Systems-of-systems are often characterized as systems where the constituent parts have some independence from the whole. Recent research has aimed at clarifying in more detail what this independence means. It has shown that independence requires the constituent systems to be agents that observe the system-of-systems from within and construct internal models of it as a basis for decisions. This view on observers as parts of the system-of-systems parallels development in the field of second-order cybernetics several decades ago, yet the connection between that field and systems-of-systems has not been explored previously. This paper, therefore, summarizes key concepts from second-order cybernetics, including the subtopic autopoiesis. It then discusses what the implications are on systems-of-systems engineering through the identification of 17 concerns. These concerns relate to the physical topology, behavior, and control of the system-of-systems. This paper shows how these concerns directly relate to the theoretical concepts of second-order cybernetics and autopoiesis, and thereby, opens the door to further exploitation of this theoretical foundation.