Explainable AI (xAI) methods are important for establishing trust in using black-box models. However, recent criticism has mounted against current xAI methods that they disagree, are necessarily false, and can be manipulated, which has started to undermine the deployment of black-box models. Rudin (2019) goes so far as to say that we should stop using black-box models altogether in high-stakes cases because xAI explanations ‘must be wrong’. However, strict fidelity to the truth is historically not a desideratum in science. Idealizations (...) – the intentional distortions introduced to scientific theories and models – are commonplace in the natural sciences and are seen as a successful scientific tool. Thus, it is not falsehood qua falsehood that is the issue. In this paper, I outline the need for xAI research to engage in idealization evaluation. Drawing on the use of idealizations in the natural sciences and philosophy of science, I introduce a novel framework for evaluating whether xAI methods engage in successful idealizations or deceptive explanations (SIDEs). SIDEs evaluates whether the limitations of xAI methods, and the distortions that they introduce, can be part of a successful idealization or are indeed deceptive distortions as critics suggest. I discuss the role that existing research can play in idealization evaluation and where innovation is necessary. Through a qualitative analysis we find that leading feature importance methods and counterfactual explanations are subject to idealization failure and suggest remedies for ameliorating idealization failure. (shrink)

The claim defended in the paper is that the mechanistic account of explanation can easily embrace idealization in big-scale brain simulations, and that only causally relevant detail should be present in explanatory models. The claim is illustrated with two methodologically different models: Blue Brain, used for particular simulations of the cortical column in hybrid models, and Eliasmith’s SPAUN model that is both biologically realistic and able to explain eight different tasks. By drawing on the mechanistic theory of computational explanation, (...) I argue that large-scale simulations require that the explanandum phenomenon is identified; otherwise, the explanatory value of such explanations is difficult to establish, and testing the model empirically by comparing its behavior with the explanandum remains practically impossible. The completeness of the explanation, and hence of the explanatory value of the explanatory model, is to be assessed vis-à-vis the explanandum phenomenon, which is not to be conflated with raw observational data and may be idealized. I argue that idealizations, which include building models of a single phenomenon displayed by multi-functional mechanisms, lumping together multiple factors in a single causal variable, simplifying the causal structure of the mechanisms, and multi-model integration, are indispensable for complex systems such as brains; otherwise, the model may be as complex as the explanandum phenomenon, which would make it prone to so-called Bonini paradox. I conclude by enumerating dimensions of empirical validation of explanatory models according to new mechanism, which are given in a form of a “checklist” for a modeler. (shrink)

Conceptions of acceptability-based moral or political justification take it that authoritative acceptability constitutes, or contributes to, validity, or justification. There is no agreement as to what bar for authoritativeness such justification may employ. The paper engages the issue in relation to (i) the level of idealization that a bar for authoritativeness, ψ, imparts to a standard of acceptability-based justification, S, and (ii) the degree of discursive purchase of the discursive standing that S accords to people when it builds ψ. (...) I argue that (i) and (ii) are interdependent: high idealization values entail low discursive purchase, while high degrees of purchase require low idealization values. I then distinguish between alethic conceptions of justification that prioritize ends that commit to high idealization values, and recognitive conceptions that favour high discursive purchase. On this basis, I argue for a moderately recognitivist constraint on idealization. To render the recognitive discursive minimum available to relevant people at the site of justification, S should set ψ low enough so that it is a genuine option for actual people to reject relevant views in ways that S recognizes as authoritative. (shrink)

Kripke (Wittgenstein on rules and private language: an elementary exposition. Harvard University Press, Cambridge Mass, 1982 ) rejected a naturalistic dispositional account of meaning (hereafter semantic dispositionalism) in a skeptical argument about rule-following he attributes to Wittgenstein (Philosophical investigation. Basil Blackwell, Oxford, 1958 ). Most philosophers who oppose Kripke’s criticisms of semantic dispositionalism take the stance that the argument proves too much: semantic dispositionalism is similar to much of our respected science in some important aspects, and hence to discard the (...) former would mean to give up the latter, which is obviously wrong. In this paper, I shall discuss and reject a recent defense of Kripke by Kusch (Analysis 65(2):156–163 2005 ; Sceptical guide to meaning and rules: defending Kripke’s Wittgenstein. McGill-Queen’s, London, 2006 ). Kusch attempts to show that semantic dispositionalism differs from the sciences, and consequently, Kripke’s attack can only target semantic dispositionalism, but not the sciences. Specifically, Kusch identifies some important features of the sciences with regard to how it employs idealization and ceteris paribus clauses, and argues that the ways in which semantic dispositionalism uses them are dramatically different. I argue that, upon close examination, the two are more similar than otherwise in each of those features. (shrink)

In the philosophy of science, increasing attention has been given to the methodological novelties associated with the study of complex systems. However, there is little agreement on exactly what complex systems are. Although many characterizations of complex systems are available, they tend to be either impressionistic or overly formal. Formal definitions rely primarily on ideas from the study of computational complexity, but the relation between these formal ideas and the messy world of empirical phenomena is unclear. Here, I give a (...) definition of complex systems that draws on algorithmic complexity theory, but also provides a way of interpreting the formal idea in an empirical setting. I then use the definition to show that two canonical forms of scientific idealization are empirically inadequate when applied to complex systems. The inadequacy of these forms of idealization is shown to be the primary reason that complex systems require novel methods. Moreover, this demand for novel methods helps explain the rise of complex systems science as an autonomous discipline. (shrink)

Natalia Carrillo and Tarja Knuuttila claim that there are two traditions of thinking about idealization offering almost opposite views on their functioning and epistemic status. While one tradition views idealizations as epistemic deficiencies, the other one highlights the epistemic benefits of idealization. Both of them treat idealizations as deliberate misrepresentations, however. They then argue for an artifactual account of idealization, comparing it to the traditional accounts of idealization, and exemplifying it through the Hodgkin and Huxley model (...) of the nerve impulse. From the artifactual perspective, the epistemic benefits and deficiencies introduced by idealization frequently come in a package due to the way idealization draws together different resources in model construction. Accordingly, idealization tends to be holistic in that it is not often easily attributable to some specific parts of the model. They conclude that the artifactual approach offers a unifying view into idealization in that it is able to recover several basic philosophical insights motivating both the deficiency and epistemic benefit accounts, while being simultaneously detached from the idea of distortion by misrepresentation. (shrink)

Abstraction and idealization are the two notions that are most often discussed in the context of assumptions employed in the process of model building. These notions are also routinely used in philosophical debates such as that on the mechanistic account of explanation. Indeed, an objection to the mechanistic account has recently been formulated precisely on these grounds: mechanists cannot account for the common practice of idealizing difference-making factors in models in molecular biology. In this paper I revisit the debate (...) and I argue that the objection does not stand up to scrutiny. This is because it is riddled with a number of conceptual inconsistencies. By attempting to resolve the tensions, I also draw several general lessons regarding the difficulties of applying abstraction and idealization in scientific practice. Finally, I argue that more care is needed only when speaking of abstraction and idealization in a context in which these concepts play an important role in an argument, such as that on mechanistic explanation. (shrink)

Idealization is a central feature of human thought. We build ideal models in the sciences, our politics is guided by pictures of impossible utopias, and our thinking about the arts and moral life is guided by images of how things might have been. In all these cases we sometimes proceed with a representation of the world that we know is not true or aim at a world we accept we cannot realize. This is the world of the "as if," (...) which the philosopher Hans Vaihinger delineated at the turn of the century, in ways he traced back to Kant. In this book, I aim to explore idealization in aesthetics, ethics, and metaphysics, as well as in the philosophy of mind, of language, of religion, and of the social and natural sciences. No one could be an expert on all of these things, but sometimes in philosophy it helps to stand back and take a broader view. On the way I hope to illuminate many issues, large and small, but there is one over-arching lesson: our best chance of understanding the world must be to have a plurality of ways of thinking about it. This book is about why we need a multitude of pictures of the world. It is a gentle jeremiad against theoretical monism. (shrink)

Philosophers of science increasingly recognize the importance of idealization: the intentional introduction of distortion into scientific theories. Yet this recognition has not yielded consensus about the nature of idealization. e literature of the past thirty years contains disparate characterizations and justifications, but little evidence of convergence towards a common position.

Are societies required to pursue continual economic growth as a matter of justice? In “The Value of Economic Growth”, Julie Rose considers three arguments in favor of the need for continual economic growth, each of which revolves around the instrumental value of economic growth for promoting an important good that is needed for a just society. In each case, Rose argues that there are mechanisms other than economic growth that could allow a society to deliver the relevant goods, and thus (...) meet the demands of justice with respect to those goods. I raise a set of issues for Rose’s argument that put pressure on the normative significance of her discussion. At the heart of these issues are ones about which possibilities Rose considers and which idealizations she makes. These issues tie into more general questions about the aims and methodology of normative work in social/political philosophy. Thus, in addition to being a contribution to the debate over the politics of economic growth, this paper can be understood partly as a case study in how reflection on these kinds of issues – ones about modality, idealization, and methodology – can matter to how we evaluate specific arguments in social/political philosophy. (shrink)

Highly idealized models, such as the Hawk-Dove game, are pervasive in biological theorizing. We argue that the process and motivation that leads to the introduction of various idealizations into these models is not adequately captured by Michael Weisberg’s taxonomy of three kinds of idealization. Consequently, a fourth kind of idealization is required, which we call hypothetical pattern idealization. This kind of idealization is used to construct models that aim to be explanatory but do not aim to (...) be explanations. (shrink)

It is proposed that we use the term “approximation” for inexact description of a target system and “idealization” for another system whose properties also provide an inexact description of the target system. Since systems generated by a limiting process can often have quite unexpected, even inconsistent properties, familiar limit systems used in statistical physics can fail to provide idealizations, but are merely approximations. A dominance argument suggests that the limiting idealizations of statistical physics should be demoted to approximations.

Questions concerning the epistemological status of computer science are, in this paper, answered from the point of view of the formal verification framework. State space reduction techniques adopted to simplify computational models in model checking are analysed in terms of Aristotelian abstractions and Galilean idealizations characterizing the inquiry of empirical systems. Methodological considerations drawn here are employed to argue in favour of the scientific understanding of computer science as a discipline. Specifically, reduced models gained by Dataion are acknowledged as Aristotelian (...) abstractions that include only data which are sufficient to examine the interested executions. The present study highlights how the need to maximize incompatible properties is at the basis of both Abstraction Refinement, the process of generating a cascade of computational models to achieve a balance between simplicity and informativeness, and the Multiple Model Idealization approach in biology. Finally, fairness constraints, imposed to computational models to allow fair behaviours only, are defined as ceteris paribus conditions under which temporal formulas, formalizing software requirements, acquire the status of law-like statements about the software systems executions. (shrink)

The focus in the literature on scientific explanation has shifted in recent years towards modelbased approaches. The idea that there are simple and true laws of nature has met with objections from philosophers such as Nancy Cartwright (1983) and Paul Teller (2001), and this has made a strictly Hempelian D-N style explanation largely irrelevant to the explanatory practices of science (Hempel & Oppenheim, 1948). Much of science does not involve subsuming particular events under laws of nature. It is increasingly recognized (...) that science across the disciplines is to some degree a patchwork of scientific models, with different methods, strategies, and with varying degrees of successful prediction and explanation. And so accounts of scientific explanation have reflected this change of perspective and model-based approaches have flourished in the explanation literature (Batterman, 2002b; Bokulich, 2008; Craver, 2006; Woodward, 2003). (shrink)

The principal task of the book series Poznan Studies in the Philosophy of the Sciences and the Humanities is to promote those developments in philosophy that respect the tradition of great philosophical ideas, on the one hand, and the manner of philosophical thinking introduced by analytical philosophy, on the other. The aim is to contribute to practicing philosophy as deep as Marxism and as caring about justification as positivism.

Although Burke, Bentham, Hegel and Marx do not often agree, all criticized certain ethical theories, in particular theories of rights, for being too abstract. The complaint is still popular. It was common in Existentialist and in Wittgensteinian writing that stressed the importance of cases and examples rather than principles for the moral life; it has been prominent in recent Hegelian and Aristotelian flavoured writing, which stresses the importance of the virtues; it is reiterated in discussions that stress the distinctiveness and (...) particularity of moral vicissitudes and query the importance of ethical theory. Recent critics of abstraction are opposed not only to theories of rights, and the Kantian notions with which these are linked, but also to consequentialist ethical theories. The two ethical theories that are most influential in the English-speaking world now both stand accused of being too abstract. (shrink)

This paper aims (a) to provide characterizations of realism and instrumentalism that are philosophically interesting and applicable to economics; and (b) to defend instrumentalism against realism as a methodological stance in economics. Starting point is the observation that ‘all models are false’, which, or so I argue, is difficult to square with the realist's aim of truth, even if the latter is understood as ‘partial’ or ‘approximate’. The three cheers in favour of instrumentalism are: (1) Once we have usefulness, truth (...) is redundant. (2) There is something disturbing about causal structure. (3) It's better to do what one can than to chase rainbows. (shrink)

This paper examines the dispute between Burge and McDowell over methodology in the philosophy of perception. Burge (2005, 2011) has argued that the disjunctivism posited by naive perceptual realists is incompatible with the results of current perceptual science, while McDowell (2010, 2013) defends his disjunctivism by claiming an autonomous field of enquiry for perceptual epistemology, one that does not employ the classificatory schemes of the science. Here it is argued that the crucial point at issue in the dispute is Burge’s (...) acceptance, and McDowell’s rejection, of the ‘Cartesian idealization’ of mind as a self-contained system. Burge’s case against disjunctivism rests on the assumption of a clearly demarcated boundary between mind and world, a picture of the mind that McDowell’s philosophy reacts against. This boundary is required for scientific, causal explanations of perceptual processing because it is a simplifying assumption that helps present scientists with a clearly demarcated object of investigation. Concurring with McDowell, I conclude that philosophers need not carve up their objects of investigation in the same way. (shrink)

The Galilean explanatory style is characterized by the search for the underlying structure of phenomena, the positing of "deep" explanatory principles, and a view of the relation between theory and data, on which the search for "crucial data" is of primary importance. In this paper, I trace the dynamics of adopting the Galilean style, focusing on the science of episodic memory. I argue that memory systems, such as episodic and semantic memory, were posited as underlying competences producing the observable phenomena (...) of memory. Considered in idealized isolation from other systems, episodic memory was taken to underlay the ability of individuals to remember events from their personal past. Yet, in reality, memory systems regularly interact, standing in many-to-many relations to actual memory tasks and experiences. Upon this backdrop, I explore a puzzle about the increasing prominence of the notion of autonoetic consciousness in Tulving's theory of episodic memory. I argue that, contrary to widespread belief, the prominence is not best explained by the purported essential link between autonoetic consciousness and episodic memory. Rather, it is explained by the fact that autonoetic consciousness, hypothesized to uniquely accompany episodic retrieval, was considered a source of crucial data, predictable only from theories positing a functionally distinct episodic memory system. However, with the emergence of a new generation of theories, positing wider memory systems for remembering and imagination, the question of the relation between episodic memory and autonoetic consciousness has been reopened. This creates a pressing need for de-idealization, triggering a new search for crucial data. (shrink)

Scientific research is almost always conducted by communities of scientists of varying size and complexity. Such communities are effective, in part, because they divide their cognitive labor: not every scientist works on the same project. Philip Kitcher and Michael Strevens have pioneered efforts to understand this division of cognitive labor by proposing models of how scientists make decisions about which project to work on. For such models to be useful, they must be simple enough for us to understand their dynamics, (...) but faithful enough to reality that we can use them to analyze real scientific communities. To satisfy the first requirement, we must employ idealizations to simplify the model. The second requirement demands that these idealizations not be so extreme that we lose the ability to describe real-world phenomena. This paper investigates the status of the assumptions that Kitcher and Strevens make in their models, by first inquiring whether they are reasonable representations of reality, and then by checking the models' robustness against weakenings of these assumptions. To do this, we first argue against the reality of the assumptions, and then develop a series of agent-based simulations to systematically test their effects on model outcomes. We find that the models are not robust against weakenings of these idealizations. In fact we find that under certain conditions, this can lead to the model predicting outcomes that are qualitatively opposite of the original model outcomes. (shrink)

This paper defends the viability of de-idealization strategies in economic modeling against recent criticism. De-idealization occurs when an idealized assumption of a theoretical model is replaced with a more realistic one. Recently, some scholars have raised objections against the possibility or fruitfulness of de-idealizing economic models, suggesting that economists do not employ this kind of strategy. We present a detailed case study from the theory of industrial organization, discussing three different models, two of which can be construed as (...) de-idealized versions of the first. We conclude that recent pessimism about de-idealization in economics is largely unfounded, and that de-idealization strategies are not only possible but also widely employed in economics. (shrink)

Contemporary public reason liberalism holds that coercion must be publicly justified to an idealized constituency. Coercion must be justified to all qualified points of view, not the points of view held by actual persons. Critics, in particular Nicholas Wolterstorff and David Enoch, have complained that idealization, by idealizing away what actual people accept, risks authoritarianism and disrespect by forcing people to comply with laws they in fact reject. I argue that idealization can withstand this criticism if it satisfies (...) two conditions. First, the standards of idealization, such as the norms of rationality and information, must be grounded in the present commitments of the large majority of members of the public. Second, the standards of idealization must be moderate; that is, they cannot be used to attribute reasons to citizens that stray too far from their actual commitments. (shrink)

In 1804, when asked by the aspiring writer Clemens Brentano why she had chosen to publish her work, Karoline von Günderrode wrote that she longed “mein Leben in einer bleibenden Form auszusprechen, in einer Gestalt, die würdig sei, zu den Vortreflichsten hinzutreten, sie zu grüssen und Gemeinschaft mit ihnen zu haben.” In light of this kind of statement, it is perhaps not surprising if, despite some exceptions, much of the still relatively scant literature on Günderrode reads her works largely in (...) terms of how they articulate and manifest Günderrode’s desires, frustrations, and character, for the most part ignoring their imaginary, creative, and intellectual aspects. This interpretation of the author’s works as biography is, in Günderrode’s case, often accompanied by an interpretation of her biography, particularly her suicide, as literary work. This paper is not the first to question the conflation of Günderrode’s life, death, and writing, but it is one of only a handful that aim to address the autopoietic element of Günderrode’s work in a way that does not reduce her writings to biographical and psychological expressions, or Günderrode herself to an image – or a legend – encapsulated by her writings and her relationship to them. This paper argues that Günderrode’s own position on what the self is has been largely neglected as a result of this conflation, and that taking this position into account changes how we understand Günderrode’s articulations of self in her writings. Thus this paper has two goals: to address difficulties in articulating and even constituting oneself sincerely when one’s efforts are unrecognized, belittled, censored, and forced to conform to the conventions of a society in which one is marginalized; and to unearth a neglected and potentially rich account of the modern self. (shrink)

Robert Batterman and others have argued that certain idealizing explanations have an asymptotic form: they account for a state of affairs or behavior by showing that it emerges “in the limit”. Asymptotic idealizations are interesting in many ways, but is there anything special about them as idealizations? To understand their role in science, must we augment our philosophical theories of idealization? This paper uses simple examples of asymptotic idealization in population genetics to argue for an affirmative answer and (...) proposes a general schema for asymptotic idealization, drawing on insights from Batterman’s treatment and from John Norton’s subsequent critique. (shrink)

Although we have already entered the twenty-first century, the sexist assumptions that undermine the professional status of women philosophers have not been fully exorcised. Notwithstanding Mary Ellen Waithe's groundbreaking multi-volume A History of Women Philosophers, doubts continue to arise over whether there has been or can be such a phenomenon as a woman philosopher. The very concept remains mired in stereotypical images. Auguste Rodin's famous statue of a naked male, generally referred to as "The Thinker," the self-chosen mascot of many (...) philosophers, is still dramatically posed outside Columbia University's Philosophy Hall. In fact it originally was named "The Poet,"... (shrink)

Chap. V: Generalizations of the Rule of Correspondence ... 71 1. Generalizations and Extensions 71 2. A Certain Generalization of the Concepts of ...