Volume 13 Number 10October 2019

thereasonerorgISSN 1757-0522

Contents

Editorial 63

Features 63

Dissemination Corner 63

News 65

Whatrsquos Hot in 66

Events 67

Courses and Programmes 67

Jobs and Studentships 68

Editorial

Dear Reasoners

welcome to the October 2019 is-sue of The Reasoner In it yoursquollread about the progress made bythe Logic of Conceivability projectin Amsterdam Yoursquoll also realisewhat you missed by not participat-ing at the ldquoBayes by the seardquo Sum-mer School and Workshop Andas usual yoursquoll read whatrsquos hot ina number of areas This monthwith a new column on Whatrsquos hotin Science Policy And owing to

the great work of our news edi-tor yoursquoll also have a chance to see reasoning-related coursesevents and openings

We hope yoursquoll enjoy the issue and if you do why not con-tribute to future ones

Hykel HosniUniversity of Milan

Features

Dissemination Corner

The Logic of ConceivabilityThe Logic of Conceivability (LoC) project studies the logic ofpropositional intentional states (believing that Obama is tallimagining that there will be a disorderly Brexit knowing thatone is not a brain in a vat) given that one Xs ie thinks (be-lieves imagines knows etc) that ϕ what further ψrsquos must oneX as a matter of logic Articles in The Reasoner by LoC re-searchers Peter and Aybuke have highlighted how theories oftopic or subject matter will be important to answer that ques-tion LoC is working towards a general modal framework inthis ballpark the theory of Topic-Sensitive Intentional Modals(TSIMs) Let me unpack

Since Hintikka we usually interpret X as a quantifier overpossible worlds restricted from the standpoint of a given world(and agent) by an accessibility relation R lsquoXϕrsquo (lsquoone Xs thatϕrsquo) is true at w just in case ϕ is true at a bunch of worlds ac-cessible from w via R Then one Xs all classical logical con-sequences of what one Xs This gives very idealized agents(the problem of lsquological omnisciencersquo) But the debate on epis-temic closure shows that itrsquos controversial whether X should beunrestrictedly closed even for logically astute computationallyunbounded agents such closure seem to leave us with littleknowledge in the face eg of Cartesian skepticism Addition-ally the standard Hintikkan framework is monotonic it doesnrsquot

63

straightforwardly model the plausible phenomenon wherebymore information can reduce knowledge or make us lose be-liefs

One can easily destroy any closure property eg by re-sorting to a purely syntactic approach whereby onersquos bodyof informationknowledgebeliefs is represented as a set offormulas with no logical closure properties Howeveronce we filter out the computational limitations of agentssome forms of closure seem intuitive even for anarchicmental states like imagination for instance ConjunctionCommutation (try to imagine that Obama is tall and thinwithout imagining that Obama isthin and tall) and ConjunctionElimination (try to imagine thatObama is tall and thin withoutimagining that hersquos tall) Oth-ers seem disputable although theyinvolve elementary logical infer-ences (Disjunction Introductionwhen you imagine that Obama is tall do you thereby imag-ine that either Obama is tall or therersquos water on Kepler-452b)and even for more disciplined states like knowledge (You knowyou have hands and therersquos no way you can have hands whilebeing a bodiless brain in a vat are you thereby in a position toknow yoursquore not a brain in a vat)

The three main ways in which TSIM theory differs from theHintikkan mainsteam are

(1) The Hintikkan operators are one-place modals TheTSIMs are two-place modals things of the form lsquoXϕψrsquoto be generically read as lsquoGiven ϕ the agent Xs that ψrsquo(One can recover some paradigmatically unary-soundingmodals such as lsquoone believes that ψrsquo by trivalizing the ϕ inXϕψ)

(2) The Xϕψrsquos are variably strict modals Variability repre-sents the (contextual) selection of information importedinto the Xed content on the basis of φ The operators turnout to be non-monotonic one may X that ψ given ϕ notX that ψ given ϕ and χ Doxasticepistemic logic in TSIMclothing becomes a kind of non-monotonic conditionallogic

(3) The Xϕψrsquos encompass a topicality or aboutness constraintcapturing their intentional features their being about cer-tain contents or topics towards which the mind is directed(and most importantly not about other contents to whichitrsquos not directed)

Work on aboutness has been burgeoning thanks among oth-ers to Humberstone Fine Yablo Aboutness is lsquothe relationthat meaningful items bear to whatever it is that they are on orof or that they address or concernrsquo (Yablo Aboutness Prince-ton UP 2014 p 1) This is their subject matter or topic Invarious forms of subject matter semantics necessarily equiva-lent sentences ϕ and ψ can differ in their propositional contentwhen they are about different things The content of a sentence(in context) is not specified (just) by the set of worlds in whichit is true but (also) by what itrsquos about Subject matter semanticsis thus in general hyperintensional lsquo2 + 2 = 4rsquo and lsquoEither StAndrews is in Scotland or notrsquo differ in content in spite of be-ing true at the same worlds for they say different things onlyone is about numbers

Mental states can bear aboutness too A core twofold insightbehind TSIM theory is that

(i) The aboutness of an intentional state Xing that ϕ is inline with that of the proposition P which makes for thecontent of ϕ

(ii) P should not be understood just as a set of possible worldsbut also in terms of what ϕ is about its topic or subjectmatter

In the basic semantics a generic TSIM lsquoXϕψrsquo gets a two-component meaning it is true at world w just in case

(1) (Truth-conditional component) ψ is true at a set ofworlds selected by a set-selection function paired to ϕ fϕThink of fϕ(w) as outputting the set of worlds which arein some sense cognitively accessible for the agent locatedat w given input ϕ (specific readings depend on imposingconditions on the fϕrsquos ndash more on this later)

(2) (Topicality component) ψ is fully on-topic with respect toϕ That is what (the proposition expressed by) ψ is aboutis included in what (ditto) ϕ is about

(1) makes of the TSIMs non-monotonic variably strictmodals (2) makes them topic-sensitive Topic-inclusion iscaptured via a simple mereology of topics determining the be-haviour of the logical vocabulary The Boolean operators aretopic-transparent they add no subject matter of their own Thetopic of notϕ it the same as that of ϕ (lsquoObama is not tallrsquo isexactly about what lsquoObama is tallrsquo is about ndash say Obamarsquosheight) conjunction and disjunction merge topics (lsquoObamais tall and handsomersquo and lsquoObama is tall or handsomersquo areboth about the same topic the height and looks of Obamarsquos)This seems right for the corresponding intentional states toowhen you think that Obama is not tall you are thinking aboutObamarsquos height and thatrsquos exactly what you are thinking aboutwhen you think that Obama is tall When you think that Obamais tall and handsome you think about Obamarsquos height andlooks so in particular you do think about Obamarsquos heightthatrsquos just part of what you are thinking about But when youthink that Obama is tall you donrsquot seem to thereby be automat-ically thinking that hersquos tall or handsome for you may not bethinking about Obamarsquos looks at all

This setting delivers a number of closure properties for Xeg where is logical consequence (defined the usual way astruth preservation at all worlds of all models)

(CE) Conjunction Elimination Xϕ(ψ and χ) Xϕψ

(CC) Conjunction Commutation Xϕ(ψ and χ) Xϕ(χ and ψ)

(CI) Conjunction Introduction Xϕψ Xϕχ Xϕ(ψ and χ)

Once one factors out forms of non-omniscience due only tolimitations in peoplersquos computing capacities these seem plau-sible for a number of states (given some ϕ) one cannot knowthat Lisa is rich and happy without knowing that shersquos rich orimagine that John is tall and thin without imagining that he isthin and tall or believe that Mary is funny and that Mary ishappy without believing that shersquos funny and happy

This setting also gives a number of invalidities eg whererArr is a strict conditional ie lsquoψrArr χrsquo says that therersquos no worldwhere ψ is true but χ is false

64

(CSC) (Failure of) Closure under Strict Conditional Xϕψ ψ rArrχ 2 Xϕχ

This also seems right for various states ndash take a salient oneknowledge Given empirical evidence ϕ you are in a positionto know that ψ you have hands Therersquos no way you can be ahandless brain in a vat if you have hands (ψrArr χ) But that em-pirical evidence may not put you in a position to know yoursquoreno brain in a vat What makes CSC fail is that the implicationfrom φ to ψ albeit necessary can take you off-topic your ev-idence does not address the topic of far-fetched skeptical sce-narios

By imposing conditions on the fϕrsquos one gets more specificTSIM operators expressing eg knowability relative to infor-mation mental simulation and conditional belief and validat-ing specific inferences More sophisticated TSIM-frameworksto be delivered soon by LoC so stay tuned

Franz BertoUniversity of St Andrews and University of Amsterdam

News

Bayes By the Sea Summer School and Confer-ence 25th August to 1st September

The second Bayes By the Sea event took place this summer atUnivpm Ancona Italy It was funded by the European Re-search Council (ERC) The event combined a summer schoolwith a conference Our aims were (1) to advance the interdisci-plinary study of Bayesianism and related topics (2) to aid theinvestigation of strategic behaviour in science To our delightthese goals were fully achieved

In the summer school component the subjects taught wereprobability theory (Philip Dawid and Serena Doria) epistemol-ogy (William Peden) philosophy of science (Stephan Hartmannand Jan Sprenger) and statistics (Teddy Seidenfeld and Mommevon Sydow) all from a Bayesian perspective There were alsoclasses on epistemic game theory (AndrEs Perea and MantasRadzvilas) The summer school consisted of lectures tutorialsexercisesgroup work and social events

The opening lecture was by Stephan Hartmann (LMU) Itoutlined some cutting-edge Bayesian philosophy Hartmannbegan by discussing Bayesianismıs proven potential as a theoryof reasoning He noted some contemporary challenges such ashow to model discoveries of causal relations or how to incorpo-rate the learning of conditionals into Bayesian updating Hart-mann described the ıdistance-basedı approach whereby onetries to minimise the difference between the prior probabilitiesand the posterior probabilities according to some measure ofdistance as a supplement to standard conditionalisation in suchcircumstances

In addition to his three lectures within the summer schoolAndrEs Perea (Maastricht University) gave a lecture in the con-ference The first talk discussed the implications of commonbelief in rationality for static games with unawareness Com-mon belief in rationality occurs when all players in a gamebelieve that every other player is rational Static games withunawareness are those in which some of the choices made byother players are hidden from a given player Perea formulateda model for such games and a formal procedure for identifyingdominant strategies in them

Teddy Seidenfeld (Carnegie Mellon University) started hislecture by noting that the requirement of finite additivity is of-ten considered to be a weakness of Bruno De Finettiıs theoryof probability In contrast Seidenfeld considered some of theadvantages of this requirement He showed how assuming fi-nite additivity enables the proof of some powerful theoremsin decision theory In short as Seidenfeld put it he took theılemonı of De Finettiıs finite additivity axiom and used it tomake ılemonadeı

The short talks brought together researchers from all overthe world and across a variety of disciplines In decision the-ory Jimin Kwon (UCSD) examined cautious decision-makingand risk-weighted expected utility theory with imprecise prob-abilities Stefano Bonzio (Univpm) offered an algebraic andgeometrical characterization of De Finettiıs celebrated theoremregarding coherent gambling Serena Doria (UniCH) investi-gated the applications of Hausdorff outer measures for definingparts of an interval-valued imprecise credence function

In formal epistemology Brett Topey (University of Salzburg)argued the ıplanning frameworkı approach suggested byphilosophers like Miriam Schoenfield will not do the work theyintend it to do Momme von Sydow (LMU) applied second-order probabilities to the challenging task of modelling be-liefs via probabilities Silvia Milano (University of Oxford)discussed updating by a rule called ıur-prior conditionaliza-tionı Miriam Bowen (University of Leeds) developed an an-swer to the Probabilistic Liar Paradox using suspended judge-ment a type of imprecise belief Richard Lohse (Universityof Konstanz) criticised Richard Pettigrewrsquos accuracy argumentfor probabilism Tamaz Tokhadze (University of Sussex) ar-gued that Timothy Williamsonıs E=K thesis is either mistakenor commits us to radical scepticism about induction WilliamPeden (UnivpmDurham) proposed a solution to the Paradoxof the Ravens via distinguishing ıconfirmation simpliciterıand ıpredictive confirmationı Barbara Osimani (Univpm) ex-plained an approach to evidence in terms of strategic signallingand applied it to the weighting of (1) evidence from a varietyof sources versus (2) otherwise comparable evidence from thesame source

There were also talks in social epistemology economics andgame theory Liam Kofi Bright (LSE) argued that it is possi-ble for a veritist (who believes that only acquiring trueavoidfalse beliefs matters for evaluating epistemic practices) to pro-hibit fraud universally even though fraud might sometimesbe conducive towards true beliefs Michele Crescenzi (Uni-versity of Helsinki) expanded models of rational consensusby relaxing the standard assumption that the state space ofagreementdisagreement is either a probability space or finiteGiacomo Sillari (Luiss Guido Carli University) investigatedhow agents can successfully coordinate when there are mul-tiple ways to coordinate action Oliver Braganza (Universityof Bonn) examined the economics of proxy measures for out-comes across a wide range of domains Pavel Janda (GdaskUniversity) explored rational strategies for game players withimperfect recall Mantas Radzvilas (LMU) used a sender-receiver game-theoretic framework to inquire into optimal ly-ing and how the incentives for lying can be modified in areassuch as pharmaceutical regulation Nicola Matteucci (Univpm)discussed Italian gambling policy in relation to regulatory cap-ture

Several common themes emerged across the talks Manyresearchers were interested in accuracy-based arguments for

65

Bayesianism Additionally the work of Italyıs own De Finetticontinued to stimulate research the formal investigation of is-sues raised by his work is still a reliable source of fresh ideas

We eagerly look forward to the third Bayes By the Sea in2020 For more information on our past events see httpwwwbayesbytheseacom This news report was sup-ported by the ERC on the project Philosophy of Pharmacol-ogy Safety Statistical standards and Evidence Amalgamation(Grant Agreement ID 639276) For more information on thisproject see httpphilpharmblogwordpresscom

William PedenPhilosophy UnivpmDurham

Calls for PapersImprecise Probabilities Logic and Rationality special issueof International Journal of Approximate Reasoning deadline 1OctoberNancy Cartwrightrsquos Philosophy of Science special issue ofTheoria deadline 1 NovemberIdealization Representation Explanation Across the Sci-ences special issue of Studies in History and Philosophy ofScience Part A deadline 15 January

Whatrsquos Hot in

Medieval Reasoning[ continuing ]

Was Latin back then more egali-tarian than English is today Mostpeople could not even read or writein their own national language andeducation (especially of women)was often close to nonexistentwhich overall by today standardsdoesnrsquot seem very egalitarian atall However on the one handpeople could have some degree ofaccess to education across social classes often by joining a re-ligious order ndash this was especially true for women (Get theeto a nunnery go) On the other hand those who did learnhow to read and write often managed to learn Latin too some-times very well at a very young age For example WilliamFitztephen (writing around 1170) describes a group of twelveto fourteen years old boys in a London churchyard disputingsome in demonstrative rhetoric others in dialectic Some rsquohur-tle enthymemesrsquo others with greater skill employ perfect syl-logisms Boys of different schools strive against one anotherin verse or contend concerning the principles of grammar orthe rules concerning past and future There are others who em-ploy the old art of crossroads in epigrams rhymes and metreWas Medieval Latin philosophy as insular as contemporary An-glophone philosophy In some sense Medieval philosophy inLatin sometimes shared a few of the shortcomings of Anglo-phone philosophy ndash eg making general claims about the na-ture or deep structures of Language itself despite those claimsbeing dependent of very specific languages and hardly gener-alisables As much as English dominates the academic scenenowadays during the Middle Ages in order to join the philo-sophical community one had to use Latin ndash and not doing so

would immediately mark someone as an amateur Moreoverfor a very long time the average Latin Medieval author wasalmost as ignorant of other languages as the average native An-glophone philosopher is only a few were fluent Arabic or He-brew and even fewer knew Greek But there are at least twomajor differences in the way Medieval Latin philosophy andcontemporary Anglophone philosophy relate to other linguistictraditions First at some point in the central to later MiddleAges the Latins produced an enormous wave of translationsof ancient classics and ldquomodernrdquo Arabic philosophy first fromArabic and then from Ancient Greek as well Second as far aswe know at the moment there was not an enormous and dis-proportionate influence of Latin philosophy on other linguisticphilosophical traditions quite the contrary Overall then LatinMedieval philosophy was neither as insular as Anglophone phi-losophy nor as ldquoimperialisticrdquo so to speak or unbalanced infavour of a group of speakers Its language was more artifi-cial and highly regimented which constituted a loss in styleand elegance but it was also a gain in precision While Iam not remotely suggesting to go back to doing philosophy inzombie Latin it seems to be the case that the undead linguafranca of our predecessors worked quite a bit better for themthan English is working for us Besidesquidquid latine dictumsit altum videtur And who can deny that

Graziana CiolaDurham University

Science Policy

The supervision of PhD research is a special form of educationthat is focused on the individual work with an expert It is en-visioned as intensive and beneficial for the student It has beenargued that while the traditional supervision system involved astrict authoritarian relationship these strings nowadays loosenTorralba J M (2019 ldquo10 ingredients for a successful supervi-sorPhD student relationshiprdquo Elsevier Connect) Still is thenew system radically different from the traditional one

The relationship student-supervisor is both fragile and inten-sive It is governed by two types of dependencies The first oneis the epistemic dependence that is mirrored in the knowledgethat the supervisor has and the student has not yet acquiredThe other dependence is the existential dependence of the stu-dent both with respect to the current job contract and futurejob opportunities Systematic solutions regulating the exclu-sive teaching relationship are lacking and it seems that every-thing that remains is the trust that two adults will find mutualrespect However there is no evidence that in the situation of asemi-regulated power position we should expect that this powerwill never be abused Frequently in communication with early-career researchers one can hear complains that the demands oftheir supervisors are too strict extensive and sometimes crossthe professional borders

These experiences many of us had We can learn aboutthem from friends during conference breaks and even overhearfrom the conversations of complete strangers During quali-tative interviews with early-career biologist working abroadour group in Belgrade learned that an additional problematicdimension comes from hiring people from different countriesEarly-career researchers coming from poorer countries oftenfeel stronger pressure to preserve their job and the supervisoris perceived as the only one who can provide the desired possi-

66

bility of remaining in a foreign country and receiving a highersalary This can increase their existential dependency on thesupervisor who is often aware of it Female participants evenreported pressure coming from the supervisor on their familyplanning ldquoWe knew it you can do everything just do not staypregnantrdquo reported a participant

Early-career scholars are not a commodity They are tal-ented educated and highly motivated people determined toinvest substantial time into their research and often teachingTheir work is hard and demanding Their role is dual bothto acquire and disseminate the knowledge Many bachelor stu-dents gain their knowledge from PhD students Moreover eventhe traditional supervision is ideally bi-directional both the su-pervisor and the student should learn from each other

Early-career researchers should be treated with respect andcare The existential dependency on the supervisor needs to beremoved from the mentoring system PhD students should begiven job contracts for the whole duration of their studies de-cent salaries and should not be dependent on a single personThe myth that the supervisor-student relationship is forever andthat it marks the whole professional life of a student needs to beactively dismissed This is a strictly professional relationshipand it is not forever In the current circumstance already mak-ing students to feel free to change their supervisors researchtopics and institutions would be a huge improvement

The other potentially destructive pattern is linking profes-sorrsquos success to the success of her students This reasoningmistake can lead some supervisors to invest a lot of resourcesin the success eg publications and later placement of theirstudents while not understanding that the ambitions of the stu-dent might be very different or that in such a way early-careerresearches from other communities get a disadvantage

The process of writing a PhD thesis is difficult and demand-ing by itself Overwork and stress in some cases even endangermental health of early-career researchers (see ldquoBeing a PhDstudent shouldnrsquot be bad for your healthrdquo (2019) Nature 569307) This seems to be a too high price to pay for a diplomaPressure on graduate students that comes from the unfavourablepower position is hard to justify

Allowing for cumulative theses or several supervisors candecrease the dependence of students Having more than onesupervisor allows student to seek help and guidance from dif-ferent sources A cumulative thesis can ideally guarantee com-pletion of the studies given that the student published a suf-ficient number of scientific articles This opens a window ofopportunity for reforming the existing system

Vlasta SikimicUniversity of Belgrade

Events

October

EPoS Experimental Philosophy of Science Aarhus UniversityDenmark 15ndash16 OctoberRECSaC Radical Enactive Cognitive Science and its CriticsNova University of Lisbon 16ndash17 OctoberBIRDS Bridging the Gap between Information Science Infor-mation Retrieval and Data Science Melbourne Australia 19October

November

RaE Reasoning About Evidence University of Ghent 4ndash6NovemberPTampO Post-Truth and the Objectivity of Epistemic NormsUniversity of Sussex 7ndash8 November

December

CML Causal Machine Learning Vancouver 13ndash14 Decem-ber

Courses and Programmes

CoursesSSA Summer School on Argumentation Computational andLinguistic Perspectives on Argumentation Warsaw Poland 6ndash10 September

ProgrammesAPhil MAPhD in Analytic Philosophy University ofBarcelonaMaster Programme MA in Pure and Applied Logic Univer-sity of BarcelonaDoctoral Programme in Philosophy Language Mind andPractice Department of Philosophy University of ZurichSwitzerlandDoctoral Programme in Philosophy Department of Philoso-phy University of Milan ItalyLogiCS Joint doctoral program on Logical Methods in Com-puter Science TU Wien TU Graz and JKU Linz AustriaHPSM MA in the History and Philosophy of Science andMedicine Durham UniversityMaster Programme in Statistics University College DublinLoPhiSC Master in Logic Philosophy of Science and Epis-temology Pantheon-Sorbonne University (Paris 1) and Paris-Sorbonne University (Paris 4)

67

Master Programme in Artificial Intelligence Radboud Uni-versity Nijmegen the NetherlandsMaster Programme Philosophy and Economics Institute ofPhilosophy University of BayreuthMA in Cognitive Science School of Politics InternationalStudies and Philosophy Queenrsquos University BelfastMA in Logic and the Philosophy ofMathematics Departmentof Philosophy University of BristolMA Programmes in Philosophy of Science University ofLeedsMA in Logic and Philosophy of Science Faculty of PhilosophyPhilosophy of Science and Study of Religion LMU MunichMA in Logic and Theory of Science Department of Logic ofthe Eotvos Lorand University Budapest HungaryMA in Metaphysics Language and Mind Department of Phi-losophy University of LiverpoolMA inMind Brain and Learning Westminster Institute of Ed-ucation Oxford Brookes UniversityMA in Philosophy by research Tilburg UniversityMA in Philosophy Science and Society TiLPS Tilburg Uni-versityMA in Philosophy of Biological and Cognitive Sciences De-partment of Philosophy University of BristolMA in Rhetoric School of Journalism Media and Communi-cation University of Central LancashireMA programmes in Philosophy of Language and Linguisticsand Philosophy of Mind and Psychology University of Birm-inghamMRes in Methods and Practices of Philosophical ResearchNorthern Institute of Philosophy University of AberdeenMSc in Applied Statistics Department of Economics Mathe-matics and Statistics Birkbeck University of LondonMSc in Applied Statistics and Datamining School of Mathe-matics and Statistics University of St AndrewsMSc in Artificial Intelligence Faculty of Engineering Uni-versity of LeedsMSc in Cognitiveamp Decision Sciences Psychology UniversityCollege LondonMSc in Cognitive Systems Language Learning and Reason-ing University of PotsdamMSc in Cognitive Science University of Osnabruck GermanyMSc in Cognitive PsychologyNeuropsychology School ofPsychology University of KentMSc in Logic Institute for Logic Language and ComputationUniversity of AmsterdamMSc in Mind Language amp Embodied Cognition School ofPhilosophy Psychology and Language Sciences University ofEdinburghMSc in Philosophy of Science Technology and Society Uni-versity of Twente The NetherlandsMRes in Cognitive Science and Humanities Language Com-munication and Organization Institute for Logic CognitionLanguage and Information University of the Basque Country(Donostia San Sebastian)OpenMind International School of Advanced Studies in Cog-nitive Sciences University of BucharestResearchMaster in Philosophy and Economics Erasmus Uni-versity Rotterdam The Netherlands

Jobs and Studentships

JobsPostdoctoral Researcher in Philosophy of Science Univer-sity of Turin deadline 8 OctoberPost-doc positions in Robot Learning Decision Making andControl Alto University deadline 15 OctoberBertrand Russell Professorship of Philosophy in TheoreticalPhilosophy University of Cambridge deadline 18 OctoberLectureshipAssociate Professor in Philosophy of ScienceUniversity of Southampton deadline 19 OctoberAssistantAssociate Professor in Statistics or Applied Prob-ability University of Nottingham deadline 24 October

68