In this paper, we present an experiment on annotating citation contexts using a rule-based approach to investigate the extent of text around citations. The study of citation contexts and their types has been a central issue in recent work on citations. We processed articles from seven PLOS journals and performed semantic annotation of citation contexts based on linguistic resources we constructed. We built on previous work on verb form analysis, n-grams, and semantic category modeling in the form of a linguistic ontology. Based on the observations obtained, we propose some directions of work for the constitution of a semantically annotated corpus. The intermediate results obtained lead us to formulate hypotheses between the IMRaD structure and certain semantic categories.