Joseph Vukov

This paper develops a new framework for thinking about how understanding comes in degrees and what it means for an agent—human or artificial—to “understand” the world. We argue that understanding a target involves appreciating its modal structure, and that growth in understanding proceeds both vertically (penetrating deeper into the levels of the target) and horizontally (appreciating more of each level’s elements). We identify four levels of understanding: mapping a target’s elements (Level 1), grasping their powers and capacities (Level 2), identifying difference makers (Level 3), and apprehending necessitators (Level 4). We then apply this framework to the ongoing debate about whether AI systems possess world models, clarifying what is required for such models to be vehicles of understanding. While we remain neutral on whether current AI systems genuinely understand, we argue that our framework illuminates key issues and charts a path for future research.

The Ouroboros Effect refers to the fact that generative artificial intelligences (AIs) are increasingly being trained on AI-generated data. The effect results in the production of “boilerplate” or “biased” output, which does not pose any significant problem for what we call bias- and boilerplate-friendly domains. However, it does limit the applicability of generative AI to what we call heterodox domains: domains that build on yet depart from traditions, value novel and original contributions, and are context-dependent. Heterodox domains may therefore be resistant to AI-integration. And, significant problems are created when AI is applied to them. This has important implications for how we understand the application of AI to domains such as art, ethics, and education, as well as other domains that count as heterodox.

Standard accounts of mental powers are forward-looking: They individuate powers through their potential manifestations. In this article, the authors argue that mental powers are backward-looking as well. The backward-looking dimension of a power—what they call its lineage—opens up and constrains possibilities for potential manifestations, and in doing so, individuates mental powers just as readily as standard, forward-looking accounts. The idea of lineages thus incorporates standard accounts of mental powers, even while broadening the theoretical, rhetorical, and practical reach of an empowered account.

Moral bioenhancement presents the possibility of enhancing morally desirable emotions and dispositions. While some scholars have proposed that moral bioenhancement can produce virtue, we argue that within a virtue ethics framework moral bioenhancement cannot reliably produce virtue. Moreover, on a virtue ethics framework, the pursuit of moral bioenhancement carries moral risks. To make this argument, we consider three aspects of virtue—its motivational, rational, and behavioral components. In order to be virtuous, we argue, a person must (i) take pleasure in doing the right thing and have the correct motivational attitudes; (ii) reason correctly about what is called for in a particular ethical dilemma; and (iii) intentionally and continuously practice and cultivate virtues. These dimensions of morality—in short: precisely those emphasized in a virtue ethics framework—cannot be consistently or reliably met using existing moral bioenhancement technology.

AI poses a real and present danger. It contains the capacity to amplify social problems, drive a wedge further into our already-polarized society, and sow seeds of distrust in communities and personal relationships. When approached without a robust sense of human dignity, AI also threatens to undermine our self-understanding. To a degree beyond any previous technology, AI can make us forget ourselves. In this new era of AI, we must consciously make a choice: to stay human. In this book I provide a map and the tools for doing just that.

In this paper, we describe causal powers realism as a conjunction of four claims: causal powers are not reducible to counterfactuals; they are empirically-discoverable; they manifest effects in conjunction with partners; and their manifestations empower further manifestations. We describe four challenges to extended mind theory and for each show how an ontology of causal powers realism either avoids or dissolves the problem. We close by suggesting that causal powers realism isn’t a competitor with extended mind theory but rather a new way to understand what it means for minds to be extended.

Religious belief is often perceived as being in conflict with science--but does it have to be? Not usually, says Joseph Vukov. In this short, accessible guide, Vukov advances three models for Christians to utilize when navigating the relationship between science and faith: conflict, independence, and dialogue. He argues that dialogue is the ideal model to follow most of the time--but not necessarily all the time. Through a philosophical approach grounded in compelling real-world examples, Vukov shows how no single model can be universally adequate, and how Christians must proceed with discernment according to the nature of the matter at hand. Considering a wide variety of illustrative issues--including cosmology, evolutionary biology, extraterrestrial life, miracles, brain death, and theoretical physics--Vukov introduces and describes each of the three models of interaction between faith and science, surveys their applications, and evaluates the effectiveness of each. Throughout, he encourages Christians to embrace a spirit of intellectual humility and remember that, at their best, faith and science converge in their relentless human pursuit of truth.

When is it rational to undergo cognitive enhancement? In the case of what I’ll call massive cognitive enhancement, my answer is never. The reason is that one must base one’s decision to undergo massive cognitive enhancement on what I’ll call either phenomenal or non-phenomenal outcomes. If the former, the choice is not rational because massive cognitive enhancements are transformative and, I’ll argue with Paul (2015), transformative experiences cannot be chosen rationally. If the latter, the choice is not rational because it ought to be based at least partly on phenomenal outcomes. This argument, however, leaves open the idea that it may nonetheless be rational to choose massive cognitive enhancement for others—for example, one’s children. The article explores this possibility, arguing that choosing enhancement for others can be rational or moral, but not both.

Embodied, embedded, enactive, and extended theorists do not typically focus on the ontological frameworks in which they develop their theories. One exception is 4E theories that embrace New Mechanism. In this paper, we endorse the New Mechanist’s general turn to ontology, but argue that their ontology is not the best on the market for 4E theories. Instead, we advocate for a different ontology: causal powers realism. Causal powers realism posits that psychological manifestations are the product of mental powers, and that mental powers are empirically-discoverable features of individuals that account for the causal work those individuals do. We contend that causal powers realism provides a unifying framework for the central commitments of 4E theories, as well as additional resources for theorizing in a 4E framework. And while New Mechanism offers some of these resources as well, we argue that causal powers realism is ultimately the better of the two.

Fabiano argues that virtue theory offers the best “safety framework” for mitigating the risks of moral enhancement (1). He advances five desiderata for an ideal safety framework and then explains how virtue theory satisfies each. Among these desiderata is the “preservation of identity” (1). Fabiano argues that moral enhancement can safely preserve personal identity when carried out within the framework of virtue theory. We suggest Fabiano's argument for this conclusion falls short, since contra Fabiano’s claim, enhancing virtues may not preserve—and could even damage—personal identity. We draw on three sources of evidence: 1) virtue theory scholarship that argues for the importance of habituation for virtue formation, 2) Focquaert and Schermer’s (2015) distinction between active and passive enhancement and attendant endorsement of more active paths to virtue enhancement, and 3) empirical research suggesting that technologies which support moral enhancements may have damaging effects on personal identity.

Double Effect Donation claims it is permissible for a person meeting brain death criteria to donate vital organs, even though such a person may be alive. The reason this act is permissible is that it does not aim at one’s own death but rather at saving the lives of others, and because saving the lives of others constitutes a proportionately serious reason for engaging in a behavior in which one foresees one’s death as the outcome. Double Effect Donation, we argue, opens a novel position in debates surrounding brain death and organ donation, and does so without compromising the sacredness and fundamental equality of human life.

Conversation is a foundational aspect of philosophical pedagogy. Too often, however, philosophical research becomes disconnected from this dialogue, and is instead conducted as a solitary endeavor. We aim to bridge the disconnect between philosophical pedagogy and research by proposing a novel framework. Philosophy labs, we propose, can function as both a pedagogical tool and a model for conducting group research. Our review of collaborative learning literature suggests that philosophy labs, like traditional STEM labs, can harness group learning models such as Positive Interdependence Theory (PIT) to engage in meaningful discussion and execute projects and research. This article distills PIT into four essential tenets which we argue support student success at both the individual and group levels. Our argument is grounded in two case studies detailing our experiences facilitating different philosophy labs, and demonstrations of how they can foster the continued evolution of philosophical research and pedagogy beyond the single-occupancy armchair.

The BRAIN 2.0 Neuroethics Report reflects on the ways in which neuroscientific research may inform our understanding of concepts such as consciousness and empathy, and how advances in this understanding might in turn affect practices such as research on non-human animal primates. Generally, the Report calls for “the integration of neuroscience and neuroethics during the remaining years of the BRAIN initiative and beyond” (NIH 2019). In responding to the Report, the articles in this issue grapple with theoretical questions about what the integration of neuroscience and neuroethics might look like, as well as how specific practices might be affected. In this commentary, we don’t grapple with these questions ourselves. Instead, we propose a framework for debates about them. Specifically, we introduce the idea of ‘morally-salient features,’ suggesting that the concept may be helpful in providing a framework for the kinds of questions addressed by the BRAIN 2.0 Neuroethics Report as well as those tackled by the authors of this issue.

Dickert et al. (2020) effectively address how factors such as time limitations, stress, and illness severity in acute conditions warrant a deeper evaluation of how current consent processes serve patients. While data suggests that patients “prefer to be asked for permission upfront rather than waiving consent” (2), consent forms themselves “are frequently long and technical, follow rigid templates, and contain language that appears to prioritize institutional protection” (1). Such findings elucidate patients’ valuation of personal agency over settling for the “benefit of the doubt” that physicians and the consent forms they provide are acting in the patients’ best interests. In response, the authors recommend revisions to consent forms in terms of content, structure, and tone to better facilitate patient understanding beyond the mere conveyance of information. We aim to build Dickert et al.’s discussion of how “consent processes serve functions beyond facilitating an informed, autonomous decision” (2), but with a broader focus on an underrecognized purpose of the consent process: namely, reaffirming a patient’s status as a capable, rational agent. Specifically, we argue that patients can exercise this agency when they thoroughly understand the consent process, per Grimm’s (2012) conception of understanding.

Rainey et al. (forthcoming) discuss the complications that arise with assigning responsibility for brain computer interface (BCI)-mediated actions. Because BCI-mediated actions can differ from non-BCI-mediated actions in terms of control and foreseeability, the authors suggest that our ethical and legal evaluation of these actions may differ in important ways. While we take no issue with the authors’ discussion or conclusion, we also recognize the difficulty of grappling with the relationship between control, foreseeability, and moral responsibility practices, even without the additional complications introduced by BCI-mediation. In this commentary, we therefore consider BCI-mediated action against the backdrop of a different--perhaps less murky--normative framework: blame.

Berger (2019) argues effectively that “representativeness is more aptly understood as a variable that is multidimensional and continuous based on relational moral authority,” and also makes some useful suggestions about how taking this observation seriously might require changes in current patterns of practice regarding surrogates. But the essay raises additional important questions about how the Best Interest Standard (BIS) should be used among unrepresented patients and other patients as well because many surrogates besides those who “have no actionable knowledge of a patient’s preferences” find themselves in positions in which they need to determine, with the physician, what is in the patient’s best interests. In this commentary, we therefore provide support for the ethical superiority of BIS judgments made by a multi-stakeholder process (rather than solo decision-makers), and sketch a process by which such judgments might be made.

According to Macklin & Natanson (2019), one reason unusual practices can be misidentified as usual care is that “instead of using pertinent, accurate information describing usual care, investigators may rely on the opinion of ‘experts’ in the field, whose information may be out of date or otherwise inaccurate." We find Macklin & Natanson’s insights about misattributed expertise crucial, and suggest their discussion can be elucidated further by characterizing it in the context of Ballantyne (2018)’s recent exploration of what he calls epistemic trespassing. What insights can be drawn from this characterization? In this commentary, we highlight two. First, the characterization elucidates the serious consequences that can result from epistemic trespassing. Second, the characterization of Macklin & Natanson’s cases as instances of epistemic trespassing offers us a potential roadmap for overcoming some of the problems the authors identify.

Discussion of the ethics of memory modification technologies has often focused on questions about the limits of their permissibility. In the current paper, I focus primarily on a different issue: when is it rational to prefer MMTs to alternative interventions? My conclusion is that these conditions are rare. The reason stems from considerations of autonomy. When compared with other interventions, MMTs do a particularly poor job at promoting the autonomy of their users. If this conclusion is true, moreover, it provides a fresh perspective on debates about the permissibility of MMTs. On the one hand, for those who would limit the use of MMTs to a narrow range of circumstances, the conclusion that MMTs are rarely preferable gives them further reason to eye MMTs with suspicion. On the other hand, for those who view MMTs as permissible in a wide range of circumstances, the conclusion may deflate their endorsement.

Prolonged survival after the declaration of death by neurologic criteria creates ambiguity regarding the validity of this methodology. This ambiguity has perpetuated the debate among secular and nondissenting Catholic authors who question whether the neurologic standards are sufficient for the declaration of death of organ donors. Cardiopulmonary criteria are being increasingly used for organ donors who do not meet brain death standards. However, cardiopulmonary criteria are plagued by conflict of interest issues, arbitrary standards for candidacy, and the lack of standardized protocols for organ procurement. Combining the neurological and cardiopulmonary standards into a single protocol would mitigate the weaknesses of both and provide greater biologic and moral certainty that a donor of unpaired vital organs is indeed dead.

Cultural psychologists often describe the relationship between mind and culture as ‘dynamic.’ In light of this, we provide two desiderata that a theory about encultured minds ought to meet: the theory ought to reflect how cultural psychologists describe their own findings and it ought to be thoroughly naturalistic. We show that a realist theory of causal powers — which holds that powers are causally-efficacious and empirically-discoverable — fits the bill. After an introduction to the major concepts in cultural psychology and describing causal power realism, we use a case study — the effects of pathogen prevalence on culture and cognition — to show the explanatory capacities of the powers framework.

Patients in a minimally conscious state (MCS) fall into a different diagnostic category than patients in the more familiar vegetative states (VS). Not only are MCS patients conscious in some sense, they have a higher chance for recovery than VS patients. Because of these differences, we ostensibly have reason to provide MCS patients with care that goes beyond what we provide to patients with some VS patients. But how to justify this differential treatment? I argue we can’t justify it solely by looking to MCS patients’ actual levels of consciousness. We must look also to the ways in which these patients are potentially conscious. Specifically, I argue that certain sensible and well-regarded policy recommendations cannot be justified by looking solely to patients’ actual levels of consciousness, but that they can be justified by looking to patients’ potential for consciousness. Insofar then as we have prima facie reason to follow the recommendations, we also have prima facie reason to view the potential for consciousness as morally salient. If that is true, however, several important implications follow: first, if the potential for consciousness is morally salient, this translates into actual and pressing obligations to patients with disorders of consciousness generally. Second, if the potential for consciousness is morally salient, this has implications that go beyond our obligations to patients with disorders of consciousness. Indeed, if the potential for consciousness is morally salient, this adds an important layer of complexity to the normative landscape, a layer we may not have sufficiently noticed before.