Master List of Empires
Remy Haring
2026-07-01
R Markdown
This is an R Markdown document. Markdown is a simple formatting syntax for authoring HTML, PDF, and MS Word documents. For more details on using R Markdown see http://rmarkdown.rstudio.com.
When you click the Knit button a document will be generated that includes both content as well as the output of any embedded R code chunks within the document. You can embed an R code chunk like this:
The goal of this project is to analyse the popular claim that the average empire lasts 250 years, which originated from Sir John Bagot Glubb’s essay “The Fate of Empires and the Search for Survival.” I aim to do this by compiling a dataset with every state that I could find reliable information for and calculating the Median. I chose the median instead of the mean because in almost every dataset, there is a strong right skew. On top of this, I do smaller datasets for each geographic region. For datasets over 15, I will visualize it with a histogram and note the IQR in the analysis and for datasets under 15, I will use a box plot and omit the IQR for reliability reasons. I will admit that the definition of empire is somewhat arbitrary and fluid but for the purposes of this project, I defined an empire as a state that rules over multiple distinct peoples, projects military or economic power beyond its core territory, and extracts resources from a periphery to a center. On top of that, I tend to divide empires based on governments (ie, dividing the Russian Empire, Soviet Union and Russian Federation) entries are split when a fundamental change in government occurs, like revolution or conquest as opposed to simple dynastic or policy changes. Finally, for the purposes of accuracy, I will be focusing on states with well documented histories and clear start and end dates.
GLOBAL
#master list of empires: all data
ggplot(df, aes(x = `Years Active`)) +
geom_histogram(bins = nclass.FD(df$`Years Active`), fill="steelblue") +
labs(title = "Distribution of the Lifespan of Empires", x = "Years Active", y = "Count") +
geom_vline(aes(xintercept = median(`Years Active`, na.rm = TRUE)),
color = "red", linewidth = 1) +
theme_minimal()
summary(subset(df)$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3.0 89.5 224.0 270.0 371.8 1406.0Sample size of 212, median of 224, IQR of 282, longest lived entry was Japan at 1406 years, shortest lived was the Peru Bolivian Confederation at 3 years. The initial bump largely consists of short lived, conquest based empires like Napoleonic France, the Timurid Empire or the Mongols. Interestingly enough, the empires around the median come from a wide variety of regions and time periods, including everything from Prussia/German Empire to the Cuman Kipchak confederation. When we examine the tail of the graph, the majority of the empires consisted of trade and economic powers like Kanem Bornu or Champa (the second longest lived empire in fact at 1279 years), empires with highly sophisticated bureaucracies like Dai Viet or the Byzantine Empire. Most fascinatingly, many of these empires had fairly static borders, most notably the Holy Roman Empire or the Kingdom of France. Already we can see that the median empire is far below the 250 mark. Another important thing to note is that the sample size for the 250 year claim was incredibly small at ten entries which were overwhelmingly dominated by European and classical era empires. To contrast, I aimed to include empires from six different continents and from every era of recorded human history. As we will see in the upcoming data, actual median varies extremely wildly depending on region and this is mainly a birds eye view at all the data I collected.
AMERICAS
# The Americas
ggplot(subset(df, Region %in% c("Post Colonial", "Mesoamerica", "Andes")),
aes(x = "Americas", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of American Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region %in% c("Post Colonial", "Andes", "Mesoamerica"))$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3.0 86.5 179.0 299.3 450.5 1200.0Median falls significantly below 250 at 179, sample size of 11, longest lived entry is the Zapotecs at 1200, shortest lived is the Peru Bolivian Confederation. The bottom part of this dataset is dominated by post colonial states like Gran Colombia and Brazil which boasted considerable influence in the region due to the absence of European colonial powers and the United States not yet projecting significant power in the region. The other half is exclusively dominated by pre-colombian states like the Purepecha Empire and League of Mayapan. It is also important to notice that the Aztecs and Inca fall into the lower half of the dataset due to the black swan event that was the arrival of the Spanish, heavily distorting the data.
# Post Colonial America
ggplot(subset(df, Region == "Post Colonial" ),
aes(x = "Post Colonial Empires", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Post Colonial American Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Post Colonial")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3.00 9.75 39.50 52.50 82.25 128.00Incredibly small sample size of 4, smallest median in the dataset at 38 years albeit incredibly skewed by the small sample size, America having the longest lifespan at 128 and the Peru Bolivian Confederation having the shortest lifespan in the dataset at 3 years. Apart from America, all the other datapoints are in South America all have double digit lifespans that mainly existed in the lack of stronger powers inthe region like Spain, Portugal, and later on America. Not to mention all three of these South American Empires didn’t industrialize: either due to not existing long enough to industrialize or, in the case of Brazil, relying on agricultural exports. The rise of the US as an economic and military giant towards the end of the 19th century sealed that vacuum and displaced their status as empires.
# Mesoamerica
ggplot(subset(df, Region == "Mesoamerica" ),
aes(x = "Mesoamerica", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Mesoamerican Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Mesoamerica")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 93.0 230.0 250.0 441.4 434.0 1200.0Sample size of 5, median right at 250, The Aztecs were the shortest at only 93, while the Zapotecs during the Monte Alban period were a major outlier at 1200 years. The Aztecs were militarily dominant but had quite loose control over their tributary states while the Zapotecs were consolidated around the Oaxaca region of modern day Mexico. Given the spotty documentation of these states, I would hesitate from drawing any decisive conclusions about this region. It is also worth noting that several of these empires had their life spans cut short by the arrival of the Spanish, heavily altering the data as it is impossible to know how long these states would have lasted without the arrival of the Spanish.
# Andes
ggplot(subset(df, Region == "Andes" ),
aes(x = "Andes", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Andean Empire", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Andes")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 105.0 302.5 500.0 391.7 535.0 570.0Sample size of 3, median at exactly 500, shortest lived entry was the Inca Empire at 105 years while the longest lived was the Chimu at 570 years, with Wari rounding out at roughly 500 years. The small sample size and wild range in lifespans heavily distort the data. Despite this, it is worth noting that all three of these states shared similarly complex administration, capable of building large scale infrastructure projects and provincial administration. Albeit the arrival of the Spanish was such a black swan event that it cut the Inca Empire short and massively distorted the data.
EUROPE
ggplot(subset(df, Region %in% c("Western Europe", "Southern Europe", "Northern Europe", "Eastern Europe", "Balkans", "Central Europe")),
aes(x = `Years Active`)) +
geom_histogram(binwidth = 75 , fill="steelblue") +
labs(title = "Distribution of the Lifespan of European Empires", x = "Years Active", y = "Count") +
geom_vline(aes(xintercept = median(`Years Active`, na.rm = TRUE)),
color = "red", linewidth = 1) +
theme_minimal()
summary(subset(df, Region %in% c("Western Europe", "Southern Europe", "Northern Europe", "Eastern Europe", "Balkans", "Central Europe"))$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 7.0 68.0 217.0 263.4 362.0 949.0Sample size of 49, longest lived is the Kingdom of France at 949, shortest lived is Nazi Germany at 7, median of 217, IQR is 294. The Bimodal Distribution suggests there are two distinct currents of European imperial history: the peak at around 0-100 year range consists of short term “Flash in the pan” conquests that either sustained themselves on conquest like Napoleonic France and Nazi Germany or emerged in political vacuums like the Serbian Empire. The second peak is around 300-400 range and consists of much more stable empires like Spain, Austria, Portugal and even Rome. The sample size of 50 further corroberates this. Trend is consistant across regions.
Given the fact that European empires are overrepresented in Glubb’s claim of empires lasting 250 years, I think it’s important to point out his sample size primarily consisted of empires in the 300-400 year peak: Rome, the Russian Empire, Spain, etc. A final thing I think is important to point out is how the actual median is 190 with a fuller picture of European empires. Moreover, the median I calculates is significantly below the claim of 250 at 217.
# Western Europe
ggplot(subset(df, Region == "Western Europe" ),
aes(x = "Western Europe", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Western European Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Western Europe")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 8.0 38.5 65.0 265.9 360.8 949.0Highest is the Kingdom of France at 949, lowest is the French Fourth Republic at 8, sample size of 10, the median sits significantly below the 250 mark at 65 due to the significant amount of short lived empires like Burgundy or Napoleonic France. The rest of the bottom two quadrants mainly consist of medieval empires like the Angevin Empire, smaller colonial empires like Belgium, highly unstable states like the French Fourth Republic or states driven by conquest like Napoleonic France. The upper two quartiles consist of more modern colonial empires like Britain, the Dutch, and the latter half of the Kingdom of France, once again showing the diversity of European imperial traditions. Given the spread of how long colonial empires last, I think colonialism functions as a multiplier for the length of empires, but it does not necessarily guarantee longevity.
# Southern Europe
ggplot(subset(df, Region == "Southern Europe" ),
aes(x = "Southern Europe", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Southern European Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Southern Europe")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 60.0 82.0 293.0 272.7 407.0 688.0Sample size of 13 longest lived empire is the Republic of Venice at 688 while the shortest lived is the Ostrogoth Kingdom at 60 and the median is significantly higher than 250 at 293, far above Western Europe’s 65 years. The higher part of the distribution is dominated by merchant republics like Venice and Genoa, and major colonial empires like Spain and Portugal. This trend of maritime and trade based empires continues appearing throughout the data. The trend of short lived conquest empires does persists but has far fewer entries: mainly the Kingdom of Sicily and the Ostrogoths. This can partially be explained by Southern Europe having more insulated geography with mountain ranges like the Pyrenees and Alps insulating the region from invasions.
# Central Europe
ggplot(subset(df, Region == "Central Europe" ),
aes(x = "Central Europe", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Central European Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Central Europe")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 7.0 147.0 229.5 302.2 403.0 844.0Sample size of 8, median sits slightly below the 250 mark, with the longest lived being the Holy Roman Empire at 844 and the lowest being Nazi Germany at 7. The trend of empires purely running on conquest seems to continue with how short lived Nazi Germany and the Hunnic Empire were, while interestingly the empire of the similarly nomadic Avars lasted significantly longer at over a century and had significantly different ways of governance. On the other hand the more dynastic empires like Prussia under the Hohenzollerns, Hungary or Austria under the Habsburgs were a lot closer to the 250 mark and had a long and dynamic history. A commonality between the four empires is that all of them ended in some major war with outside powers, the Hapsburgs and Prussia ended via WWI, the HRE by Napoleon and Hungary by the Ottomans. What’s very fascinating is how the Holy Roman Empire lasted the longest because of its highly decentralized government and having largely static borders throughout its history. However, their decentralization runs counter to a lot of other longer lived empires.
#Northern Europe
ggplot(subset(df, Region == "Northern Europe" ),
aes(x = "Northern Europe", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Northern European Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Northern Europe")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 22.0 88.0 118.0 195.8 225.8 525.0Smallest sample size in Europe at 4, with a median of 118. The longest lived entry was the Norwegian Empire at 525 and the lowest being the North Sea Empire at 22 years. I argue the small sample size is a data point in it of itself because of the sparse population and harsh terrain causing empire building to be more difficult. An interesting thing to note is that two of the four empires, the Kalmar Union and North Sea Empire are actually personal unions, where one person rules multiple separate domains while Sweden was a more standard militaristic and expansionist state, partially due to their succession of warrior kings, while Norway was a predominantly maritime empire, with colonies all over the North Sea.
ggplot(subset(df, Region == "Eastern Europe" ),
aes(x = "Eastern Europe", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Eastern European Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Eastern Europe")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 27.0 60.5 150.0 195.1 339.8 437.0The sample size of 8, median is 196 and the lowest entry is Samo’s Empire at 27 years and the longest being the Russian Empire at 437 years. It is also worth noting the diversity of imperial traditions, Great Moravia and Samo’s Empire were both early Slavic tribal confederations while the bottom two quartiles also included the current Russian Federation and the ideological Soviet Union, unlike Samo’s Empire and Great Moravia, both are highly centralized states. The upper two quartiles included the decentralized Grand Duchy of Lithuania and Kyivan Rus and the centralized Polish Lithuanian Commonwealth and Russian Empire. The prescence of highly centralized and decentralized empires in both halves of the distribution may suggest that the correlation between centralization and longevity isn’t very strong, albeit that’s for another project.
ggplot(subset(df, Region == "Balkans" ),
aes(x = "Balkans", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Balkan Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Balkans")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 164.2 287.0 323.2 377.5 809.0The median for the Balkans is at 287 with the shortest lived being the Serbian Empire at 25 years and the longest lived being the Byzantine Empire at 809 years. It also has a quite small sample size of 6, I argue that this is because the region’s geographic position made it at the crossroads of and frequently dominated by other empires like the Austrian or Ottoman Empires. There are also a few different imperial traditions at work: The classical tradition of Athens and Macedon, the medieval tradition of the Serbs and two Bulgarian Empires and the Byzantine Empire possessing the longest lived bureucratic state in European history.
AFRICA
ggplot(subset(df, Region %in% c("North Africa", "East Africa", "West Africa", "Southern Africa")),
aes(x = `Years Active`)) +
geom_histogram(binwidth = 75 , fill="steelblue") +
labs(title = "Distribution of the Lifespan of African Empires", x = "Years Active", y = "Count") +
geom_vline(aes(xintercept = median(`Years Active`, na.rm = TRUE)),
color = "red", linewidth = 1) +
theme_minimal()
summary(subset(df, Region %in% c("North Africa", "East Africa", "West Africa", "Southern Africa"))$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 63.0 179.0 267.0 359.8 472.0 1202.0Sample size of 33, median of 267, IQR is 293, longest lived entry on the continent and third longest in the dataset is Kanem Bornu at 1202 years, shortest lived entry is the Adal Sultanate at 62 years. Unimodal distribution with a strong right skew. The right skew is dominated by trade based empires like Kanem Bornu, Ghana and Benin, while many of the empires on the lowest end of the distribution had their lifespans cut short due to the arrival of European colonialism like the Zulu, Dahomey or Sokoto Caliphate. As in the European dataset, the trend of short lived conquest empires appears again remarkably consistantly across regions, some notable examples include the Almohad Caliphate, the Songhai Empire and the Adal Sultanate. Despite barely appearing in Glubb’s sample, the median for this dataset is very close to the 250 mark.
#North Africa
ggplot(subset(df, Region == "North Africa" ),
aes(x = "North Africa", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of North African Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "North Africa")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 79.0 201.0 267.0 289.3 417.5 501.0Median of 267, sample size of 11, longest lived entry was the Egyptian New Kingdom at 501 while the shortest lived was the Ayyubids at 89 years. The upper half consists mainly of ancient Nile civilizations like the Old, Middle and New Kingdoms of Egypt and neighboring Kush. All four lasted over 400 years and exerted considerable trade dominance over the Nile and were highly centralized polities. On the other hand, the lower half is dominated by Islamic Empires with the Fatimids, Ayyubids and Mamluks which all emerged in direct succession from one another: Saladin abolished the Fatimid Caliphate and the Mamluks overthrew Saladin. As for the Almohads and Almoravids, they both initially started as tribal confederations based in Morocco that later on became conservative Islamic movements. I think its important to note that the small sample size is primarily due to the sheer longevity of Ancient Egypt as well as domination by foreign empires like the Romans, Persians and later the Ottomans that prevented independent imperial powers from forming. It’s also worth noting that all but three of the entries were based in the Nile River due to the sheer economic power of the valley. On the other hand the rest of North Africa is mostly dominated by the Sahara Desert as well as the Atlas Mountains, which are far less economically productive.
#West Africa
ggplot(subset(df, Region == "West Africa" ),
aes(x = "West Africa", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of West African Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "West Africa")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 99.0 192.0 285.0 446.4 631.5 1202.0Median at 285, sample size of 10, longest lived empire is the aforementioned Kanem Bornu Empire at 1202 years and the shortest lived was the Sokoto Caliphate at 99 years. The upper side of the distribution is heavily dominated by trade empires in the interior of the region like Mali, Ghana and Kanem Bornu who had booming economies due to the trade of gold and salt, as well as their strategic location on the Trans Saharan trade network. Interestingly enough, throughout their lifespans, these empires tended to have rather static borders. The lower half consists predominantly of either conquest based empires like the Sokoto or Songhai Caliphate, or coastal states like Dahomey or Ashanti, which ended with the arrival of European colonialism. This makes geographical sense due to the coastal empires having more exposure and contact with European empires. The more interior empires were geographically more insulated from outside conquest due to the Sahara functioning as a natural barrier.
#East Africa
ggplot(subset(df, Region == "East Africa" ),
aes(x = "East Africa", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of East African Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "East Africa")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 162.0 451.5 581.0 533.5 663.0 810.0Smallest sample size in the region at 4, median of 581, longest lived entry was Axum at 810, shortest lived entry was Adal at 162. Axum and their sucessor Ethiopia dominated the region for most of its recorded history due to their role in the Red Sea trade and defensible terrain in the form of highlands. The other long lived economic empire was the Swahili speaking Kilwa Sultanate, lasting 548 years. The state was a commercial power that traded with Persia, Arabia, India and even East Asia, as well as occupying the wealthy Swahili Coast, once again adding to the trend of economic powers lasting longer than more militaristic empires. The Adal Sultanate emerged as a peer power to the Ethiopians in the 16th century but fell after numerous wars with the Ethiopians and the Oromo Migration.
#Southern Africa
ggplot(subset(df, Region == "Southern Africa" ),
aes(x = "Southern Africa", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan Southern African Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Southern Africa")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 63.0 154.2 201.0 261.9 425.5 489.0Sample size of 8, median of 201, longest lived empire was Luba at 481, shortest was Zulu at 63. The Kongo, Luba and Lunda empires had a sophisticated bureaucracy that combined the divine right of kings, provincial governments, tributaries and trade monopolies. In fact, a lot of Sub Saharan African states operated under some sort of confederation structure like Luba or Maravi, this structure boasted notable longevity and occupies the upper two quartiles of the distribution. The Zulu Empire stands out due to its highly militaristic culture and raids all across Southern Africa only to be cut short by the British. Mutapa was predominantly an economic empire that was based off of the trade of valuable goods like gold and ivory, while also boasting a strong centralized state capable of building large projects. The pattern of decentralized empires boasting considerable longevity seems to continue to hold as well as the pattern of militaristic states having shorter lifespans.
ASIA
ggplot(subset(df, Region %in% c("Western Asia", "Eurasian Steppe", "South Asia", "Southeast Asia", "East Asia", "Oceania" )),
aes(x = `Years Active`)) +
geom_histogram(binwidth = 75 , fill="steelblue") +
labs(title = "Distribution of the Lifespan of Asian Empires", x = "Years Active", y = "Count") +
geom_vline(aes(xintercept = median(`Years Active`, na.rm = TRUE)),
color = "red", linewidth = 1) +
theme_minimal()
summary(subset(df, Region %in% c("West Asia", "Eurasian Steppe", "South Asia", "Southeast Asia", "East Asia", "Oceania"))$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 7.0 85.0 210.0 252.6 319.0 1406.0By far the richest dataset in this entire project, with a sample size of 117, a median of 209 and an IQR of 234. The longest lived empire is Japan at 1406 years (measuring from the beginning of the Asuka period of 539AD with the Yamato Kingship becoming the regional hegemon of the Japanese Archipelago, and ending in 1945 with the end of WWII) while the shortest is the Northern Qi at 7. This massive variation can be explained by numerous factors from geographical diversity to many different imperial traditions from the River Civilizations of the Bronze Age, to Islamic Caliphates, to Chinese imperial bureaucracy. It’s worth noting that despite being the largest sample size, this continent’s average falls considerably short of the 250 year mark. If we look at the graph, we can see that the distribution is quite unimodal and falls sharply in quantity after the 300 mark, almost identical to the first dataset with all the empires combined.
#West Asia
ggplot(subset(df, Region %in% c("Western Asia")),
aes(x = `Years Active`)) +
geom_histogram(binwidth = 50 , fill="steelblue") +
labs(title = "Distribution of the Lifespan of West Asian Empires", x = "Years Active", y = "Count") +
geom_vline(aes(xintercept = median(`Years Active`, na.rm = TRUE)),
color = "red", linewidth = 1) +
theme_minimal()
summary(subset(df, Region %in% c("Western Asia"))$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 11.0 108.0 195.0 225.6 340.0 556.0Sample size of 33, median at 195, IQR of 232, longest lived is the Ottoman Empire at 556 and the shortest lived is the Afsharid Empire at 11. The dataset is partially bimodal as the first peak is significantly larger than the second. In terms of composition, the first peak predominantly consists of Caliphates like the Rashiduns, Abassids and Umayyads, Persian dynasties that emerged in the chaotic period of the Iranian Intermezzo like the Buyids and Saffarids, and Turkic or Mongolic states like the Seljuks, Ilkhanate or Aq Qoyunlu. What this tells us is that the period after the Abbassid caliphate collapsed but before the Ottoman Empire was a highly turbulent era in West Asia, with many different dynasties competing for influence in the region as well as migrations from Central Asia. Interestingly enough, the second, smaller spike predominantly consists of ancient empires like the Hittite and Middle Assyrian Empires, as well as pre Islamic empires in Persia like Elam, the Parthians and Sassanids. One possible reason for this state of affairs is that during the late Bronze Age, the Hittite, Babylonian, Mittani, Assyrian and Egyptian empires all recognized each other as peer powers. While on the surface this may look similar to the competition between the dynasties of the Iranian Intermezzo, I think the key difference is the Bronze Age civilizations did not share the same, unifiying tradition that the dynasties of the Iranian Intermezzo had. Outside of these two fluid and dynamic eras of the regions history, the Ottoman Empire dominated the region for 500 years such to the point that other states did not have the same influence, barring the Safavid Empire. The rivalry between those two empires prevented the Ottoman Empire from conquering Persia. Unlike other regions I’ve covered, whats driving the bimodal distribution is time period rather than type of empire. Ironically, despite Glubbs including many empires from this region into his admittedly short dataset, the overall median is significantly lower than the 250 mark, possibly because he didnt account for the full breadth of history in West Asia and focused on the major empires like the Ottomans or “Arab Empire”.
#Eurasian Steppe
ggplot(subset(df, Region == "Eurasian Steppe" ),
aes(x = "Eurasian Steppe", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Eurasian Steppe Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Eurasian Steppe")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 35.0 52.5 104.0 131.7 179.0 319.0Sample size of 15, median of 104, longest lived was the Khazar Empire at 319, while the shortest lived is the Timurid Empire at 35. The trend of conquest based empires being short lived continues as the Mongols and Timurids, two of the most famous and devastating empires in the region, appear in the lower half of the data and are often centered around a single, skilled, leader: Genghis Khan with the Mongols and Timur with the Timurids. On top of that, both empires quickly collapsed after these leaders die. A pattern across many of the empires in this section is that they fall due to a combination of internal succession disputes and external competition from other tribes. The longevity of any of these empires seems to depend on their ability to supress internal disputes. On the other hand, the Khazars sidestepped this issue, developing a complex system of local governments, tributaries as well as having the state capacity to give merchants travelling along their lands safe passage.
ggplot(subset(df, Region %in% c("South Asia")),
aes(x = `Years Active`)) +
geom_histogram(binwidth = 50 , fill="steelblue") +
labs(title = "Distribution of the Lifespan of South Asian Empires", x = "Years Active", y = "Count") +
geom_vline(aes(xintercept = median(`Years Active`, na.rm = TRUE)),
color = "red", linewidth = 1) +
theme_minimal()
summary(subset(df, Region %in% c("South Asia"))$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 17.0 138.8 228.0 242.8 319.2 550.0Sample size of 26, median of 228, IQR of 180, shortest lived empire was the Sur Empire at 17 and the longest lived was the Indo Scythian Kingdom of 550. Indian history seems to be dominated by a complex interplay between Indian and non Indian states. The histogram itself seems to have a unimodal distribution at around the 200 - 350 mark that largely consists of Indian empires like the Chalukya and Bengal Sultanate. Fascinatingly, most of the data on the higher side of the distribution consists of empires that originate from outside the Indian Subcontinent like the Mughals, Delhi and Kushan Empires while most of the lower side of the distribution are empires that are from the Indian Subcontinent. The Mughals and Delhi Sultanate made heavy use of Persian and Islamic administrative traditions that they introduced to the Indian Subcontinent. Drawing conclusions for governance for this dataset is difficult given how complex and diverse Hindu philosophy on governance is. However, if we look at geography, we can notice that like Europe, the vast majority of these empires were in intense competition over control of the subcontinent and in very close proximity to each other, like the Bahmani Sultanate and Vijayanagara, suggesting that intense competition between states could lead to lower lifespans of.
#Southeast Asia
ggplot(subset(df, Region == "Southeast Asia" ),
aes(x = "Southeast Asia", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Southeast Asian Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Southeast Asia")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 17.0 214.8 309.5 399.9 469.2 1279.0Sample Size of 12, median of 300, far above the 250 mark, longest lived was Champa at 1279 and the second longest lived empire in the dataset, shortest lived is Singhasari at 17. The lower half of the distribution consists of all three of the Burmese empires of Pagan, Taungoo and Kongbaun as well as the unstable Singhasari Kingdom and Majapahit Empire. On the other hand the higher half of the distribution has a mix of both mainland empires like Dai Viet or Ayutthaya and maritime empires like Melayu and Srivijaya. I think this dataset is important because it shows both how long lived these trade empires can be but also the fact it is not an inevitability of longevity. Also the dataset shows that land based empires can also be incredibly long lived, most notably Dai Viet lasting 750 years.
#East Asia
ggplot(subset(df, Region %in% c("East Asia")),
aes(x = `Years Active`)) +
geom_histogram(binwidth = 80 , fill="steelblue") +
labs(title = "Distribution of the Lifespan of East Asian Empires", x = "Years Active", y = "Count") +
geom_vline(aes(xintercept = median(`Years Active`, na.rm = TRUE)),
color = "red", linewidth = 1) +
theme_minimal()
summary(subset(df, Region %in% c("East Asia"))$`Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 7.0 40.0 156.0 260.7 312.0 1406.0Sample size of 31, second lowest median in the Asia section of 156, far lower than 250 years longest lived of this set and the dataset as a whole is Japan at 1406 years, shortest lived is Northern Qi at 7, the second lowest lifespan on the dataset. The histogram itself is highly skewed to the right. The peak of the histogram is near the 100 year mark, this peak is overwhelmingly dominated by various short lived states during the turbulent period of China’s early history. Qin, Northern Wei, Northern Zhou, etc. While these may be breakaway states existing in a tumultuous and fluid political environment, their population, size and intense competition with each other warrant empire status. The further right you go this trend largely declines after the rise of the Tang Dynasty in 618, Chinese dynasties were quite long lived. Apart from the Mongol Yuan Dynasty, which never fully integrated into the Chinese form of governance due to the Mongol rulers being a seperate caste from the rest of Chinese society. looking outside of China, two of the longest empires can be found in the previously mentioned Japan, and the Korean Silla kingdom, lasting 992 years. A possible explanation for this could be that Japan and Korea adopted Confucian administration from China, but not the Mandate of Heaven institution, which incentivizes a culture of revolts when the old dynasty fails.
Oceania
#Oceania
ggplot(subset(df, Region == "Oceania" ),
aes(x = "Oceania", y = `Years Active`)) +
geom_boxplot(fill = "steelblue") +
labs(title = "Distribution of the Lifespan of Oceanian Empires", x = "", y = "Years Active") +
theme_minimal()
summary(subset(df, Region == "Oceania")$ `Years Active`)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 300 300 300 300 300 300The only entry in this section is the Tui Tonga Empire at 300 years, which was a predominantly maritime empire that extracted tribute from Samoa and Fiji. This entry was only included for the sake of a complete dataset. However, I think the sheer lack of samples is an important data point in it of itself due to this region’s being mostly small islands and the lack of written records. This is not to discount oral traditions entirely, but for the purpose of a statistical analysis they arent reliable enough to include in a dataset.
FINAL THOUGHTS AND CRITIQUE I started this project with the aim to test the famous claim made by Sir John Glubb that the average empire lasts only 250 years using a holistic dataset that includes as many entries as I could find data for. However, when I look at Glubb’s original claim, it is filled with glaring flaws. Beyond its small sample size, five out of the 11 empires are from the classical era, while another five are from the early modern era. Beyond that, some of the end dates make little sense. For example, the Spanish Empire ends at 1750. This doesn’t make sense because at this point, Spain still had a large colonial empire and was going under administrative reforms. Another example is the Ottoman Empire, which Glubb says ended in 1570, this also does not make sense because this was when the Ottoman Empire launched an ultimately successful invasion of Cyprus. Another way Glubb moves the goalpost is by only counting the Romanov era of the Russian Empire, completely neglecting the history of the Rurkid Dynasty or the Tsardom of Russia. Another strange choice for the sample was lumping up the Rashidun, Umayyad, and Abassid Caliphates as simply the “Arab Empire” when in reality all three of these states were different entities that emerged from each other’s collapse. To be fair, some of these choices and dates make sense, the Acheamenid Empire, Mamluks, Roman Republic and the Neo Assyrian Empire all have fairly reasonable start and end dates. But I think that between the small sample size, cherry picked dates and merging several different political entities make it unusable for serious statistical analysis. As for my own findings, the median empire lasts 224 years, which is not wildly different from Glubb’s findings. And, to be fair, I do have my own statistical blind spots. Mainly because I didn’t include some Precolumbian or Oceanic states due to lack of documentation. All in all, I think the lesson here is that even if your findings may have some merit, what actually matters is a robust methodology, sample size and analysis.