The Invention of Power: An Uninteresting Blog Post

This may surprise some of you reading, given that it’s my main tool for publicising the blog, but I don’t go on Twitter very much anymore. This is not a decision I particularly regret. (I’ve eaten some tasty bread made of the wheat from that site, but my God there’s been a lot of chaff.) Nonetheless, I do still go back occasionally. Recently, I noticed a group of medievalists dunking on one particular book, and it annoyed me. This tetchy response wasn’t anything to do with them – I’m perpetually cranky at the moment owing to trying to get everything finished at home and also organise an international move whilst planning as much travel between the UK and Germany to spend time with my wife as possible –  but still, I did think to myself, ‘It’s not fair to make fun of a book you haven’t actually read, so I’m going to get hold of a copy and read it really sympathetically and review it for the blog and say positive upbeat things about it.’

The book was The Invention of Power: Popes, Kings and the Birth of the West by Bruce Bueno de Mesquita; and reader, the Twitter medievalists mocking it were absolutely right. This book is dreadful.

This book took up valuable time I could have spent reading about the history of Walt Disney World

Unfortunately, the precise way in which it’s bad is so basic and uninteresting that it’s almost too obvious to mention: Garbage In, Garbage Out. I could leave it there, and I’m tempted to, but I suffered through all nine chapters and by God I’m going to get some blog content out of it.

So let’s start at the beginning. What’s the book’s argument? The book argues that the exceptionalism – by which the author means both the higher relative economic performance and freer, happier societies – of Western Europe and its ‘settler colonies’ can be traced back to a series of concordats made between the pope and the various rulers of the Holy Roman Empire, England and France in the twelfth century. In accordance with the terms of these concordats, a uniform procedure for appointing bishops was put in place, whereby the Church nominated a bishop, who was then accepted or rejected by the king; in the meantime, the king would pocket the temporal revenue of the diocese. This created a series of incentives for both the popes and temporal rulers. Kings wanted secular appointees, whilst popes wanted religious appointees. If the king turns down a Church nominee, he gets more money and/or the potential for a more loyal secularly inclined bishop, but has to pay the political costs of an angry pope. Therefore, the richer a diocese, the more incentive for the kings to be pickier: in a rich bishopric, the potential gains for the kings by not accepting a papal nominee outweigh the potential risks of papal anger. By contrast, in a poor bishopric, the inverse is true, and the Church has more leverage. This creates the incentive for kings to promote economic development to create lots of rich dioceses, whilst the Church would like to retard economic development to create lots of poor dioceses. To promote economic growth, kings needed to incentivize their subjects to increase productivity. This requires making concessions to these subjects to share in the benefits of increased productivity, concessions such as parliaments and other forms of accountable governments. Thus, in areas of Europe covered by the twelfth-century concordats, we would expect to see greater economic development and more accountable institutions than in other parts of the world. This combination of economic development and accountable institutions led to the exceptional Europe and Europe Outremer of the modern world.

…Look, I’m not going to attempt to touch on all of this. As more medievalists get hold of this book, I imagine an interested reader could take to Twitter and see more and more different approaches to analysing its failure as scholarship. For the rest of this post, I’m going to consider the failure of the book as a work of the historian’s craft. Now, it’s worth noting that the author does say:

I am not a historian, let alone a medievalist or a specialist on the Concordat of Worms. Undoubtedly, readers will find here and there a wrong date or an error in some other detail of the story… [but] the big story being told here is what matters, and the big story does not depend on anecdotes or on any individual fact. Rather, it depends primarily on quantitative evidence, which I have supplemented with anecdotes to illustrate rather than evaluate the argument.

The problem is that this book’s problems don’t come from any individual error of fact, but from a complete failure of approach towards the subject. This is ‘Garbage In’: the approach the author takes to historiography, sources, and analysis inherently preclude this book from reaching any useful conclusions.

Let’s start with the historiographical garbage: Bueno de Mesquita is working with a very old-fashioned view of the medieval past. The bibliography is almost aggressively antiquated, and secondary works are deployed with very little understanding of the sources they’re based on. In a section which struck particularly close to home for me, his account of the pre-eleventh century papacy is based on Mann’s The Lives of the Popes vol. 4: The Popes in the Days of Feudal Anarchy (1910) and a Wikipedia article. (As an aside, there’s an unnerving amount of Wikipedia in this book for what is supposed to be the product of a veteran scholar – in footnote 9 for Chapter 4, for instance, we learn that his information about episcopal careers throughout the whole period he covers comes entirely from the website.) Historical thinking on this matter has changed since World War One as we have developed better understandings of our sources – notably, in this case, that the fundamental basis for a highly negative view of the tenth-century papacy comes from polemical works trying to justify the Ottonian invasion of the 960s. Sometimes, to be sure, a distance from the orthodoxies of mainstream historical scholarship can prove advantageous; but that does require knowing what they are. This is simply repeating the orthodoxies of over a century ago. 

As this suggests, the book’s handling of evidence is also very poor. The author points towards quantitative evidence to support his conclusions – in a word, data – seemingly without realising that this data is based on sources. This is not even a case of failing to perform basic source criticism. Rather, the author systematically does not engage with medieval primary sources. This is true even at the most crucial moments for his arguments. For instance, to show that his game theory logic works, Bueno de Mesquite has to demonstrate a significant correlation after the concordats between secularly inclined bishops and rich dioceses. His data for whether bishops were secularly or religiously inclined comes – as noted above – from reading Wikipedia summaries, rather than from any kind of engagement with the sources for these bishops’ careers or elections. What this means is that he has 2,709 data points where he has no idea what the evidential basis is for that data. Regula magistri is not usually considered an acceptable kind of evidence, let alone regula anyone-can-edit-orum. Similarly, he determines whether or not dioceses were rich or poor by correlating whether or not dioceses were on trade routes and modern estimates of their potential to produce high-calorie crops. Notably missing from any of this is any discussion of medieval evidence. In fact, this particular methodology is both theoretically and empirically nonsense: first, because it assumes the bishop holds a proportionate amount of the wealth produced within the borders of his diocese; second, because it ignores property held by bishoprics outside the borders of the diocese; third and most importantly – again – because it’s untested against the actual economic histories of these bishoprics. In short, the author’s claims rely neither on primary accounts of bishops’ careers, nor on even vaguely contemporary economic data. In addition to not dealing with medieval historians (modern professionals) whilst writing this medieval history book, he has also not dealt with medieval historians (contemporary authors). What is required to prove any of these claims is a systematic, source-based analysis of episcopal elections and medieval economic history. What we have is Wikipedia. 

Probably the worst failure to think about evidence comes when he talks about art – trying to prove the changing relationship of ‘secularism’ and ‘religion’, he scoured four popular art history textbooks for images from the period in question, decided if they had a ‘religious’ or ‘secular’ subject matter, and has the gall to put the results in graphs as though it shows anything about the Middle Ages!

[edit 23/04/22: this comic came up today and seemed very appropriate. (source)]

It does produce the amusing image of Charlemagne’s court as a hotbed of ‘secularism’ – because art history textbooks have a lot of pictures of Charlemagne and Louis the Pious. I also can’t prove this (because – just as in the case of secular/religious bishops and rich/poor dioceses – I can’t find a list of which data points count as what either in the book or on his website), but I did find one of the art history textbooks he used online, and based on the datapoints for the graph in question, I think he’s counting this portrait of Otto III as secular art, which is pretty funny because, y’know, it’s an illustration from a gospel book. So this book doesn’t engage with the evidence from which it is claiming to draw data and as such can only produce inherently flawed conclusions.

This brings us, finally, to the analytical garbage. Because the author does not engage with the sources for the medieval world, and because the secondary works he reads are badly out of date, his assumptions about the period are fundamentally flawed. The distinction between ‘secular’ and ‘religious’ bishops is probably the biggest pile of nonsense here. The definitions Bueno de Mesquite gives of ‘secular’ and ‘religious’ bishops is as follows:

Bishops are classified as religious (i.e., aligned with the pope and church) if their job prior to becoming bishop for the first time was a religious post, such as abbot, monk, hermit, deacon, archdeacon, or priest. Bishops are classified as secular if their prior post was as an agent of the secular authorities, such as court ambassador, chancellor, tutor to a monarch or his children, and the like, or if their biographical information indicates that they were specifically linked to or were suggested as a candidate by the secular ruler over the diocese to which they were appointed.

The whole analysis here rests on the idea that a clerical career = a pro-papal stance = an anti-royal stance (and, the inverse: secular career = pro-royal = anti-papal). This chain of assumptions might look plausible if all you’re reading is The Papacy in the Days of Feudal Anarchy, but it doesn’t stand up to any kind of familiarity with the evidence. Notably, Bueno de Mesquite rarely discusses particular examples. We get a lot of abstract hypothesising (‘a king would be motivated to do such-and-such’…) without any systematic, source-based analysis of what specific people actually did. This has the useful effect of meaning that his assumptions don’t have to be tested against contemporary sources – who needs that when you have models? And this is true again and again – bad assumption after bad assumption, leading to faulty question after faulty question, all supported with worthless ‘data’ but clothed in the trappings of science to lend it an air of false authority. 

Why does this matter? Everyone who reads this book will have their own criticisms (I haven’t come close to covering all the issues – I leave the problems with the framing of the initial question, for instance, for others to deal with) and their own reason it matters, but for me it’s to do with the intellectual value of expertise. This isn’t the same thing as ‘being an historian for a living’. Good history is like good woodwork: you don’t have to be a professional to do it, but you do have to have a grasp of the tools, materials, and techniques. The problem with this book is not that it’s wrong. It’s that it’s so obviously, basically wrong, but presented by a senior academic with such a pretence of authority that its fundamentally flawed approach to the Middle Ages runs a serious risk of being influential, making it concomitantly harder to not be wrong in the future. Good history requires craftsmanship; this has no craftsmanship, and has the potential to devalue craftsmanship as a key attribute of historians.

I don’t like being this negative. This book does have good bits. There’s a really interesting discussion of ruling coalitions in Chapter 2 – turns out the senior political scientist is very good at explaining political science theories. However, if you want that, read a political science book. The Invention of Power has two grains of wheat in a warehouse of chaff. Even worse, the ways in which it’s bad are obvious and boring and mean it can’t possibly be recommended, even to point and laugh at*.

*Except the ‘art historical’ analyses. Those are so amazingly ill-thought out that it genuinely is quite funny.

Revisiting Louis V in Aquitaine I: For Richer, for Poorer

Hello all. Recently I’ve been thinking about a topic this blog has discussed before, Louis V’s kingship in Aquitaine. As so often with the later part of the reign of Lothar, it’s got me scratching my chin, so this week and next week we’ll look at two different aspects of it and hopefully one of the blog’s readers will have something helpful to say! It will help if we’re all singing off the same hymn-sheet, and so I’ve translated all the relevant sources I know of for Louis’ Aquitanian sojourn and put them below in a PDF file*:

This leads us naturally to the first question, which is: have I missed something? If you know any source with a direct bearing on this, please do put it in the comments!

Anyway, as you can see most of what we think we know about Louis’ kingship derives from Richer, and so this post is going to be a hatchet job on his account, because I don’t think it’s at all reliable. Before I say why, it’s probably worth giving a summary of what Richer says:

  1. After the Treaty of Margut, which Lothar signed with Emperor Otto II, Hugh Capet was very angry with the king, and visited Otto in Rome to forge an alliance against him. Discord between Lothar and Hugh went on for some years.
  2. However, they were eventually reconciled.
  3. With Hugh’s backing, Lothar had his son Louis crowned king at Compiègne.
  4. Hugh then commenced to look for a spare kingdom for Louis.
  5. However, some unnamed people decided to take over this task themselves.
  6. They advised Lothar’s wife Queen Emma that Louis should marry Adelaide, recently widowed from Raymond, duke of the Goths.
  7. Thanks to her connections, Louis would be able to win control of Aquitaine.
  8. Thanks to Geoffrey Grisegonelle, count of Anjou, this was achieved.
  9. Adelaide received Lothar and Louis at Vieille-Brioude, where she and Louis were married and she (and maybe Louis but I don’t think that’s grammatically necessary) was crowned queen.
  10. However, from the beginning none of this helped them exercise any real power.
  11. Also, they didn’t like each other because he was young and she was an old woman.
  12. So the marriage fell apart.
  13. Louis, without any councillors, gave himself over to a life of debauchery and lost everything.
  14. Lothar heard about this, came back to Brioude, and picked Louis up.
  15. Adelaide went to Provence and married Count William of Arles.
A manuscript folio of Richer’s autograph (source)

So, the problems with this account fall into three kinds, which we’ll go through in turn: 1) it doesn’t accord with other narrative sources; 2) it doesn’t accord with other non-narrative sources; and 3) it’s internally contradictory. Let’s start with the first point, conflict with other narrative sources; and here I’m just going to lay the contradictions out without trying to resolve them. There are two main contradictions to be had here. First, Ralph Glaber clearly places the marriage in the north, and only puts the visit to Aquitaine after it has already fallen apart. Second, Adhemar of Chabannes doesn’t think Louis and Adelaide actually separated at all, until death actually did them part. (One could put the Annales Sancti Germani minores’ statement that it was Hugh rather than Lothar who brought Louis back, but that’s more an inconsistency than an outright contradiction.)

Our second point, that Richer’s account doesn’t accord with non-narrative sources, is rather more significant, above all because his chronology is wrong. Richer wants Louis’ coronation at Compiègne to be in the early-to-mid 980s, whereas we know it was in 979, well before the 980 Treaty of Margut. He then wants the coronation at Vieille-Brioude to be immediately after Louis’ coronation, rather than (as I think we can say with confidence) in 982, three years later. He then wants Louis’ sojourn to have lasted almost two years, but to have concluded in around autumn 982 – at least, if he actually did think that Otto II’s military defeat in southern Italy happened at around the same time that Louis returned to the north. On another note, Richer also wants Adelaide to be an old woman (anus, not a very nice word), despite the fact that she can be deduced from charter evidence to be relatively young. We know her first marriage took place at such a time that she had teenage children by the mid-970s, but was also still having children in the late 980s. You can push her older, but I think a more plausible chronology is more compressed, putting her date of birth in the late 940s, making her in her mid-thirties in the early 980s – still about twenty years older than Louis, but not exactly the crone Richer wants her to be. The safe conclusion, I think, is that Richer knows basically nothing about the absolute chronology of the events he is describing and not all that much about their relative chronology.

Finally, let’s talk about the internal contradictions in Richer’s narrative. We’ll start with the two smaller holes in Richer’s plot logic: first, Hugh Capet’s concern is supposed to be that co-kingship detracts from royal honour, but this is not a concern in Book IV when Hugh does it himself with his own son; and, second, the original plot is supposed to be against Hugh despite his and Lothar nominally being allied at this time. (This last one may not in fact be a contradiction – but only if it’s a rhetorical construction of Richer’s designed to help his presentation of Lothar as unusually deceitful.) These two are appetisers for the main contradiction, which is that in III.94, Louis’ kingship is a dead letter from the get-go, whilst in III.95 his moral failings and bad decisions turn what had started as a going concern into a failed venture. The takeaway here, I think, is that Richer has no idea why Louis’ kingship failed, or when. It seems to me that Richer is filling in the holes from his limited knowledge with moralistic tropes from other sources. In fact, the parallels between Richer’s account of Louis V in Aquitaine and the Astronomer’s account of Louis the Pious’ kingship in Aquitaine seem striking to me. Like Louis V, the Astronomer’s Louis the Pious was sent into Aquitaine at a young age, and had to beware of ‘foreign customs’ and privation in his domestic affairs (Richer doesn’t borrow a lot of the Astronomer’s language, but mos peregrinorum and res familiaris are shared between the two authors in the same contexts.) Richer’s editor, Hoffmann, believes that Richer did know the Astronomer based on verbal echoes elsewhere, so this is plausible. What Richer’s story reads as, then, is supposition based on historical parallels: Louis the Pious had good instructors, Louis V didn’t; so the latter-day Louis’ kingship becomes a dark mirror of his ancestor’s. (Notably, such concerns were live when Richer was writing: his main practical criticism – that Louis started wearing Aquitanian clothing – is something his contemporaries and those in the generation immediately after him were concerned about in their own times.)

What, then, did Richer actually know? The vague chronology and the account’s place in the narrative makes me think we’re dealing with the chronicler’s own memories. All of this took place approximately twenty years before Richer was writing, and he would have been perhaps in his mid-to-late twenties, or early thirties, at the time. In short, we’re dealing with decades-old memories from someone who was not heavily involved in these events, which is probably why he can remember the feast date of Louis V’s accession but only roughly when it happened. Richer can remember that Geoffrey of Anjou was involved, but not why (he was Adelaide’s brother); similarly, he can remember that something significant happened in Vieille-Brioude but isn’t necessarily clear what. He knows the marriage failed, but doesn’t know why and fills the void with stereotypes; he knows the reign failed, but doesn’t know why and fills the void with moral tropes.

What, then, can we corroborate from other sources? Unsurprisingly, most of the broad outlines. We can be sure that Louis and Adelaide were in fact married and that the marriage was unsuccessful (a datum found in Glaber and, implicitly in the case of the latter half, in Adhemar). We can also say that Lothar visited Brioude (from royal diplomas). We can guess (based on the Fleury charter in the translations) that Louis was set up with a sub-kingdom. We can, somewhat surprisingly, corroborate Richer’s idea that the plan was to use Adelaide’s family connections to win support (Glaber again, although his idea of how is notably fuzzier than Richer’s). We might also be able to say that Louis stayed in Aquitaine and had to be brought home, if we can take the Annales Sancti Germani minores seriously, although the fact that its chronology is also wildly off does not inspire confidence. A two-year sojourn is also likely (assuming Lothar’s 982 diplomas come from the initial journey down; we can place him back in Aquitaine in 983/4 thanks to Adhemar and we know from the letters of Gerbert of Rheims that the two kings were back in the north by early 984 anyway), but a good part of me wants to ascribe that to coincidence. (It is also worth saying, although it doesn’t directly bear on this incident, that Adelaide’s marriages to both Raymond and William are also attested elsewhere).

What can we not corroborate, then; for what is Richer our only source? A surprisingly large amount, from the motivation for the marriage (that the aim was from the start to set up a sub-kingdom, that the plan was aimed against Hugh Capet), to the setting up of Louis’ regime (notably the marriage at Vieille-Brioude and the fact that Adelaide was crowned at all). Above all, we have no idea why Louis’ marriage or his kingship failed – and looking at the context surrounding that will be our next post.

*There are very good extant translations of both Richer and Ralph Glaber which I normally use myself, but for copyright purposes these translations are my own.

(also unspoken through all of this is that Richer is of course a very rich source for mentalités…)