The control and mechanical systems of an embodied agent should be tightly coupled so as to emerge useful functionalities such as adaptivity. This indicates that the mechanical system as well as the control system should be responsible for a certain amount of computation for generating the behavior. However, there still leaves much to be understood about how such "computational offloading" from the control system to the mechanical system can be achieved. In order to intensively investigate this, here we particularly focus on the "softness" of the body, and show how the computational offloading derived from this property is exploited to simplify the control system and to increase the degree of adaptivity. To this end, we employ a two-dimensional amoeboid robot as a practical example, consisting of incompressive fluid (i.e. protoplasm) covered with an outer skin composed of a network of real-time tunable springs. Preliminary simulation results show that the exploitation of the "long-distant interaction" stemming from "the law of conservation of protoplasmic mass" allows us to simplify the control mechanism; and that adaptive amoeboid locomotion can be realized without the need of a central controller. The results obtained are expected to shed light on how control and mechanical systems should be coupled, and what the "brain-body-interaction" carefully designed brings to the resulting behavior.