Reconsidering Demographics in Pendragon

vegas · September 17, 2023

Preamble: For many gaming groups, demographics is not an important topic at all, because the focus stays zoomed-in on a select group of knights. But for some of us that want to zoom-out, either for world building or to include a strategic view and even let their players participate in Britain-level war and politics, having a model of the demographics of the setting is nice. This post is for those people.

Executive summary: The demographic data stated and implied in 5e sources don't all fit together. Morien suggests we should assume a much larger population of Britain versus Book of Uther. I would suggest this creates new problems and severely over-shoots at fixing the discrepancies. Instead, real-world demographics and history suggest we should keep the population of Logres as indicated in BoU and reduce the average size of households in manor populations to reconcile the data.

Introduction: Elsewhere we read...

On 8/10/2023 at 6:46 AM, Morien said:

The other point is looking at the total population of Logres (which should be closer to 1+ million, given the number of knights and demographics, the half a million would work for adults, but you need the kids, too), we can get a rough estimate of about the same ~400-500 people per knight.

and...

On 8/10/2023 at 10:33 AM, Morien said:

KAP 5.2, p. 77: "A village of about a hundred households is part of your manor." Household ~ 5 people, so about 500 people.

BoUther, p. 7: "Logres has about 500,000 total people". Hence my correction that this should be 'adult people', and the need to add ~ 750 000 kids for a total population of over 1 million. But 500 000 adults is roughly 250 000 household, so when divided by 2600 or so knights, you get about that 100 households from KAP 5.

Morien correctly points out all the demographic data given don't make sense. It is easy to see the problem: if Logres has 2600 knights and 475,000 commoners (95% of the population) that works out to ~183 commoners per knight, which doesn't easily reconcile with the number of commoners on a manor. Morien's proposal is to grow the population of Logres by 2.5x to 1.25 million to get the ratio of commoners to knights to 480 per knight which fits closer to the target.

First a minor correction: KAP 5.2 p 43 tells us there are 420 commoners per manor, not 500, though the 500 figure is used elsewhere in KAP and 420 is within his range. Nevertheless the point stands that it is a long ways from 183 commoners per knight in BoU.

One issue with this approach is that the language in BoU pretty clearly is meant to be the total population, not just adults, but hey, something has to give. Morien's solution, however, introduces new "problems". First, it makes the population unhistorically large. If there are 1.25 million in Logres, then there are about 2.5 million on Britain (Logres is half the population of Britain, 5.2e p 18) and add in Ireland and you have a population over 3 million on the British Isles. The Uther years are supposed to represent the dark ages and Anarchy the 1000s (GPC pp 25 & 70) and historically the British Isles had a population somewhere between 0.5 and 2 million people in that time period [1]. Second, increasing the population to 1.25 million in Logres implies a very small militarized force. With one knight there are another 3 foot giving a total force under arms of 2610 * 4 = 10,440 or just 0.8% of the population in times of internal wars and foreign invasion. By contrast if we leave the Logres population at 0.5 million, we don't have either of these problem: the total population of the British Isles comes out to ~1 to 1.5 million (reasonable) and the total men under arms in Logres comes in at 2% (reasonable for the setting).

Alternative Solution: Instead of increasing the total population, I would propose decreasing the number of commoners per household in manors from 4.2 down to 2.85 during the Uther and Anarchy period. For a population with a life expectancy at birth of 35 years (historically reasonable and implied in the BoE survival tables) you would expect to see about 0.67 children per adult in a steady-state population.[2] (The population of the British Isles doubles once every 250 years in the dark ages; that <0.3% annual growth rate is practically steady-state.)[1] In addition, We can expect 30% of rural households will have a single-head. This is observed in historically[3] and is suggested by the high adult death rate (for example, about 1/3 of couples will have a partner die before their children reach 21.) Put these facts together and you get 285 adults on a manor (30 single-head of households, 70 married-couples, for 170 adults plus 0.67 children per adult for 114 children ~ 285 total population for 100 households). Multi-generational households obviously could increase the average commoners per household versus this simple model, but apparently this was not the norm, at least in later periods where data permits evaluation.[4] Studies from later periods also estimate rural households sizes directly at about 3.5, but population growth rates are much higher in these periods, making these estimates less applicable to our setting.[3] In any event 4.2 heads per household is not really supportable for this time period, while ~2.85 and certainly less than 3.5 seem more reasonable estimates.

Of course 285 commoners per manor still doesn't get us all the way down to the 183 commoners per knight. We can close the gap if we remember that only 20% (or less, but I will use 20% for simplicity) of knights are actually landed. In those cases they need the full 285 commoners to fund a knight, his court and 3 foot. But 80% of knights are household knights and need only 5.5L (4L for the knight and squire, 1.5L for the associated foot) meaning it only takes 157 commoners to fund a knight and one manor supports 1.8 knights. Take the weighted average of these cases and you get... 183 commoners per knight, exactly the target. (20% * 285 + 80% * 157 = 183).[5][6]

Now Morien seems to disagree and writes we need 1.5 children per adult (from the quote above) and that we need more monks and clergy too (quote below):

On 11/10/2021 at 4:32 AM, Morien said:

* Although the population figure in BoU seems to be just for the adults, as otherwise there won't be enough children to replace the knights in the next generation. Doubling the number of people in all categories to account for the kids would take care of that, and also allow for more monks and village priests.

Demographics and history don't support these assertions. A steady-state population of 15,000 nobles will have 9,000 adults and 6,000 children (that is, people under 21 are 40% of the population.)[2] The 9000 adults yield 4500 adult males (medieval sex ratios were actually tilted heavily toward men - let us ignore that detail) from which we need 2600 knights. 60% of adult noble males becoming knights is large, but it is not the case "there won't be enough children to replace the knights" as the population is in steady-state with this age distribution. This is not exactly consistent with BoU which claims 15,000 nobles would have 6,000 adult males (and by extension <6000 females and 3000+ children), but such a distribution would not be self-replacing. Hence it is not the 15,000 which is the problem, but distribution claimed in BoU that needs a small tweak from 6000 to ~4500 adult males to work demographically.

As for needing more monks and priests, at 10,000 people, that is over 18 clergy per Hundred in Logres. Of course they are going to be clustered together in monasteries and cathedrals, but it is not at all obvious that we need to "allow for more" when historically several rural villages would share a single priest among them. Remember too that the clergy have a much smaller portion of people under 21 as a person is not part of the clergy until taking vows at age ~15, and they draw their numbers from commoners and nobles. It seems there are plenty of holy men for our setting without growing the population by another 2.5x

[1] Josiah C. Russell, "Population in Europe:, in Carlo M. Cipolla, ed., The Fontana Economic History of Europe, Vol. I: The Middle Ages, (Glasgow : Collins/Fontana, 1972), 25-71.

[2] United Nations Model Life Tables, https://www.un.org/development/desa/pd/data/model-life-tables. 0.67 children per adult can be observed from life expectancy tables for E(0)=35.

[3] Maryanne Kowaleski, "Medieval People in Town and Country: New Perspectives from Demography and Bioarchaeology", Speculum, Volume 89, Number 3, July 2014.

[4] Stone, Lawrence, The family, sex and marriage in England, 1500-1800, p 92, 1979.

[5] It is something of a happy accident that the figure came out exactly to 183 commoners per knight. In the aggregate, this is not really correct, as I am just following Morien's procedure of focusing on the rural manor population, but we are ignoring that there is another 10% of commoners in market towns & cities, that some manors make 10% and 20% premiums over the standard manor, and there are sources of income besides manor income detailed in BoU - forests, natural resources, court fees, etc. - to fund the army. There is a lot of detail in BoU yet not enough to reconcile all the figures. We have to content ourselves with some back-of-the-envelope calculations to check the reasonableness, which in my opinion is perfectly adequate for a game about knights in a fictional setting.

[6] If we apply the same 20/80 ratio of vassal to household knights to Morien's we end up with .2 * 420 + .8 * 231 = 269 commoners per knight, while his larger population now has 455 commoners per knight. He over-shoots the need by 70%.

Morien · September 17, 2023

6 hours ago, vegas said:

Preamble: For many gaming groups, demographics is not an important topic at all, because the focus stays zoomed-in on a select group of knights. But for some of us that want to zoom-out, either for world building or to include a strategic view and even let their players participate in Britain-level war and politics, having a model of the demographics of the setting is nice. This post is for those people.

Executive summary: The demographic data stated and implied in 5e sources don't all fit together. Morien suggests we should assume a much larger population of Britain versus Book of Uther. I would suggest this creates new problems and severely over-shoots at fixing the discrepancies. Instead, real-world demographics and history suggest we should keep the population of Logres as indicated in BoU and reduce the average size of households in manor populations to reconcile the data.

Nice analysis. My original idea was to simply reduce the households to 40-50 or so instead of 100 peasant households per manor. However, I think this caused some other issues, not the least leading to a very top-heavy societal pyramid. Yours is making the peasant household smaller, almost halving them, if I followed your arguments.

Some quick comments below. Keeping in mind that I don't claim to have even a Masters in demographics or even history. Just a hobbyist's interest, and liking to tinker with the numbers for them to make sense to me.

6 hours ago, vegas said:

The Uther years are supposed to represent the dark ages and Anarchy the 1000s (GPC pp 25 & 70) and historically the British Isles had a population somewhere between 0.5 and 2 million people in that time period [1].

You cannot use the GPC's representation and overwrite BoU assumptions where that population number is given. The society in BoU is clearly early-to-mid-1100s, not Migration Era - Dark Ages. If you take the population estimates derived from the Domesday Book (so for 1086), you can get values up to 1.9 million for England (which, admittedly, is bigger than Logres, so you'd need to scale things down a bit) (Source: https://warwick.ac.uk/fac/soc/economics/seminars/seminars/conferences/venice3/programme/english_medieval_population.pdf p 16, Table 1, quoting Harvey 1988, pp. 48-49.). Taking the annual increase of 0.58% between 1086 and 1190 (the above reference, p. 18, Table 4), and putting us, say, 40 years forward to 1126, we'd expect another increase of ~26%. So you can pretty much half the population and still get ~1.2 million.

So I don't think 1.25 million or so is beyond the pale for ~1100s equivalent Logres.

6 hours ago, vegas said:

Second, increasing the population to 1.25 million in Logres implies a very small militarized force. With one knight there are another 3 foot giving a total force under arms of 2610 * 4 = 10,440 or just 0.8% of the population in times of internal wars and foreign invasion. By contrast if we leave the Logres population at 0.5 million, we don't have either of these problem: the total population of the British Isles comes out to ~1 to 1.5 million (reasonable) and the total men under arms in Logres comes in at 2% (reasonable for the setting).

This I would agree with, although it does ignore a few things.

One is the peasant levy, which admittedly has been steadily de-emphasized, but there is also a mention of a county militia in BoU (e.g., p. 11), into which all free men belong. So in addition to the knights and footsoldiers, there is an additional pool of armed men in the case of emergencies, especially defensive ones. Now, how many men would belong into this militia? If we start from 100 households, and assume even 10% of free men households, this is an extra 1% to add to the 0.8% you calculated above for a total of 1.8%, not that far from 2% you got. Granted, it is a militia, not a standing army, but I would imagine it getting called up pretty often, especially if there are local raids.

The other thing ignored is that the 1 knight and 3 footsoldiers is the norm, not necessarily the absolute dictum. It is possible that some barons would be bringing a different ratio, or the King himself might keep a larger core of infantry. Each knight removed adds 8 footsoldiers, almost tripling the numbers per manor. Although not quite, since you have ignored the squire, who would count as a combatant as per BoB2 rules, and hence you'd have to multiply by 5 (or 11 for a full infantry manor), not 4. So even the starting number should be 1%.

Finally, there are also mercenaries, who could be hired to bolster the armies when needed, bringing the numbers up, although not by a full 100%, I wouldn't think.

Point is, there are other inputs and variables that can be tweaked to get closer to the 2%, if that is the value you seek, not just the population.

6 hours ago, vegas said:

In any event 4.2 heads per household is not really supportable for this time period, while ~2.85 and certainly less than 3.5 seem more reasonable estimates.

This seems like a very low number compared to the rule-of-thumb 5 people per household that I have seen in general use (such as in the Table 1 of the reference I gave above). Obviously, whether you use ~3 or ~5 has a big impact on the final population numbers, assuming that the number of households stays the same.

6 hours ago, vegas said:

Of course 285 commoners per manor still doesn't get us all the way down to the 183 commoners per knight. We can close the gap if we remember that only 20% (or less, but I will use 20% for simplicity) of knights are actually landed. In those cases they need the full 285 commoners to fund a knight, his court and 3 foot. But 80% of knights are household knights and need only 5.5L (4L for the knight and squire, 1.5L for the associated foot) meaning it only takes 157 commoners to fund a knight and one manor supports 1.8 knights. Take the weighted average of these cases and you get... 183 commoners per knight, exactly the target. (20% * 285 + 80% * 157 = 183).[5][6]

Ah, here you misunderstand the BotE/W economics.

The household knights do not spring up from smaller manors, but the same size as the vassal knights. The only difference is what happens with the excess income: does it go to supporting the vassal knight's family and 'court', or to the higher noble's family and court? Thus, you cannot scale the manor size (assuming that population is directly proportional to the income anyway, which is an assumption we both implicitly or explicitly make) down to 55% for the household knights.

Same problem in your footnote [6], you cannot take 231 as the population size for household knight manors when using 420 for the vassal knights.

In short, taking your numbers as gospel, 285 commoners per manor (assuming one manor per knight), it would still result in ~750 000 commoners and a total population of about 765 000, using BoU numbers for nobility and clergy. Or we could scale those up to keep the percentages the same, in which case we'd add about 15 000 clergy and 22 500 nobles (giving a bit more space for kids and esquires), for a total population of almost 800 000, or +60% compared to BoU. Granted, it is still much less than my suggested +100% or +150% (for 1 million and 1.25 million respectively).

6 hours ago, vegas said:

Now Morien seems to disagree and writes we need 1.5 children per adult (from the quote above) and that we need more monks and clergy too (quote below):

Demographics and history don't support these assertions. A steady-state population of 15,000 nobles will have 9,000 adults and 6,000 children (that is, people under 21 are 40% of the population.)[2] The 9000 adults yield 4500 adult males (medieval sex ratios were actually tilted heavily toward men - let us ignore that detail) from which we need 2600 knights. 60% of adult noble males becoming knights is large, but it is not the case "there won't be enough children to replace the knights" as the population is in steady-state with this age distribution. This is not exactly consistent with BoU which claims 15,000 nobles would have 6,000 adult males (and by extension <6000 females and 3000+ children), but such a distribution would not be self-replacing. Hence it is not the 15,000 which is the problem, but distribution claimed in BoU that needs a small tweak from 6000 to ~4500 adult males to work demographically.

The population increase between 1086 - 1190 was about 0.58% (the above reference, p. 18, Table 4), and as I have said, this is a much better model for BoU and later Arthurian Periods rather than the Dark Ages model. Thus, you should have more kids than the steady state model you propose, and this would imply more (currently living which is an important distinction) minors than the 40% you propose. In the above reference (p. 16, Table 2), the estimate for kids below 15 is ~40%. I don't think the years 15-21 would push it to 60%, but you'd probably get closer to 50%. Given that some people marry in their teens already, this is probably OK as far as population growth is concerned. So the 'average' peasant household would probably be something like: husband, wife, 1 older teen or elder, and 2 kids (0-14). Depending a bit how you wish to juggle the number, you could claim that as a 50% split or even 60/40 if you put kids to pre-15 bin (or even pre-14, to keep it to the start of the 'apprenticeship' age at 14).

60% of adult noble males being knights is quite high, though. A quick back of the envelope calculation says that if a 21 year old knight lives to 42, he should have had 3 squires during that time. If he dies, he is replaced by a (younger) squire, so you would expect something like 1 knight and 2 esquires in the adult noble males, or about 33% knights. Obviously, that is a very rough estimate, but I think it is a better one than 60%.

Edited September 17, 2023 by Morien

Morien · September 17, 2023

I suspect that my 3 kids per 2 adults was coming from three different places.

One, to keep things simple and assume 5 people per household = mom, dad and 3 kids.

Second, a simplification of assuming that the generation of adults is the same as the generation of kids, i.e. the adults would mostly have died off by their mid-40s.

And thirdly, I think I got confused between replacement rate and the number of currently living kids. Given about one third mortality rate for kids, you'd need at least 3 kids per couple to be born, so that the next adult generation would be 2 adults again. Obviously (in hindsight) this is not the same as having 1.5 currently living kids for every adult. Although I admit I am surprised by the 0.67 number, which seems very low to me. I am probably discounting the older adults?

Alright, I just ran my code for the BotE family survival rates (life expectancy at birth ~ 34 years, the revised table in BotE v.1.3.2), and I get about 48% of the population being younger than 21. So this would imply roughly 1 kid per adult. When I run it for kids younger than 14, I get 33.6%, so close to 1/3rd as to not matter. Which is actually pretty close to the lower estimates for the younger kids in my reference, which is pleasing.

So, roughly, from the BotE family survival rates:
2/6 kids and younger teens (0-13)
1/6 older teens (14-20)
2/6 adults (21-45)
1/6 elders (46+)

(Which, by the way, if you ignore the elders and count the older teens as kids, gets back to my original assumption of 2 adults and 3 kids. But yeah, I think my original thinking there was overly simplistic anyway.)

Obviously the death rates and such in BotE are for game purposes rather than coming from actual real world data, even though it tries to get it at least roughly right.

EDIT: I was a bit confused between the versions of BotE, but the above is right for v.1.3.2. The earlier version (1.1, I think) had a hellaciously high child mortality, leading to a very wide pyramid as the kids kept dying off.

The actual (rounded) percentages:
Kids and younger teens (0-13): 34%
Older teens (14-20): 13%
Adults (21-45): 38%
Elders (46+): 16%

Edited September 17, 2023 by Morien

mj6373 · September 17, 2023

Not sure how worth mentioning this is, but even if you get the commoner population right, Pendragon's logistics regarding knight upkeep (and the implications for sustainable proportions) aren't going to match real life because Pendragon glosses over a lot of the practical realities that made the feudal system arrange itself the way it was.

Specifically, manor-holding knights were closer to norm than exception. Pendragon treats the infeudation and subinfeudation of land as practically an optional thing that lords just do to reward their greatest supporters, and in fact that they'll be penalized for doing very much of, which couldn't be further from the truth. Medieval transportation and communication logistics make it much more efficient to have knights independently managing a manor or few and using the income to provide their own military upkeep locally to bring to bear for yearly military service than it is to make all your distant holdings bring you everything they can carry in taxes and then personally funding your whole military commitment from there. A lord would have some household knights on the payroll, of course, particularly those they can afford directly from the income of their primary/central estate that doesn't have to travel far, because having full time knightly "employees" around is useful, keeps your estate safer, communicates strength and largesse, etc. But pushing beyond that, using income from beyond your local range to keep more household knights and footmen on your personal payroll, makes your ratio of "commoner labor|military force" worse, not better.

That's not to say lords were keeping as little land beyond their central estate as demesne as possible, because even though it *does* lower the total military output capacity of your holdings, you *need* enough strength not tied to your vassals that you aren't easy for a strong vassal to rebel against. But the idea that a lord would keep upwards of 80% of their land as demesne, and get Honor hits for going below that because it makes them seem *weak*? Very silly.

The whole "80+% of knights are household knights" idea doesn't even make sense for the class's sustainability. Bachelor and mercenary knights generally don't have any knighted children (they usually can't afford to care for a wife and kids to begin with, much less the exorbitant costs of knighting one of them), and landed knights generally only knight their first and second sons, the Heir and Spare. The refill population of household and mercenary knights comes overwhelmingly from second sons of landed knights whose older brothers didn't die before having their own kids, which by itself essentially guarantees that landed knights and living heirs thereof should make up more than 50% of the knightly class.

Of course, sourcing your knights differently can change those facts to some extent. While I don't use this idea in my campaigns, Book of Knights & Ladies suggests one potential difference between early and later Periods being that early Periods are more open to having lords errantly foist knighthood upon commoners who impress them, and being less interested in whether those common-born knights have a position lined up to fill with their new knightly status. If such knights are common enough, they'd at least contribute to a much higher proportion of mercenary knights. (Personally, I run it almost exactly the opposite way - knighting some rando who impressed you is an enormous show of largesse and positive acknowledgement of virtue that seems fitting more to the prosperous and generous times of Arthur, *doubly* so if they're not being knighted to fill some urgent need on the knighting lord's part.)

Morien · September 17, 2023

Speaking of demographics and historical parallels...

There is also the population growth that should be happening in Arthur's times, assuming that it parallels the historical. Then again, the timeline is so telescoped that who knows what is going on, it would lead to a very ahistorical yearly population growth, and again, need more kids. But if BoB2 is right and Logres ought to have 9000 knights by X Period, then you need to get a lot more people and cultivate a whole lot more lands. Anyway, this kind of population explosion could easily turn the original £10 manors into something like £30 mini-estates, which would not only allow more knights to be supported, but easier to fund knighting them, too.

1 hour ago, mj6373 said:

That's not to say lords were keeping as little land beyond their central estate as demesne as possible, because even though it *does* lower the total military output capacity of your holdings, you *need* enough strength not tied to your vassals that you aren't easy for a strong vassal to rebel against. But the idea that a lord would keep upwards of 80% of their land as demesne, and get Honor hits for going below that because it makes them seem *weak*? Very silly.

I don't know where Greg got that 20/80 split, so I can't comment on that.

I do agree with the main thrust of your argument, although my assumption is that the nobles with more scattered landholdings would be doing a 'progress' through their lands during the peaceful years at least. Similar to the itinerant Royal Court, but in smaller scale. Thus, the people would go where the food is, rather than vice versa, which would help to alleviate some of these issues. As well as having some local household knights assigned to outliers, as you pointed out as well.

That being said, it could be nice to have a military incentive to have vassal knights. It would actually be pretty easy to accomplish, too, as a quick house-rule.

For example, let's give the vassal knights an extra £1 'own manor efficiency' bonus. Now, the Liege Lord could, if he wishes, to demand some extra footsoldiers from the vassal knight, say two, which would cost that £1. This means that he himself could lower his own SD by 2 footsoldiers, meaning that for every vassal manor he hands off, he gains an extra £1 to spend as he wishes. Or he could keep his army the same, meaning that each vassal manor would bring him extra 2 footsoldiers for the army.

Or if you want to go OP with it and really encourage subinfeudation, just double the landholding when it is enfeoffed to a vassal knight. Thus, a £300 baron who subinfeudates half of his lands, could bring a force of 45 knights to the field, rather than 30. However, in this case, his vassals would outnumber him 2:1, leaving him in a pickle if they dislike his rule.

Obviously, both of those houserules would necessitate extra bookkeeping and ruling where the line really is. Like maybe caput major estate would benefit from the 'own manor' bonus, but nothing bigger would. And if someone has manors all over Logres or even the same county, only their 'home manor' would get this bonus. Which would encourage the lieges to actually get a bunch of single manor knights.

Anyway, just off the cuff brainstorming there.

vegas · September 18, 2023

On 9/17/2023 at 4:08 AM, Morien said:

You cannot use the GPC's representation and overwrite BoU assumptions where that population number is given. The society in BoU is clearly early-to-mid-1100s, not Migration Era - Dark Ages.

This becomes a question of the designer's intention. If the GPC is obsolete, and we are modeling the Uther era as 12th Century Britain, then for sure, the population should be larger as you suggest. The number of knights and armies should be larger too. Your points about levy and mercenaries are fine, but 2600 cavalry is less than what Harold and William each had at the battle of Hastings in the 11th Century, and that was not Harold's full army! If Uther is 12C Britain, then armies must be even larger.

I will say, I am surprised by your take that the BoU is not intended to be compatible with the GPC. The conceit of Pendragon has always been to project back on the Arthurian world a society and technology that did not exist, representing the genre from a later time, and this has to be doubly so for the dark ages where we have scant historical record to work from and can't really speak about society (royal, noble or otherwise) in that time frame in terribly meaningful way.

Regardless, for the purposes of this demographics exercise, I take the GPC as is for now and am using a dark age setting as my load star. An imaginary dark ages with Arthurian tropes and types is what I want for my setting and what the GPC describes.

On 9/17/2023 at 4:08 AM, Morien said:

This [~2.85 to less than 3.5 household size] seems like a very low number compared to the rule-of-thumb 5 people per household that I have seen in general use (such as in the Table 1 of the reference I gave above). Obviously, whether you use ~3 or ~5 has a big impact on the final population numbers, assuming that the number of households stays the same.

As you note later on, household size and population growth rates are connected. Faster population growth rates requires larger households (more offspring) for the same life expectancy. This exact point is made in your source on p 5, section 2. "A Benchmark for 1377." Also note, the text in that section 2; the historians listed in that Table 1 are all just guessing. They know there is an under-count of children in their data, and they are plugging in different estimates for the under-count driving the household size and ending population for their estimates. That is fine, guessing is required. But the high middle ages was a period of extremely fast population growth (see your source, Table 8 C) while growth rates in the dark ages were less than a 1/10th as fast. Household sizes have to be adjusted down accordingly. You will find no literature about dark age households sizes because there is no data like the 1377 poll tax. We just have to model it and guess too, but when we guess, it should reflect the growth rates for our time period.

I haven't given a lot of thought to appropriate household size for 12C setting (yet). For a dark ages setting, it has to be smaller way smaller than 4.2 or 5.

On 9/17/2023 at 4:08 AM, Morien said:

Ah, here you misunderstand the BotE/W economics.

The household knights do not spring up from smaller manors, but the same size as the vassal knights. The only difference is what happens with the excess income: does it go to supporting the vassal knight's family and 'court', or to the higher noble's family and court?...

...In short, taking your numbers as gospel, 285 commoners per manor (assuming one manor per knight), it would still result in ~750 000 commoners

No, I do understand - all manors are the same size, but knights cost different amounts - and I explained in my endnote [5] that I know this calculation is wrong and called it a "happy accident". I'm just following what you were doing and pretended that all knights are funded by manors and seeing what happens. It turns out it works perfectly, though of course we have to then find funds for court costs and lavish lifestyles.

But also, no, 285 commoners per manor does not result in 750,000 commoners, because this simple calculation (285 * 2610 ~750K) ignores too much.

I avoided going down the path of trying to reconcile all the income sources to expenses in the OP, because it involves a lot of detail and yet there isn't actually enough data to complete the tie out, so it becomes a frustrating exercise in minutia that I spared readers. (Of course I had already tried to do it!) But I can point the way: BoU p 9 shows L12.2 of income in Logres per knight, which is a clue there are substantial sources of income in addition to rents from basic L10 manors. Some of them can be enumerated: 10% higher rents in market town Hundreds, 20% higher in ports, royal income from natural resources & forests. We can tally all those up using the assumption of a commoner population of 495,000, 90% of whom are on manors, and 285 heads per manor, and conservatively they provide ~L18,500, leaving ~L7,500 needing to be "found" to fund 2600 knights.

Some of the sources of the extra L7,500 are noted without being quantified. The largest of these is church income - tithes, gifts, and labor of monks and clerics (there is a number listed for bishop and abbot rents but it is difficult to interpret). Tithing by peasants alone should be way more than enough to fund the 290 knights of the robe. For this high-level analysis, we can treat that as if it is L2,900 leaving L4,600 to explain. We know 10% of the population is the cities, ports, and market towns, and they are productive and taxed too, but not how much. If we assume they provide per capita as much as peasants in port Hundreds, that is another ~L2,000 reducing our gap to L2,600. We know the royal share of forest income, but not the baron's share (and we know they get a share from example honours); there is income earned from court profits and other liberties which are described at length in BoU without quantification; there are manors with productive investments that generate extra income. Those are the sources we know of from the books, and if we allow ourselves, we can come up with more. What we have quantified with conservative assumptions got us 90% of the way there to L26,000. Make slightly less conservative assumptions and add some value to these items that we haven't quantified at all and you can bridge the last 10% and justify funding 2600 knights with 495,000 commoners.

On 9/17/2023 at 4:08 AM, Morien said:

60% of adult noble males being knights is quite high, though. A quick back of the envelope calculation says that if a 21 year old knight lives to 42, he should have had 3 squires during that time. If he dies, he is replaced by a (younger) squire, so you would expect something like 1 knight and 2 esquires in the adult noble males, or about 33% knights. Obviously, that is a very rough estimate, but I think it is a better one than 60%.

I agree that 60% is high (and so is the 40% suggested in BoU), but it fits with the fiction that all first born sons become knights, and some later born sons will be too.

And you are right, there aren't enough noble squires to go around. That is probably why in the literature so many knights have a dwarf for their squire! More seriously, I think the fix is just to acknowledge that there are more lower-class squires among the household knights, while of course the 20% vassal knights (ie. the PKs) are going to disproportionately get the noble squires. 5e doesn't deal with these issues because it assumes the PK is a vassal. If 6e moves the assumption to them being HH knights, I hope things like the squire tables and marriage tables get adjusted accordingly for these lower status knight.

On 9/17/2023 at 4:58 AM, Morien said:

Although I admit I am surprised by the 0.67 number, which seems very low to me. I am probably discounting the older adults?

Alright, I just ran my code for the BotE family survival rates (life expectancy at birth ~ 34 years, the revised table in BotE v.1.3.2), and I get

Forgetting older adults is common. A LE(0)=35 population is going to have LE(21)>>14 (in fact, it is about 35) so there will be lots of older adults and many widows/widowers.

I would discourage you from using the game table survival rates for modeling the population as a whole. It seems nice for internal consistency, and is probably easy for you since it is on your computer, but will not end up giving reasonable population-level statistics because a game table is necessarily limited by our dice mechanics, and you are going to end up propagating errors which don't matter at the PC-level but will mess up your population statistics. (In addition we really should use different assumptions for the peasantry than the nobility, but that is another matter.) If you just look at a LE(0)=35 life table, you see that ~40% of the population is under 21, ie. ~.66 children per adult.

On 9/17/2023 at 11:26 AM, mj6373 said:

The whole "80+% of knights are household knights" idea doesn't even make sense for the class's sustainability. Bachelor and mercenary knights generally don't have any knighted children (they usually can't afford to care for a wife and kids to begin with, much less the exorbitant costs of knighting one of them), and landed knights generally only knight their first and second sons, the Heir and Spare. The refill population of household and mercenary knights comes overwhelmingly from second sons of landed knights whose older brothers didn't die before having their own kids, which by itself essentially guarantees that landed knights and living heirs thereof should make up more than 50% of the knightly class.

Yes, this is a demographic problem too. It can't be both 80% HH knights AND HH knights don't marry; the population of knights would collapse. I don't really have a view on whether or not 80% HH knights is historical, I just take it as a setting given, but then because of the demographic reality, I assume nearly all HH knights do marry, and when they retire, their eldest son takes their place to keep the population of knights roughly stable. Your solution of 50% (or less) HH knights works too. One way or another, something has to give.

On 9/17/2023 at 1:33 PM, Morien said:

There is also the population growth that should be happening in Arthur's times, assuming that it parallels the historical. Then again, the timeline is so telescoped that who knows what is going on, it would lead to a very ahistorical yearly population growth, and again, need more kids. But if BoB2 is right and Logres ought to have 9000 knights by X Period, then you need to get a lot more people and cultivate a whole lot more lands.

Yes, the population growth in the GPC is another impossibility. To me, the answer is just don't model the demographics required to move BETWEEN eras. It would be ridiculously weird the speed of population growth required. Instead, just like the technology magically advances between eras, I would just magically advance the population statistics too, and make each era is self-consistent and consistent with its historical analog, but give up on trying to tie the change between different eras together demographically through birth rates, households sizes, etc.

Edited September 18, 2023 by vegas

Oleksandr · September 19, 2023

10 hours ago, vegas said:

but 2600 cavalry is less than what Harold and William each had at the battle of Hastings in the 11th Century, and that was not Harold's full army! If Uther is 12C Britain, then armies must be even larger.

It should be noted that medieval chronicles usually greatly exaggerated number of troops, on both sides. It's usually assumed that real armies was much smaller.

10 hours ago, vegas said:

Some of them can be enumerated: 10% higher rents in market town Hundreds, 20% higher in ports,

Don't forget +10% in counties with cities, + forest income represented as another +10% to agricultural income of manor. Additionally, there are factor of land fertility, which can affect income substantially. In "Saxons!" numbers are +/- 25%, while here👇 we can see a noticeable difference between settlements of same sizes (and in the same hundreds, so bonuses should be equal):

http://web.archive.org/web/20170226164946/http://gspendragon.com/swans_hundred.pdf

http://web.archive.org/web/20170226165038/http://gspendragon.com/ambrius_hundred.pdf

With all this bonuses it should be easy to feed all required knights even with smaller population,😉

Morien · September 19, 2023

15 hours ago, vegas said:

This becomes a question of the designer's intention. If the GPC is obsolete, and we are modeling the Uther era as 12th Century Britain, then for sure, the population should be larger as you suggest. The number of knights and armies should be larger too. Your points about levy and mercenaries are fine, but 2600 cavalry is less than what Harold and William each had at the battle of Hastings in the 11th Century, and that was not Harold's full army! If Uther is 12C Britain, then armies must be even larger.

Harold explicitly had the fyrd, which is the free man militia equivalent. His army was mostly huscarls and fyrd, and famously lacking in cavalry. The local fyrd can easily supply Harold with enough men.

As for William's forces, I have seen estimates as high as 3000 for cavalry, but was all cavalry knights? I doubt it. Even if we go by Pendragon's 1 knight + 1 squire and ignore any non-knight cavalry, 1500 knights + 1500 squires = 3000 cavalry in total.

15 hours ago, vegas said:

Regardless, for the purposes of this demographics exercise, I take the GPC as is for now and am using a dark age setting as my load star. An imaginary dark ages with Arthurian tropes and types is what I want for my setting and what the GPC describes.

OK, that is your prerogative, certainly. My argument was for the Anglo-Norman ~early 12th century, and it seems that we agree that the higher population is appropriate for the later historical period.

I'll just address a couple of points below for completeness.

15 hours ago, vegas said:

No, I do understand - all manors are the same size, but knights cost different amounts - and I explained in my endnote [5] that I know this calculation is wrong and called it a "happy accident". I'm just following what you were doing and pretended that all knights are funded by manors and seeing what happens. It turns out it works perfectly, though of course we have to then find funds for court costs and lavish lifestyles.

The upkeep for the (ordinary standard of living) knights is the same, regardless of their status. True, the vassal knights are often married as soon as they can and thus would soon be providing for the family as well, but the knight+squire+horses is the same £4 per year.

I agree that assuming that each manor of £50 has ~ 100 households, and that every household has ~5 members is a HUGE assumption. But one has to start from somewhere.

But the Agricultural lands table in BoU p.9 actually leads to a higher number of people, not smaller. It is Assized Rents, so without any bonuses from anything else. If we assume that the typical PK manor is typical, so with £10 Assized Rents and 100 households, we can divide the ~£32000 AR by £10 and get about 3200 'manor-equivalents'. And then multiplying by ~300 people per manor, we get a total commoner population of about 960 000. Add the nobility and clergy (in proportion) and you get very close to 1 million, even with a 3-person average household.

The above is just to show that you can easily argue for larger population sizes if you want to, but it all rests on your starting assumptions.

(I have a few more things to say, but it needs to wait until I am back.)

(OK, continuing...)

As an aside, BoU, p. 9, Table 1.2 also shows that while the barons of the robe get about £5000 in AR, their SD requirement is more like 60% of what the barons of the sword are required to bring. No doubt the excess goes to specifically to the upkeep of clergy and maintenance of churches and such, which is also where I would expect the tithes to go.

I admit that I couldn't quite follow your version of the budget... I am sure I could, once I reread it a few times more, but it is not really that important as it all hinges on the assumptions that we both admit are contradictory (your pref for a Dark Ages population and mine for 12th century) or very uncertain (the average AR attributed per person).

Instead, I'd rather talk about something else, an alternative way to extrapolate the total population. With its own caveats and assumptions.

Let me quickly do some more clean-up:

I agree that trusting on gamefied survival tables for noble npcs and extrapolating that to a commoner demographics is a somewhat suspect. Certainly not something I would ever suggest doing for an actual paper studying historical medieval population. That being said, it did give us some numbers that are at least broadly consistent with the numbers that researchers have used for real data, like 35-40% of the population being under 15.

Speaking of internal consistency, that is the other way of doing it, building it from knights out, assuming that the number of knights mentioned in BoU and the percentages per social class are correct.

If we take 2610 knights as gospel, and assume, as I do, that each knight trains about 3 (mostly) noble squires during his active career, we should see two adult male esquires per knight, or a total number of noble men being 3xknights = ~8000 (rounded for convenience). Add an equal number of noble females, and we get 16000 adult nobles. Using your 0.67 kids per adult, we'd get ~27000 nobles. Then applying the BoU class percentages, we can add about 18 000 clergy to this, for a total of 45 000 nobles+clergy. Since they are together 5% of the total population, we can just multiply by 20 and get 900 000 total population for Logres. Again very close to about 1 million.

If you vary those assumptions, say if the knights live longer and train 4 squires instead of 3, and the noble families have more kids, like 1 per adult, the number would become ~20 000 adult nobles, 40 000 nobles, and multiplying by 33.3 (since nobles are 3% of the total) = ~ 1.3 million.

Conversely, if you assume that the number of knights vs. esquires is closer to 1:1 (whether dwarf or lifer squires are common, or that the knights are dying left and right, meaning more dead men's shoes to fill), then you can bring the numbers down by a third, to about 600 000 which is pretty close to BoU number and your preference (with 0.67 kids per adult, too). Assuming 1 noble kid per 1 noble adult would make it about 700 000.

Again, depending what assumptions you use and plug in, you can easily get values from 600 000 to 1.3 million.

Edited September 19, 2023 by Morien

Morien · September 19, 2023

16 hours ago, vegas said:

Yes, this is a demographic problem too. It can't be both 80% HH knights AND HH knights don't marry; the population of knights would collapse. I don't really have a view on whether or not 80% HH knights is historical, I just take it as a setting given, but then because of the demographic reality, I assume nearly all HH knights do marry, and when they retire, their eldest son takes their place to keep the population of knights roughly stable. Your solution of 50% (or less) HH knights works too. One way or another, something has to give.

My solution for this is to soften the statement on the marriage of the household knights. It is not that they don't EVER marry. It is that they don't USUALLY marry young, or if they do, it is not within their class, but some commoner with very little to offer in the way of dowries.

So, my solution to this is roughly as follows:

Vassal knights provide, on average, 2 knights for the next generation, the heir and the spare. So that is 2*20% = 40% of the next generation.

Household Knight officers DO marry, and provide at least one knight for the next generation. While there are not that many officer spots in general, they tend to be awarded to older guys, so they don't stick around that long (and thus more HHKs get the turn on the chair). Or if they do, they have the wherewithal to get a spare, too. Call that 20% of the next generation.

Roughly a fourth of the household knights (so 20% of the total) do marry commoners (supporting that family from loot/largesse from family/friends/liege or her own work, even) or noblewomen (working as ladies-in-waiting, etc), and generally manage to get a son, who inherits the father's equipment. So this is another 20%.

So you roughly end up with a bit less than half of the HHKs who marry and the rest who do not (dying young, not distinguishing themselves enough for the Liege to help in the support of a family). This proportion might still be a bit high, given that BotW says (p. 7) that they don't usually marry (but note that it doesn't say that they NEVER marry). However, it is about half of the HHKs during their whole lifetime. If we assume that it is mainly the older ones who marry, in any snapshot of the HHK pool, the unmarried, younger cohorts would be the majority. Good enough for me.

Finally, the esquire officers generally do marry and get families, especially the more distinguished ones, so now and again, there are promising youths who manage to get sponsored by the liege (perhaps when knighting his own son, knighting some of the son's childhood friends as well) or by not-too-corrupt graft over the years by the officer (you know, gifts for your consideration kinda thing...). This would be the last 20% or so.

Of course, given that there are at the very least as many esquires as there are knights (or close enough), as well as the younger brothers of the ones who got knighted, they help to fill in the squire pool and the esquire ranks later in life. And there is some bleed in from richer merchants, common soldiers of obvious prowess (especially in Uther Period) as well.

Edited September 19, 2023 by Morien

vegas · September 19, 2023

This is a fun (to me) conversation, but I have to prep for my game session (more fun and more important!) so I will try to brief (obviously hard for me!)...

3 hours ago, Morien said:

....As for William's forces, I have seen estimates as high as 3000 for cavalry, but was all cavalry knights? I doubt it....

I didn't mean to start a side-bar on the size of the Battle of Hastings, only make the point by illustration that 2600 knights is very small for our imaginary 12C Kingdom of Logres when we look at an 11C invasion force from the little duchy of Normandy managed more. There were not yet knights in the sense of Pendragon in the 11C, so I am not sure how parse this farther and distinguish "cavalry" and "cavalry knights". If you don't find this illustration compelling, oh well, but for me it drives home the problem.

3 hours ago, Morien said:

The upkeep for the (ordinary standard of living) knights is the same, regardless of their status...

I am 99.9% sure we both understand perfectly, but somehow we are talking past each other here. Yes, their upkeep is the same, the associated foot cost is the same; the difference is the vassal has his own court, family, and discretionary spend so a vassal costs £10 while HH knight (and associated foot) cost £5.5.

3 hours ago, Morien said:

But the Agricultural lands table in BoU p.9 actually leads to a higher number of people, not smaller. It is Assized Rents, so without any bonuses from anything else...

Yes, it does if we take the table literally. It however cannot mean what it is labeled or there is an error in the math because it breaks a lot else otherwise.

If it is literally rents, then yes there are on the order of 3,100 manors and 310,000 households. If we are using 4.2/household, you need over 1.5M people when you add in the non-manor commoners, clergy, & nobles. And if there is £31,800 of manor income AND income from forests, mining, urban populations, court fees, etc. then we break the economic model too. £31,800/2600 knights is already £12.2/knight. If you pour all these other sources of income on top and you aren't spending 55% army, 25% court, etc. (As you can tell, I have already been swimming in the numbers of BoU to try to make sense of things.)

I suspect someone erred and £31,800 is intended to be total "government" (kings, vassals, & clergy) income of Logres in part because everything just works if it is. £31,800 less clergy portion is £26,800, divide by 2600 knights is ~£10/knight. That is too suspiciously perfect to think that was not how somebody built that table.

Regardless, one way or another, something has to give. For me, wanting to keep the total Logres population at 500,000 and 2600 knights, I assume that was government income rather than assized rents. The other solution is to triple the population and add more knights & foot. Pick your poison.

5 hours ago, Morien said:

If we take 2610 knights as gospel, and assume, as I do, that each knight trains about 3 (mostly) noble squires during his active career,...

I will have to give this a think. It seems a little like having the tail wagging the dog since we are making squire training the driver of population. There are so many things that could influence that number - violent deaths of knights (and squires!), prevalence of esquires and commoner squires - it seems a volatile place to build out from, but that is just a quick take.

3 hours ago, Morien said:

So, my solution to this is roughly as follows:

Vassal knights provide, on average, 2 knights for the next generation, the heir and the spare. So that is 2*20% = 40% of the next generation...

Yeah, I assume close to the same but simpler. I have all the spares marry but only provide one HH knight (dad retires, get the gear), then in the 3rd generation you have 60% vassal, spare, & 1st cousin. Repeat with cousins marrying providing one HH knight and its 80% vassals, their brothers, 1st, & 2nd cousins, and the last 20% are either distant cousins or great soldiers that got promoted. Half of your bottom 40% (2nd cousins & further removed from a great family) either don't marry or their kid ends up a mercenary knight or less.

Morien · September 19, 2023

14 minutes ago, vegas said:

This is a fun (to me) conversation, but I have to prep for my game session (more fun and more important!) so I will try to brief (obviously hard for me!)...

No problem. I think we are reaching a consensus here. 🙂

14 minutes ago, vegas said:

I didn't mean to start a side-bar on the size of the Battle of Hastings, only make the point by illustration that 2600 knights is very small for our imaginary 12C Kingdom of Logres when we look at an 11C invasion force from the little duchy of Normandy managed more. There were not yet knights in the sense of Pendragon in the 11C, so I am not sure how parse this farther and distinguish "cavalry" and "cavalry knights". If you don't find this illustration compelling, oh well, but for me it drives home the problem.

The estimates for the cavalry of William range from 1000 to 3000, and we don't know if every cavalryman was a knight, is what I was trying to say. Also, William hired a lot of mercenaries from all over France: his forces were assigned into three battles based on their origins, and only the center one was Normans. So no, I don't think Duchy of Normandy had 3000 knights, nor even 2610.

But yeah, it is getting off-topic a bit.

14 minutes ago, vegas said:

I am 99.9% sure we both understand perfectly, but somehow we are talking past each other here. Yes, their upkeep is the same, the associated foot cost is the same; the difference is the vassal has his own court, family, and discretionary spend so a vassal costs £10 while HH knight (and associated foot) cost £5.5.

My point is simply this:
Vassal knight: £5.5 for the army, £4.5 for his family and court and discretionary spending = 1 knight per £10.
Household knight: £5.5 for the army, £4.5 for the liege's family and court and discretionary spending = 1 knight per £10.

You need to compare like with like here. The barons are not supporting two household knights and 4 infantry out of every £10 demesne manor, just 1 HHK and 3 infantry, same army as for a vassal knight manor.

14 minutes ago, vegas said:

I suspect someone erred and £31,800 is intended to be total "government" (kings, vassals, & clergy) income of Logres in part because everything just works if it is. £31,800 less clergy portion is £26,800, divide by 2600 knights is ~£10/knight. That is too suspiciously perfect to think that was not how somebody built that table.

OK, now I understand somewhat better how you did your accounting.

I don't think it is the case, but fair enough. I can see how you would come to that conclusion, and it would be somewhat pleasing if that were the case. But as I said, I don't think it is. For one, the Clergy is clearly providing knights, too. I don't think the table is infallible, either. For example, we know that Ulfius has lightened SD requirement, but we don't see it reflected in the Great Nobles row. Could be that Uther hit Gorlois with some extra servitium debitum, just to mess with him, but even more likely is that Greg didn't update the table. It also isn't an exact match with BotW, p. 38, Table 3.1.

Anyway, I wouldn't sweat too much about it. Do what works in your campaign.

14 minutes ago, vegas said:

Regardless, one way or another, something has to give. For me, wanting to keep the total Logres population at 500,000 and 2600 knights, I assume that was government income rather than assized rents. The other solution is to triple the population and add more knights & foot. Pick your poison.

Yeah, agreed.

14 minutes ago, vegas said:

I will have to give this a think. It seems a little like having the tail wagging the dog since we are making squire training the driver of population. There are so many things that could influence that number - violent deaths of knights (and squires!), prevalence of esquires and commoner squires - it seems a volatile place to build out from, but that is just a quick take.

Yeah, which is why I said you could get easily anything from 600 000 to 1.3 million with just varying the assumptions a bit. As you said earlier, it is probably just easier to decide what kind of a historical analog you are going for and what would be a reasonable population for it, and then go for it.

14 minutes ago, vegas said:

Yeah, I assume close to the same but simpler. I have all the spares marry but only provide one HH knight (dad retires, get the gear), then in the 3rd generation you have 60% vassal, spare, & 1st cousin. Repeat with cousins marrying providing one HH knight and its 80% vassals, their brothers, 1st, & 2nd cousins, and the last 20% are either distant cousins or great soldiers that got promoted. Half of your bottom 40% (2nd cousins & further removed from a great family) either don't marry or their kid ends up a mercenary knight or less.

Good point about the Spares marrying, need to make sure that there is a cadet branch ready to go if the main one snuffs it. This would alleviate the issues for the non-spare Household knights marrying, by the way. Granted, you still need to find a way to pay for it, but there are usually some ways to do it, at least for some of them. But yeah, there would be some incentive for a liege to cultivate the spares, too, just in case they happen to inherit, it is better to have a loyal, devoted vassal than one with an axe to grind.

mj6373 · September 22, 2023

I would say that this most recent point is salient to the extent it's less a separate explanation for a different portion of knighthood and moreso an overlap/explanation for earlier cases. While there's certainly real room for people to distinguish themselves and earn exceptional rewards, when we're talking patterns, those come down to incentives and power structures. If a lord is choosing from among his household knights who to make an officer, or who to set aside extra money to support a family for, he will, when able, be far more likely to choose one of the "second son of a landed guy" household knights, because this may improve relations with the family and make for a grateful future vassal if something goes wrong. And while continued dependence on one's parents/siblings is undesirable and therefore probably uncommon, second son knights will also have greater security simply from having a wealthier father/brother to ask for help.

So whatever scarce resources of officer positions and marriage income and son-knighting-funds are available to supply to household knights, they will overwhelmingly go to those household knights in the "second son of landed knight" camp, assuming any are available.

Just my two cents.

Uqbarian · October 19, 2023

Oh man, one of my favourite topics and I missed it!

It's all fuzzy and there's a lot of room for variation, but another thing that I think lends support to the higher population suggestions is the town populations.

Book of Uther p. 93 says that burghers (town dwellers) are about 5% of the total population of Logres. (In context, I assume this includes families.)

(Not all the inhabitants of towns will be burghers, but I think it's reasonable to assume that most are. My assumption is that slaves make up the only significant non-burgher town population for counting purposes.)

There are 64 market towns and cities. (BoU p. 98 says there are 64 in Britain, but given the list and the associated map, 'Britain' here evidently means the kingdom of Logres.) Of these, 3 are great cities, 5 are large cities, and the remaining 56 are market towns.

(Some of the market towns are small cities, but I don't think BoU says which are which. It's worth remembering that 'market town' and 'civitas' are legal categories, whereas 'city' is more of a size category in this context. For example, there could be a settlement somewhere that has only 12 permanent residents but is officially a market town, and there are definitely town-sized settlements that don't have market town status. On the other hand, 'burgher' is also a legal class rather than a social class. I'll assume for convenience that residents of non-market towns do not count as burghers.)

Using figures from BoU p. 96:

3 great cities, each about 10K pop: 30K

5 large cities, each about 5K pop: 25K

56 market towns (including small cities), say average 750 pop: 42K*

Total urban pop: 97K

*This is probably an underestimate, as non-city market towns are given as 625-750 and small cities are 750-1500. This also doesn't include towns without market town status.

Regarding slavery, my personal preference is to assume it doesn't exist in Logres, but it is there in the book, so I'll try to factor it in. Looking again at BoU p. 93, up to 20% of the commoner population are slaves, and slaves are found in the 3 great cities (Londinium, Venta and Glevum) and 2 of the large cities (Camulodunum and Durnovaria). (Slaves are also found in Eburacum, but that's outside Logres.) Taking 20% out of those cities removes 8K from the urban pop count (2K for each of 3 great cities, 1K for each of 2 large cities), leaving us with 89K.

If we assume this non-slave urban population is roughly equivalent to the burgher population, and if we accept that the burgher population is 5% of the total population, that suggests a total population of roughly 1.78 million.

If we instead assume that the urban population contains a significant number of non-burgher and non-slave inhabitants -- let's say an arbitrary 14K because we can see where this will get us -- we're left with 75K burghers and a total population of roughly 1.5 million.

vegas · October 20, 2023

Ah, you haven't missed anything Morien is always here and I'm often lurking. 😉

I like your idea to add another dataset for perspective and consistency, but the challenge of this approach leaps out; it is a hazardous methodology to try to estimate a larger whole by dividing from an estimate of a small fraction. Small errors in either the percentage of urban dwellers or their number will result in large errors in your estimate of the total population. So we must take this with a heaping of salt.

I also think you are misreading BoU when you equate burghers with town dwellers to get your 5% urban estimate. You recognize this isn't strictly correct in your parenthetical, but proceed to assume that most inhabitants of towns are burghers. "Social Classes" in BoU pp 93-94 section makes explicit there are others in urban spaces. The people living in towns are not just the "Townsmen" who we read are "the burghers, bourgeois,... petty town bureaucrats.... [the] merchants and traders" but also there are "Craftsmen" who include "common craftsmen and artisans" that are "in villages, towns, and cities" and "specialized ones" who "live only in cities." And while they are not mentioned in BoU, if we are world building we have to recognize there are other classes present as well: the burghers have servants in their homes and laborers in their enterprises. Servants historically came from unmarried commoner women who migrated from rural areas, and commoner women and men performed urban labor. The towns and cities will not only have slaves owned by the burghers, but also host runaway slaves and serfs who seek a better life by begging, laboring and hoping to apprentice. Taken as a whole, the market town and city populations are going to be just a diverse in social classes as the rural areas, with the burghers representing only the very top of the urban social pyramid. This breaks the key assumption of the methodology, that we can back into the total population by dividing the estimated urban population by 5% burghers. Not only would I expect burghers are not "most [of the] inhabitants of towns", if we recognize they are the upper stratus of urban populations, we would expect they are in the minority.

Morien · October 20, 2023

My reading is closer to Uqbarian's. Greg was notoriously lax with his language, but the burghers/city dwellers/burgesses/townsmen clearly include all who are citizens of the city/market town, including the manufacturers/craftsmen living in the town rather than in villages. Not just the upper crust administrators.

That being said, there is a clear internal inconsistency between ~5% burghers being ~90k, and ~5% of the population being non-commoners with 25k. Given that I don't like the 25k number to begin with, it is clear which way I am leaning. 😛 No doubt there are some escaped serfs / slaves in the cities as well, but most of the population would be a bit too many, IMHO.

In the end, I seriously doubt Greg spent too much time worrying about this.

vegas · October 20, 2023

Certainly Greg was sloppy with definitions of all sorts, but I gotta disagree with you on what burghers "clearly" includes. In plain English the word might refer to the urban middle/upper class or any city dweller. Which sense did he mean? It is not clear! But each time he used the word, it is associated with the upper class and other classes that are city dwellers get other descriptors. And as you just wrote, the math doesn't work unless there are city dwellers in addition to the burghers, suggesting burghers (the legal class) are city dwellers but not all city dwellers are burghers (the legal class).

On this view, it is certainly (as you suggest) the case that a minority of the urban population are slaves and escaped slaves (as slaves are a minority of the population as a whole). The majority of urban populations will be yeomen and escaped/released serfs: servants, laborers, craftsmen, etc. There will be a population of clergy, but again a minority. Burghers (the legal class) are the merchants, traders, the enterprise owners, and they too are a minority of the total city dwellers. That model fits quite nicely with a 500K total population, 25K burgher population, and another ~50K of other city dwellers (exactly how many depends on your estimate of the total urban population.)

Edited October 20, 2023 by vegas

Uqbarian · October 21, 2023

Yeah, that's fair. I hadn't thought about non-slave servants in town households, which would be another significant part of the urban population. Also, BoU p. 94 does clearly equate 'townsmen' (as a social class) with 'burghers' (the legal class), and it further describes the social class as just the merchants and traders, both of which support your side. On the other hand, p. 93 does say 'while some [burghers] are farmers, most are manufacturers, merchants and traders', so the class isn't just the merchants and traders. There are ways to reconcile that inconsistency, but I think they can be turned to either side of the argument. So I haven't really shed any light on anything. 🙂

Uqbarian · October 21, 2023

Just for fun, I had a poke at clergy numbers.

BoU p. 7 lists the clergy as about 10K in number, or roughly 2% of the population.

BoU also lists 26 BC abbeys, 37 RC abbeys and cathedrals, and 12 pagan sites. However, only one pagan site is specifically listed as a ‘pagan abbey’ (Stone Abbey of the Giant), and one is listed as a ‘[c]ommunity of practicing Pagans’ (Longvale), so let’s say there are 2 pagan abbey-equivalents for a total of 65 abbey-equivalents in Logres.

From BoE p. 86, abbeys typically have in the range of 20 to 80 monks but can be larger. (BoE has a separate entry for nunneries, but BoU says these are ‘always attached to an abbey’, so I’ll assume these are included in the abbey count for convenience. I'm also assuming the 'monk' counts include priests who live in the abbeys, distinct from village priests.)

At a minimum of 20 monks (or monk-equivalents) per abbey-equivalent: 1300 monks
At an average of 50 monks per abbey-equivalent: 3250 monks
At 80 monks per abbey-equivalent: 5200 monks

(There might be a few large abbeys with hundreds of monks and nuns, but I don’t think there’d be enough of those to skew the average above 80.)

Priories are ‘too numerous to list’ (BoU p. 77). That suggests to me there are at least twice as many priories as there are abbeys, so I’ll take 130 priories as a baseline. Priories are listed under British Christianity, but there’s no reason to suppose Roman Christians don’t have them too, and the Salisbury county map shows two RC priories.

(I briefly thought the Salisbury map might help for counting priories. In Salisbury county, BoU lists 2 BC abbeys (Ambrius and the Rock) and 1 RC abbey (Borders). Looking at the county map in BoW pp. 206-7, I make out 4 priories (2 RC, 2 BC), but the map doesn’t show the abbeys of Borders or the Rock (one might say they’re hidden inside the castle icons), so we don’t know how many priories exist within other castles or towns on that map.)

From BoE p. 89, priories typically have 5 to 15 monks. Let’s say 10 monks per priory across 130 priories, for 1300 monks.

Adding that to our average abbey figure (3250 monks) gives a total of 4550 monks.

How about village priests? From BoU p. 75, canons (RC priests) ‘outnumber monks significantly’. British Christianity doesn’t have canons, but I think it’s reasonable to suppose that BC monks who conduct services in villages similarly outnumber BC monks in monasteries. I’ll call members of both of these categories (i.e. secular clergy for both RC and BC) ‘priests’ for convenience.

If the village priests don’t outnumber monks ‘significantly’ but are instead roughly equal in number, we could say there are 4550 priests to give us 9100 monks plus priests, or about 10K to match the BoU population estimate. ('Significantly outnumber' could mean priests in abbeys + village priests outnumber actual monks, but I don't think that makes sense in the context. However, we could fit more village priests within the 10K by assuming fewer priories and/or a lower abbey average.)

If instead there are 1.5 village priest-equivalents per monk-equivalent, i.e. 6825 village priests, we have 11 375 clergy all up.* Say we round to 12K to allow for the foreign churches (based in London), a few hundred hermits and anchorites, and/or a couple of hidden pagan orders. That can still fit within the rough 2% estimate from a rough 500K total population.

(One thing that jumps out at me, though, is that even at 4550 'village' priests, there should easily be enough for at least one priest per manor.)

For those of us who favour a larger total population, we can assume a high-end abbey average, more priories, more priests relative to monks, more pagan clergy, and/or a lower proportion of the total population (though for the last, I wouldn’t want to go below 1.5%).

*EDIT: And that's probably still underestimating secular clergy. I don't have a copy of J C Russell's paper on clergy numbers, but a citation at this Stack Exchange page has 24.9K secular clergy versus 10.6K religious (i.e. monastic) clergy, though that is for 1377. I haven't found a good eleventh-century breakdown online.

Edited October 21, 2023 by Uqbarian

vegas · October 21, 2023

Funny thing, I have tried to look at this too because one of my players wanted to be a monk, but didn't sync it up to the setting-wide statistics. I like what you are doing here, but some comments.

It is unlikely in population statistics that the distribution will be normal within a range or that the average is the mid-point of the range. Instead we would expect something looking more like a Power Law or Benford's Law or Zipf's distribution with the vast majority of locations to being very small. The range will have will have a mode close to the minimum, the average pulled higher by the right-hand tail of the distribution, but both of the points below the mid-point of the range. Similarly for priories (which I suspect would be distributed on the same curve as abbeys, with some overlap as priories are effectively small abbeys, even though being slavishly literal the BoU descriptions would not allow for that.)

If I was going to choose some figures to fit the 500K population, 10K clergy figures,... [redacted]

EDIT: I looked up the paper mentioned in your link (THE CLERICAL POPULATION OF MEDIEVAL ENGLAND, by JOSIAH COX RUSSELL, Traditio, Vol. 2 (1944), pp. 177-212) and it paints a more complex picture. I'll come back and post more later.

A related but slightly different point, I would assume that priests tend to be collected together in noble courts and urban cathedrals so villages remain relatively under-served. The clergy in Logres tend not to choose to reside with the commoners in the field such that several villages often share a single priest.

Edited October 24, 2023 by vegas

Darius West · October 23, 2023

On 9/17/2023 at 3:06 PM, vegas said:

Executive summary: The demographic data stated and implied in 5e sources don't all fit together. Morien suggests we should assume a much larger population of Britain versus Book of Uther. I would suggest this creates new problems and severely over-shoots at fixing the discrepancies. Instead, real-world demographics and history suggest we should keep the population of Logres as indicated in BoU and reduce the average size of households in manor populations to reconcile the data.

Remember that Pendragon is set in a period when the population is rural in a way that modern demography doesn't handle well. There will be people living in forests, hills, plains, and mountains in far greater profusion than the urban centers will suggest. This is the age of the Manorial Economy, when things were extremely decentralized and your manor house could provide almost everything a knight needed, and the main thing that designated a town or a city was how many days a week the local market was open for. Urban centers likely held under 10% of the population during this time and even a village of 400 people was a pretty big concentration of capital, with most people living close to a hamlet at best. But... for all that, rural population density was a lot higher than today.

Edited October 23, 2023 by Darius West

vegas · October 29, 2023

On 10/21/2023 at 12:38 AM, Uqbarian said:

*EDIT: And that's probably still underestimating secular clergy. I don't have a copy of J C Russell's paper on clergy numbers, but a citation at this Stack Exchange page has 24.9K secular clergy versus 10.6K religious (i.e. monastic) clergy, though that is for 1377. I haven't found a good eleventh-century breakdown online.

It has taken me a while to get back to this, but I have some info for you. I have read the paper and it has the 11C data you are looking for. Also, the ratio of 2.5:1 in the 14C is not quite as simple as that. Here goes with the data.

1086: 4,500 secular clergy; 150 nuns*, 740 monks: 890 total; 5:1 secular to monastic

1377: 24,900 secular clergy; 2,054 nuns, 3,205 monks, 2122 regular canons, 3242 mendicants: 10,623 total; 2.3:1 secular to monastic.

The 1086 data comes from the Domesday Book and the 1377 data from the poll tax. Both of these are minimum estimates rather than complete censuses, but hopefully the under-counts are roughly equal between groups so we can use the ratios.

Also the note the details; BoU seems to treat canons as secular clergy in the discussion. Some of the nuns would have been canons and of course the regular canons are identified. Assuming the ratio of regular canons of nuns matches that of the males, we could recalculate the ratio of secular to monastic as 3.2:1.

*in the text he suggests 200, in the table he uses 150. Obviously this is a guesstimate on his part, but I reproduced his figure in the table.

Uqbarian · October 29, 2023

Neat! Thanks for looking that up. 🙂

Edited October 29, 2023 by Uqbarian

Oleksandr · November 5, 2023

On 10/20/2023 at 8:20 PM, vegas said:

The towns and cities will not only have slaves owned by the burghers, but also host runaway slaves and serfs who seek a better life by begging, laboring and hoping to apprentice. Taken as a whole, the market town and city populations are going to be just a diverse in social classes as the rural areas, with the burghers representing only the very top of the urban social pyramid. This breaks the key assumption of the methodology, that we can back into the total population by dividing the estimated urban population by 5% burghers. Not only would I expect burghers are not "most [of the] inhabitants of towns", if we recognize they are the upper stratus of urban populations, we would expect they are in the minority.

I'm not quite familiar with british definitions, but in Central Europe term "burghers" also include all craftsmen. Additionally, it should be pointed out that cities evolved greatly over time. For most of middle ages burghers was absolute majority of urban population. Mass influx of rural population into cities only happens when middle ages ended, due to exponentially increasing taxation - dissolution of classic feudalism and shift of aristocracy from warrior caste to court dandyes increased demand to luxury, and expenses.

P.s. 5% is quite low for medieval burghers...

On 9/17/2023 at 8:06 AM, vegas said:

Second, increasing the population to 1.25 million in Logres implies a very small militarized force. With one knight there are another 3 foot giving a total force under arms of 2610 * 4 = 10,440 or just 0.8% of the population in times of internal wars and foreign invasion.

Another important note, most research suggest that medieval armies was incredibly tiny (not including sporadic militia). Our problem with perception is that after middle ages armies was bloated to giant sizes, for many it's hard to believe that it wasn't always so.

Additionally, assumption of field armies being composed from "1 knight (and squire) + 2 foot troops" lances are based on much later examples. It seems that earlier it was "1 knight (and squire) + mounted sergeant".

Reconsidering Demographics in Pendragon

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Join the conversation