Top Banner
Professor dr. J. J. Meyer Menno Lievers
43

PHILOSOPHY OF AI

Jan 19, 2016

Download

Documents

qamra

PHILOSOPHY OF AI. Professor dr. J. J. Meyer Menno Lievers. WHAT IS THINKING?. Conceptual problem? Empirical problem? What is the task of philosophy? Is philosophy a science?. Conceptual problem?. Solution: conceptual analysis (Bennett & Hacker). Empirical problem?. - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: PHILOSOPHY OF AI

Professor dr. J. J. MeyerMenno Lievers

Page 2: PHILOSOPHY OF AI

WHAT IS THINKING?Conceptual problem?Empirical problem?What is the task of philosophy?Is philosophy a science?

Page 3: PHILOSOPHY OF AI

Conceptual problem?Solution: conceptual analysis (Bennett &

Hacker)

Page 4: PHILOSOPHY OF AI

Empirical problem?

No point in doing philosophy. Do research!

Page 5: PHILOSOPHY OF AI

The problemJonathan Bennett, Rationality (1964!)Honey-bees

Page 6: PHILOSOPHY OF AI

Descartes (1596-1650)Dualism of mind and bodyExtension is the essence of bodyThinking is the essence of mind

Page 7: PHILOSOPHY OF AI

Descartes on thoughtThinking occurs with the aid of ideas

(Language of thought…)Ideas are innate (inborn)Methodological solipsism - one can describe

what a subject thinks by taking into account the thinking subject in isolation from his/her environment

Epistemological internalism: a subject can only be said to know that P, iff she can justify P

Page 8: PHILOSOPHY OF AI

Locke (1632 - 1704)No innate ideas (à la connectionism…)Thinking is a mental act, not the essence of

the mindThinking requires perception

Page 9: PHILOSOPHY OF AI

Kant (1724 - 1804)Synthesis of Descartes and Locke‘Thoughts without content are empty;

intuitions without concepts are blind.’Copernican turn: the way we perceive reality

is a product of our thoughts (thus Harnad’s grounding problem…)

Page 10: PHILOSOPHY OF AI

Frege (1848 - 1925)Linguïstic idealism: the way we perceive

reality is a product of our languageThoughts are linguistic

Page 11: PHILOSOPHY OF AI

Dummett (1925 - )The Basic Tenet of Analytical PhilosophyAn account of language does not

presuppose an account of thoughtAn account of language yields an account

of thoughtThere is no other adequate means by

which an account of thought may be given

Page 12: PHILOSOPHY OF AI

Thoughts are linguisticThe structure of our thoughts equals the

structure of sentences of our languageThe normative rule we ought follow in

thinking are the rules of our languageThinking obeys the rules of logic‘What is meaning?’ becomes the central

question in philosophy.

Page 13: PHILOSOPHY OF AI

What is meaning?Meaning resides in your head (Descartes,

Locke)Meaning is a relation between words and

things in reality

Page 14: PHILOSOPHY OF AI

Putnam’s Twin-earthEarth: water is H2OTwin-earth: water is XYZWhat is the meaning of ‘water’?What do twins think about?Semantic externalism

Page 15: PHILOSOPHY OF AI

ExternalismSemantic externalism: after having been

baptized reality determines whether a word has been used correctly or not

Externalism in the philosophy of mind: the content of thoughts is determined by the environment of the thinker

Epistemological externalism: we can attribute possession of knowledge that P to a subject whether she can justify P or not

Page 16: PHILOSOPHY OF AI

The ontology of the mindDualismMaterialism

Page 17: PHILOSOPHY OF AI

Advantages of dualismFree willRationality and normativityCreativityInner experienceQualitative aspects of perception

Page 18: PHILOSOPHY OF AI

Objections to dualismCausal interaction between body and mindMental causation?What is a mental substance?

Page 19: PHILOSOPHY OF AI

MaterialismIdentitytheoryGrandma’s neuronPain = the firing of C-fibersLeibniz’s Law:

(x)((x=y)-->(F)(Fx<-->Fy))

Page 20: PHILOSOPHY OF AI

Token-token materialismMultiple realisibilitySupervenienceEmergence

Page 21: PHILOSOPHY OF AI

Putnam’s machine functionalismTuring machineFunctional descriptionIdentity mental state determined by input,

relations with other mental states, and outputAntireductionism: psychology remains an

autonomous science

Page 22: PHILOSOPHY OF AI

Problems for functionalismQualiaChinese room Inverted spectrumAbsent qualiaChinese people (functional characterization

of mental states too wide)

Page 23: PHILOSOPHY OF AI

Develoments after functionalismComputational model of thoughtsLanguage of thought hypothesisModularity of mindRise of subpersonal psychology

Page 24: PHILOSOPHY OF AI

ModulesDomainspecificMandatory operationLimited central accessInput systems are fastShallow outputTypical diseaseprocessesInformationally encapsulatedSpecific neural structure

Page 25: PHILOSOPHY OF AI

Rise subpersonal psychologySubdoxastic statesModulesImplicit knowledge

Page 26: PHILOSOPHY OF AI

Back to meaningTheory of meaning is a theory of

understandingConsequence: Philosophy of language is

imbedded in the philosophy of mindMeaning is being analysed as a ‘way of

thinking’

Page 27: PHILOSOPHY OF AI

Gareth EvansExternalismGenerality constraint: you can only attribute

to a thinking subject possession of the concept F, if and only if he or she cannot only entertain the thought that Fa, but also that Fb

Non-conceptual content

Page 28: PHILOSOPHY OF AI

Back to ontology:eliminative materialismScientific realismImpossibility of inter-theoretical reductionTheory-ladenness of perceptionMeaning-holismFolk-psychology is a false theory

Page 29: PHILOSOPHY OF AI

Consequences for AI:1. Symbol System HypothesisEmploy a rich, recursive compositional

language to represent realityBuild an adequate representation of

reality within a universal symbol systemUse input to construct representations of

the environment in response to stimuliProcess input (possibly into output)Output is a symbolic representation of

adequate, suitable responses to the input

Page 30: PHILOSOPHY OF AI

Philosophical presuppositions of the SSHInternalismToken-token materialism?Thinking is symbolmanipulationInnate representations?Methodological dualismThinking normative/logical?

Page 31: PHILOSOPHY OF AI

Connectionist systems

Page 32: PHILOSOPHY OF AI

Connectionist systemsAdaptive (empirism)Thoughts are not propositional

(compositional)Externalism?Normativity of thinking?Biologically real?No innate mental properties/knowledge?

Page 33: PHILOSOPHY OF AI

AI versus NeurofilosofieIntelligence is a biological phenomenonRepresentation is a product of our biological

constitutionBiological materialism‘Hard’ AI presupposes supervenience and

that is nonsense

Page 34: PHILOSOPHY OF AI

COURSE SET-UPMAPPING GREAT DEBATES:

CAN COMPUTERS THINK?

WEB-SITE: http://www.macrovu.com/CCTGeneralInfo.html

Page 35: PHILOSOPHY OF AI

SEVEN QUESTIONS1. CAN COMPUTERS THINK?

Page 36: PHILOSOPHY OF AI

SEVEN QUESTIONS1. CAN COMPUTERS THINK?2. CAN THE TURING-TEST DETERMINE

WHETHER COMPUTERS CAN THINK?

Page 37: PHILOSOPHY OF AI

SEVEN QUESTIONS1. CAN COMPUTERS THINK?2. CAN THE TURING-TEST DETERMINE

WHETHER COMPUTERS CAN THINK?3. CAN PHYSICAL SYMBOL SYSTEMS

THINK?

Page 38: PHILOSOPHY OF AI

SEVEN QUESTIONS1. CAN COMPUTERS THINK?2. CAN THE TURING-TEST DETERMINE

WHETHER COMPUTERS CAN THINK?3. CAN PHYSICAL SYMBOL SYSTEMS

THINK?4. CAN CHINESE ROOMS THINK?

Page 39: PHILOSOPHY OF AI

SEVEN QUESTIONS1. CAN COMPUTERS THINK?2. CAN THE TURING-TEST DETERMINE

WHETHER COMPUTERS CAN THINK?3. CAN PHYSICAL SYMBOL SYSTEMS

THINK?4. CAN CHINESE ROOMS THINK?

Page 40: PHILOSOPHY OF AI

SEVEN QUESTIONS5’. CAN CONNECTIONIST NETWORKS

THINK?

Page 41: PHILOSOPHY OF AI

SEVEN QUESTIONS5’. CAN CONNECTIONIST NETWORKS

THINK?5’’.CAN COMPUTERS THINK IN IMAGES?

Page 42: PHILOSOPHY OF AI

SEVEN QUESTIONS5’. CAN CONNECTIONIST NETWORKS

THINK?5’’.CAN COMPUTERS THINK IN IMAGES?6. DO COMPUTERS HAVE TO BE CONSCIOUS

TO THINK?

Page 43: PHILOSOPHY OF AI

SEVEN QUESTIONS5’. CAN CONNECTIONIST NETWORKS

THINK?5’’.CAN COMPUTERS THINK IN IMAGES?6. DO COMPUTERS HAVE TO BE CONSCIOUS

TO THINK?7. ARE THINKING COMPUTERS

MATHEMATICALLY POSSIBLE?