Ascertaining the Uncertainty in Astrobiology: Leveraging
mathematical and computational models to aid with rational decision making in astrobiology

Abstract

Uncertainty is a defining feature of astrobiology, shaping our ability to apply mathematical tools and computational models for decision making. This thesis examines how uncertainty affects these methods and explores ways to navigate it within a mathematical and computational framework. While uncertainty over key probabilities — such as abiogenesis or technological advancement — limits tools like maximising expected utility and Bayesian statistics, it does not render decision making impossible. Instead, this thesis highlights both the challenges and potential strategies for addressing uncertainty in astrobiology.
Five primary conclusions emerge:

C1: I propose and defend a new definition of biosignature: any phenomenon for which biological processes are a known possible explanation and whose potential abiotic causes have been reasonably explored and ruled out (Gillen et al., 2023, p.1228). This is strong enough to be meaningful but leaves room for uncertainty over the list of possible explanations captured by the problem of unconceived alternatives (Stanford 2001, 2006a). This conclusion is discussed in Chapter Two and Chapter Three of this thesis.

C2: I propose a new corresponding definition of potential biosignature: any phenomenon for which biological processes are a known possible explanation but whose potential abiotic causes have not yet been reasonably explored and ruled out (Gillen et al., 2023, p.1238). This is argued for in Chapter Two and Chapter Three.

C3: In light of the vast and unexplored potential research areas constituting the field of astrobiology, theoretical
arguments exist for funding high-uncertainty, high-payoff research. These arguments do not apply to high-risk, high
payoff research. This conclusion is discussed in Chapter Four and Chapter Five of this thesis.

C4: High uncertainty surrounds key fundamental probabilities in astrobiology, such as the probability of abiogenesis.
This uncertainty means that astrobiologists disagreeing about fundamental probabilities might be beneficial. This
conclusion is found in Chapter Six.

C5: The deployment of computer models of scientific communities, such as astrobiology, requires a close
correspondence between the functional form of working real-world features and the functional form of working model
features. This is consistent with a structural realist view of computer modelling. This conclusion is discussed in Chapter
Seven.

Despite the challenges posed by uncertainty, it need not paralyse scientific progress. Recognising true uncertainty (as opposed to calculatable risk) can guide research toward fascinating and obscured discoveries. Moreover, computational models, while valuable, cannot produce certain conclusions from uncertain inputs. Yet, when designed with structural realism in mind and hence a strong correspondence to real-world systems, they remain useful tools for modelling epistemic communities in astrobiology.