The Life & Mind Seminar Network

DST and Autopoiesis: joined by a common gap

Posted in General by Ezequiel on August 4, 2008

It is great that Tom and others have continued the discussion on this topic on Thursday afternoon (which unfortunately I had to miss) and it is also great that Susan Oyama has sent her own clarifications on the matter. It so happens that I was almost finished with a post of my own on this issue (the relation between DST and autopoiesis) when all of this happened. So rather than rewrite it all I will post it more or less as I intended it (but I’ll make sure to remove unnecessary duplications of what’s been said already). Here it goes.

Susan Oyama gave a very enjoyable keynote address at the EpiRob conference in Brighton last Thursday. She provided a clear description of some central concepts of Developmental Systems Theory (DST) and made some useful clarifications (remarking, for instance, on the difference between the organism and the developmental system roughly as the difference between an entity and the causal system that gives rise to this entity). But a running theme throughout the talk was the question of the relation between DST and the theory of autopoiesis as formulated by Maturana and Varela. A look at this relation was inevitable as the central metaphor that Susan adopted to guide her exposition of DST was that of development as a process “without roof” (a phrase used by Amy Cohen to describe Francisco Varela as a person and as a thinker). When applied to the scientific theories themselves, this polysemic metaphor (no roof = no end, no final determinant, no boundary, etc.), I would like to argue, is more appropriate for DST than autopoiesis, but in both cases, the result (of the respective absence and presence of roof) is the same: these perspectives have less scientific bite than we – who take them seriously – would expect. They seem more like worldviews rather than theories. Why is this so? I would like to come back to this in my comment.

Susan explicitly said that the relation between these theories (or worldviews) has been a puzzling problem for her (and through conversations with Varela she found he seemed to have a similar problem). Despite a general and strong agreement on many points and common demons, at the moment where a concrete connection between the two perspectives was sought, they couldn’t quite find how to proceed. At this point, I would not take this as evidence of the two perspectives being incompatible. That’s too rushed. That is a term that I would reserve for two attempts at resolving a similar problem such that they would lead to different views as to what is the solution. Here we would seem to be at a previous stage: are these views even trying to do comparable things?

Motivations and roofs.

Just a few bold observations:

What is autopoiesis trying to do as a theory? It attempts to do two things: to provide a systemic definition of life and to derive the implications of this definition for an understanding of cognition.

What is autopoiesis reacting to? The disappearance from modern biology of the organism as a useful scientific entity. And representationalist views of cognition.

What is DST trying to do? To present developmental processes as central to the understanding of life (at all timescales both shorter and longer than that of development itself). To break down persistent dichotomies that have dominated the discourse about development (most crucially, nature/nurture). To offer a view of developmental systems as causally extended beyond organismic boundaries.

What is DST reacting to? The conceptual and political perils of biological determinism.

DST has (by design, I dare say) no roofs. It does not want to privilege causal components of a developmental system, it does not even want to propose an “interactive” view that would soften the dichotomies while conserving the assumptions that gave rise to them. It wants to dissolve those dichotomies by taking seriously everything that happens in development, whether its in the genes, the physiology of the organism, its medium, its social and cultural engagements and so forth. If tomorrow we were to find that an unexpected causal link has been verified between the development of certain rainforest species and certain climatic regularities in the North Pole, DST would be ready to embrace this discovery.

(Classical) autopoietic theory, by contrast, is like a bunker. It not only has walls and roofs, it practically has no doors! One has to teleport into it by a leap of faith in the hope that eventually it all will make sense. And to many it does. The theory is presented and constructed so that it has apparently no gaps, no idea under-developed, no future issues, only simple logical excursions into its (unintuitive) consequences. You get the whole theory, for the price of a definition.

In each case, I think I’ve described a strength. But I’m sure the negative implications are also clear. Both theories share a similar problem: they have little friction. Autopoeisis because it is perceived as too self-enclosed, DST because it is perceived as too dilute. Not many people are holding their breaths for the experiment that will lend support to either theory and produce a massive intellectual migration out of the mainstream. (In fact, for the case of autopoiesis, it is not even clear what would count as experimental support).


The perception of autopoiesis as having an internalist emphasis is well justified. However, if we open up the bunker, it is possible to question the uniqueness of this interpretation. It all comes from what I believe is a sound distinction made be Maturana and Varela (but in later years championed mainly by Maturana) which is then followed by a problematic appreciation. This is the distinction between the constitutive and the relational domains. It is a sound systemic distinction between processes that sustain the system as it is in the “here and now” and processes in which the system as a whole relates to its environment. Discourse about constitution (not to be confused with processes that gave rise to the system – they are rather processes that fully describe how the system operates right at this moment) and discourse about relations between the whole and its medium should not be reduced to each other. That’s one main message of autopoiesis. But these discourses are unhappily referred to as “non-intersecting domains“. Maturana takes this literally. It is indeed a form of dualism: two substances that say nothing about each other, their coordination brought about (in Malebranchean fashion) not by God, but by evolution (a conclusion pregnant in the habitual phrase “otherwise the system disintegrates”).

I think that this strict separation has justified the internalist view of autopoiesis. Because where do we locate “mind”? If the system is running with its own internal dynamics and is only perturbed by the interactions, which may or may modulate the internal dynamics while conserving autopoiesis, then all that is important about the system’s relation with the outside is happening inside, in its constitutive domain. Any connection to the relational domain is merely contingent (it just happens to make the system viable).

What I believe the life and mind version of enactivism is achieving (e.g., Moreno, Barandiaran, Thompson and Di Paolo) is an alternative view. Non-reducibility between domains of discourse does not mean the total separation into non-intersecting substances. Agency for this view is on the relational domain, but initiated by constitutional processes that drive the regulation of relations with the environment. Mind is therefore at this boundary, not inside, nor external. The achievement of this alternative take (which is based on autopoiesis but eventually overcomes it while conserving its most fundamental implications) is that of opening up the question of the origins of normativity (a question raised by Hans Jonas and later addressed by Weber and Varela, 2001, Di Paolo, 2005 and Thompson 2007). Why do things matter? Who do they matter to? These are questions that bridge the relational and constitutional domains. And they should! Values are the appreciation (differential dynamics) by an adaptive self-sustaining system of the consequences of its interactions (relations) for its own viability (constitution).

A view of agency developed *from* autopoiesis need not be perceived as internalist. A window may be opened in the roof.

Joined by the gap?

If, as I believe is possible, both autopoiesis and DST were to be developed in directions that would provide them with more teeth (more friction) this would likely bring them closer since it is in both cases the problem of normativity that lurks in their immediate horizon. (Non-classical) autopoieisis would open up the bunker and let some interaction in (at the cost of a less finished flavour to its discourse and the possibility of provisional developments that do not quite close upon themselves but remain open to further change). DST would add a few structures, maybe not a roof, but perhaps a way to properly bring the organism into the picture as an asymmetric centre in the complex causal web of developmental processes.

Autopoiesis has indeed been developed in this direction in recent years. I’m not aware of similar developments in DST. But when I asked the question of normativity after Susan’s talk, she said in passing that she didn’t want to go into a debate about normality/abnormality. “That’s my non-answer”, she said but actually it was a perfect answer, because it really made it clear to me that such a development would not only be in the direction of introducing dichotomies (roofs) but also quite against the what DST originally reacted to: the idea that a conceptual framework could be used to lend support to potentially dangerous politics.

But if we need scientific ideas to help us understand our deepest problems, we need to move back into non-neutral terrain. Understanding the biological and social roots of normativity, perspectives, and autonomy is a way of re-connecting science with a discourse on ethics, values and politics. Just because scientific ideas have a history of being misused, we should not conclude that they should be encapsulated outside a discourse of values. Scientific metaphors are often criticized because of the hidden political values they embrace. But from such criticisms it would seem as if the aim of science would be to be value-less. This leads to the post-modern paralysis. All science represents a particular perspective and the discourse of a community with a history and specific interests. If we want to rebel against those interests, we shouldn’t accuse science of not being as value-free as it claims but to offer a science that best expresses the alternative values we are after.

Hence, I think DST should venture into a discourse about normativity. It would be especially sensitive to do this with the care that is needed because of its sophisticated ethical and political awareness. When has a developmental process gone wrong? What (if anything) should be done in such cases? What does it mean not to be able to self-actualize? From DST and from (non-classical) autopoiesis, we can formulate (maybe converging) answers to these questions.

For this, I believe there can only be the possibility of building bridges with the new enactive developments of autopoiesis that are beginning to complete the original aim of the theory: to ground a theory of cognition (sensemaking) though a proper naturalization of values and points of view. But this development is possible because autopoiesis does posit some boundaries to life (not so much the actual physical membrane of the cell, which has received sometimes too much emphasis) but the organizational and dynamical boundaries that ground the identity of the organism. The formal process of material transformation that is metabolism and the higher level processes that ground other forms of identity. For norms can only refer to what is good or bad to someone not to an ever receding causal system. We just need to know how to handle boundaries without leading to dichotomies. One way to do this is always keep a temporal horizon in view. Something that classical autopoiesis lacks, but DST can clearly contribute.


One Response

Subscribe to comments with RSS.

  1. tomfroese said, on August 10, 2008 at 1:14 pm

    Inspired by the previous posts, this comment is another attempt to explore what it would mean to enter into an explicit discourse about normativity in relation to autopoietic theory (AT) and developmental systems theory (DST).

    I find it interesting that while AT and DST appear to take incompatible perspectives on the issue of normativity, they seem to do so because of a more or less shared background motivation, namely due to social and ethical concerns. Perhaps this observation can be explained by taking a look at what Susan called their “sensitive dependence on initial conditions”. What are these?

    On the one hand, (if I understand it correctly) there is the DST attempt to establish a discourse in science which does not lend itself to be appropriated by any social power structures that aim to marginalize any aspect of our world at the expense of another. For example, its conceptual framework cannot be used to justify any value judgments regarding the normality/abnormality of any phenomena. This response to a politically charged history of science, especially biology and social science, puts DST at odds with any project that aims to understand normativity in a scientific manner.

    On the other hand, there is the AT attempt to establish a discourse in science which can help us to understand normativity without explaining it away. For example, can we say anything about how our biological existence relates to the fact that we experience a meaningful perspective in a world that has significance for us? This response to a disenchanted history of cognitive science, especially cognitivism, puts AT at odds with any project that aims *not* to understand normativity in a scientific manner.

    I’m aware that this is only a rough characterization of some very subtle issues, but I have a feeling that this apparent difference in motivation could help us to better understand the relationship between DST and AT. How deep does this difference go? Can we find a common denominator? I want to make two comments in this regard:

    1. We should not forget that part of the reason that AT has what Ezequiel calls “less scientific bite” is because it is inherently non-absolutist. It claims that the kind of cognitive domain in which an agent can act is co-determined by the organization of that agent and its environment, as well as their history of structural coupling. Any cognitive act is a relational phenomenon. When we apply this understanding to ourselves, we are forced to accept that AT cannot claim absolute validity. Moreover, it turns AT into a theory that continually deconstructs the foundational distinctions which make it possible in the first place.

    I think that it is for this reason that Maturana is always fond of reminding us that “everything said is said by an observer”. This seems to fit well with DST’s ethical and social motivations to bracket problematic distinctions.

    2. While deconstructing distinctions in this manner is important, it could be argued that this is only part of what it takes to clean up our discourse, and that it needs to be complemented by an understanding of the manner in which these distinctions arise in the first place. More importantly, it seems to me that such an understanding is only made possible by making the role of value judgements explicit in our discourse. Without this explicit discourse the reasons for deconstructing a particular distinction will remain opaque and unconvincing to anyone who has not reflected about her/his value-laden premises for accepting that distinction as independently valid.

    From this point of view we could say that AT provides a scientific grounding for the notion of ‘framing’, i.e. that there is no such thing as value-free science. This is a point which is surely compatible with the DST worldview.

    One last point: we know from history that we need to be very careful about having an explicit scientific discourse about normativity. Nevertheless, I think we should differentiate between explicitly accepting the general claim that we all enjoy a meaningful perspective on a world that has significance for us, and the more specific claim regarding what is good/bad for us and others.

    The latter is clearly more ethically problematic, though its potential for abuse in this context is reduced by the self-relativization of AT (see comment 1.). Still, I believe that the worry surrounding more specific normative claims should not make us resist asserting the general claim that we live according to value judgments. This is important both in terms of combating the prevalent nihilism which plagues modern/post-modern civilization, as well as for raising awareness of the responsibility we must take for our actions.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: