Explanation is a foundational goal in the exact sciences. Besides the contemporary considerations on ‘description’, ‘classification’, and ‘prediction’, we often see these terms in thriving applications of artificial intelligence (AI) in chemistry hypothesis generation. Going beyond describing ‘things in the world’, these applications can make accurate numerical property calculations from theoretical or topological descriptors. This association makes an interesting case for a logic of discovery in chemistry: are these induction-led ventures showing a shift in how chemists can problematize research questions? (...) In this article, I present a fresh perspective on the current context of discovery in chemistry. I argue how data-driven statistical predictions in chemistry can be explained as a quasi-logical process for generating chemical theories, beyond the classic examples of organic and theoretical chemistry. Through my position on formal models of scientific explanation, I demonstrate how the dawn of AI can provide novel insights into the explanatory power of scientific endeavors. (shrink)

Among the factors necessary for the occurrence of some event, which of these are selectively highlighted in its explanation and labeled as causes — and which are explanatorily omitted, or relegated to the status of background conditions? Following J. S. Mill, most have thought that only a pragmatic answer to this question was possible. In this paper I suggest we understand this ‘causal selection problem’ in causal-explanatory terms, and propose that explanatory trade-offs between abstraction and stability (...) can provide a principled solution to it. After sketching that solution, it is applied to a few biological examples, including to a debate concerning the ‘causal democracy’ of organismal development, with an anti-democratic (though not a gene-centric) moral. (shrink)

We develop and apply a multi-dimensional account of explanatory depth towards a comparative analysis of inflationary and bouncing paradigms in primordial cosmology. Our analysis builds on earlier work due to Azhar and Loeb (2021) that establishes initial conditions fine-tuning as a dimension of explanatory depth relevant to debates in contemporary cosmology. We propose dynamical fine-tuning and autonomy as two further dimensions of depth in the context of problems with instability and trans-Planckian modes that afflict bouncing and inflationary approaches (...) respectively. In the context of the latter issue, we argue that the recently formulated trans-Planckian censorship conjecture leads to a trade-off for inflationary models between dynamical fine-tuning and autonomy. We conclude with the suggestion that explanatory preference with regard to the different dimensions of depth is best understood in terms of differing attitudes towards heuristics for future model building. (shrink)

The philosophical debate around the impact of machine learning in science is often framed in terms of a choice between AI and classical methods as mutually exclusive alternatives involving difficult epistemological trade-offs. A common worry regarding machine learning methods specifically is that they lead to opaque models that make predictions but do not lead to explanation or understanding. Focusing on the field of molecular biology, I argue that in practice machine learning is often used with explanatory aims. (...) More specifically, I argue that machine learning can be tightly integrated with other, more traditional, research methods and in a clear sense can contribute to insight into the causal processes underlying phenomena of interest to biologists. One could even say that machine learning is not the end of theory in important areas of biology, as has been argued, but rather a new beginning. I support these claims with a detailed discussion of a case study involving gene regulation by microRNAs. (shrink)

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealised explanation game in which players collaborate to find the best explanation for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to (...) explore overlapping causal patterns of variable granularity and scope. We characterise the conditions under which such a game is almost surely guaranteed to converge on a optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions. (shrink)

When it comes to epistemic normativity, should we take the good to be prior to the right? That is, should we ground facts about what we ought and ought not believe on a given occasion in facts about the value of being in certain cognitive states (such as, for example, the value of having true beliefs)? The overwhelming answer among contemporary epistemologists is “Yes, we should.” This essay argues to the contrary. Just as taking the good to be prior to (...) the right in ethics often leads one to sanction implausible trade-offs when determining what an agent should do, so too, this essay argues, taking the good to be prior to the right in epistemology leads one to sanction implausible trade-offs when determining what a subject should believe. Epistemic value—and, by extension, epistemic goals—are not the explanatory foundation upon which all other normative notions in epistemology rest. (shrink)

Definition of the problem The umbrella term “explicability” refers to the reduction of opacity of artificial intelligence (AI) systems. These efforts are challenging for medical AI applications because higher accuracy often comes at the cost of increased opacity. This entails ethical tensions because physicians and patients desire to trace how results are produced without compromising the performance of AI systems. The centrality of explicability within the informed consent process for medical AI systems compels an ethical reflection on the trade- (...) class='Hi'>offs. Which levels of explicability are needed to obtain informed consent when utilizing medical AI? Arguments We proceed in five steps: First, we map the terms commonly associated with explicability as described in the ethics and computer science literature, i.e., disclosure, intelligibility, interpretability, and explainability. Second, we conduct a conceptual analysis of the ethical requirements for explicability when it comes to informed consent. Third, we distinguish hurdles for explicability in terms of epistemic and explanatory opacity. Fourth, this then allows to conclude the level of explicability physicians must reach and what patients can expect. In a final step, we show how the identified levels of explicability can technically be met from the perspective of computer science. Throughout our work, we take diagnostic AI systems in radiology as an example. Conclusion We determined four levels of explicability that need to be distinguished for ethically defensible informed consent processes and showed how developers of medical AI can technically meet these requirements. (shrink)

We argue that deliberative decision-making that is inclusive, transparent and accountable can contribute to more trustworthy and legitimate decisions on difficult ethical questions and political trade-offs during the pandemic and beyond.

Most mental disorders affect only a small segment of the population. On the reasonable assumption that minds or brains are prone to occasional malfunction, these disorders do not seem to pose distinctive explanatory problems. Depression, however, because it is so prevalent and costly, poses a conundrum that some try to explain by characterizing it as an adaptation—a trait that exists because it performed fitness-enhancing functions in ancestral populations. Heretofore, proposed evolutionary explanations of depression did not focus on thought processes; (...) instead, they emphasized that it facilitates navigation of adverse social circumstances or promotes immune response to infectious agents. According to a new hypothesis, the “analytical rumination hypothesis” (ARH), however, depression’s crucial adaptive trait is rumination—negative, intrusive thought. ARH holds that, (i) social dilemmas trigger depressed mood; (ii) depressed mood induces changes in body systems that facilitate ruminative analysis aimed at solving dilemmas; and, (iii) depressive rumination is a fitness-enhancing trait that was selected for in evolutionary time. Jointly, (i)~(iii) imply that we should not think of rumination as a disorder; instead, it is a trade-off, an eminently rational one. In the same way that fever solves a problem—coordination of the immune system in response to infection—so too does depressive rumination solve a problem, a social dilemma, albeit at the cost of inducing anhedonia and other maladies. But they argue that the cost is worthwhile, something that should be endured “until the problem is solved.” First, we argue that there are two distinct types of rumination, brooding and pondering; the former is associated with a disposition for depression, not the latter. But only the latter has the problem-solving capabilities that ARH requires. Second, recent brain imaging studies of depression reveal resting state hypoactivity in lateral regions and hyperactivity in paralimbic regions; this asymmetric pattern correlates with heightened levels of brooding, self-focused rumination. In other words, on the personal level, patients are trapped within self, isolated from the external world and suffused with negative affect; on the subpersonal level, this pattern is reflected by an asymmetric pattern of lateral vs. paralimbic resting state activity. Third, we proceed to conjecture that rational responses (e.g., pondering) to social dilemmas are those that strike a balance between internal and external considerations in the process of belief formation. Fourth, because the asymmetric resting state activity blocks those who suffer with depression from accessing and processing potentially positive stimuli from the external world, the capacity for rational, analytic response—hence, problem-solving—is constrained. Fifth, it follows that, although there might be conditions for which suffering should be endured rather than pharmacologically alleviated, depression is not one of those. Indeed, in view of the effects of the asymmetric resting state pattern, it is unlikely that depressive rumination would have been useful even for ancestral populations. (shrink)

The aim of this study is to justify the belief that there are biological normative mechanisms that fulfill non-trivial causal roles in the explanations (as formulated by researchers) of actions and behaviors present in specific systems. One example of such mechanisms is the predictive mechanisms described and explained by predictive processing (hereinafter PP), which (1) guide actions and (2) shape causal transitions between states that have specific content and fulfillment conditions (e.g. mental states). Therefore, I am guided by a specific (...) theoretical goal associated with the need to indicate those conditions that should be met by the non-trivial theory of normative mechanisms and the specific models proposed by PP supporters. In this work, I use classical philosophical methods, such as conceptual analysis and critical reflection. I also analyze selected studies in the field of cognitive science, cognitive psychology, neurology, information theory and biology in terms of the methodology, argumentation and language used, in accordance with their theoretical importance for the issues discussed in this study. In this sense, the research presented here is interdisciplinary. My research framework is informed by the mechanistic model of explanation, which defines the necessary and sufficient conditions for explaining a given phenomenon. The research methods I chose are therefore related to the problems that I intend to solve. In the introductory chapter, “The concept of predictive processing”, I discuss the nature of PP as well as its main assumptions and theses. I also highlight the key concepts and distinctions for this research framework. Many authors argue that PP is a contemporary version of Kantianism and is exposed to objections similar to those made against the approach of Immanuel Kant. I discuss this thesis and show that it is only in a very general sense that the PP framework is neo-Kantian. Here we are not dealing with transcendental deduction nor with the application of transcendental arguments. I argue that PP is based on reverse engineering and abduction inferences. In the second part of this chapter, I respond to the objection formulated by Dan Zahavi, who directly accuses this research framework of anti-realistic consequences. I demonstrate that the position of internalism, present in the so-called conservative PP, does not imply anti-realism, and that, due to the explanatory role played in it by structural representations directed at real patterns, it is justified to claim that PP is realistic. In this way, I show that PP is a non-trivial research framework, having its subject, specific methods and its own epistemic status. Finally, I discuss positions classified as the socalled radical PP. In the chapter “Predictive processing as a Bayesian explanatory model” I justify the thesis according to which PP offers Bayesian modeling. Many researchers claim that the brain is an implemented statistical probabilistic network that is an approximation of the Bayesian 7 rule. In practice, this means that all cognitive processes are to apply Bayes' rule and can be described in terms of probability distributions. Such a solution arouses objections among many researchers and is the subject of wide criticism. The purpose of this chapter is to justify the thesis that Bayesian PP is a non-trivial research framework. For this purpose, I argue that it explains certain phenomena not only at the computational level described by David Marr, but also at the level of algorithms and implementation. Later in this chapter I demonstrate that PP is normative modeling. Proponents of the use of Bayesian models in psychology or decision theory argue that they are normative because they allow the formulation of formal rules of action that show what needs to be done to make a given action optimal. Critics of this approach emphasize that such thinking about the normativity of Bayesian modeling is unjustified and that science should shift from prescriptive to descriptive positions. In a polemic with Shira Elqayam and Jonathan Evans (2011), I show that the division they propose into prescriptivism and Bayesian descriptivism is apparent, because, as I argue, there are two forms of prescriptivism, i.e. the weak and the strong. I argue that the weak version is epistemic and can lead to anti-realism, while the strong version is ontic and allows one to justify realism in relation to Bayesian models. I argue that a weak version of prescriptivism is valid for PP. It allows us to adopt anti-realism in relation to PP. In practice, this means that you can explain phenomena using Bayes' rule. This does not, however, imply that they are Bayesian in nature. However, the full justification of realism in relation to the Bayesian PP presupposes the adoption of strong prescriptivism. This position assumes that phenomena are explained by Bayesian rule because they are Bayesian as such. If they are Bayesian in nature, then they should be explained using Bayesian modeling. This thesis will be substantiated in the chapters “Normative functions and mechanisms in the context of predictive processing” and “Normative mechanisms and actions in predictive processing”. In the chapter “The Free Energy Principle in predictive processing”, I discuss the Free Energy Principle (hereinafter FEP) formulated by Karl Friston and some of its implications. According to this principle, all biological systems (defined in terms of Markov blankets) minimize the free energy of their internal states in order to maintain homeostasis. Some researchers believe that PP is a special case of applying this principle to cognition, and that predictive mechanisms are homeostatic mechanisms that minimize free energy. The discussion of FEP is important due to the fact that some authors consider it to be important for explanatory purposes and normative. If this is the case, then FEP turns out to be crucial in explaining normative predictive mechanisms and, in general, any normative biological mechanisms. To define the explanatory possibilities of this principle, I refer to the discussion of its supporters on the issue they define as the problem of continuity and discontinuity between life and mind. A critical analysis of this discussion and the additional arguments I have formulated have allowed me to revise the explanatory ambitions of FEP. I also reject the belief that this principle is necessary to explain the nature of predictive mechanisms. I argue that the principle formulated and defended by Friston is an important research heuristic for PP analysis. 8 In the chapter “Normative functions and mechanisms in predictive processing”, I start my analyzes by formulating an answer to the question about the normative nature of homeostatic mechanisms. I demonstrate that predictive mechanisms are not homeostatic. I defend the view that a full explanation of normative mechanisms presupposes an explanation of normative functions. I discuss the most important proposals for understanding the normativity of a function, both from a systemic and teleosemantic perspective. I conclude that the non-trivial concept of a function must meet two requirements which I define as explanatory and normative. I show that none of the theories I have invoked satisfactorily meets both of these requirements. Instead, I propose a model of normativity based on Bickhard's account, but supplemented by a mechanistic perspective. I argue that a function is normative when: (1) it allows one to explain the dysfunction of a given mechanism; (2) it contributes to the maintenance of the organism's stability by shaping and limiting possible relations, processes and behaviors of a given system; and when (3) (according to the representational and predictive functions) it enables explaining the attribution of logical values of certain representations / predictions. In such an approach, a mechanism is normative when it performs certain normative functions and when it is constitutive for a specific action or behavior, despite the fact that for some reason it cannot realize it either currently or in the long-term. Such an understanding of the normativity of mechanisms presupposes the acceptance of the epistemic hypothesis. I argue that this hypothesis is not cognitively satisfactory, and therefore the ontic hypothesis should be justified, which is directly related to adopting the position of ontic prescriptivism. For this reason, referring to the mechanistic theory of scientific explanations, I formulate an ontical interpretation of the concept of a normative mechanism. According to this approach, a mechanism or a function is normative when they perform such and such causal roles in explaining certain actions and behaviors. With regard to the normative properties of predictive mechanisms and functions, this means that they are the causes of specific actions an organism carries out in the environment. In this way, I justify the necessity of accepting the ontic hypothesis and rejecting the epistemic hypothesis. The fifth chapter, “Normative mechanisms and actions in predictive processing”, is devoted to the dark room problem and the related exploration-exploitation trade-off. A dark room is the state that an agent could be in if it minimized the sum of all potential prediction errors. I demonstrate that, in accordance with the basic assumption of PP about the need for continuous and long-term minimization of prediction errors, such a state should be desirable for the agent. Is it really so? Many authors believe it is not. I argue that the test of the value of PP is the possibility of a non-trivial solution of this problem, which can be reduced to the choice between active and uncertainty-increasing exploration and safe and easily predictable exploitation. I show that the solution proposed by PP supporters present in the literature does not enable a fully satisfactory explanation of this dilemma. Then I defend the position according to which the full explanation of the normative mechanisms, and, subsequently, the solution to the dilemma of exploration and exploitation, involves reference to the existence of constraints present in the environment. The constraints 9 include elements of the environment that make a given mechanism not only causal but also normative. They are therefore key to explaining the predictive mechanisms. They do not only play the role of the context in which the mechanism is implemented, but, above all, are its constitutive component. I argue that the full explanation of the role of constraints in normative predictive mechanisms presupposes the integration of individual models of specific cognitive phenomena, because only the mechanistic integration of PP with other models allows for a non-trivial explanation of the nature of normative predictive mechanisms that would have a strong explanatory value. The explanatory monism present in many approaches to PP makes it impossible to solve the problem of the dark room. Later in this chapter, I argue that the Bayesian PP is normative not because it enables the formulation of such and such rules of action, but because the predictive mechanisms themselves are normative. They are normative because they condition the choice of such and such actions by agents. In this way, I justify the hypothesis that normative mechanisms make it possible to explain the phenomenon of agent motivation, which is crucial for solving the dark room problem. In the last part of the chapter, I formulate the hypothesis of distributed normativity, which assumes that the normative nature of certain mechanisms, functions or objects is determined by the relations into which these mechanisms, functions or objects enter. This means that what is normative (in the primary sense) is the relational structure that constitutes the normativity of specific items included in it. I suggest that this hypothesis opens up many areas of research and makes it possible to rethink many problems. In the “Conclusion”, I summarize the results of my research and indicate further research perspectives. (shrink)

This paper presents two challenges faced by many initiatives that try to diversify undergraduate philosophy curricula, both intellectually and demographically. Trade-offs involve making difficult decisions to prioritise some values over others (like gender diversity over cultural diversity). Backfires involve unintended consequences contrary to the aims and values of diversity initiatives, including ones that compromise more general philosophical values. I discuss two specific backfire risks, involving the critical and political dimensions of teaching philosophy. Some general practical advice is offered (...) along the way. (shrink)

Teleological accounts of sensory normativity treat normal functioning for a species as a standard: sensory error involves departure from normal functioning for the species, i.e. sensory malfunction. Straightforward reflection on sensory trade-offs reveals that normal functioning for a species can exhibit failures of accuracy. Acknowledging these failures of accuracy is central to understanding the adaptations of a species. To make room for these errors we have to go beyond the teleological framework and invoke the notion of an ideal (...) observer from vision science. The notion of an ideal observer also sheds light on the important distinction between sensory malfunction and sensory limitation. (shrink)

There is growing concern that decision-making informed by machine learning (ML) algorithms may unfairly discriminate based on personal demographic attributes, such as race and gender. Scholars have responded by introducing numerous mathematical definitions of fairness to test the algorithm, many of which are in conflict with one another. However, these reductionist representations of fairness often bear little resemblance to real-life fairness considerations, which in practice are highly contextual. Moreover, fairness metrics tend to be implemented in narrow and targeted toolkits that (...) are difficult to integrate into an algorithm’s broader ethical assessment. In this paper, we derive lessons from ethical philosophy and welfare economics as they relate to the contextual factors relevant for fairness. In particular we highlight the debate around acceptability of particular inequalities and the inextricable links between fairness, welfare and autonomy. We propose Key Ethics Indicators (KEIs) as a way towards providing a more holistic understanding of whether or not an algorithm is aligned to the decision-maker’s ethical values. (shrink)

This research is concerned with understanding the factors behind the trade-off between child labor and child schooling, given the well-documented links between the two. It examines parents' behavior in their decision-making on their children's schooling or practicing child labor. Depending on qualitative research methods including 28 semi-structured interviews and two focus group discussions conducted in the rural areas of Bangladesh in 2020, this study reveals the following: subsistence needs compel households, particularly the ultra-poor and the female-headed, to trade (...) off child labor with schooling; due to higher demand of labor, parents engage their children into work instead of schooling; parents of labor-intensive occupations tend to trade off child labor with schooling; sexual division of labor remains obvious; finally, credit constraints and cultural beliefs have negative impacts on parental decision-making on child schooling. Interventions aiming to reduce child labor and increase schooling in these rural areas must remain mindful of the socio-economic and cultural needs. (shrink)

This article offers a characterization of what I call multiple-models juxtaposition, a strategy for managing trade-offs among modeling desiderata. MMJ displays models of distinct phenomena to...

In this article, I seek to assess the extent to which a ‘trope-theoretic’ version of Theism is a better theory than that of a theory of Atheism, as posited by Graham Oppy. This end will be achieved by utilising the systemisation of the theoretical virtues proposed by Michael Keas (as further modified by an application of the work of Jonathan Schaffer), the notion of a trope, introduced by D.C. Williams, and an aspect, proposed by Donald L.M. Baxter, which will establish (...) the basis of the trope-theoretic account of Theism that will be at the centre of our analysis. This assessment will ultimately show that Theism, rather than that of Atheism (Naturalism), can successfully achieve the trade-off between minimising theoretical commitments and maximising explanatory power. And thus, given this, the best theory of Theism—namely, that of ‘trope-theoretic Theism’—is to be privileged over that of the best theory of Atheism—namely, that of ‘Oppyian Naturalism’—and is able to provide grounds for a decisive reassessment of the cogency of Agnosticism. (shrink)

In the context of a global pandemic, there is good health-based reason for governments to impose various social distancing measures. However, such measures also cause economic and other harms to people at low risk from the virus. In this paper, I examine how to make such trade-offs in a way that is respectfully justifiable to their losers. I argue that existing proposals like using standard QALY (quality-adjusted life-year) valuations or WELLBYs (wellbeing-adjusted life-years) as the currency for trade- (...) class='Hi'>offs do not allow such justification, because they give weight to utilities that are irrelevant in a life-and-death context. Drawing on work on restricted aggregation in ethics, I articulate an alternative framework for balancing claims arising from prospects of different kinds of harm and benefit, and show how it can be applied to reasoning about trade-offs in a pandemic context. (shrink)

In the debate on how we ought to define political freedom, some definitions are criticized for implying that no one can ever be free to perform any action. In this paper, I show how the possibility of freedom depends on a definition that finds an appropriate balance between absence of interference and protection against interference. To assess the possibility of different conceptions of freedom, I consider the trade-offs they make between these two dimensions. I find that pure negative (...) freedom is clearly possible. Republican freedom might also be possible, though its protection requirement is too vague for a definitive verdict. Finally, the recently proposed ‘freedom as independence’ is impossible since it is an attempt to avoid the unavoidable trade-off. (shrink)

Abstract: Recently, the experimental philosopher Joshua Knobe has shown that the folk are more inclined to describe side effects as intentional actions when they bring about bad results. Edouard Machery has offered an intriguing new explanation of Knobe's work—the 'trade-off hypothesis'—which denies that moral considerations explain folk applications of the concept of intentional action. We critique Machery's hypothesis and offer empirical evidence against it. We also evaluate the current state of the debate concerning the concept of intentionality, and argue (...) that, given the number of variables at play, any parsimonious account of the relevant data is implausible. (shrink)

This paper examines two prominent formal trade-offs in artificial intelligence (AI)---between predictive accuracy and fairness, and between predictive accuracy and interpretability. These trade-offs have become a central focus in normative and regulatory discussions as policymakers seek to understand the value tensions that can arise in the social adoption of AI tools. The prevailing interpretation views these formal trade-offs as directly corresponding to tensions between underlying social values, implying unavoidable conflicts between those social objectives. In (...) this paper, I challenge that prevalent interpretation by introducing a sociotechnical approach to examining the value implications of trade-offs. After providing a set-theoretic understanding of these model trade-offs, I argue that their implications cannot be understood in isolation from the technical, psychological, organizational, and social factors that shape the integration of AI models into decision-making pipelines. Specifically, I identify three key considerations---validity and instrumental relevance, compositionality, and dynamics---for contextualizing and characterizing these implications. These considerations reveal that the relationship between model trade-offs and corresponding values depends on critical choices and assumptions. Crucially, judicious sacrifices in one model property for another can, in fact, promote both sets of corresponding values. The proposed sociotechnical perspective thus shows that we can and should aspire to higher epistemic and ethical possibilities than the prevalent interpretation suggests, while offering practical guidance for achieving those outcomes. Finally, I draw out the broader implications of this perspective for AI design and governance, highlighting the need to broaden normative engagement across the AI lifecycle, develop legal and auditing tools sensitive to sociotechnical considerations, and rethink the vital role and appropriate structure of interdisciplinary collaboration in fostering a responsible AI workforce. (shrink)

Lane et al. propose an integrative model for the reconsolidation of traces in their timely and impressive article. This commentary draws attention to tradeoffs between accuracy and self-narrative integrity in the model. The tradeoffs concern the sense of agency in memory and its role in both implicit and explicit memory reconsolidation, rather than balances concerning degrees of emotional arousal.

Tailoring the design of robot bodies for control purposes is implicitly performed by engineers, however, a methodology or set of tools is largely absent and optimization of morphology (shape, material properties of robot bodies, etc.) is lag- ging behind the development of controllers. This has become even more prominent with the advent of compliant, deformable or "soft" bodies. These carry substantial potential regarding their exploitation for control – sometimes referred to as "mor- phological computation" in the sense of offloading computation (...) needed for control to the body. Here, we will argue in favor of a dynamical systems rather than com- putational perspective on the problem. Then, we will look at the pros and cons of simple vs. complex bodies, critically reviewing the attractive notion of “soft” bodies automatically taking over control tasks. We will address another key dimension of the design space – whether model-based control should be used and to what extent it is feasible to develop faithful models for different morphologies. ----- This paper was also published in the 2014 AISB proceedings http://aisb50.org/representation-of-reality-humans-animals-and-machines/ - http://doc.gold.ac.uk/aisb50/. (shrink)

This chapter argues that rather than viewing transparency as a right, we should regard it as a finite resource whose allocation involves tradeoffs. It then argues that those tradeoffs should be resolved by using a multi-principle approach to distributive justice. The relevant principles include maximizing welfare, maximizing autonomy, and giving priority to the worst off. Finally, it examines some of the implications for law of recognizing the tradeoffs presented by transparency proposals.

Autonomy is often considered a core value of Western society that is deeply entrenched in moral, legal, and political practices. The development and deployment of artificial intelligence (AI) systems to perform a wide variety of tasks has raised new questions about how AI may affect human autonomy. Numerous guidelines on the responsible development of AI now emphasise the need for human autonomy to be protected. In some cases, this need is linked to the emergence of increasingly ‘autonomous’ AI systems that (...) can perform tasks without human control or supervision. Do such ‘autonomous’ systems pose a risk to our own human autonomy? In this article, I address the question of a trade-off between human autonomy and system ‘autonomy’. (shrink)

This paper examines two prominent formal trade-offs in artificial intelligence (AI)---between predictive accuracy and fairness, and between predictive accuracy and interpretability. These trade-offs have become a central focus in normative and regulatory discussions as policymakers seek to understand the value tensions that can arise in the social adoption of AI tools. The prevailing interpretation views these formal trade-offs as directly corresponding to tensions between underlying social values, implying unavoidable conflicts between those social objectives. In (...) this paper, I challenge that prevalent interpretation by introducing a sociotechnical approach to examining the value implications of trade-offs. After providing a set-theoretic understanding of these model trade-offs, I argue that their implications cannot be understood in isolation from the technical, psychological, organizational, and social factors that shape the integration of AI models into decision-making pipelines. Specifically, I identify three key considerations---validity and instrumental relevance, compositionality, and dynamics---for contextualizing and characterizing these implications. These considerations reveal that the relationship between model trade-offs and corresponding values depends on critical choices and assumptions. Crucially, judicious sacrifices in one model property for another can, in fact, promote both sets of corresponding values. The proposed sociotechnical perspective thus shows that we can and should aspire to higher epistemic and ethical possibilities than the prevalent interpretation suggests, while offering practical guidance for achieving those outcomes. Finally, I draw out the broader implications of this perspective for AI design and governance, highlighting the need to broaden normative engagement across the AI lifecycle, develop legal and auditing tools sensitive to sociotechnical considerations, and rethink the vital role and appropriate structure of interdisciplinary collaboration in fostering a responsible AI workforce. (shrink)

To address the rising concern that algorithmic decision-making may reinforce discriminatory biases, researchers have proposed many notions of fairness and corresponding mathematical formalizations. Each of these notions is often presented as a one-size-fits-all, absolute condition; however, in reality, the practical and ethical trade-offs are unavoidable and more complex. We introduce a new approach that considers fairness—not as a binary, absolute mathematical condition—but rather, as a relational notion in comparison to alternative decisionmaking processes. Using US mortgage lending as an (...) example use case, we discuss the ethical foundations of each definition of fairness and demonstrate that our proposed methodology more closely captures the ethical trade-offs of the decision-maker, as well as forcing a more explicit representation of which values and objectives are prioritised. (shrink)

Engineers fine-tune the design of robot bodies for control purposes; however, a methodology or set of tools is largely absent, and optimization of morphology is lagging behind the development of controllers. This has become even more prominent with the advent of compliant, deformable or ‘soft’ bodies. These carry substantial potential regarding their exploitation for control—sometimes referred to as ‘morphological computation’. In this article, we briefly review different notions of computation by physical systems and propose the dynamical systems framework as the (...) most useful in the context of describing and eventually designing the interactions of controllers and bodies. Then, we look at the pros and cons of simple versus complex bodies, critically reviewing the attractive notion of ‘soft’ bodies automatically taking over control tasks. We address another key dimension of the design space—whether model-based control should be used and to what extent it is feasible to develop faithful models for different morphologies. (shrink)

Engineers fine-tune the design of robot bodies for control purposes, however, a methodology or set of tools is largely absent, and optimization of morphology (shape, material properties of robot bodies, etc.) is lagging behind the development of controllers. This has become even more prominent with the advent of compliant, deformable or ”soft” bodies. These carry substantial potential regarding their exploitation for control—sometimes referred to as ”morphological computation”. In this article, we briefly review different notions of computation by physical systems and (...) propose the dynamical systems framework as the most useful in the context of describing and eventually designing the interactions of controllers and bodies. Then, we look at the pros and cons of simple vs. complex bodies, critically reviewing the attractive notion of ”soft” bodies automatically taking over control tasks. We address another key dimension of the design space—whether model-based control should be used and to what extent it is feasible to develop faithful models for different morphologies. (shrink)

Non-technical summary: -/- When developing and deploying negative emissions technologies (NETs), little attention has been paid to where. On the one hand, one might develop NETs where they are likely to contribute most to global mitigation targets, contributing to a global climate solution. On the other hand, one might develop NETs where they can help support development on a regional basis, justified by regional demands. I defend these arguments and suggest that they reflect the values of efficiency and responding to (...) need, respectively. To the extent that these values conflict, they introduce what I call the Need-Efficiency Trade-off Effect (‘NET Effect’). -/- Technical summary: -/- Unlike other geoengineering methods, the effectiveness of negative emissions technologies (NETs) tends to be sensitive to regional siting. This paper argues that this point raises morally and legally important implications by identifying a trade-off between ‘efficiency’ and ‘need’. First, it introduces two arguments justifying NETs: one focused on contributions to global mitigation and the other focused on contributions to regional development. Second, reflecting the two arguments, the paper discusses the moral values of efficiency and need, respectively. For instance, if the strategy is to try to use NETs to maximize expected mitigation contributions to reflect efficiency, then deployment should occur in regions with the best prospects for success (e.g. Western countries). However, if the strategy is to try to use NETs to improve the chances of simultaneous development and mitigation to respond to need, then deployment should occur in regions with limited development and expected growth of demand for NETs (e.g. Asian countries). When these values conflict, I call that a Need-Efficiency Trade-off Effect (‘NET Effect’). The paper concludes by considering the NET Effect in the context of bioenergy with carbon capture and storage as well as direct air carbon capture and storage. [Open access]. (shrink)

The Market Failures Approach (MFA) is one of the leading theories in contemporary business ethics. It generates a list of ethical obligations for the managers of private firms that states that they should not create or exploit market failures because doing so reduces the efficiency of the economy. Recently the MFA has been criticised by Abraham Singer on the basis that it unjustifiably does not assign private managers obligations based on egalitarian values. Singer proposes an extension to the MFA, the (...) Justice Failures Approach (JFA), in which managers have duties to alleviate political, social, and distributive inequalities in addition to having obligations to not exploit market failures. In this paper I describe the MFA and JFA and situate them relative to each other. I then highlight a threefold distinction between different types of obligations that can be given to private managers in order to argue that a hybrid theory of business ethics, which I call the MFA +, can be generated by arguing that managers have obligations based on efficiency and duties based on equality to the extent that these latter obligations do not lead to efficiency losses. This argument suggests a novel theoretical option in business ethics, elucidates the issues that are at stake between the MFA and the JFA, and clarifies the costs and benefits of each theory. (shrink)

Minimum wages are usually assumed to be inefficient as they prevent the full exploitation of mutual gains from trade. Yet advocates of wage regulation policies have repeatedly claimed that this loss in market efficiency can be justified by the pursuit of ethical goals. Policy makers, it is argued, should not focus on efficiency alone. Rather, they should try to find an adequate balance between efficiency and equity targets. This idea is based on a two-worlds-paradigm that sees ethics and economics (...) as two inherently conflicting ways of thinking. We, however, believe that this view of the relationship between ethics and economics is fundamentally flawed and blurs our understanding of how an ethically responsible regulation of the labour market should be conducted. In drawing on an economic-ethical approach that resolves the antinomy between efficiency and equity, we show that ethics and economics are, in fact, two sides of the same coin and that minimum wage legislation can only be ethically responsible, if it is at the same time economically efficient. In other words, we can have our cake and eat it too. On the basis of our approach, we develop two simple game theoretical models for different types of labour markets and derive policy implications from an economic-ethical viewpoint. We suggest that under the assumption of perfectly competitive labour markets a tax-funded wage subsidy is preferable over minimum wages, as it makes everyone better off. If, however, employers have monopsony power in the wage setting process, the minimum wage is justifiable under certain conditions. (shrink)

This paper disputes the notion, endorsed by much of narrative theory, that the reading of literary narrative is functionally analogous to an act of communication, where communication stands for the transfer of thought and conceptual information. The paper offers a basic typology of the sensorimotor effects of reading, which fall outside such a narrowly communication-based model of literary narrative. A main typological distinction is drawn between those sensorimotor effects pertaining to the narrative qua verbal utterance (verbal presence) and those sensorimotor (...) effects pertaining to the imaginary physical world(s) of the story (direct presence). While verbal presence refers to the reader's vicarious perception of the voices of narrators and characters, direct presence refers to the emulated sensorimotor experience of the imaginary worlds that the narrators' and characters' utterances refer to. The paper further elaborates on how, by which kinds of narrative content and structure, direct presence may be prompted. The final section addresses some of the observational and historical caveats that must be attached to any theoretical inquiry made into the sensorimotor effects of reading. As a preliminary for further research, a few ideas about the model's potential for empirical validation are put forward. A brief, tentative history of the sensorimotor benefits of literary narrative reading is then outlined. (shrink)

We argued that explanatoriness is evidentially irrelevant in the following sense: Let H be a hypothesis, O an observation, and E the proposition that H would explain O if H and O were true. Then our claim is that Pr = Pr. We defended this screening-off thesis by discussing an example concerning smoking and cancer. Climenhaga argues that SOT is mistaken because it delivers the wrong verdict about a slightly different smoking-and-cancer case. He also considers a variant of SOT, called (...) “SOT*”, and contends that it too gives the wrong result. We here reply to Climenhaga’s arguments and suggest that SOT provides a criticism of the widely held theory of inference called “inference to the best explanation”. (shrink)

We demonstrate how real progress can be made in the debate surrounding the enhanced indispensability argument. Drawing on a counterfactual theory of explanation, well-motivated independently of the debate, we provide a novel analysis of ‘explanatory generality’ and how mathematics is involved in its procurement. On our analysis, mathematics’ sole explanatory contribution to the procurement of explanatory generality is to make counterfactual information about physical dependencies easier to grasp and reason with for creatures like us. This gives precise (...) content to key intuitions traded in the debate, regarding mathematics’ procurement of explanatory generality, and adjudicates unambiguously in favour of the nominalist, at least as far as explanatory generality is concerned. (shrink)

We argue elsewhere that explanatoriness is evidentially irrelevant . Let H be some hypothesis, O some observation, and E the proposition that H would explain O if H and O were true. Then O screens-off E from H: Pr = Pr. This thesis, hereafter “SOT” , is defended by appeal to a representative case. The case concerns smoking and lung cancer. McCain and Poston grant that SOT holds in cases, like our case concerning smoking and lung cancer, that involve frequency (...) data. However, McCain and Poston contend that there is a wider sense of evidential relevance—wider than the sense at play in SOT—on which explanatoriness is evidentially relevant even in cases involving frequency data. This is their main point, but they also contend that SOT does not hold in certain cases not involving frequency data. We reply to each of these points and conclude with some general remarks on screening-off as a test of evidential relevance. (shrink)

In ‘Towards a Solution to the Sorites Paradox’, Graham Priest gives us a new account of the sorites based on fuzzy logic. The novelty lies in the suggestion that truth-value assignments should themselves be treated as fuzzy objects, i.e., objects about which we can make fuzzy identity statements. I argue that Priest’s solution does not have the explanatory force that Priest advocates. That is, it does not explain why we find the existence of a cut-off point counter-intuitive. I also (...) argue that this sort of explanation calls for a general theory that goes beyond the special case of linguistic vagueness, for the phenomenon is at bottom not linguistic. (shrink)

We argue in Roche and Sober (2013) that explanatoriness is evidentially irrelevant in that Pr(H | O&EXPL) = Pr(H | O), where H is a hypothesis, O is an observation, and EXPL is the proposition that if H and O were true, then H would explain O. This is a “screening-off” thesis. Here we clarify that thesis, reply to criticisms advanced by Lange (2017), consider alternative formulations of Inference to the Best Explanation, discuss a strengthened screening-off thesis, and consider how (...) it bears on the claim that unification is evidentially relevant. (shrink)

Critics have recently argued that reliabilists face trade-off problems, forcing them to condone intuitively unjustified beliefs when they generate lots of true belief further downstream. What these critics overlook is that reliabilism entails that there areside-constraintson belief-formation, on account of which there are some things you should not believe, even if doing so would have very good epistemic consequences. However, we argue that by embracing side-constraints the reliabilist faces a dilemma: she can either hold on to reliabilism, and with (...) it aforementioned side-constraints, but then needs to explain why we should allow the pursuit of justification to get in the way of the acquisition of true belief; or she can deny that there are side-constraints – and in effect give up on reliabilism. We'll suggest that anyone moved by the considerations that likely attract people to reliabilism in the first place – the idea the true belief is good, and as such should be promoted – should go for the second horn, and instead pursue a form ofepistemic utilitarianism. (shrink)

There is a general presumption against arming outlaw states. But can that presumption sometimes be overturned? The argument considered here maintains that outlaw states can have legitimate security interests and that transferring weapons to these states can be an appropriate way of promoting those interests. Weapons enable governments to engage in wrongful oppression and aggression, but they also enable them to fend off predators in a manner that can be beneficial to their citizens. It clearly does not follow from the (...) fact that a state is oppressive or aggressive that it will never be a victim of wrongful aggression itself, and while an outlaw state’s primary aim in repelling such aggression will often be the preservation of its own power, its defensive manoeuvres will sometimes also serve its citizens’ interests. In short, supplying weapons to outlaw states may sometimes contribute to the protection of innocents. (shrink)

Internet trolling involves making assertions with the aim of provoking emotionally heated responses, all while pretending to be a sincere interlocutor. In this paper, I give an account of some of the epistemic and psychological dimensions of trolling, with the goal of better understanding why certain kinds of trolling can be dangerous. I first analyze how trolls eschew the epistemic norms of assertion, thus covertly violating their conversation partners’ normative expectations. Then, drawing on literature on the “explore/exploit trade-off,” I (...) argue that trolling is a kind of exploratory behaviour. Specifically, it involves exploring the consequences of violating an interlocutor’s expectation that one will follow the epistemic norms of assertion. To defend this account, I argue that it explains various facts about trolls and the kind of pleasure they get from their activities. My account provides a deeper understanding of why trolling can be dangerous: namely, it explains why certain trolling behaviours contribute to radicalizing people into extremist or hateful ideologies, as well as why online platforms where trolls run amok gradually become polluted with extreme, hateful speech. (shrink)

The overall aim of this article is to reorient the contemporary debate about epistemic consequentialism. Thus far the debate has to a large extent focused on whether standard theories of epistemic justification are consequentialist in nature and therefore vulnerable to certain trade-off cases where accepting a false or unjustified belief leads to good epistemic outcomes. We claim that these trade-offs raise an important—yet somewhat neglected—issue about the epistemic demands on inquiry. We first distinguish between two different kinds (...) of epistemic evaluation, viz., backing evaluation and outcome evaluation, and then go on to outline and discuss a consequentialist metatheory about the right combinations of decision procedures to adopt in inquiry. Note that the piece is exploratory in the following sense: we try to explore epistemic evaluation in consequentialist terms, which involves stating a form of epistemic consequentialism, but also pointing to what non-consequentialist alternatives might be. Rather than trying to argue decisively for a particular conclusion, we aim to outline various intricate issues in an underexplored area of theorizing. In the course of doing this, we’ll transpose some well-known themes from discussions of consequentialism in ethics to the current debate about consequentialism in epistemology, e.g., agent-neutrality, options, and side-constraints. (shrink)

There is a trade-off between specificity and accuracy in existing models of belief. Descriptions of agents in the tripartite model, which recognizes only three doxastic attitudes—belief, disbelief, and suspension of judgment—are typically accurate, but not sufficiently specific. The orthodox Bayesian model, which requires real-valued credences, is perfectly specific, but often inaccurate: we often lack precise credences. I argue, first, that a popular attempt to fix the Bayesian model by using sets of functions is also inaccurate, since it requires us (...) to have interval-valued credences with perfectly precise endpoints. We can see this problem as analogous to the problem of higher order vagueness. Ultimately, I argue, the only way to avoid these problems is to endorse Insurmountable Unclassifiability. This principle has some surprising and radical consequences. For example, it entails that the trade-off between accuracy and specificity is in-principle unavoidable: sometimes it is simply impossible to characterize an agent’s doxastic state in a way that is both fully accurate and maximally specific. What we can do, however, is improve on both the tripartite and existing Bayesian models. I construct a new model of belief—the minimal model—that allows us to characterize agents with much greater specificity than the tripartite model, and yet which remains, unlike existing Bayesian models, perfectly accurate. (shrink)

This article introduces and defends the “pathological complexity thesis” as a hypothesis about the evolutionary origins of minimal consciousness, or sentience, that connects the study of animal consciousness closely with work in behavioral ecology and evolutionary biology. I argue that consciousness is an adaptive solution to a design problem that led to the extinction of complex multicellular animal life following the Avalon explosion and that was subsequently solved during the Cambrian explosion. This is the economic trade-off problem of having (...) to deal with a complex body with high degrees of freedom, what I call “pathological complexity.” By modeling the explosion of this computational complexity using the resources of state-based behavioral and life history theory we will be able to provide an evolutionary bottom-up framework to make sense of subjective experience and its function in nature by paying close attention to the ecological lifestyles of different animals. (shrink)

Which is better: a guarantee of a modest amount of moral value, or a tiny probability of arbitrarily large value? To prefer the latter seems fanatical. But, as I argue, avoiding such fanaticism brings severe problems. To do so, we must decline intuitively attractive trade-offs; rank structurally identical pairs of lotteries inconsistently, or else admit absurd sensitivity to tiny probability differences; have rankings depend on remote, unaffected events ; and often neglect to rank lotteries as we already know (...) we would if we learned more. Compared to these implications, fanaticism is highly plausible. (shrink)

In this paper I explore the relationship between skill and sensitivity to reasons for action. I want to know to what degree we can explain the fact that the skilled agent is very good at performing a cluster of actions within some domain in terms of the fact that the skilled agent has a refined sensitivity to the reasons for action common to the cluster. The picture is a little bit complex. While skill can be partially explained by sensitivity to (...) reasons – a sensitivity often produced by rational practice – the skilled human agent, because imperfect, must navigate a trade-off between full sensitivity and a capacity to succeed. (shrink)

This paper outlines an account of political realism as a form of ideology critique. Our focus is a defence of the normative edge of this critical-theoretic project against the common charge that there is a problematic trade-off between a theory’s groundedness in facts about the political status quo and its ability to consistently envisage radical departures from the status quo. To overcome that problem we combine insights from three distant corners of the philosophical landscape: theories of legitimacy by Bernard (...) Williams and other realists, Frankfurt School-inspired Critical Theory, and recent analytic epistemological and metaphysical theories of cognitive bias, ideology, and social construction. The upshot is a novel account of realism as empirically-informed diagnosis- critique of social and political phenomena. This view rejects a sharp divide between descriptive and normative theory, and so is an alternative to the anti- empiricism of some approaches to Critical Theory as well as to the complacency towards existing power structures found within liberal realism, let alone mainstream normative political philosophy, liberal or otherwise. (shrink)

Recent work on interpretability in machine learning and AI has focused on the building of simplified models that approximate the true criteria used to make decisions. These models are a useful pedagogical device for teaching trained professionals how to predict what decisions will be made by the complex system, and most importantly how the system might break. However, when considering any such model it’s important to remember Box’s maxim that "All models are wrong but some are useful." We focus on (...) the distinction between these models and explanations in philosophy and sociology. These models can be understood as a "do it yourself kit" for explanations, allowing a practitioner to directly answer "what if questions" or generate contrastive explanations without external assistance. Although a valuable ability, giving these models as explanations appears more difficult than necessary, and other forms of explanation may not have the same trade-offs. We contrast the different schools of thought on what makes an explanation, and suggest that machine learning might benefit from viewing the problem more broadly. (shrink)

Where does the impetus towards ethical theory come from? What drives humans to make values explicit, consistent, and discursively justifiable? This paper situates the demand for ethical theory in human life by identifying the practical needs that give rise to it. Such a practical derivation puts the demand in its place: while finding a home for it in the public decision-making of modern societies, it also imposes limitations on the demand by presenting it as scalable and context-sensitive. This differentiates strong (...) forms of the demand calling for theory from weaker forms calling for less, and contexts where it has a place from contexts where it is out of place. In light of this, subjecting personal deliberation to the demand turns out to involve a trade-off. (shrink)

Why would philosophers interested in the points or functions of our conceptual practices bother with genealogical explanations if they can focus directly on paradigmatic examples of the practices we now have?? To answer this question, I compare the method of pragmatic genealogy advocated by Edward Craig, Bernard Williams, and Miranda Fricker—a method whose singular combination of fictionalising and historicising has met with suspicion—with the simpler method of paradigm-based explanation. Fricker herself has recently moved towards paradigm-based explanation, arguing that it is (...) a more perspicuous way of reaping the same explanatory pay-off as pragmatic genealogy while dispensing with its fictionalising and historicising. My aim is to determine when and why the reverse movement from paradigm-based explanation to pragmatic genealogy remains warranted. I argue that the fictionalising and historicising of pragmatic genealogy is well motivated, and I outline three ways in which the method earns its keep: by successfully handling historically inflected practices which paradigm-based explanation cannot handle; by revealing and arguing for connections to generic needs we might otherwise miss; and by providing comprehensive views of practices that place and relate the respects in which they serve both generic and local needs. (shrink)