The medieval as franchised protection

By Lorenzo

In my post on the French Revolution as Chinese dynastic cycle, I denied that China as had a medieval period as such. Accepting the Naito Hypothesis that Song dynasty China (960-1279) was the first modern society, I hold that China went from its late antiquity to the early modern without a medieval period. This post sets out how I define the medieval and why I hold China did not have a full medieval period.

What is medieval?

What is meant by the notion of a medieval period? Is it merely the period between a classical period and the modern? The middle not-classical, not-modern period? If that is so, then every society which had a classical period and made it to the modern without moving directly from one to the other has a medieval period.

So, if we hold the period of Classical China to end with the collapse of the Han dynasty in 220 and Song China (960-1279)–with its paper money, meritocratic bureaucratised autocracy, tax-paid soldiers, public rituals but private religion, scholar-gentry replacing vanishing landed aristocracy, competitions for better technology-was the first modern society, then the medieval period in China extends from 220 to 960.

We are still left with defining the classical (not so hard) and defining the modern (rather more fraught) but the medieval is defined simply as a by-product of defining the classical and the modern.

C6th Eastern Roman ivory given to Frankish queen

C6th Eastern Roman (“Byzantine”) ivory given to a Frankish queen

Acceptance of the notion of Late Antiquity–a period neither classical nor medieval but of transition between, and which tends to replace the previous term “Dark Ages”–is an implicit rejection of this negative definition of the medieval. The underlying notion seems to be that there had not been enough of an institutional break to make the period of Late Antiquity medieval. In particular, there are still centralised state(s) raising armies of tax-paid soldiers even though the rising manorialism and the shift to cavalry-based armies are elements that will become much more dominant in the medieval period.

Once we have weak states–early on, not much more than chiefdoms–relying on beholden warriors for their military forces, then we are more fully transitioning into the medieval. Apparently arriving at the medieval just as those beholden warriors become horse-mounted and typically individually financed by specific localities, or beholden to someone who is.

Note that this is a step that so-called “medieval” China never took. Yes, there was manorialism and yes there was a land-owning military aristocracy. But they are typically officers in mass infantry armies or formal cavalry units. They are no more “medieval” than the Junkers of Prussia, the noble-gentry of Poland or the C18th and C19th dvoriane service nobility of Russia. So-called “medieval” China looks much more like late antiquity Mediterranean or early-modern Eastern Europe than it looks like medieval Europe.

Or like medieval Japan, medieval Islam, medieval Northern India or the Iranian plateau before the Muslim conquest. Which were all–with the partial exception of Japan–times and places where the armoured, mounted warrior was the dominant military and social element.  And Japan is an exception only in the sense that the samurai were not specifically mounted. Japan is also a very mountainous archipelago–not an ideal place for raising and using cavalry armies.


The importance of the beholden warriors of medieval Europe being a horseman can be seen in the languages of Europe. The terms for knight mean either rider in Germanic languages:

ritter (German) ridder (Dutch), ridder (Danish), riddare (Swedish).

Or derive from the Vulgar Latin caballarius (horseman–in English, cavalier) in Romance languages:

caballero (Spanish), cavaliere (Italian), chevalier (French), cavaleiro (Portugese), cavaler (Romanian).

Mounted beholden warriors defeating infantry beholden warriors.

Mounted beholden warriors defeating infantry beholden warriors.

Foreign observers also noticed the importance of the horsemen. The great French medievalist Marc Bloch quotes the “Arab emir Ousama” (Usama ibn Munqidh) as noting of the Crusader kingdoms:

… all pre-eminence belongs to the horsemen. They are in truth the only men who count. Theirs it is to give counsel; theirs to render justice (p.291).

The English word knight actually comes from the Old English cniht (boy or servant) and is connected to the German knecht (servant, bondsmen, vassal). By 1100 (i.e. after the Norman Conquest) it had come to mean “military follower of king or other superior” (i.e. retainer).

In Japan, the bushi or warriors came to be referred to as samurai which derives from saburau meaning “those who serve close attendance to the nobility” (i.e., retainer) with bushi and samurai being effectively synonymous by the end of the C12th (early in Japan’s medieval period). So, the terms for armoured beholden warrior have the same derivations in meaning in the two archipelago medieval societies.

Expensive warriors

It is useful to make a distinction between warriors and soldiers:

  • The archetypal warrior pays for his (it is usually his) own equipment, started training in arms and horsemanship very young, his allegiance is personal with an honour ethic because personal reputation is central.
  • The archetypal soldier enters military service as an adult, has his equipment and training paid for, his allegiance is formal-institutional rather than personal with a duty ethic for fitting into a unit within wider structures.

There are gradations between these archetypes that can be thought of as the ends of a spectrum. So, Roman legionaries were soldiers; knights, samurai, rajputssipahi, and so on were warriors.

Does not come cheap.

Does not come cheap.

A soldier is an armed employee. To have soldiers, there has to be a sufficiently centralised tax base and administrative structure able to support paying, training and equipping soldiers.

Warriors need income sources, but these can be local and personal, though there were tax-paid warriors—as, for example, under the Umayyad Caliphate (661-750). And a warrior could have soldiers in his service, given sufficient income.

The training and equipping of a soldier has to be paid for upfront. That of a warrior is typically paid for by his family, as a cross-generational transfer (each generation is trained by the previous and trains the next).

The training and equipping of an armoured, mounted warrior is very expensive. So an arrangement which spread the cost over the life-time of the warrior was clearly desirable. Moreover, an armoured–especially mounted–warrior can dominate a locality quite effectively. So, beholden warriors who extracted income directly from a locality represented considerable saving in administrative and management costs and upfront equipping and training costs.

To put it another way, they acted as franchised protection. Rather than paying for a tax-extracting structure, a training structure, an equipping structure and a salary paying-structure, it was all bundled together in some sort of fief arrangement. The technology of the armoured, especially mounted, warrior encouraged a fief arrangement (using the term ‘fief’ quite broadly to cover any sort of local income for military service). Especially if administrative resources (such as literacy) were scarce.

The model soldier.

The model soldier.

Mass infantry armies, on the other hand, have economies of scale in equipping and training and are not particularly good at dominating specific localities.  So they will either be farmer-militias (such as the citizen-soldiers of the Greek polis or early Republican Rome, the fubing system of Sui and early Tang China or the fyrd of Anglo-Saxon England) or tax-paid employees (as in the Late Republican and Imperial Roman, Chinese armies from the later Tang dynasty onwards and modern armies).

The forms of the fief varied considerably. In the Middle East, where centralised control was at a premium due to the perennial manage-the-pastoralist-nomads problem, the iqta tax-fief evolved.  In Northern India, where central control was also at a premium but brahmins tended to dominate village justice yet there was significant religious diversity, the jagir tax-and-justice fief evolved. In Europe, where central control was difficult and social stability was at a premium, the full land-fief evolved. Japan developed a complicated system of all sorts of claims on the income from land on an income-for-service basis before eventually evolving a much simpler system of provincial overlords with dependent samurai.

The land-holding-provides-mounted-armoured-warrior system originally evolved on the Iranian plateau. Not surprising, since it was Iranian peoples who evolved the mounted armoured lancer in the first place.

So, a medieval society in the full sense has beholden warrior societies providing franchised protection. In other words, societies with fief-warriors. Which raises the dread term “feudalism“. Between the (unhelpful) Marxist usage (which divorces the concept from its original etymology and its military focus), the Elizabeth Brown-Susan Reynolds critique and the serious differences in the form of fiefs across the above societies, the term probably has become too fraught to be useful.

Swearing fealty (or doing homage) as public acknowledgement.

Swearing fealty (or doing homage) as public acknowledgement.

On the Reynolds’ critique, her point that feudal overlordship was territorially defined, so there was no choice of feudal overlord, is much less powerful than she seems to think.  Taking the legal definition of a contract as a basis:

The requisite elements that must be established to demonstrate the formation of a legally binding contract are (1) offer; (2) acceptance; (3) consideration; (4) mutuality of obligation; (5) competency and capacity; and, in certain circumstances, (6) a written instrument.

Then the question becomes acceptance of certain structuring of the relationship, not whether there was another overlord on offer. If the subordinate is providing neither taxes nor service, why should the superior respect his (or her) claims to the land? If the superior is not going to recognise the property rights of the servitor, why should they provide service? Accepting of service for recognition of property rights means that there is offer and acceptance, consideration, mutuality, competence and capacity wrapped up in a publicly acknowledged package. It may be the only deal on offer, but it is still a deal, a way of structuring their relationship. One which minimises the administrative, management and upfront equipping and training costs of the overlord.

Back to China

China from the collapse of the Han dynasty in 220 to the rise of the Sui Dynasty in 581 had manorialism, warlordism, break-up of empire, trade regression, landed aristocracy. As did the period from the Late Tang dynasty to the rise of the Song Dynasty in 960. So, the post-Han pre-Song period (220-960) had many of the features of the medieval: but, then, so did the Late Western Roman Empire. Post-Han pre-Song China, like the Late Empire, had the germs of a medieval order. Unlike Europe after the collapse of the Western Roman Empire, it did not quite make it into a medieval order.

And the reason China did not quite make it was it never took the final step of the locality-financed mounted armoured warrior; of franchised protection operating at village level. Chinese armies remained predominantly infantry armies and the balance between administrative capacity and training and equipping costs of infantry never tipped over into that final step.

Going for mass effect.

Going for mass effect.

The interaction with the nomads was probably crucial here, as in so much of China’s history (pdf). Mass infantry was the Chinese advantage over nomad mobility, particularly when armed with massed crossbows. The selenium deficiency of Chinese soil made raising horses effectively impossible in much of China and difficult and expensive elsewhere. Basing status and social order on a resource one did not sufficiently control (horses) was not sensible. Especially given the difficulty in breeding up heavy lancer horses. Hence the landowning military aristocracies of China before their eclipse under the Song dynasty were much more like the Junkers of early modern Prussia–estate-holding officers leading predominantly infantry forces–than the barons of medieval Europe–castle-holding armoured mounted warriors leading other armoured mounted warriors beholden to them.

China moving to the Japanese path of beholden armoured warriors not necessarily mounted did not make sense because mass was the answer to nomad mobility, particularly in the flat terrain of Northern China. Mass infantry augmented by massive wall-building.

Moreover, China, like its rival nomad confederations (pdf), was able to generate enough internal trade to create a revenue source whose economies of scale and scope were able to overcome perennial diseconomies of scale in control over territory. Especially as Chinese administration relied on high levels of village self-management.

The Chinese state(s) had neither the revenue/management cost pressure nor the training and equipping cost pressure to shift to locally-financed beholden armoured warriors. So it didn’t, thereby avoiding having a medieval period in the sense that Europe, Japan, Islam, Northern India and the Iranian plateau all evolved into medieval societies and so had “real” medieval periods.

ADDENDA: I would now differentiate more clearly between a medieval period and a medieval society. China can be held to have had a medieval period, it was never a medieval society.


  1. Mark
    Posted August 9, 2013 at 12:13 pm | Permalink

    HI Lorenzo,
    As an economic historian I am really enjoying these recent posts. An intermediate case you might consider is the Byzantine Empire.

    The Eastern Roman empire relied on a professional military system. But from the 7th century onwards a kind of fief system emerged in the form of the themes for similar reasons to for those you suggest in your post.

    However, this did not develop into feudalism. The Byzantine army became more professionalized again in the 10th century and like China it is arguable that the Byzantine empire did not become fully medieval (at least until it was captured by the Venetians in 1204).

  2. Posted August 9, 2013 at 1:57 pm | Permalink

    [email protected] Glad you have been enjoying them. As for:

    An intermediate case you might consider is the Byzantine Empire.

    I gave a paper to the HTAV (History Teachers of Australia-Victoria) conference in July on what is meant by ‘medieval’ , with an Appendix entitled Late Sassanid and Thematic Roman hybrid systems. So, I agree! 🙂

    In fact, I might just post it.

  3. kvd
    Posted August 9, 2013 at 3:50 pm | Permalink

    the medieval period in China extends from 220 to 960, so I looked up the population at those rough boundaries, and came up with approximately 40M, going to 60M. For whole-of-Europe (with definition, somewhere, of 476 to 1453 AD), the population figures are estimated as 27M rising to 50M – but if you stick with Greece thru Italy, Spain, France, Germany/Scandinavia and British Isles, the numbers are 22M rising to 42M .

    Comparing the respective land areas, you get roughly 10M square kilometers for both (give or take) but the over-riding geography is markedly different – China (an exaggeration) ‘monolithic’, Europe not at all – but more importantly, coastlines (China 10,000k, Europe 80-90,000K) which seems to be remarked upon as an important factor for both trade and outwards exploration.

    Which is all background for a couple of questions on your very interesting post:

    1) Do you think that the beholden warrior societies providing franchised protection necessarily ‘scales up’ – given that the end European population was only just then reaching the starting point of the Chinese population?

    2) Where do you place the quite marked ‘geographical and sea-access disparity’ between the two areas in order of importance, compared to other factors?

    Nothing about horses, although I do like them, and thanks for a very interesting post!

  4. Posted August 10, 2013 at 8:29 am | Permalink

    [email protected] In answer to the “scaling up” issue, I see no reason why not. The Mughal Empire used income-and-service bundling, over a comparable area and comparable population. Bureaucracy and tax-paid soldiers involve more people, rather than less.

    That Europe was a peninsula of peninsulas is very important in its history. I don’t think it was an accident that the society which evolved institutions most like Europe’s was also a mountainous archipelago.

    It was simply geographically easier to unify China though I agree with Peter Turchin and others, the nomads were the most important reason for the greater rate of Chinese unity. So, the geography made unity much more possible, the nomads made it much more common.

  5. kvd
    Posted August 11, 2013 at 7:31 am | Permalink

    Lorenzo, slightly off-topic, I found this paper worth a read, particularly for the ‘factors’ as summarised on the last page. Makes me wish I had the ability to suss out what the resulting conclusions were.

  6. Posted August 12, 2013 at 8:31 am | Permalink

    [email protected] You could just buy the book 🙂

  7. kvd
    Posted August 13, 2013 at 6:07 am | Permalink

    Thank you Lorenzo – I shall. Hope its got lots of graphs.

  8. Posted August 13, 2013 at 7:35 am | Permalink

    [email protected] The graph to beat all graphs!

    Reminds me of the classic graph of Napoleon’s Russian campaign, a pioneering graphical presentation of data.

Post a Comment

Your email is never published nor shared. Required fields are marked *