The Life & Mind Seminar Network

Developmental systems theory and autopoiesis

Posted in General by Tom Froese on August 1, 2008

Hi everyone,

Yesterday many of us had the great pleasure to listen to Susan Oyama give a talk about her work on developmental systems theory (DST) as part of the 8th Epigenetic Robotics conference held at Sussex University. Of particular interest was her attempt to compare and contrast DST with the autopoietic tradition started by Maturana and Varela.

Since Susan’s talk led to many heated discussions afterwards, I thought it might be a good idea to try to draw out some of these ideas in a more principled manner. So here are some questions that I think are important to consider:

1. Does DST need autopoietic theory?

From what I heard in the talk, I think it does. What the autopoietic tradition offers is a definition of life (e.g. “a system that self-constitutes an identity under precarious conditions”) that can at the same time also be used as the foundation for a biological grounding of normativity. Such a (living) system, i.e. one whose being is due to its own continuous activities, and which could thus become a different kind of system at any time (living or non-living), we can say to have some kind of concern for its existence. Of course, this claim regarding normativity does not follow as a logical necessity from autopoiesis, but it does lend a certain coherence to the life/mind continuity thesis.

I have to admit that I’m much more familiar with autopoietic theory rather than DST, so I’m just curious to know: how does DST distinguish the living from the non-living? This is where autopoiesis might help, especially because it introduces an important asymmetry into the developmental system. Since a living system self-produces its identity (within a domain of necessary conditions/processes), it also establishes its domain of possible interactions at the same time. Without this locus of activity it is hard (if not impossible) to account for a living system’s intentional directedness toward its world.

2. Does autopoietic theory need DST?

Again, I would agree with Susan that this is also the case. Indeed, the notion of autopoiesis has often been used in the context of a radically individualist constructivism, while completely ignoring the material/energetic flow that necessarily opens the autopoietic system onto a world. And as DST shows so well, it’s not just a matter of physics either! The necessary conditions for life extend well beyond the demands of organic matter all the way to the social domain.

While the autopoietic tradition has acknowledged the role of external factors to some extent in such notions as structural coupling and, in a more historical sense, structural drift, it could greatly benefit from DST which has spent a considerable amount of effort to better understand this web of involvement and interdependent relationships.

3. Are the current forms of DST and autopoietic theory compatible?

I think that DST and autopoietic theory are actually quite compatible. If we wanted to understand them as complementary aspects that could be incorporated into a more encompassing theoretical framework, then there might be at least two ways to go about this, depending on whether you start from DST or autopoiesis.

On the one hand, from the point of view of DST, the autopoietic tradition could be of help to define exactly what is the living system within the developmental systemic whole. Susan said: “A developmental system is a causal system. It contains the entity it produces but is not coextensive with it”. What is that entity? It is what is being produced by the developmental process. But it is also part of that process that produces it (though, of course, not co-extensive). When we distinguish that entity, might it not then appear to be self-producing, that is, autopoietic?

On the other hand, from the point of view of autopoietic theory, in the literature there is a considerable lack of attention to the complex network of relations that a living system is involved in, many of which are also necessary conditions for its existence. However, the recent development in enactive cognitive science toward using autopoiesis as a way to ground normativity is starting to expose this lack quite clearly: we’ve started to talk about meaning, but what kind of meaning is there for the organism? What differentiates it? On what does it depend? Here DST appears to me to be able to provide an appropriate context which could enable us to better structure the space of meaning for a living organism.

Accordingly, the two traditions might be able to meet each other in their respective blindspots and move the debate to a meta-level where each of them appears as a necessary complement to the other.

I would be very happy to hear in what way you agree/disagree with these comments.



16 Responses

Subscribe to comments with RSS.

  1. Susan Oyama said, on August 1, 2008 at 8:47 pm

    I’m interested and pleased that my visit to Sussex occasioned some discussion of these issues. At Tom’s request I’ve sent the text of my talk, so you will presumably have access to it soon.
    For now, just a few initial responses to Tom’s remarks. While I didn’t pose the question, Do DST and the autopoietic approach need each other, you are of course free to chew on it. I myself did not, however, claim or imply anything about any such needs, so Tom’s opinion in #2 should stand on its own rather than being offered as an agreement with me. (I should also say here that I speak only for myself.) Still, I did end by wondering out loud whether some sort of productive dialogue might be possible, and the chapter on development and evolution in the work by Evan Thompson I mentioned (Mind in life: Biology, phenomenology, and the sciences of mind. 2007. Cambridge, MA: Belknap/Harvard University Press, ch. 7) is in part a result of exchanges he, Varela, and I had over the years, so I’m interested in seeing others’ thoughts.
    One thing I have pointed out to other groups (including the ad hoc gathering after our last session on Thursday, some of whom might be interested in joining this online effort—perhaps they can be invited) on this issue of relations with what I would consider more internalist perspectives like those of Maturana and Varela, Goodwin, and to some extent Kauffman, is that I have never been concerned to define life, identity, or organisms. Neither have I focused on the origin of life. I’m not uninterested in these matters, but am quite happy to proceed in my own writings with a commonsense, rough-and-ready notion of organism, always with the caveat that if anyone wants to get rigorous about defining them all sorts of difficulties are sure to present themselves. As I said the other day, I have traveled a particular intellectual path, largely circling around nature-nurture problems, and have never felt that the work I was doing required anything more. I am, after all, primarily interested in the way people understand and present these processes, not in formulating some grand Theory of Life of my own. I’ll take the organism as you see it and move on from there. That said, I realize some may take issue with this relatively modest-sounding characterization of my project.
    Marion Lamb thought I was weaseling a bit by using “entity,” but this word isn’t an attempt to avoid saying “organism” (in fact I talk about organisms quite a lot), or to dodge defining organisms, but rather a way of leaving open the possibility of addressing infra- and supra-organismic scales. Think of Stern’s gastrulating cells, or of groups.
    That’s it for now. I’ll sign off and that I enjoyed the various talks I had with people during my brief stay at Sussex., and will be delighted if this generates or adds to a bit of ferment.
    Susan Oyama

    ps: Someone mentioned an innovative low-cost housing unit called, I thought, Spaceships. It was, he said, in the park adjacent to the campus, but I didn’t have time to go look at it. A cursory web search came to naught. If there’s an image on the web, do let me know.

  2. Mike Beaton said, on August 2, 2008 at 3:41 pm

    It was interesting that Susan’s talk was presented as if she felt that she was giving a view which would likely generate a negative backlash, from those working in the field – whereas I think for the ‘Sussex’ crowd, at least, the general reaction was: ‘yes, that all sounds very clear thinking and reasonable’!

    On the autopoiesis question, I think there was a feeling that something in Susan’s approach at least hinted at a less internalist approach to ‘the organism’, than anything in autopoiesis. It seems that in SO’s approach, ‘the organism’ is only defined/definable as a part of that system within which it is seen to be embedded.

    In autopoiesis (as I understand it) we say whether something is an agent, or not, by looking at whether there are the right kind of self-constituting loops within the organism’s structure. In SO’s approach (as I understand it, from much too little acquaintance) we say whether something is an evolving organism by looking at whether there are certain relationships between the organism-as-a-whole and its environment.

    This hints (perhaps no more than hints) that there similarly might be a definition of ‘agent’ which is concerned entirely with the mutual interactions of whole-agent and environment, and not at all with sub-system, self-constituting loops. It hints that the definition of agent which can be made compatible with Susan’s DST is not the definition of agent which autopoiesis offers.


    PS It is ‘Earthship’:

  3. matthewegbert said, on August 4, 2008 at 11:21 am

    To me, DST is an argument concerning how we should (not) study things. What type of conclusions we can (not) draw from certain experimentation. For example DST says that: we can not attribute causality to single processes/events/things, as causality is generally not a linear arrow of cause->effect, but is instead a tangled cyclical network. Stated more positively: we must remain aware that there is a tangled network of effects that extend across the boundaries of what we study. Elements in this network are of equal causal importance, because it can effect the other elements — some directly, some indirectly.

    DST does not make a case about what systems we should study or what even what developmental systems are. As Simon McGregor asked Susan Oyama after the conference, “What is _NOT_ a developmental system?” This is an important question that as far as I know is not addressed in DST. Autopoiesis is one way that we can identify (define?) systems of interest. To me, so far, it is a good starting point with plenty of room for improvement. I know that others disagree.

    So far, DST and autopoiesis seem to be quite compatible. However, if we look more closely, we find conflict between the two theories — in particular in the notion of asymmetry between organism and everything else. Does such asymmetry exist in an
    objective sense? Where autopoiesis asserts the existence of an asymmetry between organism and everything else, DST asserts the absence of any such asymmetry in the causal domain.

    DST: Any mechanisms “within” an organism are affected by mechanisms “outside” the organism and vice versa — so how can we have an objective asymmetry?

    AUTOPOIESIS: The unity defines its interactions with the environment, thereby producing an asymmetry of that which produces the interactions that that which does not.

    DST: But if the unity is in a tangled causal network, inseparable from its environment, then how can you say that the organism creates its interaction with the environment? Is it not just as valid to say that the environment produces the organism thereby producing the organisms interaction with the environment? Everything affects everything else — there is no _cause_, as such, for anything!

    This issue to me in unresolved for the time being. I’m unsatisfied with both arguments!..for the following reasons:

    My interpretation of DST almost replaces the idea of an organism with a “developmental system” — but this DS seems to have no starting point — temporally, spatially, or causally. Again the question “What is (or is not) a developmental system?” If a DS is a tangled causal network — how do you define the network? What is the first item from which you infer the rest of the system? Perhaps ultimately the lack of answer to this question does not matter — but for now it sits uneasily with me.

    Similarly, I like the idea that autopoiesis theory can resolve this “no start problem” of DST by producing an objective asymmetry — but I am not yet entirely convinced that
    it can. Perhaps somebody has a convincing argument that this is the case?

    I like Ezequiel’s comment about DST being too dilute and autopoiesis being too self-enclosed. They are both great ideas, but difficult to apply. I’m keen to develop models that perhaps loosen the theory slightly, but in the process allow us to perhaps gain some traction for the theories.


  4. Simon McGregor said, on August 4, 2008 at 11:47 am

    I think Tom’s questions are appropriate ones, although unsurprisingly my answers are not the same as his. In this post I’ll concentrate on question 1.

    1. Does DST need autopoietic theory?
    I suspect the answer to this depends on two sub-questions:
    – 1a. Does DST need to address the question of what an organism is?
    – 1b. Does autopoiesis usefully address the question of what an organism is?

    I don’t personally think the issue of normativity is relevant because I consider it to be an idiosyncratic concern of the autopoiesis community. DST does not purport to deal with normativity, qualia or for that matter quantum gravity. And why should it?

    On the other hand, DST explicitly involves (one or more?) “organism” in the system of interest, so it is not unreasonable to probe this term further. So:

    Q 1a. Does DST need to address the question of what an organism is?
    A Yes.

    I think Susan’s instinct to leave this one to common sense is a mistake. One unarguable contribution of the autopoiesis project has been to demonstrate that the definition of “organism” is nontrivial. DST researchers should be worried about what hidden assumptions they are making by using the concept of an organism at all. Putting one’s faith in common sense is a dangerous move, particularly for a project which relies upon deconstruction of common-sense categories such as nature / nurture. DST needs to establish that it isn’t building on quicksand.

    Q 1b. Does autopoiesis usefully address the question of what an organism is?
    A Only if photocopiers are also organisms.

    In my opinion Susan completely hits the nail on the head here when she says that autopoiesis privileges internal causation over external causation. Because we are so used to making attributions of causal locus, we forget they are not formally justified: there is no “the cause” of an event in a coupled system; rather, there are multiple internal and external causes. Whether an organism (perturbed by its environment) creates and maintains its own boundary, or whether the environment (perturbed by an organism) creates and maintains the boundary, is a subjective question, not an operational distinction.

    Years ago I began a stubborn campaign of resistance to the idea that autopoiesis is well-founded, on the basis that it couldn’t (as stated) distinguish operationally between an amoeba and a photocopier. Although I haven’t changed my opinion on the formal autopoietic nature of photocopiers, I have become less convinced that this is a weakness of the theory. As Susan indicated, we should be prepared to broaden our horizons and let go of our prejudices.

    On the other hand, I still don’t buy the alleged continuity of life and mind!


  5. Susan Oyama said, on August 6, 2008 at 7:28 pm

    More on Tom’s initial post:

    I have the impression that my points about theoretical approaches’ sensitive dependence on initial conditions (the role of history, that is, and of the particular issues that animated M&V and DST) were of less interest than a head-to-head comparison of the two perspectives. So be it.

    To fill in my wary reaction to normativity: Again, think of my path, through the nature-nurture opposition in various disicplines but also in the society at large. So much debate is motivated by the desire to say what is right, what is natural, what should be, what is good for people or other animals, what is good for society, etc., etc., and to give judgments the authority of a ‘biological base.’ The appeal to biology in such debates is often treacherous. It carries instant legitimacy (SCIENCE), enough ambiguity, smoke and mirrors to deflect hard questions, and emotional/moral resonance to spare. That’s just part of what makes me cautious. I would agree that there’s no logically necessary link between autopoiesis and normativity here, but then I’d ask what the connection is, and why DST needs to heed it. The features of autopoiesis in Tom’s last paragraph are just what I was trying to get at as the sorts of conceptual mismatches with DST that seem to stand in the way of anything more than friendly neighborhood relations between us.

    As for the extended developmental influences, I would say not just that the social world is necessary for development but that individuals are socially constituted, not in the sense that they are formed from the outside-in (an more than they are formed from the inside-out) but in the sense that sociality is not added on but is implicated in the very formation of individuals. In ch. 10 of Evolution’s Eye I tried to get at some of the conceptual issues troubling this question.

    On appearing to be self-producing: I’m not sure where this gets us. Part of DST’s point is that the appearance of self-making is in certain senses deceptive. And a whirlpool also appears to make itself.

    These reactions of mine are not to be taken as a dismissal of the possibility of rapprochement; I’d like to see where this goes and what others make of it. Evan Thompson (2007) certainly seems to feel there’s no serious barrier to integration, but I don’t know the literature well enough to say how much he departs from M&V (Maturana & Varela), and I’ve only read a few chapters, quickly.

    On Mike’s post: Thanks for the tipoff on Earthship.

    I’ll cop to sometimes taking defenses-up stance, largely out of force of habit. My audiences are various, and I did have some feeling that Maynard Smith’s ghost might defend information and other aspects of the standard view. I must say, however, the ‘that’s completely crazy, and besides, it makes my work impossible’ response seems to have diminished over the years, and part of the pleasure of being in Sussex was encountering all the readied minds.

    On agency: think on, and I’ll read with interest. Agency is a horrifically hard problem, in my opinion, and as I think we agreed, it tends to be linked to notions of responsibility.

    On Ezequiel’s post:

    Yes, I think it’s useful to look at what the two perspectives are trying to accomplish, although I’d say that I at least was reacting to more than biological determinism; rather, it was to the whole ramified system of assumptions and inferential habits involved in nature-nurture (biology-culture, body-mind etc., etc.) oppositions, and thus to the usual ways of connecting development and evolution, as suggested in fact by your section on ‘What is DST trying to do?’

    The comment on experimental support helps make a point about the meaning of ‘theory’ in DST. I’ve often said that ‘perspective’ or ‘approach’ is better if you understand ‘theory’ to be a machine for grinding out predictions (Ezequiel’s ‘worldview’ would do too). A looser, more ample notion of theory, however, fits us fine. (See also Godfrey-Smith in Cycles of Contingency.) While it’s not clear what a disconfirming or confirming experiment could be, however, there is another function a theoretical viewpoint can serve, which is to reframe questions, methods, and findings. This is what I was alluding to in the talk, when I cited Jablonka-type inheritance and research from areas like ecology and niche construction.

    You read my reaction to your question about normativity correctly, I think, as is also clear from my response to Tom. (Again I’d refer to that chap. in E’s E., as well as 8, 11.) The point about the fantasy of a value-free science is also well-taken. In fact, I think my writings are already value-saturated. Entering into an explicit ‘discourse about normativity,’ however, is another matter, and the necessity of doing so is not clear to me. Neither is the path to be taken. (Francisco, of course, would have said we lay down our path in walking.)

    Part of the notion of representations in the DNA is that there is intentionality in the processes realizing the phenotype. A hallmark of intentionality is the possibility of mistake (see Sterelny & Griffiths, Sex and Death). If we deviate from the plan in our genes (or from the adapted phenotype ‘designed’ by a quasi-intentional natural selection) then we can not only make mistakes, we can be mistakes: results that fail to realize the ‘correct’ outcome. I always thought that M&V were trying to avoid the whole area of comparative judgment when they refused numerosity as an evolutionary criterion and declined to talk about degrees of adaptedness.

    So, as you see, I’m intrigued but cautious.

    On Matthew’s post:
    This leads nicely to Matthew’s point that there are lots of cautionary admonitions in DST, or at least my writings: don’t conclude this, don’t infer that, don’t exclude this other thing. Given the state of affairs at which those writings were initially directed, I’d say that was a service in itself. I think this group wants what some would call more ‘positive’ messages from DST, and that’s what these posts seem largely to be about.

    BTW: I wouldn’t talk about causal factors being of ‘equal causal importance,’ because this implies a metric is already in place. Rather I’d caution against prejudging importance and inappropriately generalizing it or drawing conclusions from it. In a particular study, with certain conditions, you can say X made more difference to the outcome than Y did. In that limited context I wouldn’t object to saying X had more effect on the results. I wouldn’t want to move from there to saying that therefore Xs are more important than Ys in general, or that the resulting phenotypes are made more by Xs than by Ys, or that they were prefigured in the X, etc. From this point of view we should no more say causes are equally important than we should elevate one class above others in general, though if a factor is developmentally relevant then it is ‘equal’ in the sense that it is as much a part of the developmental complex as another.

    I agree that saying how to justify a judgment of asymmetry is both hard and important to exploring the relations between the two perspectives.
    On temporal unboundedness of a DS: the operation of developmental complexes is continuous from one generation to another, and this is an important point to make when the background assumption is that an organism arises de novo from the genes. Organisms come into being in worlds that predate them and that help make them.

    I disagree that DST replaces (or comes close to doing so) organisms with DSs. If you wish to talk about organisms, you can do so (see next section), but if you want to talk about their development, you must cast your net wider. I don’t think I could contrast the boundaries of organisms with those of their DSs, as I did in my talk, if I were replacing one by the other. I may have a difference with Griffiths and Gray, who at one point say there’s no difference between organism and environment. I’m inclined to read that as overstatement in service of some more limited point, but have not discussed it with them. For me, you have to distinguish the organism because the rest of the DS is defined with respect to it. That’s also how you limit a DS; if it has no developmental impact, a factor is not in the DS. It’s our interest in the organism’s development that starts the ball rolling, no?

    On Simon’s post:
    What kinds of ‘hidden assumptions’ may cling to the concept of the organism? Some, like causal/productive self-sufficiency, would be headed off by the rest of DST’s package, I’d think; what others should I be worrying about?

    I didn’t actually say I put my ‘faith in common sense.’ My point was that I’m not committed to a definition such that I need to insist on it in order to have a discussion. Whatever definition I used would be tentative. I’m not aware of any place where this has been a problem, and in fact, in my exchanges with people, in person and in print, I’ve never been asked to accept or reject some particular definition of an organism, and yet have had countless productive talks about lots of issues. In general I’m willing to talk til I can’t: that is, til things grind to a halt because we’re talking past each other or I’m too confused (or depressed) to proceed, etc.


  6. Simon McGregor said, on August 11, 2008 at 11:56 am

    Re: hidden assumptions in the concept of “organism”…

    One way of looking at DST (if I understand it right) is to say, “When we take a scientific perspective on any physical system which interacts with its environment, we have to understand the system’s behaviour holistically: without attributing causal primacy to either the system or its environment. This applies to what we call living systems just as much as to everything else.” I don’t see this as problematic.

    However, some of Susan’s comments (e.g. “a whirlpool is not a DS”) seem to suggest that DST is supposed to be a specific theory of living systems. That’s OK if one has a non-arbitrary definition of organism which excludes whirlpools. It’s even OK if one has conclusive grounds for believing that such a definition is possible.

    But if there’s any doubt that we can reliably distinguish between the animate and the inanimate, then there is a problem. It’s not that DST might be an inappropriate theory of living systems; it’s that the very concept of a “theory of living systems” might be founded on a false distinction (between living and nonliving).

    Another way of looking at it is this: if DST implicitly defines “organism” by reference to our intuitions, it is importing a potentially unscientific dichotomy. If on the other hand it genuinely doesn’t have *any* preconceptions about what “organism” means, then “organism” in DST is a placeholder noun which could be replaced by “object” or “machine”, and a rock hanging from a string becomes a DS.

  7. Susan Oyama said, on August 14, 2008 at 9:39 pm

    Susan again, with some backward-looking bits on past postings, followed by responses to more recent ones. Being a newbie at this, I would appreciate pointers on how to respond to a multi-threaded set of posts. I’m treating Ezequiel’s site and the Life & Mind together, for instance, but maybe that’s not optimal.

    Rereading my response to Matthew I see it could be read as saying DST’s messages are primarily negative or critical. While this may be the case for many questions, including many normative ones, there is a constructive, ‘positive’ side as well, having to do with the articulation of an alternative approach and all the findings that can be reinterpreted within that approach, as well as the research that can be generated from it.

    Simon: I don’t think I actually said “a whirlpool is not a DS.” I was talking about what benefit there could be to saying that a thing can seem to organize itself, pointing out that because a whirlpool can give this impression, that impression doesn’t distinguish the living from the nonliving in the way that I think AT would want it to. This would seem consistent with your point about copiers.

    People like Goodwin and Mae-Wan Ho used to talk about the physical processes involved in a variety of nonliving self-organizing systems, like convection cells. (This makes me wonder: were they ever asked to distinguish the living systems they were concerned with from the nonliving ones they used to illustrate the physical principles at work in the living? With what results?) Are you suggesting that any theory, including DST and AT, that doesn’t make that distinction is defective, or perhaps that the search for the distinction is misguided, and that we ought to be talking about certain kinds of physical systems regardless of whether some other scheme classifies them as living? Or are you getting at something else?

    Nathaniel and Ezekiel: Your exchange about boundaries was fascinating and provocative, though I didn’t follow every point. I suspect that the distinction between the constitutive and the relational domains is key. Ezequiel says ‘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.‘ I’m not convinced it can be maintained. In fact, you could argue one main thrust of DST had been to merge the two. Earlier I talked about being humans being socially constituted, for instance—for me, relations with the ‘outside’ are constitutive, not just external relations of already-constituted beings.

    Does Ezequiel’s definition of homeostasis as matter of a variable being ‘regulated by the totality of this organization’ help bridge this gap? Systemic interdependence in development seems to meet this criterion as well as the usual examples of physiological homeostasis. (Waddington talked of homeorhesis as well as homeostasis.)

    On ‘a science that best expresses the alternative values we are after’: I’d venture to say my values leak out of every paragraph I write. Again, see chapters in Evolution’s Eye.

    On the question, ‘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?,’ here’s where I get very careful. What criteria do we use? Part of the power of the gene’s eye view was that it called into question the assumption that the notion of ‘interests’ was not exhausted by tracking organismic integrity or survival, but even without it we can look at salmon battering themselves against rocks, spawning and disintegrating, and see this assumption put in doubt. So yes, either the organism disintegrates or it doesn’t, but does that opposition define its good? Or, instead of looking at what aids survival, should we look at the choices the organism makes itself, including those that sacrifice its life? Some breeders let chickens choose the size of their cages. Should we assume organisms choose what is good for them? A rat chooses sweet water over something more nutritious, and it chooses a zap to a certain place in its brain over food and water. Humans choose all manner of things other humans find abominable and hurtful. What if someone argues she is actualizing herself most fully by having her limbs amputated? I have just never seen a set of criteria that can adjudicate all the conflicting desires, considerations, and interests at the level of individuals, let alone at other levels.

    But rather than ranting on in this way, which surely introduces nothing new, I’ll wait to see what emerges from this discussion. Meanwhile, I’m sending to Tom a paper called Speaking of Nature that touches on some of the issues on the table. I’m hoping he can make it available, along with the talk I gave at the conference. (The latter may already be accessible someplace; as I’ve indicated, I’m skilless at such matters, and welcome instruction.) You can skip the DST primer and skip to the second half.

    I would, by the way, like to hear more about time, as in ‘One way to do this is always keep a temporal horizon in view.’

    Also, to Nathaniel’s musing about whether all of Earth’s plants must be part of a worm’s autopoietic system, how about saying that the notion of developmental relevance could place O2 in the DS, but if the organism’s actions or physiology affect the O2 or CO2 levels through effects on the plants, then that loop can be included, at least at the aggregate level, as indeed it is in contemporary discourse on climate change.

    Tom and others:
    There is, sprinkled through these threads, an implication that DST is an attempt to formulate a view without bias or dichotomies or, as Tom says, to ‘establish a discourse in science which does not lend itself to be appropriated’ (or, from Ezequiel, that DST is roofless by design).This turns things around a bit. I at least did not set out with these goals in mind. I realize that DST could look as though it was produced from a starting point of ‘one cause bad, many causes good,’ or ‘dichotomy bad, system good,’ or some such. Actually, I started out being irritated and dissatisfied with what seemed to me to be pseudoexplanations and the ways certain distinctions were justified. They didn’t make sense, were inconsistent, either internally or with other such distinctions, there was slippage from one discourse to another, massive ambiguity, and so on. It was the experience of trying to follow trains of reasoning or to make definitions work that eventually led to the formulations you’re familiar with. Causal parity wasn’t a guiding principle, but parity of reasoning did turn out to be a useful analytic tool.

    The political/moral connections were there to goad me on; whenever I wondered whether there was any point to all this irritable nitpicking they reminded me that the complexes I was confronting were part and parcel of our everyday methods of dealing with ourselves and others, that they had flesh-and-blood consequences. As I indicate above, I don’t avoid drawing normative lessons from DST because they might be dangerous, but because I don’t see how it can be done.

    There is another issue that appears in several posts, which I’ll try to identify with a second quote from Tom’s latest. He says ‘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.’

    I think there’s a difference between trying to derive or justify normative judgments by using a theoretical framework and trying to ‘understand normativity in a scientific manner.’ (Crudely put, I suppose, prescription vs. description, not that there’s a clean distinction.) The latter seems to me to be relatable to DST’s habit of parsing the world in terms of causal relevance to a focal entity, and is consistent with an epistemological stance that always takes the knower into account. This, in turn, as Tom observes, is one of DST’s, or least my, areas of agreement with AT.

    I’m not sure, furthermore, that any ‘nihilism which plagues modern/post-modern civilization,’ if nihisim it be, has to do with denying ‘that we live according to value judgments.’ Rather, isn’t it more a disquiet that because we do value and devalue, are valued and devalued, according to often widely varying and mutually incompatible judgments, we often wish for an overarching framework that will sort everything out, and that we wish in vain?

  8. ezequieldipaolo said, on August 15, 2008 at 8:36 pm

    Two quick notes in reply to Susan’s last comment. (It deserves a more thorough reply but I think these notes are important in clarifying possible misunderstandings).

    1. Constitutive/Relational

    Maybe the word “constitutive” is not helping much here. I’ve used “operational” in the past after M&V, but that can be confusing too.

    [In the following, please, consider the word system to refer to the conceptual construct that we use to describe the phenomena of interest; I’m not talking about “systems out there”, but “here” as ideas, diagrams, equations, theories, pen-and-paper stuff.]

    The distinction is unproblematic if we take it literally:

    Domain A: The domain of contemporaneous processes that give rise to whatever transformation the system undergoes at a particular time, given its state at that time.

    Domain B: The domain of relations (current and historical) that the system as a whole enters into.

    A and B are different domains. In dynamical systems terms, we make this distinction all the time: variables, parameters and laws of transformation belong to domain A, parametrical coupling, history, function, purpose, etc. belong to domain B.

    One can safely choose *any* system in these terms. The only extra “rule” is that the choice should lead to a state-determined system (M&V’s structure-determined system) to be able to use this description scientifically. What does this mean? That our delimitations of the system give rise to a level of description such that freezing the conditions in domain B, phenomena in domain A is determined by the current state of affairs.

    What could happen is that even when we “freeze” all relations in domain B, what happens in A remains undetermined (a same state leading to 2 different outcomes). Then we don’t have a state-determined system and we must re-describe. [stochastic systems can easily be incorporated into this notion.]

    Back to the main issue. I called A constitutive and B relational. But of course, if we use the word constitution in its genetic (historical) sense, we’re describing events that involve both domains. I don’t disagree. (We have even written a paper describing how the autonomy of social interaction is involved in the constitution of our social roles, De Jaegher & Di Paolo, 2007). So here I think I may have muddled the waters by not clarifying how I was using this term in this context.

    Back to another main issue. Surely, my delimitation of the “system” is no better that yours! No. Indeed. That’s perfectly fine. If our descriptions are different, we are indeed talking about different systems (even when both of them are trying to make sense of similar phenomena in order to perhaps answer the very same question). This difference can be perfectly legitimate and it is often the case that one and the same person can make use of this shift in points of view. This freedom of perspective in no way undermines the distinction between domains which is simply shifted to new domains A and B.

    I often hear two quick reactions to this observer-dependence: 1. Whatever you are describing is arbitrary. 2. It makes no sense to privilege some systemic description over others. Here I take a pragmatic perspective. “whatever one is describing” is a mutually constrained, evolving thing. Constrained by our perspectives, conceptual constructs and our interaction with the world. Change any of these, and the likely result will be a different view. At no point does this entail an arbitrarily different view. Which one to choose? Depends on our goals and traditions. Simple. Moreover, some systemic distinctions may give rise to the realization of certain properties that do not belong to other systemic description. We can choose to describe the living cell from the point of view of fluxes of matter and energy. It will not look like a living cell. We can choose to describe it from the point of view of processes of material transformation – we may realize it is autopoietic (“it” here refers to the entity defined by the systemic description that allowed us to make this claim). [Whoever questions the very coherence of the concept of “organism” based on this situation is missing the epistemological point: *all* concepts can be subject to the same treatment, so the concept of life does not suffer in increased ambiguity due to “shift in perspective” arguments – cars, houses, rocks and electrons will have a similar fate.]

    Relation between domains A and B.
    M&V use the unfortunate phrase: these domains are non-intersecting. Reading how this is used in the text, I think that what they mean is they are not reducible to each other (at least this interpretation helps me make better sense of their story). Non-intersection brings an unacceptable separation. Indeed, just consider the previous story of constitution through interaction, surely, there is more than an external relation between what happens in one domain and what happens in the other. M&V make an implicit appeal to evolution here. If the domains are uncoordinated, the system dies. I have written recently how this is unsatisfactory. The story involves the notion of the re-inscription of agency in metabolism. (Briefly presented at the Alife conference during my talk on Life in Time).

    However, what if find interesting is that while the domains interact (or participate in each other) and the distinction can be eventually questioned, this questioning leads to a sublation (Aufhebung) of the distinction, that is an overcoming whilst conserving it. Not quite a dissolving of the domains. [This, I’m afraid will be part of another story, but I just want to signal that I see the relation between these two domains as quite complex, and yet I have never seen the need to question the distinction itself.]

    Susan says that you could argue that DST has attempted to merge the two domains. If find this a bit confusing. I think the distinction holds as soon as we use any “distinction”, the basic cognitive operation of focusing on the value of a difference so as to act differentially with respect to it. What DST has done, in my reading, is to draw attention to the fact that certain distinctions are not comprehensive in generating appropriate explanations (as some people have proposed). That other distinctions also generate explanations of relevance for the phenomena of interest (potentially better ones) and that certain distinctions may have well outlived their usefulness.

    But the systemic distinction between domains A and B becomes operative as soon as we draw *any* boundary, regardless of whether we choose to re-draw it in the next sentence. I’m not talking necessarily about lines around entities and contentious boundaries, but about just any distinction like when we use phrases like “this mechanism”. “This mechanism does this or that” or “operates at this rate” these phrases belong in domain A – “this mechanism interacts with this or that”, “is more efficient that X”, or “it fulfills this or that function”, “it was evolved because of X”, we’re in domain B.

    2. A science of norms = A science of viewpoints = A science of autonomy

    When I say that we need to develop a science of normativity based on the central figure of autonomy, I’m not saying at all that the result of this science should be a list of norms of general applicability, nor any list of norms whatsoever. Quite on the contrary, a theory of normativity must first of all give us a naturalization of what it means to have point of view, what it means for things to matter to an autonomous system. And for this, inevitably, this theory must refer back to the system that has a point of view and to whom things matter. The multiplicity of norms and points of view is therefore built into the very question that this theory attempts to answer. It is a theory that addresses the question: why are there such things as values in our world? Not a theory that proclaims that these values are better than those.

    This is not anything new. Kurt Goldstein, Georges Canguilhem, Helmut Plessner, FJJ Buytendijk, Erwin Straus and others have attempted these connections, and in very interesting ways that were never quite developed afterwards. For instance, Canguilhem makes use of Goldstein’s notion of self-actualization to question whether health and illness should be defined as a statistical correlation of species-specific patterns or whether the organism in question should be the provider of this norm. (I may suffer from some respiratory disease, with a particular name, that will only be manifested once I start scaling mountains – since I don’t scale mountains, am I sick? Do I need treatment?).

    Hans Jonas, I think, is the ultimate philosopher of biology that attempts this. F Varela only found about his work relative late and was surprised at the striking parallels with what he described as his project of grounding sensemaking and teleology in autopoiesis. And for phenomenologists like Merleau-Ponty, it’s normativity all the way down to the very simplest act of perception, which our body regulates to achieve maximal grip.

    So, the clarification is, a science of normativity is a the scientific enquiry into why things matter, why are some systems autonomous, why do they have a perspective.

    [I now see that my paragraph on the “the science that best expresses the alternative values” probably didn’t help in clarifying what I meant by a science of normativity. I, of course, stand by that paragraph, but its meaning is general. It so happens that I believe that a science that can help us make sense of such things as autonomy, perspectives, and values, best represents those alternative values for me. The lack of a proper scientific account of these concepts had led people (not just scientists, but also policy-makers and the public in general) to believe these concepts are illusions, and that not ethical commitment is due when someone claims for instance that their autonomy is trampled on by the over-grammaticalization of their jobs – “they’re just being unscientific!”]

  9. Mike Beaton said, on August 16, 2008 at 11:48 am

    I would very much agree with Susan Oyama’s suggestion that a creature’s own choices are more important, for determining its own interests, than anything about whether or not those choices promote its continued physical well-being.

    Just a note about rats’ choices under experimental conditions: rats don’t always choose the sweetened water…

    Bruce Alexander’s ‘Rat Park’ experiments are very striking. A very rough overview – rats were forced to consume only sweetened, heroin-laced water. They were subsequently released into a lovely environment with lots of space, and things to explore and play with. They were then given a free choice between drinking the opiate laden water, or plain tap water. Most rats chose to stop drinking the opiate laden water, and to drink the plain tap water, even though they showed some initial withdrawal symptoms. Alexander argued that rats in many normal experimental conditions continue to ‘self medicate’ because the experimental conditions are so horrible.

  10. Nathaniel Virgo said, on August 16, 2008 at 3:56 pm

    Hi Susan,

    I think replying to everything in one post is fine personally. I’m working on a longer reply to Ezequiel’s comment in the other thread, but I wanted to touch on the point about time scales here. I think that whether we see a system as organisationally closed or open (or whether we see “feedback loops” or “enabling factors”) has a lot to do with time scales. On the time scale of a single worm’s lifespan there is no appreciable effect of its respiration on the level of oxygen in the atmosphere, so the presence of oxygen looks like an enabling factor, part of the background against which the worm is able to maintain its autonomy. But if our time scale is much longer then we start to see the cycling of oxygen. As you say, on an aggregate level worm respiration probably does have an effect on plant growth (in general more CO2 in the atmosphere = faster plant growth) and of course earthworms also have a much more direct effect on plants by breaking up and aerating the soil, so it starts to seem reasonable to include that feedback loop in the system. So whether we describe the cycling as part of the system or part of its background can depend on which time scale we’re interested in. I’m not sure whether this is what Ezequiel meant by “keeping a temporal horizon in view,” but it seems to me like a promising way to address this sort of issue.

  11. Susan Oyama said, on August 18, 2008 at 8:25 pm

    Ezequiel: Thank you. I’m not sure I follow it all, especially about domains, but I’ll keep working on it.

    Is there a relation between domains A and B and M&V’s internal and external points of view?

    Your clarification of the issue of values helps, and it sounds like we share a preference for a quite pragmatic stance.

    I didn’t actually say ‘a creature’s own choices are more important, for determining its own interests, than anything about whether or not those choices promote its continued physical well-being.’ Your expansion of the rat example is useful here, and I agree with your implication that these are nice findings that highlight the tacit choices that inform laboratory work, from the choice of species (docile, rapidly reproducing, developmentally robust to ambient variations, etc.) to animal housing to the research methods. I juxtaposed all those questions not to argue that organisms’ choices are a better guide to their ‘good’ but that there’s a multiplicity of criteria that I don’t find a stable way of evaluating. Because we think heroin’s bad, and love the Rat Park (I did, at least) we cheer the rat that’s restored to its senses, so to speak, by the enriched environment. But that of course opens us to questions about our attitudes toward heroin and environments with cool stuff in them.

    The point I’d rather make is that, precisely, rats’ choices can depend on their disparate experiences. I’m willing to bet you can influence chickens’ choices of cage size by varying their antecedent experiences. So again the task of evaluation is shifted to those experiences.

    A skilled interviewer or composer of a questionnaire knows how to get a particular response by preceding a question with a calibrated set of other questions and by careful wording. (My guess is that one powerful way of steering a response is by subtly invoking particular reference groups, and thus, by altering in a real sense who/what the person is at that moment. This speaks to my last point, about focusing on individuals.)

    Far from arguing that we listen to the organism, I was expressing skepticism that any such dictum can provide a legitimate guide. We routinely overrule our childrens’, pets’, psychiatric patients’, and demented parents’ choices—for their own good; indeed many would consider us at least irresponsible, possibly criminal or immoral if we did not. But even if we accept the legitimacy of such infringements (how would we justify them?), each of these presents us with a huge border area where judgment falters. Is there an age of reason beyond which we really should allow our children to chart their own courses, no matter what? When might we retract that privilege? (How crazy do they have to be, for instance? And who’s ‘we’?)

    Related to these questions is a worry that the issue, couched in this way, skews the inquiry by talking about individuals, more or less decontextualized and alone. Am I misreading you?

    Nathaniel: Yes. Even at the level of a single organism, you could have a respiratory loop; anyone in a closed space depletes the air supply. Here’s a nice piece from the New York Times:

    Come Up for Air? Not These Insects, Which Carry a Bubble as a Lung

    John Bush and Morris Flynn
    The water boatman uses a thin layer of air on its body to breathe underwater.

    Published: August 11, 2008
    To swim underwater, hundreds of species of insects like the water boatman bring along their own air bubbles.

    What is even more remarkable is that the air bubbles automatically refill with oxygen, allowing the bugs to swim indefinitely without coming to the surface. Some insects even hibernate underwater all winter.

    “It turns out this bubble functions as an external lung,” said John W. M. Bush, an associate professor of applied mathematics at the Massachusetts Institute of Technology and an author of a paper describing the mechanics of the air bubbles in the Aug. 10 issue of The Journal of Fluid Mechanics.

    Similar to lotus leaves and Gore-Tex fabric, the abdomen of an underwater-swimming insect is rough and waxy, which repels water molecules. With the water repelled, a thin layer of air is trapped along the abdomen. The roughness comes from tiny hairs, about 100 microns long and 1 micron wide and spaced 10 microns apart.

    As insects breathe in and out through tiny holes in the abdomen, oxygen decreases in the bubble and carbon dioxide increases — but the changes in pressure draw oxygen from the surrounding water into the bubble and push out carbon dioxide.

    Biologists had noticed this ability in insects long ago. In the new paper, Dr. Bush and Morris Flynn, a former instructor at M.I.T., calculated details of the insects’ diving ability based on the structure of the hairs, including how deep they can go before water pressure overcomes the water repellence and collapses the bubble.

    “That’s the math problem we solved, when these lungs can function,” Dr. Bush said.
    In principle, it would be possible to build a structure with an expansive hairy, waxy surface to provide an inexhaustible oxygen tank for a human diver. But scaling this mechanism from millimeter and centimeter-size insects to much larger people would be impractical. A more plausible application might be for powering an underwater vehicle that would draw oxygen from the water for a fuel-cell-driven motor.

  12. Mike Beaton said, on August 19, 2008 at 1:43 pm

    Dear Susan,

    No, you’re not misreading me at all. Your points are well made, and well taken. My own personal starting point is the broadly Western-scientific, decontextualised and alone, view of the self, and I’m trying to find my own way back to contextualised meaning.

    I’m still tempted to push my line, somehow or other. Do we really think that children allowed to pursue their every whim, end up happier than children given clear but kind guidance? I very much doubt it (and I suppose you’d be more than happy to agree… given my framing of the question!).

    So what am I looking for? A notion of that which is better for the agent (animal, child, etc.), from it’s own point of view, even if it doesn’t know it!

    But maybe this line isn’t tenable, in the end. (And it’s certainly moving off the topic of the other lines of discussion here – but thank you very much for responding!)


  13. Anna Dumitriu said, on August 19, 2008 at 8:42 pm

    As I mentioned at the Niklas Luhmann wrote extensively on Institutional Autopoiesis so he’s worth a google. Here’s a link

  14. Anna Dumitriu said, on August 20, 2008 at 3:46 pm

    I enjoyed the seminar and have been pondering things in my humble artists brain.

    My view is Autopoiesis is a pretty useless concept IF it’s not directly related to a definition of life, as far as anything I am interested in. By applying it to wider systems (where it can’t really be applied in that way) we are just playing definition games and attempting to define wider systems as living systems when they are only metaphorically so (yes the guitar is part of a system of music creation but it is not living – there is something else that makes things living and that’s the exciting question, what is that something else? What is life? (and that’s what Varela was trying to answer). A jazz band is a complex system, but not closed, not autopoietic or even autonomous – ‘just’ a complex system that constantly interacts with wider and wider systems and onwards to Gaia and beyond.

    I’m sorry to harp on about my favourite topic but lets get back to bacteria! There was much support for that in post seminar discussion and quite a lot of people who doubted the autopoietic jazz band concept (as I have dubbed it). But an autopoietic bacterium is another matter and far closer to the original ideas.

    I also think the boundary, membrane or whatever is a by product of the processes. It’s not the important thing but that it occurs is. But the boundary in question is the cell wall.

    In fact in the discussion the question of reification (especially in terms of the notion of money/economics as a system) becomes very pertinent but no-one really mentioned that. IF the boundaries are blurred and even more are only boundaries because we have decided to describe them as such then what are we trying to do? The membrane describes itself by existing which I think is what Nathaniel and Mike were trying to get to grips with, it’s not arbitrary, it stops bits falling out and bits dropping in, otherwise we’d all just be bits of code like viruses and they cannot ‘self create’.

  15. John Stewart said, on August 26, 2008 at 11:10 am

    I am absolutely delighted to see that there is such lively and fruitful discussion between DST and autopoiesis – my two favourite theories in biology, I cannot imagine they are other than compatible and complementary.
    The point I want to make is a very basic one, but does not seem to have been mentioned by others. It is this: THE prototypical example of autopoiesis is a MONOCELLULAR (procaryotic) organism (or even a tesselation automaton?!); whereas DST requires development, ie a MULTICELLULAR organism.
    To my mind, the important weaknesses of each theory lie less in their differences; but rather that each does not actually go fully to the bottom of its own prototypical domain. Thus: autopoeisis is still not properly or fully worked out, as a theory of single-celled organisms! (eg: what is the minimal autopoietic organism? what about the actual origin of life on Earth? etc etc). And DST is still not properly or fully worked out, as a theory of ontogeny (the fantastic process that leads from a fertilized egg-cell, through the wonders of embryogenesis, to maturation, aging and death). The challenge to DST is to account in detail for the fantastic REGULARITY of ontogeny (within any one species); it is our failure to do this, that lends the disastrous notion of “genetic programme” its fatal attraction.
    A last word on the relation: Maturana himself once said, if I remember right, that it is not at all settled whether a multicellular organism is actually properly autopoietic in its own right (as apart from being made up of cells, each of which IS clearly autopoietic).

  16. Susan Oyama said, on September 16, 2008 at 1:27 pm

    What happened at the workshop that was held after I left Sussex?

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 )

Google+ photo

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

Connecting to %s

%d bloggers like this: