Engineering has advanced for more than a century by improving materials, components, and the systems that integrate them. As engineered systems become increasingly interconnected, autonomous, and adaptive, a growing share of their behavior is determined not by any single component but by the relationships among components, subsystems, missions, and the dynamic environments in which they operate. This paper proposes Relationship Engineering: a candidate research direction in which measurable relationships are treated as explicit objects of engineering design, analysis, and optimization, alongside the components and systems that current practice already addresses so effectively. The proposal is offered as a complement to established disciplines — systems engineering, model-based systems engineering, network science, control theory, complexity science, and cybernetics — each of which already engages part of this territory, rather than as a replacement for any of them. We situate the proposal historically, review what the neighboring disciplines contribute, and ask whether a distinct layer of practice organized explicitly around the relationship as a unit of analysis would yield research and design leverage that the individual disciplines, pursued separately, do not. We introduce a minimal vocabulary and a tentative construct, the relationship unit, intended to make relationships specifiable and comparable. To keep the discussion concrete, we use transportation — and, within it, the relationship between a vehicle and the atmosphere through which it moves — as a single demonstration domain. We close with a research agenda, a candid discussion of limitations and anticipated criticism, and an invitation to collaborate. The paper advances no claim of inevitability; it proposes a direction and argues that the direction is worth investigating.