We propose a nonmonotonic Description Logic of typicality able to account for the phenomenon of the combination of prototypical concepts. The proposed logic relies on the logic of typicality ALC + TR, whose semantics is based on the notion of rational closure, as well as on the distributed semantics of probabilistic Description Logics, and is equipped with a cognitive heuristic used by humans for concept composition. We first extend the logic of typicality ALC + TR by typicality inclusions of the (...) form p :: T(C) v D, whose intuitive meaning is that “we believe with degree p about the fact that typical Cs are Ds”. As in the distributed semantics, we define different scenarios containing only some typicality inclusions, each one having a suitable probability. We then exploit such scenarios in order to ascribe typical properties to a concept C obtained as the combination of two prototypical concepts. We also show that reasoning in the proposed Description Logic is EXPTIME-complete as for the underlying standard Description Logic ALC. (shrink)

We propose a nonmonotonic Description Logic of typicality able to account for the phenomenon of combining prototypical concepts, an open problem in the fields of AI and cognitive modelling. Our logic extends the logic of typicality ALC + TR, based on the notion of rational closure, by inclusions p :: T(C) v D (“we have probability p that typical Cs are Ds”), coming from the distributed semantics of probabilistic Description Logics. Additionally, it embeds a set of cognitive heuristics for concept (...) combination. We show that the complexity of reasoning in our logic is EXPTIME-complete as in ALC. (shrink)

Combining typical knowledge to generate novel concepts is an important creative trait of human cognition. Dealing with such ability requires, from an AI perspective, the harmonization of two conflicting requirements that are hardly accommodated in symbolic systems: the need of a syntactic compositionality (typical of logical systems) and that one concerning the exhibition of typicality effects (see Frixione and Lieto, 2012). In this work we provide a logical framework able to account for this type of human-like concept combination. We propose (...) a nonmonotonic Description Logic of typicality called TCL (Typicality-based Compositional Logic). (shrink)

In this paper we present a framework for the dynamic and automatic generation of novel knowledge obtained through a process of commonsense reasoning based on typicality-based concept combination. We exploit a recently introduced extension of a Description Logic of typicality able to combine prototypical descriptions of concepts in order to generate new prototypical concepts and deal with problem like the PET FISH (Osherson and Smith, 1981; Lieto & Pozzato, 2019). Intuitively, in the context of our application of this logic, (...) the overall pipeline of our system works as follows: given a goal expressed as a set of properties, if the knowledge base does not contain a concept able to fulfill all these properties, then our system looks for two concepts to recombine in order to extend the original knowledge based satisfy the goal. (shrink)

The goal of this paper is to review and critically discuss the philosophical aspects of probabilistic seismic hazard analysis (PSHA). Given that estimates of seismic hazard are typically riddled with uncertainty, diferent epistemic values (related to the pursuit of scientifc knowledge) compete in the selection of seismic hazard models, in a context infuenced by non-epistemic values (related to practical goals and aims) as well. We frst distinguish between the diferent types of uncertainty in PSHA. We claim that epistemic and nonepistemic (...) considerations are closely related in the selection of the appropriate estimate of seismic hazard by the experts. Finally, we argue that the division of scientifc responsibility among the experts can lead to responsibility gaps. This raises a problem for the ownership of the results (“no one’s model” problem) similar to the “problem of many hands” in the ethics of technology. We conclude with a plea for a close collaboration between philosophy and engineering. (shrink)

The aim of this paper is to clarify under what conditions a philosophical problem arises. I will describe two ways in which we might perceive a question as a problem. First, when we fnd ourselves inclined to believe in propositions that appear incompatible with each other. Second, when we fnd ourselves inclined to believe in propositions that seem incompatible with our desires. I will discuss both of these cases and articulate a didactic strategy – the Clash Strategy – which can (...) be used in order to plan and facilitate dialogic inquiry in the classroom. This strategy suggests introducing stimuli that will help students appreciate the clash that drives the philosophical question to be explored during the dialogue. In summary, this paper aims to ofer insights to all teachers interested in promoting dialogic inquiry, whether philosophical or otherwise, in education. (shrink)

This book discusses Aristotle’s notions of essence and substance as they are developed in Metaphysics ZH. I examine Aristotle's argument at length and defends an unorthodox interpretation according to which his motivation is to provide an answer against a conflation between criteria for existential priority (delivering substances as primary beings) and criteria for explanatory priority (delivering essences as primary principles).

Translation of Aristotle's Physics I-II into Portuguese, with commentaries. Tradução para o português dos livros I e II da Física de Aristóteles, com comentários.

This is an introductory handbook for some of the main themes around the notion of predication in Aristotle. It does not aim at being exhaustive, but only sketches some important lines about the subject; it contains an introductory essay, besides the translation (into Portuguese) and commentary of basic texts (such as Posterior Analytics I-22, Categories 1-5, Interpretation 1-6 etc.).

Translation of Aristotle's Posterior Analytics I into Portuguese, with a few notes, experimental glossary and introduction. The translation, which was made at 2003/4, was preliminary and its publication was intended to provide a didactic tool for courses as well as a provisional resource in research seminars. It needs some revision. I am currently working (slowly...) on the revision of the translation and a new revised one will surely appear at some point.

Translation of Aristotle's Metaphysics IV and VI, with notes. The translation is preliminary and intended as a provisional teaching tool to be also used in seminars and discussions with peers in order to reach a more elaborated version.

Translation of Aristotle's Posterior Analytics II into Portuguese, with a few notes, experimental glossary and introduction. The translation, which was made at 2002 (with a new printing in 2004), was preliminary and its publication was intended to provide a didactic tool for courses as well as a provisional resource in research seminars. It needs some revision. I am currently working (slowly...) on the revision of the translation and a new revised one will surely appear at some point.

Translation of Aristotle's Metaphysics I-III into Portuguese, with a few notes and introduction. The translation, which was made at 2007, is preliminary and its publication was intended to provide a didactic tool for courses as well as a provisional resource in research seminars. It needs some revision. I am currently working (slowly...) on the revision of the translation and a new revised one will surely appear at some point.

Translation of Aristotle’s Metaphysics IX and X (Theta & Iota) into Portuguese, with a few notes, experimental glossary and introduction. The translation, which was made at 2004, is preliminary and its publication was intended to provide a didactic tool for courses as well as a provisional resource in research seminars. It needs some revision. I am currently working (slowly...) on the revision of the translation and a new revised one will surely appear at some point.

In this paper we introduce three methods to approach philosophical problems informationally: Minimalism, the Method of Abstraction and Constructionism. Minimalism considers the specifications of the starting problems and systems that are tractable for a philosophical analysis. The Method of Abstraction describes the process of making explicit the level of abstraction at which a system is observed and investigated. Constructionism provides a series of principles that the investigation of the problem must fulfil once it has been fully characterised by the previous (...) two methods. For each method, we also provide an application: the problem of visual perception, functionalism, and the Turing Test, respectively. (shrink)

In the paper it is argued that bridging the digital divide may cause a new ethical and social dilemma. Using Hardin's Tragedy of the Commons, we show that an improper opening and enlargement of the digital environment (Infosphere) is likely to produce a Tragedy of the Digital Commons (TDC). In the course of the analysis, we explain why Adar and Huberman's previous use of Hardin's Tragedy to interpret certain recent phenomena in the Infosphere (especially peer-to-peer communication) may not be entirely (...) satisfactory. We then seek to provide an improved version of the TDC that avoids the possible shortcomings of their model. Next, we analyse some problems encountered by the application of classical ethics in the resolution of the TDC. In the conclusion, we outline the kind of work that will be required to develop an ethical approach that may bridge the digital divide but avoid the TDC. (shrink)

Translation of Aristotle's Metaphysics V.1-8 with a few notes.

In this monograph we contrast two prominent models of education, Competence-Based Education (CBE), more recent and currently dominant in most school systems of the world, and Bildung-Oriented Education (BOE), once the basis of school systems of Northern Europe. CBE interprets learning as the acquisition of clearly definable and allegedly measurable competences, and is supported by supranational organisations, such as the OECD, which approach education from the perspective of the human capital theory. BOE instead characterises learning holistically as aimed at the (...) progressive articulation of a meaningful ‘big picture’ in the student’s mind. Emerged in Northern Europe under the influence of the ideals of the Enlightenment and Neo-Humanism of individual and collective autonomy and responsibility, BOE was subsequently enriched with a hermeneutical approach to teaching and socio-political objectives of emancipation and solidarity originating within Critical Theory. We argue that, in spite of its celebrated ‘scientificity’, CBE is internally incoherent and unreliable, contributes to structural forms of oppression and injustice, fosters social pathologies, and fails to provide students with the kind of intellectual autonomy they need as both human beings and citizens of our complex post-industrial societies. We also defend BOE from objections made by critical theorists, poststructuralists and postcolonial thinkers, and argue that it is a coherent and flexible model of education that endows students with autonomy and responsibility, and can help heal social pathologies. (shrink)

This chapter discusses the first part of Aristotle's Posterior Analytics A-33, 88b30-89a10. I claim that Aristotle is not concerned with an epistemological distinction between knowledge and belief in general. He is rather making a contrast between scientific knowledge (which is equivalent to explanation by the primarily appropriate cause) and some explanatory beliefs that falls short of capturing the primarily appropriate cause.

I discuss what Aristotle means when he say that scientific demonstration must proceed from necessary principles. I argue that, for Aristotle, scientific demonstration should not be reduced to sound deduction with necessary premises. Scientific demonstration ultimately depends on the fully appropriate explanatory factor for a given explanandum. This explanatory factor is what makes the explanandum what it is. Consequently, this factor is also unique. When Aristotle says that demonstration must proceed from necessary principles, he means that each demonstration requires the (...) principle that is the necessary one for the fully appropriate explanation of its explanandum. This picture also provides a key to understand Aristotle's thesis that scientific explanation depends on essences: it is the essence of the attribute to be explained that should be stated as the fully appropriate explanatory factor. (shrink)

Tradução preliminar em pré-print. A Tradução definitiva é a de 2009.

In Posterior Analytics 71b9 12, we find Aristotle’s definition of scientific knowledge. The definiens is taken to have only two informative parts: scientific knowledge must be knowledge of the cause and its object must be necessary. However, there is also a contrast between the definiendum and a sophistic way of knowing, which is marked by the expression “kata sumbebekos”. Not much attention has been paid to this contrast. In this paper, I discuss Aristotle’s definition paying due attention to this contrast (...) and to the way it interacts with the two conditions presented in the definiens. I claim that the “necessity” condition ammounts to explanatory appropriateness of the cause. (shrink)

In this paper, I present an interpretation of Wittgenstein's remarks on the experience of wonder at the existence of the world. According to this interpretation, Wittgenstein's feeling of wonder stems from perceiving the existence of the world as an absolute miracle, that is, as a fact that is in principle beyond explanation. Based on this analysis, I will suggest that Wittgenstein's experience is akin to what has been described by other authors such as Coleridge, Pessoa, Heidegger, Scheler, Sartre, and Hadot, (...) among many others. Through a comparison between Wittgenstein and Coleridge on the experience of existence, I shall highlight some core features of this experience, chief among them the use of the notion of nothingness in clarifying what is understood in the experience, the role of intuition, and the presence of a specific pathos. As a whole, this paper aims to provide a contribution to the project of a phenomenology of the experience of existence and intends to create a bridge for a dialogue between Wittgenstein and the phenomenological tradition on the enigma of existence. (shrink)

Translation of Aristotle's Parts of Animals Book I into Portuguese, with full commentaries.

One of the most pressing challenges that occupy the Russellian panpsychist’s agenda is to come up with a way to reconcile the traditional argument from categorical properties (Seager, 2006; Alter & Nagasawa, 2015) with H. H. Mørch’s dispositionalism-friendly argument from the experience of causation (2014, 2018, 2020) — on the way to a unitary, all-encompassing case for the view. In this regard, Mørch claims that, via the commitment to the Identity theory of properties, one can consistently hold both panpsychist arguments (...) without contradiction (2020: 281) — I shall refer to such proposal as Reconciliation. In my paper, I shall argue that this is not the case. To this extent, I will first consider H. Taylor’s argument that the Identity theorists have the exact same resources as the dispositionalists (as, after careful enquiry, their views on the metaphysics of properties turn out to coincide (2018: 1438)), and thus contend that Reconciliation fails to obtain. Then, I will suggest that one can avoid the problem and reconcile the arguments by adopting a different version of the powerful qualities view, namely the Compound view — and thus advance a reformulated version of the claim, i.e. Reconciliation*. Finally, even though pursuing my proposed solution might expose Russellian panpsychism to the risk of epiphenomenalism, I shall conclude that such specific form of epiphenomenalism is a rather benign one, and thus that, via Reconciliation*, the constitution of a unitary case for panpsychism as a positive proposal (and not as a mere alternative to dualism and physicalism) can be achieved. (shrink)

We present epistemic multilateral logic, a general logical framework for reasoning involving epistemic modality. Standard bilateral systems use propositional formulae marked with signs for assertion and rejection. Epistemic multilateral logic extends standard bilateral systems with a sign for the speech act of weak assertion (Incurvati and Schlöder 2019) and an operator for epistemic modality. We prove that epistemic multilateral logic is sound and complete with respect to the modal logic S5 modulo an appropriate translation. The logical framework developed provides the (...) basis for a novel, proof-theoretic approach to the study of epistemic modality. To demonstrate the fruitfulness of the approach, we show how the framework allows us to reconcile classical logic with the contradictoriness of so-called Yalcin sentences and to distinguish between various inference patterns on the basis of the epistemic properties they preserve. (shrink)

Translation of Aristotle's Metaphysics Lambda into Portuguese.

Tradução dos livros Z e H da Metafísica de Aristóteles, com introdução e notas. Translation of Aristotle's Metaphysics Books VII and VIII into Portuguese, with Introduction and Notes.

Translation of passages of Aristotle's De Anima into Portuguese. The passages are these: I.1, I.4 (the 'Rylean passage'); II.1-6; III.1-8. The translation is preliminary.

I discuss an important feature of the notion of cause in Post. An. 1. 13, 78b13–28, which has been either neglected or misunderstood. Some have treated it as if Aristotle were introducing a false principle about explanation; others have understood the point in terms of coextensiveness of cause and effect. However, none offers a full exegesis of Aristotle's tangled argument or accounts for all of the text's peculiarities. My aim is to disentangle Aristotle's steps to show that he is arguing (...) in favour of a logical requirement for a middle term's being the appropriate cause of its explanandum. Coextensiveness between the middle term and the attribute it explains is advanced as a sine qua non condition of a middle term's being an appropriate or primary cause. This condition is not restricted either to negative causes or to middle terms in second‐figure syllogisms, but ranges over all primary causes qua primary. (shrink)

This chapter has two main sections. In the first section I briefly sketch Immanuel Kant’s moral theory as laid out in his Groundwork of the Metaphysics of Morals (1785). I explain Kant’s claim that morality must be grounded on what he calls a categorical imperative and examine his three formulations of this categorical imperative. In the second section I explain the distinction between “deontological” and “teleological” ethical theories. Kantian ethics is often presented as the paradigm example of a deontological ethical (...) theory, but I question whether Kant’s ethics should be understood as purely deontological. (shrink)

We discuss a well-known puzzle about the lexicalization of logical operators in natural language, in particular connectives and quantifiers. Of the many logically possible operators, only few appear in the lexicon of natural languages: the connectives in English, for example, are conjunction _and_, disjunction _or_, and negated disjunction _nor_; the lexical quantifiers are _all, some_ and _no_. The logically possible nand (negated conjunction) and Nall (negated universal) are not expressed by lexical entries in English, nor in any natural language. Moreover, (...) the lexicalized operators are all upward or downward monotone, an observation known as the Monotonicity Universal. We propose a logical explanation of lexical gaps and of the Monotonicity Universal, based on the dynamic behaviour of connectives and quantifiers. We define update potentials for logical operators as procedures to modify the context, under the assumption that an update by \( \phi \) depends on the logical form of \( \phi \) and on the speech act performed: assertion or rejection. We conjecture that the adequacy of update potentials determines the limits of lexicalizability for logical operators in natural language. Finally, we show that on this framework the Monotonicity Universal follows from the logical properties of the updates that correspond to each operator. (shrink)

ABSTRACTLinguistic evidence supports the claim that certain, weak rejections are less specific than assertions. On the basis of this evidence, it has been argued that rejected sentences cannot be premisses and conclusions in inferences. We give examples of inferences with weakly rejected sentences as premisses and conclusions. We then propose a logic of weak rejection which accounts for the relevant phenomena and is motivated by principles of coherence in dialogue. We give a semantics for which this logic is sound and (...) complete, show that it axiomatizes the modal logic KD45 and prove that it still derives classical logic on its asserted fragment. Finally, we defend previous logics of strong rejection as being about the linguistically preferred interpretations of weak rejections. (shrink)

I discuss in this paper the six requirements Aristotle advances at Posterior Analytics A-2, 71b20-33, for the premisses of a scientific demonstration. I argue that the six requirements give no support for an intepretation in terms of “axiomatization”. Quite on the contrary, the six requirements can be consistently understood in a very different picture, according to which the most basic feature of a scientific demonstration is to explain a given proposition by its appropriate cause.

I have two aims in this paper. First, I argue that, in Aristotle’s theory of the four causes, there is a basic and common feature by which all causes are causes: they all work in a triadic framework in which they explain why a given attribute holds of a given underlying thing. Secondly, I argue against a version of “compatibilism” according to which each kind of cause is complete in its own domain and does not compete with any other kind. (...) I claim that there are priority relations according to which some kinds of cause are subordinated to others, even if these relations do not hold in every cases. (shrink)

According to Jim Pryor’s dogmatism, when you have an experience with content p, you often have prima facie justification for believing p that doesn’t rest on your independent justification for believing any proposition. Although dogmatism has an intuitive appeal and seems to have an antisceptical bite, it has been targeted by various objections. This paper principally aims to answer the objections by Roger White according to which dogmatism is inconsistent with the Bayesian account of how evidence affects our rational credences. (...) If this were true, the rational acceptability of dogmatism would be seriously questionable. I respond that these objections don’t get off the ground because they assume that our experiences and our introspective beliefs that we have experiences have the same evidential force, whereas the dogmatist is uncommitted to this assumption. I also consider the question whether dogmatism has an antisceptical bite. I suggest that the answer turns on whether or not the Bayesian can determine the priors of hypotheses and conjectures on the grounds of their extra-empirical virtues. If the Bayesian can do so, the thesis that dogmatism has an antisceptical bite is probably false. (shrink)

The combination of panpsychism and priority monism leads to priority cosmopsychism, the view that the consciousness of individual sentient creatures is derivative of an underlying cosmic consciousness. It has been suggested that contemporary priority cosmopsychism parallels central ideas in the Advaita Vedānta tradition. The paper offers a critical evaluation of this claim. It argues that the Advaitic account of consciousness cannot be characterized as an instance of priority cosmopsychism, points out the differences between the two views, and suggests an alternative (...) positioning of the Advaitic canon within the contemporary debate on monism and panpsychism. (shrink)

Collective volume with papers by alumni and students from the Campinas Aristotle Group.

The analytical notion of ‘scientific style of reasoning’, introduced by Ian Hacking in the middle of the 1980s, has become widespread in the literature of the history and philosophy of science. However, scholars have rarely made explicit the philosophical assumptions and the research objectives underlying the notion of style: what are its philosophical roots? How does the notion of style fit into the area of research of historical epistemology? What does a comparison between Hacking’s project on styles of thinking and (...) other similar projects suggest? My aim in this paper is to answer these questions. Hacking has denied that his project of styles of thinking falls into the field of historical epistemology. I shall challenge his remark by tracing out the connections of the notion of style with historical epistemology and, more in general, with a tradition of thought born in France in the beginning of twentieth-century. (shrink)

When people want to identify the causes of an event, assign credit or blame, or learn from their mistakes, they often reflect on how things could have gone differently. In this kind of reasoning, one considers a counterfactual world in which some events are different from their real-world counterparts and considers what else would have changed. Researchers have recently proposed several probabilistic models that aim to capture how people do (or should) reason about counterfactuals. We present a new model and (...) show that it accounts better for human inferences than several alternative models. Our model builds on the work of Pearl (2000), and extends his approach in a way that accommodates backtracking inferences and that acknowledges the difference between counterfactual interventions and counterfactual observations. We present six new experiments and analyze data from four experiments carried out by Rips (2010), and the results suggest that the new model provides an accurate account of both mean human judgments and the judgments of individuals. (shrink)

I discuss three kinds of relationship between ends and means (or "things that promote ends") in the Aristotelian ethical theory, in order to clarify how moral virtues and phronesis are related both in adopting ends and in determining means for virtuous actions. Phronesis seems to be mainly charged with determining means for an end given by the moral virtues, but it must involve some conception of ends too. Phronesis cannot be parasitic on moral virtue concerning the conception of ends, for (...) otherwise it will lack intrinsic moral value. I argue that the intrinsic moral value of phronesis can be better understood through a certain kind of relation between means and ends that has not received much attention. (shrink)

I discuss the exact meaning of the thesis according to which the object of scientific knowledge is necessary. The thesis is expressed by Aristotle in the Posterior Analytics, in his definition of scientific knowledge. The traditional interpretation understands this definition as depending on two parallel and independent requirements, the causality requirement and the necessity requirement. Against this interpretation, I try to show, through the examination of several passages that refer to the definition of scientific knowledge, that the necessity requirement specifies (...) more exactly the causality requirement: what cannot be otherwise is the explanatory relation between the explanandum and the cause by which it is what it is. (shrink)

This chapter argues in favour of three interrelated points. First, I argue that demonstration (as expression of scientific knowledge) is fundamentally defined as knowledge of the appropriate cause for a given explanandum: to have scientific knowledge of the explanandum is to explain it through its fully appropriate cause. Secondly, I stress that Aristotle’s notion of cause has a “triadic” structure, which fundamentally depends on the predicative formulation (or “regimentation”) of the explanandum. Thirdly, I argue that what has motivated Aristotle to (...) choose the syllogism as a demonstrative tool was precisely the fact that syllogisms are apt to express causal relations in their triadic structure. Instead of complaining against Aristotle’s preference for the syllogisms as demonstrative tools, I argue that Aristotle was fully aware of the advantages of regimenting the explanandum into a predication. One of these advantages is to abandon a purely extensional standpoint and to highlight the importance of the notion of relevancy in explanation. (shrink)

Este livro é um 'ancestral' em pré-print do meu livro de 2006, Introdução à Teoria da Predicação em Aristóteles (ISBN 978-85-268-0716-1), publicado pela Editora da Unicamp (ver https://www.academia.edu/6912408/Introdu%C3%A7%C3%A3o_%C3%A0_teoria_da_predica%C3%A7%C3%A3o_em_Arist %C3%B3teles). O ancestral foi felizmente muito citado, mesmo depois da aparição do livro definitivo em 2006. -/- This is an ancestor (in pré-print) of my 2006 Book, 'Introdução à Teoria da Predicação em Aristóteles' (ISBN 978-85-268-0716-1), published by Editora da Unicamp (see https://www.academia.edu/6912408/Introdu%C3%A7%C3%A3o_%C3%A0_teoria_da_predica%C3%A7%C3%A3o_em_Arist %C3%B3teles). The ancestor was cited by many, even after the definitive book (...) appeared in 2006. (shrink)

Constitutivism argues that the source of the categorical force of the norms of rationality and morality lies in the constitutive features of agency. A systematic failure to be guided by these norms would amount to a loss or lack of agency. Since we cannot but be agents, we cannot but be unconditionally guided by these norms. The constitutivist strategy has been challenged by David Enoch. He argues that our participation in agency is optional and thus cannot be a source of (...) categorical demands. In this paper, I defend the viability of constitutivism by showing that agency is indeed a special ‘inescapable’ enterprise. Agency has the largest jurisdiction, and it is closed under rational assessment. This inescapability does not exempt constitutivism from raising the question whether agents have reason to be agents, but this question has to be taken up within agency. If this question is answered affirmatively, then—I argue—the criteria of practical correctness are self-ratifying in a non-circular way. This is sufficient to show the viability of the constitutivist strategy. Whether agents have conclusive reasons to be agents, however, is a matter to be addressed in the terms of particular versions of constitutivism. (shrink)

Analyzing geometric properties of high-dimensional loss functions, such as local curvature and the existence of other optima around a certain point in loss space, can help provide a better understanding of the interplay between neural network structure, implementation attributes, and learning performance. In this work, we combine concepts from high-dimensional probability and differential geometry to study how curvature properties in lower-dimensional loss representations depend on those in the original loss space. We show that saddle points in the original space are (...) rarely correctly identified as such in expected lower-dimensional representations if random projections are used. The principal curvature in the expected lower-dimensional representation is proportional to the mean curvature in the original loss space. Hence, the mean curvature in the original loss space determines if saddle points appear, on average, as either minima, maxima, or almost flat regions. We use the connection between expected curvature in random projections and mean curvature in the original space (i.e., the normalized Hessian trace) to compute Hutchinson-type trace estimates without calculating Hessian-vector products as in the original Hutchinson method. Because random projections are not suitable to correctly identify saddle information, we propose to study projections along dominant Hessian directions that are associated with the largest and smallest principal curvatures. We connect our findings to the ongoing debate on loss landscape flatness and generalizability. Finally, for different common image classifiers and a function approximator, we show and compare random and Hessian projections of loss landscapes with up to about 7×10^6 parameters. (shrink)

We present an inferentialist account of the epistemic modal operator might. Our starting point is the bilateralist programme. A bilateralist explains the operator not in terms of the speech act of rejection ; we explain the operator might in terms of weak assertion, a speech act whose existence we argue for on the basis of linguistic evidence. We show that our account of might provides a solution to certain well-known puzzles about the semantics of modal vocabulary whilst retaining classical logic. (...) This demonstrates that an inferentialist approach to meaning can be successfully extended beyond the core logical constants. (shrink)

The analytical notions of ‘thought style’, ‘paradigm’, ‘episteme’ and ‘style of reasoning’ are some of the most popular frameworks in the history and philosophy of science. Although their proponents, Ludwik Fleck, Thomas Kuhn, Michel Foucault, and Ian Hacking, are all part of the same philosophical tradition that closely connects history and philosophy, the extent to which they share similar assumptions and objectives is still under debate. In the first part of the paper, I shall argue that, despite the fact that (...) these four thinkers disagree on certain assumptions, their frameworks have the same explanatory goal – to understand how objectivity is possible. I shall present this goal as a necessary element of a common project -- that of historicising Kant's a priori. In the second part of the paper, I shall make an instrumental use of the insights of these four thinkers to form a new model for studying objectivity. I shall also propose a layered diagram that allows the differences between the frameworks to be mapped, while acknowledging their similarities. This diagram will show that the frameworks of style of reasoning and episteme illuminate conditions of possibility that lie at a deeper level than those considered by thought styles and paradigm. (shrink)

Lexical innovations (e.g., zero-derivations coined on the fly by a speaker) seem to bear semantic content. Yet, such expressions cannot bear semantic content as a function of the conventions of meaning in force in the language, since they are not part of its lexicon. This is in tension with the commonplace view that the semantic content of lexical expressions is constituted by linguistic conventions. The conventionalist has two immediate ways out of the tension. The first is to preserve the conventionalist (...) assumption and deny that lexical innovations bear semantic content. The second is to dynamicize the conventionalist assumption, that is, argue that presentations of unattested expressions trigger an augmentation of the standing semantic resources of the language and instantiate content as a result of this underlying update. Building on a comparison with the production of novel onomatopoeic words, iconic pseudowords and pro-speech gestures, the paper argues that the issue is best addressed by suspending the conventionalist assumption, and describes the metasemantic implications of the claim. (shrink)