Capabilities are widely used in the design of software systems to ensure security. A system of capabilities can become a mess in the presence of objects and functions: objects may leak capabilities and functions may capture capabilities. They make reasoning and enforcing invariants in capability-based systems challenging if not intractable. How to reason about capability-based systems formally? What abstractions that programming languages should provide to facilitate the construction of capability-based systems? Can we formulate some fundamental capability disciplines as typing rules? In this paper we propose that stoicity is a useful property in designing, reasoning and organizing capabilities in systems both at the macro-level and micro-level. Stoicity means that a component of a system does not interact with its environment in any way except through its interfaces. As an incarnation of this idea, we introduce stoic functions in a functional language. In contrast to normal functions, stoic functions cannot capture capabilities nor non-stoic functions from the environment. We formalize stoic functions in a language with mutable references as capabilities. In that setting, we show that stoic functions enjoy non-interference of memory effects. The concept of stoic functions also shows its advantage in ffect polymorphism and effect masking when used to control side effects of programs.