The Lives of Empires

Wednesday, June 18th, 2014

Sir John Glubb studied the lives of empires:

If we desire to ascertain the laws which govern the rise and fall of empires, the obvious course is to investigate the imperial experiments recorded in history, and to endeavour to deduce from them any lessons which seem to be applicable to them all.

The word ‘empire’, by association with the British Empire, is visualised by some people as an organisation consisting of a home-country in Europe and ‘colonies’ in other continents. In this essay, the term ‘empire’ is used to signify a great power, often called today a superpower. Most of the empires in history have been large landblocks, almost without overseas possessions.

We possess a considerable amount of information on many empires recorded in history, and of their vicissitudes and the lengths of their lives, for example:

Glubb Lives of Empires

This list calls for certain comments.

(1) The present writer is exploring the facts, not trying to prove anything. The dates given are largely arbitrary. Empires do not usually begin or end on a certain date. There is normally a gradual period of expansion and then a period of decline. The resemblance in the duration of these great powers may be queried. Human affairs are subject to many chances, and it is not to be expected that they could be calculated with mathematical accuracy.

(2) Nevertheless, it is suggested that there is suf?cient resemblance between the life periods of these different empires to justify further study.

(3) The division of Rome into two periods may be thought unwarranted. The ?rst, or republican, period dates from the time when Rome became the mistress of Italy, and ends with the accession of Augustus. The imperial period extends from the accession of Augustus to the death of Marcus Aurelius. It is true that the empire survived nominally for more than a century after this date, but it did so in constant confusion, rebellions, civil wars and barbarian invasions.

(4) Not all empires endured for their full lifespan. The Babylonian Empire of Nebuchadnezzar, for example, was overthrown by Cyrus, after a life duration of only some seventy-four years.

(5) An interesting deduction from the ?gures seems to be that the duration of empires does not depend on the speed of travel or the nature of weapons. The Assyrians marched on foot and fought with spears and bow and arrows. The British used artillery, railways and ocean-going ships. Yet the two empires lasted for approximately the same periods.

There is a tendency nowadays to say that this is the jet-age, and consequently there is nothing for us to learn from past empires. Such an attitude seems to be erroneous.

(6) It is tempting to compare the lives of empires with those of human beings. We may choose a ?gure and say that the average life of a human being is seventy years. Not all human beings live exactly seventy years. Some die in infancy, others are killed in accidents in middle life, some survive to the age of eighty or ninety. Nevertheless, in spite of such exceptions, we are justi?ed in saying that seventy years is a fair estimate of the average person’s expectation of life.

(7) We may perhaps at this stage be allowed to draw certain conclusions:

(a) In spite of the accidents of fortune, and the apparent circumstances of the human race at different epochs, the periods of duration of different empires at varied epochs show a remarkable similarity.

(b) Immense changes in the technology of transport or in methods of warfare do not seem to affect the life-expectation of an empire.

(c) The changes in the technology of transport and of war have, however, affected the shape of empires. The Assyrians, marching on foot, could only conquer their neighbours, who were accessible by land — the Medes, the Babylonians, the Persians and the Egyptians.

The British, making use of ocean-going ships, conquered many countries and subcontinents, which were accessible to them by water — North America, India, South Africa, Australia and New Zealand — but they never succeeded in conquering their neighbours, France, Germany and Spain.

But, although the shapes of the Assyrian and the British Empires were entirely different, both lasted about the same length of time.


  1. AAB says:

    He’s right, those dates are really arbitrary.

    The Ottoman Empire lasted up until the 20th century.

    The British Empire started way back in the Medieval period when the various Anglo-Norman monarchs tried to colonise Wales, Scotland and Ireland. Ireland was continually treated like a colony from the time of the plantations onwards.

  2. Rollory says:

    The British example is more defensible than the Ottoman one. England was an offshore backwater as far as Europe was concerned until about the 1600s, at which point the privateers started really making their mark raiding Spanish shipping. (Yes yes yes the 100 Years’ War but that was 1) not a successful imperial venture, 2) a dynastic question of who would rule France, not one of England establishing hegemony over multiple neighbors). The conquests of Ireland and Scotland were not at all the same scale as the sort of accomplishments Glubb is talking about.

    One could argue whether distinguishing the Ottomans from the Seljuks is appropriate in world-historical terms. It was the Turks in general who reduced the Eastern Empire to a shadow of its former self; by the time the Ottomans got started most of the work had been done. (Reading Gibbon’s Decline and Fall, the overall pattern that presented itself to me for the Eastern Empire was: 300 years of losing ground to the Arabs, 300 years of moderate recovery under Basil and his successors, then 300 years of final collapse to the Turks.)

    If one does treat the Ottomans as separate it would be more appropriate to start their ascension at around 1400, and the final decline in the 1800s, which gives them 400-500 years of imperial behavior.

    Similarly the Roman period is indeed arbitrarily split, and cut short; Aurelian and Diocletian were masters of a unified imperial world which a subject of Augustus would have found largely familiar, and a contemporary of Marius would not have found incomprehensible. The civilization-ending crises didn’t really reach fruition until Valens and Valentinian, late in the 300s.

    All that said, the basic pattern of behavior and of the temper and character of the peoples of the various empires that he describes is worth considering and retains applicability, even if the timescales aren’t consistent.

  3. Alrenous says:

    No mention of Byzantium, skepticism rods enable. Conservatively estimated, it lasted for 800 years.

    Not to say there’s no pattern here. I just think we could learn at least as much from the outlier.

  4. Rollory says:

    One could make the argument that Glubb’s overall argument is a descriptive one, of a recurring pattern of human behavior approximately 250 years in wavelength. With that in mind, the Roman and Ottoman examples can be viewed as macroscopical waves composed of ones on the scale he is talking about. The triple 300-year pattern I mentioned above for the Eastern Empire would seem to fit that.

    There is the risk of excessively cherry-picking data and making it fit the conclusions one is seeking. But I don’t see a rigorous way to measure that. You just have to look at it and decide if it is a real pattern or a Rorschach effect.

  5. T. Greer says:

    Glubb seems to be attempting the same thing that Peter Turchin did in War and Peace and War, but in a less precise and more sloppy fashion.

    Turchin makes the point in above that we really should not talk of the ‘rise and fall’ of the Roman empire, but the ‘rises and falls’ of the Roman empire. The empire was structurally different enough after each one of these ‘falls’ and ‘rises’ to justify splitting them up for historical analysis (a comparison could be made with Chinese transition from the Qin to the Han, starkly divided in our minds, but really less different from each other than the Romans of Antonius and the Romans of the Tetrarchy).

    Taking his same standards to the east:

    Kingdom/Dynasty of Qin – 340 BC-206 BC (134y)
    W. Han Dynasty – 206 BC-9 AD (215y)
    E. Han Dynasty – 22AD-189 AD (167y)
    Tang Dynasty, mark I – 618-755 (137y)
    Tang Dynasty, mark II – 763-884 (121y)
    N. Song Dynasty – 960-1127 (167y)
    S. Song Dynasty – 1128-1276 (148y)
    Jin Dynasty – 1115-1234 (106y)
    Ming Dynasty – 1370-1630 (260y)
    Qing Dynasty – 1681-1850 (169y)

    The average for these Chinese empires is about 60-80 years less than the European/Middle Eastern ones he lists.

    Not familiar with India’s empires to do the same calculations off my head.

  6. Kangxi says:

    T. Greer:

    Ming Dynasty – 1370-1630 (260y)
    Qing Dynasty – 1681-1850 (169y)

    Why did you pick those end dates? Usually I’ve seen:

    Ming – 13?? – 1644
    Qing – 1644 – ~1910

  7. T. Greer says:


    Glub disqualifies the Roman empire after the 3rd century because it is wracked by civil war and chaos, so I figure same standard should apply here. In the case of the Qing, that means starting at the end of the Three Feudatory revolts and ending with the Taiping Rebellion. Same story with the Ming — cut off the periods of chaos at the beginning and the end.

Leave a Reply