Soil pH varies by several units among ecosystems. While soil pH is known to be a key driver of plant species composition, we still have a poor understanding of how it affects carbon cycling processes. For instance, soil pH, or its associated chemistry in terms of base cations and organic acids, may affect decomposition rates of dead matter directly, by controlling decomposer composition and activity and, indirectly, by controlling the traits of the plant species and thereby the afterlife effects of those traits on litter decomposition. Leaf and litter pH may play a role in this control. Based on the very limited empirical data available, we hypothesized that variation in species traits including leaf (litter) pH, within and between ecosystems contrasting in soil pH, would have stronger effects on leaf litter decomposition rates than variation in soil chemistry would. We tested this hypothesis by carrying out a ‘common garden’ litterbed experiment in subtropical SW China, in which leaf litters of the 30 predominant plant species from mid‐successional forest on acidic sandstone (soil pH around 4.0) and calcareous soil (pH around 7.5), respectively, were incubated and their decomposition rates measured over two harvests in 14 months, both in soil plus litter matrix from their ‘home’ forest and in those from the ‘away’ forest. We found that leaf (litter) trait variation among species and plant functional types, headed by species' dry matter content but also including tissue pH, was the strongest driver of variation in leaf litter decomposition rates. Surprisingly, however, while these effects of interspecific trait variation were very strong among species from the same site, there was no overall difference in litter decomposability between the species from the acidic versus calcareous site. Equally surprising was that this strong difference in pH of soil substrate plus litter matrix from an acidic sandstone site versus a calcareous karst site did not directly affect leaf litter decomposition rates across a given species set. This first attempt to disentangle the multiple potential direct and indirect ways in which soil and leaf (litter) acidity might be related to litter decomposition rates has important implications for our understanding of soil–plant feedbacks. Based on our forest‐based study, we predict that soil–plant feedbacks via acidity are unlikely to be strong in ecosystems with wide‐ranging species in terms of their leaf functional traits, including leaf pH. A freePlain Language Summarycan be found within the Supporting Information of this article.