Blended modeling aims to enhance the development of complex software-intensive systems by seamless integration of textual and graphical modeling. Manual provision of the synchronization transformations between notations is considered to be a tedious and error-prone task. Moreover, as synchronized languages and/or notations evolve, transformations become redundant as the implicit mappings they represent are no longer accurate. This research proposes a language-and notation-Agnostic solution for the automatic generation of synchronization transformations between arbitrary domain-specific modeling languages. We contribute with i) a mapping modeling language defined appositely for the definition of explicit mapping rules in a mapping model and ii) higher-order transformations for the automatic generation of synchronization transformations driven by mapping models. The approach is successfully validated against two use cases. © 2022 Owner/Author.