Hubert Dreyfus

from Wikipedia, the free encyclopedia
Dreyfus (2011)

Hubert L. Dreyfus (born October 15, 1929 in Terre Haute , Indiana ; † April 22, 2017 in Berkeley ) was an American philosopher and professor of philosophy at the University of California, Berkeley . He mainly dealt with phenomenology , existential philosophy , the philosophical effects of artificial intelligence (AI) and the philosophy of psychology and literature .

Family, education, job

Hubert Dreyfus (left) in front of his Berkeley house in 1976.

Hubert Lederer Dreyfus was the son of Stanley S. and Irene Lederer Dreyfus and was born in Terre Haute , Indiana . He studied at Harvard University (BA 1951, MA 1952) and received his doctorate in 1964 from the Norwegian philosopher Dagfinn Føllesdal .

Dreyfus taught at the Massachusetts Institute of Technology between 1960 and 1968 . In 1965 he worked for the Rand Corporation and was visiting professor at the Johann Wolfgang Goethe University in Frankfurt am Main and at Hamilton College in Clinton (Oneida County, New York) . In 1968 he became an Associate Professor of Philosophy at the University of California, Berkeley ; from 1972 to 1994 full professor at the University of California, Berkeley, and from 1994 until his death professor at the Graduate School of the University of California, Berkeley. He has held numerous visiting professorships at national and international universities.

background

Dreyfus became known for his criticism of artificial intelligence and as the author of the book What Computers Can't Can't ( 1972) . The limits of artificial intelligence . A revised version was released in 1979 under the title What Computers Still Can not Do .

In 1964 Dreyfus published his book Alchemy and Artificial Intelligence , which attacked the work of Allen Newell and Herbert A. Simon . Both were the leaders in the field of artificial intelligence. Dreyfus not only questioned the results of the two, but also criticized their basic requirements (intelligence arises through the manipulation of physical symbols according to appropriate formal rules). He believed the AI's research program was doomed to fail. However, in 1967 he was the first person to lose to a chess program (to MacHack VI by Richard Greenblatt ).

In 1980 Dreyfus and his brother Stuart Dreyfus released a skill acquisition model .

Dreyfus and Heidegger

In addition to his work in the field of artificial intelligence, Dreyfus became known for his work on the European philosophers, particularly Martin Heidegger and Michel Foucault , which he made available to analytically trained philosophers.

Dreyfus was part of the first US translation group for Heidegger's Being and Time . This happened in the 1950s when John Daniel Wild (1902–1972) had parts of Being and Time translated in his seminars at Harvard, Northwestern University and Yale . Dreyfus was then a Harvard graduate student. When he met the Canadian philosopher Charles Taylor on a visit to Oxford in 1952 , the first small group of people interested in Heidegger formed.

Dreyfus' book What Computers Can't Do (1972) is largely shaped by a reception of Heidegger's philosophy. For Dreyfus, Heidegger's concept of the "world" (see the article Being and Time ) stands in the way of the possibility of artificial intelligence for two reasons:

First of all, we never encounter individual things in the world to which we then attach a meaning, but for us things are always integrated into a whole of meaning: the hammer is there to hammer, nails hold the boards together for a house, the house serves that Living, protection against storms, etc. etc. Dreyfus ties in with Heidegger's terms “stuff”, “stuff wholeness” and “context of events”. However, this understanding of things can never be made fully explicit in propositional form ( apophantic "as", know-what ), but is always only accessible to a being who "always does" practical handling of things (hermeneutic "as") , know-how ).

Second, for Dreyfus, dealing with the world did not essentially consist in calculating possibilities in practical situations. We try z. For example, when we see the sun, not touching it, we are not talking to stones: however, assuming the representatives of artificial intelligence, all this would have to be done by a computer brain. Dreyfus did not consider this to be feasible due to the infinite number of possibilities to be calculated, because such a machine needed an infinite set of rules to deal with things in the world.

Dreyfus had quite a number of students who relate to Heidegger, such as B.

  • Charles B. Guignon ( Heidegger and the Problem of Knowledge (1983)),
  • John Richardson ( Existential Epistemology: A Heideggerian Critique of the Cartesian Project (1986)),
  • John Haugeland ( Dasein's Disclosedness (1989)),
  • William Blattner ( Heidegger's Temporal Idealism (1999)).

It is characteristic of the Heidegger reception , which is linked to Dreyfus , that he and his students turn almost exclusively to epistemological phenomena. Heidegger's concept of being-in-the-world is opposed to Edmund Husserl - Searl's view of intentionality, but the entire second half of being and time , i.e. the themes of temporality and historicity, is neglected.

Dreyfus' criticism of the AI

Dreyfus' criticism of artificial intelligence (AI) was based on how he viewed the four primary assumptions of AI research. The first two assumptions he criticized are called biological and psychological assumptions. The biological assumption is that the brain can be compared to computer hardware and the mind can be compared to computer software. The psychological assumption is that the mind works by making discrete calculations (in the form of algorithmic rules) on discrete representations of symbols.

Dreyfus claimed that the credibility of the psychological assumption is based on two others: the epistemological and the ontological assumption. The epistemological assumption is that all activity (of living and inanimate objects) can be described in advance by mathematical rules and laws. Researchers in the field (of AI) argue that intelligence is nothing more than following formal laws. The ontological assumption states that reality actually consists of a set of independent, atomic (indivisible) facts (see fuzzy logic ). This is justified by the fact that people use internal representations of reality.

Based on these two assumptions, researchers in the field (AI) claim that knowledge comes from manipulating internal symbols through internal rules. Therefore, human behavior is largely free of context. Hence, a true scientific psychology is possible that describes in detail the internal rules of the human mind, just as the laws of physics describe the external laws of the physical world. But it was precisely this key assumption that Dreyfus denied. In other words, he claimed that we will never understand our own behavior in the same way as we understand objects, for example in physics or chemistry, that is, by viewing ourselves as a thing whose behavior is predicted by objective and context-free scientific laws can be. According to Dreyfus, a context-free psychology is a contradiction in terms.

Dreyfus' arguments against this position stem from the phenomenological tradition (especially the work of Martin Heidegger ). Heidegger argued that our being is very contextual, as opposed to the cognitive view on which AI is based, which is why the two context-free assumptions are wrong. Dreyfus did not deny that we can view human (or other) activity as governed by law, just as we can view reality as a collection of atomic facts - if we want to. But from there it means a big leap towards the definition: Just because we can or want to see things that way, therefore it is an objective fact and therefore the case. Indeed, Dreyfus argued that they are not (necessarily) the case and therefore any research program that assumes it is very quickly run into profound theoretical and practical problems. Therefore, the current endeavors of researchers in this area are doomed to fail.

It is important to emphasize that Dreyfus did not argue for a fundamental impossibility of (strong) AI, but for the inadequacy of so-called classical research approaches. He was convinced that the possibility of human-like intelligence presupposes a correspondingly human-like presence in the world. A device designed for this purpose therefore needs a body, as well as a social context (e.g. a society) and a manipulable environment, as meaning and meaning arise only in the targeted interaction of these components and thus a basis for intelligent action at all. There are similar approaches in the cognitive sciences, for example by representatives of embodiment and z. B. Rodney Brooks also in the field of robotics .

Honors

The Erasmus University Rotterdam awarded Dreyfus an honorary doctorate: for his brilliant and very influential work in the field of artificial intelligence and for his equally outstanding contribution to the analysis and interpretation of European philosophers in the twentieth century. He was also a member of the American Academy of Arts and Sciences since 2001 .

Fonts

  • Alchemy and Artificial Intelligence. 1964.
  • Continental Philosophy: An Introduction.
  • What Computers Can't Do: The Limits of Artificial Intelligence. 1972, ISBN 0-06-090613-8 . German: What computers can't. The limits of artificial intelligence . Athenaeum, Königstein / Ts. 1985. ISBN 3-7610-8369-6
  • What Computers Still Can't Do: A Critique of Artificial Reason. 1979, ISBN 0-262-54067-3 . New edition: MIT Press, 1992, ISBN 0-262-04134-0 .
  • with Paul Rabinow : Michel Foucault. Beyond structuralism and hermeneutics . Univ. of Chicago Press, Chicago 1982 (German: Michel Foucault. Beyond structuralism and hermeneutics . Athenaeum, Frankfurt / M. 1987. ISBN 3-610-00732-X )
  • with Stuart Dreyfus : Mind Over Machine. Free Press, 1986 (German: Artificial Intelligence. On the Limits of the Thinking Machine and the Value of Intuition . Rowohlt, Reinbek b. Hamburg 1987. ISBN 3-499-18144-4 )
  • Being in the World: Division 1. 1991.
  • On the internet. Routledge 2001, ISBN 0-415-22807-7 .
  • Internet. 2002.
  • with Sean Kelly: All Things Shining. Reading the Western Classics to Find Meaning in a Secular Age. New York 2011.
  • with Charles Taylor : Retrieving Realism . Harvard University Press, New York 2015 (German: The recovery of realism . Suhrkamp, ​​Berlin 2016. ISBN 978-3518586853 )

literature

  • George Lakoff , Mark Johnson : Philosophy in the Flesh: the Embodied Mind and its Challenge to Western Thought. Basic Books, 1999, ISBN 0465056733
  • Mark A. Wrathall (Ed.): Heidegger, Coping, and Cognitive Science: Essays in Honor of Hubert L. Dreyfus. MIT Press, Cambridge (Mass.) / London, 2000, ISBN 978-0-262-73128-7

Web links

Individual evidence

  1. ↑ In Memoriam: Hubert Dreyfus. American Philosophical Association , April 22, 2017, accessed April 29, 2017 . Yasmin Anwar: Hubert Dreyfus, preeminent philosopher and AI critic, dies at 87.Berkeley News , April 24, 2017, accessed on April 29, 2017 .
  2. ^ Hubert L. Dreyfus Curriculum Vitae . University of California, Berkeley, as of January 16, 2013, accessed April 29, 2017.
  3. ^ A b Charles B. Guignon: Heidegger, American pragmatism and analytical philosophy . In: Dieter Thomä (Ed.): Heidegger Handbook: Life - Work - Effect. Metzler, Stuttgart, 2003, ISBN 978-3-476-01804-5 , p. 459.
personal data
SURNAME Dreyfus, Hubert
ALTERNATIVE NAMES Dreyfus, Hubert L.
BRIEF DESCRIPTION American philosopher
DATE OF BIRTH October 15, 1929
PLACE OF BIRTH Terre Haute , Indiana
DATE OF DEATH April 22, 2017
Place of death Berkeley