Autonomy of cognition (part 2) — self-determination and conventionalism

Are organsms self-governed system? Do they produce and follow their own rules?

In the first part of the article, in reference to interaction and selective coupling of autopoietic systems in their environment, we compared two different understandings of autonomy: a strictly biological-material conception of autonomy, endorsed by Autopoietic Theory initiators, and a relational sensorimotor notion of autonomy implied in Enactivism. Despite their specific differences, both approaches acknowledge organizational closure as a necessary condition for autonomy.

Self-determination

Along with organizational closure, another critical aspect of autonomy, bringing out even more clearly the tension between the two approaches, is the concept of self-determination associated to autopoietic systems. In particular, the two theories have different ideas about the influence of external processes on an inner system (and vice versa) and about the presence or absence of a source of norms of adaptive behaviours. In other words, the issue can be formulated as follows: in which way the environment is able to influence the organism (and vice-versa)? What kind of causal relation can exist between external factors and a sensorimotor system? Should we talk about normativity?

Self-determination in Autopoietic Theory

In AT, autonomous systems are structurally determined, which means that any interaction is determined by their own structure and then specified by their dynamic constitution. Factors which are external to the system can never instruct any process of the inner system, what they can do is to trigger or initiate structural changes but without actually specifying them. This is to say that environmental events can produce a perturbation of the living system, but always according to the fixed structural dynamics of the latter. Furthermore, it must be said that in AT this form of structural determinism is not a specificity of organizationally closed systems but a common aspect of any physical system. [8]

Therefore, we can say that in AT — the operational domain and relational domain are strictly distinguished and they can never intersect — meaning that the sensorimotor interaction with the environment (relational domain) cannot be causally determinant for the closed process dynamics outlined in the previous post (operational domain), but indeed, it can only be influent in a relational way. It follows that the AT idea of determination leaves no room for any form of normativity, which indeed is explicitly refused by AT proponents, reducing it to a neutral sort of nomology. [2]

Self-determination in Enactivism

In the case of Enactivism, because of the existence of a variety of different developments of this approach [9], tracing a well-defined and univocal position about this topic is a little more difficult. However, generally speaking, we can say that, despite also for enactivists autonomous systems are self-determining unities and their adaptive behaviour is determined by their own rules, the principle of ‘non-intersection of the two domains’ mentioned above doesn’t seem to be fully accepted. Indeed, enactivists comes to consider the relational domain in a way constitutive of cognition, thus legitimating a specific ontological realm of cognitive systems and providing those latter with an intrinsic directionality (teleology) toward the environment.

By accepting this view, many enactivists endorse the emergence of a proper form of normativity, so thinking autopoietic systems as self-governed unities. This means that living beings would be able to generate their own norms and to behave according to those. This introduces de facto the idea of behaviour “with a purpose”, ascribing to organisms norms, goals and the ability to discern good and bad actions, sounding very much like having mental rules or a conceptual guidance aimed at self-maintenance of the system. [2]

However, a first problematic aspect in front of which the enactivist word splits apart is the following: if autonomy is what distinguishes a unity (i.e. an organism) within the environment, enabling the emergence of its cognitive agency, what exactly should we indicate as ‘embodiment of autonomy’? Which aspect of the organism, over its adaptive interaction, has to be considered as foundation of cognition, then as well as centre and source of the supposed normativity?

Autonomist enactivism or radical enactivism?

Within the enactive framework we can distinguish two broad positions on this matter. The first one could be called ‘autonomist enactivism’, and closely reflecting the concept of autonomy presented so far [6][3] takes a strict physiological-metabolic approach, opting for a neuro-centred (nervous system) view of autonomy and then of cognition; the second one, often called ‘radical enactivism’ [4], stressing the purely dynamic nature of cognition and in a way bypassing the drastic notion of “intentional autonomy”, considers sensorimotor coupling (continuous embodied interaction) as the core of cognitive development, so in a way clearing the field of any form of ‘content of cognition’, refusing what we previously mentioned as ‘mental rules’ or ‘meaning’(and representation).

Monism or Pluralism?

The assumption that autonomy (operational closure + self-determination) produces normativity begs also the question of how many “kinds” of autonomy can actually emerge and then how many levels and kinds of norms regulate an organism’s life. In this regard, among various positions we can discern two main directions — a monist approach and a pluralist approach –. Taking a monist approach means saying that, regardless how complex the organism is, only one kind of autonomy emerges, therefore, the whole organism’s self-maintenance system — from cell metabolism and homeostasis to intentional adaptive behaviour — is governed only by the norms generated at physical and autopoietic level and in any case always in causal subordination to that unique level of normativity. On the other hand, according to a pluralist viewpoint, in order to explain complex forms of self-maintenance, preservation and adaptation, such as multicellular homeostasis, ecological behaviour and social interaction, is necessary to conceive multiple types of autonomy emerging at different levels, producing at each level a new normative order (and a phenomenological domain). [2]

A conventionalist approach

To summarize, the issues highlighted in the first and second part of this article, which are only few of the problematic aspects related to the promising enactive view, are all multiple facets of the same original problem of explaining the embodiment of selective coupling, Namely, the issue of the naturalization of complex forms of adaptivity exhibited by complex systems like animals throughout their ecological interaction, producing cognitive and behavioural phenomena such as sense-making, individual-social agency or learning, integrating those latter with physiological-biological autopoietic processes.

Despite what it seems to be a not-negotiable position offered by AT and an unsolvable tension between AT and EN (and between different EN approaches), a suggestion for a reasonable way out is already provided in the origin of AT-EN approach and have been highlighted in the critical literature dealing with the issues presented so far ([7][2]) — taking a conventionalist approach. A conventionalist viewpoint, according to which cognition and intentional behaviour would only be arbitrary categories defined by the observer, doesn’t necessarily lead to discard any concept of cognition or adaptive coupling. In fact, acknowledging the impossibility to demonstrate the ontological implications of relational adaptive dynamics in strict naturalistic terms, doesn’t mean excluding their explanatory function. On the contrary, it could be said that concepts like adaptivity, selective coupling and cognition are essential — commentaries or explanatory propositions –about natural processes in the organism’s (or other autopoietic system) relational domain. [5] Following this line, those concepts would have a central scientific role of empirically and naturalistically based conventional concepts, similarly to notions like ‘species’ and ‘clade’ in biology.

In this perspective, as we shall see in future posts, theoretical models and useful concepts have been introduced in order to formalize a plausible framework for describing the essential dynamic nature of ecological interaction and the emergence of adaptive coupling in harmony with biological autopoiesis. For instance, with intent to overcome the issue of compatibility between autopoietic mono-autonomy (naturalist monism) and multi-autonomy-normativity agency (phenomenological pluralism), more complex forms of agency have been introduced, describing organisms as multi-level hierarchical-heterarchical systems in which multiple levels of normativity can emerge throughout embodied sensorimotor dynamics, under constraints of biological normativity but not reducible to it (formalized as complex models of non-linear neurodynamics). The notion of habit for example has been introduced to indicate — a self-sustaining pattern of sensorimotor coordination that is formed when the stability of a particular mode of sensorimotor engagement is dynamically coupled with the stability of the mechanisms generating it’’. [1]

[1] Barandiaran X. Mental life: A naturalized approach to the autonomy of cognitive agents. Unpublished doctoral dissertation, University of the Basque Country, Spain. 2008.

[2] Barandiaran XE. Autonomy and enactivism: Towards a theory of sensorimotor autonomous agency. Topoi. 2017 Sep 1;36(3):409–30.

[3] Di Paolo EA, Thompson E. The enactive approach. The Routledge handbook of embodied cognition. 2014 Jul 2:68–78.

[4] Hutto DD, Myin E. Radicalizing enactivism: Basic minds without content. Mit Press; 2012 Dec 14.

[5] Maturana H. Ultrastability… autopoiesis? Reflective response to Tom Froese and John Stewart. Cybernetics & Human Knowing. 2011 Jan 1;18(1–2):143–52.

[6] Thompson E. Mind in life. Harvard University Press; 2010 Sep 30.

[7] Villalobos M, Palacios S. Autopoietic theory, enactivism, and their incommensurable marks of the cognitive. Synthese. 2019 Sep 3:1–7.

[8] Villalobos M, Ward D. Living systems: Autonomy, autopoiesis and enaction. Philosophy & Technology. 2015 Jun 1;28(2):225–39.

[9] Ward D, Silverman D, Villalobos M. Introduction: The varieties of enactivism. Topoi. 2017 Sep 1;36(3):365–75.