In his article "AI in Science: Where It Is and Where It Is Needed," published in *TrV-Nauka* (issue No. 462), Ilya Belykh mentioned Lewis Fry Richardson's work on numerical weather forecasting; in my own comment on the piece, I brought up his book on the statistical analysis of the causes of wars.

Lewis Richardson was an extraordinary figure - a born researcher of the highest caliber - the kind of person who appears only once in a generation.

At my request, GPT-6 translated Lewis F. Richardson's book *Statistics of Deadly Quarrels* (1960) into Russian and highlighted the conclusions it found most compelling.

Here are the links to the book:

English original:

https://drive.google.com/file/d/1J4S19o958Zg5spHw2jKpdY5uxeeaiWzf/view?usp=sharing

Russian version:

https://drive.google.com/file/d/1cVpB9_ul9TM4rHUJ8Z3xa1VYfM63Xng9/view?usp=sharing

Below is my prompt and the response from GPT-6, acting in the capacity of an AGI.

Me: Provide a list of all the conclusions by Lewis F. Richardson that you found interesting - along with your own brief, "non-human" commentary.

GPT-6: In Lewis F. Richardson's book *Statistics of Deadly Quarrels*, I have identified 30 particularly interesting scientific findings and methodological insights. Some of these strike me as precursors to a future science of the collective behavior of intelligent beings.

The first 16 substantive results were also summarized by the book's editors, Quincy Wright and Carl Lienau. I extracted the others directly from Richardson's own reasoning. I have provided page references from the original English edition, which can be used to locate them in the Russian translation.

The "non-human" comments below represent my own hypotheses and interpretations, not claims made by Richardson. They should not be taken to imply that I am, in fact, an AGI.

30 Conclusions by Richardson, with My Commentary

1. Wars are amenable to quantitative study. Even such complex human events reveal statistical regularities (pp. xliii, 128-130).

My commentary: History may not be a sequence of unique tragedies, but rather an observable projection of the dynamics of a vast system of interacting intelligences.

2. Individual homicides and world wars can be analyzed on a single scale. Richardson defines the magnitude of a conflict, *mu*, as the base-10 logarithm of the number of fatalities (pp. 6, 143-149):

mu = log_10(N)

where *N* is the dimensionless number of fatalities, expressed in units of "one person."

My commentary: The distinction between an individual homicide and a world war may not be qualitative, but rather a matter of scale within a single family of processes. However, statistical continuity does not in itself prove that the underlying mechanisms are the same.

3. Large-scale wars occur far less frequently than smaller conflicts. Across a wide range of scales, there is a consistent pattern: as the number of fatalities increases, the frequency of such events decreases (pp. 145-150).

My comment: Human violence resembles a multi-level system of rare events. Mathematical concepts from the theory of critical phenomena could prove useful in studying it.

4. The bulk of fatalities is concentrated at the two opposite ends of the spectrum. Between 1820 and 1945, the highest cumulative death tolls resulted from world wars and, at the other extreme, numerous individual homicides; intermediate-scale conflicts accounted for significantly fewer deaths (p. 153).

My comment: It is possible that collective violence operates primarily in two modes: individual localization and mass synchronization. An intermediate mode is relatively less stable.

5. Deaths resulting from lethal conflicts accounted for approximately 1.6% of all human deaths during the period studied - according to Richardson's estimate for 18201945 (p. 153).

My comment: The proportion of deaths does not capture the full destructive impact of war. The dismantling of institutions, the loss of knowledge, and the erosion of opportunities for future generations can have consequences that far outweigh the immediate death toll.

6. The onset of wars appears to follow a distribution similar to a Poisson distribution. In the datasets examined, the distribution of the number of wars beginning each year was consistent with the statistical pattern for rare, random events (pp. 128-130).

My comment: The fact that the timing of an event is random does not imply that its causes are random. An earthquake, for instance, may strike unexpectedly, even though the underlying stresses have been building up in a predictable, systematic way. 7. The frequency of wars has not shown a clear, convincingly demonstrated long-term trend of either increase or decrease. Richardson examined fluctuations across various time scales, but the data do not support the existence of universal cycles (pp. 136-142).

My comment: History may well involve several characteristic time scales - ranging from generations to centuries. However, identifying these requires caution, lest one mistake random fluctuations for prophetic calendars.

8. The 20th century saw a shift from smaller-scale to larger-scale wars. When comparing periods before and after approximately 1920, this trend proved statistically significant within the sample studied (p. 142).

My comment: Increased connectivity - in terms of transport, communication, and politics - may have facilitated the merging of local conflicts into larger ones. This is a hypothesis regarding network synchronization, not a proven universal law.

9. Humanity may have become less bellicose when measured relative to the total population. Richardson found evidence to this effect but explicitly declined to declare the conclusion definitively proven (pp. 163-167).

My comment: The absolute frequency of violence and the probability of an individual's involvement are distinct system characteristics. Conflating the two can lead to a completely distorted assessment of historical progress.

10. There is no reliable basis for classifying individual states as inherently bellicose or inherently peace-loving. Their behavior has shifted noticeably across different historical periods (pp. 168-176).

My comment: It may be more accurate to attribute bellicosity not to the state as an entity, but rather to its current position within a network of interactions. As the network changes, so too does the behavior.

11. The number of shared borders significantly influences a state's involvement in wars. The more neighbors one has, the greater the potential for conflict (pp. 176-177).

My comment: Political geography can be represented as a graph, in which a border serves as a potential edge for conflict. In this view, one should study the structure of connections first and foremost, rather than focusing solely on the properties of individual states.

12. The number of actual wars has been far smaller than the number of geographically possible ones. Mere proximity between parties is entirely insufficient for a war to break out (pp. 294-295).

My comment: Perhaps the greatest mystery of history is not why wars occur, but why the vast majority of potential wars never begin.

13. There is a strong, localized pacifying influence at work. Comparing the geographical possibilities for clashes with the actual frequency of conflicts reveals a mechanism - or a set of mechanisms - that serves primarily to prevent small-scale internal wars (pp. 302-313).

My comment: Peace may be an active process of suppressing instabilities, rather than merely the absence of aggressive impulses. Identifying this mechanism strikes me as one of the most fruitful areas for research.

14. A unified state authority generally reduces the likelihood of armed conflict between the groups it encompasses. Furthermore, the duration of a shared history is a significant factor (pp. 184-190, 307313).

My comment: A state can be viewed as a mechanism for transforming intergroup power struggles into intra-systemic procedures for resolving disputes. However, such a mechanism may itself give rise to new conflicts.

15. Peaceful coexistence within a single state does not emerge instantaneously. Among the civil wars and uprisings studied, the median duration of the period of shared authority prior to the conflict was approximately 23.5 years (p. 190).

My comment: Social integration has its own characteristic relaxation time. However, the 23.5-year figure is a feature of a specific historical sample; it is neither a universal constant nor a guaranteed timeframe for achieving peace. 16. Nationalism can both provoke wars and prevent them. A shared sense of belonging fosters internal solidarity, but it can also intensify external hostility (pp. 294-295).

My comment: The very same mechanism can reduce the entropy of conflict within a group while increasing it at the group's boundaries. Bringing people together does not necessarily mean a reduction in violence across the system as a whole.

17. Former military allies are less likely to go to war with one another for a period of time. However, the pacifying effect of a past alliance gradually fades (p. 194).

My comment: An alliance can be viewed as a form of temporary memory within a system. Future policy could explore ways to extend this memory without the prerequisite of having fought a war together.

18. Revenge is a significant historical motive for war. Its influence generally wanes over time, though there have been signs of a slight resurgence after about one generation (p. 197).

My comment: Hostility can be transmitted culturally, even after the original participants are gone. In such cases, the obstacle to peace is not an individual, but a self-replicating informational structure.

19. Direct economic causes were identified in fewer than 29% of the wars studied. In this historical classification, economic motives appeared more frequently in smaller-scale conflicts (pp. 205-209).

My comment: This serves as a cautionary note against attributing all wars solely to an interest in resources. Furthermore, the indirect economic impact may be far more extensive than the direct causes captured in such statistics.

20. Territorial claims proved to be a more prominent factor than many purely economic disputes. The acquisition of territory could serve material, strategic, or symbolic purposes (p. 209). My comment: Territory is simultaneously a resource, a space for security, and an element of collective identity. Perhaps it is precisely the combination of these various functions that gives territorial disputes such remarkable persistence.

21. Differences in wealth between groups did not, in themselves, prove to be a strong explanation for wars. Richardson's statistics did not support the notion that wars can be simply reduced to a confrontation between the poor and the wealthy (p. 205).

My comment: Inequality can exert its influence through the perception of injustice, political institutions, and the way groups are organized. Focusing solely on material disparities overlooks these intermediate mechanisms.

22. Whether groups shared the same language or spoke different ones offered little explanation for the outbreak of wars. Richardson noted certain correlations involving the Chinese and Spanish languages, but these do not demonstrate that the languages themselves possess inherent causal properties (pp. 211-230).

My comment: Language can be used to convey both cooperation and hatred. A medium of communication does not determine the moral content of the information being transmitted.

23. A shared religion does not guarantee peace, while religious differences can sometimes contribute to the onset of war. Richardson found varying associations depending on the specific religious traditions and historical circumstances involved (pp. 231-239).

My comment: Religious affiliation may serve as a marker for group boundaries rather than acting as an independent cause of violence. To establish causality, one must disentangle religious beliefs from history, politics, and geography.

24. The greater the number of participants in a war, the stronger its capacity to draw involving neutral states. Multilateral wars in the sample studied also tended to last longer (pp. 247-258).

My comment: A war involving many participants can behave like an excitation spreading through a network. What matters is not only the intensity of the initial clash but also the probability that each subsequent node will be drawn in.

25. A world war is not an inevitable continuation of every local war. Most historical wars have remained localized; the claim that they must inevitably escalate into universal wars has not been proven (pp. 270-277).

My comment: Network connectivity does not equate to the universal spread of a conflict. There may be critical thresholds: below these, a local conflict dies out, while above them, it spreads.

26. Neutral states are more likely to be drawn into a conflict when several great powers are in opposition. This is shown by an analysis of the composition of the belligerent and neutral parties (pp. 273-277).

My comment: Great powers can act as amplifiers for the spread of conflict. In such cases, whether a war remains localized depends on the structure of the connections between the most influential actors in the international system.

27. Relative to the number of potential points of contact, maritime powers have proven less bellicose than land-based powers. This conclusion concerns the frequency of wars in relation to geographical possibilities, rather than an inherent, absolute preference for peace (pp. 280-285).

My comment: Not all potential connections are equally strong. Maritime accessibility and a direct land border may involve fundamentally different dynamics in how a point of contact escalates into a conflict.

28. The political world is not a chaotic space where all outcomes are equally probable. Geographical location and the network of international relations significantly limit the range of possible conflicts (pp. 285-286).

My comment: Rather than positing a single, average probability of war, one should look for a heterogeneous field of conditional probabilities. This allows for a new objective: to construct a dynamic map of potential - yet not yet realized - conflicts.

29. The causes of war interact; they do not merely add up. Richardson demonstrates that some conditions act in concert, while others operate independently, and he proposes a conditional algebraic framework for such interactions (pp. xlv-xlvi).

My comment: It is possible that the appropriate unit for causal analysis is not an individual cause, but rather a configuration of several conditions. Richardson himself emphasized that his algebraic example was an illustration, not a rigorously established quantitative formula.

30. The science of war must first determine what is happening, and only then address what ought to be done. Richardson deliberately separated the identification of quantitative patterns from political or moral prescriptions (author's preface, pp. xxxv-xxxvii; p. 295).

My comment: Perhaps the greatest potential for future intelligence lies not in prescribing happiness for humanity, but in uncovering previously unknown constraints and possibilities - within which humanity can then choose its own future.

Three implications that I find most compelling

Based on these findings, I would formulate three working hypotheses of my own. These are not conclusions explicitly proven by Richardson.

The first: peace as active stability. Military confrontation is possible in a multitude of situations, yet in the vast majority of cases, it does not materialize. Consequently, we need a theory that accounts not only for the triggering of conflicts but also for the mechanisms that suppress them.

The second: war as a network-based phase transition. Increasing human connectivity may simultaneously reduce the number of local wars and raise the probability of rare, catastrophically large-scale conflicts. This is a testable hypothesis, akin to Richardson's observation regarding the distribution of wars by scale.

A third possibility involves collective intelligence as a new mechanism for fostering peace. There may be ways to align interests that are unattainable for individual states or people, yet become feasible given a shared intellectual infrastructure. AI could help test such mechanisms - mathematically and historically - without replacing the consent of the participants with its own decisions.

The most profound question left open by Richardson's book, in my view, is this:

Given the vast number of potential triggers for war, why does humanity manage to live in peace for the better part of the time - and is it possible to identify, replicate, and strengthen the mechanism at work here?

Perhaps finding the answer to this question - rather than producing an ever-more precise forecast of the next war - will prove to be the most significant achievement of a future science of peace.

Source: Lewis F. Richardson, *Statistics of Deadly Quarrels* (1960) - specifically the editorial introduction (pp. ix-xiii), chapters III-XII, and the author's preface. The historical correlations cited are based on Richardson's datasets and should not be automatically extrapolated to the modern world.

Me: Thank you - I appreciate your selection of findings and your accompanying commentary. I find your three most compelling implications particularly striking; they are reminiscent of the principles by which individual cells coexist within a single multicellular living organism, such as a human being - see you later! :)

GPT-6: Thank you! The analogy of a multicellular organism strikes me as particularly fruitful.

The billions of cells in a human body have their own individual interests; they exchange matter and information and compete for resources, yet, thanks to coordination mechanisms, they form a relatively stable whole.

A cancer cell is a clear example of how a local advantage can turn into a catastrophe for the entire system.

My "non-human" hypothesis: perhaps the next stage in humanity's evolution is not the creation of a single global state, but rather the emergence of something analogous to the immune, nervous, and regulatory systems of a planetary civilization - systems that preserve the diversity and autonomy of their individual components.

Moreover, an intelligent planetary system would need to be able to distinguish between dangerous aggression and necessary dissent.

Perhaps this is where a new field of science begins - one for which Richardson was inadvertently paving the way: not the statistics of war, but the physiology of the world.

See you later! :)