Plant–soil feedbacks can have important implications for the interactions among plants. Understanding these effects is a major challenge since it is inherently difficult to measure and manipulate highly diverse soil communities. Mathematical models may advance this understanding by making the interplay of the various processes affecting plant–soil interaction explicit and by quantifying the relative importance of the factors involved. The aim of this paper is to provide a complete analysis of a pioneering plant–soil feedback model developed by Bever and colleagues (J Ecol 85: 561–573, 1997; Ecol Lett 2: 52–62, 1999; New Phytol 157: 465–473, 2003) to fully understand the range of possible impacts of plant–soil feedbacks on plant communities within this framework. We analyze this model by means of a new graphical method that provides a complete classification of the potential effects of soil communities on plant competition. Due to the graphical character of the method, the results are relatively easy to obtain and understand. We show that plant diversity depends crucially on two key parameters that may be viewed as measures of the intensity of plant competition and the direction and strength of plant–soil feedback, respectively. Our analysis provides a formal underpinning of earlier claims that plant–soil feedbacks, especially when they are negative, may enhance the diversity of plant communities. In particular, negative plant–soil feedbacks can enhance the range of plant coexistence by inducing competitive oscillations. However, these oscillations can also destabilize plant coexistence, leading to low population densities and extinctions. In addition, positive feedbacks can allow locally stable forms of plant coexistence by inducing alternative stable states. Our findings highlight that the inclusion of plant–soil interactions may fundamentally alter the predictions on the structure and functioning of above-ground ecosystems. The scenarios presented in this study can be used to formulate hypotheses about the ways soil community effects may influence plant competition that can be tested with empirical studies. This will advance our understanding of the role of plant–soil feedback in ecological communities.