The logical omniscience problem, whereby standard models of epistemic logic treat an agent as believing all consequences of its beliefs and knowing whatever follows from what else it knows, has received plenty of attention in the literature. But many attempted solutions focus on a fairly narrow specification of the problem: avoiding the closure of belief or knowledge, rather than showing how the proposed logic is of philosophical interest or of use in computer science or artificial intelligence. Sentential epistemic logics, as (...) opposed to traditional possible worlds approaches, do not suffer from the problems of logical omniscience but are often thought to lack interesting epistemic properties. In this paper, I focus on the case of rule-based agents, which play a key role in contemporary AI research but have been neglected in the logical literature. I develop a framework for modelling monotonic, nonmonotonic and introspective rule-based reasoners which have limited cognitive resources and prove that the resulting models have a number of interesting properties. An axiomatization of the resulting logic is given, together with completeness, decidability and complexity results. (shrink)

This paper proposes an approach to investigate norm-governed learning agents which combines a logic-based formalism with an equation-based counterpart. This dual formalism enables us to describe the reasoning of such agents and their interactions using argumentation, and, at the same time, to capture systemic features using equations. The approach is applied to norm emergence and internalisation in systems of learning agents. The logical formalism is rooted into a probabilistic defeasible logic instantiating Dung’s argumentation framework. Rules (...) of this logic are attached with probabilities to describe the agents’ minds and behaviours as well as uncertain environments. Then, the equation-based model for reinforcement learning, defined over this probability distribution, allows agents to adapt to their environment and self-organise. (shrink)

Agents which perform inferences on the basis of unreliable information need an ability to revise their beliefs if they discover an inconsistency. Such a belief revision algorithm ideally should be rational, should respect any preference ordering over the agent’s beliefs (removing less preferred beliefs where possible) and should be fast. However, while standard approaches to rational belief revision for classical reasoners allow preferences to be taken into account, they typically have quite high complexity. In this paper, we consider belief (...) revision for agents which reason in a simpler logic than full first-order logic, namely rule-based reasoners. We show that it is possible to define a contraction operation for rule-based reasoners, which we call McAllester contraction, which satisfies all the basic Alchourrón, Gärdenfors and Makinson (AGM) postulates for contraction (apart from the recovery postulate) and at the same time can be computed in polynomial time. We prove a representation theorem for McAllester contraction with respect to the basic AGM postulates (minus recovery), and two additional postulates. We then show that our contraction operation removes a set of beliefs which is least preferred, with respect to a natural interpretation of preference. Finally, we show how McAllester contraction can be used to define a revision operation which is also polynomial time, and prove a representation theorem for the revision operation. (shrink)

Syntactic logics do not suffer from the problems of logical omniscience but are often thought to lack interesting properties relating to epistemic notions. By focusing on the case of rule-based agents, I develop a framework for modelling resource-bounded agents and show that the resulting models have a number of interesting properties.

The developing academic field of machine ethics seeks to make artificial agents safer as they become more pervasive throughout society. In contrast to computer ethics, machine ethics is concerned with the behavior of machines toward human users and other machines. This study aims to use an action-based ethical theory founded on the combinational aspects of deontological and teleological theories of ethics in the construction of an artificial moral agent (AMA).,The decision results derived by the AMA are acquired via (...) fuzzy logic interpretation of the relative values of the steady-state simulations of the corresponding rule-based fuzzy cognitive map (RBFCM).,Through the use of RBFCMs, the following paper illustrates the possibility of incorporating ethical components into machines, where latent semantic analysis (LSA) and RBFCMs can be used to model dynamic and complex situations, and to provide abilities in acquiring causal knowledge.,This approach is especially appropriate for data-poor and uncertain situations common in ethics. Nonetheless, to ensure that a machine with an ethical component can function autonomously in the world, research in artificial intelligence will need to further investigate the representation and determination of ethical principles, the incorporation of these ethical principles into a system’s decision procedure, ethical decision-making with incomplete and uncertain knowledge, the explanation for decisions made using ethical principles and the evaluation of systems that act based upon ethical principles.,To date, the conducted research has contributed to a theoretical foundation for machine ethics through exploration of the rationale and the feasibility of adding an ethical dimension to machines. Further, the constructed AMA illustrates the possibility of utilizing an action-based ethical theory that provides guidance in ethical decision-making according to the precepts of its respective duties. The use of LSA illustrates their powerful capabilities in understanding text and their potential application as information retrieval systems in AMAs. The use of cognitive maps provides an approach and a decision procedure for resolving conflicts between different duties.,This paper suggests that cognitive maps could be used in AMAs as tools for meta-analysis, where comparisons regarding multiple ethical principles and duties can be examined and considered. With cognitive mapping, complex and abstract variables that cannot easily be measured but are important to decision-making can be modeled. This approach is especially appropriate for data-poor and uncertain situations common in ethics. (shrink)

A well-trod debate at the nexus of cognitive science and linguistics, the so-called past tense debate, has examined how rules and exceptions are individually acquired. However, this debate focuses primarily on individual mechanisms in learning, saying little about how rules and exceptions function from a sociolinguistic perspective. To remedy this, we use agent-based models to examine how rules and exceptions function across populations. We expand on earlier work by considering how repeated interaction and cultural transmission across speakers affects the (...) dynamics of rules and exceptions in language, measuring linguistic outcomes within a social system rather than focusing individual learning outcomes. We consider how population turnover and growth effect linguistic rule dynamics in large and small populations, showing that this method has considerable potential particularly in probing the mechanisms underlying the linguistic niche hypothesis. (shrink)

- In decision theory, an agent is deciding how to value a gamble that results in different outcomes in different states. Each outcome gets a utility value for the agent.

Zusammenfassung Der Geschäftsbereich Leistungssport des Deutschen Olympischen Sportbundes agiert als Common Agent, nämlich der Sportverbände und des Bundes­ministern des Innern bei der Verteilung der Sportfördermittel des BMI an die olympischen Spitzenverbände im DOSB. In diesem Beitrag wird theoriegeleitet ein Hypothesensystem entwickelt, welches beschreibt, nach welchen Kriterien der GBL bei der Verteilung der Mittel vorgeht. Es zeigt sich, dass neben den Bewertungskriterien, die im so genannten Förderkonzept 2012 formalisiert werden, diskretionäre Handlungsspielräume bestehen. Diese werden quantifiziert und mittels des Hypothesensystems interpretiert. Es (...) wird diskutiert, wie die Problematik einer Common Agency entschärft werden kann. (shrink)

Capacities for abstract thinking and problem‐solving are central to human cognition. Processes of abstraction allow the transfer of experiences and knowledge between contexts helping us make informed decisions in new or changing contexts. While we are often inclined to relate such reasoning capacities to individual minds and brains, they may in fact be contingent on human‐specific modes of collaboration, dialogue, and shared attention. In an experimental study, we test the hypothesis that social interaction enhances cognitive processes of rule‐induction, which (...) in turn improves problem‐solving performance. Through three sessions of increasing complexity, individuals and groups were presented with a problem‐solving task requiring them to categorize a set of visual stimuli. To assess the character of participants’ problem representations, after each training session, they were presented with a transfer task involving stimuli that differed in appearance, but shared relations among features with the training set. Besides, we compared participants’ categorization behaviors to simulated agents relying on exemplar learning. We found that groups performed superior to individuals and agents in the training sessions and were more likely to correctly generalize their observations in the transfer phase, especially in the high complexity session, suggesting that groups more effectively induced underlying categorization rules from the stimuli than individuals and agents. Crucially, variation in performance among groups was predicted by semantic diversity in members’ dialogical contributions, suggesting a link between social interaction, cognitive diversity, and abstraction. (shrink)

Agent-Based Models are useful to describe and understand social, economic and spatial systems' dynamics. But, beside the facilities which this methodology offers, evaluation and comparison of simulation models are sometimes problematic. A rigorous conceptual frame needs to be developed. This is in order to ensure the coherence in the chain linking at the one extreme the scientist's hypotheses about the modeled phenomenon and at the other the structure of rules in the computer program. This also systematizes the model design (...) from the thematician conceptual framework as well. The aim is to reflect upon the role that a well defined ontology, based on the crossing of the philosophical and the computer science insights, can play to solve such questions and help the model building. We analyze different conceptions of ontology, introduce the 'ontological test' and show its usefulness to compare models. Then we focus on the model building and show the place of a systematic ABM ontology. The latter process is situated within a larger framework called the 'knowledge framework' in which not only the ontologies but also the notions of theory, model and empirical data take place. At last the relation between emergence and ontology is discussed. (shrink)

According to an influential characterisation of reflective equilibrium, it is a kind of algorithm for licensing explicitly normative claims in philosophical inquiries. Call this the machine-view of reflective equilibrium. The machine-view implies a causal relation between input and output data that is devoid of human agency in any significant sense. In this paper, I argue for a neo-Aristotelian alternative view. According to this view, the judgement that is called forth in the decision procedure of reflective equilibrium is a rational response (...) of an epistemic agent, which integrates the whole of her character as a knower, both cognitive and affective, and which cannot be reduced to the explicit following of rules or an algorithm. Unlike the machine-view, this alternative account is able to support the dynamic structure and the normative status of the decision procedure. (shrink)

Since e-Commerce has become a discipline, e-Contracts are acknowledged as the tools that will assure the safety and robustness of the transactions. A typical e-Contract is a binding agreement between parties that creates relations and obligations. It consists of clauses that address specific tasks of the overall procedure which can be represented as workflows. Similarly to e-Contracts, Intelligent Agents manage a private policy, a set of rules representing requirements, obligations and restrictions, additionally to personal data that meet their user’s (...) interests. In this context, this study aims at proposing a policy-based e-Contract workflow management methodology that can be used by semantic web agents, since agents benefit from Semantic Web technologies for data and policy exchanges, such as RDF and RuleML that maximize interoperability among parties. Furthermore, this study presents the integration of the above methodology into a multi-agent knowledge-based framework in order to deal with issues related to rules exchange where no common syntax is used, since this framework provides reasoning services that assist agents in interpreting the exchanged policies. Finally, a B2C e-Commerce scenario is presented that demonstrates the added value of the approach. (shrink)

Practical reasoners are resource-bounded—in particular they require time to derive consequences of their knowledge. Building on the Timed Reasoning Logics (TRL) framework introduced in [1], we show how to represent the time required by an agent to reach a given conclusion. TRL allows us to model the kinds of rule application and conflict resolution strategies commonly found in rule-based agents, and we show how the choice of strategy can influence the information an agent can take into (...) account when making decisions at a particular point in time. We prove general completeness and decidability results for TRL, and analyse the impact of communication in an example system consisting of two agents which use different conflict resolution strategies. (shrink)

In the literature of belief revision, it is widely accepted that: there is only one revision phase in belief revision which is well characterized by the Bayes’ Rule, Jeffrey’s Rule, etc.. However, as I argue in this article, there are at least four successive phases in belief revision, namely first/second order evaluation and first/second order revision. To characterize these phases, I propose mainly four rules of belief revision based on agent’s criteria, and make one composition rule (...) to characterize belief revision as a whole. (shrink)

Computer science advocates institutional frameworks as an effective tool for modelling policies and reasoning about their interplay. In practice, the rules or policies, of which the institutional framework consists, are often specified using a formal language, which allows for the full verification and validation of the framework (e.g. the consistency of policies) and the interplay between the policies and actors (e.g. violations). However, when modelling large-scale realistic systems, with numerous decision-making entities, scalability and complexity issues arise making it possible only (...) to verify certain portions of the problem without reducing the scale. In the social sciences, agent-based modelling is a popular tool for analysing how entities interact within a system and react to the system properties. Agent-based modelling allows the specification of complex decision-making entities and experimentation with large numbers of different parameter sets for these entities in order to explore their effects on overall system performance. In this paper we describe how to achieve the best of both worlds, namely verification of a formal specification combined with the testing of large-scale systems with numerous different actor configurations. Hence, we offer an approach that allows for reasoning about policies, policy making and their consequences on a more comprehensive level than has been possible to date. We present the institutional agent-based model methodology to combine institutional frameworks with agent-based simulations). We furthermore present J-InstAL, a prototypical implementation of this methodology using the InstAL institutional framework whose specifications can be translated into a computational model under the answer set semantics, and an agent-based simulation based on the jason tool. Using a simplified contract enforcement example, we demonstrate the functionalities of this prototype and show how it can help to assess an appropriate fine level in case of contract violations. (shrink)

Rule violations have usually been studied from a third-person perspective, identifying situational factors that render violations more or less likely. A first-person perspective of the agent that actively violates the rules, on the other hand, is only just beginning to emerge. Here we show that committing a rule violation sensitises towards subsequent negative stimuli as well as subsequent authority-related stimuli. In a Prime-Probe design, we used an instructed rule-violation task as the Prime and a word categorisation task (...) as the Probe. Also, we employed a control condition that used a rule inversion task as the Prime. Probe targets were categorised faster after a violation relative to after a rule-based response if they related to either, negative valence or authority. Inversions, however, primed only negative stimuli and did not accelerate the categorisation of authority-related stimuli. A heightened sensitivity towards authority-related targets thus seems to be specific to rule violations. A control experiment showed that these effects cannot be explained in terms of semantic priming. Therefore, we propose that rule violations necessarily activate authority-related representations that make rule violations qualitatively different from simple rule inversions. (shrink)

Artificial Social Agents (ASAs), which are AI software driven entities programmed with rules and preferences to act autonomously and socially with humans, are increasingly playing roles in society. As their sophistication grows, humans will share greater amounts of personal information, thoughts, and feelings with ASAs, which has significant ethical implications. We conducted a study to investigate what ethical principles are of relative importance when people engage with ASAs and whether there is a relationship between people’s values and the ethical (...) principles they prioritise. The study uses scenarios, embedded with the five AI4People ethical principles (Beneficence, Non-maleficence, Autonomy, Justice, and Explicability), involving ASAs taking on roles traditionally played by humans to understand whether these roles and behaviours (including dialogues) are seen as acceptable or unacceptable. Results from 268 participants reveal the greatest sensitivity to ASA behaviours that relate to Autonomy, Justice, Explicability, and the privacy of their personal data. Models were created using Schwartz’s Refined Values as a possible indicator of how stakeholders discern and prioritise the different AI4People ethical principles when interacting with ASAs. Our findings raise issues around the ethical acceptability of ASAs for nudging and changing behaviour due to participants’ desire for autonomy and their concerns over deceptive ASA behaviours such as pretending to have memories and emotions. (shrink)

Weisberg and Muldoon’s epistemic landscape model (ELM) has been one of the most significant contributions to the use of agent-based models in philosophy. The model provides an innovative approach to establishing the optimal distribution of cognitive labor in scientific communities, using an epistemic landscape. In the paper, we provide a critical examination of ELM. First, we show that the computing mechanism for ELM is correct insofar as we are able to replicate the results using another programming language. Second, we (...) show that small changes in the rules that determine the behavior of individual agents can lead to important changes in simulation results. Accordingly, we claim that ELM results are robust with respect to the computing mechanism, but not necessarily across parameter space. We conclude by reflecting on the possible lessons to be gained from ELM as a class of simulations or cluster of models. (shrink)

Multi-level institutional designs with distributed power and authority arrangements among federal, state, regional, and local government agencies could lead to the emergence of differential patterns of socioeconomic and infrastructure development pathways in complex social–ecological systems. Both exogenous drivers and endogenous processes in social–ecological systems can lead to changes in the number of “basins of attraction,” changes in the positions of the basins within the state space, and changes in the positions of the thresholds between basins. In an effort to advance (...) the theory and practice of the governance of policy systems, this study addresses a narrower empirical question: how do intergovernmental institutional rules set by federal, state, and regional government agencies generate and sustain basins of attraction in funding infrastructure projects? A pattern-oriented, agent-based model (ABM) of an intergovernmental network has been developed to simulate real-world transportation policy implementation processes across the federal, the state of Vermont, regional, and local governments for prioritizing transportation projects. The ABM simulates baseline and alternative intergovernmental institutional rule structures and assesses their impacts on financial investment flows. The ABM was calibrated with data from multiple focus groups, individual interviews, and analysis of federal, state, and regional scale transportation projects and programs. The results from experimental simulations are presented to test system-wide effects of alternative multi-level institutional designs, in particular different power and authority arrangements between state and regional governments, on the emergence of roadway project prioritization patterns and funding allocations across regions and towns. (shrink)

This paper is concerned with adaptive learning and coordination processes. Implementing agent-based modeling techniques (Learning Classifier Systems, LCS), we focus on the twofold impact of cognitive and environmental complexity on learning and coordination. Within this framework, we introduce the notion of Adaptive Learning Agent with Rule-based Memory (ALARM), which is a particular class of Artificial Adaptive Agent (AAA, Holland and Miller 1991). We show that equilibrium is approached to a high degree, but never perfectly reached. We also (...) demonstrate that memorization and learning capacities depend upon the relative discordance between the cognitive complexity of agents' mental models and the degree of stability of the environment. (shrink)

The agent-relative/agent-neutral distinction is one of the most important in contemporary moral theory. Yet, providing an adequate formal account of it has proven difficult. In this article I defend a new formal account of the distinction, one that avoids various problems faced by other accounts. My account is based on an influential account of the distinction developed by McNaughton and Rawling. I argue that their approach is on the right track but that it succumbs to two serious objections. I (...) then show how to formulate a new account that follows the key insights of McNaughton and Rawling’s approach yet avoids the two objections. (shrink)

Agent-based models derive the behavior of artificial socio-economic entities computationally from the actions of a large number of agents. One objection is that highly idealized ABMs fail to represent the real world in any reasonable sense. Another objection is that they at best show how observed patterns may have come about, because simulations are easy to produce and there is no evidence that this is really what happens. Moreover, different models may well yield the same result. I will (...) rebut these objections by focusing on an often neglected, but crucial function of ABMs. Building on Gelfert’s account of the exploratory uses of scientific models I show that, in the absence of an accepted underlying theory, successful ABMs lend inductive support to assumptions concerning certain structural feutures of the behavioral rules employed. One core step towards this goal is what I call multiple-model robustness analysis. (shrink)

In this paper, we propose an empirically based Monte Carlo bus-network model as a test bed to simulate intervention strategies to overcome the inefficiencies of bus bunching. The EMB model is an agent-based model which utilizes the positional and temporal data of the buses obtained from the Global Positioning System to constitute a set of empirical velocity distributions of the buses and a set of exponential distributions of interarrival time of passengers at the bus stops. Monte Carlo sampling (...) is then performed on these two derived probability distributions to yield the stochastic dynamics of both the buses’ motion and passengers’ arrival. Our EMB model is generic and can be applied to any real-world bus network system. In particular, we have validated the model against the Nanyang Technological University’s Shuttle Bus System by demonstrating its accuracy in capturing the bunching dynamics of the shuttle buses. Furthermore, we have analyzed the efficacy of three intervention strategies: holding, no-boarding, and centralized-pulsing, against bus bunching by incorporating the rule set of these strategies into the model. Under the scenario where the buses have the same velocity, we found that all three strategies improve both the waiting and travelling times of the commuters. However, when the buses have different velocities, only the centralized-pulsing scheme consistently outperforms the control scenario where the buses periodically bunch together. (shrink)

Given that artificial moral agents—such as autonomous vehicles, lethal autonomous weapons, and automated financial trading systems—are now part of the socio-ethical equation, we should morally evaluate their behavior. How should artificial moral agents make decisions? Is one moral theory better suited than others for machine ethics? After briefly overviewing the dominant ethical approaches for building morality into machines, this paper discusses a recent proposal, put forward by Don Howard and Ioan Muntean (2016, 2017), for an artificial moral agent (...) based on virtue theory. While the virtuous artificial moral agent has various strengths, this paper argues that a rule-based utilitarian approach (in contrast to a strict act-utilitarian approach) is superior because it can capture the most important features of the virtue-theoretic approach while realizing additional significant benefits. Specifically, a 2-level utilitarian artificial moral agent incorporating both established moral rules and a utility calculator is especially well-suited for machine ethics. (shrink)

In this paper we address the modelling of reasoning processes in natural agents. We focus on a very basic kind of non-monotonic inference for which we identify a simple and plausible underlying process, and we develop a family of logical models that allow to match this process. Partial worlds models, as we call them, are a variant of Kraus, Lehmann and Magidor’s cumulative models. We show that the inference relations they induce form a strict subclass of cumulative relations and (...) tackle the issue of providing sound and complete sets of rules to characterise them. Taking inspiration from the work of Gabbay and Schlechta, we analyse the question in terms of definable sets of partial worlds and conclude that completeness is probably unreachable using a standard propositional language. This brings us to enrich our language with an additional connective, which makes it possible to distinguish between two kinds of disjunction in the partial worlds context. Within this renewed framework, we provide two representation theorems: one for inference relations induced by precisification-free smooth models, and the other for inference relations induced by precisification-free ranked models. Finally we give an aperçu of how partial worlds models lend themselves to the modelling of learning. (shrink)

This article presents a rule-utilitarian theory which lies much closer to the social contract tradition than most other forms of consequentialism do: calculated-rates rule preference utilitarianism. Being preference-utilitarian allows the theory to be grounded in instrumental rationality and the equality of agents, as opposed to teleological assumptions about impartial goodness. The calculated-rates approach, judging rules’ consequences by what would happen if they were accepted by whatever number of people is realistic rather than by what would happen if (...) they were accepted universally or by exactly 90% of the population, allows it to select rules based not just on their ability to give good advice to their followers but also on their ability to attract followers in the first place. The result is a theory that, although fully utilitarian and not at all pluralistic or intuitionist, nevertheless offers a principled justification for giving some weight to seemingly non-utilitarian considerations: Lockean natural rights, Kantian respect for autonomy, and Scanlonian distributive justice. (shrink)

In normative multi-agent systems, the question of “how an agent identifies norms in an open agent society” has not received much attention. This paper aims at addressing this question. To this end, this paper proposes an architecture for norm identification for an agent. The architecture is based on observation of interactions between agents. This architecture enables an autonomous agent to identify prohibition norms in a society using the prohibition norm identification (PNI) algorithm. The PNI algorithm uses association (...) class='Hi'>rule mining, a data mining approach to identify sequences of events as candidate norms. When a norm changes, an agent using our architecture will be able to modify the norm and also remove a norm if it does not hold in the society. Using simulations of a park scenario we demonstrate how an agent makes use of the norm identification framework to identify prohibition norms. (shrink)

When people learn normative systems, they do so based on limited evidence. Many of the possible actions that are available to an agent have never been explicitly permitted or prohibited. But people will often need to figure out whether those unspecified actions are permitted or prohibited. How does a learner resolve this incompleteness? The learner might assume if an action-type is not expressly forbidden, then acts of that type are permitted. This closure principle is one of Liberty. Alternatively, the (...) learner might assume that if an action-type is not expressly permitted, then acts of that type are prohibited. This closure principle would be one of Residual Prohibition (Mikhail, 2011). On the basis of principles of pedagogical sampling (e.g., Shafto, Goodman, & Griffiths, 2014), we predicted that participants would infer the Liberty Principle (LP) when trained on prohibitions, and they would infer the Residual Prohibition Principle when trained on permissions. This is exactly what we found across several experiments. We also found a bias in favor of Liberty insofar as participants trained on both a prohibition and a permission rule tended still to infer the LP. However, we also found that if an action is potentially harmful, this diminishes the tendency to infer the LP. (shrink)

Public reason liberalism includes an ideal of political stability where justified institutions reach a kind of self-enforcing equilibrium. Such an order must be stable for the right reasons — where persons comply with the rules of the order for moral reasons, rather than out of fear or self-interest. John Rawls called a society stable in this way well-ordered. In this essay, I contend that a more sophisticated model of a well-ordered society, specifically an agent-based model, yields a richer and (...) more attractive understanding of political stability. An agent-based model helps us to distinguish between three concepts of political stability — durability, balance, and immunity. A well-ordered society is one that possesses a high degree of social trust and cooperative behavior among its citizens (durability) with low short-run variability (balance). A well-ordered society also resists destabilization caused by noncompliant agents in or entering the system (immunity). Distinguishing between these three concepts complicates the necessary reformulation of the idea of a well-ordered society. Going forward, public reason theorists must now distinguish between types of assurance, specify heretofore unknown aspects of reasonable behavior, and reconceive of the nonideal preconditions for forming a stable, ideal social order. (shrink)

While Aristotle in the Nicomachean Ethics does not provide a guide for action in the form of rules for a decision process as deontological or consequentialistethical theories purport to do, he does present a description of the virtuous agent and the virtues that this agent exercises in his choices of action. In this paper Iargue that Aristotle’s mature virtuous agent characteristically exercises the virtue of wisdom (sophia) as well as the practical virtues of character and intelligence in his choices of (...) action and that students of virtue (those sincerely interested in becoming virtuous by acting as the virtuous agent does) can derive certain action-guiding rules from a description of these three virtues and how they are exercised by the mature virtuous agent in any given choice of action. (shrink)

In this article, through a comparative analysis of Dewey's and Laozi's relevant accounts, I examine a pragmatic insight concerning moral rules and moral experience to the effect that (i) fixed and formulated moral rules should not be taken as the final absolute moral authority, and (ii) attention needs to be paid to the moral agent's own moral experience that responds to the felt demands in concrete situations. The purpose of this paper is to enhance understanding the crucial points of the (...) pragmatic insight and to look at how, in certain complementary ways, Dewey's and Laozi's distinct approaches could contribute to the pragmatic insight and learn from each other. I endeavour to show several points: (1) The pragmatic insight has its distinct metaphysical foundations in Dewey's and Laozi's accounts, whose combination could enhance each other's visions and overcome each other's limitations; (2) Both Dewey and Laozi reject some sharp dualism to look at the nature of moral experience that responds to the felt demands in concrete situations; in so doing, their distinct focuses on different aspects, or developing stages, of such moral experience could be complementarily coordinated into a whole; (3) Their characterisations of the pragmatic insight are also based upon their distinct but related naturalistic perspectives to human moral foundation; Laozi's approach could provide some constructive insight for and due natural limitations on Dewey's account. (shrink)

We are moving towards a future where Artificial Intelligence (AI) based agents make many decisions on behalf of humans. From healthcare decision-making to social media censoring, these agents face problems, and make decisions with ethical and societal implications. Ethical behaviour is a critical characteristic that we would like in a human-centric AI. A common observation in human-centric industries, like the service industry and healthcare, is that their professionals tend to break rules, if necessary, for pro-social reasons. This (...) behaviour among humans is defined as pro-social rule breaking. To make AI agents more human-centric, we argue that there is a need for a mechanism that helps AI agents identify when to break rules set by their designers. To understand when AI agents need to break rules, we examine the conditions under which humans break rules for pro-social reasons. In this paper, we present a study that introduces a ‘vaccination strategy dilemma’ to human participants and analyzes their response. In this dilemma, one needs to decide whether they would distribute COVID-19 vaccines only to members of a high-risk group (follow the enforced rule) or, in selected cases, administer the vaccine to a few social influencers (break the rule), which might yield an overall greater benefit to society. The results of the empirical study suggest a relationship between stakeholder utilities and pro-social rule breaking (PSRB), which neither deontological nor utilitarian ethics completely explain. Finally, the paper discusses the design characteristics of an ethical agent capable of PSRB and the future research directions on PSRB in the AI realm. We hope that this will inform the design of future AI agents, and their decision-making behaviour. (shrink)

In a recent paper I argued that agent causation theorists should be compatibilists. In this paper, I argue that compatibilists should be agent causation theorists. I consider six of the main problems facing compatibilism: (i) the powerful intuition that one can't be responsible for actions that were somehow determined before one was born; (ii) Peter van Inwagen's modal argument, involving the inference rule (β); (iii) the objection to compatibilism that is based on claiming that the ability to do (...) otherwise is a necessary condition for freedom; (iv) "manipulation arguments," involving cases in which an agent is manipulated by some powerful being into doing something that he or she would not normally do, but in such a way that the compatibilist's favorite conditions for a free action are satisfied; (v) the problem of constitutive luck; and (vi) the claim that it is not fair to blame someone for an action if that person was determined by forces outside of his or her control to perform that action. And in the case of each of these problems, I argue that the compatibilist has a much more plausible response to that problem if she endorses the theory of agent causation than she does otherwise. (shrink)

This thesis proposes that an account of first-person reference and first-person thinking requires an account of practical knowledge. At a minimum, first-person reference requires at least a capacity for knowledge of the intentional act of reference. More typically, first-person reasoning requires deliberation and the ability to draw inferences while entertaining different 'I' thoughts. Other accounts of first-person reference--such as the perceptual account and the rule-based account--are criticized as inadequate. An account of practical knowledge is provided by an interpretation (...) of GEM Anscombe's account in her landmark monograph "Intention". (shrink)

Foundational theories of mental content seek to identify the conditions under which a mental representation expresses, in the mind of a particular thinker, a particular content. Normativists endorse the following general sort of foundational theory of mental content: A mental representation r expresses concept C for agent S just in case S ought to use r in conformity with some particular pattern of use associated with C. In response to Normativist theories of content, Kathrin Glüer-Pagin and Åsa Wikforss propose a (...) dilemma, alleging that Normativism either entails a vicious regress or falls prey to a charge of idleness. In this paper, I respond to this argument. I argue that Normativism can avoid the commitments that generate the regress and does not propose the sort of explanation required to charge that its explanation has been shown to be problematically idle. The regress-generating commitment to be avoided is, roughly, that tokened, contentful mental states are the product of rule-following. The explanatory task Normativists should disavow is that of explaining how it is that beliefs and other contentful mental states are produced. I argue that Normativism, properly understood as a theory of content, does not provide this kind of psychological explanation, and therefore does not entail that such explanations are to be given in terms of rule-following. If this is correct, Normativism is not the proper target of the dilemma offered by Glüer-Pagin and Wikforss. Understanding why one might construe Normativism in the way Glüer-Pagin and Wikforss must, and how, properly understood, it avoids their dilemma, can help us to appreciate the attractiveness of a genuinely normative theory of content and the importance of paying careful attention to the sort of normativity involved in norm-based theories of content. (shrink)

While the sociality of software agents drives toward the definition of institutions for multi agent systems, their autonomy requires that such institutions are ruled by appropriate norm mechanisms. Computational institutions represent useful abstractions. In this paper we show how computational institutions can be built on top of the RoleX infrastructure, a role-based system with interesting features for our aim. We achieve a twofold goal: on the one hand, we give concreteness to the institution abstractions; on the other hand, (...) we demonstrate the flexibility of the RoleX infrastructure. (shrink)

There is an increasing need for norms to be embedded in technology as the widespread deployment of applications such as autonomous driving, warfare and big data analysis for crime fighting and counter-terrorism becomes ever closer. Current approaches to norms in multi-agent systems tend either to simply make prohibited actions unavailable, or to provide a set of rules which the agent is obliged to follow, either as part of its design or to avoid sanctions and punishments. In this paper we argue (...) for the position that agents should be equipped with the ability to reason about a system’s norms, by reasoning about the social and moral values that norms are designed to serve; that is, perform the sort of moral reasoning we expect of humans. In particular we highlight the need for such reasoning when circumstances are such that the rules should arguably be broken, so that the reasoning can guide agents in deciding whether to comply with the norms and, if violation is desirable, how best to violate them. One approach to enabling this is to make use of an argumentation scheme based on values and designed for practical reasoning: arguments for and against actions are generated using this scheme and agents choose between actions based on their preferences over these values. Moral reasoning then requires that agents have an acceptable set of values and an acceptable ordering on their values. We first discuss how this approach can be used to think about and justify norms in general, and then discuss how this reasoning can be used to think about when norms should be violated, and the form this violation should take. We illustrate how value based reasoning can be used to decide when and how to violate a norm using a road traffic example. We also briefly consider what makes an ordering on values acceptable, and how such an ordering might be determined. (shrink)

How situated embodied agents may achieve goals using knowledge is the classical question of natural and artificial intelligence. How organisms achieve this with their nervous systems is a central challenge for a neural theory of embodied cognition. To structure this challenge, we borrow terms from Searle's analysis of intentionality in its two directions of fit and six psychological modes (perception, memory, belief, intention-in-action, prior intention, desire). We postulate that intentional states are instantiated by neural activation patterns that are stabilized (...) by neural interaction. Dynamic instabilities provide the neural mechanism for initiating and terminating intentional states and are critical to organizing sequences of intentional states. Beliefs represented by networks of concept nodes are autonomously learned and activated in response to desired outcomes. The neural dynamic principles of an intentional agent are demonstrated in a toy scenario in which a robotic agent explores an environment and paints objects in desired colors based on learned color transformation rules. (shrink)

Given the complexity of human social life, it is astonishing to observe how quickly children adapt to their social environment. To be accepted by the other members, it is crucial to understand and follow the rules and norms shared by the group. How and from whom do young children learn these social rules? In the experiments, based on the crucial role of affective social learning and dominance hierarchies in simple rule understanding, we showed 15-to-23-month-olds and 3-to-5-year-old children videos (...) where the agents’ body size and affective cues were manipulated. In the dominant rule-maker condition, when a smaller protagonist puts an object in one location, a bigger agent reacts with a positive reaction; on the contrary, when the smaller protagonist puts an object in another location, the bigger agent displays a negative reaction. In the subordinate rule-maker condition, the roles are shifted but the agents differ. Toddlers expect the protagonist to follow the rules (based on anticipatory looks), independent of the dominant status of the rule-making agent. Three-to-five-year-old pre-schoolers overall perform at the chance level but expect the protagonist to disobey a rule in the first trial, and obey the rule in the second trial if the rule-maker is dominant. (shrink)

How should we update de dicto beliefs in the face of de se evidence? The Sleeping Beauty problem divides philosophers into two camps, halfers and thirders. But there is some disagreement among halfers about how their position should generalize to other examples. A full generalization is not always given; one notable exception is the Halfer Rule, under which the agent updates her uncentered beliefs based on only the uncentered part of her evidence. In this brief article, I provide (...) a simple example for which the Halfer Rule prescribes credences that, I argue, cannot be reasonably held by anyone. In particular, these credences constitute an egregious violation of the Reflection Principle. I then discuss the consequences for halfing in general. (shrink)

This article explores how high-functioning children with autism navigate in the social world, specifically how they orient in the realm of norms and standards. In particular, this investigation focuses on rule violations episodes and sheds light on how these children account for their conduct and position themselves in the moral framework. This analysis shows that high-functioning children with autism can actively engage in discourse about norms and transgressions in an initiatory capacity, thereby displaying a mastery of social rules as (...) a guide for appropriate conduct and interpretation of others’ behavior. Furthermore, this article argues that these social skills are linked with the ability to operate with sequentially based understandings: Prior courses of action constitute for the autistic children the fundamental source for reaching an understanding of the normative mechanics of everyday life, and subsequently for constructing their own lines of conduct and themselves as moral agents. (shrink)

Normative moral theories are frequently invoked to serve one of two distinct purposes: (1) explicate a criterion of rightness, or (2) provide an ethical decision-making procedure. Although a criterion of rightness provides a valuable theoretical ideal, proposed criteria rarely can be (nor are they intended to be) directly translated into a feasible decision-making procedure. This paper applies the computational framework of bounded rationality to moral decision-making to ask: how ought a bounded human agent make ethical decisions? We suggest agents (...) ought to follow moral maxims: principles that approximate rightness in many situations, but that can be overridden in specific, precisely describable circumstances. While this intuitive idea has been proposed many times before, we provide a precise model of how maxim consequentialism functions as an approximation to an act-consequentialist criterion of rightness, while maintaining the flexibility and defeasibility that has eluded most forms of rule consequentialism. Furthermore, while our overarching aim is to propose a new normative standard of moral decision-making, we demonstrate how maxim consequentialism can also function as a descriptive account of human behavior. We conclude by noting that different criteria of rightness may lead to different maxim-based ethics. (shrink)

R. M. Hare has argued for and defended a ‘two-level’, view of moral agency. He argues that moral agents ought to rely on the rules of ‘intuitive moral thinking’ for their ‘everyday’ moral judgments. When these rules conflict or when we do not have a rule at hand, we ought to ascend to the act-utilitarian,‘critical’ level of moral thinking. I argue that since the rules at the intuitive level of moral thinking necessarily conflict much more often than Hare (...) supposes, and since we often do not have ready-made rules for our moral judgments, we must necessarily use critical moral thinking very frequently. However, act-utilitarian judgements at this level will sharply conflict with our strongly held ‘intuitive’ moral convictions. I show that Hare's attempt to balance these two aspects of moral judgment requires us to simultaneously adopt two conflicting sets of moral standards, and thus an attempt to inculcate such standards constitutes a ‘schizophrenic’ moral education. Finally, I briefly outline an alternative conception of moral education, based on Aristotelian phronesis. (shrink)

I propose an account of desire that reconciles two apparently conflicting intuitions about practical agency. I do so by exploring a certain intuitive datum. The intuitive datum is that often when an agent desires P she will seem to immediately and conclusively know that there is a reason to bring P about. Desire-based theories of reasons seem uniquely placed to explain this intuitive datum. On this view, desires are the source of an agent’s practical reasons. A desire for P (...) grounds conclusive knowledge of a reason to bring P about because that desire makes it true that there is a reason to do so. However, this implies that a basic desire for P can never be in error about there being at least some reason to bring P about. We have the conflicting intuition that basic desires sometimes rationally count for nothing. The guise of the good explains this intuition about the fallibility of desires. On this view, a desire for P represents P as good in some respect. Desires and reasons are independent, so a desire might misrepresent one’s reasons. But this independence is usually taken to rule out that desires ever provide conclusive knowledge of reasons. Capturing the intuition about conclusive knowledge rules out capturing the intuition about fallibility, and vice versa. I propose an epistemological disjunctivist version of the guise of the good that reconciles fallibility with the possibility of conclusive knowledge. (shrink)

For many commentators, Confucian ethics is a kind of virtue ethics. However, there is enough textual evidence to suggest that it can be interpreted as an ethics based on rules, consequentialist as well as deontological. Against these views, I argue that Confucian ethics is based on the roles that make an agent the person he or she is. Further, I argue that in Confucianism the question of what it is that a person ought to do cannot be separated (...) from the question of what it is to be a person, and that the latter is answered in terms of the roles that arise from the network of social relationships in which a person stands. This does not mean that Confucian ethics is unlike anything found in Western philosophy. Indeed, I show that many Western thinkers have advanced a view of ethics similar to the Confucian ethics as I interpret it. (shrink)

Agents need to be able to change their beliefs; in particular, they should be able to contract or remove a certain belief in order to restore consistency to their set of beliefs, and revise their beliefs by incorporating a new belief which may be inconsistent with their previous beliefs. An influential theory of belief change proposed by Alchourron, G¨ardenfors and Makinson (AGM) [1] describes postulates which a rational belief revision and contraction operations should satisfy. The AGM postulates have been (...) perceived as characterising idealised rational reasoners, and the corresponding belief change operations are considered unsuitable for implementable agents due to their high computational cost [3]. The main result of this paper is showing that an efficient (linear time) belief contraction operation nevertheless satisfies all but one of the AGM postulates for contraction. This contraction operation is defined for a realistic rule-based agent which can be seen as a reasoner in a very weak logic; although the agent’s beliefs are deductively closed with respect to this logic, checking consistency and tracing dependencies between beliefs is not computationally expensive. Finally, we give a non-standard definition of belief revision in terms of contraction for our agent. (shrink)

Assuming that moral discourse is prescriptive, what distinguishes it from other types of prescriptive discourse? To say, as Hare does, that it is its overridingness, is subject both to criticisms that assume that overridingness could, in principle, be used to distinguish one type of prescriptive discourse from another, and then show that it is doubtful that moral discourse is overriding, and to the criticisms that claim that it is in principle impossible to use overridingness to distinguish one type of prescriptive (...) discourse from another, because all of them are bound to be on a par in respect of overridingness. ;It is also impossible to distinguish moral discourse from other types of prescriptive discourse by saying that in it we use arguments based on imaginatively putting oneself in the shoes of others, because such arguments are used in prudential discourse as well. However, we can account for the distinction, if we realize that such arguments can be performed only on certain designated argument-places, and that in moral claims argument-places of two different types are designated for the purpose: those for moral agents and those for moral patients; while in prudential claims argument-places of only one type are designated: those for prudential agents. ;If this account is accepted, this raises a number of further questions. Examination of these questions leads to a form of relativism about membership in the sets of moral agents and moral patients. This form of relativism, however, leaves considerable room for rational discussion of membership in these sets and is compatible with the rejection of relativism about the content of moral rules. (shrink)