Deep neural networks have become increasingly successful in applications from biology to cosmology to social science. Trained DNNs, moreover, correspond to models that ideally allow the prediction of new phenomena. Building in part on the literature on ‘eXplainable AI’, I here argue that these models are instrumental in a sense that makes them non-explanatory, and that their automated generation is opaque in a unique way. This combination implies the possibility of an unprecedented gap between discovery and explanation: When unsupervised models (...) are successfully used in exploratory contexts, scientists face a whole new challenge in forming the concepts required for understanding underlying mechanisms. (shrink)

This book explores the prospects of rivaling ontological and epistemic interpretations of quantum mechanics (QM). It concludes with a suggestion for how to interpret QM from an epistemological point of view and with a Kantian touch. It thus refines, extends, and combines existing approaches in a similar direction. -/- The author first looks at current, hotly debated ontological interpretations. These include hidden variables-approaches, Bohmian mechanics, collapse interpretations, and the many worlds interpretation. He demonstrates why none of these ontological interpretations can (...) claim to be the clear winner amongst its rivals. Next, coverage explores the possibility of interpreting QM in terms of knowledge but without the assumption of hidden variables. It examines QBism as well as Healey’s pragmatist view. The author finds both interpretations or programs wanting in certain respects. As a result, he then goes on to advance a genuine proposal as to how to interpret QM from the perspective of an internal realism in the sense of Putnam and Kant. -/- The book also includes two philosophical interludes. One details the notions of probability and realism. The other highlights the connections between the notions of locality, causality, and reality in the context of violations of Bell-type inequalities. (shrink)

Deep Neural Networks (DNNs) are becoming increasingly important as scientific tools, as they excel in various scientific applications beyond what was considered possible. Yet from a certain vantage point, they are nothing but parametrized functions fθ(x)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varvec{f}_{\varvec{\theta }}(\varvec{x})$$\end{document} of some data vector x\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varvec{x}$$\end{document}, and their ‘learning’ is nothing but an iterative, algorithmic fitting of the parameters to data. Hence, what could be special about (...) them as a scientific tool or model? I will here suggest an integrated perspective that mediates between extremes, by arguing that what makes DNNs in science special is their ability to develop functional concept proxies (FCPs): Substructures that occasionally provide them with abilities that correspond to those facilitated by concepts in human reasoning. Furthermore, I will argue that this introduces a problem that has so far barely been recognized by practitioners and philosophers alike: That DNNs may succeed on some vast and unwieldy data sets because they develop FCPs for features that are not transparent to human researchers. The resulting breach between scientific success and human understanding I call the ‘Actually Smart Hans Problem’. (shrink)

Computer simulations are nowadays often directly involved in the generation of experimental results. Given this dependency of experiments on computer simulations, that of simulations on models, and that of the models on free parameters, how do researchers establish trust in their experimental results? Using high-energy physics (HEP) as a case study, I will identify three different types of robustness that I call conceptual, methodological, and parametric robustness, and show how they can sanction this trust. However, as I will also show, (...) simulation models in HEP themselves fail to exhibit a type of robustness I call inverse parametric robustness. This combination of robustness and failures thereof is best understood by distinguishing different epistemic capacities of simulations and different senses of trust: Trusting simulations in their capacity to facilitate credible experimental results can mean accepting them as means for generating belief in these results, while this need not imply believing the models themselves in their capacity to represent an underlying reality. (shrink)

The question of where, between theory and experiment, computer simulations (CSs) locate on the methodological map is one of the central questions in the epistemology of simulation (cf. Saam Journal for General Philosophy of Science, 48, 293–309, 2017). The two extremes on the map have them either be a kind of experiment in their own right (e.g. Barberousse et al. Synthese, 169, 557–574, 2009; Morgan 2002, 2003, Journal of Economic Methodology, 12(2), 317–329, 2005; Morrison Philosophical Studies, 143, 33–57, 2009; Morrison (...) 2015; Massimi and Bhimji Studies in History and Philosophy of Science Part B: Studies in History and Philosophy of Modern Physics, 51, 71–81, 2015; Parker Synthese, 169, 483–496, 2009) or just an argument executed with the aid of a computer (e.g. Beisbart European Journal for Philosophy of Science, 2, 395–434, 2012; Beisbart and Norton International Studies in the Philosophy of Science, 26, 403–422, 2012). There exist multiple versions of the first kind of position, whereas the latter is rather unified. I will argue that, while many claims about the ‘experimental’ status of CSs seem unjustified, there is a variant of the first position that seems preferable. In particular I will argue that while CSs respect the logic of (deductively valid) arguments, they neither agree with their pragmatics nor their epistemology. I will then lay out in what sense CSs can fruitfully be seen as experiments, and what features set them apart from traditional experiments nonetheless. I conclude that they should be seen as surrogate experiments, i.e. experiments executed consciously on the wrong kind of system, but with an exploitable connection to the system of interest. Finally, I contrast my view with that of Beisbart (European Journal for Philosophy of Science, 8, 171–204, 2018), according to which CSs are surrogates for experiments, arguing that this introduces an arbitrary split between CSs and other kinds of simulations. (shrink)

Despite remarkable efforts, it remains notoriously difficult to equip quantum theory with a coherent ontology. Hence, Healey (2017, 12) has recently suggested that ‘‘quantum theory has no physical ontology and states no facts about physical objects or events’’, and Fuchs et al. (2014, 752) similarly hold that ‘‘quantum mechanics itself does not deal directly with the objective world’’. While intriguing, these positions either raise the question of how talk of ‘physical reality’ can even remain meaningful, or they must ultimately embrace (...) a hidden variables-view, in tension with their original project. I here offer a neo-Kantian alternative. In particular, I will show how constitutive elements in the sense of Reichenbach (1920) and Friedman (1999, 2001) can be identified within quantum theory, through considerations of symmetries that allow the constitution of a ‘quantum reality’, without invoking any notion of a radically mind-independent reality. The resulting conception will inherit elements from pragmatist and ‘QBist’ approaches, but also differ from them in crucial respects. Furthermore, going beyond the Friedmanian program, I will show how non-fundamental and approximate symmetries can be relevant for identifying constitutive principles. (shrink)

The workshop “Machine Learning: Prediction Without Explanation?” brought together philosophers of science and scholars from various fields who study and employ Machine Learning (ML) techniques, in order to discuss the changing face of science in the light of ML's constantly growing use. One major focus of the workshop was on the impact of ML on the concept and value of scientific explanation. One may speculate whether ML’s increased use in science exemplifies a paradigmatic turn towards mere pattern recognition and prediction (...) and away from science’s traditional aim of explanation. In contrast, certain conceptions of explanation, such as statistical explanation, could turn out to fit well with the achievements of present-day ML and concede an explanatory value to these achievements after all. It is an open question how to explain ML successes themselves, and this question was in the focus of several talks in the workshop. Based on the topics raised, we will discuss the talks’ contents in more detail as organized into (i) practitioners’ perspectives, (ii) explanations from ML, (iii) explanations of ML, (iv) societal implications, and (v) global and historical perspectives. (shrink)

Bei Kant tauchen viele antike Autoren und Theorien auf. In dieser ersten Gesamtbearbeitung zum Thema wird gegen eine langjährige Forschungsmeinung gezeigt, dass nicht so sehr Platon und Aristoteles als vielmehr der hellenistischen Philosophie die entscheidende Rolle für sein Denken zukommt. Anhand der drei Kritiken werden Konstanzen und Umbrüche seines Antikeverständnisses herausgearbeitet und in ihren Konsequenzen für die Kantdeutung beleuchtet.

In this paper, we pursue three general aims: (I) We will define a notion of fundamental opacity and ask whether it can be found in High Energy Physics (HEP), given the involvement of machine learning (ML) and computer simulations (CS) therein. (II) We identify two kinds of non-fundamental, contingent opacity associated with CS and ML in HEP respectively, and ask whether, and if so how, they may be overcome. (III) We address the question of whether any kind of opacity, contingent (...) or fundamental, is unique to ML or CS, or whether they stand in continuity to kinds of opacity associated with other scientific research. (shrink)

Tim Maudlin has claimed that EPR’s Reality Criterion is analytically true. We argue that it is not. Moreover, one may be a subjectivist about quantum probabilities without giving up on objective physical reality. Thus, would-be detractors must reject QBism and other epistemic approaches to quantum theory on other grounds.

Two powerful arguments have famously dominated the realism debate in philosophy of science: The No Miracles Argument (NMA) and the Pessimistic Meta-Induction (PMI). A standard response to the PMI is selective scientific realism (SSR), wherein only the working posits of a theory are considered worthy of doxastic commitment. Building on the recent debate over the NMA and the connections between the NMA and the PMI, I here consider a stronger inductive argument that poses a direct challenge for SSR: Because it (...) is sometimes exactly the working posits which contradict each other, i.e., that which is directly responsible for empirical success, SSR cannot deliver a general explanation of scientific success. (shrink)

Quantum Mechanics notoriously faces a measurement problem, the problem that the unitary time evolution, encoded in its dynamical equations, together with the kinematical structure of the theory generally implies the non-existence of definite measurement outcomes. There have been multiple suggestions to solve this problem, among them the so called many worlds interpretation that originated with the work of Hugh Everett III. According to it, the quantum state and time evolution fully and accurately describe nature as it is, implying that under (...) certain conditions multiple measurement outcomes that are seemingly mutually exclusive can be realized at the same time – but as different 'worlds' contained in a global, quantum mechanical structure, sometimes referred to as 'the multiverse'. The many worlds interpretation has, however, been confronted with serious difficulties over the course of its development, some of which were solved by the advent of decoherence theory. The present thesis critically investi- gates the state of play on a key remaining problem of the many worlds interpretation, the problem of the meaning and quantification of probabilities in a quantum multiverse. Recent attempts of deriving the pivotal statistical ingredient of quantum mechanics, Born’s rule, from either principles of decision theory or from quantum mechanics alone, supplemented with a few general premises about probability are analyzed and their premises are scrutinized. It will be argued that, though both approaches yield promising results, they both ultimately fail to clearly establish the validity of Born’s rule in the context of the many worlds interpretation. It is hence suggested that further research on this problem is indicated. (shrink)

Quantum mechanics notoriously faces the measurement problem, the problem that if read thoroughly, it implies the nonexistence of definite outcomes in measurement procedures. A plausible reaction to this and to related problems is to regard a system's quantum state |ψ> merely as an indication of our lack of knowledge about the system, i.e., to interpret it epistemically. However, there are radically different ways to spell out such an epistemic view of the quantum state. We here investigate recent developments in the (...) branch that introduces hidden variables λ in addition to the quantum state |ψ> and has its roots in Einstein's views. In particular, we confront purported achievements of a concrete model that has been considered to serve as evidence for an epistemic view of the envisioned kind, as well as specific no-go results and their import. It will be argued that while an epistemic account of the particular kind is not straightforwardly ruled out by the no-go results, they demonstrate that the evidential character of the model(s) discussed rests on a rather shaky foundation, and that they make some achievements widely recognized in the literature appear worthy of doubt. (shrink)

Background: Recent research has demonstrated that depressed individuals show impairments in inhibiting irrelevant emotional material, and that these impairments are linked to rumination. Cognitive inhibition, however, is not a unitary construct but consists of several components which operate at different stages of information processing. The present study was designed to assess two components of inhibition and examine their relation to depression and rumination in a sample of clinically depressed and healthy control participants. Methods: Twenty-two individuals diagnosed with a current depressive (...) episode and 27 never-disordered control participants completed an Emotional Flanker Task to assess individual differences in interference control and a modification of the Working Memory Selection Task to assess individual differences in the ability to discard no longer relevant emotional material from working memory. Participants completed self-report measures to assess depressive symptoms and rumination. Results: Clinically depressed compared to control participants showed significantly reduced interference control of irrelevant negative information. The groups, however, did not differ in their ability to discard no longer relevant negative information from working memory. In contrast, rumination was associated with difficulty removing no longer relevant negative material from working memory but not with deficits in interference control. Conclusions: Results highlight the importance of differentiating among components of inhibition to gain a better understanding of cognitive mechanisms underlying depression and rumination. (shrink)

The buck-passing account of value (BPA) is very fertile ground that has given rise to a number of interpretations and controversies. It has originally been proposed by T.M. Scanlon as an analysis of value: according to it, being good ‘is not a property that itself provides a reason to respond to a thing in certain ways. Rather, to be good or valuable is to have other properties that constitute such reasons’. Buck-passing stands in a complicated relation to the fitting-attitude analysis (...) of normativity that reaches back to the work of Brentano and Ewing. Proponents of FA-analysis hold that, as a matter of the conceptual analysis of ‘good’, something is good if it is fitting to have pro-attitudes of a certain kind towards it. The ambition of FA-analysis is to explain all practical normativity in terms of the fittingness of attitudes, which – while being itself a normative concept – is the only normative concept which grounds all others. Wlodek Rabinowicz and Toni Rønnow-Rasmussen welcomed BPA as a contemporary version of FA-analysis, and others have followed this suggestion. But whether BPA should be regarded as a version of FA-analysis is a moot point. In section I, I show that BPA is not as obviously a successor of the fitting-attitude analysis of value as some have thought. The much discussed wrong-kind-of-reasons (for short: WKR) problem afflicts buck-passing only in so far as it incorporates a version of FA analysis, or at any rate is expressed in terms of reasons for attitudes. There can be a buck-passing account of value which is not affected by the problem: one that limits the account to reasons for actions (rather than attitudes). However, insofar as BPA does inherit elements of FA analysis, it also has a WKR problem. In section II, I discuss this problem and its solution. I show that it has been misidentified in the current literature, and that – once we understand the problem correctly – its solution is likely to be unavailable to the buck-passer. Hence we should reject any account of BPA that incorporates FA analysis. That leaves us with versions which do not: versions that formulate BPA in terms of reasons for actions only, rather than reasons for attitudes. Finally, in section III, I discuss why buck-passing seemed to be appealing to begin with, and whether a version of BPA that does not incorporate FA analysis is a viable contender of the account – beyond the WKR problem. (shrink)

A deep dive into the influences of Hegelian thought on the work of revolutionary and postcolonial theorist Frantz Fanon Hegel is most often mentioned – and not without good reason – as one of the paradigmatic exponents of Eurocentrism and racism in Western philosophy. But his thought also played a crucial and formative role in the work of one of the iconic thinkers of the ‘decolonial turn’, Frantz Fanon. This would be inexplicable if it were not for the much-quoted ‘lord-bondsman’ (...) dialectic – frequently referred to as the ‘master-slave dialectic’ – described in Hegel's The Phenomenology of Spirit. Fanon takes up this dialectic negatively in contexts of violence-riven (post-)slavery and colonialism; yet in works such as Black Skin, White Masks and The Wretched of the Earth he upholds a Hegelian-inspired vision of freedom. The essays in this collection offer close readings of Hegel’s text, and of responses to it in the work of twentieth-century philosophers, that highlight the entangled history of the translations, transpositions and transformations of Hegel in the work of Fanon, and more generally in colonial, postcolonial and decolonial contexts. (shrink)

Large scale experiments at CERN’s Large Hadron Collider rely heavily on computer simulations, a fact that has recently caught philosophers’ attention. CSs obviously require appropriate modeling, and it is a common assumption among philosophers that the relevant models can be ordered into hierarchical structures. Focusing on LHC’s ATLAS experiment, we will establish three central results here: with some distinct modifications, individual components of ATLAS’ overall simulation infrastructure can be ordered into hierarchical structures. Hence, to a good degree of approximation, hierarchical (...) accounts remain valid at least as descriptive accounts of initial modeling steps. In order to perform the epistemic function Winsberg Model-based reasoning in scientific discovery. Kluwer Academic/plenum Publishers, New York, pp 255–269, 1999) assigns to models in simulation—generate knowledge through a sequence of skillful but non-deductive transformations—ATLAS’ simulation models have to be considered part of a network rather than a hierarchy, in turn making the associated simulation modeling messy rather than motley. Deriving knowledge-claims from this ‘mess’ requires two sources of justification: holistic validation :253–262, 2010; in Carrier M, Nordmann A Science in the context of application. Springer, Berlin, pp 115–130, 2011), and model coherence. As it turns out, the degree of model coherence sets HEP apart from other messy, simulation-intensive disciplines such as climate science, and the reasons for this are to be sought in the historical, empirical and theoretical foundations of the respective discipline. (shrink)

In ‘Reichenbach's cubical universe and the problem of the external world’, Elliott Sober attempts a refutation of solipsism à la Reichenbach. I here contrast Sober's line of argument with observati...

It is an assumption common to many theories of rationality that all practical reasons are based on a person's given desires. I shall call any approach to practical reasons which accepts this assumption a "Humean approach". In spite of many criticisms, the Humean approach has numerous followers who take it to be the natural and inevitable view of practical reason. I will develop an argument against the Humean view aiming to explain its appeal, as well as to expose its mistake. (...) I focus on just one argument in favour of the Humean approach, which I believe can be constructed as the background idea of many Humean accounts: the argument from motivation. (shrink)

With the author(s)’ decision to opt for Open Choice the copyright of the article changed on 26 May 2021 to ©The Author(s) 2021 and the article is forthwith distributed under a Creative Commons Attribution.

Howson famously argues that the no-miracles argument, stating that the success of science indicates the approximate truth of scientific theories, is a base rate fallacy: it neglects the possibility of an overall low rate of true scientific theories. Recently a number of authors has suggested that the corresponding probabilistic reconstruction is unjust, as it concerns only the success of one isolated theory. Dawid and Hartmann, in particular, suggest to use the frequency of success in some field of research \ to (...) infer a probability of truth for a new theory from \. I here shed doubts on the justification of this and similar moves and suggest a way to directly bound the probability of truth. As I will demonstrate, my bound can become incompatible with the assumption specific testing and Dawid and Hartmann’s estimate for success. (shrink)

Putnam coined what is now known as the no miracles argument “[t]he positive argument for realism”. In its opposition, he put an argument that by his own standards counts as negative. But are there no positive arguments against scientific realism? I believe that there is such an argument that has figured in the back of much of the realism-debate, but, to my knowledge, has nowhere been stated and defended explicitly. This is an argument from the success of quantum physics to (...) the unlikely appropriateness of scientific realism as a philosophical stance towards science. I will here state this argument and offer a detailed defence of its premises. The purpose of this is to both exhibit in detail how far the intuition that quantum physics threatens realism can be driven, in the light also of more recent developments, as well as to exhibit possible vulnerabilities, i.e., to show where potential detractors might attack. (shrink)

In this paper, I argue that a person can have a reason to do what she cannot do. In a nutshell, the argument is that a person can have derivate reasons relating to an action that she has a non-derivative reason to perform. There are clear examples of derivative reasons that a person has in cases where she cannot do what she (non-derivatively) has reason to do. She couldn’t have those derivative reasons, unless she also had the non-derivative reason to (...) do what she cannot do. I discuss a number of objections to this view, in particular two: (1) The objection that if there were reasons to do what one cannot do, many of those would be ‘crazy reasons’, and (2) the worry that if there were such reasons, then agents would have reasons to engage in futile deliberations and tryings. I develop an explanation of ‘crazy reasons’ that shows that not all reasons to do the impossible are crazy and only those that are need to be filtered out, and, regarding the second objecting, I show that the reasons for trying as well as for taking the means to doing something—instrumental reasons in a broad sense—are different from the reasons for performing the action in the first place. They are affected by impossibility, and we can explain why that is so. The view I argue for is that a person may have a reason to do what she cannot do, but she does not have a reason to try to do so or to take means to realizing the impossible. (shrink)

A recent no-go theorem (Frauchiger and Renner in Nat Commun 9(1):3711, 2018) establishes a contradiction from a specific application of quantum theory to a multi- agent setting. The proof of this theorem relies heavily on notions such as ‘knows’ or ‘is certain that’. This has stimulated an analysis of the theorem by Nurgalieva and del Rio (in: Selinger P, Chiribella G (eds) Proceedings of the 15th international conference on quantum physics and logic (QPL 2018). EPTCS 287, Open Publishing Association, Waterloo, (...) 2018), in which they claim that it shows the “[i]nadequacy of modal logic in quantum settings” (ibid.). In this paper, we will offer a significantly extended and refined reconstruction of the theorem in multi-agent modal logic. We will then show that a thorough reconstruction of the proof as given by Frauchiger and Renner requires the reflexivity of access relations (system T). However, a stronger theorem is possible that already follows in serial frames, and hence also holds in systems of doxastic logic (such as KD45). After proving this, we will discuss the general implications for different interpretations of quantum probabilities as well as several options for dealing with the result. (shrink)

In this essay I examine the debate between Jürgen Habermas and Charles Taylor on the post-secular state. I argue that, although their views on the relation of religion and politics converge in certain respects, a profound difference remains between their overall approaches. Their disagreement on the epistemic status of religious as opposed to secular moral reasons, and on the role religious arguments can play in the public sphere testify to a deeper schism. Thus what might at first seem like a (...) quarrel about details proves to be a fundamental philosophical divide on the issue of modernity. I conclude that Taylor’s model of post-secularism is more promising as an approach to the challenge posed by growing religious and cultural diversity, for, if understood as a version of “reiterative universalism,” it avoids both moral relativism and Eurocentrism. (shrink)

In this paper I investigate whether the phenomenon of quantum decoherence, the vanishing of interference and detectable entanglement on quantum systems in virtue of interactions with the environment, can be understood as the manifestation of a disposition. I will highlight the advantages of this approach as a realist interpretation of the quantum formalism, and demonstrate how such an approach can benefit from advances in the metaphysics of dispositions. I will also confront some commonalities with and differences to the many worlds (...) interpretation, and address the difficulties induced by quantum non-locality. I conclude that there are ways to deal with these issues and that the proposal hence is an avenue worth pursuing. (shrink)

In this paper, I comment on Waismann’s view of ‘motivational explanations’ as he develops it in his unfinished, posthumously published essay ‘Will and Motive’. According to a traditional view, when we act, the motive is an internal psychological state of which we can know through introspection, and it triggers or causes the action. Thus the motive causally explains an independent event which is the action. As Waismann sees it, everything here is false. The motive is (1) not an internal psychological (...) state. (2) We do not know it through introspection. (3) It does not trigger an action, and (4) there is no causal explanation in which the motive is the cause and the action the effect. I analyze Waismann’s arguments for these claims and work out his positive view. Waismann draws our attention to two features: the uncertainty and instability of motives, and the fact that they are often ascribed on the basis of the action even when the agent wasn’t aware of them prior to acting. Might this hold only for certain motives: viz. ‘Triebfedern’ (drives), in particular those that are unconscious, but not for motives in the sense of ‘Beweggrund’ (reason, purpose)? If action explanations in the light of an agent’s purposes or reasons are unaffected by Waismann’s arguments for the instability and uncertainty of motives, then perhaps the traditional view would not be false for this kind of explanation: it is not uncertain, or unstable. However, I understand Waismann as trying to show (in particular through his interpretation of Dostojevki’s Raskolnikov) that even if a reason for which a person acts or a purpose is something that the agent is aware of prior to action, we get the same kind of instability and uncertainty that occurs with unconscious motives. Here is why: agents are normally aware of quite a number of reasons for and against acting in a certain way. But for any one of those reasons, they may deny – as Raskolnikov does – that that is the reason for which they did it. It was a reason, all right, and even one that they considered, but not the one for which they acted. Thus the problems of instability and uncertainty that Waismann points out affects motives not only insofar as they are unconscious and mere drives. They affect motives in the sense of ‘Beweggrund’ as well. (shrink)