Recognition of cultural diversity has led to an increased recognition of the contributions that various knowledge sources can offer in understanding health issues. These diverse perspectives can enhance our approach to defining and promoting health. The challenge is how to establish an adequate dialogue between them. We use fuzzy cognitive mapping (FCM) for this purpose. A fuzzy cognitive map is a graph that represents causal relationships between factors (nodes) connected by arrows indicating the direction of the causal relationship. When we work with communities, the map's authors indicate the factors they consider affect the occurrence of a particular health issue. Then, they indicate how these effects work by drawing arrows from each cause to its outcomes. The arrows receive weights from one for the weakest causal relationship to five for the strongest. Weights are positive if the relationship is direct (an increase in the cause increases the outcome) or negative if the relationship is inverse. We use four approaches for the analysis of maps, starting with pattern matching tables to compare their content. Using fuzzy transitive closure we calculate the maximum influence of a factor on others through direct and indirect connections. Reduction summarises complex maps to facilitate their communication and discussion. Finally, using measures of centrality for each factor, we identify the most influential ones. We present our experience of dialogue with two groups of traditional midwives in Guerrero State, Mexico, about factors that contribute to healthy maternity in their communities. A second example presents the perspectives of traditional communities in Bauchi State, Nigeria, to discuss the causes of short childbirth intervals. Using FCM, we have developed models that integrate the local knowledge of diverse cultural groups into the formal process of epidemiological studies. This offers a useful tool to facilitate dialogue in health research.