In this article we investigate the multilingualism of references in the Semantic Scholar Open Research Corpus (S2ORC). While this dataset contains peer-reviewed papers from different disciplines, written mainly in English, we identify the languages used in the references, their linguistic groups and their distribution over time. The results allow us to observe the dominance of English in science, as well as the relative proportions of the other 34 languages, representing over 2.4 million of the cited sources, and their linguistic groups. We show that the relative share of non-English citations has been increasing since 2000. However, the vast majority of citations in non-English publications are to English sources. We discuss some of the limitations of this study, mainly related to the nature of the dataset, which is biased towards English, and the quality of the language detection tool.