Causal understanding is a defining characteristic of human cognition. Like many animals, human children learn to control their bodily movements and act effectively in the environment. Like a smaller subset of animals, children intervene: they learn to change the environment in targeted ways. Unlike other animals, children grow into adults with the causal reasoning skills to develop abstract theories, invent sophisticated technologies and imagine alternate pasts, distant futures and fictional worlds. In this Review, we explore the development of (...) human-unique causal learning and reasoning from evolutionary and ontogenetic perspectives. We frame our discussion using an ‘interventionist’ approach. First, we situate causal understanding in relation to cognitive abilities shared with non-human animals. We argue that human causal understanding is distinguished by its depersonalized (objective) and decontextualized (general) representations. Using this framework, we next review empirical findings on early human causal learning and reasoning and consider the naturalistic contexts that support its development. Then we explore connections to related abilities. We conclude with suggestions for ongoing collaboration between developmental, cross-cultural, computational, neural and evolutionary approaches to causal understanding. (shrink)

Discovering high-level causal relations from low-level data is an important and challenging problem that comes up frequently in the natural and social sciences. In a series of papers, Chalupka etal. (2015, 2016a, 2016b, 2017) develop a procedure forcausal feature learning (CFL) in an effortto automate this task. We argue that CFL does not recommend coarsening in cases where pragmatic considerations rule in favor of it, and recommends coarsening in cases where pragmatic considerations rule against it. We propose a (...) new technique, pragmatic causal feature learning (PCFL), which extends the original CFL algorithm in useful and intuitive ways. We show that PCFL has the same attractive measure-theoretic properties as the original CFL algorithm. We compare the performance of both methods through theoretical analysis and experiments. (shrink)

This article provides a conceptual account of causal understanding by connecting current psychological research on time and causality with philosophical debates on the causal asymmetry. I argue that causal relations are viewed as asymmetric because they are understood in temporal terms. I investigate evidence from causal learning and reasoning in both children and adults: causal perception, the temporal priority principle, and the use of temporal cues for causal inference. While this account does not (...) suffice for correct inferences of causal structure, I show it to serve as a preliminary understanding of causal concepts as asymmetric, that later incorporates other types of evidence. This approach supplies causal models with an asymmetric concept of causation that underlies hypotheses about causal structure, as I will illustrate from the framework of the knowledge-based causal induction model. I further argue for an integrating perspective, showing how the understanding of causes as preceding their effects underlies both psychological models and philosophical debates over time and the causal asymmetry, particularly regarding problem cases such as simultaneous causation or backwards causation, and the conceptual connection between causation and action. (shrink)

How feasible are conceptual engineering projects of social concepts that aim for the engineered concept to be widely adopted in ordinary everyday life? Predominant frameworks on the psychology of concepts that shape work on stereotyping, bias, and machine learning have grim implications for the prospects of conceptual engineers: conceptual engineering efforts are ineffective in promoting certain social-conceptual changes. Specifically, since conceptual components that give rise to problematic social stereotypes are sensitive to statistical structures of the environment, purely conceptual change (...) won’t be possible without corresponding world change. This tradition, however, tends to ignore that concepts don’t only encode statistical, but also causal information. Paying attention to this feature of concepts, I argue, shows that conceptual engineering is not only possible. There is an imperative to conceptually-engineer. (shrink)

This study explores a liberatory epistemic virtue that is suitable for good learning as a form of liberating socially situated epistemic agents toward ideal virtuousness. First, I demonstrate that the weak neutralization of epistemically bad stereotypes is an end of good learning. Second, I argue that weak neutralization represents a liberatory epistemic virtue, the value of which derives from liberating us as socially situated learners from epistemic blindness to epistemic freedom. Third, I explicate two distinct forms of epistemic (...) transformation: constitutive and causal epistemic transformation. I argue that compared with the ideal conception of epistemic virtue, constitutive epistemic transformation that involves good learning has a transcendent value in light of agents constantly renewing their default epistemic status toward ideal virtuousness. (shrink)

Proponents of cognitive penetration often argue for the thesis on the basis of combined intuitions about categorical perception and perceptual learning. The claim is that beliefs penetrate perceptions in the course of learning to perceive categories. I argue that this "diachronic" penetration thesis is false. In order to substantiate a robust notion of penetration, the beliefs that enable learning must describe the particular ability that subjects learn. However, they cannot do so, since in order to help with (...) learning they must instruct learners to employ previously existing abilities. I argue that a better approach recognizes that we can have sophisticated causal precursors to perceptual learning, but that the learning process itself must operate outside of cognitive influence. (shrink)

This chapter begins with a discussion of the significance of studies of aspects of tool use in understanding causal cognition. It argues that tool use studies reveal the most basic type or causal understanding being put to use, in a way that studies that focus on learning statistical relationships between cause and effect or studies of perceptual causation do not. An overview of the subsequent chapters is also presented.

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)

It is commonly thought that Hume endorses the claim that causal cognition can be fully explained in terms of nothing but custom and habit. Associative learning does, of course, play a major role in the cognitive psychology of the Treatise. But Hume recognizes that associations cannot provide a complete account of causal thought. If human beings lacked the capacity to reflect on rules for judging causes and effects, then we could not (as we do) distinguish between accidental (...) and genuine regularities, and Hume could not (as he does) carry out his science of human nature. One might reply that what appears to be rule-governed behavior might emerge from associative systems that do not literally employ rules. But this response fails: there is a growing consensus in cognitive science that any adequate account of causal learning must invoke active, controlled cognitive processes. (shrink)

Modularity theorists have challenged that there are, or could be, general learning mechanisms that explain theory-of-mind development. In response, supporters of the ‘scientific theory-theory’ account of theory-of-mind development have appealed to children's use of auxiliary hypotheses and probabilistic causal modeling. This article argues that these general learning mechanisms are not sufficient to meet the modularist's challenge. The article then explores an alternative domain-general learning mechanism by proposing that children grasp the concept belief through the progressive alignment (...) of relational structure that occurs as a result of structural-comparison. The article also explores the implications of the proposed account for Fodor's puzzle of conceptual learning. (shrink)

As the demand for explainable deep learning grows in the evaluation of language technologies, the value of a principled grounding for those explanations grows as well. Here we study the state-of-the-art in explanation for neural models for natural-language processing (NLP) tasks from the viewpoint of philosophy of science. We focus on recent evaluation work that finds brittleness in explanations obtained through attention mechanisms.We harness philosophical accounts of explanation to suggest broader conclusions from these studies. From this analysis, we assert (...) the impossibility of causal explanations from attention layers over text data. We then introduce NLP researchers to contemporary philosophy of science theories that allow robust yet non-causal reasoning in explanation, giving computer scientists a vocabulary for future research. (shrink)

Learning mathematics through distance education can be challenging, with the “logical” and “rewarding” nature proving difficult to measure. This article aimed to articulate an argument explaining the “logical” and “rewarding” nature of online mathematics learning, elucidating their causal factors. Existing data from the literature that involving students at Visayas State University, Philippines, were utilized in this study. The study used statistical measures to capture descriptions from the data, and quantile regression analysis was employed to forecast the predictors (...) of the logicality and rewarding nature of learning mathematics at a distance. Results indicate that learning mathematics in the new normal is perceived as “logical” and moderately “rewarding”. The regression and correlation analyses revealed a significant positive association between the perceived “logical” nature and the rewarding experience of learning mathematics. The constructed statistical models depicted that the determinants of logicality in learning mathematics include family income, money spent on the internet, learning environment, household size, social status, and health. Moreover, causal factors such as family income, money spent on the internet, learning environment, leisure activities, social status, and health significantly determine the rewarding nature of learning mathematics online. Conclusively, institution must support college students with their online learning needs. Furthermore, mathematics instructors should create lively and exciting online discussions that boost their logical thinking. Providing problem-solving task that are intrinsically rewarding can contribute to a more fulfilling learning experience. (shrink)

In this article it is argued that the optical metaphor and critical practice of diffraction further developed by Donna Haraway and Karen Barad might be no less significant than the widely spread notion of reflection, when the questions of various practices of knowledge are addressed. By considering Paul Ramsden’s approach to learning/teaching and its underlying theory in higher education alongside Karen Barad’s methodology of diffraction, it is shown that Ramsden’s understanding of learning/teaching is rather based on the theoretical (...) assumptions of diffractive practice. His notion that teaching/learning are closely related and actively shaping each other, and that learners are not disconnected from the environment and their previous experiences with the subject matter and learning process itself, adds to Barad’s onto-epistemological position that knowers know the world at the same time as being the part of the world in its ongoing intra-activity. Ramsden’s understanding of relation is diffractive, because it is not about predefined binary entities and their fixed identities, but about layers and entanglements of various previous experiences and reactions to the learning environment. In addition, looking at learning/teaching processes through a different perspective also leads to a different approach to teaching and other ways of problem-solving. Both Ramsden and Barad distrust homologies, analogies, and causality-based conceptions of knowledge sharing. Instead, the ability to respond to an always new learning/teaching environment is assessed, which implies a diffractive type of sensitivity to the context, iterative process of re-turning, and the creation of dangerously indeterminate relationships and commitments. In this way, some of Barad’s philosophical notions, i.e., the diffraction pattern, intra-activity, returning, and others, also may acquire new practical content. (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)

This paper critically reviews and characterizes the student's causal-explanatory understanding; this is done as a step toward explicating the problematic of evolution education as it concerns the cognitive difficulties in understanding Darwin's theory of natural selection. The review concludes that the student's understanding is fundamentally different from Darwin's, for the student understands evolutionary change as necessary individual transformation caused by the transformative action of various physical and behavioral factors. This is in complete contrast to Darwin's (and even the Darwinian's, (...) for that matter) understanding of evolutionary change as a change caused by accumulative selection. Hence, to understand natural selection, the student has to learn to “see” how the accumulative selection causes evolutionary change. (shrink)

The use of counterfactuals for considerations of algorithmic fairness and explainability is gaining prominence within the machine learning community and industry. This paper argues for more caution with the use of counterfactuals when the facts to be considered are social categories such as race or gender. We review a broad body of papers from philosophy and social sciences on social ontology and the semantics of counterfactuals, and we conclude that the counterfactual approach in machine learning fairness and social (...) explainability can require an incoherent theory of what social categories are. Our findings suggest that most often the social categories may not admit counterfactual manipulation, and hence may not appropriately satisfy the demands for evaluating the truth or falsity of counterfactuals. This is important because the wide- spread use of counterfactuals in machine learning can lead to misleading results when applied in high-stakes domains. Accordingly, we argue that even though counterfactuals play an essential part in some causal inferences, their use for questions of algorithmic fairness and social explanations can create more problems than they resolve. Our positive result is a set of tenets about using counterfactuals for fairness and explanations in machine learning. (shrink)

Within the discourse surrounding mind-body interaction, mental causation is intimately associated with non-reductive physicalism. However, such a theory holds two opposing views: that all causal properties and relations can be explicated by physics and that special sciences have an explanatory role. Jaegwon Kim attempts to deconstruct this problematic contradiction by arguing that it is untenable for non-reductive physicalists to explain human behavior by appeal to mental properties. In combination, Kim’s critique of mental causation and the phenomenal concept strategy serves (...) as an effectual response to the anti-physicalist stance enclosed within the Knowledge Argument and the Zombie Thought Experiment. (shrink)

We investigate flexibility and problem solving in skilled action. We conducted a field study of mountain bike riding that required a learner rider to cope with major changes in technique and equipment. Our results indicate that relatively inexperienced individuals can be capable of fairly complex 'on-the-fly' problem solving which allows them to cope with new conditions. This problem solving is hard to explain for classical theories of skill because the adjustments are too large to be achieved by automatic mechanisms and (...) too complex and rapid to be achieved by cognitive processes as they are usually understood. A recent theory, Mesh, can explain these results because it posits that skill-specific cognitive abilities develop during skill learning and that control typically involves an interplay between cognitive and automatic mechanisms. Here we develop Mesh further, providing a detailed explanation for these problem solving abilities. We argue that causal representation, metacognitive awareness and other forms of performance awareness combine in the formulation and control of action strategies. We also argue that the structure of control present in this case is inconsistent with Bratman's model of intentions, and that, in the face of high uncertainty and risk, intentions can be much more labile than Bratman recognises. In addition, we found limitations and flaws in problem solving which illuminate the representations involved. Finally, we highlight the crucial role of social and cultural learning in the development of complex skills. (shrink)

A new model of the development of temporal concepts is described that assumes that there are substantial changes in how children think about time in the early years. It is argued that there is a shift from understanding time in an event-dependent way to an event-independent understanding of time. Early in development, very young children are unable to think about locations in time independently of the events that occur at those locations. It is only with development that children begin to (...) have a proper grasp of the distinction between past, present, and future, and represent time as linear and unidirectional. The model assumes that although children aged 2 to 3 years may categorize events differently depending on whether they lie in the past or the future, they may not be able to understand that whether an event is in the past or future is something that changes as time passes and varies with temporal perspective. Around 4 to 5 years, children understand how causality operates in time, and can grasp the systematic relations that obtain between different locations in time, which provides the basis for acquiring the conventional clock and calendar system. (shrink)

In temporal binding, the temporal interval between one event and another, occurring some time later, is subjectively compressed. We discuss two ways in which temporal binding has been conceptualized. In studies showing temporal binding between a voluntary action and its causal consequences, such binding is typically interpreted as providing a measure of an implicit or pre-reflective “sense of agency”. However, temporal binding has also been observed in contexts not involving voluntary action, but only the passive observation of a cause-effect (...) sequence. In those contexts, it has been interpreted as a top-down effect on perception reflecting a belief in causality. These two views need not be in conflict with one another, if one thinks of them as concerning two separate mechanisms through which temporal binding can occur. In this paper, we explore an alternative possibility: that there is a unitary way of explaining temporal binding both within and outside the context of voluntary action as a top-down effect on perception reflecting a belief in causality. Any such explanation needs to account for ways in which agency, and factors connected with agency, have been shown to affect the strength of temporal binding. We show that principles of causal inference and causal selection already familiar from the literature on causal learning have the potential to explain why the strength of people’s causal beliefs can be affected by the extent to which they are themselves actively involved in bringing about events, thus in turn affecting binding. (shrink)

Considerable evidence proves that causal learning and causal understanding greatly enhance our ability to manipulate the physical world and are major factors that distinguish humans from other primates. How do we enable unintelligent robots to think causally, answer the questions raised with "why" and even understand the meaning of such questions? The solution is one of the keys to realizing artificial intelligence. Judea Pearl believes that to achieve human-like intelligence, researchers must start by imitating the intelligence of (...) children, so he proposed a "causal inference engine" to help future artificial intelligence make causal inference, pass the Minimal Turing Test, and even become a moral subject who can discern good from evil. This study attempts to provide some insights into the development of children's education from basic assumptions and construction goals of artificial intelligence, and to reflect on the causal model of artificial intelligence through children's education. (shrink)

Empirical research has recently emerged as a key method for understanding the nature of causation, and our concept of causation. One thread of research aims to test intuitions about the nature of causation in a variety of classic cases. These experiments have principally been used to try to resolve certain debates within analytic philosophy, most notably that between proponents of transference and dependence views of causation. The other major thread of empirical research on our concept of causation has investigated the (...) role that normative considerations play in causal judgments. These experimental results suggest that philosophical accounts of our concept of causation should take a broader view of what might be relevant. For both lines of research, we describe some of the significant experiments and outline key philosophical morals that have been drawn, all while pointing out various limitations. We conclude by considering other kinds of empirical research that should be philosophically interesting for those studying the nature of causation and our concept of causation. In particular, we point towards the need for philosophical research about causal perception, causal reasoning, and causal learning, as well as ways in which this research could play a role in prescriptive metaphysics. (shrink)

Causal models of agents have been used to analyse the safety aspects of machine learning systems. But identifying agents is non-trivial -- often the causal model is just assumed by the modeler without much justification -- and modelling failures can lead to mistakes in the safety analysis. This paper proposes the first formal causal definition of agents -- roughly that agents are systems that would adapt their policy if their actions influenced the world in a different (...) way. From this we derive the first causal discovery algorithm for discovering agents from empirical data, and give algorithms for translating between causal models and game-theoretic influence diagrams. We demonstrate our approach by resolving some previous confusions caused by incorrect causal modelling of agents. (shrink)

Much of the modern philosophy of causation has been governed by two ideas: (i) causes make their effects inevitable; (ii) a cause is something that makes a difference to whether its effect occurs. I focus on explaining the origin of idea (ii) and its connection to (i). On my view, the frequent attempts to turn (ii) into an analysis of causation are wrongheaded. Patterns of difference-making aren't what makes causal claims true. They merely provide a useful test for (...) class='Hi'>causal claims. Moreover, what justifies us in using them as a test is idea (i). That's how (i) and (ii) are connected. (shrink)

Orthodox causal decision theory is unstable. Its advice changes as you make up your mind about what you will do. Several have objected to this kind of instability and explored stable alternatives. Here, I'll show that explorers in search of stability must part with a vestige of their homeland. There is no plausible stable decision theory which satisfies Savage's Sure Thing Principle. So those in search of stability must learn to live without it.

Advancements in machine learning have fuelled the popularity of using AI decision algorithms in procedures such as bail hearings, medical diagnoses and recruitment. Academic articles, policy texts, and popularizing books alike warn that such algorithms tend to be opaque: they do not provide explanations for their outcomes. Building on a causal account of transparency and opacity as well as recent work on the value of causal explanation, I formulate a moral concern for opaque algorithms that is yet (...) to receive a systematic treatment in the literature: when such algorithms are used in life-changing decisions, they can obstruct us from effectively shaping our lives according to our goals and preferences, thus undermining our autonomy. I argue that this concern deserves closer attention as it furnishes the call for transparency in algorithmic decision-making with both new tools and new challenges. (shrink)

Recent breakthroughs in artificial intelligence have primarily resulted from training deep neural networks (DNNs) with vast numbers of adjustable parameters on enormous datasets. Due to their complex internal structure, DNNs are frequently characterized as inscrutable ``black boxes,'' making it challenging to interpret the mechanisms underlying their impressive performance. This opacity creates difficulties for explanation, safety assurance, trustworthiness, and comparisons to human cognition, leading to divergent perspectives on these systems. This chapter examines recent developments in interpretability methods for DNNs, with a (...) focus on interventionist approaches inspired by causal explanation in philosophy of science. We argue that these methods offer a promising avenue for understanding how DNNs process information compared to merely behavioral benchmarking and correlational probing. We review key interventionist methods and illustrate their application through practical case studies. These methods allow researchers to identify and manipulate specific computational components within DNNs, providing insights into their causal structure and internal representations. We situate these approaches within the broader framework of causal abstraction, which aims to align low-level neural computations with high-level interpretable models. While acknowledging current limitations, we contend that interventionist methods offer a path towards more rigorous and theoretically grounded interpretability research, potentially informing both AI development and computational cognitive neuroscience. (shrink)

Necessity and sufficiency are the building blocks of all successful explanations. Yet despite their importance, these notions have been conceptually underdeveloped and inconsistently applied in explainable artificial intelligence (XAI), a fast-growing research area that is so far lacking in firm theoretical foundations. Building on work in logic, probability, and causality, we establish the central role of necessity and sufficiency in XAI, unifying seemingly disparate methods in a single formal framework. We provide a sound and complete algorithm for computing explanatory factors (...) with respect to a given context, and demonstrate its flexibility and competitive performance against state of the art alternatives on various tasks. (shrink)

This paper attempts to answer a central question in unsupervised learning: what does it mean to “make sense” of a sensory sequence? In our formalization, making sense involves constructing a symbolic causal theory that both explains the sensory sequence and also satisfies a set of unity conditions. The unity conditions insist that the constituents of the causal theory – objects, properties, and laws – must be integrated into a coherent whole. On our account, making sense of sensory (...) input is a type of program synthesis, but it is unsupervised program synthesis. Our second contribution is a computer implementation, the Apperception Engine, that was designed to satisfy the above requirements. Our system is able to produce interpretable human-readable causal theories from very small amounts of data, because of the strong inductive bias provided by the unity conditions. A causal theory produced by our system is able to predict future sensor readings, as well as retrodict earlier readings, and impute (fill in the blanks of) missing sensory readings, in any combination. In fact, it is able to do all three tasks simultaneously. We tested the engine in a diverse variety of domains, including cellular automata, rhythms and simple nursery tunes, multi-modal binding problems, occlusion tasks, and sequence induction intelligence tests. In each domain, we test our engine's ability to predict future sensor values, retrodict earlier sensor values, and impute missing sensory data. The Apperception Engine performs well in all these domains, significantly out-performing neural net baselines. We note in particular that in the sequence induction intelligence tests, our system achieved human-level performance. This is notable because our system is not a bespoke system designed specifically to solve intelligence tests, but a general-purpose system that was designed to make sense of any sensory sequence. (shrink)

While more species are being identified as cultural on a regular basis, stark differences between human and animal cultures remain. Humans are more richly cultural, with group-specific practices and social norms guiding almost every element of our lives. Furthermore, human culture is seen as cumulative, cooperative, and normative, in contrast to animal cultures. One hypothesis to explain these differences is grounded in the observation that human children across cultures appear to spontaneously over-imitate silly or causally irrelevant behaviors that they observe. (...) The few studies on over-imitation in other species are largely taken as evidence that spontaneous over-imitation is not present in other species. This leads to the over-imitation hypothesis – that the differences between human culture and animal cultures can be traced to the human unique tendency to over-imitate. In this paper, we analyze the current state of the literature on animal over-imitation and challenge the adequacy of the over-imitation hypothesis for the differences between humans and animal cultures. To make this argument, we first argue that the function of human over-imitation is norm-learning and that over-imitation, like skill-learning, should be subject to selective social learning effects. Then we review the empirical evidence against animal over-imitation and argue that these studies do not take into account the relevant variables given the normative and selective nature of over-imitation. We then analyze positive empirical evidence of over-imitation in great apes and canids from the experimental literature and conclude that the current body of evidence suggests that some canids and primates may have the capacity for over-imitation. This paper offers a methodological suggestion for how to study animal over-imitation, and a theoretical suggestion that over-imitation might be much more widely found among species. The larger implication for claims about human uniqueness suggests that if we do find widespread evidence of over-imitation across species, many of the current theories of human uniqueness that focus on human hyper-cooperation or social norms may have only identified a difference of degree, not of kind, between humans and other animals. (shrink)

Since at least the mid-1980s claims have been made for rationality in rats. For example, that rats are capable of inferential reasoning (Blaisdell, Sawa, Leising, & Waldmann, 2006; Bunsey & Eichenbaum, 1996), or that they can make adaptive decisions about future behavior (Foote & Crystal, 2007), or that they are capable of knowledge in propositional-like form (Dickinson, 1985). The stakes are rather high, because these capacities imply concept possession and on some views (e.g., Rödl, 2007; Savanah, 2012) rationality indicates self-consciousness. (...) I evaluate the case for rat rationality by analyzing 5 key research paradigms: spatial navigation, metacognition, transitive inference, causal reasoning, and goal orientation. I conclude that the observed behaviors need not imply rationality by the subjects. Rather, the behavior can be accounted for by noncognitive processes such as hard-wired species typical predispositions or associative learning or (nonconceptual) affordance detection. These mechanisms do not necessarily require or implicate the capacity for rationality. As such there is as yet insufficient evidence that rats can reason. I end by proposing the ‘Staircase Test,’ an experiment designed to provide convincing evidence of rationality in rats. (shrink)

Nowadays, there is a growing interest in transdisciplinary approach to university curriculum development, transdisciplinary Studies are about the realms, goals, and goals of the transition field. The early phases of transdisciplinary in higher education curriculum can be complex and so there are challenges to the definition and operationalization this approach to the university curriculum. In this paper, in respect to the different perspectives on the subject, the conceptual framework and the model of the curriculum based on the causal layered (...) learning are explained and the challenges and obligations of the subject are conceptualized based on four different types of curricular issues: Philosophical and pedagogical considerations, issues and responsibilities affecting the teaching process, issues and responsibilities affecting students learning, issues and responsibilities affecting course design and issues and responsibilities affecting learning assessment. (shrink)

In a recent article, Erasmus, Brunet, and Fisher (2021) argue that Artificial Neural Networks (ANNs) are explainable. They survey four influential accounts of explanation: the Deductive-Nomological model, the Inductive-Statistical model, the Causal-Mechanical model, and the New-Mechanist model. They argue that, on each of these accounts, the features that make something an explanation is invariant with regard to the complexity of the explanans and the explanandum. Therefore, they conclude, the complexity of ANNs (and other Machine Learning models) does not (...) make them less explainable. In this reply, it is argued that Erasmus et al. left out one influential account of explanation from their discussion: the Unificationist model. It is argued that, on the Unificationist model, the features that makes something an explanation is sensitive to complexity. Therefore, on the Unificationist model, ANNs (and other Machine Learning models) are not explainable. It is emphasized that Erasmus et al.’s general strategy is correct. The literature on explainable Artificial Intelligence can benefit by drawing from philosophical accounts of explanation. However, philosophical accounts of explanation do not settle the problem of whether ANNs are explainable because they do not unanimously declare that explanation is invariant with regard to complexity. (shrink)

There is widespread recognition at universities that a proper understanding of science is needed for all undergraduates. Good jobs are increasingly found in fields related to Science, Technology, Engineering, and Medicine, and science now enters almost all aspects of our daily lives. For these reasons, scientific literacy and an understanding of scientific methodology are a foundational part of any undergraduate education. Recipes for Science provides an accessible introduction to the main concepts and methods of scientific reasoning. With the help of (...) an array of contemporary and historical examples, definitions, visual aids, and exercises for active learning, the textbook helps to increase students’ scientific literacy. The first part of the book covers the definitive features of science: naturalism, experimentation, modeling, and the merits and shortcomings of both activities. The second part covers the main forms of inference in science: deductive, inductive, abductive, probabilistic, statistical, and causal. The book concludes with a discussion of explanation, theorizing and theory-change, and the relationship between science and society. The textbook is designed to be adaptable to a wide variety of different kinds of courses. In any of these different uses, the book helps students better navigate our scientific, 21st-century world, and it lays the foundation for more advanced undergraduate coursework in a wide variety of liberal arts and science courses. Selling Points Helps students develop scientific literacy—an essential aspect of _any_ undergraduate education in the 21 st century, including a broad understanding of scientific reasoning, methods, and concepts Written for all beginning college students: preparing science majors for more focused work in particular science; introducing the humanities’ investigations of science; and helping non-science majors become more sophisticated consumers of scientific information Provides an abundance of both contemporary and historical examples Covers reasoning strategies and norms applicable in all fields of physical, life, and social sciences, _as well as_ strategies and norms distinctive of specific sciences Includes visual aids to clarify and illustrate ideas Provides text boxes with related topics and helpful definitions of key terms, and includes a final Glossary with all key terms Includes Exercises for Active Learning at the end of each chapter, which will ensure full student engagement and mastery of the information include earlier in the chapter Provides annotated ‘For Further Reading’ sections at the end of each chapter, guiding students to the best primary and secondary sources available Offers a Companion Website, with: For Students: direct links to many of the primary sources discussed in the text, student self-check assessments, a bank of exam questions, and ideas for extended out-of-class projects For Instructors: a password-protected Teacher’s Manual, which provides student exam questions with answers, extensive lecture notes, classroom-ready Power Point presentations, and sample syllabi Extensive Curricular Development materials, helping any instructor who needs to create a Scientific Reasoning Course, ex nihilo. (shrink)

Le déficit d’explicabilité des techniques d’apprentissage machine (AM) pose des problèmes opérationnels, juridiques et éthiques. Un des principaux objectifs de notre projet est de fournir des explications éthiques des sorties générées par une application fondée sur de l’AM, considérée comme une boîte noire. La première étape de ce projet, présentée dans cet article, consiste à montrer que la validation de ces boîtes noires diffère épistémologiquement de celle mise en place dans le cadre d’une modélisation mathématique et causale d’un phénomène physique. (...) La différence majeure est qu’une méthode d’AM ne prétend pas représenter une causalité entre les paramètres d’entrées, qui peuvent être de plus de haute dimensionnalité, et ceux de sortie. Nous montrons dans cet article l’intérêt de mettre en œuvre les distinctions épistémologiques entre les différentes fonctions épistémiques d’un modèle, d’une part, et entre la fonction épistémique et l’usage d’un modèle, d’autre part. Enfin, la dernière partie de cet article présente nos travaux en cours sur l’évaluation d’une explication, qui peut être plus persuasive qu’informative, ce qui peut ainsi causer des problèmes d’ordre éthique. (shrink)

Le déficit d’explicabilité des techniques d’apprentissage machine (AM) pose des problèmes opérationnels, juridiques et éthiques. Un des principaux objectifs de notre projet est de fournir des explications éthiques des sorties générées par une application fondée sur de l’AM, considérée comme une boîte noire. La première étape de ce projet, présentée dans cet article, consiste à montrer que la validation de ces boîtes noires diffère épistémologiquement de celle mise en place dans le cadre d’une modélisation mathéma- tique et causale d’un phénomène (...) physique. La différence majeure est qu’une méthode d’AM ne prétend pas représenter une causalité entre les paramètres d’entrées et ceux de sortie. Après avoir proposé une clarification et une adaptation des notions d’inter- prétabilité et d’explicabilité telles qu’on les rencontre dans la littérature déjà abondante sur le sujet, nous montrons dans cet article l’intérêt de mettre en œuvre les distinctions épistémologiques entre les différentes fonctions épistémiques d’un modèle, d’une part, et entre la fonction épistémique et l’usage d’un modèle, d’autre part. Enfin, la dernière partie de cet article présente nos travaux en cours sur l’évaluation d’une explication, qui peut être plus persuasive qu’informative, ce qui peut ainsi causer des problèmes d’ordre éthique. (shrink)

This paper examines the question of whether Large Language Models (LLMs) like ChatGPT possess minds, focusing specifically on whether they have a genuine folk psychology encompassing beliefs, desires, and intentions. We approach this question by investigating two key aspects: internal representations and dispositions to act. First, we survey various philosophical theories of representation, including informational, causal, structural, and teleosemantic accounts, arguing that LLMs satisfy key conditions proposed by each. We draw on recent interpretability research in machine learning to (...) support these claims. Second, we explore whether LLMs exhibit robust dispositions to perform actions, a necessary component of folk psychology. We consider two prominent philosophical traditions, interpretationism and representationalism, to assess LLM action dispositions. While we find evidence suggesting LLMs may satisfy some criteria for having a mind, particularly in game-theoretic environments, we conclude that the data remains inconclusive. Additionally, we reply to several skeptical challenges to LLM folk psychology, including issues of sensory grounding, the "stochastic parrots" argument, and concerns about memorization. Our paper has three main upshots. First, LLMs do have robust internal representations. Second, there is an open question to answer about whether LLMs have robust action dispositions. Third, existing skeptical challenges to LLM representation do not survive philosophical scrutiny. (shrink)

This book gives an account of work that I have done over a period of decades that sets out to solve two fundamental problems of philosophy: the mind-body problem and the problem of induction. Remarkably, these revolutionary contributions to philosophy turn out to have dramatic implications for a wide range of issues outside philosophy itself, most notably for the capacity of humanity to resolve current grave global problems and make progress towards a better, wiser world. A key element of the (...) proposed solution to the first problem is that physics is about only a highly specialized aspect of all that there is – the causally efficacious aspect. Once this is understood, it ceases to be a mystery that natural science says nothing about the experiential aspect of reality, the colours we perceive, the inner experiences we are aware of. That natural science is silent about the experiential aspect of reality is no reason whatsoever to hold that the experiential does not objectively exist. A key element of the proposed solution to the second problem is that physics, in persistently accepting unified theories only, thereby makes a substantial metaphysical assumption about the universe: it is such that a unified pattern of physical law runs through all phenomena. We need a new conception, and kind, of physics that acknowledges, and actively seeks to improve, metaphysical presuppositions inherent in the methods of physics. The problematic aims and methods of physics need to be improved as physics proceeds. These are the ideas that have fruitful implications, I set out to show, for a wide range of issues: for philosophy itself, for physics, for natural science more generally, for the social sciences, for education, for the academic enterprise as a whole and, most important of all, for the capacity of humanity to learn how to solve the grave global problems that menace our future, and thus make progress to a better, wiser world. It is not just science that has problematic aims; in life too our aims, whether personal, social or institutional, are all too often profoundly problematic, and in urgent need of improvement. We need a new kind of academic enterprise which helps humanity put aims-and-methods improving meta-methods into practice in personal and social life, so that we may come to do better at achieving what is of value in life, and make progress towards a saner, wiser world. This body of work of mine has met with critical acclaim. Despite that, astonishingly, it has been ignored by mainstream philosophy. In the book I discuss the recent work of over 100 philosophers on the mind-body problem and the metaphysics of science, and show that my earlier, highly relevant work on these issues is universally ignored, the quality of subsequent work suffering as a result. My hope, in publishing this book, is that my fellow philosophers will come to appreciate the intellectual value of my proposed solutions to the mind-body problem and the problem of induction, and will, as a result, join with me in attempting to convince our fellow academics that we need to bring about an intellectual/institutional revolution in academic inquiry so that it takes up its proper task of helping humanity learn how to solve problems of living, including global problems, and make progress towards as good, as wise and enlightened a world as possible. (shrink)

In domains like stock brokerage, clinical psychiatry, and long‐term political forecasting, experts generally fail to outperform novices. Empirical researchers agree on why this is: experts must receive direct or environmental learning feedback during training to develop reliable expertise, and these domains are deficient in this type of feedback. A growing number of philosophers resource this consensus view to argue that, given the absence of direct or environmental philosophical feedback, we should not give the philosophical intuitions or theories of expert (...) philosophers greater credence than those of novice philosophers. This article has three objectives. The first is to explore several overlooked issues concerning the strategy of generalizing from empirical studies of non‐philosophical expertise to the epistemic status of philosophical expertise. The second is to explain why empirical research into a causal relationship between direct learning feedback and enhanced expert performance does not provide good grounds for abandoning a default optimism about the epistemic superiority of expert philosophical theories. The third is to sketch a positive characterization of learning feedback that addresses developmental concerns made salient by the empirical literature on expert performance for specifically theory‐driven or “armchair” domains like philosophy. (shrink)

It is important for the theory of knowledge to understand the factors involved in the generation of the capacities of knowledge. In the history of modern philosophy, knowledge is generally held to originate in either one or two sources, and the debates about these sources between philosophers have concerned their existence, or legitimacy. Furthermore, some philosophers have advocated scepticism about the human capacity to understand the origins of knowledge altogether. However, the developmental aspects of knowledge have received relatively little attention (...) both by past philosophers and in current philosophical discussions. This dissertation provides a historical approach to this developmental problem of knowledge by interpreting the developmental theories of knowledge of Maine de Biran (1766–1824) and Henri Bergson (18591941) from the perspective of a theory of the ‘generative factors of knowledge.’ It first studies the philosophies of Maine de Biran and Bergson separately and then brings together and compares the metaphilosophical aims drawn from these philosophers. The dissertation’s novel analysis, provided by its theory and structure, has far-reaching consequences. From a wide point of view, it fills in considerable scholarly gaps and provides great opportunities for future research in the study of the history of philosophy. From more specific points of view, it provides its most decisive contributions in such metaphysical and epistemological topics as the nature of causality, self-generated activity, the role of effort in knowing and learning, the complementary relationship between philosophy and science, and the non-conceptual basis of knowledge. (shrink)

People’s beliefs about normality play an important role in many aspects of cognition and life (e.g., causal cognition, linguistic semantics, cooperative behavior). But how do people determine what sorts of things are normal in the first place? Past research has studied both people’s representations of statistical norms (e.g., the average) and their representations of prescriptive norms (e.g., the ideal). Four studies suggest that people’s notion of normality incorporates both of these types of norms. In particular, people’s representations of what (...) is normal were found to be influenced both by what they believed to be descriptively average and by what they believed to be prescriptively ideal. This is shown across three domains: people’s use of the word ‘‘normal” (Study 1), their use of gradable adjectives (Study 2), and their judgments of concept prototypicality (Study 3). A final study investigated the learning of normality for a novel category, showing that people actively combine statistical and prescriptive information they have learned into an undifferentiated notion of what is normal (Study 4). Taken together, these findings may help to explain how moral norms impact the acquisition of normality and, conversely, how normality impacts the acquisition of moral norms. (shrink)

According to recent social interactionist accounts in developmental psychology, a child's learning to talk about the past with others plays a key role in memory development. Most accounts of this kind are centered on the theoretical notion of autobiographical memory and assume that socio-communicative interaction with others is important, in particular, in explaining the emergence of memories that have a particular type of connection to the self. Most of these accounts also construe autobiographical memory as a species of episodic (...) memory, but its episodic character, as such, is not typically seen as falling within the remit of an explanation in social interactionist terms. I explore the idea that socio-communicative interaction centered on talk about the past might also have an important role to play, quite independently of considerations about the involvement of the self in memory, in accounting for the emergence of memories that are episodic in character, i.e., memories that involve the recollection of particular past events. In doing so, I also try to shed light on a distinctive role that talk about the past plays in socio-communicative interaction. (shrink)

Epistemologists have begun paying attention to the phenomenon of _core cognition_ from developmental psychology. Core cognition posits innate automatic cognitive modules that enable children to quickly grasp and learn certain concepts. A key element of core cognition is sometimes named _core knowledge_ because it encodes the constraints, parameters, and concepts that are required for core cognition modules to function. Until now, no successful epistemological account of it has been presented, and it is difficult to integrate into standard accounts of epistemology (...) given that it is only implicitly believed, not accessible to explicit cognitive processing, and innate. In this paper I propose an account of the epistemology of core cognition, focussing on the epistemic status of this core knowledge. I argue that, rather than being knowledge, or some ordinary justified belief, it consists of Wittgensteinian _hinge certainties_. These are the implicit presupposition that we need for our epistemology to function. I illustrate the argument with the core cognition of causality. Finally, I propose that even though core knowledge consists of unjustified hinges, we are epistemically entitled to trust them to be accurate. (shrink)

Teaching while managing instructional workload is causal to the teaching-learning process. The study determined the teachers' instructional workload management to the teachers' teaching efficacy. It was conducted in a community college in Tangub City, Misamis Occidental. The descriptive-correlational design was used in the study. There were 15 program heads and 361 students who served as the respondents selected through a stratified random sampling technique. The researcher-made Teachers' Instructional Workload Management and Teachers' Teaching Efficacy Questionnaires were used as research (...) instruments. Mean, Standard Deviation, and Pearson Product-Moment Correlation Coefficient were used as statistical tools to analyze the data gathered. Results revealed that the teachers were very good in their instructional workload management. The teachers' teaching efficacy was very high. Some aspects of the teachers' instructional workload management like the preparation of instructional materials influenced the teachers' administration of assessment tasks and classroom management. Their conduct of consultation also influenced the administration of assessment tasks. Moreover, their submission of academic reports influenced the administration of assessment tasks and school resource utilization. The workload management of teachers play a crucial role in shaping teachers' teaching efficacy. It is recommended for teachers to manage their workload proactively since it can impact their teaching efficacy. (shrink)

Turing’s much debated test has turned 70 and is still fairly controversial. His 1950 paper is seen as a complex and multilayered text, and key questions about it remain largely unanswered. Why did Turing select learning from experience as the best approach to achieve machine intelligence? Why did he spend several years working with chess playing as a task to illustrate and test for machine intelligence only to trade it out for conversational question-answering in 1950? Why did Turing refer (...) to gender imitation in a test for machine intelligence? In this article, I shall address these questions by unveiling social, historical and epistemological roots of the so-called Turing test. I will draw attention to a historical fact that has been only scarcely observed in the secondary literature thus far, namely that Turing’s 1950 test emerged out of a controversy over the cognitive capabilities of digital computers, most notably out of debates with physicist and computer pioneer Douglas Hartree, chemist and philosopher Michael Polanyi, and neurosurgeon Geoffrey Jefferson. Seen in its historical context, Turing’s 1950 paper can be understood as essentially a reply to a series of challenges posed to him by these thinkers arguing against his view that machines can think. Turing did propose gender learning and imitation as one of his various imitation tests for machine intelligence, and I argue here that this was done in response to Jefferson's suggestion that gendered behavior is causally related to the physiology of sex hormones. (shrink)

We review some of the main implications of the free-energy principle (FEP) for the study of the self-organization of living systems – and how the FEP can help us to understand (and model) biotic self-organization across the many temporal and spatial scales over which life exists. In order to maintain its integrity as a bounded system, any biological system - from single cells to complex organisms and societies - has to limit the disorder or dispersion (i.e., the long-run entropy) of (...) its constituent states. We review how this can be achieved by living systems that minimize their variational free energy. Variational free energy is an information theoretic construct, originally introduced into theoretical neuroscience and biology to explain perception, action, and learning. It has since been extended to explain the evolution, development, form, and function of entire organisms, providing a principled model of biotic self-organization and autopoiesis. It has provided insights into biological systems across spatiotemporal scales, ranging from microscales (e.g., sub- and multicellular dynamics), to intermediate scales (e.g., groups of interacting animals and culture), through to macroscale phenomena (the evolution of entire species). A crucial corollary of the FEP is that an organism just is (i.e., embodies or entails) an implicit model of its environment. As such, organisms come to embody causal relationships of their ecological niche, which, in turn, is influenced by their resulting behaviors. Crucially, free-energy minimization can be shown to be equivalent to the maximization of Bayesian model evidence. This allows us to cast natural selection in terms of Bayesian model selection, providing a robust theoretical account of how organisms come to match or accommodate the spatiotemporal complexity of their surrounding niche. In line with the theme of this volume; namely, biological complexity and self-organization, this chapter will examine a variational approach to self-organization across multiple dynamical scales. (shrink)

Stephen Yablo’s notion of proportionality, despite controversies surrounding it, has played a significant role in philosophical discussions of mental causation and of high-level causation more generally. In particular, it is invoked in James Woodward’s interventionist account of high-level causation and explanation, and is implicit in a novel approach to constructing variables for causal modeling in the machine learning literature, known as causal feature learning (CFL). In this article, we articulate an account of proportionality inspired by both (...) Yablo’s account of proportionality and the CFL account of variable construction. The resulting account has at least three merits. First, it illuminates an important feature of the notion of proportionality, when it is adapted to a probabilistic and interventionist framework. The feature is that at the center of the notion of proportionality lies the concept of “determinate intervention effects.” Second, it makes manifest a virtue of (common types of) high-level causal/explanatory statements over low-level ones, when relevant intervention effects are determinate. Third, it overcomes a limitation of the CFL framework and thereby also addresses a challenge to interventionist accounts of high-level causation. (shrink)

Sensorimotor theory claims that what you do and what you know how to do constitutes your visual experience. Central to the theory is the claim that such experience depends on a special kind of knowledge or understanding. I assess this commitment to knowledge in the light of three objections to the theory: the empirical implausibility objection, the learning/post-learning objection and the causal-constitutive objection. I argue that although the theory can respond to the first two objections, its commitment (...) to know-how ultimately renders it vulnerable to the third and arguably most serious objection. I then suggest that sensorimotor theory has two options: concede the causal-constitutive objection or challenge it. I shall argue for the latter. I will claim that a radical sensorimotor theory offers the best means of responding to this objection. (shrink)

This essay explores a problem for Nyāya epistemologists. It concerns the notion of pramā. Roughly speaking, a pramā is a conscious mental event of knowledge-acquisition, i.e., a conscious experience or thought in undergoing which an agent learns or comes to know something. Call any event of this sort a knowledge-event. The problem is this. On the one hand, many Naiyāyikas accept what I will call the Nyāya Definition of Knowledge, the view that a conscious experience or thought is a knowledge-event (...) just in case it is true and non-recollective. On the other hand, they are also committed to what I shall call Nyāya Infallibilism, the thesis that every knowledge-event is produced by causes that couldn’t have given rise to an error. These two commitments seem to conflict with each other in cases of epistemic luck, i.e., cases where an agent arrives a true judgement accidentally or as a matter of luck. While the Nyāya Definition of Knowledge seems to predict that these judgements are knowledge-events, Nyāya Infallibilism seems to entail that they aren’t. In this essay, I show that Gaṅgeśa Upādhyāya, the 14th century Naiyāyika, solves this problem by adopting what I call epistemic localism, the view that upstream causal factors play no epistemically significant role in the production of knowledge. (shrink)

Implicit stochastic models, including both ‘deep neural networks’ (dNNs) and the more recent unsupervised foundational models, cannot be explained. That is, it cannot be determined how they work, because the interactions of the millions or billions of terms that are contained in their equations cannot be captured in the form of a causal model. Because users of stochastic AI systems would like to understand how they operate in order to be able to use them safely and reliably, there has (...) emerged a new field called ‘explainable AI’ (XAI). When we examine the XAI literature, however, it becomes apparent that its protagonists have redefined the term ‘explanation’ to mean something else, namely: ‘interpretation’. Interpretations are indeed sometimes possible, but we show that they give at best only a subjective understanding of how a model works. We propose an alternative to XAI, namely certified AI (CAI), and describe how an AI can be specified, realized, and tested in order to become certified. The resulting approach combines ontologies and formal logic with statistical learning to obtain reliable AI systems which can be safely used in technical applications. (shrink)