Abstract

Lynch et al., in an article in this issue, argue that an entire microbiome is rarely, if ever, the right target of analysis for causal explanations in microbiome research. They argue, using interventionist criteria of proportionality, specificity and stability, for restricting causal claims to the smallest subset of microbes—a causal core—that generate the effect of interest. A further question remains: what kind of interactions generate a consortium of microbes that can operate as causal agents in this manner? Here I introduce two possible kinds of such consortia: ‘trojan horses’ and ‘syntrophic’ individuals.