This paper has been written in order to clarify the way that some potentially very useful physical devices can be thought about in the context of a wider system. These devices mechanically change state as their operating environment changes and can be directly used to act on that environment or another physically adjacent environment. The devices are autonomous and require no power. They are usually regarded as being so basic that they are not worth thinking about in a sophisticated 21st Century context.
The fact that they require no power makes them valuable and the fact that they are autonomous puts them into the same class of object as a person or an autonomous robot that can exist in an environment and act on it. Probably the most important issue with regard to these devices is to understand where, how and if the concept of an internal model separated from physical operation is useful when we construct our own model of their behaviour. We can examine this question by looking at conceptualisations of the Watt’s Governor. When we place these objects in a wider system it becomes important to know what additional physical modifications are required in order for them to modify their automatic operation.
These modifications can allow communication between this type of device and entities that are often conceptualised as having a discrete internal model (people) or entities where this conceptualisation is embodied in a discrete digital control device.