Privacy requirements are difficult to elicit for any given software engineering project that processes personal information. The problem is that these systems require personal data in order to achieve their functional requirements and privacy mechanisms that constrain the processing of personal information in such a way that the requirement still states a useful functionality.We present privacy patterns that support the expression and analysis of different privacy goals: anonymity, pseudonymity, unlinkability and unobservability. These patterns have a textual representation that can be instantiated. In addition, for each pattern, a logical predicate exists that can be used to validate the instantiation. We also present a structured method for instantiating and validating the privacy patterns, and for choosing privacy mechanisms. Our patterns can also be used to identify incomplete privacy requirements. The approach is illustrated by the case study of a patient monitoring system.