We show how a symbolic approach to the semantics of process algebras can be fruitfully applied to the modeling and analysis of partially unspecified biological systems, i.e., systems whose components are not fully known, cannot be described entirely, or whose functioning is not completely understood. This adds a novel deductive perspective to the use of process algebras within systems biology: the investigation of the behavioural or structural properties that unspecified components must satisfy to interact within the system. These can be computationally inferred, extending the effectiveness of the in silico experiments. The use of the approach is illustrated by means of case studies.