Autopoiesis is the natural semiosis that occurs during the embodiment of any system of cognition. In other words, the description of how it self-organizes - as distinguished from the process itself of which the description is only ever approximation.

For instance, the knitting of noisy signals from the eye into pictures in the brain, as an organism grows.

The impossibility of observing the process by any means other than testing of the cognitive system that results at each stage implies a particular terminology caution: When anything other than a single monolithic natural life system is involved, i.e. a model of same, the entity should be described as an autopoietic network of parts. This may be larger, consisting of persons, or smaller, consisting of cells, than the whole system observed to have the power of autopoiesis in the natural world.

The scientific study of autopoiesis is considered central in most credible sources to the so-called living systems perspective:

A typical scientific guide to the subject includes works by Humberto F. Maturana who was one of those who coined the word to describe a strictly limited sort of self-aware, self-organized system that could be observed in controlled settings.

Tenative use of the concept to describe natural systems and wild cognition, up to a maximum sort of gaiapoiesis, has been attempted, but, it is not clear whether this is valid.

Use of the concept in business literature is far less controlled and often confuses it with heteropoiesis: the growing-together of multiple cognitions in say a social setting, by commitment or by more mechanical bonds:

A more breathless and ultra-mechanistic approach, common in the US, relates auto- with hetero- under assumptions that a molecular economy practicing an adaptive doctrine and perhaps with access to matter compilers - advanced molecular engineering - would lead to such a rapid economic life cycle that there would effectively be no easy to determine difference between socially demanded "hetero" and biological "auto" forms of cognition. This could only be credible within a mechanistic paradigm however, e.g. that of Christopher Meyer and Stan Davis, which seems to assume a military-industrial ontology. An extreme version of this is the mechanosynthesis presupposed by K. Eric Drexler's Engines of Creation as central to any molecular economy.