This chapter contains sections titled: Reincarnation/Rebirth Karma Causality Problem of Evil Determinism, Freedom, and Moral Responsibility Karma and Release Transfer of Merit Recent Developments Works cited.

Conscious macrostates are usually assumed to be emergent from the underlying physical microstates comprising the brain and nervous system of biological organisms. However, a major problem with this assumption is that consciousness is essentially nonmeasurable unlike all other proven emergent properties of physical systems. In an earlier paper, using a no-go theorem, it was shown that conscious states cannot be comprised of processes that are physical in nature (Reason, 2019). Combining this result with another unrelated work on causal emergence (...) in physical systems (Hoel, Albantakis and Tononi, 2013), we show in this paper that conscious macrostates are not emergent from physical systems and they also do not supervene on physical microstates. An important implication of our work is that there must be some form of violation of energy conservation in biological systems that are conscious. (shrink)

Many tasks in statistical and causal inference can be construed as problems of entailment in a suitable formal language. We ask whether those problems are more difficult, from a computational perspective, for causal probabilistic languages than for pure probabilistic (or “associational”) languages. Despite several senses in which causal reasoning is indeed more complex—both expressively and inferentially—we show that causal entailment (or satisfiability) problems can be systematically and robustly reduced to purely probabilistic problems. Thus there is (...) no jump in computational complexity. Along the way we answer several open problems concerning the complexity of well-known probability logics, in particular demonstrating the ${\exists \mathbb {R}}$ -completeness of a polynomial probability calculus, as well as a seemingly much simpler system, the logic of comparative conditional probability. (shrink)

Much has been written on the role of causal notions and causal reasoning in the so-called 'special sciences' and in common sense. But does causal reasoning also play a role in physics? Mathias Frisch argues that, contrary to what influential philosophical arguments purport to show, the answer is yes. Time-asymmetric causal structures are as integral a part of the representational toolkit of physics as a theory's dynamical equations. Frisch develops his argument partly through a (...) critique of anti-causal arguments and partly through a detailed examination of actual examples of causal notions in physics, including causal principles invoked in linear response theory and in representations of radiation phenomena. Offering a new perspective on the nature of scientific theories and causal reasoning, this book will be of interest to professional philosophers, graduate students, and anyone interested in the role of causal thinking in science. (shrink)

In psychiatry, pharmacological drugs play an important experimental role in attempts to identify the neurobiological causes of mental disorders. Besides being developed in applied contexts as potential treatments for patients with mental disorders, pharmacological drugs play a crucial role in research contexts as experimental instruments that facilitate the formulation and revision of neurobiological theories of psychopathology. This paper examines the various epistemic functions that pharmacological drugs serve in the discovery, refinement, testing, and elaboration of neurobiological theories of mental disorders. I (...) articulate this thesis with reference to the history of antipsychotic drugs and the evolution of the dopamine hypothesis of schizophrenia in the second half of the twentieth century. I argue that interventions with psychiatric patients through the medium of antipsychotic drugs provide researchers with information and evidence about the neurobiological causes of schizophrenia. This analysis highlights the importance of pharmacological drugs as research tools in the generation of psychiatric knowledge and the dynamic relationship between practical and theoretical contexts in psychiatry. (shrink)

Causal reasoning is primarily concerned with what would happen to a system under external interventions. In particular, we are often interested in predicting the probability distribution of some random variables that would result if some other variables were forced to take certain values. One prominent approach to tackling this problem is based on causal Bayesian networks, using directed acyclic graphs as causal diagrams to relate post-intervention probabilities to pre-intervention probabilities that are estimable from observational data. However, (...) such causal diagrams are seldom fully testable given observational data. In consequence, many causal discovery algorithms based on data-mining can only output an equivalence class of causal diagrams. This paper is concerned with causal reasoning given an equivalence class of causal diagrams, represented by a ancestral graph. We present two main results. The first result extends Pearl 's celebrated do-calculus to the context of ancestral graphs. In the second result, we focus on a key component of Pearl's calculus---the property of invariance under interventions, and give stronger graphical conditions for this property than those implied by the first result. The second result also improves the earlier, similar results due to Spirtes et al. (shrink)

Causal reasoning is an aspect of learning, reasoning, and decision-making that involves the cognitive ability to discover relationships between causal relata, learn and understand these causal relationships, and make use of this causal knowledge in prediction, explanation, decision-making, and reasoning in terms of counterfactuals. Can we fully automate causal reasoning? One might feel inclined, on the basis of certain groundbreaking advances in causal epistemology, to reply in the affirmative. The aim (...) of this paper is to demonstrate that one still has good skeptical grounds for resisting any conclusions in favour of the automation of causal reasoning. If by causal reasoning is meant the entirety of the process through which we discover causal relationships and make use of this knowledge in prediction, explanation, decision-making, and reasoning in terms of counterfactuals, then one relies besides on tacit knowledge, as might be constituted by or derived from the epistemic faculty virtues and abilities of the causal reasoner, the value systems and character traits of the causal reasoner, the implicit knowledge base available to the causal reasoner, and the habits that sustain our causal reasoning practices. While certain aspects of causal reasoning may be axiomatized and formalized and algorithms may be implemented to approximate causal reasoning, one has to remain skeptical about whether causal reasoning may be fully automated. This demonstration will involve an engagement with Meno’s Paradox. (shrink)

The main focus of this paper is the question as to what it is for an individual to think of her environment in terms of a concept of causation, or causal concepts, in contrast to some more primitive ways in which an individual might pick out or register what are in fact causal phenomena. I show how versions of this question arise in the context of two strands of work on causation, represented by Elizabeth Anscombe and Christopher Hitchcock, (...) respectively. I then describe a central type of reasoning that, I suggest, a subject has to be able to engage in, if we are to credit her with causal concepts. I also point out that this type of reasoning turns on the idea of a physical connection between cause and effect, as articulated in recent singularist approaches of causation. (shrink)

Background: Causal reasoning as a way to make a diagnosis seems convincing. Modern medicine depends on the search for causes of disease and it seems fair to assert that such knowledge is employed in diagnosis. Causal reasoning as it has been presented neglects to some extent the conception of multifactorial disease causes. Goal: The purpose of this paper is to analyze aspects of causation relevant for discussing causal reasoning in a diagnostic context.

A novel attempt to explain why teens and adults often struggle with scientific explanation even thought young children clearly possess impressive causal reasoning skills.

I argue that one central aspect of the epistemology of causation, the use of causes as evidence for their effects, is largely independent of the metaphysics of causation. In particular, I use the formalism of Bayesian causal graphs to factor the incremental evidential impact of a cause for its effect into a direct cause-to-effect component and a backtracking component. While the “backtracking” evidence that causes provide about earlier events often obscures things, once we our restrict attention to the cause-to-effect (...) component it is true to say promoting (inhibiting) causes raise (lower) the probabilities of their effects. This factoring assumes the same form whether causation is given an interventionist, counterfactual or probabilistic interpretation. Whether we think about causation in terms of interventions and causal graphs, counterfactuals and imaging functions, or probability raising against the background of causally homogenous partitions, if we describe the essential features of a situation correctly then the incremental evidence that a cause provides for its effect in virtue of being its cause will be the same. (shrink)

A formal theory of causal reasoning is presented that encompasses both Pearl’s approach to causality and several key formalisms of nonmonotonic reasoning in Artificial Intelligence. This theory will be derived from a single rationality principle of causal acceptance for propositions. However, this principle will also set the theory of causal reasoning apart from common representational approaches to reasoning formalisms.

Not only has the philosophical debate on causation been gaining ground in the last few decades, but it has also increasingly addressed the sciences. The biomedical sciences are among the most prominent fields that have been considered, with a number of works tackling the understanding of the notion of cause, the assessment of genuinely causal relations and the use of causal knowledge in applied contexts. Far from denying the merits of the debate on causation and the major theories (...) it comprises, this paper is meant as a stimulus for theorists of causation in the philosophy of biomedicine, with a focus on clinical matters. Without aiming at putting forward an original theory of causation and starting from the narration of two actual but paradigmatic cases at the joints between biomedical research and clinical practice, we want to point out that some pathological situations addressed by molecular medicine actually prove resistant to some of our major epistemological accounts of causal explanation. Given this scenario, which is very frequent in our hospitals, our analysis aims to provide a stimulus for the debate among theorists of causation in biomedicine interested in real cases in science in practice. We believe that this might in turn encourage some more general rethinking of the complex intertwinement of science, philosophy of science and ethics, as well as of the role of philosophy of science for clinical medicine itself. (shrink)

This paper explores some interactions between normative/ philosophical/theoretical theorizing about causation and empirical research into causal reasoning and judgment of the sort conducted by psychologists and others. I attempt to extract some general morals regarding the kinds of interactions between the empirical and the more traditionally philosophical that in my experience have been most fruitful. I also compare the experimental work on which I focus with some of the research strategies employed in experimental philosophy.

Recent work by Marchisella exposed a gap in the literature on causal non-monotonic reasoning: what is needed is an approach whose primary motivation is the formal representation of the way in which humans typically reason with cause and effect. We extend the work of Marchisella, and propose a new type of causal non-monotonic reasoning, _Plausible Causal Reasoning_, which fills the gap in the literature. We propose some new principles which help characterise Plausible Causal (...) class='Hi'>Reasoning, and suggest a family of non-monotonic logics which provides a suitable starting point for its formalisation. (shrink)

This paper argues that according to Adam Smith the workings of (anything but extremely simple) sympathetic judgment (s) presuppose and crucially depend on counterfactual causal reasoning in the sympathetic process. In particular it argues for four related claims: (i) that according to Smith that the sympathetic process depends on a type of causal reasoning that goes well beyond the kind of simulationist theory standardly attributed to him; (ii) that the Smithian imagination in the sympathetic process works (...) by way of counterfactual reasoning and that even the feelings we ought to feel as a consequence of the sympathetic process need not be actual, but counterfactual; (iii) that Smithian agents are non-trivially understood as belonging to the causal order of nature; (iv) that Smitian judgments of propriety are intrinsically judgments about the proportionality of causal relations. (shrink)

Economists reason causally. Like many other scientists, they aim at formulating justified causal claims about their object of study. This thesis contributes to our understanding of how causal reasoning proceeds in economics. By using the research on the causes of unemployment as a case study, three questions are adressed. What are the meanings of causal claims? How can a causal claim be adequately supported by evidence? How are causal beliefs affected by incoming facts? In (...) the process of answering these semantic, epistemic and dynamical questions, the richness of causal reasoning in economics is brought to the foreground. In the part on semantics, it is argued that the meaning of causal generalizations in policy-oriented economics is best captured by using the concept of ‘inferential relation’ rather than the one of ‘referential relation’. This part contributes to develop an inferentialist semantics by distinguishing among types of inferential relations. In the part on epistemology, it is argued that justification in sciences like economics often relies, and ought to rely, on evidential variety―i.e. the combination of evidence from multiple sources. An interpretation of independence among evidential sources is offered: two evidential sources are fully independent if and only if the reasons why they would be unreliable do not overlap. It is then shown that the variety-of-evidence thesis does not hold under extreme circumstances. The part on dynamics focuses on deviant-case study. A storyline is provided for what deviant-case study aims at and how it proceeds in what is characterized as an ‘eclectic science’. (shrink)

When explaining the causes of human behavior, genes are often given a special status. They are thought to relate to an intrinsic human 'essence', and essentialist biases have been shown to skew the way in which causation is assessed. Causal reasoning in general is subject to other pre-existing biases, including beliefs about normativity and morality. In this synthesis we show how factors which influence causal reasoning can be mapped to a framework of genetic essentialism, which reveals (...) both the shared and unique factors underpinning biases in causal reasoning and genetic essentialism. This comparison identifies overlooked areas of research which could provide fruitful investigation, such as whether normative assessments of behaviors influence the way that genetic causes are ascribed or endorsed. We also outline the importance of distinguishing reasoning processes regarding genetic causal influences on one's self versus others, as different cognitive processes and biases are likely to be at play. (shrink)

A theory of categorization is presented in which knowledge of causal relationships between category features is represented in terms of asymmetric and probabilistic causal mechanisms. According to causal‐model theory, objects are classified as category members to the extent they are likely to have been generated or produced by those mechanisms. The empirical results confirmed that participants rated exemplars good category members to the extent their features manifested the expectations that causal knowledge induces, such as correlations between (...) feature pairs that are directly connected by causal relationships. These expectations also included sensitivity to higher‐order feature interactions that emerge from the asymmetries inherent in causal relationships. Quantitative fits of causal‐model theory were superior to those obtained with extensions to traditional similarity‐based models that represent causal knowledge either as higher‐order relational features or “prior exemplars” stored in memory. (shrink)

How do we make causal judgments? Many studies have demonstrated that people are capable causal reasoners, achieving success on tasks from reasoning to categorization to interventions. However, less is known about the mental processes used to achieve such sophisticated judgments. We propose a new process model—the mutation sampler—that models causal judgments as based on a sample of possible states of the causal system generated using the Metropolis–Hastings sampling algorithm. Across a diverse array of tasks and (...) conditions encompassing over 1,700 participants, we found that our model provided a consistently closer fit to participant judgments than standard causal graphical models. In particular, we found that the biases introduced by mutation sampling accounted for people's consistent, predictable errors that the normative model by definition could not. Moreover, using a novel experimental methodology, we found that those biases appeared in the samples that participants explicitly judged to be representative of a causal system. We conclude by advocating sampling methods as plausible process‐level accounts of the computations specified by the causal graphical model framework and highlight opportunities for future research to identify not just what reasoners compute when drawing causal inferences, but also how they compute it. (shrink)

The present paper deals with the tools that can be used to represent causation and to reason about it and, specifically, with their diversity. It focuses on so-called “causal probabilities”—that is, probabilities of effects given one of their causes—and critically surveys a recent paper in which Joyce argues that the values of these probabilities do not depend on one’s conception of causation. I first establish a stronger independence claim: I show that the very definition of causal probabilities is (...) independent of one’s conception of causation. Second, I investigate whether causal probabilities indeed take the same values under their different possible definitions. (shrink)

Gene–environment covariance is the phenomenon whereby genetic differences bias variation in developmental environment, and is particularly problematic for assigning genetic and environmental causation in a heritability analysis. The interpretation of these cases has differed amongst biologists and philosophers, leading some to reject the utility of heritability estimates altogether. This paper examines the factors that influence causal reasoning when G–E covariance is present, leading to interpretive disagreement between scholars. It argues that the causal intuitions elicited are influenced by (...) concepts of agency and blame-worthiness, and are intimately tied with the conceptual understanding of the phenotype under investigation. By considering a phenotype-specific approach, I provide an account as to why causal ascriptions can differ depending on the interpreter. Phenotypes like intelligence, which have been the primary focus of this debate, are more likely to spark disagreement for the interpretation of G–E covariance cases because the concept and ideas about its ‘normal development’ relatively ill-defined and are a subject of debate. I contend that philosophical disagreement about causal attributions in G–E covariance cases are in essence disagreements regarding how a phenotype should be defined and understood. This moves the debate from one of an ontological flavour concerning objective causal claims, to one concerning the conceptual, normative and semantic dependencies. (shrink)

Causal pluralism can be defended not only in respect to causal concepts and methodological guidelines, but also at the finer-grained level of causal inference from a particular source of evidence for causation. An argument for this last variety of pluralism is made based on an analysis of causal inference from randomized experiments (RCTs). Here, the causal interpretation of a statistically significant association can be established via multiple paths of reasoning, each relying on different assumptions (...) and providing distinct elements of information in favour of a causal interpretation. (shrink)

According to a view widely held among philosophers of science, the notion of cause has no legitimate role to play in mature theories of physics. In this paper I investigate the role of what physicists themselves identify as causal principles in the derivation of dispersion relations. I argue that this case study constitutes a counterexample to the popular view and that causal principles can function as genuine factual constraints. 1. Introduction2. Causality and Dispersion Relations3. Norton's Skepticism4. Conclusion.

We present a general formalism of causal reasoning that encompasses both Pearl’s approach to causality and a number of key systems of nonmonotonic reasoning in artificial intelligence.

Understanding causation in social-ecological systems (SES) is indispensable for promoting sustainable outcomes. However, the study of such causal relations is challenging because they are often complex and intertwined, and their analysis involves diverse disciplines. Although there is agreement that no single research approach (RA) can comprehensively explain SES phenomena, there is a lack of ability to deal with this diversity. Underlying this diversity and the challenge of dealing with it are different causal reasonings that are rarely explicit. Awareness (...) of hidden assumptions is essential for understanding how the causal reasoning of an RA is constituted, and for promoting the integration, translation, or juxtaposition of different RAs. We identify the following elements as particularly relevant for understanding causal reasoning: methods, frameworks and theories, accounts of causation, analytical focus, and causal notions. We begin with the idea that one of these elements typically figures as an entry point to an RA. This entry point is particularly important because it generates a path dependence that orients causal reasoning. In a subsequent step, when an approach is applied, causal reasoning concretizes as a result of a particular constellation of the remaining elements. We come to these insights by studying the application of four different RAs to the same social-ecological case (the collapse of Baltic cod stocks in the 1980s). On the basis of our findings we developed a guide for the analysis of causal reasoning by raising awareness of the assumptions, key elements, and the relations between these key elements for a given RA. The guide can be used to elicit the causal reasoning of RAs, facilitate interdisciplinary collaboration, and support disclosure of ethical/political dimensions that underlie management/governance interventions that are formulated on the basis of causal findings of research studies. (shrink)

Causal cognition – how we perceive, represent and reason about causal events – are fundamental to the human mind, but it has rarely been approached in its cultural specificity. Here, we investigate this core concept among Wichi people, an indigenous group living in Chaco Forest. We focus on the Wichi, because their epistemological orientations and explanatory frameworks about ecosystem differ importantly from those documented among most Western majority-culture populations. We asked participants to reason about causes of events that (...) involve the hunhat lheley (inhabitants of the earth: humans, non-human animals, plants and spiritual beings) and other entities of their ecosystem (e.g., lagoon). We find a native ontological framework that encompasses three interacting organizing principles. This new evidence highlights ways in which native categories guide causal reasoning. Our research challenge long-held assumptions that dichotomies – nature-culture or natural-supernatural – are universal features of the human mind. (shrink)

Identifying causal relations from correlational data is a fundamental challenge in personality psychology. In most cases, random assignment is not feasible, leaving observational studies as the primary methodological tool. Here, we document several techniques from behavior genetics that attempt to demonstrate causality. Although no one method is conclusive at ruling out all possible confounds, combining techniques can triangulate on causal relations. Behavior genetic tools leverage information gained by sampling pairs of individuals with assumed genetic and environmental relatedness or (...) by measuring genetic variants in unrelated individuals. These designs can find evidence consistent with causality, while simultaneously providing strong controls against common confounds. We conclude by discussing several potential problems that may limit the utility of these techniques when applied to personality. Ultimately, genetically informative designs can aid in drawing causal conclusions from correlational studies. (shrink)

Artificial Neural Networks have reached “grandmaster” and even “super-human” performance across a variety of games, from those involving perfect information, such as Go, to those involving imperfect information, such as “Starcraft”. Such technological developments from artificial intelligence (AI) labs have ushered concomitant applications across the world of business, where an “AI” brand-tag is quickly becoming ubiquitous. A corollary of such widespread commercial deployment is that when AI gets things wrong—an autonomous vehicle crashes, a chatbot exhibits “racist” behavior, automated credit-scoring processes (...) “discriminate” on gender, etc.—there are often significant financial, legal, and brand consequences, and the incident becomes major news. As Judea Pearl sees it, the underlying reason for such mistakes is that “... all the impressive achievements of deep learning amount to just curve fitting.” The key, as Pearl suggests, is to replace “reasoning by association” with “causal reasoning” —the ability to infer causes from observed phenomena. It is a point that was echoed by Gary Marcus and Ernest Davis in a recent piece for theNew York Times: “we need to stop building computer systems that merely get better and better at detecting statistical patterns in data sets—often using an approach known as ‘Deep Learning’—and start building computer systems that from the moment of their assembly innately grasp three basic concepts: time, space, and causality.” In this paper, foregrounding what in 1949 Gilbert Ryle termed “a category mistake”, I will offer an alternative explanation for AI errors; it is not so much that AI machinery cannot “grasp” causality, but that AI machinery (qua computation) cannot understand anything at all. (shrink)

Many of our decisions pertain to causal systems. Nevertheless, only recently has it been claimed that people use causal models when making judgments, decisions and predictions, and that causal Bayes nets allow us to formally describe these inferences. Experimental research has been limited to simple, artificial problems, which are unrepresentative of the complex dynamic systems we successfully deal with in everyday life. For instance, in social interactions, we can explain the actions of other's on the fly and (...) we can generalize from limited observations to predict future actions and their consequences. Our main argument is that none of these inferences (i.e., induction, generalization, explanation, and prediction) can be achieved without causal reasoning. As a case in point we use the popular television series desperate housewives and show how causal Bayes nets are able to explain the inferences made in social contexts. Crucially, causal Bayes nets also allow us to understand why we can infer so much from so little when making sense of a protagonist's behavior. (shrink)

We examine the vexed question of connections between logical and probabilistic reasoning. The reason for making such a connection are examined. We give an account of recent work which uses loglinear models to make the connection. We conclude with an analysis of various existing approaches combining logic and probability.

This is a commentary on Mathias Frisch's book Causal Reasoning in Physics (Cambridge 2014). This commentary was presented at the 2016 Pacific Division Meeting of the American Philosophical Association in a session sponsored by the Society for the Metaphysics of Science.

The Southern Journal of Philosophy, EarlyView.

Causal Bayes nets have been developed in philosophy, statistics, and computer sciences to provide a formalism to represent causal structures, to induce causal structure from data and to derive predictions. Causal Bayes nets have been used as psychological theories in at least two ways. They were used as rational, computational models of causal reasoning and they were used as formal models of mental causal models. A crucial assumption made by them is the Markov (...) condition, which informally states that variables are independent of other variables that are not their direct or indirect effects conditional on their immediate causes. Whether people’s inferences conform to the causal Markov and the faithfulness condition has recently been investigated empirically. A review of respective research indicates that inferences frequently violate these conditions. This finding challenges some uses of causal Bayes nets in psychology. They entail that causal Bayes nets may not be appropriate to derive predictions for causal model theories of causal reasoning. They also question whether causal Bayes nets as a rational model are empirically descriptive. They do not challenge, however, causal Bayes nets as normative models and their usage as formal models of causal reasoning. (shrink)

The present studies investigate how the intentions of third parties influence judgments of moral responsibility for other agents who commit immoral acts. Using cases in which an agent acts under some situational constraint brought about by a third party, we ask whether the agent is blamed less for the immoral act when the third party intended for that act to occur. Study 1 demonstrates that third-party intentions do influence judgments of blame. Study 2 finds that third-party intentions only influence moral (...) judgments when the agent's actions precisely match the third party's intention. Study 3 shows that this effect arises from changes in participants' causal perception that the third party was controlling the agent. Studies 4 and 5, respectively, show that the effect cannot be explained by changes in the distribution of blame or perceived differences in situational constraint faced by the agent. (shrink)

This paper aims to bridge philosophical and psychological research on causation, counterfactual thought, and the problem of backtracking. Counterfactual approaches to causation such as that by Lewis have ruled out backtracking, while on prominent models of causal inference interventionist counterfactuals do not backtrack. However, on various formal models, certain backtracking counterfactuals end up being true, and psychological evidence shows that people do sometimes backtrack when answering counterfactual questions in causal contexts. On the basis of psychological research, I argue (...) that while ordinarily both kinds of counterfactuals may be employed, non-backtracking counterfactuals are more easily used in causal inference because they are consistent with temporal order information embedded in the mental simulation heuristic, and they match reasoners’ experience of causation. While this approach is incompatible with the ambitions of counterfactual theories that seek to establish the non-backtracking interpretation as the only legitimate one, it can provide support for perspectival views on causation and open further inquiry on the functions of causal and counterfactual thought in the context of causal models. (shrink)

Many animal species use tools, but human technical engagement is more complex. We argue that there is coevolution between technical engagement and advanced forms of causal cognition in the human lineage. As an analytic tool, we present a classification of different forms of causal thinking. Human causal thinking has become detached from space and time, so that instead of just reacting to perceptual input, our minds can simulate actions and forces and their causal consequences. Our main (...) thesis is that, unlike the situation for other primate species, an increasing emphasis on technical engagement made some hominins capable of reasoning about the forces involved in causal processes. This thesis is supported in three ways: We compare the casual thinking about forces of hominins with that of other primates. We analyze the causal thinking required for Stone Age hunting technologies such as throwing spears, bow hunting and the use of poisoned arrows, arguing that they may serve as examples of the expansion of casual cognition about forces. We present neurophysiological results that indicate the facilitation of advanced causal thinking. (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)