| 英文摘要: | To the Editor —
Academic disputes are often contentious, but 'conflict studies' are especially so1. Raleigh et al. criticize2 our recent papers3, 4, 5 for “anchoring a modern form of environmental determinism”, claiming that our focus on environmental conditions “removes violence from its local, social and political contexts” and that our results imply that “poor people act violently for natural reasons”. Both claims are based on misunderstandings of our work and the methods we employ.
Our findings show that the relationship between extreme temperatures and violence is observed in both rich and poor populations alike5 and we have consistently highlighted the importance of socio-economic settings. For instance, we demonstrated that the effects of global climatic variation on civil conflicts in the tropics are lower, but still positive, for relatively richer countries4. Similarly, in our meta-analysis of the literature we find the magnitude of the effects of climate anomalies on intergroup conflict to be over three times larger than the magnitude of the effects on interpersonal conflict5. This implies that both the local socio-economic context and the type of conflict being examined are important. Even our earliest work studied how the effects of climate on conflict differed across local economic, political, social, geographic and demographic conditions6.
None of this work claims that high temperatures or other climatic variations are necessary or sufficient to trigger violence at any scale. To use an analogy, drunkenness may increase traffic accidents, but not all traffic accidents involve drunk drivers and not all drunk drivers have traffic accidents.
All of us believe that political and economic factors influence the occurrence of conflict. Indeed all of our statistical models acknowledge these factors — as Raleigh and colleagues themselves note. We agree with the sentiment that a complete conceptual model of conflict must include political, social and economic factors in addition to environmental causes. The research community will get there faster by first acknowledging the empirical facts that can be credibly established by data, including our observation that hot weather is related to violence at various scales. |