merits, or because  I'm reviewing it, so if I think that a book is awesome, I'm not reviewing it. Unless "awesome" counts as a review. So I'm two for two on the Penguin History of Britain series. Fleming is great, and David Mattingly's An Imperial Possession: Britain in the Roman Empire is great. I am very much reminded how the Oxford History of the United States is now going one for three for me right now. (Yes to this, BTW.) Pick up your socks, Oxford! As the reading suggests, the end of the Roman Empire and the emergence of Anglo-Saxon Britain bulk (disproportionately) large in my world history project.

Why? Because of that old story about how invading Anglo-Saxons turned Britain into England. I don't believe it, and don't think that you should, either. Much more importantly, I don't think that you should believe the story that is being told with it.

Fall of France, 7, Hardware: Queen of the Desert

So, the British army had played around with tanks large and small during the 1920s. Activity hit a low ebb during the Depression, and began to recover slightly after 1933. This was all immediately politicised by the journalism of one B. H. Liddell Hart, so I am going to skip over that for now. After all, this is a hardware story!


So let's go see what a hardware shop looks like.

This is the Vulcan Foundry, founded in 1832 and builder of many, many locomotives by 1937. In a rearmament boom-led recovery, you would think that that would be a good place to be. After all, the army needs self-propelled heavy metal, right? And locomotives were an old, old business for British industry. In the clover, right?

Wrong. The locomotive business had a severe overcapacity problem, and, in general, the companies that could be expected to give way were the international exporters. Locomotives are big, and big things can't really be made by "mass production." They're basically built up of parts made in essentially the same way that replacements are made and damaged parts repaired. This has the implicit consequence that they can be profitably built in any major railway work shop. Well, steam engines can. Modern electrics and diesel-electrics are a more complicated matter, because the manufacturer comes to be more about the engine than the heavy metal.

I throw this out because the interwar was a sad era for British locomotive builders. In some visions of how an industrial strategy might work, the British government had no business propping up these kinds of firms with sweet inside deals. Admittedly, Vulcan Foundry actually survived the interwar and the end of steam, but it is hard to say whether that might not be because they won a contract to build a tank for the army in 1937. On the other hand, Vulcan won large private sector contracts right up until the beginning of the global slump, and delivered more than 100 locomotives in 1933.

Now, about this contract, Postan, Hay and Scott suggest that this reflects inadequate British industrial capacity or something, and Correlli Barnett, as expected, takes the ball and runs with it. There are no tank builders in Britain, so the government has to give the contract to inexperienced locomotive builders who stumble about trying to make ultra-modern war machines in their Dickensian plants!

Put it another way: Vickers-associated historians complain about arms contracts not going to Vickers. So what's going on at the Vickers plant? After Vickers had the specification of the A10 changed right out from  under them into a "cruiser tank," they had nothing much to offer. Accordingly, the Vickers Drawing Office rushed into action, designing an "infantry tank" that only weighed 16 tons, so that it could run on the same automotive plant as the A10. It's hard to know what to make of this. The Valentine has a reputation for being a cramped and compact vehicle. And there is a reason for this. The problem with building a tank is that the  more armour you put on a box, the heavier it is, so that the smaller the box, the lighter will be the tank, and conversely, the larger the box, the much-more-heavy-than-otherwise will be the resulting tank. And the smaller the box. So if if the Valentine was going to be heavily armoured, yet light, it was not going to have much room inside the box for much of anything, whether that be guns, crew, or engine. What's more, the smaller the box, the harder it is to give it low ground pressure and good fording depth, leading to a distinct lack of operational manoevrability.

And yet the Russians,who knew their tanks, loved the Valentine. Thinking about this, the most obvious reason I can think of for this is that with that kind of weight, and the modest speed, it must have been a pretty sweet drive. That's a virtue that doesn't get a great deal of play, but it is real enough, and the fact that Vickers was an experienced tank builder suggests that it was intentional. Leading to the suspicion that the Valentine was to be pitched on the strength of that virtue, perhaps for "colonial" operations on bad Northwest Frontier roads?

But, rush as they might, Vickers only got the proposal to the War Office on 10 February 1938,and it was not approved for another year. In that time, Vickers completed deliveries of its two-person, heavily armoured "Matilda I," which was basically intended to be a mobile, offensive pill box, and became obsolete the moment that enemy armoured counterattack became possible. Meanwhile, the Mechanisation Board, after worrying the matter over for some time, approved a design for a "heavy infantry tank," and, yes, put it out to contract with Vulcan Foundry. The first two resulting "Matilda IIs" appeared in service in September, 1939. (Wikipedia says that the first Matilda II was built in 1937, but unless this was a mockup made in sheet steel, this is a little hard to see.) Later contracts were placed at 5 other locomotive builders, and just under 3000 Matilda IIs were built through 1943, at which point it was replaced in service by the Churchill tank, and in production by 450 "Austerity" locomotives needed to keep British industry working and for Europe after the invasion.

The extent of the production contracts suggest the raw productive capacity of the Vulcan Foundry. And there was certainly nothing else to distract them. The economic slump of 1938 refuted the idea that the armaments boom of 1936--7 was going to drive up inflation and cut exports by taking up engineering industry capacity. On the contrary, 1938 was a year of  rising unemployment, stagnant wages, falling steel consumption and outright misery in the north.  That said, armament procurement was not trivial. Engineering industry armaments production had risen from 43.8 million in 1936—7 to 76.7m in 1937—8 and 109.5 in 1938—9, compared with a total sector-wided value output of 473.6 million output in 1935. Armaments took up much of what would have otherwise been a severe drop in engineering sector productive capacity, and the editorial board of Engineering turned to worry about post-rearmament excess capacity. Which is all very well, but with war coming, perhaps not the most urgent of issues.

So, returning to the question of picking the wrong companies. Would everything have been fine if the British government had just given the Matilda II contract to Vickers, instead of sending a bag of money off to Third World Liverpool in hopes of staving off cannibalism? Well, no. The Valentine was approved for production in April 1939, and, as was often done in the last years before the war, was sent directly to production. An advanced batch was available for testing in May 1949, and the first regular production models appeared in the summer. That's pretty impressive compared with, say, the T-34, which was field-tested in a preliminary test vehicle in 1939, had a prototype in January 1940, and only appeared in a production type in September 1940, or, for that matter, the history of the Matilda II, which had its design finalised perhaps at the  "end of 1937" and was still only available in limited numbers in May 1940.


The thing is, though, that British industry delivered 67 infantry tanks prewar, 63 in the first quarter of the war, 46 in the second, and 121 in the third. So given that only 140 Matilda Is were ordered, divided into 60 vehicles in the first (enough for one regiment. Hmm....) and, presumably, another 80 in the next, simple math tells us that virtually the first Vulcan order for only 65 Matilda IIs,and some of the subsequent increase to 140 vehicles had been completed by June of 1940. Production and delivery to the front are distinct events, so the 21 that fought in France might have been the only ones that were really ready, but the country was able to spare two regiments worth by August, sending them off to the Middle East to become the  "Queen of the Desert."


True, people carp. And well they should. The nickname is awfully self-congratulatory, and the Matilda had its problems even before 1943, when it was functionally obsolete. There were odd design choices. It had two engines, for example, developed from an AEC bus power plant. Why? Could "Britain" not build sufficiently powerful diesels? On the contrary, British maritime diesel engine makers would have been perfectly happy to build one, as evidenced by Thorneycroft and Paxman sniffing around the business. The thing was that the Matilda was a 27 ton, heavily armoured monster. That the monsters would get even bigger was not obvious. The divided power plant was considered necessary to make it work at all, by giving separate power to each track. Given that, 190hp made it go quite adequately fast, and the Paxman 600hp diesel called, on first inspection, for some monstrous land-hulk. That Dr. Merritt had adequately solved the transmission problem was not obviously apparent to Walter Wilson's backers even in 1946, with the successful experience of the Cromwell and Churchill behind them, so it is no wonder that they were not thinking about a 30 ton tank tearing about the countryside at high speed. (To their credit, the Russians were.)


So I'm not seeing an issue around the engine. People also complain that the castings on the Matilda were too big and difficult to mass produce. Again, given that entire turrets were cast, I'm not sure I see the argument. If there is one thing that you go to locomotive builders for, it is experience in making big pieces of steel capable of taking heavy stresses! Anyway, "difficult" is not the same as "impossible," as a production total of 3000 machines tends to suggest.

What I am seeing an issue around is this: an order for 65 tanks? In 1938? And why weren't they all in France on 10/05/40? Frankly, talking about how Vulcan Foundry might not have been the right contractor seems to me to duck these more important issues.

Electric Santayana III: Let's Not Think About Rob Schneider Any More, If That's Not too Meta

I hadn't intended to blog about this again any time soon. But then there's this.


"Britain was slow to move from the old industries of the first Industrial Revolution into modern sectors like electrical engineering, which impeded the adoption of mass-production methods. It also failed to adopt precision machinery that depended on electricity, which prevented it from producing machined components for use in assembling typewriters, cash registers, and motor vehicles. The same story can be told about other new industries like synthetic chemicals, dyestuffs, and telephony, in all of which Britain failed to establish a foothold." 


(Barry Eichengreen of the University of California, Berkeley, from here. Thanks to Brad Delong for bringing this to my attention. Hey! That's me boiling dry down in the comments!) It's hard to know what to make of this. The criticisms themselves float in a timeless space that may refer to the 1880s or the 1930s. It's only when Eichengreen goes on to discuss their causes that we nail down the chronology:

"A final explanation for Britain’s failure to keep up makes economic policy the culprit. Britain failed to put in place an effective competition policy. In response to the collapse of demand in 1929, it erected high tariff walls. Sheltered from foreign competition, industry grew fat and lazy . . . . Herein lies the most convincing explanation for British decline."

Professor Eichengreen isn't setting himself up as a historian of British industry here. He has a Very Important Message for modern Americans, and a cautionary tale from somewhere far away and long ago serves his rhetoric well. Who really cares whether the story is true or not? Well, if Professor Eichengreen substituted "Atlantis" for Britain in the example above, he wouldn't get much traction! It's the appearance of historical fact here that gives his warning its persuasive power.  And he's made judicious use of weasel words here so that historical criticism is hard. Britain was "slow" to move into modern industries like electrical engineering. So Britain could have had an electrical engineering industry. It's just later than soon. When I start talking about the 1930s, the good Professor can wave at the map of the past that exists only in his mind, he can say, "that's not what I meant!"

He does, however, say that British industry got fat and lazy in the 1930s, after the global tariff wars of 1929. He doesn't say that electrical engineering and "telephony," did not exist by this time, but I assume that I'm safe in thinking that they must have been particularly lazy. Leaving aside the question of why tariffs and export subsidies made American and German industry fit and active, I can put British industry on the treadmill, haul out some calipers, and maybe learn a thing or two after the jump.

November 11th

http://www.youtube.com/watch?v=6TEhzpcCRYQ

Fall of France, 6: Hardware

When we talk about things like tanks in World War II, there is a very strong temptation to talk about hardware. Because hardware is cool. You can spend a lifetime obsessing on the details of World War II hardware, turn it into a career, and become an important authority with valuable things to say about it. Or you can become a professional obscurantist and lead generations of naive readers drastically astray. (Speaking of people who take Corelli Barnett far too seriously.)

It's telling the difference between the two that's the problem. Many military historians actually eschew too-close engagement with military technology, after one too many encounters with "rivet counters." I don't think that's a very helpful attitude, myself. If you're going to make an analytic point stand on the differing ways that the Germans and British used their heavy AA guns, then you need to know that the difference between the 3.7" (94mm) and 88mm AA guns. At the same time, when you read someone arguing that you need to understand a particular partial differential equation in order to "get" British tank transmissions in World War II, then you have an excuse. I can only think of one historian off the top of my head who has attempted to understand (a set of)   PDEs in a historical way. It's a fascinating exercise, but it's far from clear who the audience for such an undertaking is supposed to be.

So when we talk about the history of the British tank, we're put in a schizophrenic place. How can you not talk about hardware? When the British official history tackled this problem, the editors made the very creditable decision to put the same medieval economic historian who wrote the history of British war production in charge of Design and Development of Weapons. Throw in two industrial historians (D. Hay and J. D. Scott), and you've got a team that can do justice to the questions that historians bring to the project as well as engineers.

One hopes. In reality, you've got bias built in. Just picking a medieval economic historian in 1950s Britain pretty much guaranteed that you'd get a quasi-Marxist ready and able to beat back the then-active school of technical journalists arguing that "British free enterprise won World War II, so what the heck's all this nationalisation about?") As for Hay and Scott, we've oh-so-cleverly picked the official historians of the Vickers group of companies. (IIRC, Google failing me right now). Again, there is wisdom to this, in that the Vickers group have designed and built a disproportionate number of British weapons over the years. The problem is that along the way, Vickers has fought off any number of outsiders intent on invading their turf.

I'm not going to argue against the company's expertise and insight into tank-building problems, Between 1927 and 36, between 22,500 and 93,750 pounds were spent annually on new tank development, and Vickers was the only firm involved. In the process, the government-funded research complex that we customarily denote as Woolwich Arsenal developed a process for welding hardened steel plate and a nickel-chromium-molybdenum steel suitable for welding. This was to armour Vickers tanks, to be sure, but the spin-off implications are not small, and the crumbs were well-positioned to fall Vickers' way. At the same time, while Dr. Merritt might be working on a tank transmission, he was not a Crown employee, but rather David Brown's top boffin. The article linked to above notwithstanding, I do not believe that we have a full understanding of what was going on here. As far as I can see, the Merritt-Brown transmission could not work without carbonitrided gears, and the potential of this technology ran far beyond tanks to everything from mine elevators to ship engines to the heavy-duty machine tools used for making armour. (I like these recursive formulations a lot, you can tell.)

So did Vickers get an inside look at these technologies and an opportunity to exploit them? I think so, although I've no intention of trying to prove it now. Then, in October 1936, the War Office published three specifications, for a light tank, cruiser, medium, and infantry tank, stipulating 14mm armour on light tanks, 30mm on cruisers and 60mm for infantry. The light tank was also required to have a gun and regenerative steering, and this put Vickers' existing 5.25t, 32mph Mk VI out of the running, even though, since the replacement was not needed urgently, due to a conceptual shift to a "scout car," it in practice became the last major British light tank of the war.

The Cruiser tank requirement came at the same time as the Wavell Mission to see the Red Army manoeuvres and was probably inspired by it. In spite of that, Vickers' A9 prototype satisfied the requirement. So, unbelievably, did its A10, supposedly an infantry tank, but ruled out for that role by the new armour requirements. In short, every tank that Vickers was then producing somehow met the army's needs. But what of the new tanks? Well, the Nuffield Motor Group was commissioned to build the new cruiser, in spite of having to buy existing outside engine and tank designs --both American in origin. The London Midland and Southern Railway shop got involved with a proposal for a tank with a British-designed diesel engine, later dropped in favour of a development of the  flat Meadows engine that had gone into all the old Vickers tanks on the grounds that tanks should be low to the ground. Which was an official and well-found War Office position that you'd think would rule out orders for the American Liberty V-12 that went into the Nuffield tank!

 Meanwhile Vickers designed the Matilda I as a fast entry into the infantry tank sweepstakes, of which crazy little design perhaps more anon. But the big contract went to a locomotive shop just outside Liverpool called the  Vulcan Foundry, which ended up producing one of the real bruisers of 1940, Matilda II. In theory, though, there should have been one "tank circus" per corps in France in 1940, which would have amounted to more than 400 of these monsters. Instead, there were fewer than 30. So the question is, why? 

Well, it's not much of a question, because Correlli Barnett already has an answer: British industry was in decline! Vulcan would be swept off into the ashcan of history in the later '50s, almost the last of the prime mover builders of the industrial North, thereby freeing up Liverpool lads to go into, I don't know, rock and roll or something. Computers? Computers would be good.

Only, I think that's the wrong answer. Next up: I interrupt this interminable lecture series with an even more boring film!

On Decline

So the United States is the greatest empire that the world has ever seen. But are its best days behind it? Spain and Britain are empires that have risen and declined, and in some ways America is like a combo Spain-and-Britain deal. And so, asks oh-so-smart Yale Professor Paul Kennedy, isn't America soon going to decline and be overtaken by Japan?

Japan? Oh, sorry. Got 1987 confused with 2010 again. Seems Paul Kennedy got himself a job at Yale on the strength of  a pretty solid early writing career, notably dilating on the "rise and fall of British naval mastery" (thesis: "you kids stop mucking around and get out of that there Mediterranean afore you get an infection!") . Then he hit the sweetspot with a book about the imminence of American decline. That's what we expect of a man using up one of the precious few chairs at Yale. Unfortunately, he then returned to the library and produced the monstrous Preparing for the Twenty-First Century. My copy of same was free, because someone moving  out of a basement apartment up the street left his copy lying on the grass behind him. Which seems like a pretty generous review to me --my own first diagnosis was incipient cerebral shutdown, but Professor Kennedy has gone on to write seriously and appropriately in the last decade, and Preparing is no doubt just the product of a book tour, which would drive anyone batty.

The interesting point here is that Kennedy is a Briton, born in Northumberland, saddest and most autumnal of Anglo-Saxon kingdoms (where now the glory of Bamburgh, the holiness of Lindisfarne? In whom survives the blood of Ida?) and the progression from writing about the Royal Navy to the decline of empires is naturally British. There being more money in Yale than at the University of East Anglia in the 1980s, it is perhaps not surprising that he exported his act across the Atlantic, at just the right time for Americans to receive it with open arms. With American declinism in the news right now, all I'm really saying is that these things are cyclic. Like empires, y'know? First you have your principal declinists, then a golden age of philosophical declinism followed by an influx of severe military officers, then a crisis in the literature, followed by the emergence of the dominant names.

And somewhere in there, at least one declinist names his horse a consul. In his 1986 book, Collapse of British Power: The Audit of War: The Illusion and Reality of Britain as a Great Power, a sometime journalist turned amateur historian named Correlli Barnett made an impressively detailed argument that pretty much everything that happened in 1939--1945 proved that Britain was in "decline." As far back as the immediate wake of the industrial revolution, Britain had turned its back on engineering, business, science, realistic foreign policy, military staff planning, and practically everything that is rational. Then it blithely skipped through the 20th Century until it ran into the ruthless, warlike rationalism of Prussian militarism filtered through the state-planned dynamism of National Socialism and....

Oops. Won the war. But the point is, it shouldn't have. Or it should have won the war even more. Or it was all down to American assistance. Or, mostly, all of the above. It is clause 2 that fascinated me as a boy, and in some ways fascinates me still because it is so apparently, obviously true. Imagine replacing the arsenal of British power as it existed in 1939 with the one that existed in 1959. Don't even throw nuclear bombs in there. With Victors and Centurions and Hunters, the British armed forces would have kicked Nazi butt! And wasn't Britain 20 years ahead of the rest of the world industrially in 1820?* So, really, the dispiriting events of 1939--45, where Britain had to play second fiddle to those darn Americans was down to someone, or many someones, who screwed up in 1820--1939.

Audit of War is only loosely about that, however. It faces the rather difficult problem that what with Spitfires and radar and all, World War II certainly looks like a war fought by an industrially superior power, and the book is Barnett's extended demonstration that it really isn't so. So that is why Barnett talks about monocoque fuselages, Ebbw Vale, four wheel drive, and vacuum tubes at great length. It leaves the historian a little ill at ease. Barnett seems to know a lot about these subjects, and it is hard to know just where to start with unpicking the tapestry he weaves. In my graduate student days, discussions of Audit (which I really, really wanted to discuss) tended to be turned away with waves in the direction of Martin Wiener, whose attempt to discover the identity of Barnett's villains seemed at least open to critique.

Now, Wiener's attempt at cultural history is just plain weak. (One word, Dr. Wiener: prosopography.) And Audit of War is a pretty hard book to attack. I've spent much of the last two decades preparing to write a substantive reply, and I still get pretty important technico-industrial details wrong in public. I mean, there's aluminum and electrical cable and steelmaking and organic chemistry and aerodynamics and mine engineering I really have no idea how Barnett achieved the level of mastery required to make such confident claims ...No. Wait. I do. And you don't need Audit to see it, too. Before Audit, Barnett rehearsed his argument with a brief section of his book The Swordbearers that "proved" that the British Grand Fleet at Jutland was technically inferior to the German High Seas Fleet, and this happens to be a subject on which you can find an easily predigested summary of the technical issues. Battleships soaked up a lot of government money in the years before 1914, and there was a lot of journalism on this subject. Shipyards that got contracts (and the officers who directed the contracts to them) leaked one line to journalists ("our battleships are awesome"), and shipyards that didn't get the contracts (and opposition politicians) fed another line to different journalists ("our battleships suck.")

In the final analysis, it is pretty hard to sort out these claims to the last detail, but you can go to a high level review of the arguments (I think you'll find a good one, Brassey's Naval Annuals, in with all the other detail here.)  A point by point review of the things that Barnett apparently thinks are true about the Grand Fleet in 1916 demonstrates that he just swallows every criticism, no matter how inaccurate, and rejects every defence, even when such things are easily tested in the more recent technical literature that Audit of War implies that he has mastered. This is not research. It is indictment.

Which we already get from Wiener's critique, anyway. And, actually, even further back, from Lord Snow's "two cultures" argument. "The politicians just don't respect us engineers and scientists. They're all flouncing about quoting Latin, and they just pat us and say, 'there's a good boy,' when we say we should build something totally awesome!") Wiener and Barnett are pretty clear on what's to blame for all of this: labour unions, public schools, Nonconformism, Gladstone.... I'm surprised that they don't mention Home Rule. Or maybe they do. The important point is that Britain had a chance to embrace Prussian-style technocracy, and went all wet.

And this brings me back to my last few posts, where I've talked about Brabazon and Lord Weir and the National Grid scheme and even Messrs. Balfour and Atlee. Because it looks to me like Britain between the wars was bloody well as technocratic as all get out. This leads me to the unsurprising conclusion that we worry most about what we care most about. In short, it is likely to be a technocracy that worries that science is not getting enough respect from policy makers, and a world-bestriding empire that worries that it is about to go the way of Nineveh and Tyre.

Now if you'll excuse me, I think I'll go check out one of my own links and read some more about fire control at Jutland. Cool stuff!


*Well, no. But that's another story.

Electric Santayana II: Douce Gigawatt

So let's talk about a time when a government facing a "business slump" launched a  major infrastructural project, and it turned out well.

Our idea of the Nineteenth Century is the Gaslit Age, so it is amazing in retrospect that the gaslight was only fully tamed for indoor use with the development of mantles in the 1880s. By the time that Welsbach's first white light mantle was in production, Thomas Swan had been installing his lightbulbs around the U.K. for more than a decade.
By 1890, electrical lighting was making specialised inroads against gas. And by specialised, Hannah Leslie, the great historian of the industry means "indoor lighting." The typical British electrical power station had 8 machines of about 0.1 MW (134hp), all reciprocating steam engines running local DC supply grids. That's a whole lot of obsolete technology all wrapped up together, but these little plants burning slack coal on the canal side were still enough to light as many as 26,000 Swan lamps (ie. "lightbulbs") each. Let's put our expectations in the ol' time machine and send them back to 1890. That's what a big deal in electricity looks like in 1890. For those of our expectations with some engineering background, I'll note that this was at 10lbs/coal per kWh. 


with AC distribution, turbines instead of pistons, and better cables, there was almost unlimited room for expansion. Overall sales rose from 40 Gwh in 1895 to 1400 in 1910, the year that industrial power demand overtook illumination. In the course of WWI, the existing generating plant was put to new uses, run at overload and without maintenance, and been a boon to war production generally. What didn't happen was significant investment in new generating plant, something that greatly concerned the business-background technocrats that Liberal Prime Minister David Lloyd George brought into government to wage the war. Which is why Eric Geddes proposed a technocratic, top-down solution to the problems of the electrical supply industry problems in 1919, which got run out of Parliament.


Why? Because "socialism" versus "free enterprise." And also because David Lloyd George was a bit of a dick. (Geddes, by all accounts, was a lot of a dick, but not in a bad way when you need to get things done, whereas Lloyd George's dickishness tended to get in the way of his schemes in the long run.) Anyway, this set up the new Labour Government to take a new look at the problem in 1925. "Socialism? We're all over that!" Interestingly enough, the young Labour MP most involved actually had abackground in the electrical supply business --one Clement Atlee. More big names: as already noted, the guy they put in charge of thew new look investigation was another of Lloyd George's "men of push and go," but of an even higher calibre than Geddes, William Weir. 


And here the story out to end, you'd think, because the first Labour Ministry was shortlived, and Weir reported to a huge new Conservative majority in March of 1926. Weir himself was an industrialist. The MP that pushed himself forward as the Conservative spokesmen on matters electrical-supply related, George Balfour, was building his own private grid in the north of Scotland to power an aluminum plant, and was convinced that socialistic central planning would just hobble the industry. It's not as though the ultimate outcome of the Weir scheme, a national supply grid, was visualised at the outset. The Weir idea was, in fact, just to link up 140 main "flagship" generating plants with a system of high-voltage main lines to allow interregional, coordinated, electricity markets for a standardised product (50Hz AC). The fact that all the regions of the UK (except northern Scotland) were to be integrated meant, in effect, that this would be the infrastructure of a national grid, but a lot of engineering problems would have to be solved first.  A government that wouldn't even let the Royal Airship Factory build a zeppelin without commissioning a private-sector rival was not going to go for something so Bolshevik ...was it?


And then Clement Atlee received a visit from a similarly junior MP opposite. And not just any MP, because while the future Baron Brabazon was a man of the technocratic future, no-one could call him wet, as his favoured political party of the 1930s would suggest. And yet here he was going behind Balfour's back to push the Weir scheme through Parliament.


The substantive consequences are clear enough. The scheme was completed, on time and on budget, by the end of 1930. A perhaps over-optimistic estimate was that building the grid created 120,000 jobs directly and indirectly, more than all the other government make-work schemes of the 1920s put together. The National Grid came into operation late in 1938. Meanwhile, on the generation side, the development of those 140 flagship generating plants galvanised a great deal of the 750 million pound investment into the British electrical engineering industry in the interwar decades that made it the leading sector of Britain's leading industry [Hannah, 148]. There was the development of the larger part of Scotland's hydroelectric potential. 


And the failure to develop adequate new generating capacity to cope with runaway demand during rearmament and World War II underlined the timeliness of it all. Why not build these things when aggregate demand is slack? In retrospect, the biggest mistake was the failure to invest in several new generating schemes, including coal-burning plants in Ebbw Vale, a tidal project in the Severn estuary, and more dams in Scotland. More generating plant, built when private sector investment was at a low ebb, would have paid off in spades during World War II. (So would have the new port on the Severn that was part of the tidal power scheme.) The timing was not completely perfect. In retrospect, building ought to have peaked during the 1931 crisis instead of the year before, but that would have taken a neat bit of prophesy back in the spring of 1926.


And speaking of prophesy, I have not answered the reason "why." What was the meaning of this amazing outbreak of bipartisanship, with Conservatives implementing electrical supply "socialism?" The instrumental answer is that by the summer of 1926, the Weir scheme was hugely popular. The Daily Mail, the 1920s' version of Fox News, supported it! What's going on here? It would seem that the key motivator was --fears of British decline. The year was replete with warnings, complete with bogus statistics, that various countries ranging from America to Germany to Shanghai and Belgium were all more electrified than Britain, which was clearly declining in all directions everywhere. "On dune and headland sink the pyres," and all that. 


I'm not usually much of a friend to declinism. I find it a frightening example of the way that metaphor becomes argument. But this time it seems to have been a powerful motivator for a valuable public initiative in infrastructural development. Or maybe we can put it all down to Tutmania. It turns that electrical pylons are named after Egyptian temple architectural elements, because that was the cool thing that the kids were down with in the late '20s. (Hannah, 118.)


I did not know that.


Edit: who's got time to edit blog posts? Thing is, though, I kind of buried the lede here.

Fall of France, 5: The Circus

A few years ago, I decided to read the interwar issues of Army Quarterly looking for smart British army officers saying military-scientific things. I found some.

That's not my point here, though. I've been blithering about military-scientific work and the "lesson learned" variety of primarily official military history for years. What I learned is that, astonishingly enough, the soldiers of the interwar period thought about the First World War. A lot. And they analysed it, and they tried to understand what had gone right, and what had gone wrong. And that they were basically right on the mark. To put it another way, reading Army Quarterly brought back the experience of reading John Terraine and realising that forever after I would need to look at history as a place where people aren't stupid. That, in general, people aren't stupid. Did I have the vague impression that a new history of the First World War was needed, or am I just trying to take credit for someone else's good idea? Probably the last.

In particular, I have a vivid idea of how one went about winning a battle in 1918. More vivid, perhaps, for a morning person because it is an image of an early morning fog in spring, with a bright dawn above, the roads muffled and populated by guns, their steel tyres wrapped in cloth, and silent men move forward, pontoons shouldered. Beautiful Aurora, goddess of the dawn and new possibilities ---and steel sleeting death. They go together, right? Shut up.

If all goes well, the guns will open up on the enemy in complete surprise. Machine guns and mortars will keep the enemy on the front line at the bottom of their trenches, while field-calibre guns will lay down barrages to block reinforcements. Bigger guns will conduct suppressive fire against enemy batteries, while the very longest-ranged guns will interdict crossroads far to the rear. Under cover of fog and fire, the tanks will lurch into no-man's land to occupy the critical terrain, while bullets, shrapnel and fragments deny it to the enemy. As that fire lifts, the infantry go in to assault the enemy's fortifications, while the barrage "creeps" ahead of them, progressively expanding the ground gained until it reaches the limit of the "bite" of ground that this offensive aims to take from the enemy. In a perfect world only imagined by theorists, the "bite" reaches the limits of the enemy's defence in depth, and a new wave of tanks, armoured personnel carriers, horse cavalry and who-knows-what-else is unleashed.

That last part never happened in WWI. It would be pretty foolish for the enemy to deliberately build its whole defensive zone within the range brackets of your field artillery. You have to stretch them pretty thin for that to work. Even then, nature inconveniently interposes a river line roughly every 5 kilometers in Europe. When your way of war depends on truly massive guns, then the pace of your advance is dictated by your bridging equipment. Either get better equipment, or build some fantasy weapon that can wade across rivers.

And while silence and fog are the means of stealth, effective aerial and other reconnaissance is a double-edged sword. If you know the enemy's disposition, how can you guarantee that they don't know yours? Because if they do know yours, they will reinforce their defences with their own guns, tanks and infantry, and you will be sending your troops into a massacre. Maybe the enemy air force attacks the roads, and your troops don't even arrive at the front at all. You have to fool the enemy.

Now, there are stunts and tricks that help you do that, but when you get right down to it, the key is not to assume that the enemy is stupid, but to give them the least possible opportunity. You bring up your guns and tanks and men at the last minute. In a perfect world, your offensive assets are relaxing far in the rear one day, giving the enemy a colossal crack the next.

Offensive assets? Clearly, that includes your average rifleman. Will he march up to the front? Take a train? A truck? When does he start? When does he get there? With hundreds of thousands of men involved, it's no wonder that commanders-in-chief delegate these things to their division commanders. But the crack is pretty much administered by these guys, and, when you get right down to it, by these. To achieve strategic surprise, you need to get all of this stuff to the front at the same time, so they need something like a "division," too. In the cozy world of prewar, there were specialists to take care of such things, but post-1914, it's a different era.

And that's where circusses come in. It's not a very military word, but it stands for the best known warrior of the World War, Major Manfred v. Richthofen, who used to take Jagdgeschwader 1 from one sector of the front to another as needed. When the brightly painted planes of Richthoffen's "flying circus" showed up over a sector of trenches, you knew that the eyes of the German High Command were on it. And reaching for a label for the similar formations of guns and tanks that he imagined would be held in the hands of a future commander, the promising young Canadian officer E. L. M. Burns labelled them "circusses" in 1938.

It's a coinage that has not stuck. Historical amnesia portrays it as a war lesson learned by the artillery branch and an armoured division. The value of Burns' "circus" concept is that it captures the fact that this systematic grouping of military resources has its roots in postwar analysis of the battles of 1918, and was carried over into the British army's planning for war in 1940.

In particular, Lord Gort did have a "circus" of tanks: 1st Army Tank Brigade. And this raises, or ought to raise a question: why did he have a "circus," an innovation of the postwar era, and not the armoured division he was supposed to have, needed to have, had been allocated for his use in 1874?

Electric Santayana

As in, blah learn from history blah.

So, a recent thread at Edge of the American West went in the direction of current proposals for infrastructure investment as a way of getting the American economy going again.

Now, as I much as I have personal reason to seethe about the way that our American connection is holding the Canadian economy back, I'm not sure that infrastructure spending will serve in a timely fashion.

That's not to say that it is a bad idea, and the example that leaped to mind -and no sooner mind than that awful empty comment box-- was the creation of the British National Grid, which was initiated during the 1926 slump and completed in 1938, thereby getting two economic downturns for the price of one.

So, if I have a factilicious take on the creation of the Electrical Grid that might inspire modern-day infrastructural-spending advocates, this would be the place for it, right?

Well, no. First of all, I need to have some pithy postings. Second, all I have right now is an inchoate idea about what will go into the "Industry" section of "I Will Run Away to the Air": Britain, the Air War, and the Industrial History of Strategy, 1919-1945, in the event that I ever get to write it.

Now, I do promise to write such a posting, but it will require a library day, and I'm not keen on making that trip today, on account of having to do laundry and then go to bed early, since it is inventory day at work tomorrow and I will be in awfully early, and anyway there might be someone working tomorrow --not to say more on that score at risk of jinxing myself.

So here's Wikipedia, instead. And further evidence that Lord Weir was a time traveller from a nightmarish, Nazi zombie ruled future, come back to change history.

Fall of France, 4: The Armoured Division, I

So let's imagine an infantry division on the march. It's at point A, and has been told off to march to Point B. Intelligence is that the enemy is at Point C, and GHQ has decided that the enemy can't have Point B. They may or may not be planning to march on Point B, but the fact remains that they might. If they reach Point B first, they will most probably advance on Point A. These are the sorts of things that GHQs think about.

So imagine it marching down a road. We can basically assume a typical country road with two lanes and two narrow shoulders between the ditches, because whatever the particulars, the narrowest stretch of road sets the pace. That's enough room for 4 men to march abreast, with one lane left free for other uses. The division will be spread out to allow extra space between each section, platoon, company, battalion and brigade, and, in total, will need about 1000 meters of road per thousand men. In total, then, a 12,000 man division will stretch out over a half-day's combat march. If the enemy is run into along the road, the division will come into battle in a single day.

That's why there are divisions in the first place. So that the army can be divided up by the number of available roads.

Okay, but imagine that the division is marching down, say, from the Ardennes plateau of Belgium towards the river Meuse. The high hills on the French side of the river dominate the scene, and the road follows a long and gentle slope. What's wrong with this picture? Well, in the Bastille Day parade of 1899, the French public was treated to a display of the latest military technology of the army of the Republic: the Soixante-Quinze. This new gun had a hydro-pneumatic recuperator that allowed it to fire without jumping out of position, a rapid-loading mechanism that allowed it to be fired as quickly as it recuperated, up to 20 rounds per minute with a fresh crew, and an automatic calculating sight, so that gun-aimers only needed to keep a telescopic sight laid on the target with a pair of adjusting dials. The new cordite propellant gave a muzzle velocity of 500 m/s, and the 16lb shrapnel shells, made thin-walled of the new high quality steels being made in recuperative, regenerative Siemens-style furnaces and machined by the new generation of powered machine tools, could hold 300 balls. (In the old days, a "shrapnel shell" contained a small amount of high explosive and a large number of musket-style balls that were "fired" out the front of the shell by a timed fuze.)

The result, provided that the initial aim was right, was that a single "75" on the French hills opposite, more than 5 kilometers away, could, without warning, open up on the advancing division with the equivalent firepower of 20 machine guns!

One often hears that the generals of 1914 underestimated the firepower of the machine gun. Bearing in mind that the 75 had been invented since the machine gun, these kinds of numbers tell us why. It didn't hold a candle to the new generation of artillery.

Obviously the solution to a problem like this was not to hope that it didn't happen, but rather to deploy the division in skirmish order at the rim of the valley so as not to give the enemy artillery this kind of target. Now, I've picked this specific example for its relevance to the Fall of France, but we should not get hung up on technology. From longbowmen on a hill facing the road exiting the forest of Crecy down to the a divisional artillery network controlled by a single radio-equipped forward observer in 1945, the tactical problem hasn't changed, only the scale. But scale is important here. If a bunch of longbowmen control 300 yards of road debouching from a forest, you either march your army up that road in a column and try to run the longbowmen over, or you stop and deploy into a long and shallow line that minimises your exposure while threatening the English flanks. If you figure that it will take an hour to shake out your lines in the middle of the woods, it will take an hour to get back into column. So the English can wait until you're deployed, and scamper. They've gained an hour in their retreat. If you need to disperse into skirmish formation and advance forward painfully slowly every time there might be a single enemy gun within 6000 (or, in 1940, 15,000) yards of your road, then all the enemy has to do is send a battery of guns forward, and it will have all the time in the world to get to Point B before you.

So there has to be another component to your solution. You have to make sure that the enemy isn't in that position. Military thinkers talk about "screens," and "security," and "reconnaissance," but it comes down to this. Somebody has to go see if the enemy is out there, and maybe shoot them if they are. And this means something crucial. You have to have a part of your army that is much faster than marching infantry, if the infantry are to march as fast as they can.

It's quite possible that we can name the man who brought this to the world's attention in 1595BC: Mursili I led a Hittite army all the way down to Babylon and sacked the city, ending the Amorite dynasty of Hammurabi and bringing back the temple idol of Marduk and the scribes who inadvertently started Indo-European by adapting the old Sumerian script to the mutable, pre-literate tongues of Anatolia. It's such an extraordinary and unprecedented achievement that it is hard to believe that it was not the war chariot's definitive baptism of fire. (A chariot "invention" event has been proposed for the Eurasian steppe in the late 2000s BC, but without rejecting the location of the event, Russian scholar Elena Kuzmina raises serious doubts about the chronology here.)

It's hard to emphasise more the world-historical importance of the horse, and thus our need to understand its agronomy. Attempts to sidestep the complex details by waves in the direction of compound bows and stirrups as all-explanatory devices not only do a disservice to world history, they help us understand the issues at stake in the Cavalry Division Debate, as I herewith christen it.

The first problem to appreciate is the matter of scale. How many horsemen would a divisional commander want to have under his command before he felt safe in making a pell-mell advance on Point B? Think about 3000 fresh-faced boys looking up in the last second of their lives at the whizzing shells about to bloom, and you realise that it is a trick question. The answer is, "all of them." That said, men on horses spread out more than men on foot. It's safe to say that a 30 year-old infantry officer can command 1--200 men, and a 35 year-old 600--1000, and the "company" and "battalion" organisations cluster around these strengths. Their cavalry equivalent are the squadron of 1--150 men, and the regiment (or "brigade" in the Nineteenth Century "Light Brigade" sense) is 4--600 sabres, as the old-timey military historians liked dto say to distinguish mounted men from foot soldiers ("rifles.") Above that, and you get into the realm of a cavalry division.

So, should a division commander settle for a squadron, which is to say, 150 sabres to screen the front and flanks of a column 15 kilometers long? No. Nor is he going to get a division, so he settles for a regiment. Or, hmm. Does he? The thing is, countries like Britain, France and Germany have huge populations of more than 40 millions and, as 1899 turns into 1900, mass conscription or some thought of imposing it. There's no reason for them to field fewer than 50 divisions, and the ambitious can aim at 100. 100 divisions x 500 horses/division=50,000 horses. Which does not seem so much until you remember that all the guns and wagons have to be drawn by horse as well. In fact, the Soixante-Quinze has set a new bar by requiring 6 horses. (Plus more additional horses to haul all that ammunition.)

And horses do not grow on trees. Great Britain, an unusually rich country with unusually rich farmers and a great deal of excellent stud land, has about 2 million horses. France, by contrast, a country with very little good stud land, has some trouble reaching a ratio of 1 horse/20 inhabitants. Nor can send your breeding stock to war, nor remove too many horses from agriculture. There is, in short, not likely to be enough horses to go round. (Except for Russia and India, which are plugged into the Inner Eurasian horse export market.)

So what are we going to do? For starters, infantry division commanders are not going to be allowed to have more than a squadron of horses or so. Many divisions will have their horses taken away entirely, especially if they aren't marching in the first wave. To provide additional security for the first wave, those horses will be concentrated in cavalry divisions and placed under army commanders. One can even imagine bolder, more sweeping notions. If Attila could ride across Europe and into France with an entire cavalry army, why could a German cavalry army not do the same, when one day in the future the inevitable war broke out between the two countries. (It did finally happen in the late summer of 1914 after the Schlieffen Plan had well and truly failed. And it worked out only a little better for the Germans than it did for Attila.)

The next thing you can try is find an alternative to the horses. Late Nineteenth Century technology was overflowing with possibilities. It is no coincidence that Count Zeppelin was a former cavalry officer, and across the border in France, Clement Adler got money for his experimental aeroplane from the French cavalry authorities. By the time 1914 rolled around, such divisional cavalry squadrons and regiments as were still about were stretched with companies of bicyclists, and armoured cars showed up as soon as war was declared, and promptly demonstrated the obvious problem that wheeled vehicles can't go where cavalry needs to go.

Nor were these visions the only ones. What if modern science could make foot soldiers as fast as horses? There's a great book by Rabinbach that covers these issues in pre-WWI France, and the Fascist-era's jogging soldiers, the celeri, are well known in their own right. In Britain, you have the cult of the Peninsular War-era Light Division, well developed by the middle of the Nineteenth Century and still going strong. It would be a little odd for the British to give up the advantage that their plentiful horseflesh conferred upon them, and so there was also the thought that rifleman+horse might turn out to be a superior form of warrior --a mounted rifleman. We don't really have a history of the idea of the mounted rifleman, although it has fascinated the English-speaking world since the 1840s. (Stephen Badsey is working to fix that.) Why? Well, the cavalry was seen as an aristocratic, Tory/Conservative, arm, while the riflemen were seen as a "middle class," Whig/Liberal arm. So it's all down to partisan politics --but if we take Daniel Szechi seriously, perhaps very consequential party politics.

So, next question. What happens when the horses go away, as they did in Britain during the interwar period, when the number fell to 200,000 by the mid-30s? One answer is that you say good riddance. You could make an absolutely brilliant field gun, far superior to the Soixante-Quinze, if you didn't have to worry about it being pulled by horses. A good tractor could pull a heavier load, faster, and never needed to be put out to grass to recover from a limp or whatnot. Okay, imagine that all of your wagon needs are being replaced by trucks. Wagons don't have to be hauling necessities about all the time. Usually, they end up loaded with footsore stragglers during most of the march to Point B. So systematise that: form the divisions' trucks up into a company, and have them lift --oh, say, an infantry brigade from the rear of the column to the front, then return down and pick up the new rear brigade and move it to the front...

And you get infantry that marches faster than cavalry. It still can't screen itself --but the point is, cavalry can't screen it, because the divisional cavalry falls behind. So that's it. The end of divisional (horse) cavalry. Also, the end of entire armies, as they march blindly into the fire ambush.

No, of course not. You need to put the cavalry in vehicles that are faster than horses. As I've already shown, there were a host of reasons for wanting to do that, anyway. It's just that by the mid-30s, it was inevitable. And if it was inevitable for the divisional cavalry, it was also inevitable for the British Army's Cavalry Division. To do the necessary screening work that the division had always done, it would need to be put into tanks.

Pretty straightforward? Right? Wrong. B. H. Liddell Hart threw a fit at the time, and had not recovered when he sat down to shape the history of World War II. Why? He had a host of arguments, of which the most telling in some people's minds is evidently the army's decision that the Inspector-General of Cavalry should be the commander of the division. Hart argued that the new "Mobile Division" should be under a tank man, who had ridden around on AFVs a lot, and thus understood them, whereas riding around on horses makes your brain freeze into arteriosclerotic stupidity.

The Army didn't make its decision in the dark. The IGC was to be the commander of the Cavalry Division for the same reason that commander of military districts would become GOCs of divisions upon mobilisation. He already had a headquarters and a staff, so he could organise the darn thing! If the particular IGC of 1937 or 1938 was too stupid to command an armoured division (and some proof, or even a name might be forthcoming from Hart), the army could always appoint another IGC. And since Hart's choice to command, Percy Hobart, was a arsehole and petty tyrant utterly unsuited to a field command, whatever his merits as an organiser, he is no-one to be throwing stones.

But in reality, Hart is just throwing up obstacles to obscure the real issues at stake.

Those Wacky Hittites: Postscript

I launched a fairly wild notion in the course of my text explosion on the Hittites t'other day.
[Yes! I will consider the possibility that lithium might be right for me!] Which is that historical linguistics might be radically wrong about what Proto-Indo-European (PIE!) looks like, essentially because we've put Sanskrit in the wrong place in its development. There's actually much more substantive critiques of older reconstructions of PIE, and competing new ones, not that you'd know it from the sorts of places where most people learn about ancient Aryans, which is still trapped in the tar pits with Georges Dumezil. Hating on certain people is reason enough to make my point as starkly as possible.

Sanskrit was invented out of "Prakrit" by a guy named Panini, who probably wrote after 100AD.

Hey, look! I just got a billion Indians, a hundred historical linguists, and 10 very old PIEists mad at me!

But, dudes, seriously. The first dated Sanskrit texts are the Buddhist modified Sanskrit literature produced in Kashmir about 100AD. This is called "Buddhist modified" because it doesn't follow Panini's grammar. So, the first "modified" Sanskrit comes prior to the first "orthodox" Sanskrit, an inscription commemorating the building of a dam in about 150AD. BMS looks a great deal like the variety of "Prakrits," i.e., dialects, that are documented prior to 150AD. The Prakrits are all, of course, derived from Sanskrit. And we know this because.... well, because!

Okay, no. The "proof" is that the Vedic literature that forms one of the bases of the Hindu religion dates to 1500BC. Well, how do we know that? The earliest copy of the Vedas is in a manuscript written in 1450, and the first Vedic text dates to the mid-1100s. So we need to find a way to push this back another 2500 years. The claim is that the Vedas were handed down by an oral tradition. Anyone who has ever played "Telephone" will find that hard to swallow, but the advocates say that the Vedas were sacred, and the super-important-importantness of religion trumps all human frailty. (Wendy Doniger's excelllent recent book has a less-than-excelllent shaggy dog story about how Vedic ritual practitioners needed to keep every word of it secret from the impure --who, what, weren't allow to eavesdrop on the rituals?-- and so didn't write it down, or only wrote it in code, or something. Until they just stopped doing it.) Note that this theory also needs to explain the textless propagation of a massive commentary on the original Vedas. What, apart from not wanting to offend people (a noble objective, but one that can only be carried so far) obligates us to believe this, centuries after we abandoned the idea that the first five books of the Bible were written by Moses?

And it is wrong at another level. There is language change in the Rigveda. The (European) author who compiled it says that the different manuscripts before him rarely had variant readings, but then we are looking at manuscripts coming out of a four-century written tradition by this point. And, reluctant as they are to face the issue, modern scholars acknowledge that extended portions of the Rigveda are incomprehensible gibberish. It may not necessarily "Telephone" gibberish, but that's the way to bet. Here's a brief (and old) summary of some arguments about reading the Rigveda.

But back up here to the whole "1500BC" thing. Where do we get this date? Wikipedia maintains the pure pull-it-out-of-your-assness of the actual date. I've already mentioned that it is a backdating from the Trojan War. Which we don't even think happened any more! And the backdating happened like this: the great Nineteenth Century Indologist Max Mueller was asked whether this dating made sense in the light of his work. Mueller replied that the last possible date for the composition of the last of the Vedic literature (the Upanishads) was the date of the Buddha, who was clearly reacting to them. Buddha lived about 550BC. And there were 5 or so major generational stages to be detected in the Vedic literature, and 200 years was a perfectly reasonable length of time for the changes in each stage, so 1500BC was A-OK in his book!

Dudes: you don't even have a date. You have a guy who says that the date isn't inconsistent with the numbers. Except he's wrong. The modern accepted account puts the Buddha ahead 150 years from Mueller. And yet that's only because they're choosing to use something. We could easily go later, although that provokes continuing scholarly resistance, even though archaeology suggests 400BC as an earliest limit. And then there's the uncomfortable fact that if we put the Buddha after the date of the Persian, never mind Macedonian invasions, we change the cultural context completely.

Moving on, it is observed that the Rigveda reads a great deal like Zoroaster's Gathas, the authentic original core of the Zoroastrian holy texts, the Avestas. I'd be less impressed if it weren't for that bit about the Bible. But more than that, the Book of Mormon reads just like the King James Bible. So clearly the Book of Mormon dates to 1000BC! Well, no. The KJV is a recent translation, the Book of Mormon a deliberate imitation. Is this a valid comparison? Well, the Avestas as we have them come from manuscripts collected in Bombay in the late 1700s. They can be physically dated to about 11ooAD, and the tradition, is that they were compiled by a commission set up by Ardashir I (226--241). The language here is a little complicated. We usually refer to the eastern Indo-European languages as forming two larger families of Indo-Iranian (spoken in Iran-like places), and Indo-Aryan, spoken in India. The fact that Urdu-speaking Pakistani truck drivers can make themselves understood in Teheran and Delhi suggests that the border distinguishes these languages less than nationalism (see further: talking to Croatian linguists about "Serbo-Croatian"). So I am going to go with the eminently scholarly Indo-Whatsit to describe a grab-bag of languages and dialects, of which Sanskrit is attested in 150AD, Avestan about the same time. And when we go earlier, we learn surprising things.

Now, there is every reason to be suspicious of Ardashir's motives. He was the founder of a new, Sassanian state, and needed to formulate an appropriate ideology for it. In doing so, he had to sideswipe the Arsacid (Parthian) dynasty/regime that had ruled in Iran for over 400 years by his time, because they were still around, his rivals, and Roman allies. That gave him a motive for skipping back to the Achaemenians as a model, and notwithstandinghis traditional prominence in the story, a Sassanian state religion of Zoroastrianism only comes into clear focus along side the Arsacid official state church in 301AD. If Zoroastrianism had not existed, the Sassanids would have had to invent it. Only, both can be true.

That is, the Achaemenians were a natural alternative source of legitimacy, and even if the Sassanians knew nothing else of them (or knew things that did not work for them), they would have had the same Greek literature that we have. So they would have the names, some stories, and the figure of Zoroaster. That's not saying that these are the only sources. It's just saying that these are the sources that we know they had. What we do not have is any doxology, nor the slightest evidence that the Parthians had any doxology. Zoroaster is mentioned by Plato. So is Pythagoras. Does Zoroastrianism have anything more to do with Zoroaster than neo-Pythagoreanism to Pythagoras?

So what is the actual earliest-evidence of --well, let's not call it "Sanskrit," or "the Vedas," but something more neutral, such as "Indo-Iranian context?" The answer goes back to the Hittites, again. In about 1380BC, Suppililiuma III signed a treaty with King Shattiwaza of the neighbouring Mitanni state. Treaties were a big deal, so you called on as many gods to witness them as you could:

"[T]he Storm-god, Lord of Heaven and Earth, the Moon-god and the Sun-god, the Moon-god of Harran, heaven and earth, the Storm-god, Lord of the kurinnu of Kahat, the Deity of Herds of Kurta, the Storm-god, Lord of Uhušuman, Ea-šarri, Lord of Wisdom, Anu, Antu, Enlil, Ninlil, the Mitra-gods, the Varuna-gods, Indra, the Nasatya-gods, Lord of Waššukanni, the Storm-god, Lord of the Temple Platform (?) of Irrite, Partahi of Šuta, Nabarbi, Šuruhi, Ištar, Evening Star, Šala, Belet-ekalli, Damkina, Išhara, the mountains and rivers, the deities of heaven and the deities of earth."

Varuna, Indra, and the Mitra and Nasatya-gods appear in the Rig-Veda! We also have Indo-Whatsit names for both Mittani and Kassite (the guys in charge at Babylon in these days) kings, and some Indo-Whatsit vocabulary is used in the horse-training manual of Kikkuli the Mittanian, a Hittite-era bestseller preserved in multiple copies in Bogazkoy archives. Clearly, then, the Mittani had been conquered by an Indo-Whatsit speaking tribe of expert charioteers! The Indo-Whatsit vocabulary in Mittani is a superstrate. Curiously, the Hittite gods mentioned immediately previously are all "Hattic." The Hittites used this unique language in some religious texts, adopted their gods, and even gave Hattic names to their kings. This, of course, proves that invading Indo-European Hittites conquered the Hattic people, and their language is presumably a substrate74.

Opposite conclusions from identical datum suggest that the conclusion is dictating the explanation. In this case, the Indo-Europeans as invaders. Now, there is precious little reason for this a priori. Besides the Hittites and the Indo-Whatsit rulers of the Mittani and perhaps Kassites, we have more Indo-European languages attested in this part of the world. West and south of the Hittites, various independent states speak the near-Hittite Luwian language. Across the cozy Aegean Sea, in Greece and down to Crete, we have an early dialect of Greek recorded in the famous Linear B tablets. Looking up the Euphrates and Tigris as they break into the mountains through one tributary valley and satellite mountain spur after another, we find unexcavated city states everywhere receiving strings of horses raised on the mountains on either flank, some carrying tin carried down by road. Somewhere in here, there is no reason that PIE could not have been first spoken, somewhere near where the Hurrian-speaking kingdom of Urartu/Armenia emerged.

But that's not what we argue. The invasion happened. And re-happened. For example, the Hittites must have invaded and overthrew the Hattians about 2200BC. and some Anatolian cities were burnt at about this time! What other possibly explanation is there than a migrant horde? Trevor Bryce supports my snark, but allows for another invasion, about 1800BC. And about 1500BC, the Linear A writing people of Knossos in Crete were supplanted by those afore-mentioned early Greek writers, the "Mycenaeans." Clearly they invaded from the north, and would soon get civilised enough to appear in Hittite records as a recognisable form of Homer's Achaeans. More Greeks would in turn overthrow the Mycenaeans and much else about 1150BC in a "Sea Peoples" invasion. We also have a preinvasion in the east, where history used to usher the "Persians" onto the stage about 800BC, leaving the Indo-Whatsists of Mitanni to be a precocious offshoot, although people don't do that any more. (For an approach to early Iranian history I like much better, see here.) Invasion: it happened, and they came from the Black Sea.

You would think that no-one ever proposed an alternative. But there is one, the one I've already noted, the "Anatolian homeland." It is even Colin Renfrew's version. Renfrew now admits that he's wrong in imagining a scenario in which agriculture, after being invented in Iraq, crossed the mountains to Anatolia, hooked up with some PIE-speakers, and spread with them all over Europe, perhaps at the same time that neighbours speaking Ural-Altaic, Elamo-Dravidian, and Afro-Asiatic speakers were making for the territories in different directions, spreading agriculture with their language families over all the Earth. Although nearly Biblical, the dates (c. 6000BC) are too early to be plausible, and clearly at least the regions at the extremes of the language-family range came to feature these languages as a result of adoption rather than physically replacement.

This does not, however, dispose of the hippy-uncle-car-trip issue. Where better to put the PIE homeland than in the middle of the road, as opposed to the muddy, uncrossable ditch that we mean when we vaguely point to the Pontic steppes, and which has always directed traffic north and south, rather than east and west? This leads to J. P. Mallory's disproof, which is almost proof masquerading as disproof. Mallory says that Anatolia can't be the homeland, because language families can only evolve in geographic spaces large enough to avoid contamination by other language families, and the ancient Middle East was filled to the brim with lnguage families and isolates. (It's in here somewhere.) Seriously? How did Hattic, Hurrian, Kassite, Sumerian, Elamite and Afro-Asiatic come to evolve? Do we need to find vast, remote language-spaces for each of them? Of course not: Mallory's implicit claim is that Indo-European is special. And he's right.


But not the way he wants to be. The discovery of Indo-European led to a rush of other language family discoveries. that culminated with Edward Sapir (1884--1939) travelling from one reserve to the next and finding vast, sprawling Indian language families wherever he looked, and making vast claims about language and human consciousness. Many of Sapir's families now look like confirmation bias, and what followed made this blatant to the point where there's been something of a reaction, and linguists now argue that Indo-European in particular is the exception to the rule. "Language families" are not a generalisable heuristic. The majority of languages do not belong to language families. Those that do require special explanation. Colin Renfrew is right, but the fact that most "isolate" languages are identified in non-agricultural societies does not mean that agriculture is the explanation.

Now, I've already supplied an explanation for the spread of Indo-European, that is hardly original to me. That it was entangled with the spread of a prestige technology, the war chariot, and the necessary prerequisites, bronze and the domesticated horse is clear, the residual question being the nature of the entanglement. I say the language was spread by the technology, in the form of migrating specialists teaching local charioteers. The invasion thesis holds that the language was spread by the technology in the form of proto-Nazi panzer divisions, chariot-riding conquerors. This is the so-called "Kurgan hypothesis," based on Marija Gimbutas' reformulation, where the spread of Indo-European could be linked to the spread of a particular kind of burial that is much more archaeologically visible than chariots. More recently, we have this formidable iteration of these ideas.

Why do I disagree with this? I have no problem with the idea that languages "migrate." I speak English, which migrated to North America along with English speakers. Clearly, people migrate, and when they get to their new home, they have to speak some language, and the one they already speak is one option. But that assumes the homogenous movement of large numbers of speakers of the same language, and that is not the usual rule. In my own region, early migrants came speaking many languages. To bride the gap between native and imported languages, they began speaking a mixed languages, or "pidgin," called Chinook. Chinook soon gave way to English, the written language of government and education. This happens in most places, but there are a number of cases where it has not, enough that we can study creoles. What kind of a language is a creole?

The stakes here for linguistic theorists are big. Imagine two tribes, each using a different range. For one or another reason, one range has a few bad years. that group faces starvation. Meanwhile, the other tribe can exploit the resources that the failing tribe can't. So the lucky tribe invites the tribe to live with them. They camp together, the children play and intermarry, and, a generation later, the camp splits to cover the old ranges. If creoles can rapidly evolve into a new language, then a Martian ethnologist who visits at intervals of a century will suppose that the old tribes are extinct, and replaced by two new tribes (likely with new names, but quite possibly adopting older ones) who speak different dialects of an entirely new language. Now, there is no question that this process happens. We can point to the Kiowas and Seminoles and enough other cases to need a technical term: "ethnogenesis." Ethnogenesis also generates new languages by "Creolisation," and if this happened routinely, the language map will not only contain few language families, but we will be able to say much less about history on the basis of studying language families. To dispense with the argument, we will need to argue that Creolisation is a self-limiting phenomena. This is pretty clearly wrong. Creoles are just normal language that emerge through non-genetic processes.

So what makes language families? A specific technological package spread Indo-European, but other packages do not have their own language. There is just precisely no "agricultural revolution" language family. But then most of our evidence is written, and we still have not taken this seriously. The texts are not incidental. They are what we need to study.

In the Middle East, writing starts out in Egypt and lower Iraq at about the same time (3100BC) and the same context of state building. The evidence of contact-era Peru is that this cultural contexts will have possessed multiple symbolic inventories matched to their storage media. The problem lay in stabilising reference, so that any given symbol reliably meant something specific. The choice made, inevitable or not, was that at least two symbolic inventories, of cuneiform scratches on clay tablets and heiroglyphs (painted cartoons), were linked to specific spoken language by rendering some symbols as words, others as syllables. Egyptian and Sumerian have been described as the languages of Egypt and lower Iraq respectively, but I see no reason not to problematise this. All three languages fit a syllabic writing system better than many others, notably Hittite and Indo-Whatsit. The script might well have chosen the language.

Under the first imperial state of which we have written records, the Sumerian written system was formalised to write an Afro-Asiatic language called Akkadian, after the Akkadian rulers of what is now Baghdad, in about 2300BC. The rise of the Akkadian state has been plausibly linked to the "Amorites," who are either the first barbarians to trouble a civilised state, or, more likely, an economic classification. By this argument, the success of city-level organisation has led tothe rise of industrial-scale textile production, which has in turn led to large scale sheepherding on the margins of early civilisation. The new economic context generates new political forms, and Akkadian spreads up and down the rivers to other city states, but not over the mountains into Turkey. If I were pressed for a reason for this, I would suggest that the practice of writing things down on anything as cumbersom as a dry clay tablet may have seemed a little more trouble than it was worth absent a coercive managerial apparatus. (So speaks the sometimes middle manager, all-too often exhausted in his attempts to impose ill-considered top-down iniatives. Hey, consultants and clueless top-level management, quit it!)

That said, Akkadian represents the first script-language problem. It, too, is a syllabic language, and the script can directly borrow the syllable-symbols of written Sumerian. But it does not do so universally. Many Sumerian words are adopted directly into Akkadian. By the time the next empire falls (Neo-Babylonians, c. 1595BC), the world is finally ripe for the idea of adapting cuneiform to new languages. Thus we get the menagerie already encountered, of written Hittite, Elamite, Hurrian, Hattic, Aramaic and Greek, while the Egyptians made limited use of cuneiform Akkadian in diplomacy.

This is interesting because it is easy to imagine a context that was nothing but Akkadian specialists writing letters for kings. The "serious" literature of medieval Europe is all in Latin, even though there were no more Latin speakers in a way in which it would have been unthinkable for the Middle-to-Late Bronze Age to use Akkadian. On the contrary, on the periphery of the great power system we find archives written in native languages in adapted scripts, such as Linear A and Linear B.

This is not the case for the Hittites, obviously. They wrote a great deal of Hittite in cuneiform, to the point where the main problem we have in understanding the language is the use of "Sumerograms," and syllable symbols where they are not quite appropriate. Thus, written Hittite contains many superfluous letters, and we lack some pretty basic words, including the numbers! (Of note here is that one of the key pieces of evidence that Kikkuli was translating out of Indo-Whatsit into Hittite was that he uses ordinal forms [i.e. "eighth"] that look Indo-Whatsit. If we lack the words for Hittite numbes, do we also lack Hittite ordinals?) Of Greek, we have even less. Fortunately, Greek developed into one of the best-attested languages on Earth. Hittitte, not so much. It vanished, buried in the archives of its abandoned cities. That, however, should not be melodramatically overinterpreted. The Hittite state did heiroglyphics, too, and it chose to use Luwian, not Hittite in them. For, like all empires before nationalism, the Hittite regime saw that to favour any of the many languages of its subject peoples was bad politics. To the extent that there was a "family" language used within the ruling family, it could be an exception. But, as far as we know, that language was Hattic for the Hittites. In any case, Luwian was related to Hittite, and had a posthumous relationship with the Hittite state. For by 7/600BC, there were "neo-Hittite" states in northern Syria and adjacent parts of Turkey that claimed the legacy of the old Hittite Empire, and they wrote in Luwian. Our first Greek historian, Herodotus may not "remember" the Hittites, but others can.

That said, historical memory is an interesting thing here. The Neo-Hittite states were gone by Herodotus' time, destroyed by the Assyrians and absorbed by the Persians. He would have known Luwian through its use in another, more present power, an empire that stretched from its capital at Sardis on the Aegean coast of Turkey deep inland --the tyranny of Asia, Lydia. The Lydian form of Luwian is distinctly its own language of "Lydian, and" closely related to Carian, a language Herodotus knew the the Lydian "tyranny" well. He was fascinated with it. All the Greeks were. It had been overthrown by the Persians sixty years before Herodotus' birth, an event that shaped all subsequent Greek history, since it brought many "Greek" city states under Persian rule. Herodotus was in no position to delve into the ancient past of Lydia, since he couldn't read heiroglyphic Luwian or get a reading of them, or at least did not bother to seek one out. Yet Herodotus was not entirely bereft of curiosity. He tells us at least one story that intimates that he was on the same path as we are right now.

This story has it that Pharaoh wishes to learn which is the first human language. It is a curious question to consider, given that he speaks the language of the gods, but on the other hand there is the interests of science to consider. Pharaoh does not order the compilation of massive vocabularies and grammars. How we wish he had! Instead, he has some children raised without any human contact, hard as that is to imagine before the invention of the Gameboy. Eventually, the children come to spontaneously demand "bekos," or bread in the language of the Phrygians of west-central Turkey. Phrygian, language of King Midas, is thus the original language of humanity. We know that this is wrong, but more importantly, we know why Herodotus says this. "Phrygian origins" are a way of getting at the idea that the political legitimacy of the state can be derived of origins and first things. (Midas was a great and ancient king who was a great lover, perhaps of his mother or sister. Thus, he "fathers" the state. Or something like that.) This has nothing to do with the actual history of the Phrygians. Which is not to say that we should not inquire.

Phrygian is an Indo-European language. We know it because the Phoenicians spread the idea of alphabetic scripts in the early 700sBC. It was an excellent way of writing Indo-European languages, and came attached to reasons for wanting to write them that gave rise to a new literature. From the Greek tradition we get the Iliad and the Odyssey. . From the Aramaics of northern Iraq, we get a large literature winnowed down in time pretty much to Biblical texts. (Aramaic-writing scribes got around.) From Phoenicians, Aramaics, Lydians, Carians and Phrygians we get a few surviving fragments. Chances are, this is an accident of preservation. The Phrygians did it earlier and were a bigger deal than the Greeks, as at the time they were a major player, and even get mentioned in the Assyrian archives. Then, they were devastated by a Cimmerian invasion about 700AD or so, and overshadowed by their Lydian neighbours, while going forward Greek-writing states would do better than either.

So who were the Phrygians? They are usually brought onto the scene as invaders from Europe during the dark years iof the early Iron Age, but this strikes a suspect narrative in its own right (it makes these teachers and relatives of the Greeks into Europeans rather than "Asiatics") and entirely unsourced, except in one of the multitude of internally contradictory migration tales the old Greek authors liked to tell. Phrygia sits on old Hittite territory, and short of their running some preliterate Einsatzgruppen, assimilated Hittite elements and so are in some sense inheritors of the Hittite state. I'm sure that historical linguists have solid reasons for not arguing that Phrygian is descended from Hittite, and perhaps one could explain. (Though if the two languages are related, that would open up a can of worms, because Greek is often said to be related to Phrygian, and possibly also Armenian and Old Persian(!))

So, about Old Persian, which is in fact the next major piece of textual interest. Sometime between 522 and 486, the Persian emperor Darius commissioned a great inscription by the side of a major road coming down from the Iranian plateau to the Iraqi plain, at a place called Bisitun. It contains a lengthy trilingual, or text in three languages: Aramaic, Elamite, and Old Persian. The former played much the same part in the Persian Empire as elsewhere in the Middle East, as a language of learned comment and administration that overshadowed lol tongues. The second is an isolate language spoken in Khuzistan, the appendage of the Iraqi plain squeezed into the escarpement of the Iranian plateau down in the southwest of modern Iran. The third is ...well, let's not get ahead of ourselves!

So Darius is known to us from Gree history as the third Persian emperor, and the first not (closely) related to the empire's founder, and he has a fascinating tale to tell. He knows, we all know how the Persian Empire was founded. Even two thousand years later a full account could be dug up in old Babylon, called the Nabonidus Chronicle, while Herodotus has his own version. So we know that a man called Cyrus launched a series of campaigns in which he conquered Lydia and the Neo-Babylonian Empire. (Herodotus adds a third victim, the kingdom of the Medes. We certainly have no reason to doubt that Darius ruled the Medes. It's just that the conquest is not corroborated by the incomplete Chronicles.) Anyway, Cyrus campaigned in various directions until he died in battle. Since kings and generals die in battle far less often than one might suppose, this might cause me to suspect something, were my reserves of suspicion not needed at the next turn in the story. Cyrus was succeeded by his son, Cambyses, an equallly effective campaigner who soon added Egypt to the collection. Unfortunately, there is a well-established axis of relations between the Aegean coastal states and Egypt. From now on, revolution in Egypt will stir things up in the Greek-speaking world.

Cambyses, Darius tells us, died soon after conquering Egypt. At this point, a sinister usurper emerges and takes the throne, and General Darius overthrows him, defeats rebels in all directions, and assumes the throne of the empire, which he will rule for 36 years, creating, as in all long reigns, considerable historical inertia down a certain path. But which path? Even if we did not have Herodotus's account, which says that it was Darius who was the illegitimate usurper, the way that Darius dwells on being born a king and a son of a king would tell us something. The Bisitun inscription is not meant to be read by mortals (although of course everyone at the time would know what it said), but by gods, and gods can be awfully gullible at times. Darius was saying these things to make them so. Conclusion: the whole story that Darius belongs to the Achaemenian family of Cyrus is wrong, because Cyrus was not an Achaemenian.

So what has this to do with language? Since Darius was a Persian, it is, of course, logical that the Bisitun text was in Old Persian, a major language of his empire. Right? Wrong. The major languages of the old Persian Empire, by count of surviving documents, were Aramaic and Elamite. The Persepolis Archives have as many documents in Old Persian as they have in Phrygian (one each, so this is hardly a statistically interesting comparison.) Old Persian is used on coins, on a few royal monuments, and carved, probably after the fact, onto some images associated with Cyrus. Backing up to the Nabonidus Chronicles, we find Cyrus described as the "King of Anshan," a traditional label for Elamite players in lowland politics. If Cyrus were to fill out a Census form, I suspect that he would tick the box for "Elamite," and while Darius would tick the box for "Persian," there would not be that many of them.

Not that Cyrus would situated himself in ethnic terms, for all that the greater part of his army would have been Elamite speakers. On the contrary, he was located within great power politics. His mother was a daughter of the King of the Medes, his grandmother a sister of Croesus, last king of Lydia. These were the alliances that the ancient powers had to make to rid themselves of the shadow of Assyria and overwhelm its cities in 620--605BC. Sardis seems like a long way away from Khuzistan, to be sure, but there is a sense in which the Median Empire was in the middle of them.

But wait. Hold up for a moment. You can get out your ancient history map and see the Median Empire illustrated. What you cannot find is independent evidence that it existed. There was an anti-Assyrian confederacy of "Medes," but, such as it was, it consisted of all the highland tribes arrayed in a semi-circle around the Assyrian homeland, assembled into one great coalition of grievances after the eruption of the Cimmerians --whom the Assyria identify as an Indo-European speaking people up there in the mountains of Kurdistan. This might not be an ethnic label, any more than "Tea Partiers" represents a rebellious tribe living in Teapartia. So who were the kings of the Medes from whom Cyrus was descended? Go forward a bit into Herodotus, and we find Croesus again, making war on the Medes by besieging their great city of Pteria. Herodotus is a little vague about where Pteria was, but until recently no-one doubted that it was in central Turkey. Go back far enough and Bogazkoy used to get singled out. More recently, another very large hilltop city not too far east of Bogazkoy, but destroyed and abandoned at the right time, has been singled out: Kerkenes. The problem that has led scholars to retract their identification? Kerkenes was a Phyrgian city.

Woo. Spooky. Phyrgians, Greeks, Hittites, Medes, Persians, Lydians, Cimmerians --they're all blurring together into one composite Early Iron Age entity, exchanging influences, princesses, and presumably words and jargon. It sure was much simpler when we could derive them all from the Pontic Steppe and assume that the Greeks went west to heroic, virtuous Europe, while the Persians went east, to servile, decadent Asia!

But there are a couple of other curiosities about the Bisitun inscription. First of all, it barely mentions "India," even though "India" was a key part of the early Persian conquests according to Herodotus. Either this land of millions of square miles and vast wealth hardly even impinged on the consciousness of the king, or, more likely, he is referring to the region on the right bank of the Indus --Afghanistan, Baluchistan and parts of eastern Iran that get often get that label later. The idea of Indian would be transferred to the whole subcontinent much later. This opens up the possibility that the Greeks thought the same way, which is very interesting, because after Alexander the Great conquered the Achaemenian empire, the eastern parts fell in to the hands of the Seleucid dynasty and various "satraps." In about 255BC, the remote but important province of Bactria (it's on the only practical road for bringing elephants out of India) rebelled under the leadership of a "Greek satrap" named Diodotus. Actually, ethnic labels aren't very helpful here. All we know directly is that the "Greco-Bactrian" state had Greek-style cities, used Greek on its coins. It might have been a racially-pure garrison apartheid state, or something more congenial. Anyway, Diodotus defeated Arsaces, semi-legendary founder of the Parthian state, the Seleucids, and a bunch of others who got in his way in the course of establishing this Greco-Bactrian state, which lasted for several hundred more years, and is well known from archaeology and Indian tradition, as well as to an extent in the classics.

What makes this interesting is that at the same time, and in the very places that Diodotus claimed to rule, we know of another empire: that of the man who calls himself Priyadarsi and sometimes Devadutta in a series of monumental inscriptions in Aramaic and also the "Prakrit," ie. dialect of Sanskrit, called Kharosthi. Buddhist sources of the so-called Mahayana tradition especially call this king "Ashoka," and describe him as a virtuous ruler living a hundred years after the Buddha (that is, 250BC, in this tradition.) He founded many stupas, spread the Buddhist word, and in various ways was the prototype Buddhist world emperor, or Cakravartin ("Wheel-Spinning King.") And now that we have the Kharosthi manuscripts recovered around Ghandhara in the Northwest Frontier Province of Pakistan, we know that this is not mere legend. Or, rather, it is a legend that goes back almost to 100BC. And as with Herodotus' slightly more ancient world, it is hard to believe that there were not more symbolically stored historical traditions accessible to them than there are today. Just as the Sassanian kings destroyed a huge Elamite inscription and mural in order to make their own at Nashq-i-Rustam, so history was lost as history was made.

The names are also direct Kharosthi translations of "Diodotus." Were Diodotus and Asoka the same person? Probably: it's a less interesting question than some have supposed, because having a Greek name doesn't make the Beloved of the the Gods any more a Greek than having a Kharosthi name makes him an Indian. What is interesting is that we have linked the first manifestation of Indo-Whatsit languages in the Indian subcontinent not to some unspeakably ancient oral tradition, but to the royal propaganda of a powerful and successful ruler on the edge of the Greco-Persian continuum, embedded in a tradition of using the purposeful creation and manipulation of languages and scripts in the interest of establishing historical legitimacy. Here is a language of writing, law and power, a language that can displace local tongues and creoles and, in time, become the language and literature of two great world civilisations.

Because it would be naive to think that these developments will have no effect on the further development of Greek or Latin. Indeed, it would be a mistake to think that these developments have any centre, except perhaps in the Anatolian middle. Rather, ideas, words, scripts and grammars are moving east and west, north and south along the artery roads throughout this period to create the Indo-European continuum.

Sorry about the abruptness of the ending. I've got to put this to bed --it's taken far too much of my time already.