Engineering complex systems is challenged by problems of language, increasingly as systems become more complex. Prominent are (1) integration of subsystems using communication networks, and (2) integration of work of engineers, product planners, suppliers, and customers. Both “hard engineering” and “soft process” cases encounter some of the same language problems in different settings. This paper briefly notes aspects of the structure of language and meaning and how they relate to these system engineering contexts. A practical, model-based technique, the Protocol Information Model (PIM), is described, to integrate both “hard” engineered systems and “soft” work processes which use “shared communication channels”.