The term was introduced by the evolutionary biologists Humberto Maturana and Francisco Varela in their groundbreaking book Autopoiesis and Cognition: The Realization of the Living (1972) which set out to understand the nature and organization of living systems from the perspective of their unitary character. Maturana and Varela boldly assume that there is an organization that is common to all living systems, regardless of their actual component parts, and it is this abstract system that they set out to define. They also maintain that all living systems are machines, which they mean in the most literal sense---i.e. it is a form of organization that can only be explained in terms of the relations it institutes, not its components. A machine may be realizable by a variety of different components, a fact that does not alter its nature as a machine. Similarly, the uses to which a machine may be put may vary considerably, but do not in any way determine the nature of the machine. Their approach is purely mechanistic inasmuch as they will not consider the possibility of forces not found in the physical universe, but their interest in relations means that their work is necessarily speculative. Their most important claim is that autopoiesis is necessary and sufficient to characterize a living system---i.e. wherever autopoiesis is found there is a living system, and, vice versa, wherever a living system is found there will be autopoiesis. The autopoietic machine has quite distinctive characteristics. To begin with, autopoietic machines are homeostatic---this means all feedback is internal to the machine, and, as such, they exist in a balanced state of constant self-regeneration of their own components. They are not static, but maintain themselves through a constant process of change. Autopoietic machines do not have inputs and outputs that are external to them, although they can be disturbed or disrupted by processes occurring outside them. Autopoietic machines are also autonomous---all changes they undergo are for the sole purpose of maintaining themselves. Autopoietic machines have individuality---their organizational structure does not vary, even if their specific components do; they are also topological unities; their operations are boundary-defining; they reproduce only themselves. For all these reasons, a factory cannot be considered an autopoietic machine, because its processes of production are not oriented towards producing components of its own system and its outputs contribute nothing to the regeneration of the system; similarly, a crystal cannot be considered an autopoietic machine, because its processes, again, do not regenerate the components of the system, but instead convert media of one kind into a different state, and the process ends when the media are exhausted. The simplest and clearest example of an autopoietic machine is the cell. German philosopher Luhmann, Niklas adapted Maturana and Varelaâs thinking in order to produce an account of social systems as autopoietic machines. See also sympoiesis.