2022-01-21

a generalised evolutionary theory of knowledge

introduction

The concept of evolution originally arose in Biology, yet, it has managed to find its way into a particular branch of contemporary epistemic¹ thought known as Evolutionary Epistemology (EE) (Bradie and Harms 2020). Having breached its disciplinary boundary by finding application within the domain of Philosophy, evolution as a concept, has been elevated to the rank of superconcept, according to Wilson (Wilson 2010). And rightfully so, having proven valuable to “one of the greatest philosophers of science of the twentieth century” (Thornton 2021), Karl Popper, as evidenced by works such as “Objective Knowledge: an evolutionary approach” (Popper 2012).

However, although I will offer a brief overview of the central themes of EE, the focus of this essay is not in highlighting the significance of the ideas this particular branch has generated. Rather, I hope to illustrate the power of tackling knowledge through an evolutionary lens, by forwarding a plausible, generalised description of how knowledge in individual knowers² can change over time. The aim is to abstract evolutionary epistemic ideas to extend their applicability beyond organisms to account for potential inorganic or artificial systems. In the process, I hope to touch on the role played by evolution in this generalised theory, as well as how such a theory relates to existing works in EE.

I will begin by exclusively outlining the features of EE necessary for contextualizing my argument, before presenting it. After which, I will address the potential problems and weaknesses it faces, briefly touching on how these can be approached or even resolved.

a brief and incomplete overview of Evolutionary Epistemology (EE)

EE exists as a branch of naturalized epistemology (NE) (Quine 1969), a school of thought born out of the failures of positivism³ and logical-positivism³ and the perceived intractability of the reference problem⁴ (Gontier 2006). Seeking to address these shortcomings, NE took a non-traditional approach in the view that philosophical investigations into knowledge should be closely connected to “empirical (’natural’) science” (Rysiew 2021). In this way, EE is a particular instantiation of NE that centers around the conjecture that epistemology is to be understood as a “product of biological evolution” and thus, should be “studied from within evolutionary biology” (Gontier 2021).

Importantly, due to the fact that there are “disagreements on what evolution in general is” (Gontier 2021), there are different types of EE to reflect the variety of evolutionary theories presented by biologists (Millstein 2021). In general, these can be distinguished by the perspective from which evolution is taken (environment, organism, genes…). In the case of this essay, I will be endorsing ideas drawn from the non-adaptationist or constructivist approach that regards evolution from the point of view of the organism, more on this to come.

It’s also worth noting that attention within EE is divided into two main programs: the Evolution of Epistemological Mechanisms (EEM) and the Evolutionary Epistemology of Theories (EET). The former examines the evolution of physical cognitive structures that underly cognitive activity such as organisms’ brains, sensory systems and motor systems (Bradie and Harms 2020). The latter tracks the evolution of the product of these structures like beliefs, ideas, cultural norms and language (Bradie and Harms 2020). It is also the program of which this essay is a part, which is to say that I will not be addressing the nature of underlying mechanisms, only assuming their existence.

Finally, another distinction is drawn between what are known as phylogenetic and ontogenetic considerations (Bradie and Harms 2020). Both terms are imported from evolutionary biology and refer to the scope at which development is considered. While phylogeny tracks the development of traits across generations, ontogeny concerns itself with the development of traits over the lifetime of individual organisms. Concerning the theory in this essay, note that it falls under an ontogenetic analysis as it considers how knowledge changes within a singular system and not its generations.

a generalised theory of knowledge

The crux of evolutionary epistemic theories lies in their respective conceptions of the organism-environment relationship. After all, this is one of the foundational tenets of biology, let alone evolution. However, without entering into discussions surrounding the nature of biological individuals⁵, my stipulation is that an evolutionary theory of knowledge can be generalised to the relationship between an abstract knower and an environment, and needn’t be restricted to the domain of organic systems/organisms.

What I mean to suggest by this is that whether a system is organic (ie: is an organism) is independent of its ability to produce knowledge⁶ through an evolutionary process. Put another way, knowledge can be a product of an evolutionary process in an inorganic or artificial system by means of its relationship with its environment and the interactions it has with it.

My hope is to demonstrate this by presenting an abstracted evolutionary theory of knowledge, inspired by features of Lewontin’s (Lewontin 2001) dialectic approach⁷.

the theory

Since I am trying to offer a generalised description, the classical entities of “organism” and “environment” are abstracted. To reflect this, I will introduce the term “knower” to replace that of “organism”, but make no changes with regards to “environment”. Instead, I will provide an extremely simplistic interpretation of one to both maximise its generality, but also keep away from the metaphysical intricacies of a fully defined one. Having done that, I will detail how evolution can be applied to the knowledge produced through knower-environment interactions, ending on how agency determines the modes of evolution (Millstein 2021).

no longer just an organism but a ‘knower’

One can only speak of the development of knowledge in organisms, as EE does, after having assumed that organisms are entities capable of knowledge in the first place. Hence, a more generalised term for the entities under consideration in this theory will be that of a “knower”. This refers to any abstract entity with the capacity to produce knowledge. By which I mean the entity has the sufficient underlying mechanisms to produce knowledge (whatever these may be), as deemed necessary by the EEM⁸.

Importantly, under this theory, it is assumed that such mechanisms would necessarily include a capacity for a knower to interface with an environment, in the form of both receiving external input, but also generating externalised output. This seems self-evident because for an entity to gain knowledge of the environment it is part of requires there be a way for the two to be brought into relation with one another. This is achieved through the two necessary interfacing properties I will ground under interactability.

Additionally, a knower’s ability to output to an environment grants it agency, which is defined to be its capacity to act on an environment (Schlosser 2019). Because of this, the term “agent” shall use be used interchangeably with “knower” from here on.

the environment

For the sake of simplicity, I will define an environment as a collection of interactable objects that can accommodate a potential agent. Crucially, this environment is objective in so far as its existence is independent of that of the knower. This isn’t to say that the state of the environment does not change or is entirely autonomous⁹, only that an environment can be described without any agents within it.

the dialectical knower-environment relationship

By introducing interactability as a core property of both agents and environments, we allow for their relationship to become dialectical and not simply dualistic, because each party can act on the other. In a traditional biological context, Richard Levins calls this their “mutual interpenetration” (Hahlweg and Hooker 1989).

And as anticipated, this construction of the knower-environment relationship reflects the original biological constructivist/non-adaptationist¹⁰ approach (Gontier 2021) where “…the claim that the environment of an organism is causally independent of the organism, and the changes in the environment are autonomous and independent of changes in the species itself, is clearly wrong.” (Lewontin 2001).

Having said that, this raises the question of how knowledge products can arise and develop from only interactability and a dialectical relationship between knower and environment?

knowledge as representation

According to Wuketits: “Organisms do not simply correspond to their surroundings and do not get everything that is ‘out there’ but rather form their own ‘picture’ of what is around and react adequately, according to the specific requirement of their lives.” (Wuketits 2001)

Through interacting with an environment an agent’s cognitive mechanisms can generate knowledge in the form of a representation or model of their environment. Or in the case of organisms, what Wuketits calls a “`picture`”.

Naturally, these representations will be inaccurate and or incomplete representations of an agent’s environment. Especially, when one considers the resolution and limitations of the mechanisms by which an agent interfaces with its environment, as well as the latter’s relative complexity and scale.

In consequence, unless an agent takes the form of Laplace’s demon¹¹, the actions an agent’s knowledge will inspire will not always result in the observable external effect that was intended.

conflict as the driver of change

An agent acts on the knowledge of its environment that it has produced¹². Consider the case where the representation of its environment stands diametrically opposed to its true nature. That is, there is no aspect of the description of the state of the environment that overlaps with an agent’s model of it. In this case, a knower’s agency is effectively reduced to zero by completely restricting its capacity to fulfill its underlying intention. This is intuitive when one considers action as the fulfillment of intention (Schlosser 2019) and acknowledges that an agent would be acting on a model of an environment that in no way exists.

Hence, even in less dramatic cases, the inevitable discrepancies between an environment and an agent’s representation/knowledge of it, result in conflicts that arise when an agent acts on its environment. These conflicts undermine the agency of the knower providing an opportunity for the state of its knowledge to change in response.

The way in which an agent will resolve these conflicts by modifying its knowledge, or put in more evolutionary terms: the selective methods by which existing beliefs an agent has of its environment are preserved, and new ones adopted, are determined by its intentionality¹³.

agency determines evolutionary modes

At first, Darwin didn’t employ the term “evolution” but referred to the process as “descent with modification” (Darwin 2004). In spite of seeming like a fairly general account of simple change over time, evolution is special in that it describes change as a process of iterative modifications under specific modes (Millstein 2021). These modes offer explanations of which modifications occur and under what conditions, effectively detailing the process of “selection”. These range from genetic drift to mutation, with the most famous mode natural selection. In its case, there needs to be heritability and variation in a collection of organisms, so that the “struggle for existence” may act on the variations to “select” those conducive to survival (Gildenhuys 2019).

In the case of our theory, survival isn’t necessarily the foundational urge or intention behind the actions of an agent. On the contrary, the theory aims to assume no particular intention. What this means is that because differences between environment and representation can only be revealed to an agent through its actions, the kind of actions it takes will determine which conflicts arise and which part of its representation is undermined.

In consequence, what “selects” the parts of an agent’s knowledge to be modified and to be preserved is in fact determined by the intention that lies behind its actions, whatever it may be. In this way, a knower’s agency can be maximised and its conflicts with its environment minimised, through the process of developing its knowledge.

problems and weaknesses

Unfortunately, no part of the theory that I have laid out is immune to criticism.

The most obvious vulnerabilities lie in its dependence on a hierarchy of contentious philosophical ideas. Its naturalist epistemological underpinning exposes it to the more general attacks directed towards naturalism (Papineau 2021). Likewise, there are other views within EE on which evolutionary perspective should be taken. To the extent that if non-adaptationist approaches were to be undone, the theory in this essay would be quite easily undermined.

Tangentially, the argument presented also seems incomplete in parts.

How does it account for more “complex” knowledge products like language and culture, through nothing but selective processes underlying the agent-environment relationship?

What’s more, what of derivable knowledge? If a knower can arrive at new knowledge from existing knowledge without interacting with an environment wouldn’t this mean that an agent’s knowledge could develop independently of its environment?

addressing problems and weaknesses

Any body of knowledge has interrelated parts that lead to complex dependencies. Although I won’t argue that the ideas underlying this theory can be brought into question, a person cannot be expected to defend the history of all the ideas their work builds on. Doing so would grind any form of progress to a snail’s pace. Hence, if this theory is to be undone by the ideas that inspired it, so be it; it will be one amongst many to be abandoned. Until then, there exist works dedicated to the defense of particular philosophical positions like that of naturalized epistemology (Quine 1969).

As for the production of more complex structures of knowledge such as language. I would argue that these are entirely enabled by specific mechanisms. That is, there is no relationship or interaction any agent can have with its environment that would produce complex structures like language, unless specific mechanisms responsible for such structures are present within the agent.

Finally, I would concede that it is conceivable for a knower to update its knowledge independently of its environment, solely through its internal mechanisms. However, even if one assumes this to be the case this does nothing to dispel the presented notion that the development of knowledge can be explained through environment interactions. If anything, it could be seen as offering insight into another type of knowledge development in agents. Either way, I would still stand by the argument that the way in which its knowledge will be modified, will be governed by its agency with regards to a particular environment.

conclusion

Regardless of the shortcomings of the theory I presented, I hope to have offered insight into how the concept of evolution has found its way at the forefront of contemporary epistemology. At the same time, I hope to have offered a small glimpse into how evolution is and can be applied in this branch of Philosophy, as well as the the value of its contributions in doing so.

References