6
ON 28 JANUARY 814, at the age of sixty-five (considerable for a Carolingian male, whose lifespans normally weighed in at more like fifty years), the emperor Charlemagne shuffled off his mortal coil. He was also unusually tall. A nineteenth-century estimate put him at about 190 cm, but an X-ray and CT scan of his skeleton performed in 2010 suggested 184 cm. Either result would put him in something like the ninety-ninth percentile for height in his period. Einhard also gives an arresting image of the king in later middle age:
He had a round head, large and lively eyes, a slightly larger nose than usual, white but still attractive hair, a bright and cheerful expression, a short and fat neck, and he enjoyed good health, except for the fevers that affected him in the last few years of his life. Toward the end, he dragged one leg. Even then, he stubbornly did what he wanted and refused to listen to doctors, indeed he detested them, because they wanted to persuade him to stop eating roast meat, as was his wont, and to be content with boiled meat.1
Aside from physical ailments, his latter years had their fair share of other sorrows too, although there was a nasty way in which the most obvious of them greatly simplified the politics of his empire.
As Charlemagne began to age, and even before, the great issue facing his empire was, predictably, succession after the death of its founder. Charlemagne had a first go at settling matters in 806, when a great assembly of Frankish magnates agreed to the arrangements set out in a formal document whose text still survives: the Divisio Regnorum. Charlemagne had five wives (just the one at a time), many concubines, and lots of children: the best guesses come in at a round dozen. The settlement of 806, however, focused on three adult male heirs: in order of birth, Charles the Younger, Pippin and Louis. All three were to have access to Italy, but otherwise they were to rule separate realms. Louis had long been established in Aquitaine, Pippin was given much of the old Lombard kingdom in northern Italy, and Charles the largest territories including the old Frankish heartlands in Neustria and Austrasia (Figure 12). No mention was made of the imperial title, but it is usually assumed that it was destined for Charles, because of the evident dominance of his position.
But Pippin of Italy then died in 810 and Charles the Younger in 811, which left Charlemagne with just the one adult son: Louis, destined to be known as Louis the Pious. Succession arrangements were revised accordingly in 813, when Louis was brought to Aachen from Aquitaine, whose designated king he had been since 781 when he was only three. At another great ceremony, Louis was not only named heir to everything, but anointed as emperor. No doubt the old king was saddened by the deaths of his sons, but he was also perfectly well aware of his own, his father’s, and his grandfather’s successions, where attempted divisions between brothers had generated massive political instability. He may have been able to comfort himself, therefore, with the thought that God, who had obviously created the empire, was now directing affairs in His characteristically mysterious way, to ensure its untroubled replication in the next generation.2
Almost exactly seventy-five years later, however, in February 888, the nobility of the western half of the Frankish world chose a non-Carolingian as king: Eudes or Odo. A Carolingian restoration followed in the person of Charles the Simple, great-grandson of Louis the Pious, but Charles was king only of West Francia and never emperor, while by the time of Charles’ death in 929 East Francia and Italy had each gone their separate ways under non-Carolingian leadership. For all the success of his own lifetime, Charlemagne’s imperial creation came and went within a century, whereas the Roman Empire, by comparison, lasted – depending on what you count – for over 500 years certainly, and maybe 1,000. Even within West Francia, moreover, Carolingian rule was destined not to last out the tenth century, being definitively replaced from 987 by Hugh Capet, grandson of the intrusive Odo’s brother Robert, whose line would then rule in an astonishing sequence of uninterrupted father-to-son successions down to 1328. So why did the Carolingian Empire not last longer? As with the overturning of Merovingian rule, we are dealing with a much larger phenomenon than the mere replacement of one dynasty with another. Rather like the mid-to-later seventh century, the mid-to-later ninth century – and indeed afterwards – saw a substantial fragmentation of central control more generally. To understand the swift disappearance of the dynasty’s control over large parts of its former empire, it is necessary to get to the heart of how Charlemagne and Louis the Pious governed the enormous territories which God had given them.
THE GODFATHER (PART 1)
Many explanations have been offered for the empire’s lack of durability, not least the character of Charlemagne’s son and heir: Louis the Pious. Without looking very hard, it’s possible to find characterizations of him in the sources as a humourless and unwarlike individual who would have preferred life in a monastery to occupying the most powerful throne of his day. It’s then not a difficult step to do a little amateur psychology and come up with variations of the round peg in a square hole. And this is exactly the kind of explanation of which W. B. Yeats, while not thinking remotely of the Carolingians, would have profoundly approved. In his view, the centre couldn’t hold because ‘the best lack all conviction, while the worst are full of passionate intensity’. But all these characterizations of Louis were post de facto rationalizations, penned by individuals struggling to explain the political troubles that affected his later years, when he was locked in a wrestling match with his three older sons: Lothar, Pippin and Louis the German. These three were Louis’ sons with his first wife Irmingard, and in 817 he had publicized a succession settlement for them, the Ordinatio Imperii, declaring that he would have no further heirs. By the late 820s, however, he had remarried and had another son, Charles (destined to be known as Charles the Bald, but presumably, aged four, he wasn’t yet so), and manoeuvred to include him in the succession. This generated the usual dogfight, which saw Louis generally supported by his younger namesake, but opposed by Lothar and Pippin. In the worst moment, however, all three older sons were united against him, and the emperor was forced into a formal penance and virtual abdication at Soissons in 833. This was followed by a period of monastic confinement until Lothar’s overweening ambition led Louis the German to change sides again, and the emperor – formally rearmed in a ceremony at St Denis in 835 – emerged resplendent. When you add into this picture of a father who, despite everything, could not control his sons, a formal act of public penance that Louis had voluntarily performed in 822 at Attigny to atone for previous sins, then a seductive image of personal insufficiency becomes hard to resist.
Before buying it, however, it’s important to see what Louis was actually doing penance for in 822. When he came to power in 814, he had no brothers left to fight, but faced a different and equally intractable problem. Since the age of three, his world had been Aquitaine, which he had ruled first in name and then increasingly in fact. He had been to Aachen on only a handful of occasions, and was known to none of the great magnates of the Frankish heartlands, the backbone around which the empire as a whole was constructed. His problem, therefore, was to assert his authority in a world which was used to operating entirely without him. Louis showed both sufficient political acumen to understand the issue, and total ruthlessness in dealing with it. In particular, he swept out of the palace and into a selection of the nearest available monasteries all his many sisters, with whom their father had liked to surround himself, not to mention his concubines, their illegitimate offspring and all the human detritus of Charlemagne’s lengthy rule. This was done not quietly but with great fanfare, Louis declaring himself to have rid the palace of countless whores. In parallel propaganda campaigns, he established a model monastery within three miles of the palace to set a determined new tone for his regime, declaring, in 814, that he’d found the running of the empire to be replete with ‘oppression’ and ‘corruption’.
Piety, therefore, was Louis’ path to self-assertion and it never stopped him from doing what needed to be done. His eldest brother Charles had not married, but Pippin of Italy had a son, Bernard, who was already seventeen at the time of Charlemagne’s death. In true Carolingian style, Louis totally ignored his nephew’s claims in the Ordinatio Imperii (which simply didn’t mention him) and then when he came north of the Alps to negotiate in 817, had Bernard imprisoned and blinded, the unfortunate youth dying two days later from the after-effects. The emperor who did penance in 822 was a thoroughly self-satisfied one who’d taken a firm grip on the reins of power, and could now afford to square things up with both his conscience and the Almighty. Don’t be misled by the nickname: Louis the Pious was absolutely as hard a bastard as his father, and it is of a piece with this Louis that the crisis with his older sons was actually caused by him being overly aggressive. The three brothers, it seems, were originally willing (at least for the moment) to tolerate their father’s desire to insert their non-hirsute half-sibling into the succession, but Louis wrapped up in this a second process where he sought to humiliate two leading Frankish magnates, counts Hugh of Tours and Matfrid of Orleans, who were close confidants particularly of Lothar, the former Lothar’s father-in-law. It was at this point that Lothar smelled a rat and the rebellion began.3
As soon as you look at the detail, therefore, Louis the Pious ceases to look very different from his father in the general character of his rule. Charlemagne, as we will explore in much more detail in the next chapter, devoted huge energy to processes of Christian religious reform, and both father and son deployed a calculated mixture of bribery, GBH and Christian piety to achieve their desired ends. And this, broadly speaking, was what the job of early medieval north European ruler required.
Carolingian government was, in its essentials, fairly straightforward, and did not differ much from what had been inherited from the later Merovingians, although there were many local variations in detail across the wide reaches of the empire. The imperial landscape was divided into numerous local administrative units – pagi or counties – of which, it has been calculated, there were in total some 600–700. They varied greatly in size, wealth and importance, but the ability of any Carolingian to rule depended on a range of revenues and services that were levied on their inhabitants by their designated, appointed head, usually a count. Unlike the later Roman Empire, which operated a developed central bureaucracy, comprising 3,000 very senior posts in each half of the empire, together with numerous medium- and lower-level functionaries, its Carolingian successor functioned with almost no central bureaucracy at all, other than a few tens of officials, who often had multiple roles rather than a specialized task.
For an older strand of scholarship, this lack of a central government machine, and of sufficient lay elite literacy to power one, rather than the more personal failings of Louis the Pious, lay at the heart of Carolingian failure. But this myth too has gone the way of all flesh. It is true that bureaucratic methods were a bit rudimentary in Charlemagne’s time, but this was more to do with established governmental habits (or the lack of them) rather than a structural problem caused by any basic lack of literacy. Charlemagne and Louis the Pious (and their Merovingian predecessors in fact) issued many orders on paper, and communicated by letter on all manner of subjects. And their magnates were perfectly capable of dealing with what was sent to them. You could always employ a clerical scribe of some kind to pen any reply that might be necessary, and, anyway, old visions of an illiterate lay elite in the early Middle Ages have been overturned. These were Christian men and women, conscious too – whatever their original ethnic background – of a classical heritage where education and civilization were inextricably linked. They certainly learned enough Latin to read their Bibles, therefore, even if, unlike their late Roman predecessors, they were largely learning it at home from their mothers, rather than at school from a professional teacher. As a result, not all of their endings were absolutely correct – thus hastening on the grammatical simplifications which would turn Latin into its various medieval Romance derivatives outside of the Church – and they learned largely passively: to read, that is, rather than to write.4
This did generate a massive cultural shift from late Roman norms where the elite had learned to read and to write in a rigid form of classical Latin, to one where the only writing, by and large, was done by clergymen and the secular elite stuck to a little edifying reading. But there is no evidence that any of this seriously hampered the effectiveness of Carolingian rule. Slightly paradoxically, the evidence stacks up in the other direction, since bureaucratic habits waxed stronger in the Carolingian world as actual monarchical power was on the wane, at least in the field of legal affairs.
Roman ideologies focusing on the importance of written law in marking out a higher order of human society were alive, kicking, and thoroughly understood at Charlemagne’s court. The overall legal situation within the empire was complicated, however, by the fact that every major region had its own pre-existing law code. Charlemagne was careful to respect these traditions while nonetheless taking every opportunity to appropriate for himself the image of divinely directed imperial legal reformer, after the pattern of Justinian, by reissuing new versions of all the old texts in his imperial name, not least at the first major assembly he held after the coronation, at Lorsch in 802. In fact, the texts were little changed and the whole process was mostly gone through for its propaganda value. At the same time, and particularly after the imperial coronation again, much more practical updating regulations on a whole series of matters – from the moral and religious to the highly pragmatic and immediate – were added to the pile of legally valid regulations operating within the empire. These were often preceded by discussion at the regular assemblies where the ruler met his great men, and, once publicized, were supposed to apply throughout the empire, providing a body of regulations that was supplementary to what was already in the older law codes. If much of this was being put down on paper from 802 onwards, initial Carolingian bureaucratic rustiness shows up in the recording and dissemination of the texts of these new decisions.
Some were recorded on paper as a series of rulings in texts known as capitularies, of which a great many survive from the later years of Charlemagne and the reign of Louis the Pious. But we also know that there were many assemblies and lots of important decisions being taken in the early years of Charlemagne. At that point, however, most of them were not being written down, and, even later in Charlemagne’s reign, those decisions that were recorded were not always done so in the same way by all participants at a given assembly. At least, variant texts sometimes survive from the same assembly, so single, authoritative listings of rulings were not always generated.
Not until the middle of the reign of Louis the Pious, moreover, did anyone start systematically to collect the capitularies as a body of material. In the mid-820s, a first attempt was made by Ansegius, abbot of the important monastery of St Wandrille (quite possibly by royal request, though we don’t know this), but his efforts show up the limitations of what could be done at that stage, given the weaknesses of past practice. His collection was not able to find all the capitulary texts that have come down to us via alternative routes, and he misattributes some of those that he did find. Up to the 820s, therefore, the empire was only groping towards regularized bureaucratic habits in legal matters. By the time of Louis’ son Charles the Bald, these teething troubles had been ironed out. His West Frankish assemblies customarily generated a single, authoritative text of the new decisions, all new sets of rulings were systematically added to the pile of existing capitularies, and it was now commonplace, where appropriate, for new decisions to be cross-referenced to older ones, making all this material so much easier to use.5
So, even if it started out a little rusty, Carolingian bureaucracy eventually got the hang of regularized government by paper, and there is no sign that the less intensive patterns of elite lay literacy characteristic of the early Middle Ages were any great hindrance to the process. But the fascinating point is that these bureaucratic habits were reaching their peak in West Francia under Charles the Bald, at precisely the same time as central royal power was beginning to weaken. Indeed, that part of Charlemagne’s empire which best maintained traditions of a very powerful central authority into the tenth century (even if not in the hands of a member of the dynasty) was East Francia, and there the written capitulary tradition failed to get off the ground at all. This is not telling us that central government power was inversely related to the strength of its bureaucratic structures, I think, merely that these structures did not play any very crucial role in the Carolingian governmental system, in so far as there was one.
Lacking much in the way of a central bureaucracy, the rule of Charlemagne and Louis the Pious relied on the count of the pagus as the workhorse of the system, who was responsible almost single-handedly for making it all work. First and perhaps foremost, it was his job to extract any revenues owed to the king. He was sometimes responsible for the income from any royal estates that fell within his jurisdiction, he collected any customs and tolls due from markets or fairs, again passing on to the king his designated share, and he was in addition the chief legal officer charged with presiding over regular courts, and, again, passing on to the king his rightful percentage of any fines levied. To reimburse him for all this effort, the count too was entitled to a percentage of most of the revenues he raised, and might also be temporarily granted – while in post – the annual revenues from a convenient piece of royally owned fisc land.
Counts were also responsible for turning out most of the royal army. Carolingian monarchs maintained in and around their palaces a hard core of what were essentially military professionals: a mixture of hard-bitten lifers and the younger male offspring of some of their leading landowners on temporary attachment. The bulk of Carolingian armies (as of their Merovingian predecessors) consisted not of these professional forces, however, but of local landowners and a designated portion of their dependants. The landowners’ obligation to serve was not infinite, and may have been limited to no more than three months in any given year, but, when the royal command came, it was the count’s job to mobilize the liable individuals and rendezvous with the king at the stipulated time and place. The contingent also had to be both properly armed and supplied with many of the necessary foodstuffs for the upcoming campaign.
Since counts were themselves, for the most part, from the landowning community which they led, rather than complete outsiders, Carolingian government was essentially a partnership between ruler and local landowning community more generally, which relied on the count doing his various jobs to make the whole thing work. A certain amount of checking up had to be routine, done partly by the king in person in the areas of the realm where he was customarily to be found. For Charlemagne in his later years and Louis for much of his reign, this generally meant the Frankish heartlands either side of the Rhine in northern Europe. For all the prodigious distances he certainly covered in his life, even the younger Charlemagne had not been a properly itinerant monarch aiming to criss-cross the entirety of his territories in a regular pattern, and many outlying areas of the empire had only ever seen an occasional royal progress. For much of the empire, the checking role was played by missi, who, as the system became increasingly regularized under Louis the Pious, generally consisted of a pair of high officials: one a cleric, the other a layman.
But even checking had its limits, not least sociopolitical ones, since missi were for the most part connected to the same kind of aristocratic-cum-gentry landowner networks as the counts, and the extent to which they were willing to interfere might well depend on pre-existing relationships. The timing of Bishop Wulfad of Bourges’ attempt to win control of an estate in Burgundy, for instance, has been plausibly attributed to the fact that one of his kinsmen had recently been appointed missus there (although in point of fact his attempt failed). Fundamentally, therefore, effective rule depended on the prevalence of excellent relationships between the king and his counts, and, if relations were not good, there were all kinds of ways that counts could lob a spanner in the royal works without having to go as far as open rebellion. When it came to military service, for instance, all you had to do was just turn up a bit late, or advance to slightly the wrong rendezvous, and that might be enough to derail a campaign.6
What really held the empire of Charlemagne together, therefore, were the bonds of personal loyalty and connection that its rulers were able to form with the local landowning elites, who filled the position of count and otherwise dominated their home communities. By Christmas Day 800, Charlemagne’s position was an extremely strong one, and no count was about to step too obviously out of line. But it hadn’t always been that way. If his brother had not died so unexpectedly, for instance, so that there had been two Carolingians vying for magnate loyalty over an extended period, then his political situation would have been very different. One recent study, focusing on the realm built by one of Charlemagne’s grandsons east of the Rhine, has offered the rule of thumb that it took about ten years on average for a ruler in early medieval (as opposed to Roman) conditions to build up the relationships and bonds of loyalty with local leaders which meant that he was really in charge of his kingdom. Ten years, that is, to identify a body of men from among the landowning magnate class who, if appointed counts, would on the whole serve you loyally and efficiently.
In Charlemagne’s case, the early death of his brother coupled with the unexpected scale and ease of his victory over the Lombards probably speeded things up, but the basic political process was the same at the start of every reign, whatever the size of unit concerned. Louis the Pious, for instance, went through exactly the same thing on a smaller scale when he was made sub-king in Aquitaine: hence his problem when suddenly translated to Aachen after his brothers’ deaths. What you had to do during these ten years – or five if you were really lucky – was, first, hand out lots of gifts of one kind or another to as many key landowners as possible, which included appointing the right men to all the posts of count within your disposal. This showed not least that you were the generous lord that early medieval ideologies absolutely required. Where we have a good run of charters (documents recording gifts of various kinds) for an early medieval king, they consistently demonstrate that kings had to give away a great deal at the beginning of their reigns, for this is when you recruited the loyal servants who would run the localities in your favour. Louis the Pious, for instance, at the same time as he was sweeping the whores and corruption out of Aachen, made over a hundred donations by charter in the first three years of his reign. This dropped to an average of twelve per year subsequently. If, like Charlemagne, you could work a major victory into these years as well, accompanied by some large-scale expropriations from the losers, then you could accelerate the process accordingly, have lots of extra prestige and a great deal extra to give away.
What you distributed in these years was not just land, although that was the main form of capital wealth in this overwhelmingly agricultural world, so everyone naturally wanted it. But there were lots of other ways of showing favour too: giving individuals the rights to hold a market (and hence collect the tolls), for instance, or your support in a law suit, or you might even arrange a desirable marriage for one of their offspring. With a little imagination, a new king could find countless ways to build up close relations with a functioning network of his landowning magnates.
What you didn’t do, however, was just give stuff away willy-nilly. It had to be targeted, which meant you had get to know your magnates well, in order to have a strong sense of the price of their loyalty and of which ones you could really trust with the important post of count. Some might be content with the kind of gift which would just cause another to start plotting against you, and everyone’s price would be higher when there was more than one Carolingian to choose from. And some you could never trust, however much you gave them. This kind of information you found out through the cycles of meeting and contact which punctuated the ceremonial year, and, perhaps above all, when out on campaign.
By the time Charlemagne took the throne, it had long been traditional for kings to hold annual assemblies early in the campaigning season (the same assemblies that, later, often generated capitulary texts). This was a great moment not only for reaching decisions on the great matters of the day, but also for gathering information, and mutual exchanges of favour. It was also customary for the assembly to be followed by an actual campaign. In the ninety years from the accession of Charles Martel down to 803/4, a major Frankish army was in the field against foreign foes in all but five: three during 749–51 when Pippin was pulling the strings necessary to have himself crowned, 759 and 790. And when there was no campaign, this was always commented upon in the sources. Campaigning, especially the successful variety which marked Charlemagne’s reign, was a great moment for forging the practical bonds of loyalty which would also make it possible for the ruler to win the peace, since they were customarily followed by distributions of whatever booty had come to hand.
Even so, humans being human, you could pretty much guarantee that not every magnate would be equally satisfied, or even satisfied at all. Competing magnates from the same locality were one obvious problem, since if you showed favour to one, the other would be alienated, and sometimes, given the levels of brotherly love around in the upper classes, if you promoted one brother that would make his sibling a lifelong foe. The other thing that had to be done, therefore, especially early in a reign, was root out the malcontents. Tough on rebellion, and even tougher on the causes of rebellion was an excellent Carolingian motto, and no self-respecting king would let the grass grow under his feet on that one. Charlemagne moved heaven and earth – well, crossed the Alps – to remove his brother’s family as a potential magnet for the disgruntled. He then brutally put down two subsequent revolts, one by eastern magnates still smarting from their forced incorporation into the Carolingian machine, the other surrounding one of his sons, Pippin the Hunchback, who was being left out of all the succession plans. Pippin himself ended up in a monastery but his supporters were all executed.7
Up close and personal, therefore, the Father of Europe celebrated by one of his court poets looks a lot more like the Godfather of Europe. What Charlemagne did, essentially, was share a series of revenues generated in the different localities of his empire on a percentage basis with selected members of the local landowning magnate class. For this to work, the localities had to be run by men the ruler could trust, and he could use a combination of generosity, comradeship and intimidation to make sure that he received his due share. All of this worked beautifully (though still not without some serious hiccups) when there was only one Carolingian around and regular victories were lubricating both the reputation and coffers of the emperor. How the machine started first to misfire and then actually to break apart when those conditions ceased to apply emerges with striking clarity in the era of Charlemagne’s grandsons.
‘NO SLAUGHTER WAS EVER WORSE’
Hindsight is ever a wonderful thing, but sometimes it’s not really necessary. Even at the time, contemporaries were well aware that 25 June 841 marked the day when the struggle for power within Charlemagne’s empire reached an entirely unprecedented level of ferocity. The three surviving sons of Louis the Pious lined up and straightforward, outright battle was joined near the small Belgian village of Fontenoy. No bluffs, no blackmail, no blindings, and no long-term rest cures in monastic establishments: the gloves came off. And the result was a slaughter of Frank upon Frank which could be expressed either prosaically, as it was in the Chronicles, or with a great deal of poetic hand wringing, as in the verses of the otherwise unknown Engelbert who fought there for Lothar, the eldest of the sons:
No slaughter was ever worse on any field of war;
The law of Christians was shattered by this shedding of blood,
Whence the company of hell and the mouth of its three-headed
dog rejoice …
Let not that accursed day be counted in the calendar of the year,
Rather let it be erased from all memory,
May the sun’s rays never fall there, may no dawn ever come to
[its] twilight.
Engelbert had fought on the losing side, but even the victors ordered three days of Masses to atone for the horror of inter-Christian civil war, and it wasn’t mere sour grapes which prompted our poet’s sad reflections. The battle of Fontenoy marked the first in a series of catastrophic Carolingian failures to transmit power efficiently from one generation to the next.
The failure had its roots in the same situation which had led Louis’ sons to rebel against him in his own lifetime (page 250). By the time of their father’s death on 20 June 840, one of Irmingard’s children, Pippin, had died, and this had enabled Louis to make the perhaps as yet still hirsute Charles the Bald king in Aquitaine. The dead Pippin had left a more or less adult son, yet another Pippin, however, so the year up to Fontenoy saw two brothers, a half-brother and a nephew negotiating, feinting and blustering until they backed themselves into a corner by a little village in Belgium, and their warriors paid the price. Or, as Engelbert put it:
O what grief and wailing! The dead lie there naked,
While vultures, crows and wolves savagely devour their flesh:
They shake since they lack graves and their corpses lie there
to no end.
Any temptation to blame the whole thing on the sexual incontinence of the father needs to be resisted with the utmost determination.8
The three – relatively – peaceful transmissions of power from one sole Carolingian ruler to another sole Carolingian over the 125-year period covered by the reigns of Charles Martel and Louis the Pious (714 to 840) had required an astonishing amount of luck. Charles Martel was able to grab power so easily only because his two half-brothers had died. Pippin’s path was eased by his brother’s departure for Italy, Charlemagne’s by his brother’s early death, and Louis’ by the convenient deaths of no less than two older brothers. These deaths and departures did not remove all potential causes of dynastic dispute, but they left one individual in each generation with a decisive advantage as measured in the perceptions of the magnates – the equivalents of the aides and second-rank leaders who size up the Godfather’s sons so carefully (page 18) – and each had then rammed home that advantage by acting decisively to remove younger collaterals. This longer view of Carolingian dynastic history offers two important perspectives, which should reduce virtually to zero any temptation to lump the blame on Louis the Pious.
First, even without the addition of Charles the follicularly challenged, his three other sons – never mind that they had emerged from the same womb – were a racing certainty to come into conflict at some point: not necessarily on the battlefield, maybe, but certainly a major stand-off of some kind. Charles’ addition did not materially add to the problem, as the detail itself shows, since at Fontenoy he and the younger Louis the German lined up together against Lothar. Second, at each stage, the job of dynastic simplification was undertaken only after the death of the previous king, not in his own lifetime, as when Louis the Pious had taken out Bernard of Italy, or when Charlemagne and his father dealt with their brothers’ families. Sorting out dynastic complexity, you might say, was the job of each generation when it came to power, not the previous one beforehand, which again extracts Louis from any hook on to which you might feel inclined to hang him.
There were excellent reasons why this had to be so. To start with, death is always an unexpected guest. This is true enough now, but was even more so in the early Middle Ages. A king could not go about pruning his heirs, not only because he – presumably – loved his children (even if they didn’t love each other) but also because you never knew who was going to survive. In 806, Charlemagne had three adult sons, but, by the time of his death less than eight years later, he was down to one. And if more than one heir posed problems, having none at all was the ultimate nightmare (as we shall see in a moment). You might control your daughters’ marriages, since you wouldn’t want multiple grandchildren with a half-decent claim to the throne. The smart money is that Charlemagne had this in mind, as much as his own love for them, when he wouldn’t let his daughters marry and leave his court (page 251). But it was foolish to go around killing sons. Equally important, what fundamentally mattered in succession disputes were the magnates’ perceptions of the candidates’ abilities and characteristics, as it was the magnates’ choices of allegiance which actually decided how things would turn out. Not enough of either of the two Carlomans’ magnates were ready to fight off the ambitions of Pippin and Charlemagne in defence of their former kings’ offspring in 748–9 and 771–2, so power passed quickly and relatively peacefully to one Carolingian alone. In that sense, Louis did his job perfectly well. He made sure that he passed on a decent stock of adult male heirs to the next generation, and it was then up to them to sort out what would happen next.
At this point, the dynasty’s luck ran out, so that instead of coming down to just the one well-established adult male heir, who could clear out all the collaterals fairly easily, the 840s saw two very well-established rulers in the persons of Lothar and Louis the German, and a third pretty-well-established one in Charles the Bald. Indeed, it was probably the fact that Charles could exploit the political no-man’s-land offered up by the head-to-head between his two half-brothers which allowed him to survive long enough to take a firm grip on the Aquitanian power base which his father had provided for him only two years before his own death. As a result, and Fontenoy notwithstanding, no one could achieve a decisive advantage, so the provisional split left by the father was confirmed by treaty at Verdun in 843 (Figure 14). Lothar retained the imperial title with which his father had endowed him back in 817, and was a touch richer than his brothers, but all three took firm possession of extensive realms. The younger Pippin, son of the dead Pippin, remained in Bordeaux, and his struggle with Charles the Bald carried on for a while, but, by 848, it was clear that his half-uncle was the man more likely to. Magnate support therefore eroded away. By 851/2, Pippin was enjoying full board and lodging at the monastery of Saint-Médard at Soissons, and although he would escape again, and even team up with some Vikings in the 860s, he was never able to attract enough magnate support to make himself a viable rival for Charles the Bald.9
We don’t need to immerse ourselves – fortunately – in all the complex details of what happened next. In some ways, the brothers exhibited entirely contradictory behavioural traits. On the one hand, there was a huge emphasis in their public statements on an ideology of fraternal love: the importance of sharing nicely. Nor was it all bullshit. They held an astonishing seventy fraternal summit meetings, some of which even generated agreed joint policies. At the same time, they kept a very firm eye on the main chance and proceeded to exploit any opportunity that came their way to achieve a material advantage over their brothers. There were endless small- to medium-sized wars, therefore, such as the moment in 854 when Louis the German sent his second son, Louis the Younger, to Aquitaine in the hope of undermining his half-brother’s control there.
This particular round of conflict further solidified matters since it ended with both Charles and Louis expelling magnates from their kingdoms whom they felt were overly inclined to support the other side.
The issue which really stirred things up, however, was the fate of the emperor Lothar’s middle kingdom. Lothar died first, on 29 September 855, long before Louis and Charles. He was succeeded by his three sons but one of them soon died without heirs, and his lands were split between the remaining two: Lothar II and Louis II of Italy. The problem here was that Lothar II had no male heirs by his wife Teutberga, so wanted to divorce her and marry the concubine with whom he had already had a son. Seeing an opportunity, Louis the German and Charles the Bald opposed the divorce tooth and nail (although Louis II was more amenable). When Lothar II died in 869, the matter was unresolved and a further fraternal summit at Mersen in 870 led to the uncles dividing up most of the territory while paying off Louis II with a moderate cut (Figure 14).
For all of these adventures, it was necessary to bid for magnate support, whether to ensure that your own landowners were solidly on side, to attempt to seduce over the loyalties of those in your half-brother’s kingdom, or to appeal to the rudderless: those left over after a king died without issue. Similar processes of repeated appeal were also necessary when, over time, a second dimension of political activity came into play as the children of Louis the German and Charles the Bald themselves came of age. Charles the Bald had six sons with two wives, of whom two made it to a politically active adulthood: Louis the Stammerer and Carloman. Louis had three in this category: Carloman (again), Louis the Younger and Charles the Fat. As true Carolingians, these princes expected a share of power while the fathers were still alive, and were entirely prepared to rebel either if they received nothing, or if what they did get failed to match their expectations. They could usually find some landowner support amongst those who were not doing as well as others in the current configuration of their fathers’ regime, and, of course, any member of this next generation had a living half-uncle, who was always ready to stir up a little manure for his half-brother. So Louis the German supported the rebellions of Charles the Bald’s sons, particularly that of Carloman, who refused to accept the clerical destiny his father had mapped out for him and pushed the point to such an extent in the early 870s that his father had him blinded and confined to a monastery (although he subsequently escaped again and lived out his days in his half-uncle’s kingdom). Charles, likewise, had been happy to get his retaliation in first, when the other Carloman led the way against Louis the German in the early 860s and his brothers followed suit.10
While Charles and Louis, having weathered the early storms of vulnerability, were themselves pretty much immoveable from c.850 onwards, the course of Frankish politics was thus rarely stable, both for inner dynastic reasons, and also because Viking invasion was adding an interesting extra dimension to life at the top. Paradoxically, a third and critical dose of instability was provided by the kings’ very longevity. This was all very well while they lived, but the problem came after their deaths: that of Louis the German on 28 August 876 aged seventy, closely followed by Charles the Bald on 6 October 877 at the age of fifty-four. It has long been noticed that the passing of these two old warhorses was followed by a clutch of rapid deaths among Carolingian males: no less than seven more in just five years from 879, beginning with Charles the Bald’s chosen heir, Louis the Stammerer. The dead included all three sons of Louis the German, and here their father’s longevity was the prime culprit. On average, Carolingian males who survived childhood made it to fifty or thereabouts. Louis made it to seventy, so that, at the time of this death, his sons were already forty-six, forty-one and thirty-seven. That they should have died soon after him is thus not so surprising.
And Charles the Bald, too, was at fault, though in a different way. Presumably having learned from his own and his brothers’ behaviour, he looked to smooth the path for his one chosen son and heir, Louis the Stammerer, physically and somewhat brutally removing potential alternatives from among his own offspring, such as the blinded Carloman. But as Charlemagne’s example showed, you could never be sure that one heir was enough, since if the average age was fifty, this means about half would die younger. Sure enough, Charles’ strategy misfired. Instead of letting each generation look after itself, his attempts at assistance for Louis the Stammerer just left things in a mess when Louis himself died in 879 at the age of thirty-three. The Stammerer had two teenage boys, but they both died without male heirs shortly afterwards: another Louis in 882, and another Carloman at the end of 884.
By the mid-880s, therefore, the dynasty was rapidly running out of heirs, which generated a final brief moment of imperial unity, when Louis the German’s youngest son Charles the Fat captured the allegiance of a critical mass of magnates from both his father’s and his half-uncle’s old realms. I was tempted to call this section German, Bald, and Fat in his honour, but decided this might be just a little flippant. Born in 839, Charles was already in his forties by the time he gained recognition in West Francia in 884, and his best days were behind him. Whether this contributed to his celebrated failure to break the grip of a large force of Vikings around Paris in 886 is unclear, but it certainly had a major impact on what happened next. As his grip in West Francia began visibly to weaken, the illegitimate son of his brother Carloman, Arnulf of Carinthia, launched a coup d’état against him in the autumn of 887. Charles suffered a stroke before he could respond, and from early in 888, Louis the German’s old kingdom was in the hands of an illegitimate grandson.11
The really significant development, however, was what happened at this point in West Francia. OK, so Arnulf wasn’t fully legitimate, but he was a Carolingian, and had the advantage of being a strapping twenty-seven-year-old with – hopefully – another two good decades of political life left in him. The West Frankish magnates, however, chose an entirely different path, electing one of their own number as king instead, an aristocratic landowner by the name of Odo. Nor was this the first moment that some West Frankish aristocrats had opted for a non-Carolingian. On the death of Louis the Stammerer in 879, one group of bishops and nobles from the regions around the rivers Rhone and Saône met in synod at Mantaille and elected as their king Boso of Provence. His best royal connection was the fact that his aunt Teutberga had been married to Lothar II (the very woman that he had been so desperate to divorce), while he himself had served both Charles the Bald and then his son. In fact, the surviving Carolingians, although in the middle of their great mortality-fest, found sufficient energy and co-ordination more or less to stomp on this usurper. Odo, by contrast, prospered – up to a point – winning great prestige by sweeping the Vikings out of the Paris basin. In due course, a portion of his nobility decided that they preferred to support a Carolingian, Louis the Stammerer’s posthumous son Charles the Simple, and, on Odo’s death in 898, the entire kingdom came back into Carolingian hands, and would stay so for the best part of another century.12
So why all the hype about Odo? Why make all that fuss about February 888 when it proved to be just a blip? Well, in an obvious sense, it was just a blip: within a decade, the royal title reverted to the Carolingian line. But in another, it was anything but. Especially when you add Boso into the picture, the phenomenon becomes pretty extraordinary. For well over a century, Carolingian kings-cum-emperors had reigned untouched, but suddenly we find two aristocrats within the same decade seeing no reason why they shouldn’t put themselves forward to be king, and other landowners who were entirely happy to support them. The clear blue water which the combined careers of Charles Martel, Pippin the Short and Charlemagne had managed to put between the Carolingians and what were originally their fellow aristocrats had all but disappeared. This brute fact did not go unnoticed. As one contemporary commentator, Regino of Prum, put it:
After Charles [the Fat’s] death, the kingdoms … were loosened from their bodily structure into parts and now looked to no lord of hereditary descent, but each set out to create a king for itself out of its own guts. This event roused many impulses to war, not because Frankish princes of sufficient nobility, strength and wisdom to rule kingdoms were lacking, but because among themselves an equality of generosity, dignity, and power increased discord. No one so surpassed the others that they considered it fitting to submit themselves to follow his rule.13
Nor did the return of the Carolingians in the person of Charles the Simple materially change the situation. He was not by any means powerless (nor simple in a pejorative sense; it should probably be translated ‘Charles the Straightforward’). But it is the case that the tenth-century Carolingian monarchs of West Francia were in a much less dominant position vis-à-vis their own nobility than Charlemagne and Louis the Pious had been. The gap in wealth between themselves and the greatest aristocratic families was nothing like so large, they controlled fewer monasteries and bishoprics, the effective geographical range of their power was confined to the Île-de-France and some close outliers, and elsewhere a quasi-monarchical power (including such old royal monopolies such as the right to hold courts and mint coins) was being wielded instead by the dominant aristocratic families of particular regions (although the patterns of intra-regional power were far from stable from one generation to the next).14
In short, the tenth-century Carolingian monarchs of West Francia look much more like their later Merovingian forebears than the dynasty of Carolingian godfathers who had dominated Europe for a hundred years. In their – relative – enfeeblement, they stand in stark contrast not only to Charlemagne, but also to the line of rulers who picked up the mantle of Louis the German and Arnulf of Carinthia in East Francia. There, dynastic accident meant that the Carolingian line again died out, in September 911 on the demise at age eighteen of Arnulf’s only legitimate son, the himself childless Louis the Child.
But, despite a similar emergence within East Francia, just as in West Francia, of powerful regional nobilities from the mess of the ninth-century Carolingian political process, a powerful central monarchy persisted there alongside the dukes. Indeed, in the persons of what was originally the family line of the Dukes of Saxony, it would even regenerate empire when Otto I had himself crowned in Rome on 2 February 962. To understand why the Carolingian political process of the ninth century created these substantially different outcomes in East and West Francia, we need to think in a little more detail about how it intersected with the networks of loyal magnates which had made Charlemagne and Louis the Pious so powerful.
THE END OF TAXATION
When, in the later ninth and tenth centuries, we see Carolingian rulers struggling to exercise as much influence over the localities of their regna as Charlemagne or Louis the Pious had done, this is telling us that the levers of power at their disposal – gifts and fear – had become less potent. It has precisely the same meaning as it did when we compared Merovingian kings of the later seventh and early eighth centuries with their sixth-century predecessors. The phenomenon has nothing whatsoever to do with any degeneration in the Carolingian genetic stock, and everything to do with a structural shift in the relative quantities of resources being controlled at the centre and out in the localities.
As soon as you think about the political processes which unfolded in Charlemagne’s empire in the ninth century in the light of this key factor, you can immediately see why, over time, control of various assets would have tended to pass out of royal hands. For one thing, dividing the empire between more than one heir immediately reduced the financial distance between the value of the stock of key assets held by any individual ruler (actual land, rights to draw revenues from ecclesiastical institutions or markets, valuable rights to appoint to important positions, lay and ecclesiastical, etc.), and that of those of their greatest magnates. In the Treaty of Verdun in 843, the royal fisc (the sum total of these assets) was divided between each of Louis the Pious’ three sons. This immediately and dramatically narrowed the gap between any one of them and their greatest magnates. The eventual reallocation of Lothar’s lands bought Louis the German and Charles the Bald considerable reinforcement, but they were also having to hand out lands to their children, and never again, except for that brief moment under Charles the Fat, would all the fisc of Charlemagne’s entire empire fall to a single Carolingian. And not only did Charles the Fat not live long enough – that crucial five to ten years – to make it stick, but, by 885, there is every reason to think that the overall total value of the fisc had anyway undergone substantial reduction.
The reasons for thinking this are straightforward. By the 880s, we have magnates who clearly think of themselves, and were thought of by others, as essentially the peers of their Carolingian monarchs. Not only the famous Odo, but also Boso of Provence were considered ripe for royal promotion. Obviously, such individuals would take over the remaining royal fisc if successful in their bid for power, but their own personal wealth must have been more than a little monarch-like anyway. They also had their East Frankish equivalents in the region’s various dukes who were pretty much immoveable from c.900 onwards, so that, there too, magnates were looking much more king-like than they had been a hundred years before: a clear sign of and in itself that substantial quantities of assets had been transferred.15 Equally important, the political conditions of the ninth century are more than enough to explain why a substantial transfer should have occurred. Broadly speaking, the prevailing situation from the later 820s onwards was one of intense political rivalry on at least two levels: first between two or more (and often several) reigning monarchs, and then second between them and other potential royal wannabes, in the form, frequently, both of their own children and of collateral relatives. It does look as though Charles the Bald, at least, learned the lesson and tried – as it turned out unfortunately to ill effect – to limit the amount of political competition that he bequeathed to the next generation, but until the great Carolingian cull of the later 870s, the general political context was one of multiple Carolingians bidding for the support of the militarized landowning magnate class. Without their support, no bid for a throne stood any chance of success.
The price of magnate loyalty was not remotely fixed. If you were faced with a truly scary individual like Charlemagne after he had conquered the Lombards, then you were probably pretty grateful if he gave you anything at all. At the other end of the scale, a collateral pretender currently without a kingdom would have to promise the earth to win support for an attempted coup d’état, and half-installed monarchs facing several rivals would have to pay something in between to shore up the foundations of their rule. But the political narrative of the ninth century makes it obvious that the general Carolingian fear quotient was heading downwards, and correspondingly that the political price of sufficient loyal followers to avoid a lengthy monastic holiday was on the up. In these conditions, there is every reason to accept what the political narrative is telling us, that the building-blocks of power – land, revenues and rights over key appointments – were passing out of royal hands at a steady, and probably increasing rate.
Two additional factors hurried things along. First, the main foreign policy problem facing the Frankish world – and indeed its Anglo-Saxon counterpart north of the Channel – was Viking raiding. The chief military difficulty was Viking speed of movement. Vikings came by ship and often brought horses with them too, or captured them when they arrived, so they moved around with great speed. Countering fast-moving raids was extremely difficult in Carolingian conditions, when messages had first to be transmitted over considerable distances to a royal court, and orders then sent out to military forces which were often not already assembled, since part-time militarized landowners would normally be living on their home estates. One obvious stratagem, therefore, was to devolve power to a trusted regional magnate in endangered territories (basically anywhere within range of a navigable stretch of river), who would be much closer to the action and able to organize a swifter response. Charles the Bald, in particular, faced with the many rivers piercing his territories adopted this reasonable measure in systematic fashion. But one of his trustees was a certain Robert the Strong, who was given a key command on the Loire, and the Odo made king in 888 was Robert’s son. A family of trusted subordinates in one generation might gather sufficient rewards from this service, therefore, to morph into potential usurpers in the next.
This much had been just as true in the early Merovingian era, where the Carolingians had started life as the Merovingians’ most trusted supporters in Austrasia. In the ninth century, however, a second factor stacked the deck still further in the favour of local landowning magnates: castles. We’re still talking prototype, largely wooden models here, not the great stone keeps whose gutted remains still dot much of the landscape of Western Europe. But, in the ninth century, elite home improvement was taking the form of fortification for the first time, and the habit spread fast. In part, it was another response to Viking mobility. If you couldn’t hope to mobilize quickly enough to counter them with large numbers, an alternative approach was to use a smaller numbers of warriors more efficiently, by making it difficult for the Vikings to get their hands on anything worth stealing. Fortified centres could serve as more general refuges and could also be held by relatively few men against relatively many.
Charles the Bald, in particular, was very keen on a whole range of fortifications – not only standard-type strongpoints but also fortified bridges at key points on river systems. Once the habit had been picked up and generalized by the greater magnates, however, this made it more difficult to intimidate a rebellious subject, since a castle could be held against a king just as easily as against Vikings.16
Older generations of commentators, convinced – as everyone was in the era of nationalism – that central state power was inherently better than regional magnate power, saw this unfolding process as a great tragedy, and the Carolingian Empire as a whole, and West Francia in particular, as might-have-beens brutally cut short by the wicked plots of magnates determined to resist their rightful rulers. In this historiographical world, magnates and kings were usually portrayed locked in a death-struggle for control of the landscape and historical destiny, with France and Germany as the Promised Land to which all the better, centralizing trends would eventually lead. More recent writing, less obsessed with thinking of things as either ‘good’ (where they seem to move things on towards a modern world dominated by nation states) or ‘bad’ (where they don’t), has pointed out that, most of the time, kings and landowning elites can be found co-operating and certainly cohabiting more or less happily, and, having experienced twentieth-century totalitarianism, we are also well aware that there is nothing inherently better in central as opposed to regional power. As a result, much of the heat has gone out of the issue and, instead of local magnates bitterly determined to throw off central power, we are happy with a world where most of the prominent individuals contesting power in the late ninth century were – like Odo or Boso – from families with impeccable records as Carolingian loyalists.
All this is fair enough, but any temptation to replace an old vision of magnates determined to undermine royal power with a new one in which loyalists just pick up the pieces of royal collapse should be resisted. Positing too clear a dichotomy between loyalty and rebellion in ninth-century conditions would be misleading. Being loyal to a dynasty never stopped anyone from profiting from their loyalty. By definition, those best placed to profit from a process whereby assets were passing out of central control would usually be loyalists, for they will have been in the right position to gain most from any process of transfer. In exactly the same fashion, the Carolingian line started out as the prime supporters of Merovingian rule in Austrasia. And, as early Carolingian history shows so well, being a loyalist at one moment never precluded more self-assertive activity at the next. The patterns of behaviour displayed by the line of Robert the Strong, therefore, are not remotely odd, and find their match in other magnate lines besides. What they actually do is mark out an excellent recipe for success: strong initial loyalty to the ruling dynasty bringing enough wealth and power for subsequent members of a magnate line then to make themselves independent of central control.17 And, of course, should things go really well, you might even find yourself in a position to try to seize the throne itself.
Rather than worrying too much about magnate morals and the extent to which the beneficiaries from the erosion of central assets were loyalists or usurpers, there is a more interesting point to be made. The fact that the political processes of the ninth century had such a profound effect on the balance between centre and locality is actually telling us something important about the fragility of central power in early medieval conditions. Unless the transfer of power from one generation of the ruling dynasty to the next was very straightforward, then the subsequent process of regime-building was always likely to transfer assets away from the centre. And ideal transfers of power were not that easy to arrange. The death of two half-brothers, a further brother’s religious bent and the early death of another were required to make it easy for Charles Martel, Pippin and then Charlemagne to monopolize power from early on in their reigns, and two more sudden fraternal deaths were necessary to clear the path for Louis the Pious. The amount of luck involved in these four generations of unproblematic succession was staggering, and could not go on indefinitely. And once that luck gave out, a cycle of asset transfer was always going to set in to undermine the – or in fact any – dynasty’s overall power. This had nothing to do with the capacity of individual members of the dynasty, or with the greed of particular magnates. It was the product of some key structural changes which meant that Carolingian kings and emperors had considerably fewer levers of power in their hands than any late Roman predecessor such as Justinian.
Both Carolingian and Roman emperors faced, it is worth emphasizing, the same governmental problem. Trying to run geographically vast empires with communications technologies where pretty much everything moved at about forty kilometres a day meant that, in most respects, localities would run themselves. The central authority just didn’t have the governmental capacity to involve itself intensely in local affairs. The trick in both the Roman and the Carolingian context, therefore, was how to devolve power in such a way that it did not allow unavoidable local autonomy to degenerate into a dangerous degree of local independence, at which point the empire would fragment. But if the problem was the same, the weaponry that Charlemagne and Louis the Pious could deploy was considerably less effective than that available 500 years before.
Most important, they had neither the legal-cum-political right nor the administrative capacity to draw systematic and substantial tax revenues from their empire’s agricultural production, by far the largest sector of any pre-modern economy, employing upwards of 90 per cent of the total population. The Roman world was divided into city-based administrative units – called civitates (singular civitas) – which consisted both of an urban core and of a – sometimes very large – rural hinterland. Extensive records of productive capacity and ownership were kept in the relevant civitas, from which tax liabilities for the surrounding countryside were calculated, and officers appointed in the civitas were responsible for collecting and passing on the designated tax levies. These taxes could be taken in the form either of actual produce or of cash according to the requirements and orders of the imperial government. These – relatively – massive annual revenues were spent above all on supporting a large professional army, consisting of several hundred thousand soldiers. The best guess is that something like two-thirds of the annual tax take was spent on the army, the rest being devoted to prestige projects of various kinds and the maintenance of that – again relatively – massive central Roman bureaucracy which dwarfs anything we see in the Carolingian period, or for many a century afterwards.
This articulated, interdependent structure of systematic taxation, professional armies, and overarching bureaucratic apparatus placed some highly effective levers of power in the hands of Roman emperors. At worst, if all else failed, the professional army, independent of allegiance to any local community, could be turned loose on rebellious subjects to compel their adherence to the imperial system. Such a high level of dissent surfaced only with extreme rarity, however, because the other levers of power exercised such an effective pull on local loyalties. The need to survey and measure local agricultural production, in order to tax it effectively, meant that imperial officials were all over local society like a rash, at least once every fifteen years when the tax rates were reassessed. But the fact that these imperial officials tended to be larger landowners, often with established affiliations to the local society they were investigating, highlights another important dimension of the relationship: the vast central bureaucracy was not merely there to administer things, it was also a patronage-distribution machine. The kinds of benefits you got as an imperial bureaucrat – especially as a retired one, which is when you might be called upon to run a tax reassessment for your home community – made bureaucratic jobs both desirable in themselves and, at the same time, the path to local political pre-eminence: a clever little double act which tied locally important landowners tightly into the imperial centre. If you add to this the fact that the empire’s legal system also defined and protected these landowners’ property rights, you can easily see why the army was rarely called upon to keep even rich local landowners in line. The imperial system had far too much to offer them in terms of protection and reward, with just a little constraint being applied now and again to keep them honest.18
Most of the key elements in this Roman imperial balancing act disappeared in the aftermath of imperial collapse. The tax system did not survive the fall of the empire unscathed, especially in areas north of the Loire which saw extensive Frankish settlement, since it looks as though Merovingian kings never taxed their Frankish (as opposed to their Roman) subjects. But tax registers are mentioned occasionally elsewhere in sixth-century Gaul, and Frankish kings regularly fought each other for control of particularcivitates. And when the kingdom was divided between a number of sons, sixth-century partitions often involved what look like geographically odd partitions, with individuals receiving not just discrete blocks of territory, but also clusters of civitates dispersed across the map (for an example, see Figure 14 and compare it to the nice straight lines of the ninth-century Carolingian partitions alongside. This makes most sense if individual civitates were still generating known annual tax incomes, making the purpose behind the geographically odd partitions to achieve a required balance in the annual revenues accruing to each of the parties. But by the later Merovingian and hence certainly the Carolingian periods, except perhaps in the form of limited residual dues (owed in cash, kind, or other services) attached to particular institutions, the Roman taxation structure had ceased to function. There was no attempt to tax agricultural production systematically, and the administrative structure which had made it all possible had disappeared. Royal revenues were now derived from a much more ad hoc mixture of landowning, customs tolls and an agreed percentage of judicial fines. The civitas had also been allowed to break up everywhere into a larger number of counties: a clear sign that its central function – raising taxation – had itself lapsed.
On the face of it, giving up systematic taxation rights was a ridiculous thing to do, but there were good short-term reasons for being so generous. As far as we can reconstruct it, three parallel processes made taxation rights an attractive asset for kings to give away. First, large-scale taxation was much less necessary to successor-state kings than it had been to their Roman imperial predecessors. The majority of Roman tax income was spent on maintaining a professional army. Merovingian armies (outside of semi-professional royal household forces) consisted of their greater and lesser landowners mobilized for individual campaigns, and this liability for military service was one of the key rights that kings enjoyed over their subject population. Second, all the more general evidence also indicates that the economy of the post-Roman West declined not only in overall output but also in complexity and frequency of exchange. This meant that, at the same time as taxation became less necessary, tax yields both declined overall and became increasingly difficult to turn into useful cash: as opposed to kilo after kilo of actual foodstuffs. Third, we can deduce that there was a huge desire on the part of one section of the kingdom’s population for taxation rights to be granted away. The Merovingians’ Frankish followers do not originally seem to have paid tax, and resisted all attempts to impose it with gusto – the odd lynching of tax officials is recorded in our sources – since they provided military service instead. But, from early in the sixth century at the latest, Roman landowners were also being called upon systematically to perform military service. There is a myth out there in some of the literature that fighting in the post-Roman West was done by incoming barbarians and their descendants, while surviving members of the old Roman elite hid themselves in the Church in disgust. There are some famous examples of such a response, but also of the descendants of incoming barbarians moving quickly into the Church too. And, when you go looking for it, there is an overwhelming body of evidence that the vast majority of the surviving Roman landowning aristocracy and gentry morphed not into Christian clergy, but into the early medieval warrior elite. The sixth-century Merovingian Frankish army was built around contingents from civitasterritories, before the civitates ceased to exist, and many of these had never seen a Frankish settler. These contingents were composed of local Roman landowners and a picked bunch of their largest and most aggressive retainers.
From early in the sixth century, therefore, Romans were fighting for their Frankish rulers alongside the descendants of Frankish settlers, but still had to pay taxes, whereas their new Frankish comrades did not. This must quickly have been perceived as unfair and, once it was, tax remissions will have become the favour of choice that Roman subjects wanted from their Merovingian rulers. There is no surviving documentation to allow us to follow the process in any detail. Only some ecclesiastical institutions have had a continuous existence and maintained archives from the early Middle Ages into more recent times, so the only surviving examples of Merovingian tax remission apply to monasteries and episcopal sees. I have no doubt, however, that the absence of extant remissions to lay landowners is merely reflecting the lack of any mechanism which might have preserved such records and has no greater significance than that. As the sixth century progressed, there became no easier way for a monarch to win a little political support than by granting a landowner of Roman descent the seemingly cheap favour of a little immunity from taxation, until, bit by bit, the whole tax structure eroded away. My own guess would be that these grants took the form of allowing the now tax-free landowner to form his own administrative unit free of the old civitas, so that this also explains how the old civitas network broke up into the pattern of much more numerous counties characteristic of the Carolingian period.19
This was a short-to-medium-term transformation with massive consequences for the future. In several critical respects, these changes made it significantly harder for rulers of geographically extensive entities in the early Middle Ages, such as the Carolingian Empire, to prevent local autonomy from cascading into independence. First, local elites were themselves now armed. This made it much more difficult (though certainly not impossible in individual instances) to constrain them in extremis than when Roman rulers had an entirely independent professional army at their disposal to turn on civilian local elites. Not only could armed local elites now put up more resistance, but there were also likely to be much more complex political problems involved in setting one set of armed landowners loose on another one. Second, the jewel in the crown of the local elites’ financial assets – their landed estates – were not touched by systematic royal taxation, and, once the tax structure had eroded, early medieval rulers were much less rich than their late Roman counterparts. Not only did these rulers have less need of a bureaucracy, therefore, but they actually couldn’t afford one, which further diminished the hold they might exercise on local elites. Since the late Roman bureaucracy was as much for distributing patronage as it was for actually administering things, its disappearance removed another magnate magnet from the ruler’s armoury. Third, and not least, because the local elites ran their own courts and provided, through their own armed capacity, both their own legal structure and an enforcement mechanism, the central state mechanism was no longer the ultimate guarantee of elite status through the protection it provided for property rights.
Fundamentally, the early Middle Ages saw the emergence of a new ‘smaller’ type of state structure. With no state-run professional army, no large-scale systematic taxation of agriculture, and no developed central bureaucratic structures, the early medieval state swallowed up a much smaller percentage of GDP than had its Roman predecessor. As far as we can tell, this had nothing to do with right-wing ideologies and everything to do with a basic renegotiation of centre–local relations around the brute fact that landowning elites now owed their ruler actual military service, which put their own very physical bodies on the line. Equally important, all the changes conspired together (although I’m sure none of this was ever planned) to make it much more difficult for early medieval rulers to hold together large geographical areas over the longer term.20
There was also one further, critical difference in the type of economic assets that the ruler of a smaller early medieval state structure had at his disposal. Although late Roman emperors were landowners in their own right, like their Carolingian successors, they drew the majority of their much larger overall income from tax revenues. And tax revenues were entirely renewable. If you spent your entire tax income from year one in year one, you still received exactly the same income in year two, and so on ad infinitum. By contrast, most of the Carolingians’ financial muscle came in the form of a fixed stock of capital assets, whether we’re talking land or the various rights which might generate a revenue stream (to hold a market or court, or make an appointment to a desirable office, etc.). If, to win magnate support, you gave any of these items away in year one, that gift automatically reduced your income in year two. From this followed the fact that the Roman Empire could bumble along for centuries, however messy the succession process (and often it was extremely messy), because regime-building in the Roman world did not strip out the empire’s central asset base. The Carolingians, on the other hand, required more or less untroubled successions to achieve the same outcome. And since genetic accident was only going to give you so many untroubled successions in a row, the central power of the Carolingians – or that of any imperial dynasty in early medieval conditions – was time-limited in a way that its Roman predecessor was not.
And yet, on occasion, you could still build up inter-regional, imperial-type authority in early medieval conditions. The Merovingians achieved it in the sixth century, the Carolingians in the eighth, and the Ottonians would manage the same feat in East Francia in the tenth. The reasons behind these exceptional moments, when the prevailing systemic obstacles to imperial or at least interregional authority were overcome, allow us to generate one final crucial perspective on the ‘small’ state of the early medieval West. This brings both Charlemagne’s achievements and their limitations fully into focus.
THE FIRST REICH
It is possible to tell the story of what came subsequently to be known as the first Reich – though until there was a second Reich after the Franco-Prussian War, it was just ‘the Reich’ – in any number of ways, and in greater or lesser detail. An obvious path into it is the progression of its ruling dynasties and we need a little of that for scene-setting. It was a direct descendant of the East Frankish kingdom of Louis the German, although the Carolingian line was extinguished there in 911 when Arnulf’s son Louis the Child died without issue. At that point the kingdom consisted of four duchies: Franconia, Saxony, Swabia and Bavaria, each with their own ducal lines (Figure 15). Although it was never uncontested, the position of king passed from Louis the Child to Duke Conrad I of Franconia, whose wife was related to Louis’ mother. He had to battle hard to establish his authority, however, and eventually died in 918 of wounds incurred in battle against the Duke of Bavaria: the fabulously named Arnulf the Bad. Conrad had a brother but was himself of the opinion that the man most likely to hold the monarchy together was the Duke of Saxony: Henry the Fowler (my old history teacher liked to indulge in a little bathroom humour at this point, but Henry just liked hunting). Henry first got the Franconians and Saxons to acclaim him king in 919, then constrained the Swabians and Bavarians into line. He also added a fifth duchy, Lorraine, to the kingdom before passing on the crown to his son Otto I. Otto prospered both in East Francia and beyond, even to the extent of conquering most of Italy, and this became the basis for his own imperial coronation on 2 February 962. Otto passed on the imperial mantle to eponymous sons and grandsons: Otto II and Otto III. But, on the latter’s death in 1002, the imperial baton passed to the line of Otto’s younger brother Henry, who had been installed as Duke of Bavaria, although the latest Henry (they were all called Henry to save confusion) was not crowned emperor until 1014. When he died in 1024, the Ottonian dynasty gave out entirely, and power passed to a succession of four king-emperors from the Salian line. Down the years, there followed a bewildering range of dynastic transmissions until Napoleon called time on the whole enterprise after the battle of Austerlitz in 1806, but we don’t have to worry about anything beyond the mid-eleventh century. So the pattern of dynastic transmission within the East Frankish kingdom which became the Holy Roman Empire in 962 went from Carolingians to Conradines, to Ottonians, to Salians. And since the Ottonians clearly styled themselves as the direct descendants of Charlemagne, and their spin doctors invented the concept of translatio imperii – ‘transfer of empire’ – to explain a) that it was the direct continuation of the essence of Charlemagne’s imperial authority, and b) that it was therefore God’s will (where have we met this before?), you could be forgiven for supposing that the Holy Roman Empire of Otto the Great would be the next and final destination for this book.21
But as the same history teacher also liked to say, the Holy Roman Empire was neither Holy nor Roman. Behind the joke, there is an important sense in which these latest emperors, powerful and impressive as they were (and the Salians will play an important role in the final chapter), nonetheless do not stand in the direct line of attempted Roman imperial restorations examined in this book. Theoderic, Justinian and Charlemagne can be thought of as ‘Roman’ in one or more of three separate dimensions of meaning. Their empires can reasonably be called Roman either in terms of the character of the state they created (Theoderic and Justinian), or in the coincidence between the geographical extent of their power and the bulk of the old Western Empire (Theoderic and Charlemagne), or because the state they created was the overwhelming dominant force in the Christian European landscape of its time (Justinian and Charlemagne). Thanks to late Roman precedent and its Byzantine continuation, and as we saw in the course of Charlemagne’s careful manoeuvrings towards it in the 790s, in early medieval Europe the imperial title had come to mean above all else ‘supreme leader of Christendom’.
The Holy Roman Emperors do not quite pass muster as properly Roman on any of these counts. For one thing, the empire, even at its height under Otto I, comprised only East Francia and north and central Italy. The rich and extensive territories of West Francia were not brought under its wing, so that it never rivalled Charlemagne’s creation in size. Moreover, as the empire was born, European Christendom was expanding rapidly. The reconquista was busy returning large parts of the Iberian peninsula to the Christian fold from the mid-eleventh century onwards, and the tenth century had seen the rise of the first Christian states in Poland, Bohemia, Hungary and western Russia (Figure 15). Otto and his successors were simply not in a position to echo the level of overarching dominance within the Christian and/or Roman landscapes achieved by Theoderic, Justinian and Charlemagne.22
But if, for that reason, there is no need to explore the history of the Ottonian Reich in its own right within the confines of this study, one particular feature of what was originally East Francia does demand our attention. What happened in East Francia in the later ninth and early tenth centuries, broadly speaking, was a kind of Carolingian Collapse Lite. The region saw much the same political process in action as the rest of the Carolingian world. Louis the German faced periodic rebellion from multiple offspring, aided and abetted by his half-brother’s interference. It generated, likewise, exactly the same need for all pretenders to build solid bases of support by handing out the required levels of reward to the militarized landowning classes, with similar consequences. The transfer of assets, amongst other things, created the evident wealth and political security within their own domains of the ducal lines that were an established feature of East Frankish politics from the time of Arnulf of Carinthia at the very latest. But although the creation of four powerful dukedoms (or five, counting Lorraine) looks like a perfect setting for ultimate political fragmentation, as happened in West Francia, this was not the outcome. First Conrad, then Henry I and Otto had to battle away at times to prevent it, but that final political fragmentation did not occur in the Frankish east. Why not?
There are several elements to the answer. For one thing, East Francia was slightly less affected by the great Carolingian cull of the later ninth century, and Arnulf of Carinthia deliberately returned to the political rhythms of assembly and contact that had been established by his grandfather Louis the German. So, for the vast majority of the fifty-plus years between the Treaty of Verdun and Arnulf’s death, East Frankish political life had operated in exactly the same way from the same centres of power. This generated a continuity of tradition which played a major role in shaping the mentalities of East Frankish magnates, who more naturally expected to operate together, for certain purposes at least, as a single unit. Such expectations played a major role, for instance, in their largely unanimous willingness to accept Louis the Child as king after Arnulf’s death, at a time when non-Carolingian regional magnates of West Francia were either making themselves kings, or just ignoring royal authority as irrelevant.23
They also had strategic reasons to hang together. The first substantial Slavic-speaking polity to emerge in central Europe was so-called ‘Great’ Moravia, centred on modern Slovakia, where some astounding physical remains of the architecture of its capital at Mikulcice can still be seen. Moravia was a successor state to the Avar Empire which Charlemagne destroyed, and figures substantially in East Frankish annals from the mid-ninth century onwards as the main client state beyond its south-eastern borders. The history the annals tell – like that of any of the frontier clients of the old Roman Empire – saw alternating periods of war and peace, with Moravia being drawn ever closer into the Frankish orbit. Despite its best efforts, it even accepted – eventually – an East Frankish version of Christianity, expelling the Byzantine missionary Methodius who had originally operated there with his brother Cyril. But in the 890s the Moravian kingdom was destroyed by the latest wave of nomadic horsemen to penetrate into central Europe from the great Eurasian steppes: the Magyars.
And I really mean destroyed. All the great architectural structures were abandoned, and Moravia simply disappears from the annals. What’s really significant about the Magyars for present purposes, however, is that – like the Huns and Avars before them (page 183) – their internal structures relied on large-scale downward redistributions of wealth from the leadership to smooth over potentially divisive internal political rivalries. As a result, the arrival of the Magyars could only spell danger for their new neighbours, and the early tenth-century history of Frankish Europe is marked by a new explosion of nomad violence. In 907, the Magyars inflicted a heavy defeat on the army of the Duke of Bavaria at Pressburg, and followed that up in 910 with an equally heavy defeat of the combined East Frankish army of Louis the Child at Augsburg. These victories presaged a decade and more of raiding which periodically disturbed the peace not only (if primarily) of East Francia, but ranged as far afield as Provence and northern Italy. Facing up to the Magyar threat required joint action, therefore, to stand any chance of success and a crucial factor in the rising prestige of the Saxon line was its ability – eventually – to provide effective leadership. As told by its own chroniclers, the story is that a truce Henry negotiated with the Magyars in 926 was critical to his eventual success, even though it cost him annual tribute payments. In the breathing space, a series of settlements were fortified as refuge centres for the rural population, with elements of the latter organized to provide them with effective garrisons, while the main East Frankish army was retrained and re-equipped as heavy armoured cavalry. This construction is surely hiding a degree of humiliation in the arrangements of 926, but the military reorganization was real enough and provided Henry with the wherewithal to inflict a first major defeat on the Magyars at Riade in 933 (which followed the unilateral ending of the tribute payments in 932). When the Magyars came off the reservation for the last time in 955 it further allowed Otto to inflict a massive defeat on them at the battle of the river Lech. This both prompted a major political restructuring among the Magyars, as they were forced to move towards a political economy which was not so dependent upon raiding, and propelled Otto on towards the imperial title.24
But none of this yet takes us quite to the heart of the matter, at least in terms of why Conrad should have thought, back in 918 before the Magyars had been defeated, that the Saxon was the best-placed ducal clan to maintain traditional East Frankish unity. When thinking about evolving patterns of power within any political landscape, it’s usually a good idea to focus on economic resources. Power always has other components besides, but ‘follow the money’ – Deepthroat’s famous advice to Woodward and Bernstein as they unravelled Watergate – usually works in historical analysis too, even if you’re talking not just about cash (although there was some of it), but also land, jobs, and the bundle of varied rights which, as we’ve seen, often stood in for cash in early medieval contexts. Even more transparently in East Francia, since there is little sign of any central bureaucratic apparatus at all, the careful manipulation of these resources to create loyalty within the militarized landowning class underpinned the successful exercise of central power. The rulers of East Francia thus faced the same political problem as their Western peers. How could you generate the requisite loyalty in one generation without depleting the central stock of rewards available for regime-building in the next?
The Ottonian line of Saxon dukes did not manage to reinstitute large-scale taxation of agricultural production (if they had, they really would deserve to rank as a fourth ‘Roman’ imperial moment). But they did get their hands on two other sources of much more renewable wealth than the usual early medieval bundle of land, appointments and rights, which they could use to grease the wheels of magnate loyalty without bankrupting their own position. One was highly specific. From the 920s, the silver ores of the Harz Mountains were exploited with increasing intensity and to the particular profit of the Saxon line. There are no figures for the scale of production, but it funded the Ottonians silver coinage which rapidly became the predominant currency of tenth-century Europe. New mining settlements (of which there were many, all signalled by place names ending in -rode) grew up swiftly in the region, and the result was a flow of renewable income into the Ottonian coffers, much of which could be recycled towards the magnates without undermining the holdings of lands and rights which otherwise made the dynasty politically pre-eminent.25
The dynasty’s other source of renewable wealth derived from regular expansionary warfare. In the Middle Ages in particular, but actually in most contexts, there is a key political distinction to be made between defensive and offensive warfare. The latter is usually much more popular, because the opportunities for economic benefit are often so much larger. Where you are fighting to protect your homelands, there are few – legitimate at least – opportunities for collecting booty, but foreign opposition is always fair game. And if you are, strategically, in a position not just to conduct raiding, but steadily to expand your frontier line, this creates a whole stock of new positions for your magnates to hold, as well as bringing on stream a range of new lands that you can distribute to them without reducing your own stock of assets. In the case of East Francia, it was only the Saxon duchy in c.910 which had access to an open frontier. The other duchies either bordered other parts of Francia itself or the formidable Magyars, whereas the eastern border of Saxony, beyond the Elbe, opened out into Slavic-dominated territories. This West Slavic world was currently right in the middle of a process of demographic, economic and political development which would eventually so expand the borders of European Christendom. In c.900, east of the Elbe, these processes had gone just far enough, however, to provide the armies of Henry I and his successors with a series of decent targets: areas with enough population density and general economic activity to make them worth the effort of conquest. The sources report extremely regular campaigns, and the dukes were quick to recycle the profits of conquest to secure the loyalty of their magnates. By the 950s, extensive marcher lands had been added to the duchy (Figure 15), and, within these regions, new positions and landed estates could be granted without it costing Otto a penny from the royal fisc.26
When you widen the scope to the early medieval West more generally, it becomes inescapable that there was a structural link between access to an open frontier of this kind, and periods when central as opposed to local power was in the ascendant. Not only did the fruits of expansion power the Ottonians to pre-eminence within East Francia, but they were also a crucial element in the rise of the Carolingians. It really is one of the most significant statistics of them all that Carolingian armies were in the field for eighty-five out of the ninety years from the accession of Charles Martel to 803/4. The vast majority of these campaigns were aggressive and expansionary, and the renewable wealth they liberated – in all its forms – made it possible for four generations of the dynasty to build their regimes without eroding the fixed assets of the royal fisc. The linkage works equally well for the Merovingian era – where the central authority of the dynasty was at its height during the generally expansionary sixth century – and even (if you’re interested) to Anglo-Saxon England as well. There, the kingdoms which were originally the most wealthy and powerful were those benefiting from the profits of trading and other links with the continent, above all Kent. Its later fifth- and sixth-century cemeteries throw up a great deal of gold ornamentation and its kings were politically pre-eminent in the early seventh century when Bede’s political narrative begins. As the seventh century progressed, however, political pre-eminence became increasingly the preserve of the ‘outer’ kingdoms of Wessex, Mercia and Northumbria, all of which shared borders with native sub-Romano-British kingdoms and were able to expand at their expense. This gave their kings booty to sell or recycle, not least slaves, and renewable stocks of land which they could use to attract further warriors to their train. Not long after 700, the residual wealth of Kent ceased to be a sufficient basis for political pre-eminence in the face of this military might. Instead, it became itself merely an attractive target for the predatory ambitions of the new powers in the land. In the small-state world of early medieval Europe, expansionary warfare replaced large-scale taxation as the source of renewable wealth that was necessary to maintaining a powerful central authority in anything but the very shortest of terms.
So intimate, indeed, was the link between controlling the fruits of expansion and maintaining central authority, that the inverse correlation works equally well. Whenever a previously powerful central authority ceased to benefit from the rewards of expansion, then its authority quickly eroded, because each subsequent generation was now having to fund regime-building from non-renewable assets. We have already seen this correlation at work with the ninth-century Carolingians, but it applies equally to the seventh-century Merovingians, and in Anglo-Saxon England too. There, once the Picts and northern British had organized themselves sufficiently to fend off further Northumbrian expansion in the later seventh century, the kingdom’s eighth-century rulers paled into relative insignificance in the face of internal political instability. It also applies to the Ottonians. By the 980s, processes of development had gone far enough in the Slavic world to close off the open frontier. Via a mixture of rebellion on the part of previously subdued Slavs and outright resistance from new Slavic forces, where the kingdom of Poland plays a starring role, Ottonian expansion eastwards came to a grinding halt, and the later members of the dynasty and their Salian successors quickly found it much more difficult to exercise central authority than had Henry the Fowler or Otto I.27
All of which prompts one final question: if expansion was so crucial to the longer-term exercise of central authority, filling the massive gap in royal finance created by the end of taxation, why did later Carolingian monarchs allow it to end? Various suggestions have been put forward, including the thought that it was not felt to be legitimate to continue expansion beyond the area of old Merovingian hegemony at its height. But that is very unconvincing, not least because Charlemagne’s rule actually went far beyond it in most directions. It has also been suggested that, after the imperial coronation, old age, the costs of the Saxon wars (which had taken twenty-plus years) and his interest in matters Christian between them deflected the royal eye from the ball of conquest. There is something to these other lines of thought, and the ageing Charlemagne was much more fixed in his new royal palace at Aachen, but they don’t explain why expansion did not begin anew under his successors.
A more profitable route into the problem is to consider expansionary warfare in terms of the cost–benefit equations which governed it. Expansionary warfare would bring in profits, but also involved costs, not just in financial terms (food, weaponry, etc.), but also in personal terms since some of those participating would certainly die. If you think about it in this way, then the ideal profile of an area ripe for expansion is easy enough to construct: it needs to be economically developed enough to offer a satisfying level of reward both in terms of moveable booty and potential land-grabbing, but militarily not so well organized that too many of your expeditionary army, on average, are going to die winning access to the prize. There is also the further point that the costs – emotional as well as financial – increased with distance, since the stresses and strains all become much higher and you are cut off from your home and mainstream concerns for so much longer while on campaign to some distant land.
If you run your mind round the borders of the Carolingian Empire in the early ninth century, it becomes clear, I think, why expansion slowly ground to a halt. In the south, the territories of Muslim Spain were rich and highly desirable, but the peninsula was full of well-organized Islamic polities which fought expansion tooth and nail. Even apart from the famous defeat at Roncesvalles in the Pyrenees, extensive Carolingian efforts succeeded only in expanding the frontier as far south as Barcelona. This slow progress contrasts massively, and revealingly, with, say, the conquest of the Lombards. The best bits of Italy, likewise, had already been swallowed up, and all that was left was either poor and mountainous or else still in those Byzantine hands which were not at all easy to prise open. In the Balkans, after the defeat of the Avars, further progress was blocked by a powerful mixture of the Bulgar and Byzantine empires and, anyway, the central Balkans were not the most attractive source of potential rewards that one might think of. The eastern frontier on the river Elbe, beyond Saxony, remained a possibility, and Louis the German did campaign effectively there, but that northern region was still not so economically developed as it would be a hundred years later when the Ottonians could exploit it to greater effect, while, further south, sufficiently solid Slavic structures were already emerging in Bohemia and Moravia to make expansion difficult. On every corner of the frontier, the cost–benefit equation was starting to deliver a negative answer, either because the enemy was too formidable (Spain), or because the likely benefits were not that great (the Balkans), or some combination of the two (southern Italy and the southern Elbe region).
The strategic situation was generally unripe for further expansion, but how did that generate a decision to stop mounting campaigns? It is inconceivable that Charlemagne or Louis the Pious sat down with a map, not least because they probably didn’t have one, and went through the same kind of analysis that we have just done. The answer, I think, lies in the fact that the wars of expansion were actively fought by the politically important landowning class, who provided the bulk of the army. That the Carolingian express had reached the buffers of expansion would actually have become lived experience for them, taking the highly perceptible form of an increasing number of lives being lost for a decreasing level of benefit. And since these men were politically important, some of them having contact with their rulers in the regular cycle of assemblies, then the likeliest a priori model for how expansion actually came to a halt would be that the build-up of resistance to further campaigning on the ground eventually forced the rulers’ hand. And when you go looking for it, there is just a little evidence that shows precisely this kind of process in action. It was Louis the Pious’ determination to send Hugh of Tours and Matfrid of Orleans, two of Lothar’s chief supporters, off to fight in Spain, for instance, which sparked off enough suspicion and outrage in Lothar to bring him into open revolt. In their view, and Lothar’s too, the costs of being excluded from internal political manoeuvring had become more important than any benefits to be gained from participating in expansion warfare.
And this, to my mind, adds one more important layer to the analysis. The militarized magnates of the Carolingian world were actually working with a pair of simultaneous cost–benefit equations in their heads. On the one hand, they were calculating potential gains and losses from campaigning beyond the frontier, but, simultaneously, they were comparing the kind of answer they were getting from this with a similar calculation about the potential costs and benefits of political intrigue at home. Not only was the campaigning equation delivering a much more negative answer from c.800, but from the 810s, or certainly from the divisio regni of 817 at the latest, its internal political pair was looking much more positive. From then on, while there was currently still just the one ruler, stocks and shares in the futures market in political allegiance were rising steadily with three heirs looking for supporters, a trend which the arrival of Charles the Bald only accelerated. To my mind, therefore, political pressure from the military elite brought expansion to a halt, as it became less rewarding and the potential rewards of selling your loyalty grew internally. And the latter of course both offered rewards in a very local currency which our magnates highly prized, and meant that they would not have to go trekking off either across the Elbe or beyond the Pyrenees. The real story of the end of expansion, therefore, is not so much of the halting of violence altogether as of its refocusing from external to internal enemies, and this only emphasizes how dangerous to the power of the Carolingian monarchy the end of expansion actually was.28
THE CREATION OF EUROPE AND
THE END OF EMPIRE
Thinking about them in the round, the history of these three attempts at Roman imperial restoration prompts, I think, one conclusion above all. By the end of the first millennium, it was no longer possible to create imperial dominion across the European landscape on the same geographical scale that Rome had achieved. Rome had used a demographic and economic base in the Mediterranean to conquer all of western and southern Europe and broadly intimidate much of the north-central zone of secondary development besides. By the year 1000, the rise of Islam had fractured the unity of that power base, decoupling its southern shores from the rest, and turning the rump of the East Roman Empire into medieval Byzantium, as much of a successor state in its own way as any of the Western kingdoms.
But even if the Mediterranean had remained united, the region would no longer have provided sufficient resources to dominate western Eurasia on anything like the same scale. Around the birth of Christ, when Rome exploded to imperial dominion, there had been three distinct zones of development within northern Europe, and all three were operating less intensively in demographic, economic and political terms than the new empire’s Mediterranean power base. By the year 1000, the gaps within northern Europe had narrowed dramatically and the overall strategic edge available in the earlier era to a Mediterranean-based state had disappeared. Population densities in what had been the western La Tene region a thousand years before, and in parts of the old north-central zone – in the heartlands of the Ottonian Reich as far east as the river Elbe at least – had by this point surpassed those of the Mediterranean, as the careers of Charlemagne and Otto I make so very clear. Western and north-central Europe were now the basis of states which could predate upon Italy and Rome, not vice versa. And while still lagging someway ‘behind’ as it were (not that anyone was racing), in the sense that its potential resources were as yet far from so fully mobilized, patterns of life in the third zone of northern and Eastern Europe had also been transformed out of all recognition. What had been the abode of small dispersed populations of simple Iron Age farmers – a world which was so cut off from the rest of Europe in the first half of the millennium that its archaeological remains show not the slightest signs of contact with Rome – now supported much larger populations, an increasingly diverse economy, and some robust political structures of its own.
All this amounted to a complete revolution in prevailing regional balances of strategic power, which was already so far advanced that, even if the north and east still lacked some of the complexity and intensity of agricultural and other economic activity visible in the west, its new dynasties still commanded power bases of sufficient size to fend off encroaching imperial domination from the west. In the later tenth and eleventh centuries, for instance, the Ottonian Reich became embroiled in complex diplomacy with the new state that the Piast dynasty was constructing in neighbouring Poland. Relations oscillated between friendship and competition, and the first Reich was undoubtedly the more powerful of the two. But when diplomacy gave way to conflict, the Reich did not find it possible simply to conquer the Piast state. Even though it was open to cultural influences from the Reich in the form of Christianity, the Polish kingdom was both too far away and too powerful to fall easily under its military subjugation. What we are really seeing by the end of the first millennium is that Europe – defined as a zone of more or less equal societies engaged in complex political, economic, and cultural interactions – was actually coming into existence. The old, massive disparities were a thing of the past, and, as a direct result, it was no longer possible to base a dominant empire on one corner of the European land mass which happened to have undergone a more precocious process of development than the rest.
Somewhat paradoxically, the evidence suggests that the past exercise of imperial dominion had played a fundamental role in stimulating these same extraordinary changes which now rendered empire impossible. What you see, again looking right across the first millennium, is broadly a two-stage process. The economic, social and political structures of the Germanic world went through a similar set of structural transformations on the fringes of the Roman world in its first 500 years, as the Slavic world did subsequently on the fringes of Francia in the 500 years which followed, with some extra imperial stimulation in their case being supplied by contacts with the Avars, Byzantium and, above all, Islam. In older traditions of writing about these kinds of subject, empires were usually portrayed as doing things to, or for, their less developed neighbours, bringing them such fruits of civilization as writing and Christianity. Having seen globalization at work, and moved in time beyond the great era of European imperialism, we are rightly less confident that empires always bring benefits to those with whom they come into contact. But trading with empires, picking up new farming techniques from them, receiving their diplomatic subsidies, copying their weaponry and ideologies, and organizing yourself to fend off the worst excesses of their domination, all pushed forward the sequential emergence of more developed economies and larger state structures in the Germanic and Slavic worlds in the two halves of the first millennium. The language of ‘civilization’ or ‘gift’ really doesn’t capture the process. All empires operate in their peripheries overwhelmingly for their own benefit, and not everyone in the periphery comes out ahead in the long-term transformations which follow. Rather, particular groups in the periphery are able to take advantage of the opportunities opened up by the range of new contacts with an imperial neighbour, and this is precisely what we now call globalization. But out of that particular dance there does eventually emerge much greater overall economic strength in the periphery, bigger populations, and more robust political structures centred on those who benefited most. Over the long term – and we are talking centuries of contact – Roman and Frankish imperialism thus played a central if indirect role in constructing the peripheral powers which would eventually blunt their imperial spears. Instead of seeing peripheral populations as passive recipients of imperial civilization, the story is one of particular groups in the periphery exercising agency to profit from a new situation.29 The aggregate strategic effect could not have been more profound.
As a direct result, a restored empire that captured the essence of the Roman original had become completely impossible by the year 1000. Not only had Islam broken apart ancient Mediterranean unity, and the balance of power in Western Europe shifted decisively north of the Alps, but, still more fundamentally, patterns of development were now much too equal across the broader European landscape. Thanks to this equalization of development, you might say, the scene was set for the thousand subsequent years of fruitless warfare which followed as Europe’s dynasts intermittently struggled to achieve a level of overarching dominance that was in fact impossible. In that sense, it took the nightmare of two world wars in the twentieth century before the European Dream was finally called into existence to try to put a stop to the process of endless armed competition between powers that were always too equal for there to be an outright winner.
But if a traditional empire of military and economic dominance on the old Roman scale had become impossible by the year 1000, a new kind of empire was about to be reborn from the ashes of the Carolingian and Ottonian imperial projects. Emperors of the Roman type operating on a truly Roman scale were a thing of the past, and their centre could certainly no longer hold; yet a thousand years after the birth of Christ, a rough imperial beast of an entirely new kind was slouching forward to be born, this time not towards Bethlehem as Yeats imagined it, but towards the old imperial capital itself.
You can support our site by clicking on this link and watching the advertisement.