There is a growing number of works dedicated to the definition of life. They often contain a plurality of definitional proposals belonging to different disciplines. They tend to start from a set of abstract, theoretical assertions, with little reference to the phenomenology of life.Footnote 1 Here, we want to take a quite different approach, relying on phenomenological observations rather than speculative considerations. In order to do so, we go where life has on our Earth the simplest possible expression, namely, at the level of unicellular organisms. We watch how a living cell works, and we see whether we can elicit from these phenomenological, simple observations a definition of cellular life. We wonder then whether and to what extent this cellular definition is also valid at other, macroscopic levels. We apply this procedure following and partially reorienting the research pathway covered by Maturana and Varela in the Seventies. The result is a reformulation of their definition of the living which aims to solicit further developments of the autopoietic definitional approach (Maturana and Varela 1973).

From the Phenomenology of the Cell to the Theory of Autopoiesis

Let us start from the simple question: How does the biological cell, as the minimal form of life, work?

This question, as basic as it may be, brings us to an apparent paradox. Inside any simple cell there take place thousands of transformations, but, in spite of that, the cell maintains its own identity, at least during all the homeostasis period. How is this possible? For the scientific observer the answer is simple. This depends on the cell’s capability of re-generating from the inside all the components transformed—be they ATP, glucose, aminoacids or proteins—at the expense of matter and energy coming from the medium. This is our first phenomenological observation, and the most important one. To the observer, the cell appears as a physico-chemical system which has the ability to exploit external energy and matter to carry out an internal activity of self-production and self-maintenance, consisting in the permanent re-generation from within of all its components and its own boundary.

This very simple remark provides the basis of the so-called theory of autopoiesis developed by Maturana and Varela (1973), according to which this activity of self-production—«autopoiesis»—is common to all the livings—it is what the minimal biological unit shares with other living beings.Footnote 2 As Maturana puts it in a recent interview (Poerksen 2004):

“When you regard a living system you always find a network of processes or molecules that interact in such a way as to produce the very network that produced them and that determine its boundary. Such a network I call autopoietic. Whenever you encounter a network whose operations eventually produce itself as a result, you are facing an autopoietic system. It produces itself. The system is open to the input of matter but closed with regard to the dynamics of the relations that generate it.”

Maturana here touches upon the peculiarity of the theory of autopoiesis, which is to point out the dynamical mechanism allowing the living process of self-production. The idea is that of a close chain of operations of synthesis and destruction of components, within which each operation triggers and integrates the other in such a way that the concatenation of process (re)-generates itself through the production of its own components. This theoretical perspective, usually condensed in the notion of organizational closure, constitutes the core of Maturana and Varela’s definition of the living as «autopoietic system». We can formulate it as follows:

  1. (1)

    An autopoietic system is organized as a network of production processes (transformation and destruction) of components which produces the components which, through their interactions and transformations, permanently regenerate and realize the network of processes, constituting a concrete topological unit defined by a boundary.Footnote 3

Although this definition holds for the basic autopoietic systems -i.e. the «first order autopoietic systems» given by living cells- it has broader implications. Indeed, it does not focus on the «structure» of the cell, that is, on its components, which are undergoing a continuous transformation. It focuses instead on the «organization»Footnote 4 of the life’s basic unit, that is, on the invariant functional relations which connect its elementary components into the persistent global unity able to produce itself.Footnote 5 According to Maturana and Varela, this specificity offers to (1) the capability of defining the whole living domain, as it makes (1) able to provide a general characterization of the cellular system—that is, valid for every kind of cell—and therefore a fundamental characterization of all the living beings. The idea of the two authors is quite intuitive. The cellular organization is not the only form of the biological organization, but it can be considered the basic one. It is present in the other two (known) forms of living organization, i.e. multicellular and social. According to the theory of autopoiesis, they respectively belong to (ii) «second order» and (iii) «third order autopoietic systems», corresponding to (ii) multicellular organisms and (iii) their social aggregates, whose organization does not merely contain the cellular one, but was generated by its transformative evolutions. In this sense, defining the cellular organization amounts to producing a fundamental definition of the whole biological domain: a definition of the basic organizational form of the living, in which specific organizational definitions, about multicellulars and their social aggregates, are contained—and from which they can be derived by conceptual transformation.Footnote 6

Here lies the main hypothesis of autopoietic biology, which affirms that the definition of the autopoietic system holds for every living structure: if a physico-chemical system is living, then it is autopoietic and, conversely, if it is autopoietic, it is living (Maturana and Varela 1973). Accordingly, if we substitute in (1) «autopoietic unit» with «living system», we provide a definition of cellular life which aims to be valid also for all forms of macroscopic life. At this point, a somewhat simpler form can also be proposed:

  1. (2)

    A living system is a system capable of self-production and self-maintenance through a regenerative network of processes which takes place within a boundary of its own making.

It is an original definition of life, with many specificities we cannot treat in detail here.Footnote 7 For our definitional purposes it is enough to notice that it implicitly proposes criteria to distinguish the living from the non-living. As already pointed out by Luisi et al. (1996), these criteria can be summarized as shown in Table 1.

Table 1 Criteria derived from the definition of the autopoietic system (adapted from Luisi et al. 1996)

Autopoietic Organization: A Necessary and Sufficient Notion for Defining Life?

Maturana and Varela have always claimed that the notion of autopoietic organization, here summerized in (1), is necessary and sufficient for characterizing living systems (Maturana and Varela 1973, 1984).

In the last years, independently, two groups of researchers (Bitbol and Luisi 2004; Bourgine and Stewart 2004) examined critically this statement and came to the conclusion that the notion of autopoiesis expresses a necessary, but not a sufficient condition for defining life. They agreed that all livings are characterizable as autopoietic systems, but denied the converse. In fact they were able to show the existence of artificial systems—generated not only by simulation, but also in chemical laboratory, such as self-reproducing micelles or vesicles (Luisi 2006)—that respect the autopoietic definition but cannot be considered living. In particular, these authors highlighted that some artificial systems can satisfy the criteria of distinction between living and non-living derivable from the notion of autopoiesis (see Table 1), but lack something essential to be alive: the adaptative interaction with the environment, which is described by the theory of autopoiesis as a «cognitive» interaction (Maturana and Varela 1973, 1984).

Maturana and Varela have dealt extensively with «cognition», and defined it as the recursive interaction of the autopoietic unit with the its ambience. They described it as a self-regulating coupling of the system with the environmental context: an active coordination of the internal autopoietic processes with the environmental dynamics which allows the system to conservatively react to external variations. For Maturana and Varela it corresponds to a permanent self-determinated modification of the system’s patterns of activity which, seen from the outside, appears as «intelligent adaptive behaviour» (selective assimilation of nutrients available in the environment, overcoming of obstacles while moving in the environment, etc.). Conceived this way, the act of cognition is an act which, literally, permits life, as it dynamically integrates the organic structure of the living systems in the environment. Maturana and Varela thus came to the strong conclusion that «life is cognition»—that there is no life without cognition. (Maturana and Varela 1973).

But the primary literature contains possibly confusing elements. Although the two authors have always claimed that autopoiesis and cognition are «two aspects of the same process, the process of life», in their production, the theme of cognition has been developed after the autopoietic definition of life - and in this sense separately. Consequently, it often seems difficult to understand precisely how cognition fits with autopoiesis, or how to consider the relation between autopoiesis and cognition (Bitbol and Luisi 2004).

Bitbol and Luisi argued that we cannot attach the label of «living» to a system which, even if it is autopoietic according to the criteria indicated in Table 1, lacks the cognitive aspect. Their conclusion was that a system, in order to be living, must be grounded in an internal metabolic network which is capable of interacting in an adaptive way with the environment. This reflects Maturana and Varela’s notion that the living has to have the cognitive ability to couple its own self-production process with those of the environment, an aspect of the biological systems left implicit in the autopoietic definition of life.Footnote 8 We do not have the room here to discuss all the implications. Let us just notice that, to express necessary and sufficient conditions for life, the autopoietic definition has to include a reference to cognition. How then can our previous definition(s) of life be changed, in order to take into account this new aspect?

The Proposal

We propose the following reformulation:

  1. (3)

    An autopoetic system is organized as a network of production processes which produces the components which, through their interactions and transformations, permanently regenerate the network of processes constituting the system itself as a concrete topological unit, separated from its medium by a boundary and related to it through cognitive or adaptive coupling.

Or, in the simpler version:

  1. (4)

    A living system is a system capable of self-production and self-maintenance through a regenerative network of processes which takes place within a boundary of its own making and regenerates itself through cognitive or adaptive interactions with the medium.

In doing so, we propose a definition of the living based on the theory of autopoiesis which, in its final form, also takes care of the notion of cognition, that is, of the adaptive interaction with the environment. This links autopoiesis not only to metabolism, but to evolution as well, and in a sense, to more traditional ways of thinking life.