Research

My research investigates the epistemic and ethical dimensions of artificial intelligence engineering: the philosophical implications of designing, manufacturing, and deploying autonomous intelligent systems in science and society.

My current projects extend that work in two main directions. The first rehabilitates the maker’s knowledge tradition to explain how making minds advances cognitive science. The second unpacks what AI engineers owe everyone else. In both directions I work with two kinds of tool at once: the careful conceptual analysis and logical argumentation of philosophy and the computational methods of artificial intelligence, including computer simulations, robots, and large-language-model text analysis.


Papers

Complex Cognitive Systems and the Maker’s Knowledge Tradition

When Does the Body Matter? A Dynamical Signature of Coupling Dependence in Evolved Neural Controllers
Collin Lucken
From Animals to Animats 18: Proceedings of the 18th International Conference on the Simulation of Adaptive Behavior (SAB 2026), Berlin. Springer Lecture Notes in Artificial Intelligence, forthcoming
Forthcoming
This paper presents a method for studying how continuous-time recurrent neural networks regulate behavior in virtual, highly idealized models of embodied, evolved agents, or ‘animats’. It asks the question: what patterns emerge in a simple animat’s neural dynamics when its sensorimotor coupling to the environment is severed or manipulated? The paper proposes three interventions that vary the incoming perturbations to the animat’s evolved neural controllers. It then applies these three interventions across 60 different controllers, evolved to find and move to a light source in their environment (i.e. phototaxis). The patterns of neural activity recorded in the animats during the phototaxis task are then measured using dynamical systems theory. Networks that react more to these interventions are described as those with greater embodiment dependence. In the analysis, embodiment dependence is found to be characterized by three largely independent dynamical dimensions: (1) input-driven amplification of neural activity, (2) distinctness of neural trajectories between trials, and (3) metastable-coordination richness (the number of transient coordination states the network visits).
Mindshaping and Learner-Centered Pedagogy: A Revealing Harmony
Collin Lucken & W. John Koolage
Teaching Philosophy, forthcoming
Forthcoming
The mindshaping research program, which is increasingly gaining attention in human social cognitive science, centers pedagogy as among the most important components of human cognition. By centering pedagogy as an essential social cognitive practice, the mindshaping research program is uniquely suited as a scientific framework for philosophy teaching. An explicit articulation of this connection demonstrates both the fruitfulness of the mindshaping research program and an opportunity for philosophy teachers to rethink or refine their teaching. This paper argues that philosophy teachers can usefully treat philosophy pedagogy as if it were a form of mindshaping. Our proposal is that adopting mindshaping as a ‘lens’ or regulative frame through which one conceives of philosophy teaching reorganizes what teachers see when they look at their own practice. Specifically, we show that viewing teaching through the lens of mindshaping conceived as a form of epistemic niche construction clarifies what good teaching aims to do: shape learners’ skills and dispositions to build and inhabit epistemic environments that extend their cognitive capacities.
Scientific Works and the Transfer of Technique
Collin Lucken
European Journal for Philosophy of Science
Revise and resubmit
Techniques routinely transfer between scientific domains that share no theoretical framework. Photolithographic patterning migrates from semiconductor physics to microfluidics; adaptive optics moves from astronomy to retinal imaging; cryogenic vitrification transfers from materials science to structural biology. Existing philosophical accounts presuppose that cross-domain coherence requires theoretical or social mediation. Counter to that presupposition, I introduce the concept of scientific works: deliberate material interventions that, by restructuring an organism-environment system, durably extend the set of invariants in that system that the community can reliably exploit. Drawing on Gibson’s ecological psychology and niche construction theory, I argue abductively that scientific works provide the best explanation of technique transfer. A shared material regularity is necessary for transfer but not sufficient. Transfer succeeds when the target domain lies within the range of transformations across which a technique’s invariant persists and when the receiving community restructures its own organism-environment system so as to exploit that invariant; it fails when either condition is unmet.
When Does the Body Matter? Separating Input Sensitivity from Loop Dependence in Evolved Agents
Collin Lucken
Adaptive Behavior
Revise and resubmit
Whether the body is part of cognition or merely a cause of it turns on an inference from coupling to constitution that behavioral evidence cannot settle. This paper asks a prior question: what kind of dependence on ongoing sensorimotor coupling does a controller have, and how could one tell? I argue that the divergence of neural trajectories under open-loop “ghost” conditions measures only input sensitivity. Claims of loop dependence need two counterfactual controls instead: the same perturbation applied with the loop closed and with it cut, and substitution of a controller with the agent’s own steady states but no internal dynamics (its reflex shadow). I applied these controls to several hundred recurrent-network agents evolved to find a light with two sensors or with one, and tested the core claims on a second, preregistered set. Open-loop networks forget their initial state and follow their input, the closed loop compensates for perturbation in every agent, and a reflex suffices for most two-sensor agents and for no one-sensor agent. In exploratory analyses, a frequency-response analysis of each network predicts which agents need their dynamics. Those agents pair a nearly flat steady-state map with a strong response at the frequency of the sensory oscillation their own movement produces, which is the signature of an ordinary dynamic compensator. Dependence profiles, however refined, therefore cannot separate cognitive systems from controlled ones. The coupling–constitution question turns on the kind of process involved, and what these measures contribute is an empirical description of that process.

History of Philosophy of Science

The Erotetic Logic of William Whewell’s Fundamental Antithesis
Collin Lucken
Studies in History and Philosophy of Science, forthcoming
Forthcoming
Whewell’s Fundamental Antithesis of Philosophy is a curious feature of his thought, as it seems to serve as the foundation for his entire system while also remaining especially mysterious. Recent Whewell scholarship has taken note of the importance of the Antithesis for Whewell, appealing to it to explain his views on necessity (Morrison 1997), induction (Snyder 1997, 2008; Forster 2011, Dethier 2018), and mind (van der Merwe 2023). Despite its importance, the Fundamental Antithesis has received relatively little attention compared to other topics in Whewell’s philosophy such as those just listed. In this essay, I aim to rectify that lack of attention. More specifically, I am concerned with determining how we as Whewell’s readers should approach his mysterious Antithesis. My conclusion is that the Fundamental Antithesis is best understood through erotetics, or the logic of question and answer. This is opposed to how the Fundamental Antithesis is most frequently described as being about mind and world merging.

Engineering, Artificial Intelligence, and the Common Good

An Identity-Based Ethics for Engineering
Collin Lucken
Science and Engineering Ethics, forthcoming
Forthcoming
Engineering ethics education often relies on case-study methods focused on disaster prevention and professional codes. Recent work in virtue ethics and professional identity has begun to develop a possible alternative, but the philosophical foundations of engineering’s ethical character remain underdeveloped. This paper argues that engineering constitutes a morally significant professional identity: a life-shaping practice whose internal goods are oriented toward the development of central human capabilities. Drawing on Appiah’s identity framework, MacIntyre’s theory of practices and his account of compartmentalization, the capability approach as formalized by Robeyns, and Wolfendale’s regulative-ideal account of professional integrity, I analyze engineering identity through institutional analyses of France’s École Polytechnique and the United States’ Accreditation Board for Engineering and Technology (ABET). I argue that engineering’s normative regulative ideal is the development of human capabilities, and that this orientation provides the philosophical grounding for an identity-based approach to engineering ethics education that connects ethical development to the cultivation of engineering character.
Collin Lucken & Tim Elmo Feiten
Studies in History and Philosophy of Science, Volume 110, 2025, pp. 55–64
Published
This article explores new potentials for productive dialogue between public engagement with science (PEWS) and radical embodied cognitive science (RECS). The concept of participatory sense-making from enactive cognitive science provides an account of why active, dialogical engagement in science communication is so effective. After establishing the connection between PEWS and RECS, the authors motivate the need for “participatory cognitive strategies” and present a case study showing their potential in actively involving different groups of stakeholders throughout the development of large-scale AI systems, contributing to ongoing debates about the meaning of “democratizing AI.”

Other Work

Engineering Progress in Science
Collin Lucken
PhD Dissertation, University of Cincinnati, 2025
Completed
This dissertation develops an operative account of scientific progress that foregrounds practical cognition, material construction, and engineering alongside traditional intellectual reasoning. I argue that scientific works—including instruments, models, calibration routines, and experimental systems—are epistemically significant achievements that standard accounts of progress overlook. The framework reveals how engineering practices generate distinctive forms of knowledge and contribute to scientific progress on their own terms.
A Guide for Academic Researchers Conducting Science Outreach
Melissa Jacquart, Amanda Corris, Andrew Evans, Tim Elmo Feiten, Collin Lucken & Angela Potochnik
Elements in Public Engagement with Science, Cambridge University Press
Under Contract