"Signature in counterparts" is a legal process that permits a contract between two or more parties to be brought into force by having the parties independently (possibly, remotely) sign different copies of the contract, rather than placing their signatures on a common copy at a physical meeting. The paper develops a logical understanding of this process, developing a number of axioms that can be used to justify the validity of a contract from the assumption that separate copies have been signed. It is argued that a satisfactory account benefits from a logic with syntactic self-reference. The axioms used are supported by a formal semantics, and a number of further properties of this semantics are investigated. In particular, it is shown that the semantics implies that when a contract is valid, the parties do not just agree, but are in mutual agreement (a common-knowledge-like notion) about the validity of the contract.