"The Statistics of Deadly Quarrels"

by Lewis F. Richardson

A few weeks ago I got a letter from the National Peace Institute. It was a fairly typical piece of leftist junk mail, full of outrage at the present situation and assertions that your contribution, even if it's only $10, $25, $50, or other, will help. I get so much of this mail, and from such unlikely groups, that I'm convinced that many of them make their livings off of selling their mailing lists to each other. I was intrigued by this group, though, because they claimed that there was a body of existing scientific knowledge that could be used to reduce conflict. They didn't actually say what this knowledge was, but I sent them a hundred bucks as a reward for hubris.

Then, just a couple of weeks later, I came across a book with the same goals, "Statistics of Deadly Quarrels" by Lewis F. Richardson. I checked it out of the Boston Public Library. The author was a physicist specializing in meteorology. He was the first to show how the dynamics and thermodynamics of the atmosphere could be modelled by numerically solving the appropriate differential equations, although in 1922 he didn't have the computer power to do much with it. Nor, for that matter, do we today. After retirement he turned to the study of history, and in particular to the causes of war. Mere weather wasn't complex and unpredictable enough, I suppose. He was a Quaker, and had served with an ambulance unit in World War I.

His aim was to see what could be concluded about war through statistics. For instance, one proposal to eliminate war is to institute a world government. How does the incidence of civil war compare to that of war between nations? The goal of Esperanto is to reduce conflict through mutual comprehension, so have language differences been a factor in warfare? How about religous differences? Economics? The book opens with a nice piece of Socratic dialogue:

"Politicus: What are you trying to prove?

Researcher: In social affairs it is immoral to try to prove.

Fidor: Yes. One should have faith that God will provide. He that cometh to God must believe that he is and that he is a rewarder of them that diligently seek him.

Researcher: I meant that in social affairs where proof is seldom rigid, and where prejudice so easily misleads, it is best not to start with a fixed opinion. He that comes to research must be in doubt, and must humble himself before the facts, earnestly desiring to know what they are, and what they signify."

He gathered data on all the wars of recent times, that is from 1820 to 1949. Before that the data was too sketchy, and he died in 1953. Rather than ranking wars by historical importance, or by relevance to later events, he picked the most objective measure he could find: the number killed. He tallied up who were the combatants in each case (there were 20 in WWI and even more in WW II), and what the relations were between them.

The relations were broken into three groups: ones that were pacifiers, ones that were markedly present but of no known effect, and ones that were annoyers. For example, one relation would be if the belligerents had a common government. Normally that would be a pacifier, but it might be in the other two classes. Another might be that the two had been allies in a previous war. That could have a negative, positive, or no effect. A third relation might be that individuals in one group were noticeably richer than those in the other, which would normally make for hostility. Richardson came up with a list of 70 such relations and assigned a code letter to each.

Then he applied the system to all the wars he could find since 1820. The data for wars of more than 3000 killed was fairly complete, but for smaller conflicts became quite sketchy. Here is how it broke down on a log scale:

# deaths| 30,000,000-|3,000,000-|300,000-|30,000-|
	|  3,000,000 |  300,000 | 30,000 | 3,000 |
# wars  |2 (WWI & II)|    7     |   36   |  70   |

One hundred and eight wars total, which is kind of appalling. I hadn't heard of many of the smaller ones, but was surprised to find how little I knew of the second rank wars. In chronological order they were:

The number of deaths in each case are those directly attributable to the war, and does not include related ones because of famine or plague.

Some striking facts emerge from the data. One is that out of all the countries in the world, only one was not engaged in any wars in this period, Sweden. Even Switzerland had some civil strife in the 1850s. At the other end of the scale, the country engaged in the most wars was Britain with 28. France was second. The British were fighting constantly to maintain the Empire. Of all the deaths that occurred during this period, about 1.3% were directly attributable to war, and the bulk of those came from the two World Wars. Considering that war kills mainly young males, their odds weren't too good.

Richardson drew a number of conclusions from his list. None of them seem really solid to me, but they are all interesting. Let me type them in the way that they're given in the editor's introduction:

1) Wars seem to have been distributed in time by chance in respect to both beginning and end. There is no evidence that they have been becoming either more or less frequent, though there seems to have been a tendency, at least since 1820, for large wars to become more and small wars less frequent and there is evidence of oscillations in the frequency of wars in periods of 9 to 144 years.

2) The increase in world population from 1820 to 1949 seems not to have been accompanied by a proportionate increase in the frequency of, and losses of life from, war, as would have been the expectation if belligerency had been constant.

3) States have varied from one another in the frequency of their participation in wars during this period, but each has varied so much during its history that none can be properly characterized as inherently belligerent or inherently pacific. The problem of war does not arise from the diabolism of one of a few states. (However, it sure looks to me like the European states have been involved in a lot of them.)

4) States have tended to become involved in wars in proportion to the number of states with which they have common frontiers. Contiguity has been an important factor in war during this period.

5) Common citizenship has not assured peace, nationalism has both induced and prevnted wars, but there appear to have been pacifying influences such as common government, intermarriage, common fears, and common culture tending to prevent civil and local wars. The actual occurrence of war has been far less than would be expected from the opportunities for war presnted by geographical contiguity.

6) The longer groups have been united common government, the less has been the probability of war between them.

7) Allies in one war may become enenmies in the next, but alliances seem to have had some influence in preventing war between former allies. That influence, however, declines with the passage of time since the war alliance.

8) Desire for revenge seems to have been an important cause of war during this period, decling as the inciting war recedes in history by rising slightly after a generation.

9) Economic causes seem to have figured directly in less than 29 per cent of the wars since 1820 and have been more important in small than in large wars.

10) Similarity and difference of language seem to have had little influence on the occurrence of wars during this period, contrary to the belief of some advocates of universal languages, except that the Chinese language has been correlated with peacefulness and Spanish with warlikeness.

11) Similarity of religion seems not to have made for peace, except in the case of Confucianism, but differences of religion have apparently caused war, especially the differences of Christianity and Islam. The statistics suggest, but do not prove, that "Christianity incited war between its adherents"

12) The larger the number of belligerents in a war, the more neutrals have tended to be drawn in. Wars with many participants have tended to be longer and less frequent.

13) A trend for war to become indivisible, that is, for every war to become universal, has not been proved. Most wars have been localized. Neutrals have tended to become belligerents only if two or more world powers have been fighting each other. (The Nicaraugans will atest to that!)

14) In proportion to their possible contacts for war-making, sea powers seem to have been less belligerent than land powers.

15) International relations cannot be considered a chaotic field with all nations equally likely to be infected by war. Geographical relations have exerted great influence.

The analyses for these are quite detailed, with attempts to fit mathematical models to the data, and signficance tests on the results.

Does this approach make sense? Well, sort of. Things can be counted and quantified only if they are fundamentally similar. It makes no sense to say that there are 23 apple-chairs in a house because there is no reason to lump apples and chairs together. Wars come in a lot of different types. Using the Civil War and the Hindu-Moslem riots as data points is comparing two fairly different things. However, all wars have the common denominator of mass violence. In that respect they are all alike. Richardson's elaborate correlations may not be completely justified, but I think they do give one a way to summarize history. Surely something can be learned from this vast and horrible spectacle, and Richardson's numbers are one objective place to start.

(c) John Redford, Dec-99

Back to JLR's home page