Find counter arguments by entering NameVs… or …VsName.
The author or concept searched is found in the following 38 entries.
| Disputed term/author/ism | Author |
Entry |
Reference |
|---|---|---|---|
| Absoluteness | Lewis | IV 245 Absolute/relative/vagueness/Unger/Lewis: e.g. (Peter Unger): you could say there is really nothing that is really smooth. Sure, the pavement is flat, but the desk is still smoother. Lewis pro: suppose that for no description of the relative vagueness of "smooth" and "smoother" it is true that something is smoother than something that is smooth. Score keeping/vagueness/Lewis: the correct answer is that he changed their score. The approximation rule allows Unger to create a context in which everything he says is true but that does not prove that something is wrong in our everyday contexts. Cf. >Absoluteness/Unger. |
Lewis I David K. Lewis Die Identität von Körper und Geist Frankfurt 1989 Lewis I (a) David K. Lewis An Argument for the Identity Theory, in: Journal of Philosophy 63 (1966) In Die Identität von Körper und Geist, Frankfurt/M. 1989 Lewis I (b) David K. Lewis Psychophysical and Theoretical Identifications, in: Australasian Journal of Philosophy 50 (1972) In Die Identität von Körper und Geist, Frankfurt/M. 1989 Lewis I (c) David K. Lewis Mad Pain and Martian Pain, Readings in Philosophy of Psychology, Vol. 1, Ned Block (ed.) Harvard University Press, 1980 In Die Identität von Körper und Geist, Frankfurt/M. 1989 Lewis II David K. Lewis "Languages and Language", in: K. Gunderson (Ed.), Minnesota Studies in the Philosophy of Science, Vol. VII, Language, Mind, and Knowledge, Minneapolis 1975, pp. 3-35 In Handlung, Kommunikation, Bedeutung, Georg Meggle Frankfurt/M. 1979 Lewis IV David K. Lewis Philosophical Papers Bd I New York Oxford 1983 Lewis V David K. Lewis Philosophical Papers Bd II New York Oxford 1986 Lewis VI David K. Lewis Convention. A Philosophical Study, Cambridge/MA 1969 German Edition: Konventionen Berlin 1975 LewisCl Clarence Irving Lewis Collected Papers of Clarence Irving Lewis Stanford 1970 LewisCl I Clarence Irving Lewis Mind and the World Order: Outline of a Theory of Knowledge (Dover Books on Western Philosophy) 1991 |
| Computer Games | Norvig | Norvig I 189 Computer Games/artificial intelligence/algorithms/Norvig/Russell: A game can be defined by the initial state (how the board is set up), the legal actions in each state, the result of each action, a terminal test (which says when the game is over), and a utility function that applies to terminal states. In two-player zero-sum games with perfect information, the minimax algorithm can select optimal moves by a depth-first enumeration of the game tree. The alpha–beta search algorithm computes the same optimal move as minimax, but achieves much greater efficiency by eliminating sub trees that are provably irrelevant. Usually, it is not feasible to consider the whole game tree (even with alpha–beta), so we Norvig I 190 need to cut the search off at some point and apply a heuristic evaluation function that estimates the utility of a state. Many game programs precompute tables of best moves in the opening and endgame so that they can look up a move rather than search. Games of chance can be handled by an extension to the minimax algorithm that evaluates a chance node by taking the average utility of all its children, weighted by the probability of each child. Optimal play in games of imperfect information, such as Kriegspiel and bridge, requires reasoning about the current and future belief states of each player. A simple approximation can be obtained by averaging the value of an action over each possible configuration of missing information. >Minimax algorithm. History of mechanical games: The most notorious of these was Baron Wolfgang von Kempelen’s (1734–1804) “The Turk,” a supposed chess-playing automaton that defeated Napoleon before being exposed as a magician’s trick cabinet housing a human chess expert (see Levitt, 2000)(1). Computer chess: In 1846, Charles Babbage (who had been fascinated by the Turk) appears to have contributed the first serious discussion of the feasibility of computer chess and checkers (Morrison and Morrison, 1961)(2). VsBabbage: He did not understand the exponential complexity of search trees, claiming “the combinations involved in the Analytical Engine enormously surpassed any required, even by the game of chess.” Chess: The NSS chess program (Newell et al., 1958)(3) used a simplified version of alpha-beta; it was the first chess program to do so. Alpha–beta pruning was described by Hart and Edwards (1961)(4) and Hart et al. (1972)(5). Alpha-beta was used by the “Kotok - McCarthy” chess program written by a student of John McCarthy (Kotok, 1962)(6). Knuth and Moore (1975)(7) proved the correctness of alpha-beta and analysed its time complexity. Pearl (1982b)(8) shows alpha–beta to be asymptotically optimal among all fixed-depth game-tree search algorithms. >Chess/artificial intelligence/Norvig/Russell. 1. Levitt, G. M. (2000). The Turk, Chess Automaton. McFarland and Company. 2. Morrison, P. and Morrison, E. (Eds.). (1961). Charles Babbage and His Calculating Engines: Selected Writings by Charles Babbage and Others. Dover. 3. Newell, A., Shaw, J. C., and Simon, H. A. (1958). Chess playing programs and the problem of complexity. IBM Journal of Research and Development, 4(2), 320–335. 4. Hart, T. P. and Edwards, D. J. (1961). The tree prune (TP) algorithm. Artificial intelligence project memo 30, Massachusetts Institute of Technology. 5. Hart, P. E., Nilsson, N. J., and Raphael, B. (1972). Correction to “A formal basis for the heuristic determination of minimum cost paths”. SIGART Newsletter, 37, 28–29. 6. Kotok, A. (1962). A chess playing program for the IBM 7090. AI project memo 41, MIT Computation Center. 7. Knuth, D. E. (1975). An analysis of alpha–beta pruning. AIJ, 6(4), 293–326. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Computer Games | Russell | Norvig I 189 Computer Games/artificial intelligence/algorithms/Norvig/Russell: A game can be defined by the initial state (how the board is set up), the legal actions in each state, the result of each action, a terminal test (which says when the game is over), and a utility function that applies to terminal states. In two-player zero-sum games with perfect information, the minimax algorithm can select optimal moves by a depth-first enumeration of the game tree. The alpha–beta search algorithm computes the same optimal move as minimax, but achieves much greater efficiency by eliminating sub trees that are provably irrelevant. Usually, it is not feasible to consider the whole game tree (even with alpha–beta), so we Norvig I 190 need to cut the search off at some point and apply a heuristic evaluation function that estimates the utility of a state. Many game programs precompute tables of best moves in the opening and endgame so that they can look up a move rather than search. Games of chance can be handled by an extension to the minimax algorithm that evaluates a chance node by taking the average utility of all its children, weighted by the probability of each child. Optimal play in games of imperfect information, such as Kriegspiel and bridge, requires reasoning about the current and future belief states of each player. A simple approximation can be obtained by averaging the value of an action over each possible configuration of missing information. >Minimax algorithm. History of mechanical games: The most notorious of these was Baron Wolfgang von Kempelen’s (1734–1804) “The Turk,” a supposed chess-playing automaton that defeated Napoleon before being exposed as a magician’s trick cabinet housing a human chess expert (see Levitt, 2000)(1). Computer chess: In 1846, Charles Babbage (who had been fascinated by the Turk) appears to have contributed the first serious discussion of the feasibility of computer chess and checkers (Morrison and Morrison, 1961)(2). VsBabbage: He did not understand the exponential complexity of search trees, claiming “the combinations involved in the Analytical Engine enormously surpassed any required, even by the game of chess.” Chess: The NSS chess program (Newell et al., 1958)(3) used a simplified version of alpha-beta; it was the first chess program to do so. Alpha–beta pruning was described by Hart and Edwards (1961)(4) and Hart et al. (1972)(5). Alpha-beta was used by the “Kotok - McCarthy” chess program written by a student of John McCarthy (Kotok, 1962)(6). Knuth and Moore (1975)(7) proved the correctness of alpha-beta and analysed its time complexity. Pearl (1982b)(8) shows alpha–beta to be asymptotically optimal among all fixed-depth game-tree search algorithms. >Chess/artificial intelligence/Norvig/Russell. 1. Levitt, G. M. (2000). The Turk, Chess Automaton. McFarland and Company. 2. Morrison, P. and Morrison, E. (Eds.). (1961). Charles Babbage and His Calculating Engines: Selected Writings by Charles Babbage and Others. Dover. 3. Newell, A., Shaw, J. C., and Simon, H. A. (1958). Chess playing programs and the problem of complexity. IBM Journal of Research and Development, 4(2), 320–335. 4. Hart, T. P. and Edwards, D. J. (1961). The tree prune (TP) algorithm. Artificial intelligence project memo 30, Massachusetts Institute of Technology. 5. Hart, P. E., Nilsson, N. J., and Raphael, B. (1972). Correction to “A formal basis for the heuristic determination of minimum cost paths”. SIGART Newsletter, 37, 28–29. 6. Kotok, A. (1962). A chess playing program for the IBM 7090. AI project memo 41, MIT Computation Center. 7. Knuth, D. E. (1975). An analysis of alpha–beta pruning. AIJ, 6(4), 293–326. |
Russell I B. Russell/A.N. Whitehead Principia Mathematica Frankfurt 1986 Russell II B. Russell The ABC of Relativity, London 1958, 1969 German Edition: Das ABC der Relativitätstheorie Frankfurt 1989 Russell IV B. Russell The Problems of Philosophy, Oxford 1912 German Edition: Probleme der Philosophie Frankfurt 1967 Russell VI B. Russell "The Philosophy of Logical Atomism", in: B. Russell, Logic and KNowledge, ed. R. Ch. Marsh, London 1956, pp. 200-202 German Edition: Die Philosophie des logischen Atomismus In Eigennamen, U. Wolf (Hg) Frankfurt 1993 Russell VII B. Russell On the Nature of Truth and Falsehood, in: B. Russell, The Problems of Philosophy, Oxford 1912 - Dt. "Wahrheit und Falschheit" In Wahrheitstheorien, G. Skirbekk (Hg) Frankfurt 1996 Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Cost-Benefit Analysis | Hicks | Parisi I 357 Cost-benefit analysis/Hicks: Benefit and costs are determined in principle by the measure developed by J. Hicks (1939)(1) known as the “compensating variation” (CV). In practice the measures are usually “consumer and producer surpluses” which are approximations of the Hicksian measure. Consumer surplus is approximately the amount one would pay minus the amount actually paid. Thus if a price is a cup of coffee is $1.00 and the willingness to pay for it is $2.50 (absent other vendors), the consumer surplus is $1.50. Producer surplus is the economic rent or the amount that could be taken away without affecting Parisi I 358 affecting supply. For example, considered as a whole, most of the wages of college football coaches is economic rent since, were their salaries lowered en masse, they would be unlikely to leave for other jobs. The measures correspond to willingness to pay (WTP) for gains and willingness to accept (WTA) payment to bear costs, which measures are consistent with the psychological (utility) effects associated with prospect theory, which is steeper for losses than gains, convex in losses, and concave in gains (Tversky and Kahneman, 1992)(2). >Willingness to pay/Tversky/Kahneman, >Cost-benefit analysis/Zerbe. 1. Hicks, John R. (1939). “The Foundations of Welfare Economics.” Economic Journal 49: 696. 2. Tversky, A. and D. Kahneman (1992). Journal of Risk and Uncertainty 29. Richard O. Zerbe. “Cost-Benefit Analysis in Legal Decision-making.” In: Parisi, Francesco (ed) (2017). The Oxford Handbook of Law and Economics. Vol 1: Methodology and Concepts. NY: Oxford University. |
EconHicks I John R. Hicks Mr. Keynes and the "classis"; a suggested reinterpreation 1937 Parisi I Francesco Parisi (Ed) The Oxford Handbook of Law and Economics: Volume 1: Methodology and Concepts New York 2017 |
| Designation | Geach | I 52 Naming/Denotation/Two-Names Theory/GeachVsAristoteles: Incorrect approximation of predication and naming: as if predicates were (complex) names : "on the mat". >Naming, >Predication, >Attribution, >Names, >Predicates, >Aristotle. ((s) "The man stabbing Caesar to death stabbed the one stabbed by Brutus.") Geach: Additionally, Geach would use a link. Two-names theory/Aristotle/Geach: "Socrates is a philosopher" should be true because the thing is named. GeachVs: "Philosopher" (general term) is not a name for "all (or every) philosopher". >General terms. --- I 153f Intentionality/naming/Parmenides/Geach: one cannot name anything that does not exist. (Geach pro) - ((s) Existence introduction is not arbitrary, not without premise). >Existence, >Existence statement, >Existence/Parmenides, >Introduction, >Nonexistence, >Fiction. E.g. Geach dreamed of a girl and wants to call it "Pauline". - On the other hand, acquaintance is sufficient - presence is not necessary. Problem: is the girl even more imaginary, if he has not dreamed of her? Geach: that is a sure sign that this is all nonsense. >Objects of thought, >Objects of belief. Geach with Parmenides: "There is only that what exists." GeachVsParmenides: However, one can talk about non-existent objects. - E.g. talking about absent friends without knowing that he is dead, changes the truth value, but not the fact that these are sentences. >Truth value, >Reference. Imaginary girls are not competing for identification in the dream. - If it is true of no identifiable girl that I dreamed of her, then I have not dreamed of any girl. >de re, >de dicto, >Identification. Solution: "I dreamed of a girl, but it is not true of a certain girl that I dreamed of her". - This is similar to: it is not true of a certain stamp that I want it. --- I 252 Predication/Geach: predication can be done without naming: in an if-that-sentence or in an or-sentence, a term P can be predicated of a thing without naming the thing "P". E.g. "If that what the policeman said is true, then he drove faster than 60". This does not call the policeman's sentence true. - (> Conditional). Predication/naming: centuries-old error: that the predicate is uttered by the thing. Frege: Difference >naming / >predication, >designation: to name a thing "P", a sentence must be asserted! But a property is also predicted in a non-assertive sub-clause (subset). Therefore, naming must be explained by predication, not vice versa. >Naming. |
Gea I P.T. Geach Logic Matters Oxford 1972 |
| Determinism | Davidson | McGinn I 135 Domestication Theory/McGinn: The decision connection only applies as a special case of domestication attempts, of a natural connection of a different kind, whereas the modality concerned is not fundamentally different from other modalities. Approximation to an independent model of the course of action. Such reductivist experiments can a) be deterministic (Davidson) or b) indeterministic. McGinn I 135 Davidson/Domestication Theory/McGinn: previous world conditions are a causally sufficient condition for a specific decision. Freedom is then a certain kind of causal series, namely, in which there is a corresponding set of mental precursors. >Initial conditions, >Free will, >Anomalous monism. Freedom is that desires and beliefs cause causal effects. McGinnVsDavidson: makes no sense to the concept of freedom of action, for it looks as if it is of the same kind as any other causality. The corresponding causality is then not an "event causality" but an "action causality". >Actions/Davidson. |
Davidson I D. Davidson Der Mythos des Subjektiven Stuttgart 1993 Davidson I (a) Donald Davidson "Tho Conditions of Thoughts", in: Le Cahier du Collège de Philosophie, Paris 1989, pp. 163-171 In Der Mythos des Subjektiven, Stuttgart 1993 Davidson I (b) Donald Davidson "What is Present to the Mind?" in: J. Brandl/W. Gombocz (eds) The MInd of Donald Davidson, Amsterdam 1989, pp. 3-18 In Der Mythos des Subjektiven, Stuttgart 1993 Davidson I (c) Donald Davidson "Meaning, Truth and Evidence", in: R. Barrett/R. Gibson (eds.) Perspectives on Quine, Cambridge/MA 1990, pp. 68-79 In Der Mythos des Subjektiven, Stuttgart 1993 Davidson I (d) Donald Davidson "Epistemology Externalized", Ms 1989 In Der Mythos des Subjektiven, Stuttgart 1993 Davidson I (e) Donald Davidson "The Myth of the Subjective", in: M. Benedikt/R. Burger (eds.) Bewußtsein, Sprache und die Kunst, Wien 1988, pp. 45-54 In Der Mythos des Subjektiven, Stuttgart 1993 Davidson II Donald Davidson "Reply to Foster" In Truth and Meaning, G. Evans/J. McDowell Oxford 1976 Davidson III D. Davidson Essays on Actions and Events, Oxford 1980 German Edition: Handlung und Ereignis Frankfurt 1990 Davidson IV D. Davidson Inquiries into Truth and Interpretation, Oxford 1984 German Edition: Wahrheit und Interpretation Frankfurt 1990 Davidson V Donald Davidson "Rational Animals", in: D. Davidson, Subjective, Intersubjective, Objective, Oxford 2001, pp. 95-105 In Der Geist der Tiere, D Perler/M. Wild Frankfurt/M. 2005 McGinn I Colin McGinn Problems in Philosophy. The Limits of Inquiry, Cambridge/MA 1993 German Edition: Die Grenzen vernünftigen Fragens Stuttgart 1996 McGinn II C. McGinn The Mysteriouy Flame. Conscious Minds in a Material World, New York 1999 German Edition: Wie kommt der Geist in die Materie? München 2001 |
| Evidence | Naess | I 397 Increase in evidence/evidence/naess: increase in evidence leads to more reasons to claim something, but not repetition of "it is so". No approximation of knowledge: E.g., I reach for a dummy - search for knowledge is not comparable to this. - Increase of evidence may be measurable, but not of knowledge. >Knowledge, >Justification. If understanding, remembering and seeing implies truth, then increase of evidence cannot lead to this knowledge or memory. - A distance to the truth is not determinable as in the case of grasping next to it. >Understanding, >Recognition, >Cognition, >Theory of knowledge. I 399 No approach to knowledge: Either the statement is false or it is true, then we already possess knowledge. >Truth, >Statements, >Truth values. |
Naess I Arne Naess Can Knowledge Be Reached? Inquiry 1961, S. 219-227 In Wahrheitstheorien, Gunnar Skirbekk Frankfurt/M. 1977 |
| Experience | Sellars | McDowell I 163 Experience/Sellars/McDowell: experience is not possible without concepts that play a role in the system of beliefs. What we regard as the message of experience, is already a part and no external constraint of the system. >Concepts/Sellars, >Consciousness/Sellars. Experience/experience history/Sellars: experience history not the result of impressions, but of phenomena. Phenomena/Sellars: phenomena are conceptually (in order to establish them in a rational relationship to beliefs). >Beliefs/Sellars. --- I XXXVIII Sellars: Three possible experiences: 1. This is a red object, 2. The object appears to be red (maybe it is white in red light). 3. It looks like it would be an object. (Also hallucination is possible). >Appearance/Sellars. One might assume that they cherish the same thought in all three cases, but obtain a new setting every time towards this thought. (Reductive-materialistic: Armstrong, Dennett). Ambiguity: process or result of process: experience, activities, non-descriptive content - experience object. I 44 Fact/Sellars: the fact that something seems to be red over there, is not experiencing. (Although it is a fact, of course.) >Appearance. But that does not mean that the common descriptive core might be perhaps experiencing. Facts: are experienced but are not experiencing. And also no experience. Experience: we need a name for the experience, which is not only an abbreviation for a description. Does the everyday language have such names? I 44 Experience/Sellars: ambiguity: experiencing: activity, non-descriptive content - experience object: the fact that something over there this seems to be red, is not experiencing. -However: the common descriptive core can be an experiencing - facts: are experienced, but are not an experiencing- and no experience. I 45 Experience/Sellars: Problem: sensation always seems to presuppose an object. - Possible solution: should we align sensations and propositional attitudes? >Sensations. - "Vs: this is the approximation leading to Locke/Descartes - Descartes/Locke: sensations on a stage with ideas. - SellarsVs. |
Sellars I Wilfrid Sellars The Myth of the Given: Three Lectures on the Philosophy of Mind, University of London 1956 in: H. Feigl/M. Scriven (eds.) Minnesota Studies in the Philosophy of Science 1956 German Edition: Der Empirismus und die Philosophie des Geistes Paderborn 1999 Sellars II Wilfred Sellars Science, Perception, and Reality, London 1963 In Wahrheitstheorien, Gunnar Skirbekk Frankfurt/M. 1977 McDowell I John McDowell Mind and World, Cambridge/MA 1996 German Edition: Geist und Welt Frankfurt 2001 McDowell II John McDowell "Truth Conditions, Bivalence and Verificationism" In Truth and Meaning, G. Evans/J. McDowell |
| Historiography | Droysen | Gadamer I 11 History/History/Droysen/Gadamer: As early as 1843, J. G. Droysen, the author and discoverer of the history of Hellenism, wrote: "There is probably no scientific field so remote, theoretically Gadamer I 12 justified, delimited and structured than history". Gadamer: Already Droysen had called for a Kant who in a categorical imperative of history would prove "the living source from which the historical life of mankind flows". He expresses the expectation that "the more profound concept of history will be the gravitational point in which the wild sway of the human sciences now has to gain steadfastness and the possibility of further progress"(1). The model of the natural sciences that Droysen calls up here is thus not meant in terms of content, in the sense of an approximation of the theory of science, but on the contrary in the sense that the humanities would have to be founded as an equally independent group of sciences. Droysen's [work] is an attempt to solve this problem. Cf. >Humanities/Dilthey. 1. J. G. Droysen, Historik (Neudruck 1925, hrsg. von E. Rothacker), S. 97. |
Droys I J. G. Droysen Grundriss der Historik Paderborn 2011 Gadamer I Hans-Georg Gadamer Wahrheit und Methode. Grundzüge einer philosophischen Hermeneutik 7. durchgesehene Auflage Tübingen 1960/2010 Gadamer II H. G. Gadamer The Relevance of the Beautiful, London 1986 German Edition: Die Aktualität des Schönen: Kunst als Spiel, Symbol und Fest Stuttgart 1977 |
| Idealization | Cartwright | I 105 Approach/Cartwright: approaches are a proxies. I 107 They lead away from the theory - but can lead to the truth. >Theories, >Truth. I 111 Idealization/Science/Cartwright: does not only omit, but introduces something that is not prescribed by the fundamental laws. - Approach: We cannot make corrections at the beginning. - We cannot correct backwards and thus assume to come out at a fundamental law ->crossover effects. I 119 It may be that what is the correct approach is not decided by the facts. ((s)>Non-factualism). I.e. two approaches (with different results) can be justified by the same facts - the same approach, if applied in different places, can have different results: E.g. Lamb shift: excited atom or in the base state - not a fact prescribes which is to be assumed. >Facts. I 121 Accuracy is only apparent if the initial problem is not given exactly. I 150 Idealization/Distortion/Science/Physics/Cartwright: Example a) interested in atoms: Then distortion in the description of the field (E.g. infinite number of degrees of freedom) - b) if field is examined: then infinite degrees of freedom stored in the walls of the laser cavity, etc. - realistic: is an approach that uses more bridge principles. >Bridge laws, >Principles, >Reality. Hacking I 361 Approximation/Cartwright: Problem: approximation should lead away from confusing details - but the number of possible approximations itself is confusing - most approximate equations are themselves already simplifications of equations that you could not solve. |
Car I N. Cartwright How the laws of physics lie Oxford New York 1983 CartwrightR I R. Cartwright A Neglected Theory of Truth. Philosophical Essays, Cambridge/MA pp. 71-93 In Theories of Truth, Paul Horwich Aldershot 1994 CartwrightR II R. Cartwright Ontology and the theory of meaning Chicago 1954 Hacking I I. Hacking Representing and Intervening. Introductory Topics in the Philosophy of Natural Science, Cambridge/New York/Oakleigh 1983 German Edition: Einführung in die Philosophie der Naturwissenschaften Stuttgart 1996 |
| Idealization | Feyerabend | I 81 Approximation Method/Qualitative Difficulties/Feyerabend: Examples of reduction are: reduction of the wave packet (I 82), heliocentric theory at the time of Galileo. Ad-hoc approximations to many quantitative results of the theories are not correct and surprisingly qualitatively inadequate. E.g., von Neumann replaced the semi-intuitive concepts of Dirac and Bohr with incredibly complicated concepts. The relation to experience becomes more obscure than ever. >Experience, >Observation, >Explanation, >Reduction, >Reductionism. |
Feyerabend I Paul Feyerabend Against Method. Outline of an Anarchistic Theory of Knowledge, London/New York 1971 German Edition: Wider den Methodenzwang Frankfurt 1997 Feyerabend II P. Feyerabend Science in a Free Society, London/New York 1982 German Edition: Erkenntnis für freie Menschen Frankfurt 1979 |
| Idealization | Wittgenstein | II 432 Approximation/approach/proxy/numbers/Wittgenstein: e.g. 0.49 and 0.5: the approximation 0.49 is not a different thing. Just another notation - a proof of 0.5 must be the same as the proof of 0.49. >Proofs, >Numbers, >Mathematics. |
W II L. Wittgenstein Wittgenstein’s Lectures 1930-32, from the notes of John King and Desmond Lee, Oxford 1980 German Edition: Vorlesungen 1930-35 Frankfurt 1989 W III L. Wittgenstein The Blue and Brown Books (BB), Oxford 1958 German Edition: Das Blaue Buch - Eine Philosophische Betrachtung Frankfurt 1984 W IV L. Wittgenstein Tractatus Logico-Philosophicus (TLP), 1922, C.K. Ogden (trans.), London: Routledge & Kegan Paul. Originally published as “Logisch-Philosophische Abhandlung”, in Annalen der Naturphilosophische, XIV (3/4), 1921. German Edition: Tractatus logico-philosophicus Frankfurt/M 1960 |
| Inference | AI Research | Norvig I 471 Reasoning/inference/artificial intelligence/AI research/Norvig/Russell: The three main formalisms for dealing with nonmonotonic inference—circumscription (McCarthy, 1980)(1), default logic (Reiter, 1980(2)), and modal nonmonotonic logic (McDermott and Doyle, 1980)(3) - were all introduced in one special issue of the AI Journal. Delgrande and Schaub (2003)(4) discuss the merits of the variants, given 25 years of hindsight. Answer set programming can be seen as an extension of negation as failure or as a refinement of circumscription; Norvig I 472 the underlying theory of stable model semantics was introduced by Gelfond and Lifschitz (1988)(5), and the leading answer set programming systems are DLV (Eiter et al., 1998)(6) and SMODELS (Niemel¨a et al., 2000)(7). The disk drive example comes from the SMODELS user manual (Syrjanen, 2000)(8). Lifschitz (2001)(9) discusses the use of answer set programming for planning. Brewka et al. (1997)(10) give a good overview of the various approaches to nonmonotonic logic. Clark (1978)(11) covers the negation-as-failure approach to logic programming and Clark completion. Van Emden and Kowalski (1976)(12) show that every Prolog program without negation has a unique minimal model. Recent years have seen renewed interest in applications of nonmonotonic logics to large-scale knowledge representation systems. The BENINQ systems for handling insurance-benefit inquiries was perhaps the first commercially successful application of a nonmonotonic inheritance system (Morgenstern, 1998)(13). Lifschitz (2001)(9) discusses the application of answer set programming to planning. Norvig I 473 Spatial reasoning: The earliest serious attempt to capture commonsense reasoning about space appears in the work of Ernest Davis (1986(14), 1990(15)). The region connection calculus of Cohn et al. (1997)(16) supports a form of qualitative spatial reasoning and has led to new kinds of geographical information systems; see also (Davis, 2006)(17). As with qualitative physics, an agent can go a long way, so to speak, without resorting to a full metric representation. Psychological reasoning: Psychological reasoning involves the development of a working psychology for artificial agents to use in reasoning about themselves and other agents. This is often based on so-called folk psychology, the theory that humans in general are believed to use in reasoning about themselves and other humans. ((s) Cf. >Folk psychology/Philosophical theories). When AI researchers provide their artificial agents with psychological theories for reasoning about other agents, the theories are frequently based on the researchers’ description of the logical agents’ own design. Psychological reasoning is currently most useful within the context of natural language understanding, where divining the speaker’s intentions is of paramount importance. Minker (2001)(18) collects papers by leading researchers in knowledge representation, summarizing 40 years of work in the field. The proceedings of the international conferences on Principles of Knowledge Representation and Reasoning provide the most up-to-date sources for work in this area. 1. McCarthy, J. (1980). Circumscription: A form of non-monotonic reasoning. AIJ, 13(1–2), 27–39. 2. Reiter, R. (1980). A logic for default reasoning. AIJ, 13(1–2), 81–132. 3. McDermott, D. and Doyle, J. (1980). Nonmonotonic logic: i. AIJ, 13(1–2), 41–72. 4. Delgrande, J. and Schaub, T. (2003). On the relation between Reiter’s default logic and its (major) variants. In Seventh European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty, pp. 452–463. 5. Gelfond, M. and Lifschitz, V. (1988). Compiling circumscriptive theories into logic programs. In Non- Monotonic Reasoning: 2nd International Workshop Proceedings, pp. 74–99. 6. Eiter, T., Leone, N., Mateis, C., Pfeifer, G., and Scarcello, F. (1998). The KR system dlv: Progress report, comparisons and benchmarks. In KR-98, pp. 406–417. 7. Niemela, I., Simons, P., and Syrjanen, T. (2000). Smodels: A system for answer set programming. In Proc. 8th International Workshop on Non-Monotonic Reasoning. 8. Syrjanen, T. (2000). Lparse 1.0 user’s manual.saturn.tcs.hut.fi/Software/smodels. 9. Lifschitz, V. (2001). Answer set programming and plan generation. AIJ, 138(1–2), 39–54. 10. Brewka, G., Dix, J., and Konolige, K. (1997). Nononotonic Reasoning: An Overview. CSLI Publications. 11. Clark, K. L. (1978). Negation as failure. In Gallaire, H. and Minker, J. (Eds.), Logic and Data Bases, pp. 293–322. Plenum. 12. Van Emden, M. H. and Kowalski, R. (1976). The semantics of predicate logic as a programming language. JACM, 23(4), 733–742. 13. Morgenstern, L. (1998). Inheritance comes of age: Applying nonmonotonic techniques to problems in industry. AIJ, 103, 237–271 14. Davis, E. (1986). Representing and Acquiring Geographic Knowledge. Pitman and Morgan Kaufmann. 15. Davis, E. (1990). Representations of Commonsense Knowledge. Morgan Kaufmann 16. Cohn, A. G., Bennett, B., Gooday, J. M., and Gotts, N. (1997). RCC: A calculus for region based qualitative spatial reasoning. GeoInformatica, 1, 275–316. 17. Davis, E. (2006). The expressivity of quantifying over regions. J. Logic and Computation, 16, 891– 916. 18. Minker, J. (2001). Logic-Based Artificial Intelligence. Kluwer Norvig I 570 Inference/temporal models/AI research/Norvig/Russell: (…) the basic inference tasks that must be solved: a) Filtering: This is the task of computing the belief state—the posterior distribution over the most recent state - given all evidence to date. Filtering is also called state estimation. >Belief states/Norvig. b) Prediction: This is the task of computing the posterior distribution over the future state, given all evidence to date. That is, we wish to compute P(Xt+k | e1:t) for some k > 0. Norvig I 571 c) Smoothing: This is the task of computing the posterior distribution over a past state, given all evidence up to the present. That is, we wish to compute P(Xk | e1:t) for some k such that 0 ≤ k < t. d) Most likely explanation: Given a sequence of observations, we might wish to find the sequence of states that is most likely to have generated those observations. That is, we wish to compute argmaxx1:t P(x1:t | e1:t). In addition to these inference tasks (…): Learning: The transition and sensor models, if not yet known, can be learned from observations. Just as with static >Bayesian networks, dynamic Bayes net learning can be done as a by-product of inference. Inference provides an estimate of what transitions actually occurred and of what states generated the sensor readings, and these estimates can be used to update the models. >Change/AI research, >Uncertainty/AI research. Norvig I 605 Ad a) The particle filtering algorithm (…) has a particularly interesting history. The first sampling algorithms for particle filtering (also called sequential Monte Carlo methods) were developed in the control theory community by Handschin and Mayne (1969)(1), and the resampling idea that is the core of particle filtering appeared in a Russian control journal (Zaritskii et al., 1975)(2). It was later reinvented in statistics as sequential importance sampling resampling, or SIR (Rubin, 1988(3); Liu and Chen, 1998(4)), in control theory as particle filtering (Gordon et al., 1993(5); Gordon, 1994(6)), in AI as survival of the fittest (Kanazawa et al., 1995)(7), and in computer vision as condensation (Isard and Blake, 1996)(8). The paper by Kanazawa et al. (1995)(7) includes an improvement called evidence reversal whereby the state at time t+1 is sampled conditional on both the state at time t and the evidence at time t+1. This allows the evidence to influence sample generation directly and was proved by Doucet (1997)(9) and Liu and Chen (1998)(4) to reduce the approximation error. Particle filtering has been applied in many areas, including tracking complex motion patterns in video (Isard and Blake, 1996)(8), predicting the stock market (de Freitas et al., 2000)(10), and diagnosing faults on planetary rovers (Verma et al., 2004)(11). A variant called the Rao-Blackwellized particle filter or RBPF (Doucet et al., 2000(12); Murphy and Russell, 2001)(13) applies particle filtering to a subset of state variables and, for each particle, performs exact inference on the remaining variables conditioned on the value sequence in the particle. In some cases RBPF works well with thousands of state variables. >Utility/AI research, >Utility theory/Norvig, >Rationality/AI research, >Certainty ffect/Kahneman/Tversky, >Ambiguity/Kahneman/Tversky. 1. Handschin, J. E. and Mayne, D. Q. (1969). Monte Carlo techniques to estimate the conditional expectation in multi-stage nonlinear filtering. Int. J. Control, 9(5), 547–559. 2. Zaritskii, V. S., Svetnik, V. B., and Shimelevich, L. I. (1975). Monte-Carlo technique in problems of optimal information processing. Automation and Remote Control, 36, 2015–22. 3. Rubin, D. (1988). Using the SIR algorithm to simulate posterior distributions. In Bernardo, J. M., de Groot,M. H., Lindley, D. V., and Smith, A. F. M. (Eds.), Bayesian Statistics 3, pp. 395–402. Oxford University Press. 4. Liu, J. S. and Chen, R. (1998). Sequential Monte Carlo methods for dynamic systems. JASA, 93, 1022–1031. 5. Gordon, N., Salmond, D. J., and Smith, A. F. M. (1993). Novel approach to nonlinear/non-Gaussian Bayesian state estimation. IEE Proceedings F (Radar and Signal Processing), 140(2), 107–113. 6. Gordon, N. (1994). Bayesian methods for tracking. Ph.D. thesis, Imperial College. 7. Kanazawa, K., Koller, D., and Russell, S. J. (1995). Stochastic simulation algorithms for dynamic probabilistic networks. In UAI-95, pp. 346–351. 8. Isard, M. and Blake, A. (1996). Contour tracking by stochastic propagation of conditional density. In ECCV, pp. 343–356. 9. Doucet, A. (1997). Monte Carlo methods for Bayesian estimation of hidden Markov models: Application to radiation signals. Ph.D. thesis, Université de Paris-Sud. 10. de Freitas, J. F. G., Niranjan, M., and Gee, A. H. (2000). Sequential Monte Carlo methods to train neural network models. Neural Computation, 12(4), 933–953. 11. Verma, V., Gordon, G., Simmons, R., and Thrun, S. (2004). Particle filters for rover fault diagnosis. IEEE Robotics and Automation Magazine, June. 12. Doucet, A., de Freitas, N., Murphy, K., and Russell, S. J. (2000). Rao-blackwellised particle filtering for dynamic bayesian networks. In UAI-00. 13. Murphy, K. and Russell, S. J. (2001). Rao-blackwellised particle filtering for dynamic Bayesian networks. In Doucet, A., de Freitas, N., and Gordon, N. J. (Eds.), Sequential Monte Carlo Methods in Practice. Springer-Verlag. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Information Extraction | AI Research | Norvig I 873 Information extraction/AI Research/Norvig/Russell: Information extraction is the process of acquiring knowledge by skimming a text and looking for occurrences of a particular class of object and for relationships among objects. A typical task is to extract instances of addresses from Web pages, with database fields for street, city, state, and zip code; (…).In a limited domain, this can be done with high accuracy. As the domain gets more general, more complex linguistic models and more complex learning techniques are necessary. Norvig I 874 A. Finite-state template-based information extraction: Attribute-based extraction system: (…) assumes that the entire text refers to a single object and the task is to extract attributes of that object. E.g., Manufacturer; product; price. Relational extraction systems: deal with multiple objects and the relations among them. Norvig I 875 A relational extraction system can be built as a series of cascaded finite-state transducers. E.g., FASTUS consists of five stages: 1. Tokenization, 2. Complex-word handling, 3. Basic-group handling, 4. Complex-phrase handling, 5. Structure merging. Norvig I 876 B. Probabilistic models for information extraction: When information extraction must be attempted from noisy or varied input, (…) it is better to use a probabilistic model rather than a rule-based model. The simplest probabilistic model for sequences with hidden state is the hidden Markov model, or HMM. >Bayesian networks, >Statistical learning. (…) an HMM models a progression through a sequence of hidden states, xt, with an observation et at each step. To apply HMMs to information extraction, [one] can either build one big HMM for all the attributes or build a separate HMM for each attribute. HMMs are probabilistic, and thus tolerant to noise. (…) with HMMs there is graceful degradation with missing characters/words, and [one] get[s] a probability indicating the degree of match, not just a Boolean match/fail. Norvig I 877 (…) HMMs can be trained from data; they don’t require laborious engineering of templates, and thus they can more easily be kept up to date as text changes over time. Norvig I 878 VsHMMs: Problem: One issue with HMMs for the information extraction task is that they model a lot of probabilities that we don’t really need. An HMM is a generative model; it models the full joint probability of observations and hidden states, and thus can be used to generate samples. That is, we can use the HMM model not only to parse a text and recover the speaker and date, but also to generate a random instance of a text containing a speaker and a date. Solution: All we need in order to understand a text is a discriminative model, one that models the conditional probability of the hidden attributes given the observations (the text). Conditional random field: We don’t need the independence assumptions of the Markov model - we can have an xt that is dependent on x1. A framework for this type of model is the conditional random field, or CRF, which models a conditional probability distribution of a set of target variables given a set of observed variables. Like Bayesian networks, CRFs can represent many different structures of dependencies among the variables. Norvig I 879 Ontology extraction: [different from] information extraction as finding a specific set of relations (e.g., speaker, time, location) in a specific text (e.g., a talk announcement) (…) [ontology extraction] is building a large knowledge base or ontology of facts from a corpus. This is different in three ways: First it is open-ended - we want to acquire facts about all types of domains, not just one specific domain. Second, with a large corpus, this task is dominated by precision, not recall - just as with >question answering on the Web (…). Third, the results can be statistical aggregates gathered from multiple sources, rather than being extracted from one specific text. E.g., Hearst (1992)(1) looked at the problem of learning an ontology of concept categories and subcategories from a large corpus. The work concentrated on templates that are very general (not tied to a specific domain) and have high precision (are Norvig I 880 almost always correct when they match) but low recall (do not always match). Here is one of the most productive templates: NP such as NP (, NP)* (,)? ((and | or) NP)? . Here [“such as”, “and”, “or”] and commas must appear literally in the text, but the parentheses are for grouping, the asterisk means repetition of zero or more, and the question mark means optional. Problems: The biggest weakness in this approach is the sensitivity to noise. If one of the first few templates is incorrect, errors can propagate quickly. One way to limit this problem is to not accept a new example unless it is verified by multiple templates, and not accept a new template unless it discovers multiple examples that are also found by other templates. Machine reading: (…) a system that could read on its own and build up its own database. Such a system would be relation-independent; would work for any relation. In practice, these systems work on all relations in parallel, because of the I/O demands of large corpora. They behave less like a traditional information extraction system that is targeted at a few relations and more like a human reader who learns from the text itself; because of this the field has been called machine reading. A representative machine-reading system is TEXTRUNNER (Banko and Etzioni, 2008)(2). TEXTRUNNER uses cotraining to boost its performance, but it needs something to bootstrap from. In the case of Hearst (1992)(1), specific patterns (e.g., such as) provided the bootstrap, and for Brin (1998)(3), it was a set of five author-title pairs. Norvig I 884 Early information extraction programs include GUS (Bobrow et al., 1977)(4) and FRUMP (DeJong, 1982)(5). Recent information extraction has been pushed forward by the annual Message Understand Conferences (MUC), sponsored by the U.S. government. The FASTUS finite-state system was done by Hobbs et al. (1997)(6). It was based in part on the idea from Pereira and Wright (1991)(7) of using FSAs as approximations to phrase-structure grammars. Surveys of template-based systems are given by Roche and Schabes (1997)(8), Appelt (1999)(9), Norvig I 885 Freitag and McCallum (2000)(10) discuss HMMs for Information Extraction. CRFs were introduced by Lafferty et al. (2001)(11); an example of their use for information extraction is described in (McCallum, 2003)(12) and a tutorial with practical guidance is given by (Sutton and McCallum, 2007)(13). Sarawagi (2007)(14) gives a comprehensive survey. 1. Hearst, M. A. (1992). Automatic acquisition of hyponyms from large text corpora. In COLING-92. 2. Banko, M. and Etzioni, O. (2008). The tradeoffs between open and traditional relation extraction. In ACL-08, pp. 28–36. 3. Brin, D. (1998). The Transparent Society. Perseus 4. Bobrow, D. G.,Kaplan, R.,Kay,M.,Norman, D. A., Thompson, H., and Winograd, T. (1977). GUS, a frame driven dialog system. AIJ, 8, 155–173. 5. DeJong, G. (1982). An overview of the FRUMP system. In Lehnert,W. and Ringle,M. (Eds.), Strategies for Natural Language Processing, pp. 149–176. Lawrence Erlbaum. 6. Hobbs, J. R., Appelt, D., Bear, J., Israel, D., Kameyama, M., Stickel, M. E., and Tyson, M. (1997). FASTUS: A cascaded finite-state transducer for extracting information from natural-language text. In Roche, E. and Schabes, Y. (Eds.), Finite- State Devices for Natural Language Processing, pp. 383–406. MIT Press. 7. Pereira, F. and Wright, R. N. (1991). Finite-state approximation of phrase structure grammars. In ACL-91, pp. 246–255. 8. Roche, E. and Schabes, Y. (1997). Finite-State Language Processing (Language, Speech and Communication). Bradford Books. 9. Appelt, D. (1999). Introduction to information extraction. CACM, 12(3), 161–172. 10. Freitag, D. and McCallum, A. (2000). Information extraction with hmm structures learned by stochastic optimization. In AAAI-00. 11. Lafferty, J., McCallum, A., and Pereira, F. (2001). Conditional random fields: Probabilistic models for segmenting and labeling sequence data. In ICML-01. 12. McCallum, A. (2003). Efficiently inducing features of conditional random fields. In UAI-03. 13. Sutton, C. and McCallum, A. (2007). An introduction to conditional random fields for relational learning. In Getoor, L. and Taskar, B. (Eds.), Introduction to Statistical Relational Learning. MIT Press. 14. Sarawagi, S. (2007). Information extraction. Foundations and Trends in Databases, 1(3), 261–377. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Information Value | Norvig | Norvig I 629 Information value/AI research/Norvig/Russell: One of the most important parts of decision making is knowing what questions to ask. Tests are often expensive and sometimes hazardous (both directly and because of associated delays). Their importance depends on two factors: whether the test results would lead to a significantly better (…) plan, and how likely the various test results are. Example: (…) an oil company is hoping to buy one of n indistinguishable blocks of ocean-drilling rights. (…) exactly one of the blocks contains oil worth C dollars, while the others are worthless. The asking price of each block is C/n dollars. (…)a seismologist offers the company the results of a survey of block number 3, which indicates definitively whether the block contains oil. How much should the company be willing to pay for the information? With probability 1/n, the survey will indicate oil in block 3. In this case, the company will buy block 3 for C/n dollars and make a profit of C −C/n = (n − 1)C/n dollars. With probability (n−1)/n, the survey will show that the block contains no oil, in which case the company will buy a different block. (…) the company should be willing to pay the seismologist up to C/n dollars for the information: the information is worth as much as the block itself. The value of information derives from the fact that with the information, one’s course of action can be changed to suit the actual situation. One can discriminate according to the situation, whereas without the information, one has to do what’s best on average over the possible situations. Def Information value/Norvig: (…) the value of a given piece of information is defined to be the difference in expected value between best actions before and after information is obtained. >Multi-attribute utility/AI Research, >Decision networks/Norvig. Norvig I 631 Nonnegative value of information: can [information] actually have negative expected value? Intuitively, one should expect this to be impossible. After all, one could in the worst case just ignore the information and pretend that one has never received it. This is confirmed by the following theorem, which applies to any decision-theoretic agent: The expected value of information is nonnegative. >Software agents/Norvig. Agents/AI/information value: A sensible agent should ask questions in a reasonable order, should avoid asking questions that are irrelevant, should take into account the importance of each piece of information in relation to its cost, and should stop asking questions when that is appropriate. All of these capabilities can be achieved by using the value of information as a guide. Norvig I 639 The theory of information value was explored first in the context of statistical experiments, where a quasi-utility (entropy reduction) was used (Lindley, 1956)(1). The Russian control theorist Ruslan Stratonovich (1965)(2) developed the more general theory presented here, in which information has value by virtue of its ability to affect decisions. Stratonovich’s work was not known in the West, where Ron Howard (1966)(3) pioneered the same idea. His paper ends with the remark “If information value theory and associated decision theoretic structures do not in the future occupy a large part of the education of engineers, then the engineering profession will find that its traditional role of managing scientific and economic resources for the benefit of man has been forfeited to another profession.” To date, the implied revolution in managerial methods has not occurred. Recent work by Krause and Guestrin (2009)(4) shows that computing the exact nonmyopic value of information is intractable even in polytree networks. There are other cases - more restricted than general value of information—in which the myopic algorithm does provide a provably good approximation to the optimal sequence of observations (Krause et al., 2008)(5). In some cases - for example, looking for treasure buried in one of n places – ranking experiments in order of success probability divided by cost gives an optimal solution (Kadane and Simon, 1977)(6). 1. Lindley, D. V. (1956). On a measure of the information provided by an experiment. Annals of Mathematical Statistics, 27(4), 986–1005. 2. Stratonovich, R. L. (1965). On value of information. Izvestiya of USSR Academy of Sciences, Technical Cybernetics, 5, 3–12. 3. Howard, R. A. (1966). Information value theory. IEEE Transactions on Systems Science and Cybernetics, SSC-2, 22–26. 4. Krause, A. and Guestrin, C. (2009). Optimal value of information in graphical models. JAIR, 35, 557 - 591. 5. Krause, A., McMahan, B., Guestrin, C., and Gupta, A. (2008). Robust submodular observation selection. JMLR, 9, 2761–2801. 6. Kadane, J. B. and Simon, H. A. (1977). Optimal strategies for a class of constrained sequential problems. Annals of Statistics, 5, 237–255. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Knowledge | Feynman | I 363 Principle of the Shortest Time/Fermat/"Knowledge"/Feynman: how does light know which path to take? Important: there is an area near the optimum point where, in the first approximation, there is no significant change in time. But there is in second order. Feynman: this is not about causality! >Refractive index, >Causality, >Principles. In fact, the light decides! It practically smells the right path. This is related to the wavelength. I 538 Knowledge/Saying/Language/Uncertainty Principle/Feynman: that we cannot precisely measure place and impulse does not mean a priori that we cannot talk about it. It just means we do not need to talk about it! >Uncertainty relation, >Indeterminacy, >Vagueness. --- Genz II Knowledge/Atoms/Atomism/Feynman: if all physical knowledge were destroyed and only one insight was preserved, the thesis that everything is made of atoms would be the most important one. >Atoms/Feynman. |
Feynman I Richard Feynman The Feynman Lectures on Physics. Vol. I, Mainly Mechanics, Radiation, and Heat, California Institute of Technology 1963 German Edition: Vorlesungen über Physik I München 2001 Feynman II R. Feynman The Character of Physical Law, Cambridge, MA/London 1967 German Edition: Vom Wesen physikalischer Gesetze München 1993 |
| Meaning Change | Rorty | I 293f Meaning Change/Rorty: Question: Did the Greeks refer to prudence with the expression Sophrosyne? >Reference. Rorty: This question can be rejected with the hint that there is for expectation; in a completely different culture this expression would be implantable; no particular reason. We have to make ourselves familiar with the exotic language game. >Relativism, >Cultural relativism, >Context dependence. In the case of science, however, such an attitude seems unnatural. Here we want to say that out there is something, laws to which one should refer or at least one has referred to. Rorty: "whiggistic" winner perspective: tells us, Aristotle spoke in reality of gravity, when he spoke of a natural settling movement, sailors would have, when they spoke of unicorns, referred to the horns of narwhals in reality, "heat flow" is a misleading description of the energy transfer between dancing molecules. >Theory change. I 301 Meaning/truth/existence/Change of Theories/Meaning Change/Quine/Rorty: Quineans would say, the question, whether they meant the same back then, is not raised. - It's more about the truth values. >Truth values, >Speaker meaning, >Speaker intention, >Meaning/Intending, >Assertibility. Rorty: a) Aristotle said something wrong about movement, or b) He said something true, but that was not movemnt. RortyVsAyers: with this, one will not get far if one does no longer believe in concepts like intellectual property etc. Ayers exaggerates the contrast between our and his concepts. I 315 ff Semantic change/change of theory/reference/Rorty: solution: the functioning of an expression should be better seen as the picking out of objects, than as the description of reality. - So either a) reference as a basis, or b) also accepting reference as conventional. - Searle-trawson-Criterion: "What would make most of his opinions true." I 318 Solution: distinction reference: a) philosophical - b) "Speaking about" (common sense) - Rorty: it is only about existence. - Therefore, no criterion for reference possible. I 321 RortyVsReference Theory/Theory of Reference: 1. Semantic search for the objects is hopeless. - 2. Hopeless: to strive for an epistemological refutation of skepticism. >Skepticism. --- III 103 Meaning change: Adorno/Horkheimer/Rorty: pro - PutnamVs. --- IV 131 Term/Meaning change/Conceptual change/Change of theories/Rorty: terms that got a new twist through a thinker: E.g. Aristotle: ousia Descartes: res Hume: impression. Wittgenstein: game Einstein: simultaneity. Bohr: Atom. >Theory Change, >Incommensurability. --- VI 361 Interpretation/Rorty: in such approximation efforts, the procedure is obviously anachronistic. But when that happens consciously, there is no objection. |
Rorty I Richard Rorty Philosophy and the Mirror of Nature, Princeton/NJ 1979 German Edition: Der Spiegel der Natur Frankfurt 1997 Rorty II Richard Rorty Philosophie & die Zukunft Frankfurt 2000 Rorty II (b) Richard Rorty "Habermas, Derrida and the Functions of Philosophy", in: R. Rorty, Truth and Progress. Philosophical Papers III, Cambridge/MA 1998 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (c) Richard Rorty Analytic and Conversational Philosophy Conference fee "Philosophy and the other hgumanities", Stanford Humanities Center 1998 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (d) Richard Rorty Justice as a Larger Loyalty, in: Ronald Bontekoe/Marietta Stepanians (eds.) Justice and Democracy. Cross-cultural Perspectives, University of Hawaii 1997 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (e) Richard Rorty Spinoza, Pragmatismus und die Liebe zur Weisheit, Revised Spinoza Lecture April 1997, University of Amsterdam In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (f) Richard Rorty "Sein, das verstanden werden kann, ist Sprache", keynote lecture for Gadamer’ s 100th birthday, University of Heidelberg In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (g) Richard Rorty "Wild Orchids and Trotzky", in: Wild Orchids and Trotzky: Messages form American Universities ed. Mark Edmundson, New York 1993 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty III Richard Rorty Contingency, Irony, and solidarity, Chambridge/MA 1989 German Edition: Kontingenz, Ironie und Solidarität Frankfurt 1992 Rorty IV (a) Richard Rorty "is Philosophy a Natural Kind?", in: R. Rorty, Objectivity, Relativism, and Truth. Philosophical Papers Vol. I, Cambridge/Ma 1991, pp. 46-62 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty IV (b) Richard Rorty "Non-Reductive Physicalism" in: R. Rorty, Objectivity, Relativism, and Truth. Philosophical Papers Vol. I, Cambridge/Ma 1991, pp. 113-125 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty IV (c) Richard Rorty "Heidegger, Kundera and Dickens" in: R. Rorty, Essays on Heidegger and Others. Philosophical Papers Vol. 2, Cambridge/MA 1991, pp. 66-82 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty IV (d) Richard Rorty "Deconstruction and Circumvention" in: R. Rorty, Essays on Heidegger and Others. Philosophical Papers Vol. 2, Cambridge/MA 1991, pp. 85-106 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty V (a) R. Rorty "Solidarity of Objectivity", Howison Lecture, University of California, Berkeley, January 1983 In Solidarität oder Objektivität?, Stuttgart 1998 Rorty V (b) Richard Rorty "Freud and Moral Reflection", Edith Weigert Lecture, Forum on Psychiatry and the Humanities, Washington School of Psychiatry, Oct. 19th 1984 In Solidarität oder Objektivität?, Stuttgart 1988 Rorty V (c) Richard Rorty The Priority of Democracy to Philosophy, in: John P. Reeder & Gene Outka (eds.), Prospects for a Common Morality. Princeton University Press. pp. 254-278 (1992) In Solidarität oder Objektivität?, Stuttgart 1988 Rorty VI Richard Rorty Truth and Progress, Cambridge/MA 1998 German Edition: Wahrheit und Fortschritt Frankfurt 2000 |
| Measurements | Wittgenstein | Hintikka I 77 Measuring/Wittgenstein/Hintikka: Suppose you use a pole to find out whether the bottom of a river is sandy or muddy. This pole cannot be used in the same way to find out how the pole itself is made. On the other hand, it is easily possible to use the pole to find out whether it can be used to reach the ground at all. >Circular reasoning. I 176 "...when I apply a scale to a spatial object, I apply all scale lines at the same time." I 216 ff Comparability/World/Picture/Measuring/Verification/Wittgenstein/Hintikka: One cannot compare a picture with reality if one cannot measure it against it. You have to be able to put the sentence onto reality. >Reality. I 218 Measuring/Wittgenstein: The scale must always be in the same room as the measured thing. Only the sentences of a physicalistic language can be compared with facts and therefore also represent them. II 29 Measuring/Wittgenstein: Sentence: e.g. "apply the scale three times to G, and you will get the height of G". Then the existence of G and also that of the scale belongs to the sentence. The height of G, on the other hand, is not part of the sentence. II 184 Def Time/Wittgenstein: is what is measured with a clock. >Time. If we know several measurement methods that do not contradict each other, we do not require a specific method to explain these words. But test methods give different meanings to the expression "have the same color" . II 236 Measuring/Wittgenstein: "Greenwich's scale is actually one foot long" there is no point in saying that. It is a definition. Example: Originial Meter. II 238 Measuring/Scale/flexible/Wittgenstein: For example, we could also call all the values read at different temperatures the real length. To the objection that "I like it and I don't like it" which is not a case to which "contradiction" applies, corresponds the objection that a scale is useless unless it is rigid. However, in some cases elasticity may be desired. II 354 ...besides, Russell's equal signs can be eliminated, and in this case the equations cannot be written down at all. Difference: Measuring: e.g. numerical equality of classes or Calculating: e.g. equal number of roots of a 4th degree equation: one is a measurement, the other a calculation. >Calculus. Is there an experiment to determine if two classes have the same number? This may or may not be the case for classes that cannot be overlooked. II 355 It is a damaging prejudice to believe that we are dealing with an experiment when using strokes. II 333 Pattern/Meter/prototype metre/Scale/Measurement/Wittgenstein: the pattern does not belong to the application, but to the language - the existence of the metre measure does not guarantee that there is something that is one metre long. Setting the scale is a decision, not a discovery. II 367 Measuring/Wittgenstein: it is also when I am asked whether two pieces of wood are of equal length, and then I put them on top of each other and answer yes. You might ask if I know that nothing happened to them when they were measured. The answer is that talking about equality is no longer relevant if any method of establishing equality is rejected! II 368 If you claim the equality of lengths, it means that you say something about the method of investigation. >Method, >Comparisons. II 437 Root/Construction/Wittgenstein: For example, the diagonal is turned to the number line. Without the construction √2 is not the length. This length is not an approximation. It has nothing to do with measuring by a meter. III 231 Prototype metre/Wittgenstein/Flor: precisely because it is a paradigm - a means of creating the language game, it is not something that is produced in the language game. >Language game. VI 148 Measuring/Wittgenstein/Schulte: one is to describe the measurement method, another is to find the results. But what we call "measuring" is determined by a certain consistency of measurement results. VI 165 Scale/Measuring/Tractatus/Schulte: the image is linked to reality: it reaches to it. It is like a measure of reality. Only the outermost points of the scale lines touch the object. Thesis: a system of sentences is applied to reality like a yardstick. But I apply all scale lines at the same time. |
W II L. Wittgenstein Wittgenstein’s Lectures 1930-32, from the notes of John King and Desmond Lee, Oxford 1980 German Edition: Vorlesungen 1930-35 Frankfurt 1989 W III L. Wittgenstein The Blue and Brown Books (BB), Oxford 1958 German Edition: Das Blaue Buch - Eine Philosophische Betrachtung Frankfurt 1984 W IV L. Wittgenstein Tractatus Logico-Philosophicus (TLP), 1922, C.K. Ogden (trans.), London: Routledge & Kegan Paul. Originally published as “Logisch-Philosophische Abhandlung”, in Annalen der Naturphilosophische, XIV (3/4), 1921. German Edition: Tractatus logico-philosophicus Frankfurt/M 1960 Hintikka I Jaakko Hintikka Merrill B. Hintikka Investigating Wittgenstein German Edition: Untersuchungen zu Wittgenstein Frankfurt 1996 Hintikka II Jaakko Hintikka Merrill B. Hintikka The Logic of Epistemology and the Epistemology of Logic Dordrecht 1989 |
| Neural Networks | Norvig | Norvig I 761 Neural Networks/Norvig/Russell: Literature on neural networks: Cowan and Sharp (1988b(1), 1988a(2)) survey the early history, beginning with the work of McCulloch and Pitts (1943)(3). John McCarthy has pointed to the work of Nicolas Rashevsky (1936(4), 1938(5)) as the earliest mathematical model of neural learning.) Norbert Wiener, a pioneer of cybernetics and control theory (Wiener, 1948)(6), worked with McCulloch and Pitts and influenced a number of young researchers including Marvin Minsky, who may have been the first to develop a working neural network in hardware in 1951 (see Minsky and Papert, 1988(7), pp. ix–x). Turing (1948)(8) wrote a research report titled Intelligent Machinery that begins with the sentence “I propose to investigate the question as to whether it is possible for machinery to show intelligent behaviour” and goes on to describe a recurrent neural network architecture he called “B-type unorganized machines” and an approach to training them. Unfortunately, the report went unpublished until 1969, and was all but ignored until recently. Frank Rosenblatt (1957)(9) invented the modern “perceptron” and proved the perceptron convergence theorem (1960), although it had been foreshadowed by purely mathematical work outside the context of neural networks (Agmon, 1954(10); Motzkin and Schoenberg, 1954(11)). Some early work was also done on multilayer networks, including Gamba perceptrons (Gamba et al., 1961)(12) and madalines (Widrow, 1962)(13). Learning Machines (Nilsson, 1965)(14) covers much of this early work and more. The subsequent demise of early perceptron research efforts was hastened (or, the authors later claimed, merely explained) by the book Perceptrons (Minsky and Papert, 1969)(15), which lamented the field’s lack of mathematical rigor. The book pointed out that single-layer perceptrons could represent only linearly separable concepts and noted the lack of effective learning algorithms for multilayer networks. The papers in (Hinton and Anderson, 1981)(16), based on a conference in San Diego in 1979, can be regarded as marking a renaissance of connectionism. The two-volume “PDP” (Parallel Distributed Processing) anthology (Rumelhart et al., 1986a)(17) and a short article in Nature (Rumelhart et al., 1986b)(18) attracted a great deal of attention—indeed, the number of papers on “neural networks” multiplied by a factor of 200 between 1980–84 and 1990–94. The analysis of neural networks using the physical theory of magnetic spin glasses (Amit et al., 1985)(19) tightened the links between statistical mechanics and neural network theory - providing not only useful mathematical insights but also respectability. The back-propagation technique had been invented quite early (Bryson and Ho, 1969)(20) but it was rediscovered several times (Werbos, 1974(21); Parker, 1985(22)). The probabilistic interpretation of neural networks has several sources, including Baum and Wilczek (1988)(23) and Bridle (1990)(24). The role of the sigmoid function is discussed by Jordan (1995)(25). Bayesian parameter learning for neural networks was proposed by MacKay Norvig I 762 (1992)(26) and is explored further by Neal (1996)(27). The capacity of neural networks to represent functions was investigated by Cybenko (1988(28), 1989(29)), who showed that two hidden layers are enough to represent any function and a single layer is enough to represent any continuous function. The “optimal brain damage” method (>Artificial neural networks/Norvig) for removing useless connections is by LeCun et al. (1989)(30), and Sietsma and Dow (1988)(31) show how to remove useless units. >Complexity/Norvig. Norvig I 763 For neural nets, Bishop (1995)(32), Ripley (1996)(33), and Haykin (2008)(34) are the leading texts. The field of computational neuroscience is covered by Dayan and Abbott (2001)(35). 1. Cowan, J. D. and Sharp, D. H. (1988b). Neural nets and artificial intelligence. Daedalus, 117, 85–121. 2. Cowan, J. D. and Sharp, D. H. (1988a). Neural nets. Quarterly Reviews of Biophysics, 21, 365–427. 3. McCulloch, W. S. and Pitts, W. (1943). A logical calculus of the ideas immanent in nervous activity. Bulletin of Mathematical Biophysics, 5, 115–137. 4. Rashevsky, N. (1936). Physico-mathematical aspects of excitation and conduction in nerves. In Cold Springs Harbor Symposia on Quantitative Biology. IV: Excitation Phenomena, pp. 90–97. 5. Rashevsky, N. (1938). Mathematical Biophysics: Physico-Mathematical Foundations of Biology. University of Chicago Press. 6. Wiener, N. (1948). Cybernetics. Wiley. 7. Minsky, M. L. and Papert, S. (1988). Perceptrons: An Introduction to Computational Geometry (Expanded edition). MIT Press. 8. Turing, A. (1948). Intelligent machinery. Tech. rep. National Physical Laboratory. reprinted in (Ince, 1992). 9. Rosenblatt, F. (1957). The perceptron: A perceiving and recognizing automaton. Report 85-460-1, Project PARA, Cornell Aeronautical Laboratory. 10. Agmon, S. (1954). The relaxation method for linear inequalities. Canadian Journal of Mathematics, 6(3), 382–392. 11. Motzkin, T. S. and Schoenberg, I. J. (1954). The elaxation method for linear inequalities. Canadian Journal of Mathematics, 6(3), 393–404. 12. Gamba, A., Gamberini, L., Palmieri, G., and Sanna, R. (1961). Further experiments with PAPA. Nuovo Cimento Supplemento, 20(2), 221–231. 13. Widrow, B. (1962). Generalization and information storage in networks of adaline “neurons”. In Self-Organizing Systems 1962, pp. 435–461. 14. Nilsson, N. J. (1965). Learning Machines: Foundations of Trainable Pattern-Classifying Systems. McGraw-Hill. Republished in 1990. 15. Minsky, M. L. and Papert, S. (1969). Perceptrons: An Introduction to Computational Geometry (first edition). MIT Press. 16. Hinton, G. E. and Anderson, J. A. (1981). Parallel Models of Associative Memory. Lawrence Erlbaum Associates. 17. Rumelhart, D. E., Hinton, G. E., andWilliams, R. J. (1986a). Learning internal representations by error propagation. In Rumelhart, D. E. and McClelland, J. L. (Eds.), Parallel Distributed Processing, Vol. 1, chap. 8, pp. 318–362. MIT Press. 18. Rumelhart, D. E., Hinton, G. E., and Williams, R. J. (1986b). Learning representations by back propagating errors. Nature, 323, 533–536. 19. Amit, D., Gutfreund, H., and Sompolinsky, H. (1985). Spin-glass models of neural networks. Physical Review, A 32, 1007–1018. 20. Bryson, A. E. and Ho, Y.-C. (1969). Applied Optimal Control. Blaisdell. 21. Werbos, P. (1974). Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences. Ph.D. thesis, Harvard University. 22. Parker, D. B. (1985). Learning logic. Technical report TR-47, Center for Computational Research in Economics and Management Science, Massachusetts Institute of Technology. 23. Baum, E. and Wilczek, F. (1988). Supervised learning of probability distributions by neural networks. In Anderson, D. Z. (Ed.), Neural Information Processing Systems, pp. 52–61. American Institute of Physics. 24. Bridle, J. S. (1990). Probabilistic interpretation of feedforward classification network outputs, with relationships to statistical pattern recognition. In Fogelman Souli´e, F. and H´erault, J. (Eds.), Neuro computing: Algorithms, Architectures and Applications. Springer-Verlag. 25. Jordan, M. I. (1995). Why the logistic function? a tutorial discussion on probabilities and neural networks. Computational cognitive science technical report 9503, Massachusetts Institute of Technology. 26. MacKay, D. J. C. (1992). A practical Bayesian framework for back-propagation networks. Neural Computation, 4(3), 448–472. 27. Neal, R. (1996). Bayesian Learning for Neural Networks. Springer-Verlag. 28. Cybenko, G. (1988). Continuous valued neural networks with two hidden layers are sufficient. Technical report, Department of Computer Science, Tufts University. 29. Cybenko, G. (1989). Approximation by superpositions of a sigmoidal function. Mathematics of Controls, Signals, and Systems, 2, 303–314. 30. LeCun, Y., Jackel, L., Boser, B., and Denker, J. (1989). Handwritten digit recognition: Applications of neural network chips and automatic learning. IEEE Communications Magazine, 27(11), 41– 46. 31. Sietsma, J. and Dow, R. J. F. (1988). Neural net pruning - Why and how. In IEEE International Conference on Neural Networks, pp. 325–333. 32. Bishop, C. M. (1995). Neural Networks for Pattern Recognition. Oxford University Press. 33. Ripley, B. D. (1996). Pattern Recognition and Neural Networks. Cambridge University Press. 34. Haykin, S. (2008). Neural Networks: A Comprehensive Foundation. Prentice Hall. 35. Dayan, P. and Abbott, L. F. (2001). Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems. MIT Press. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| New Growth Theory | Economic Theories | Kurz I 260 New Growth Theory/Economic theories/Kurz: (…) [there are] reasons for the recent resumption of (some special form of) long-period analysis in 'new' growth theory. Until a few decades ago, the number of commodities and, as a consequence, the time horizon in intertemporal general equilibrium theory was assumed to be finite and, therefore, arbitrary. Problem: The principal objection to the restriction to a finite number of goods is that it requires a finite horizon and there is no natural way to choose the final period. Moreover, since there will be terminal stocks in the final period there is no natural way to value them without contemplating future periods in which they will be used (McKenzie 1987: 507)(1) The introduction of an infinite horizon turned out to be critical (see also Burgstaller 1994: 43-8)(2). It pushed the analysis inevitably towards the long period, albeit only in the very special sense of steady state. Endogenous growth/Robert Lucas: This was clearly spelled out, for instance, by Robert Lucas in a contribution to the theories of endogenous growth. He observed that [Flor any initial capital K(O) > 0, the optimal capital-consumption path (K(t), c(t)) will converge to the balanced path asymptotically. That is, the balanced path will be a good approximation to any actual path 'most' of the time [and that] this is exactly the reason why the balanced path is interesting to us. (Lucas 1988: 11)(3) Lucas thus advocated a (re-)switching from an intertemporal analysis to a steady-state one. Since the balanced path of the intertemporal model is the only path analysed by Lucas, in the perspective under consideration the intertemporal model may be regarded simply as a step toward obtaining a rigorous steady-state setting. Neoclassical economics: Moreover, Lucas abandoned one of the characteristic features of all neoclassical theories, that is, income distribution is determined by demand and supply of factors of production. If we concentrate on the balanced path, capital in the initial period cannot be taken as given along with other 'initial endowments'. Since distribution cannot be determined by demand and supply of capital and labour, in Lucas's model it is determined in the following way. Labour is considered the vehicle of 'human capital', that is, a producible factor. Hence all factors are taken to be producible and the rate of profits is determined as in Chapter II of Production of Commodities by Means of Commodities (Sraffa 1960)(4). At the beginning of that chapter (S 4—5), wages Kurz I 261 are regarded as entering the system 'on the same footing as the fuel for the engines or the feed for the cattle'. In this case the rate of profits and prices are determined by the socio-technical conditions of production alone - the 'methods of production and productive consumption' (Sraffa 1960:3)(4). The introduction of several alternative processes of production does not change the result. >P. Sraffa, >Production theory, cf. >Growth/Neoclassical Economics, >Exogenous Growth/Neoclassical Economics, >Endogenous growth. 1. McKenzie, L. W. (1987) 'General Equilibrium', The New Palgrave A Dictionary of Economics, edited by J. Eatwell, P. Newman and M. Milgate, London: Macmillan, vol. 11: 498-512. 2. Burgstaller, A. (1994) Property and Prices. Toward a Unified Theory of Value, Cambridge: Cambridge University Press. 3. Lucas, R. E. (1988) 'On the Mechanisms of Economic Development', Journal of Monetary Economics, 22: 3-42. 4. Sraffa, P. (1960) Production of Commodities hy Means of Commodities, Cambridge: Cambridge University Press. Kurz, Heinz D. and Salvadori, Neri. „Endogenous growth in a stylised 'classical' model“. In: Kurz, Heinz; Salvadori, Neri 2015. Revisiting Classical Economics: Studies in Long-Period Analysis (Routledge Studies in the History of Economics). London, UK: Routledge. |
Kurz I Heinz D. Kurz Neri Salvadori Revisiting Classical Economics: Studies in Long-Period Analysis (Routledge Studies in the History of Economics). Routledge. London 2015 |
| Parameterization | Meteorology | Edwards I 393 Parameterization/meteorology/climatology/Edwards: far from expressing pure theory, analysis models are data-laden.(1) And the same can also be said of all forecast models and general circulation models. Stephen Schneider writes: . . . even our most sophisticated ‘first principles’ models contain ‘empirical statistical’ elements within the model structure. . . .We can describe the known physical laws mathematically, at least in principle. In practice, however, solving these equations in full, explicit detail is impossible. First, the possible scales of motion in the atmospheric and oceanic components range from the submolecular to the global. Second are the interactions of energy transfers among the different scales of motion. Finally, many scales of disturbance are inherently unstable; small disturbances, for example, grow rapidly in size if conditions are favorable.(2) Edwards: Hence the necessity of parameterization, much of which can be described as the integration of observationally derived approximations into the “model physics.” Schneider and others sometimes refer to parameters as “semi-empirical,” an apt description that highlights their fuzzy relationship with observational data. For the foreseeable future, all analysis models, forecast models, and climate models will contain many “semi-empirical” elements. >Wheather forecasting/Edwards, >Models/meteorology, cf. >Homogenization/climatology, >Reanalysis/climatology. Edwards I 465 Parameter: (…) the term is often used to distinguish, from dependent variables, quantities that may be more or less arbitrarily assigned values for purposes of the problem at hand” (emphasis added). So a parameter is a kind of proxy - a stand-in for something that cannot be modeled directly but can still be estimated, or at least guessed. Parameterization illustrates the interaction of computational friction with the limits of human knowledge. In an ideal climate model, the only fixed conditions would be the distribution and the altitude of continental surfaces. Virtually all other variables - sea-surface temperature, land-surface albedo (reflectance), cloud formation, etc. - would be generated internally by the model itself from the lower-level physical properties of air, water, and other basic elements of the climate system. Instead, most physical processes operating in the atmosphere require some degree of parameterization; these parameterized processes are known as the “model physics.” >Models/climatology. Parameter: (…) parameters represent a variable physical process rather than a fixed quantity. Edwards I 466 Parameterization/Example: A major parameterization in all climate models is radiative transfer. The atmosphere contains both gases (CO2, methane, nitrogen, ozone, oxygen, water vapor, etc.) and solids (particulate aerosols, ice clouds, etc.). Each one of these materials absorbs solar energy at particular frequencies. Each also emits radiation at other frequencies. Those emissions are then absorbed and re-radiated by other gases and solids. These radiative transfers play a huge role in governing the atmosphere’s temperature. Thus, models must somehow estimate how much radiation the atmosphere in a given grid box absorbs, reflects, and transmits, at every level and horizontal location. “Line-by-line models,” which combine databases of spectrographic measurements for the various gases with physical models, can carry out this summing.(3) Edwards I 469 Ad hoc parameter/example: An example of an ad hoc parameter is “flux adjustment” in coupled atmosphere-ocean circulation models (AOGCMs). The interface between the atmospheric model and the ocean model must represent exchanges of heat, momentum (wind and surface resistance), and water (precipitation, evaporation) between the atmosphere and the ocean. These fluxes - flows of energy and matter between atmosphere and ocean—are very difficult to measure empirically. Yet they profoundly affect model behavior. Modelers spoke of flux adjustments as “non-physical” parameterizations - i.e., ones not based on physical theory—but also sometimes characterized them as “empirically determined.”(4) Any given GCM’s model physics contains hundreds or even thousands of parameterizations. Edwards I 470 An entire subfield—climate model diagnosis - works out ways to isolate the origin of particular problems to specific parameterizations and their interactions. Tuning: “Tuning” means adjusting the values of coefficients and even, sometimes, reconstructing equations in order to produce a better overall model result. “Better” may mean that the result agrees more closely with observations, or that it corresponds more closely to the modeler’s expert judgment about what one modeler I interviewed called the “physical plausibility” of the change. >Models/climatology. 1. P. N. Edwards, “Global Climate Science, Uncertainty and Politics: Data-Laden Models, Model-Filtered Data,” Science as Culture 8, no. 4 (1999): 437–. 2. S. H. Schneider, “Introduction to Climate Modeling,” in Climate System Modeling, ed. K. E. Trenberth (Cambridge University Press, 1992). 3. J. T. Kiehl, “Atmospheric General Circulation Modeling,” in Climate System Modeling, ed. K. E. Trenberth (Cambridge University Press, 1992), 338. 4. 8. J. T. Houghton et al., Climate Change 1995: The Science of Climate Change (Cambridge University Press, 1996). |
Edwards I Paul N. Edwards A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming Cambridge 2013 |
| Predication | Geach | I 52 Naming/Denotation/Two-Names Theory/GeachVsAristoteles: Incorrect approximation of predication and naming: as if predicates were (complex) names : "on the mat". >Naming, >Predication, >Attribution, >Names, >Predicates, >Aristotle. ((s) "The man stabbing Caesar to death stabbed the one stabbed by Brutus.") Geach: Additionally, Geach would use a link. Two-names theory/Aristotle/Geach: "Socrates is a philosopher" should be true because the thing is named. GeachVs: "Philosopher" (general term) is not a name for "all (or every) philosopher". >General terms. I 70 Contradictory predications like "Fa" and "~Fa" refer to a common subject - there are no "contradictory subjects". >Beliefs/Geach, >Belief ascription. I 252 Predication/Geach: can be done without naming: in an if-clause or in an or-clause, a term P can be predicated from a thing without naming the thing with "P". E.g. "If that what the police officer has said, is true, then he was driving faster than 60". This does not name the police officer's sentence as true. - (> Conditional). - Predication/naming: century-old error: the predicate is predicated from the thing - Frege: Difference >Designation, >Denotation, >Predication: to name a thing "P", a sentence must be asserted! But a property is also predicated in a non-asserting sub-clause (subset) - therefore, naming must be explained by predication, not vice versa. I 290 Predication/Geach: wrong: to read "SiP" as "a thing is a predicate" - (origin: "Two-Names-Theory, Aristotle) a subject cannot be negated - sentence negation: negation of the predicate. I 291 GeachVsAristoteles: Vs "Two-Names-Theory"/TNT: confuses the relation of names to the named with relation of the predicate to what it is stated from -> false doctrine of the Trinity. I 295 Prediction/Theology/Thomas Aquinas: the expression after "as" is predicative: e.g., "Christ inasmuch as he is human". - Distinction between subject and predicate, VsTwo-Name-Theory. >Thomas Aquinas. False: Two-Name-Theory: E.g.: "His godly nature is immortal, his human nature is mortal". Aquinas: can distinguish "Christ as human is God". False: Two-name-theory: cannot do this because "human" and "God" are merely two names. >Two name theory: see above). VsOckham: E.g. for him, "humanitas" is not more than "majesty": a disguised name for a concrete thing. Problem: for Ockham, humanity is no longer human when viewed as the Son of God. VsOckham: because it is not genuinely abstract, e.g. the mayor's office becomes the mayor. >William of Ockham. I 300 Predication/Thomas Aquinas/Geach: Subject: refers to a suppositum (an "accepted") predicate: refers to a form or nature. Predication: unequal naming: E.g. "The Prime Minister became Prime Minister" - Nonsense: "which Prime Minister?". |
Gea I P.T. Geach Logic Matters Oxford 1972 |
| Principles | Feynman | I 355 Principle/Theory/Feynman: Principles can explain laws. According to what principle should the law of Snellius be clear at all? >Laws, >Natural laws. I 355 Fermat's Principle of the Shortest Time/Refraction/Foucault: from all possible paths the light could take, it takes the path that requires the shortest time. I 356 The path that leads to the object is chosen, if the mirror had not been present. (~). I 357 Refraction: Transition from air to water: e.g. (Feynman): Suppose someone stands on the shore and sees someone falling from a boat into the water. We can run much faster than we can swim. Therefore, we should choose a path with a longer route on land than the distance in the water. Exactly the same as light does! Important: there is an area near the optimum point where, in the first approximation, there is no significant change in time. But in the second order. >Refractive index. I 358 When the ratio of the speed is n, light should penetrate at such an angle that the ratio of the sinuses of the angles Θi und Θr is the ratio of the velocities of the two media. sin Θi = n sin Θr. I 358 Principle of the Shortest Time/Fermat/Feynman: Conclusions: 1) Reversibility of direction. 2) E.g. the glass block displaces the beam parallel, since the exit angle is equal to the entrance angle. E.g. sunset: the light moves slower in the denser air on the ground. The light therefore chooses the shorter path by an acute angle at the entrance. This means that we can still see the sun, even when it has already gone down. I 359 Bsp mirage I 361 Analogy/Analog/Theory/Principle of the Shortest Time/Fermat/Identity/Physics/Feynman: apparently, the principle of the shortest time and the statement that angles are equal in reflection, and that the sinuses are proportional to the angles at refraction, are the same. >Analogies, >Comparisons, >Comparability. |
Feynman I Richard Feynman The Feynman Lectures on Physics. Vol. I, Mainly Mechanics, Radiation, and Heat, California Institute of Technology 1963 German Edition: Vorlesungen über Physik I München 2001 Feynman II R. Feynman The Character of Physical Law, Cambridge, MA/London 1967 German Edition: Vom Wesen physikalischer Gesetze München 1993 |
| Quantum Field Theory | Kanitscheider | I 172 Quantum field theory/Kanitscheider: here one works, if one assumes curved space-time, with an approximation, where the metric field is not quantized itself, but is used as classical space-time background arena, i.e. the reaction of the matter fields on the space-time is neglected. But this already reveals that the interaction between matter fields and geometry produces particles from the vacuum! >Quantum mechanics, >Relativity theory, >Gravitation/Einstein, >Space-time/Einstein. |
Kanitsch I B. Kanitscheider Kosmologie Stuttgart 1991 Kanitsch II B. Kanitscheider Im Innern der Natur Darmstadt 1996 |
| Reasoning | Minsky | I 187 Reasoning/common sense/Artificial Intelligence/Minsky: Here's a rule that's part of ordinary common sense: If A depends on B, and B depends on C, then — clearly — A depends on C. The chain containing walk, drive, and fly may appear to use several different kinds of links. But although they differ in regard to vehicles, they all refer to paths through space. For generations, scientists and philosophers have tried to explain ordinary reasoning in terms of logical principles — with virtually no success. I suspect this enterprise failed because it was looking in the wrong direction: common sense works so well not because it is an approximation of logic; logic is only a small part of our great accumulation of different, useful ways to chain things together. >Logic/Minsky. |
Minsky I Marvin Minsky The Society of Mind New York 1985 Minsky II Marvin Minsky Semantic Information Processing Cambridge, MA 2003 |
| Reinforcement Learning | AI Research | Norvig I 831 Reinforcement Learning/AI Research/Norvig/Russell: In many complex domains, reinforcement learning [by reward and punishment] is the only feasible way to train a program to perform at high levels. For example, in game playing, it is very hard for a human to provide accurate and consistent evaluations of large numbers of positions, which would be needed to train an evaluation function directly from examples. Instead, the program can be told when it has won or lost, and it can use this information to learn an evaluation function that gives reasonably accurate estimates of the probability of winning from any given position. Similarly, it is extremely difficult to program an agent to fly a helicopter; yet given appropriate negative rewards for crashing, wobbling, or deviating from a set course, an agent can learn to fly by itself. A. Passive Reinforcement learning Situation: an agent is placed in an environment and must learn to behave successfully therein. A utility-based agent learns a utility function on states and uses it to select actions that maximize the expected outcome utility. A Q-learning agent learns an action-utility function, or Q-function, giving the expected utility of taking a given action in a given state. A reflex agent learns a policy that maps directly from states to actions. Exploration: an agent must experience as much as possible of its environment in order to learn how to behave in it. >Markov decision processes/Norvig. Norvig I 833 Passive reinforcement learning: A simple method for direct utility estimation was invented in the late 1950s in the area of adaptive control theory by Widrow and Hoff (1960)(1). The idea is that the utility of a state is the expected total reward from that state onward (called the expected reward-to-go), and each trial provides a sample of this quantity for each state visited. Utility: the utilities of states are not independent! The utility of each state equals its own reward plus the expected utility of its successor states. That is, the utility values obey the Bellman equations for a fixed policy. (>Values/AI Research). Problem: By ignoring the connections between states, direct utility estimation misses opportunities for learning. Norvig I 834 Adaptive Dynamic Programming /ADP: An adaptive dynamic programming (or ADP) agent takes advantage of the constraints among the utilities of states by learning the transition model that connects them and solving the corresponding Markov decision process using a dynamic programming method. Alternatively, we can adopt the approach of modified policy iteration (…), using a simplified value iteration process to update the utility estimates after each change to the learned model. Norvig I 836 Temporal difference learning/TD: All temporal-difference methods work by adjusting the utility estimates towards the ideal equilibrium that holds locally when the utility estimates are correct. Norvig I 839 B. Active reinforcement learning: A passive learning agent has a fixed policy that determines its behavior. An active agent must decide what actions to take. First, the agent will need to learn a complete model with outcome probabilities for all actions, (…). Next, we need to take into account the fact that the agent has a choice of actions. The utilities it needs to learn are those defined by the optimal policy; they obey the >Bellman equations (…).The final issue is what to do at each step. Having obtained a utility function U that is optimal for the learned model, the agent can extract an optimal action by one-step look-ahead to maximize the expected utility; alternatively, if it uses policy iteration, the optimal policy is already available, so it should simply execute the action the optimal policy recommends. Norvig I 843 Q-Learning: There is an alternative TD method, called Q-learning, which learns an action-utility representation instead of learning utilities. [A] TD [temporal difference] agent that learns a Q-function does not need a model of the form P(s’| s, a), either for learning or for action selection. For this reason, Q-learning is called a model-free method. Norvig I 845 Function approximation: simply means using any sort of representation for the Q-function other than a lookup table. The representation is viewed as approximate because it might not be the case that the true utility function or Q-function can be represented in the chosen form. Norvig I 846 Generalization: The compression achieved by a function approximator allows the learning agent to generalize from states it has visited to states it has not visited. That is, the most important aspect of function approximation is not that it requires less space, but that it allows for inductive generalization over input states. Norvig I 848 Policies: a policy π is a function that maps states to actions. (…) we could represent π by a collection of parameterized Q-functions, one for each action, and take the action with the highest predicted value (…).if the policy is represented by Q-functions, then policy search results in a process that learns Q-functions. This process is not the same as Q-learning! In Q-learning with function approximation, the algorithm finds a value of θ such that ˆQ θ is “close” to Q ∗, the optimal Q-function. Policy search: Policy search, on the other hand, finds a value of θ that results in good performance; (…). VsPolicy search: Problems: One problem with policy representations of the kind (…) is that the policy is a discontinuous function of the parameters when the actions are discrete. Solution: This means that the value of the policy may also change discontinuously, which makes gradient-based search difficult. For this reason, policy search methods often use a stochastic policy representation πθ(s, a), which specifies the probability of selecting action a in state s. Norvig I 854 History of reinforcement learning: Turing (1948(2), 1950(3)) proposed the reinforcement-learning approach, although he was not convinced of its effectiveness, writing, “the use of punishments and rewards can at best be a part of the teaching process.” Arthur Samuel’s work (1959)(4) was probably the earliest successful machine learning research. Around the same time, researchers in adaptive control theory (Widrow and Hoff, 1960)(1), building on work by Hebb (1949)(5), were training simple networks using the delta rule. The cart–pole work of Michie and Chambers (1968)(6) can also be seen as a reinforcement learning method with a function approximator. The psychological literature on reinforcement learning is much older; Hilgard and Bower (1975)(7) provide a good survey. Neuroscience: The neuroscience text by Dayan and Abbott (2001)(8) describes possible neural implementations of temporal-difference learning, while Dayan and Niv (2008)(9) survey the latest evidence from neuroscientific and behavioral experiments. Markov decision process: The connection between reinforcement learning and Markov decision processes was first made by Werbos (1977)(10), but the development of reinforcement learning in AI stems from work at the University of Massachusetts in the early 1980s (Barto et al., 1981)(11). The paper by Sutton (1988) provides a good historical overview. Temporal difference learning: The combination of temporal-difference learning with the model-based generation of simulated experiences was proposed in Sutton’s DYNA architecture (Sutton, 1990)(12). The idea of prioritized sweeping was introduced independently by Moore and Atkeson (1993)(13) and Norvig I 855 Peng and Williams (1993)(14). Q-learning: was developed in Watkins’s Ph.D. thesis (1989)(15), while SARSA appeared in a technical report by Rummery and Niranjan (1994)(16). Function approximation: Function approximation in reinforcement learning goes back to the work of Samuel, who used both linear and nonlinear evaluation functions and also used feature-selection methods to reduce the feature CMAC space. Later methods include the CMAC (Cerebellar Model Articulation Controller) (Albus, 1975)(17), which is essentially a sum of overlapping local kernel functions, and the associative neural networks of Barto et al. (1983)(18). Neural networks are currently the most popular form of function approximator. The best-known application is TD-Gammon (Tesauro, 1992(19), 1995(20)), (…). Policy search: Policy search methods were brought to the fore by Williams (1992(21)), who developed the REINFORCE family of algorithms. Later work by Marbach and Tsitsiklis (1998)(22), Sutton et al. (2000)(23), and Baxter and Bartlett (2000)(24) strengthened and generalized the convergence results for policy search. The method of correlated sampling for comparing different configurations of a system was described formally by Kahn and Marshall (1953)(25), but seems to have been known long before that. Its use in reinforcement learning is due to Van Roy (1998)(26) and Ng and Jordan (2000)(27); the latter paper also introduced the PEGASUS algorithm and proved its formal properties. Norvig I 857 Inverse reinforcement learning: Russell (1998)(28) describes the task of inverse reinforcement learning - figuring out what the reward function must be from an example path through that state space. This is useful as a part of apprenticeship learning, or as a part of doing science—we can understand an animal or robot by working backwards from what it does to what its reward function must be. Cf. >Learning, >Generalization, >Understanding. 1. Widrow, B. and Hoff, M. E. (1960). Adaptive switching circuits. In 1960 IRE WESCON Convention Record, pp. 96–104. 2. Turing, A. (1948). Intelligent machinery. Tech. rep. National Physical Laboratory. reprinted in (Ince, 1992). 3. Turing, A. (1950). Computing machinery and intelligence. Mind, 59, 433–460. 4. Samuel, A. L. (1959). Some studies in machine learning using the game of checkers. IBM Journal of Research and Development, 3(3), 210–229. 5. Hebb, D. O. (1949). The Organization of Behavior. Wiley. 6. Michie, D. and Chambers, R. A. (1968). BOXES: An experiment in adaptive control. In Dale, E. and Michie, D. (Eds.), Machine Intelligence 2, pp. 125–133. Elsevier/North-Holland. 7. Hilgard, E. R. and Bower, G. H. (1975). Theories of Learning (4th edition). Prentice-Hall. 8. Dayan, P. and Abbott, L. F. (2001). Theoretical Neuroscience: Computational and Mathematical Modeling of Neural Systems. MIT Press. 9. Dayan, P. and Niv, Y. (2008). Reinforcement learning and the brain: The good, the bad and the ugly. Current Opinion in Neurobiology, 18(2), 185–196. 10. Werbos, P. (1977). Advanced forecasting methods for global crisis warning and models of intelligence. General Systems Yearbook, 22, 25–38. 11. Barto, A. G., Sutton, R. S., and Brouwer, P. S. (1981). Associative search network: A reinforcement learning associative memory. Biological Cybernetics, 40(3), 201–211. 12. Sutton, R. S. (1990). Integrated architectures for learning, planning, and reacting based on approximating dynamic programming. In ICML-90, pp. 216–224. 13. Moore, A. W. and Atkeson, C. G. (1993). Prioritized sweeping—Reinforcement learning with less data and less time. Machine Learning, 13, 103–130. 14. Peng, J. and Williams, R. J. (1993). Efficient learning and planning within the Dyna framework. Adaptive Behavior, 2, 437–454. 15. Watkins, C. J. (1989). Models of Delayed Reinforcement Learning. Ph.D. thesis, Psychology Department, Cambridge University. 16. Rummery, G. A. and Niranjan, M. (1994). Online Q-learning using connectionist systems. Tech. rep. CUED/F-INFENG/TR 166, Cambridge University Engineering Department. 17. Albus, J. S. (1975). A new approach to manipulator control: The cerebellar model articulation controller (CMAC). J. Dynamic Systems, Measurement, and Control, 97, 270–277. 18. Barto, A. G., Sutton, R. S., and Anderson, C. W. (1983). Neuron-like adaptive elements that can solve difficult learning control problems. IEEE Transactions on Systems, Man and Cybernetics, 13, 834– 846. 19. Tesauro, G. (1992). Practical issues in temporal difference learning. Machine Learning, 8(3–4), 257– 277. 20. Tesauro, G. (1995). Temporal difference learning and TD-Gammon. CACM, 38(3), 58–68. 21. Williams, R. J. (1992). Simple statistical gradient following algorithms for connectionist reinforcement learning. Machine Learning, 8, 229–256. 22. Marbach, P. and Tsitsiklis, J. N. (1998). Simulation based optimization of Markov reward processes. Technical report LIDS-P-2411, Laboratory for Information and Decision Systems, Massachusetts Institute of Technology. 23. Sutton, R. S., McAllester, D. A., Singh, S. P., and Mansour, Y. (2000). Policy gradient methods for reinforcement learning with function approximation. In Solla, S. A., Leen, T. K., andM¨uller, K.-R. (Eds.), NIPS 12, pp. 1057–1063. MIT Press. 24. Baxter, J. and Bartlett, P. (2000). Reinforcement learning in POMDP’s via direct gradient ascent. In ICML-00, pp. 41–48. 25. Kahn, H. and Marshall, A. W. (1953). Methods of reducing sample size in Monte Carlo computations. Operations Research, 1(5), 263–278. 26. Van Roy, B. (1998). Learning and value function approximation in complex decision processes. Ph.D. thesis, Laboratory for Information and Decision Systems, MIT. 27. Ng, A. Y. and Jordan, M. I. (2000). PEGASUS: A policy search method for large MDPs and POMDPs. In UAI-00, pp. 406–415. 28. Russell, S. J. (1998). Learning agents for uncertain environments (extended abstract). In COLT-98, pp. 101–103. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Road Pricing | Economic Theories | Mause I 466f Road Pricing/Economic Theories: A basic microeconomic principle is that any relative price change has an impact on the consumption decisions of individual economic actors and triggers substitution processes. Therefore, if a road user has to pay a fee in addition to his time and operating costs for the use of a section of road during a congestion-prone period (e.g. rush hour traffic), there is an incentive to switch to an alternative point in time, an alternative route or an alternative means of transport. The optimum amount of this additional fee corresponds exactly to the additional costs for all other road users caused by each additional road user on an overused road section (cf. Yan and Lam 1996, p. 319) (1). Tolls are not about rivalry in consumption, but the problem of non-excludability. >Externalities. Since a toll changes the price ratio between different road sections or means of transport, substitution processes can be triggered for congested road sections. Mause I 467 Problems: 1. the monetary evaluation of the time costs of road users, 2. complex interdependencies within the transport network. As a result, the theoretically best solution (first-best) cannot be implemented in practice and only an approximation (second-best) can be achieved. (cf. Samll and Verhoef 2007, p. 137) (2). Value: can be divided into two aspects: a) value of the reduction in travel time, b) value of reliability in relation to the time used. (Carrion and Levinson, 2012, p. 721. (3)) Road user fee/Toll: can be seen as a product differentiation. The consumer wonders how much he/she is prepared to pay for a more convenient product variant. (see Small & Yan, 2001, p. 311). >Preference. Mause I 468 Welfare increase: Tolls can lead to an increase in welfare and must therefore be supported from an economic point of view. Problem: the introduction creates redistributive effects. If the revenue of the fee is not redistributed within road users, but goes to the state, there are (...) "losers" through the introduction of a road user charge. 1. Yan, Hai, und William H. K. Lam. 1996. Optimal road tolls under conditions of queueing and congestion. Transportation Research Part A: Policy and Practice 30 (5): 319– 332. 2. Small, Kenneth A., und Erik T. Verhoef..The economics of urban transportation. London/ New York 2007. 3. Carrion, Carlos, und David Levinson. 2012. Value of travel time reliability: A review of current evidence. Transportation Research Part A: Policy and Practice 46 (4): 720– 741. |
Mause I Karsten Mause Christian Müller Klaus Schubert, Politik und Wirtschaft: Ein integratives Kompendium Wiesbaden 2018 |
| Science | Duhem | I 294 Science/Logic/Duhem: Which logical conditions must be fulfilled in the hypothesis selection? Does the logic require that our hypothesis be, or at least be in harmony with a cosmological system? Not at all! Our physical theories do not place their pride in being explanations (metaphysical). They are not assumptions about the actual nature of material things. They have only the economic summary and classification of the experimental laws as their goal. They are independent and independent of any metaphysical system. >Systems, >Models, >Theories, >Observation, >Experiments. I 295 Does logic require that the hypotheses by induction be generalized experimental laws? No: logic cannot make impossible demands. We are therefore not reluctant to include postulates under the foundations of physics which were not supplied by the experiment. Logic also does not dictate us to introduce our hypotheses individually and to control them individually before application. That would be absurd because of holism. What conditions must now be fulfilled according to logic in the choice of hypotheses? 1. No self-contradictory theorem is acceptable. 2. No reciprocal contradiction. 3. The conclusions which the mathematical derivation can draw from the totality are, to represent the totality of the experimental laws with reasonable approximation. >Idealization. |
Duh I P. Duhem La théorie physique, son objet et sa structure, Paris 1906 German Edition: Ziel und Struktur der physikalischen Theorien Hamburg 1998 |
| Sequential Decision Making | Norvig | Norvig I 645 Sequential Decision Making/AI research/Norvig/Russell: [this is about] the computational issues involved in making decisions in a stochastic environment. Sequential decision problems incorporate utilities, uncertainty, and sensing, and include search and planning problems as special cases. >Planning/Norvig, >Decision networks/Norvig, >Decision theory/AI Research, >Utility/AI Research, >Utility theory/Norvig, >Environment/AI research, >Multi-attribute utility theory/AI research. Norvig I 649 Optimal policy: the optimal policy for a finite horizon is non-stationary. With no fixed time limit, on the other hand, there is no reason to behave differently in the same state at different times. Hence, the optimal action depends only on the current state, and the optimal policy is stationary. States: In the terminology of multi-attribute utility theory, each state si can be viewed as an attribute of the state sequence [s0, s1, s2 . . .]. >Values/AI research. Norvig I 684 Sequential decision problems in uncertain environments, also called Markov decision processes, or MDPs, are defined by a transition model specifying the probabilistic outcomes of actions and a reward function specifying the reward in each state. Norvig I 685 Richard Bellman developed the ideas underlying the modern approach to sequential decision problems while working at the RAND Corporation beginning in 1949. (…) Bellman’s book, Dynamic Programming (1957)(1), gave the new field a solid foundation and introduced the basic algorithmic approaches. Ron Howard’s Ph.D. thesis (1960)(2) introduced policy iteration and the idea of average reward for solving infinite-horizon problems. Several additional results were introduced by Bellman and Dreyfus (1962)(3). Modified policy iteration is due to van Nunen (1976)(4) and Puterman and Shin (1978)(5). Asynchronous policy iteration was analyzed by Williams and Baird (1993)(6) (…). The analysis of discounting in terms of stationary preferences is due to Koopmans (1972)(7). The texts by Bertsekas (1987)(8), Puterman (1994)(9), and Bertsekas and Tsitsiklis (1996)(10) provide a rigorous introduction to sequential decision problems. Papadimitriou and Tsitsiklis (1987)(11) describe results on the computational complexity of MDPs. Seminal work by Sutton (1988)(12) and Watkins (1989)(13) on reinforcement learning methods for solving MDPs played a significant role in introducing MDPs into the AI community, as did the later survey by Barto et al. (1995)(14). >Markov Decision Processes/Norvig. 1. Bellman, R. E. (1957). Dynamic Programming. Princeton University Press 2. Howard, R. A. (1960). Dynamic Programming and Markov Processes. MIT Press. 3. Bellman, R. E. and Dreyfus, S. E. (1962). Applied Dynamic Programming. Princeton University Press. 4. van Nunen, J. A. E. E. (1976). A set of successive approximation methods for discounted Markovian decision problems. Zeitschrift fur Operations Research, Serie A, 20(5), 203–208. 5. Puterman, M. L. and Shin, M. C. (1978). Modified policy iteration algorithms for discounted Markov decision problems. Management Science, 24(11), 1127-1137. 6. Williams, R. J. and Baird, L. C. I. (1993). Tight performance bounds on greedy policies based on imperfect value functions. Tech. rep. NU-CCS-93-14, College of Computer Science, Northeastern University. 7. Koopmans, T. C. (1972). Representation of preference orderings over time. In McGuire, C. B. and Radner, R. (Eds.), Decision and Organization. Elsevier/North-Holland. 8. Bertsekas, D. (1987). Dynamic Programming: Deterministic and Stochastic Models. Prentice-Hall. 9. Puterman, M. L. (1994). Markov Decision Processes: Discrete Stochastic Dynamic Programming. Wiley 10. Bertsekas, D. and Tsitsiklis, J. N. (1996). Neurodynamic programming. Athena Scientific. 11. Papadimitriou, C. H. and Tsitsiklis, J. N. (1987). The complexity of Markov decision processes. Mathematics of Operations Research, 12(3), 441-450. 12. Sutton, R. S. (1988). Learning to predict by the methods of temporal differences. Machine Learning, 3, 9-44. 13. Watkins, C. J. (1989). Models of Delayed Reinforcement Learning. Ph.D. thesis, Psychology Department, Cambridge University. 14. Barto, A. G., Bradtke, S. J., and Singh, S. P. (1995). Learning to act using real-time dynamic programming. AIJ, 73(1), 81-138. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Terminology | Geach | I 52 Two-Names-Theory/TNT/GeachVsAristoteles: false approximation of predication and naming: as if predicates were (complex) names: "on the mat") - ((s) E.g. "The person who stabbed Ceasar stabbed the one stabbed by Brutus"). Geach: besides, a binding element would be needed. Two-name-theory: "Socrates is a philosopher" is supposed to be true, because the same thing is named. Vs: "philosopher" (general term) is not a name for "all (or any) philosopher". I 54 Two-Classes-Theory/TCT/GeachVs: even worse than the Two-Name-Theory: the general term "philosopher" means "class of philosophers" - Socrates is then only a part of the class. Vs: The element-relation is quite different from the subclass-relation: E.g. a parliamentary committee is not a member of Parliament. But: "is a philosopher" means exactly the same in both applications. - copula: fallacy of division: as if there were two varieties of "is": one for "is a philosopher," and one for "is an element of the class of philosophers" - Geach: equivalent sentences need not be able to be divided into equivalent sub-sets - "every logician" is not equivalent to "class of logicians". I 122 Latin prose theory/Geach: the relative pronoun is treated as a connection of a binding word with a bound pronoun: "the" is translated as "so that": E.g. the king sent emissaries to make them ask for peace. E.g.(Bach-Peters phrases) solution : A boy kissed a girl, and she really loved him, but he was only pretending (this is still ambiguous). but Solution: e.g. Every true Englishman reveres __ above all ... and __ is his queen. I 239 Predicate/Terminology/Geach: I only name predicates like this if they are used as the principal functor in a proposition, otherwise "predicables". I-predicables/I-predicate/Geach: (s): those predicates which are indistinguishable with respect to the two objects in a given theory. If distinctions can be made in an extended theory, the I-predicate does not change its meaning, it is no longer an I-predicate. E.g. "uniform" for (different, but not yet differentiated) tokens of words, later tokens are distinguished, but still "uniform". I 245 "Surman"/Geach: (should be identical, if they have the same family name) are not counted twice because in different theories differently provided with predicates and thus counted differently. ((s) identified as different by the theory.) E.g. applicable in the universe with the same right: "is the same token as ..", "is the same type as ... "- "is the same lexicon entry as...". I 250 Ascription theory/Geach: Vs "acts of will": attribution of responsibility instead of causality (GeachVs) - Oxford: Thesis: to say that an action is willful is not a description of the action but an attribution. I 291 GeachVsAristoteles: Vs "Two-Name-Theory"/TNT: confuses the relation of names to named with the relation of the predicate to from what it is said. -> false Doctrine of the Trinity. |
Gea I P.T. Geach Logic Matters Oxford 1972 |
| Theories | Cartwright | I 96 Theory/Duhem: is an abstract system for summing and logical classification of a set of experimental laws. - Does not explain the laws. >Laws, >Experimental laws. I 97 Theory is an organization of our knowledge. Theories/Cartwright: are abound - their explanations are not all needed. >Explanation. I 100 Theory/Cartwright: Prediction: lies in the fundamental laws. Content: lies in the phenomenological laws. I 133 Theory Entry: preliminary state: "unprepared description": left of the as-if-operator. 1st order prepared description: requires equation. 2nd order: Investigation of the prepared description with principles. - E.g. a laser can be described quite differently. (With or without memory). According to the decision, there are bridge principles that say which equations are to be applied. >Bridge laws. Hacking I 362 Theory/Cartwright: includes no truth itself. - If truth, then by approximations! >Idealization, >Truth. |
Car I N. Cartwright How the laws of physics lie Oxford New York 1983 CartwrightR I R. Cartwright A Neglected Theory of Truth. Philosophical Essays, Cambridge/MA pp. 71-93 In Theories of Truth, Paul Horwich Aldershot 1994 CartwrightR II R. Cartwright Ontology and the theory of meaning Chicago 1954 Hacking I I. Hacking Representing and Intervening. Introductory Topics in the Philosophy of Natural Science, Cambridge/New York/Oakleigh 1983 German Edition: Einführung in die Philosophie der Naturwissenschaften Stuttgart 1996 |
| Theories | Popper | Flor II 476 Theory/Popper: not justifiable, but verifiable. 1. Of any scientific theory one cannot know that it is true. 2. A scientific-empirical theory can contradict empirically observable facts. 3. A rational attitude is characterized by a critical attitude. Theories: decide that there is an inter-subjectively ascertainable fact, which may, however, contradict the theory. Flor II 478/79 One theory has to contain one or more strictly universal statements (laws) - General statement: e.g. "all bodies attract each other". Not a strict general statement: "all items in my drawer are red". Flor II 477 Definition basic statement: E.g. "at a certain time and in a certain place occurs this or that." A basic sentence may be in contradiction to the general statement, but cannot be derived originating. And expresses an intersubjectively observable fact. >Protocol sentences. General statement: a strictly universal statement is falsifiable if there is a possible basic statement, which contradicts it. E.g. "in my kitchen on 11 June 1989, there is a green shrew". basic satement: "A green shrew does not exist". A theory is only empirical scientific, if the class of its potential falsifiers is not empty. Flor II 484 It may turn out that basic statements were false, but one can also reject boundary conditions or additional hypotheses. Decisive: the assumption of a basic statement which is inconsistent with the statements contained in the test procedure, forces not to reject the central idea of a theory in general. A new theory has to be able to solve the problems of the old theory. In addition, it must be able to solve the problems that the old could not solve. (New theory contains the old as a subset). QuineVsPopper: this is a misconception: the new theory does not contain the old as a subset, but: E.g. also in everyday life, the theory of Newton is only an approximation. --- I 121 Theory/Popper: new theories have excess content. - But then they should not be adapted ad hoc. - Lakatos: the excess content is created piece by piece, by extending the theories. |
Po I Karl Popper The Logic of Scientific Discovery, engl. trnsl. 1959 German Edition: Grundprobleme der Erkenntnislogik. Zum Problem der Methodenlehre In Wahrheitstheorien, Gunnar Skirbekk Frankfurt/M. 1977 Flor I Jan Riis Flor "Gilbert Ryle: Bewusstseinsphilosophie" In Philosophie im 20. Jahrhundert, A. Hügli/P. Lübcke Reinbek 1993 Flor II Jan Riis Flor "Karl Raimund Popper: Kritischer Rationalismus" In Philosophie im 20. Jahrhundert, A.Hügli/P.Lübcke Reinbek 1993 Flor III J.R. Flor "Bertrand Russell: Politisches Engagement und logische Analyse" In Philosophie im 20. Jahrhundert, A. Hügli/P.Lübcke (Hg) Reinbek 1993 Flor IV Jan Riis Flor "Thomas S. Kuhn. Entwicklung durch Revolution" In Philosophie im 20. Jahrhundert, A. Hügli/P. Lübcke Reinbek 1993 |
| Theory of Relativity | Feyerabend | I 80 Relativity Theory/Feyerabend: E.g. Mercury perihelion: the famous deviation is explained by the relativity theory. The explanation shows that the prerequisite for derivation is not the general theory of relativity, but, apart from relativistic assumptions, always contains classical physics! In addition, the relativistic calculation ("blackboard solution") does not refer to the planetary system in the real world, but to the completely fictitious case of a centrally asymmetric universe that contains nothing apart from its singularity in the middle. Why are such strange assumptions made? Usual answer: we are dealing with approximations. Classical physics does not occur here, because the theory of relativity would be incomplete. Both schemes result from the general theory of relativity. You just have to neglect the sizes that are all too small. So the theory of relativity is applied consistently and in the correct way. >Aproximations. I 81 FeyerabendVs: this is a useful representation of the approximation method, but it does not reflect the real situation in the general theory of relativity! The classical theory is not used because it was proved to be correct, but in the hope that it will be useful! The approximations do not arise from relativistic calculations, but are introduced to be able to apply the theory of relativity to the case! (I 82), heliocentric theory at the time of Galileo, ad-hoc approximations to many quantitative results of the theories are not correct and surprisingly qualitatively inadequate. E.g. von Neumann: replaced the semi-intuitive concepts of Dirac and Bohr with incredibly complicated concepts. The relationship to experience becomes more obscure than ever. >Experience, >Perception, >Observation, >Theories, >Method. |
Feyerabend I Paul Feyerabend Against Method. Outline of an Anarchistic Theory of Knowledge, London/New York 1971 German Edition: Wider den Methodenzwang Frankfurt 1997 Feyerabend II P. Feyerabend Science in a Free Society, London/New York 1982 German Edition: Erkenntnis für freie Menschen Frankfurt 1979 |
| Values | Bostrom | I 226 Values/superintelligence/software-agents//Bostrom: While the agent is unintelligent, it might lack the capability to understand or even represent any humanly meaningful value. Problem: It is impossible to enumerate all possible situations a superintelligence might find itself in and to specify for each what action it should take. Similarly, it is impossible to create a list of all possible worlds and assign each of them a value. Motivation: A motivation system, therefore, cannot be specified as a comprehensive lookup table. It must instead be expressed more abstractly, as a formula or rule that allows the agent to decide what to do in any given situation. ((s) Cf. the philosophical discussion of principles against content: >Principles, >Utilitarianism, >Deontology.) I 227 Utility: Creating a machine that can compute a good approximation of the expected utility of the actions available to it is an AI-complete problem. (…) a problem, a problem that remains even if the problem of making machines intelligent is solved. We can use this framework of a utility-maximizing agent to consider the predicament of a future seed-AI programmer who intends to solve the control problem by endowing the AI with a final goal that corresponds to some plausible human notion of a worthwhile outcome. E.g., The programmer has some particular human value in mind that he would like the AI to promote. (…) let us say that it is happiness. But how could he express such a utility function in computer code? Computer languages do not contain terms such as “happiness” as primitives. I 228 If we cannot transfer human values into an AI by typing out full-blown representations in computer code, what else might we try? I 230 {Possible methods for acquiring values]: -Reinforcement learning: Often, the learning algorithm involves the gradual construction of some kind of evaluation function, which assigns values to states, state–action pairs, or policies. Problem: The evaluation function, which is continuously updated in light of experience, could be regarded as incorporating a form of learning about value. However, what is being learned is not new final values but increasingly accurate estimates of the instrumental values of reaching particular states (or of taking particular actions in particular states, or of following particular policies). Insofar as a reinforcement-learning agent can be described as having a final goal, that goal remains constant: to maximize future reward. And reward consists of specially designated percepts received from the environment. Therefore, the wireheading syndrome remains a likely outcome in any reinforcement agent that develops a world model sophisticated enough to suggest this alternative way of maximizing reward. I 233 - Motivational scaffolding: It involves giving the seed AI an interim goal system, with relatively simple final goals that we can represent by means of explicit coding or some other feasible method. Once the AI has developed more sophisticated representational faculties, we replace this interim scaffold goal system with one that has different final goals. Problem: Because the scaffold goals are not just instrumental but final goals for the AI, the AI might be expected to resist having them replaced (goal-content integrity being a convergent instrumental value). This creates a hazard. If the AI succeeds in thwarting the replacement of its scaffold goals, the method fails. I 234 Further problems: (1) The motivational scaffolding (…) carries the risk that the AI could become too powerful while it is still running on its interim goal system. (2) Installing the ultimately intended goals in a human-level AI is not necessarily that much easier than doing so in a more primitive AI. I 235 -Value learning: [in order to] AI’s intelligence to learn the values (…) we must provide a criterion for the AI that at least implicitly picks out some suitable set of values. (…) the value learning approach retains an unchanging final goal throughout the AI’s developmental and operational phases. Learning does not change the goal. It changes only the AI’s beliefs about the goal. Criteria: The AI thus must be endowed with a criterion that it can use to determine which percepts constitute evidence in favor of some hypothesis about what the ultimate goal is, and which percepts constitute evidence against. Problem: creating artificial general intelligence in the first place, which requires a powerful learning mechanism that can discover the structure of the environment from limited sensory inputs. I 240 Understanding/motivation: (…) the difficulty here is not so much how to ensure that the AI can understand human intentions. A superintelligence should easily develop such understanding. Rather, the difficulty is ensuring that the AI will be motivated to pursue the described values in the way we intended. This is not guaranteed by the AI’s ability to understand our intentions: an AI could know exactly what we meant and yet be indifferent to that interpretation of our words (being motivated instead by some other interpretation of the words or being indifferent to our words altogether). Solution: the correct motivation should ideally be installed in the seed AI before it becomes capable of fully representing human concepts or understanding human intentions. I 253 [Further] value-loading techniques: - Evolutionary selection: Powerful search may find a design that satisfies the formal search criteria but not our intentions. - Value accretion: (…) the human value-accretion dispositions might be complex and difficult to replicate in a seed AI. Problem: A bad approximation may yield an AI that generalizes differently than humans do and therefore acquires unintended final goals. I 254 - Motivational scaffolding: encourage a system to develop internal high-level representations that are transparent to humans (while keeping the system’s capabilities below the dangerous level) and then to use those representations to design a new goal system. - Emulation modulation: If machine intelligence is achieved via the emulation pathway, it would likely be possible to tweak motivations through the digital equivalent of drugs or by other means. - Institution design: Various strong methods of social control could be applied in an institution composed of emulations. In principle, social control methods could also be applied in an institution composed of artificial intelligences. >Ethics/superintelligence/Bostrom, >Ethics/superintelligence/Yudkowsky, >Norms/Bostrom. |
Bostrom I Nick Bostrom Superintelligence. Paths, Dangers, Strategies Oxford: Oxford University Press 2017 |
| Variable Elimination | Norvig | Norvig I 545 Variable Elimination/Norvig/Russell: The basic idea of variable elimination—that repeated computations within the overall sum-of-products expression can be avoided by caching—appeared in the symbolic probabilistic inference (SPI) algorithm (Shachter et al., 1990)(1). Cf. the elimination algorithm (…) developed by Zhang and Poole (1994)(2). Criteria for pruning irrelevant variables were developed by Geiger et al. (1990)(3) and by Lauritzen et al. (1990)(4) (…). Dechter (1999)(5) shows how the variable elimination idea is essentially identical to nonserial dynamic programming (Bertele and Brioschi, 1972)(6), an algorithmic approach that can be applied to solve a range of inference problems in Bayesian networks - for example, finding the most likely explanation for a set of observations. This connects Bayesian network algorithms to related methods for solving CSPs (>Constraint satisfaction problems) and gives a direct measure of the complexity of exact inference in terms of the tree width of the network. Wexler and Meek (2009)(7) describe a method of preventing exponential growth in the size of factors computed in variable elimination; their algorithm breaks down large factors into products of smaller factors and simultaneously computes an error bound for the resulting approximation. >Bayesian networks/Norvig, >Uncertainty/AI research. 1. Shachter, R. D., D’Ambrosio, B., and Del Favero, B. A. (1990). Symbolic probabilistic inference in belief networks. In AAAI-90, pp. 126–131. 2. Zhang, N. L., Qi, R., and Poole, D. (1994). A computational theory of decision networks. IJAR, 11, 83–158. 3. Geiger, D., Verma, T., and Pearl, J. (1990). Identifying independence in Bayesian networks. Networks, 20(5), 507–534. 4. Lauritzen, S., Dawid, A. P., Larsen, B., and Leimer, H. (1990). Independence properties of directed Markov fields. Networks, 20(5), 491–505 5. Dechter, R. (1999). Bucket elimination: A unifying framework for reasoning. AIJ, 113, 41–85. 6. Bertele, U. and Brioschi, F. (1972). Nonserial dynamic programming. Academic Press. 7. Wexler, Y. and Meek, C. (2009). MAS: A multiplicative approximation scheme for probabilistic inference. In NIPS 21. |
Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Variable Elimination | Russell | Norvig I 545 Variable Elimination/Norvig/Russell: The basic idea of variable elimination—that repeated computations within the overall sum-of-products expression can be avoided by caching—appeared in the symbolic probabilistic inference (SPI) algorithm (Shachter et al., 1990)(1). Cf. the elimination algorithm (…) developed by Zhang and Poole (1994)(2). Criteria for pruning irrelevant variables were developed by Geiger et al. (1990)(3) and by Lauritzen et al. (1990)(4) (…). Dechter (1999)(5) shows how the variable elimination idea is essentially identical to nonserial dynamic programming (Bertele and Brioschi, 1972)(6), an algorithmic approach that can be applied to solve a range of inference problems in Bayesian networks - for example, finding the most likely explanation for a set of observations. This connects Bayesian network algorithms to related methods for solving CSPs (>Constraint satisfaction problems) and gives a direct measure of the complexity of exact inference in terms of the tree width of the network. Wexler and Meek (2009)(7) describe a method of preventing exponential growth in the size of factors computed in variable elimination; their algorithm breaks down large factors into products of smaller factors and simultaneously computes an error bound for the resulting approximation. >Bayesian networks/Norvig, >Uncertainty/AI research. 1. Shachter, R. D., D’Ambrosio, B., and Del Favero, B. A. (1990). Symbolic probabilistic inference in belief networks. In AAAI-90, pp. 126–131. 2. Zhang, N. L., Qi, R., and Poole, D. (1994). A computational theory of decision networks. IJAR, 11, 83–158. 3. Geiger, D., Verma, T., and Pearl, J. (1990). Identifying independence in Bayesian networks. Networks, 20(5), 507–534. 4. Lauritzen, S., Dawid, A. P., Larsen, B., and Leimer, H. (1990). Independence properties of directed Markov fields. Networks, 20(5), 491–505 5. Dechter, R. (1999). Bucket elimination: A unifying framework for reasoning. AIJ, 113, 41–85. 6. Bertele, U. and Brioschi, F. (1972). Nonserial dynamic programming. Academic Press. 7. Wexler, Y. and Meek, C. (2009). MAS: A multiplicative approximation scheme for probabilistic inference. In NIPS 21. |
Russell I B. Russell/A.N. Whitehead Principia Mathematica Frankfurt 1986 Russell II B. Russell The ABC of Relativity, London 1958, 1969 German Edition: Das ABC der Relativitätstheorie Frankfurt 1989 Russell IV B. Russell The Problems of Philosophy, Oxford 1912 German Edition: Probleme der Philosophie Frankfurt 1967 Russell VI B. Russell "The Philosophy of Logical Atomism", in: B. Russell, Logic and KNowledge, ed. R. Ch. Marsh, London 1956, pp. 200-202 German Edition: Die Philosophie des logischen Atomismus In Eigennamen, U. Wolf (Hg) Frankfurt 1993 Russell VII B. Russell On the Nature of Truth and Falsehood, in: B. Russell, The Problems of Philosophy, Oxford 1912 - Dt. "Wahrheit und Falschheit" In Wahrheitstheorien, G. Skirbekk (Hg) Frankfurt 1996 Norvig I Peter Norvig Stuart J. Russell Artificial Intelligence: A Modern Approach Upper Saddle River, NJ 2010 |
| Verification | Feynman | I 33 Rules/Review/Verification/Feynman: a good way leads through the less specific rules derived from it. Something can also be checked by approximation. >Theories, >Measurements, >Prediction, >Confirmation, >Idealization, >Method, >Rules. |
Feynman I Richard Feynman The Feynman Lectures on Physics. Vol. I, Mainly Mechanics, Radiation, and Heat, California Institute of Technology 1963 German Edition: Vorlesungen über Physik I München 2001 Feynman II R. Feynman The Character of Physical Law, Cambridge, MA/London 1967 German Edition: Vom Wesen physikalischer Gesetze München 1993 |
| Verification (Confirmation) | Quine | I 54 Confirmation/QuineVsPeirce: infinite confirmation is not ideal but always correctable. False analogy of the limit value of an approximation to truth. II 36 Truth/Quine: is not confirmed by evidence! They could always be reinterpreted. Truth is intrinsic, there is nothing about it. Interpretation always takes place within a theory. >Theories, >Evidence. II 43 Quine: but we still see room for intuitive confirmation. II 44 We rely more and more on confirmed observation sentences through conditioning. The categorical observation sentences form the empirical content because only through them theory is linked to observation. V 61/62 Observation/Theory/Quine: here there are two relationships: one epistemological of confirmation and one semantic, through which the sentences get their meaning. These two relations are coextensive. V 63 Observation/Quine: e.g. face, hearing, touch, smell sensation. N.B.: for their role as confirmation or also as semantic reference points, however, it is crucial that they are something socially divided. >Reference. Problem: two people will judge them differently, partly because they notice different characteristics, partly because they have different theories. V 64 Solution/Quine: one should speak neither of sensations nor of environmental conditions, but of language. ((s) >Semantic Ascent). Fodor IV 37ff Verification (Confirmation)/Quine: cannot follow from meaning - Tarski/(s) in the case of "snow is white" it can naturally not be about verification - although statements are individuated through their content, i.e. that they are essential to them, from this does not follow anything related to possible verification - ((s) but certainly about verification conditions?) - QuineVsPeirce: (Peirce thesis meaning = method of verification): it can all turn out to be wrong, i.e. ~ even the meaning - ((s)> E.g. Putnam, stars replaced by light bulbs). |
Quine I W.V.O. Quine Word and Object, Cambridge/MA 1960 German Edition: Wort und Gegenstand Stuttgart 1980 Quine II W.V.O. Quine Theories and Things, Cambridge/MA 1986 German Edition: Theorien und Dinge Frankfurt 1985 Quine III W.V.O. Quine Methods of Logic, 4th edition Cambridge/MA 1982 German Edition: Grundzüge der Logik Frankfurt 1978 Quine V W.V.O. Quine The Roots of Reference, La Salle/Illinois 1974 German Edition: Die Wurzeln der Referenz Frankfurt 1989 Quine VI W.V.O. Quine Pursuit of Truth, Cambridge/MA 1992 German Edition: Unterwegs zur Wahrheit Paderborn 1995 Quine VII W.V.O. Quine From a logical point of view Cambridge, Mass. 1953 Quine VII (a) W. V. A. Quine On what there is In From a Logical Point of View, Cambridge, MA 1953 Quine VII (b) W. V. A. Quine Two dogmas of empiricism In From a Logical Point of View, Cambridge, MA 1953 Quine VII (c) W. V. A. Quine The problem of meaning in linguistics In From a Logical Point of View, Cambridge, MA 1953 Quine VII (d) W. V. A. Quine Identity, ostension and hypostasis In From a Logical Point of View, Cambridge, MA 1953 Quine VII (e) W. V. A. Quine New foundations for mathematical logic In From a Logical Point of View, Cambridge, MA 1953 Quine VII (f) W. V. A. Quine Logic and the reification of universals In From a Logical Point of View, Cambridge, MA 1953 Quine VII (g) W. V. A. Quine Notes on the theory of reference In From a Logical Point of View, Cambridge, MA 1953 Quine VII (h) W. V. A. Quine Reference and modality In From a Logical Point of View, Cambridge, MA 1953 Quine VII (i) W. V. A. Quine Meaning and existential inference In From a Logical Point of View, Cambridge, MA 1953 Quine VIII W.V.O. Quine Designation and Existence, in: The Journal of Philosophy 36 (1939) German Edition: Bezeichnung und Referenz In Zur Philosophie der idealen Sprache, J. Sinnreich (Hg) München 1982 Quine IX W.V.O. Quine Set Theory and its Logic, Cambridge/MA 1963 German Edition: Mengenlehre und ihre Logik Wiesbaden 1967 Quine X W.V.O. Quine The Philosophy of Logic, Cambridge/MA 1970, 1986 German Edition: Philosophie der Logik Bamberg 2005 Quine XII W.V.O. Quine Ontological Relativity and Other Essays, New York 1969 German Edition: Ontologische Relativität Frankfurt 2003 Quine XIII Willard Van Orman Quine Quiddities Cambridge/London 1987 F/L Jerry Fodor Ernest Lepore Holism. A Shoppers Guide Cambridge USA Oxford UK 1992 Fodor I Jerry Fodor "Special Sciences (or The Disunity of Science as a Working Hypothesis", Synthese 28 (1974), 97-115 In Kognitionswissenschaft, Dieter Münch Frankfurt/M. 1992 Fodor II Jerry Fodor Jerrold J. Katz Sprachphilosophie und Sprachwissenschaft In Linguistik und Philosophie, G. Grewendorf/G. Meggle Frankfurt/M. 1974/1995 Fodor III Jerry Fodor Jerrold J. Katz The availability of what we say in: Philosophical review, LXXII, 1963, pp.55-71 In Linguistik und Philosophie, G. Grewendorf/G. Meggle Frankfurt/M. 1974/1995 |
| |||
| Disputed term/author/ism | Author Vs Author |
Entry |
Reference |
|---|---|---|---|
| Creary, L | Cartwright Vs Creary, L | I 62 Factual View/Causal forces/Lewis Creary/Cartwright: Creary tries to maintain separate causal laws and at the same time to save the facts-view by postulating an intermediate effect: physical laws/Creary: Thesis: there are two types of them: a) Laws on causal influence: E.g. law of gravity. Coulomb’s Law: they tell us, what forces or other causal impacts become effective in different circumstances. b) Laws about causal action: ("Action laws"): they tell us what results arise from such combinations. E.g. law of vector addition, provides satisfactory explanations. Composition/Creary: less satisfactory types: E.g. amplification, interference, elimination (predomination). I 63 Truth/Physical laws/Creary/Cartwright: according to Creary these laws are true, because they correctly describe what influences are produced. Truth/Law/CartwrightVsCreary: this is a plausible representation of the structure of many causal explanations, but it has two disadvantages: 1) often there are no general laws of interaction. The dynamics of the vector addition is fortunate in this regard. Problem: irreversible processes: flux, Laws of transport (heat transport) distribution functions. The equations in statistical mechanics do not apply in 90% of cases. (Kline, Similitude and Approximation, NY, 1969, p. 140). I 65 Creary/Cartwright: his action laws (which provide the resulting overall behavior) only apply to individual cases. CartwrightVsCreary: better correct laws like Fick’s law. Nature/Cartwright: should rather be described by many phenomenological laws which are tailored to individual situations, than ruled by first principles. (s) VsCartwright: there can be no laws for individual situations (specific situations). CartwrightVsCreary: 2) causal influence: E.g. resulting force in vector addition: Creary: Thesis: there is no force that results, but a movement (behavior). With that we can deny the reality of a resulting force. Cartwright: We both agree that there cannot be three forces: the first two, and in addition the resulting one. CartwrightVsCreary: but I assert the reality of the resulting force while Creary asserts the component forces. Causal influence/Creary: is an intermediary factor between cause and what was initially thought to be the effect. CartwrightVsCreary: this will not work in general. E.g. two laws: a) C causes E b) C’ causes E’. In addition: C and C’ together cause E’’. Then we do not want to assume three effects E, E ’and E’’, but we need to assume some other incidents F and F’ as the actual effects of the two laws a) and b). And, according to another law, these will produce E’’. CartwrightVsCreary: this can work in individual cases, but not always. I see no reason why such intermediate factors should be found all the time. These seem to me more shadowy. |
Car I N. Cartwright How the laws of physics lie Oxford New York 1983 CartwrightR I R. Cartwright A Neglected Theory of Truth. Philosophical Essays, Cambridge/MA pp. 71-93 In Theories of Truth, Paul Horwich Aldershot 1994 CartwrightR II R. Cartwright Ontology and the theory of meaning Chicago 1954 |
| Fraassen, B. van | Duhem Vs Fraassen, B. van | Cartwright I 96 Fraassen: the empirical substructure provided by the theory, should be isomorphic to that of the phenomena. DuhemVsFraassen: at best, this is possible in very rough approximation. (Cartwright ditto). |
Duh I P. Duhem La théorie physique, son objet et sa structure, Paris 1906 German Edition: Ziel und Struktur der physikalischen Theorien Hamburg 1998 CartwrightR II R. Cartwright Ontology and the theory of meaning Chicago 1954 |
| Hempel, C. | Nozick Vs Hempel, C. | II 301 Explanation/Probability/Nozick: if a fundamental probability law finds that the probability that something has P - given it also has Q - is 95%, we cannot deduce this fact from the having of the property P plus the probability law! ((s) We need to know if it also has Q). Probability Law/Hempel: thought that events that have high probabilities are explained by subsuming under a probability law. As an approximation to deduction. NozickVsHempel: what about improbable events? If we find P without Q, how do we explain that? Hempel cannot explain this . II 302 Hempel: can only explain that one or the other P occurs without Q, but not why! Nozick: but we do know that there is some kind of system that produces some Ps that are not Qs. And we explain why this thing is so through the mechanism of random operation. Nozick: the alternative would be to say that events with low probability are inexplicable. (NozickVs). Explanation/Probability/Nozick: Thesis: we have an understanding and an explanation of why something happens, even if we do not know the reasons why it happened at the time. Even if it is random, it need not be inexplicable. It may be an event of a type. |
No I R. Nozick Philosophical Explanations Oxford 1981 No II R., Nozick The Nature of Rationality 1994 |
| Principia Mathematica | Gödel Vs Principia Mathematica | Russell I XIV Circular Error Principle/VsPrincipia Mathematica(1)/PM/Russell/Gödel: thus seems to apply only to constructivist assumptions: when a term is understood as a symbol, together with a rule to translate sentences containing the symbol into sentences not containing it. Classes/concepts/Gödel: can also be understood as real objects, namely as "multiplicities of things" and concepts as properties or relations of things that exist independently of our definitions and constructions! This is just as legitimate as the assumption of physical bodies. They are also necessary for mathematics, as they are for physics. Concept/Terminology/Gödel: I will use "concept" from now on exclusively in this objective sense. A formal difference between these two conceptions of concepts would be: that of two different definitions of the form α(x) = φ(x) it can be assumed that they define two different concepts α in the constructivist sense. (Nominalistic: since two such definitions give different translations for propositions containing α.) For concepts (terms) this is by no means the case, because the same thing can be described in different ways. For example, "Two is the term under which all pairs fall and nothing else. There is certainly more than one term in the constructivist sense that satisfies this condition, but there could be a common "form" or "nature" of all pairs. All/Carnap: the proposal to understand "all" as a necessity would not help if "provability" were introduced in a constructivist manner (..+...). Def Intensionality Axiom/Russell/Gödel: different terms belong to different definitions. This axiom holds for terms in the circular error principle: constructivist sense. Concepts/Russell/Gödel: (unequal terms!) should exist objectively. (So not constructed). (Realistic point of view). When only talking about concepts, the question gets a completely different meaning: then there seems to be no objection to talking about all of them, nor to describing some of them with reference to all of them. Properties/GödelVsRussell: one could surely speak of the totality of all properties (or all of a certain type) without this leading to an "absurdity"! ((s) > Example "All properties of a great commander". Gödel: this simply makes it impossible to construe their meaning (i.e. as an assertion about sense perception or any other non-conceptual entities), which is not an objection to someone taking the realistic point of view. Part/whole/Mereology/GödelVsRussell: neither is it contradictory that a part should be identical (not just the same) with the whole, as can be seen in the case of structures in the abstract sense. Example: the structure of the series of integers contains itself as a special part. I XVI/XVII Even within the realm of constructivist logic there are certain approximations to this self-reflectivity (self-reflexivity/today: self-similarity) of impredicative qualities, namely e.g. propositions, which as parts of their meaning do not contain themselves, but their own formal provability. There are also sentences that refer to a totality of sentences to which they themselves belong: Example: "Each sentence of a (given) language contains at least one relational word". This makes it necessary to look for other solutions to the paradoxes, according to which the fallacy does not consist in the assumption of certain self-reflectivities of the basic terms, but in other assumptions about them! The solution may have been found for the time being in simple type theory. Of course, all this refers only to concepts. Classes: one should think that they are also not created by their definitions, but only described! Then the circular error principle does not apply again. Zermelo splits classes into "levels", so that only sets of lower levels can be elements of sets of higher levels. Reducibility Axiom/Russell/Gödel: (later dropped) is now taken by the class axiom (Zermelo's "axiom of choice"): that for each level, for any propositional function φ(x) the set of those x of this level exists for which φ(x) is true. This seems to be implied by the concept of classes as multiplicities. I XVIII Extensionality/Classes: Russell: two reasons against the extensional view of classes: 1. the existence of the zero class, which cannot be well a collection, 2. the single classes, which should be identical with their only elements. GödelVsRussell: this could only prove that the zero classes and the single classes (as distinguished from their only element) are fictions to simplify the calculation, and do not prove that all classes are fictions! Russell: tries to get by as far as possible without assuming the objective existence of classes. According to this, classes are only a facon de parler. Gödel: but also "idealistic" propositions that contain universals could lead to the same paradoxes. Russell: creates rules of translation according to which sentences containing class names or the term "class" are translated into sentences not containing them. Class Name/Russell: eliminate by translation rules. Classes/Principia Mathematica/Russell/Gödel: the Principia Mathematica can do without classes, but only if you assume the existence of a concept whenever you want to construct a class. First, some of them, the basic predicates and relations like "red", "colder" must be apparently considered real objects. The higher terms then appear as something constructed (i.e. something that does not belong to the "inventory of the world"). I XIX Ramsey: said that one can form propositions of infinite length and considers the difference finite/infinite as not so decisive. Gödel: Like physics, logic and mathematics are based on real content and cannot be "explained away". Existence/Ontology/Gödel: it does not behave as if the universe of things is divided into orders and one is forbidden to speak of all orders, but on the contrary: it is possible to speak of all existing things. But classes and concepts are not among them. But when they are introduced as a facon de parler, it turns out that the extension of symbolism opens the possibility of introducing them in a more comprehensive way, and so on, to infinity. To maintain this scheme, however, one must presuppose arithmetics (or something equivalent), which only proves that not even this limited logic can be built on nothing. I XX Constructivist posture/constructivism/Russell/Gödel: was abandoned in the first edition, since the reducibility axiom for higher types makes it necessary that basic predicates of arbitrarily high type exist. From constructivism remains only 1. Classes as facon de parler 2. The definition of ~, v, etc. as valid for propositions containing quantifiers, 3. The stepwise construction of functions of orders higher than 1 (of course superfluous because of the R-Axiom) 4. the interpretation of definitions as mere typographical abbreviations (all incomplete symbols, not those that name an object described by the definition!). Reducibility Axiom/GödelVsRussell: this last point is an illusion, because of the reducibility axiom there are always real objects in the form of basic predicates or combinations of such according to each defined symbol. Constructivist posture/constructivism/Principia Mathematica/Gödel: is taken again in the second edition and the reducibility axiom is dropped. It is determined that all basic predicates belong to the lowest type. Variables/Russell/Gödel: their purpose is to enable the assertions of more complicated truth functions of atomistic propositions. (i.e. that the higher types are only a facon de parler.). The basis of the theory should therefore consist of truth functions of atomistic propositions. This is not a problem if the number of individuals and basic predicates is finite. Ramsey: Problem of the inability to form infinite propositions is a "mere secondary matter". I XXI Finite/infinite/Gödel: with this circumvention of the problem by disregarding the difference between finite and infinite a simpler and at the same time more far-reaching interpretation of set theory exists: Then Russell's Apercu that propositions about classes can be interpreted as propositions about their elements becomes literally true, provided n is the number of (finite) individuals in the world and provided we neglect the zero class. (..) + I XXI Theory of integers: the second edition claims that it can be achieved. Problem: that in the definition "those cardinals belonging to each class that contains 0 and contains x + 1 if it contains x" the phrase "each class" must refer to a given order. I XXII Thus whole numbers of different orders are obtained, and complete induction can be applied to whole numbers of order n only for properties of n! (...) The question of the theory of integers based on ramified type theory is still unsolved. I XXIII Theory of Order/Gödel: is more fruitful if it is considered from a mathematical point of view, not a philosophical one, i.e. independent of the question of whether impredicative definitions are permissible. (...) impredicative totalities are assumed by a function of order α and ω . Set/Class/Principia Mathematica(1)/Russell/Type Theory/Gödel: the existence of a well-ordered set of the order type ω is sufficient for the theory of real numbers. Def Continuum Hypothesis/Gödel: (generalized): no cardinal number exists between the power of any arbitrary set and the power of the set of its subsets. Type Theory/VsType Theory/GödelVsRussell: mixed types (individuals together with predications about individuals etc.) obviously do not contradict the circular error principle at all! I XXIV Russell based his theory on quite different reasons, similar to those Frege had already adopted for the theory of simpler types for functions. Propositional functions/statement function/Russell/Gödel: always have something ambiguous because of the variables. (Frege: something unsaturated). Propositional function/p.f./Russell/Gödel: is so to speak a fragment of a proposition. It is only possible to combine them if they "fit together" i.e. are of a suitable type. GödelVsRussell: Concepts (terms) as real objects: then the theory of simple types is not plausible, because what one would expect (like "transitivity" or the number two) to be a concept would then seem to be something that stands behind all its different "realizations" on the different levels and therefore does not exist according to type theory. I XXV Paradoxes in the intensional form/Gödel: here type theory brings a new idea: namely to blame the paradoxes not on the axiom that every propositional function defines a concept or a class, but on the assumption that every concept results in a meaningful proposition if it is claimed for any object as an argument. The objection that any concept can be extended to all arguments by defining another one that gives a false proposition whenever the original one was meaningless can easily be invalidated by pointing out that the concept "meaningfully applicable" does not always have to be meaningfully applicable itself. 1. Whitehead, A.N. and Russel, B. (1910). Principia Mathematica. Cambridge: Cambridge University Press. |
Göd II Kurt Gödel Collected Works: Volume II: Publications 1938-1974 Oxford 1990 |
| Sense Data | Ryle Vs Sense Data | I 301 RyleVsSense Data. E.g. someone who says: "The headlamps are blinding!" need not be blinded himself. - It is a fallacy to say that the blinding is therefore not a property of the headlamps, but a characteristic of the sense-data of blinded individual. --- I 280 SenseData/RyleVsSense Data/Ryle: some theorists want us to turn away from robins and cheeses, and to turn to such things as looking at something and fragrances (sense data). That would mean, if I can observe this sight when I catch the sight of the robin, then I must catch something like a sight of the sight. Or a fragrance of a fragrance. Regress. E.g. When someone is watching horses racing the question is whether he had a good seat, but not if someone is itchy. Hence, also not the question whether the sensation of his itching was superficial. --- I 282 Attention/Ryle: Observing or sighting include noticing, but noticing does not include observing. RyleVsSense data/Ryle: it was wrong from the start to compare the objects of the open observation, such as cheese and robins to objects, which are supposed to be accessible only to a privileged observation: namely, my sensations. Sensations are, in fact, not objects of observation. We do not need to build a "theater of the outside world" in order to compare it to an inner "theater", the mind. The properties that observable objects can have, cannot be sensibly attributed or denied for sensations. --- I 283 It is senseless to say sensations had a position, size, temperature, color, or smell. Also no place. In any case, not in response to a question such as "Where was the robin?" In a certain sense this is already the tingling in my foot or "burning in my nose" but that is a different sense than pepper grains in my nose or pointy stones under my foot. My cringing is not hidden from the shoemaker because it is within me. On the contrary, it cannot at all be described as a needle, object within or outside public access. Or within the public object that I am myself. Likewise, letters cannot be classified as main words or time words as property words. --- I 286 Seeing/hearing/tasting/Ryle: are not used to describe "pure" or "unmixed" sensations, for we are talking about seeing horse races and listening to trains, tasting wines. Horse racing, trains and wines are not sensations, horse racings do not stop when I close my eyes, good wines are not destroyed when I have a cold. We need, therefore, apparently ways of talking about what ceases when I close my eyes, etc. Sense data/RayleVsSense data: the sight or the appearance of the horses actually changes as tears flow, or the taste of the wine is affected by the cold, and the sounds of the train become quieter when I cover my ears. It seems, then, that we can talk about "pure" sensations when we talk about "appearance" "sound" "taste", "fragrance", "tingling", etc. It also seems that with these expressions we can make distinctions and exclude those sensations which are achieved by reasoning, education, memory, presumption, habit, imagination, or association. (Covering one's ears does not help in the memory). --- I 287 SenseData/RyleVs: but what does it mean to get a momentary sight or a momentary scent? And what is the sight or smell like that you get? 1. The sight of a horse race is not a sporting event on a racing field. Just as anyone can be a witness of the race, not everyone can be a witness of the sight. (Already because of the perspective, not the point of view) Nor can anyone else have my tingling. Definition Sense Data/Ryle: (e.g. scent, sight, tingling, ringing) is private property of a single perceiver. Further, the sight of a horse race is described as a short-lived patchwork of color spots. RyleVsSense Data: but the visual appearance of things or the sight which is described as color spots can not be viewed as surfaces of ordinary flat objects. They are simply spots of color, not spots of colored plaster or colored canvas. Tradition: the representatives of the theory of sense data agree that sense data is something private, but not in that, --- I 288 that they are of a psychical nature or "in my mind". Tradition: the representatives now ask themselves the question: What is the fact that their recipient has it or receives it? Answer is simple: according to some, he should perceive or observe, some even say that people do not see a horse race, but only observe color spots. --- I 290 Sense data/deception/Ryle: It is claimed, in particular, that the theory solves paradoxes in the description of illusions. If the squinting man asserts that he sees two candles and the drunk says, he sees white mice, one can now translate that the squinting one sees two "candle appearances" and the drunken one in reality "white mouse appearances". Their only error then lies in the conclusion of existence. Next: For example a forward tilted plate looks elliptical (in fact is wrong). There is an elliptical white spot in the field of view. RyleVsSense data theory: the whole theory is based on a logical blunder, namely, the approximation of the concept of sensation to that of observation. --- I 291 Sense data/RyleVsSense data/Ryle: Having a look at a horse race is explained by having something else, namely a color combination. Regress: If having the sensation of a horse race implies having something else, then having that other, the color spot must again imply having something else, an earlier sense-date. --- I 294 SenseData/RyleVsSense data/Ryle: one cannot say meaningfully, one sees the appearance of a plate and not the plate itself, because appearance is already a main word for seeing. There is no other object, the "appearance" next to the plate. One does not say "I see a flat elliptical spot of white" but: "Maybe I see an elliptical and not slanted piece of white porcelain." E.g. We can say that a closer plane looks faster than the one more distant, but we cannot say "it has a faster look". "Looks faster" means, looks as if it was flying faster. Talking about the apparent speeds of airplanes does not mean talking about the speeds of the appearances of airplanes. --- I 298 Indeterminacy/VsSense data theory/Ryle: one who is peeking through a hole in the hedge may say he sees an area of yellow ... but will not be able to tell if it is dandelion or yellow canvas. He could only say, "I saw something yellow." Contrary to this normal use of expressions of indeterminacy, the sense data theory supports continuous talking of this kind. |
Ryle I G. Ryle The Concept of Mind, Chicago 1949 German Edition: Der Begriff des Geistes Stuttgart 1969 |
| Unger, P. | Lewis Vs Unger, P. | IV 244 Sorites/Truth Value/Vagueness/Lewis: For example Fred is a borderline case of baldness, then the sentence "Fred is bald" is perhaps without truth values. Nothing in our language makes such descriptions (delineations) right and others wrong. We can't find a limit once and for all. If a sentence is true over the entire range, it is simply true. But we treat a sentence as more or less "simply true" even if it goes beyond an area of its vagueness that is large enough. So if it is "true enough". We can usually cope with this, but not always, as the paradoxes testify: Problem: truth-preserving arguments do not always have the quality of being "true enough"! "true enough": when is one sentence true enough? It's a matter of vagueness in itself. IV 245 More importantly, it depends on the context. In other circumstances, something may not be true enough. Austin: "France is hexagonal". Standards can be loosened or tightened. Interestingly, tightening is easier than loosening: For example, if the standards were high and something is said that is true enough only under relaxed standards, and nobody contradicts, then the standards are lowered. But what is said under lowered standards may still seem imperfectly acceptable. For example, tightening standards: always manages to appear recommendable, even if it disturbs the purposes of conversation. Absolute/relative: e.g. (Peter Unger): one could say that there is actually nothing that is really level! The sidewalk is level, but the desk is more level! And so there is surely also something that is more level than the desk. One can always think of something that is even more level, etc. Problem: "level" should actually be taken as an absolute term. Then how could one deny that the table is level VsUnger: one could deny that "level" is absolute. But Unger is right about that. What he calls inconsistent really sounds that way. So I assume that in no description of the relative vagueness of "level" and "more level" it is true that something is more level than something that is level. LewisVsUnger: the correct answer is that he is changing her account. (He is changing the score on you). He's transferring the account to you. What he says is only acceptable under tightened standards of precision. IV 246 Because what he says is only acceptable under tighter standards, it is no longer true that the sidewalk is level. But that does not change the fact that it was true in the original context. Unger has not shown that the new context is somehow more legitimate than the old one. "Safe"/Unger: in an analogous way, Unger (correctly) observed that "safe" is an absolute term. Therefore, nobody is actually safe in any matter! In fact, the approximation rule allows Unger to create a context in which everything he says is true, but that doesn't show that anything we do in more everyday contexts is wrong. |
Lewis I David K. Lewis Die Identität von Körper und Geist Frankfurt 1989 Lewis I (a) David K. Lewis An Argument for the Identity Theory, in: Journal of Philosophy 63 (1966) In Die Identität von Körper und Geist, Frankfurt/M. 1989 Lewis I (b) David K. Lewis Psychophysical and Theoretical Identifications, in: Australasian Journal of Philosophy 50 (1972) In Die Identität von Körper und Geist, Frankfurt/M. 1989 Lewis I (c) David K. Lewis Mad Pain and Martian Pain, Readings in Philosophy of Psychology, Vol. 1, Ned Block (ed.) Harvard University Press, 1980 In Die Identität von Körper und Geist, Frankfurt/M. 1989 Lewis II David K. Lewis "Languages and Language", in: K. Gunderson (Ed.), Minnesota Studies in the Philosophy of Science, Vol. VII, Language, Mind, and Knowledge, Minneapolis 1975, pp. 3-35 In Handlung, Kommunikation, Bedeutung, Georg Meggle Frankfurt/M. 1979 Lewis IV David K. Lewis Philosophical Papers Bd I New York Oxford 1983 Lewis V David K. Lewis Philosophical Papers Bd II New York Oxford 1986 Lewis VI David K. Lewis Convention. A Philosophical Study, Cambridge/MA 1969 German Edition: Konventionen Berlin 1975 LewisCl Clarence Irving Lewis Collected Papers of Clarence Irving Lewis Stanford 1970 LewisCl I Clarence Irving Lewis Mind and the World Order: Outline of a Theory of Knowledge (Dover Books on Western Philosophy) 1991 |
| Whitehead, A.N. | Simons Vs Whitehead, A.N. | I 94 Bowman L. Clarke/topology/mereology/Simons: formal objections against his system cannot be put forward. It is based on Whitehead's basic concept of compound, the relata are informally understood space-time regions. I 95 Def connected/connection/Clarke/Whitehead: connected means sharing a point ((s) common point). But the points and all the other borders are no individuals. Limit/Whitehead/Clarke: the limit is no individual. Individuals/Whitehead/Clarke: individuals have no interiors. This leads to a non-classical mereology. Connection/spelling/Clarke: it is written as a small diamond with double tails up and down. Separated/disconnected/external connection/spelling/Clarke: >< y”: x is externally connected with y, = "x touches y". Non-classical mereology/Simons: here o (overlap) and < (part-relation) do not interact in the way as in the classic. Only when an object touches nothing (that means intuitive, if it is open, see above) we can treat its parts as in classical mereology. I 96 "Quasi-topologically"/Clarke: (Because there is no zero element and no boundary elements): e.g. concepts: "interior of x", "closure (completion, final, closure) of x", "outside of x", "x is open", "x is closed". Product: a product of any two open individuals is again open. Axioms: (...) I 97 Bowman L. Clarke: "Just as the linguistic domain of the classical individuals calculus is a complete Boolean algebra without zero-elements, our theorems are a closing-algebra without zero elements and without boundary elements. It is interesting that this much topology can be operated with as minimal assumptions. SimonsVsClarke: the idea of "removing" the boundary elements can be understood in two ways: a) that they "really exist" and we have an artificial limit by that I 98 (This would explain why the mereology is non-classical.) b) that these elements do not exist at all, then we miss the remainder principle (Principle Remainder, RP, see above). If we remove the interior (of a non-open individual), nothing will change! In fact, nothing is left. Closure/SimonsVsClarke: if we take any individual, its interior is a real part of its closure but there is no real part of its closure that is separate from the inside. So we have not even the weak supplement principle. We should therefore think that there are two types of individuals: a) "weak" (open) that do not touch anything and b) "strong" that are in contact with something. Nevertheless, we must not believe that there are any individuals who reconcile the difference again. We can distinguish individuals who differ only in one point but cannot determine the point. SimonsVs: this is not satisfactory. Nevertheless, if we want to perform topology without points and other limits, it is difficult to see how we can solve the problem. Solution/Simons: a philosophical approach must be more complex and allow vague approximations of sharp boundaries (> Menger, 1940, 107). |
Simons I P. Simons Parts. A Study in Ontology Oxford New York 1987 |
| Williams, B. | Putnam Vs Williams, B. | III 107 Ethics/relativism/Williams, Bernard: if an approximation of positions is really taking place in ethics, it is not because of a steering by the way things really are, whereas in the sciences this could actually be explained in this way. III 108 Reality/Williams/Putnam: We can select some of convictions of which one could say that they are maximally independent from our perspective. "The world as it appears to us" is interpreted as "the world as it appears to us in particular". For such a description only primary qualities should be chosen. III 263 footnote: Ex further terms can be derived therefrom: Ex "impulse" is defined by "mass" and "speed" whereas "speed" is defined by "time" and "location". III 109 How would we describe the world and imagine how it would be if there were no observers. In colloquial descriptions we could, of course, also include secondary qualities and speak of green grass and warm weather. According to Williams, we can readily be brought to the conclusion that we only describe how the grass had appeared to observers. Williams: thesis: our world (with observers) emerged from a world without observers. The laws are exactly the same. III 110 Therefore a description with primary qualities only should be possible. PutnamVsWilliams: enchanting, but it is true? Through evolution, no new laws of physics have emerged. But our predictions refer to phenomena that are described in the language of physics, not in the language of biology, psychology or economics. Once living beings and societies appear on the scene, actually new laws come to light, but they do not contradict the laws of physics. "Offer" and "demand" can not be described in terms of physics. III 128 Values/Williams/Putnam: even if it turns out that the color of a surface is an objective property of reflectivity, that does not impair the contrast between color characteristics and values, which Williams wanted to highlight. Putnam: but to demonstrate that the evaluation does not emerge from one eye from the nature of the eye, the complicated metaphysical explanations of Williams are unnecessary. Def values/Dewey: Evaluation results from the critique of various problem-solving processes. Absoluteness/Williams: contains ideally a "theory of knowledge and error"; contains both the possibility of the local views, as well as its own possibility. Is being eliminated virtually immediately by Williams: "this view of the world must enable to explain the possibility of their own existence". Later: withdrawal: "... which may be subject to the radical indeterminacy of interpretation ..." III 129 Austin: "this is the point at which the philosopher says it, and then comes the point at which s_he withdraws." III 130 PutnamVsWilliams: Problem: for the absolute conception, there is only one way to explain the possibility of local views and their own possibility: an prediction of future occurrences of characters and sounds. III 135 ff RelativismPutnamVsWilliams: the outright "truth of relativism" by Williams is not more coherent than the "absolute conception of the world". Williams/truth: rather carefree use of the term. Sometimes something that is "detected by the procedural manners of a linguistic community" (same perspective as Rorty, who Williams considers an opponent). II 136/137 Truth: According to Williams in the purely academic conflict "not really a problem." He believes that the members of other communities have ethical knowledge, and their beliefs are true, if they use their concepts carefully. PutnamVsWilliams: striking contradiction: Ex "right, her sitting together with her boss alone in the office is unchaste, but we do not consider chastity a virtue". In contradiction to Williams assertion that "true" and "false" could only be used in case of a real conflict. III 140 PutnamVsWilliams: Opposition: Williams would like to acknowledge the involvement of facts and values, and at the same time hold on to the "absoluteness" of scientific knowledge. Putnam: but that's impossible. It's not possible that science is absolute, but nothing else. I (k) 253 Norms/values/Bernard Williams: presumes the perspective of "some social world". On the other hand (according to Putnam) physics proposes an absolute metaphysical truth. PutnamVsWilliams: the talk of the "content" of a conviction that would be "perspective", is lacking any clear sense. That was grist to the mill of deconstructionism. Rorty VI 64 PutnamVsWilliams/Rorty: "approach to the big picture": purely dogmatic. The notion of absoluteness is incoherent. |
Putnam I Hilary Putnam Von einem Realistischen Standpunkt In Von einem realistischen Standpunkt, Vincent C. Müller Frankfurt 1993 Putnam I (a) Hilary Putnam Explanation and Reference, In: Glenn Pearce & Patrick Maynard (eds.), Conceptual Change. D. Reidel. pp. 196--214 (1973) In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (b) Hilary Putnam Language and Reality, in: Mind, Language and Reality: Philosophical Papers, Volume 2. Cambridge University Press. pp. 272-90 (1995 In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (c) Hilary Putnam What is Realism? in: Proceedings of the Aristotelian Society 76 (1975):pp. 177 - 194. In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (d) Hilary Putnam Models and Reality, Journal of Symbolic Logic 45 (3), 1980:pp. 464-482. In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (e) Hilary Putnam Reference and Truth In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (f) Hilary Putnam How to Be an Internal Realist and a Transcendental Idealist (at the Same Time) in: R. Haller/W. Grassl (eds): Sprache, Logik und Philosophie, Akten des 4. Internationalen Wittgenstein-Symposiums, 1979 In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (g) Hilary Putnam Why there isn’t a ready-made world, Synthese 51 (2):205--228 (1982) In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (h) Hilary Putnam Pourqui les Philosophes? in: A: Jacob (ed.) L’Encyclopédie PHilosophieque Universelle, Paris 1986 In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (i) Hilary Putnam Realism with a Human Face, Cambridge/MA 1990 In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam I (k) Hilary Putnam "Irrealism and Deconstruction", 6. Giford Lecture, St. Andrews 1990, in: H. Putnam, Renewing Philosophy (The Gifford Lectures), Cambridge/MA 1992, pp. 108-133 In Von einem realistischen Standpunkt, Vincent C. Müller Reinbek 1993 Putnam II Hilary Putnam Representation and Reality, Cambridge/MA 1988 German Edition: Repräsentation und Realität Frankfurt 1999 Putnam III Hilary Putnam Renewing Philosophy (The Gifford Lectures), Cambridge/MA 1992 German Edition: Für eine Erneuerung der Philosophie Stuttgart 1997 Putnam IV Hilary Putnam "Minds and Machines", in: Sidney Hook (ed.) Dimensions of Mind, New York 1960, pp. 138-164 In Künstliche Intelligenz, Walther Ch. Zimmerli/Stefan Wolf Stuttgart 1994 Putnam V Hilary Putnam Reason, Truth and History, Cambridge/MA 1981 German Edition: Vernunft, Wahrheit und Geschichte Frankfurt 1990 Putnam VI Hilary Putnam "Realism and Reason", Proceedings of the American Philosophical Association (1976) pp. 483-98 In Truth and Meaning, Paul Horwich Aldershot 1994 Putnam VII Hilary Putnam "A Defense of Internal Realism" in: James Conant (ed.)Realism with a Human Face, Cambridge/MA 1990 pp. 30-43 In Theories of Truth, Paul Horwich Aldershot 1994 SocPut I Robert D. Putnam Bowling Alone: The Collapse and Revival of American Community New York 2000 Rorty I Richard Rorty Philosophy and the Mirror of Nature, Princeton/NJ 1979 German Edition: Der Spiegel der Natur Frankfurt 1997 Rorty II Richard Rorty Philosophie & die Zukunft Frankfurt 2000 Rorty II (b) Richard Rorty "Habermas, Derrida and the Functions of Philosophy", in: R. Rorty, Truth and Progress. Philosophical Papers III, Cambridge/MA 1998 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (c) Richard Rorty Analytic and Conversational Philosophy Conference fee "Philosophy and the other hgumanities", Stanford Humanities Center 1998 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (d) Richard Rorty Justice as a Larger Loyalty, in: Ronald Bontekoe/Marietta Stepanians (eds.) Justice and Democracy. Cross-cultural Perspectives, University of Hawaii 1997 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (e) Richard Rorty Spinoza, Pragmatismus und die Liebe zur Weisheit, Revised Spinoza Lecture April 1997, University of Amsterdam In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (f) Richard Rorty "Sein, das verstanden werden kann, ist Sprache", keynote lecture for Gadamer’ s 100th birthday, University of Heidelberg In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty II (g) Richard Rorty "Wild Orchids and Trotzky", in: Wild Orchids and Trotzky: Messages form American Universities ed. Mark Edmundson, New York 1993 In Philosophie & die Zukunft, Frankfurt/M. 2000 Rorty III Richard Rorty Contingency, Irony, and solidarity, Chambridge/MA 1989 German Edition: Kontingenz, Ironie und Solidarität Frankfurt 1992 Rorty IV (a) Richard Rorty "is Philosophy a Natural Kind?", in: R. Rorty, Objectivity, Relativism, and Truth. Philosophical Papers Vol. I, Cambridge/Ma 1991, pp. 46-62 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty IV (b) Richard Rorty "Non-Reductive Physicalism" in: R. Rorty, Objectivity, Relativism, and Truth. Philosophical Papers Vol. I, Cambridge/Ma 1991, pp. 113-125 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty IV (c) Richard Rorty "Heidegger, Kundera and Dickens" in: R. Rorty, Essays on Heidegger and Others. Philosophical Papers Vol. 2, Cambridge/MA 1991, pp. 66-82 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty IV (d) Richard Rorty "Deconstruction and Circumvention" in: R. Rorty, Essays on Heidegger and Others. Philosophical Papers Vol. 2, Cambridge/MA 1991, pp. 85-106 In Eine Kultur ohne Zentrum, Stuttgart 1993 Rorty V (a) R. Rorty "Solidarity of Objectivity", Howison Lecture, University of California, Berkeley, January 1983 In Solidarität oder Objektivität?, Stuttgart 1998 Rorty V (b) Richard Rorty "Freud and Moral Reflection", Edith Weigert Lecture, Forum on Psychiatry and the Humanities, Washington School of Psychiatry, Oct. 19th 1984 In Solidarität oder Objektivität?, Stuttgart 1988 Rorty V (c) Richard Rorty The Priority of Democracy to Philosophy, in: John P. Reeder & Gene Outka (eds.), Prospects for a Common Morality. Princeton University Press. pp. 254-278 (1992) In Solidarität oder Objektivität?, Stuttgart 1988 Rorty VI Richard Rorty Truth and Progress, Cambridge/MA 1998 German Edition: Wahrheit und Fortschritt Frankfurt 2000 |
| |||
| Disputed term/author/ism | Pro/Versus |
Entry |
Reference |
|---|---|---|---|
| deductive- nomolog. | Versus | Cartwright I 100 Deductive-nomological model: Author: Hempel, Grünbaum, E. Nagel - may try to find causal relations between natural laws - Vs: Quine, Cartwright, Fraassen - CartwrightVs the inference direction general< special: does not correspond to practice - most of the equations can not be solved strictly. We need approximations. - I 105 these approaches are substitutes - I 127 arbitrary choice in approach and idealization is not dictated by facts. |
CartwrightR II R. Cartwright Ontology and the theory of meaning Chicago 1954 |
| |||
The author or concept searched is found in the following 6 theses of the more related field of specialization.
| Disputed term/author/ism | Author |
Entry |
Reference |
|---|---|---|---|
| Laws | Cartwright, N. | Cartwright: I have three connected arguments. Thesis 1: The obvious explanatory power of fundamental laws does not speak for their truth. Thesis 2: The way in which fundamental laws are used in explanations speaks for their falsity. We explain by ceteris paribus laws by merging causes by approximations that exceed what the fundamental laws dictate. Thesis 3: The appearance of truth comes from a bad explanatory model, I 4 that connects laws directly to reality. Cartwright instead: Def "Simulacrum-View"/Cartwright: from explanation: Thesis: the path from theory to reality goes like this: theory > model > phenomenological law. Phenomenological Laws/Cartwright: are true of the objects of reality (or can be). Fundamental Laws/Cartwright: are true only of the objects in the model. I 10 Asymmetry: Causal laws are asymmetric: Effect and cause cannot be interchanged. - On the other hand symmetrical: Laws of Association/Hume: e.g. length of shadow/height of mast. - Fraassen: Thesis: The explanatory asymmetries are not real. There is no fact about what explains what. CartwrightVsFraassen - Association/CartwrightVsHume: Association is not sufficient to distinguish between effective and ineffective strategies to fight malaria. I 51 Laws of Nature/Science/Cartwright: Thesis: There are no laws for cases where theories overlap. |
|
| Evolution | Kauffman, St. | I 238 Evolution / Optimization: / Kauffman solution: perhaps evolution could first design a redundant program or organism and then compress? I 240 Kauffman: Thesis: I suspect it cannot. The successive approximation to the minimum-length program is in fact only be useful if the program found at each level helps locate the shorter program at the next level. (> Fitness landscape). |
|
| Self-Organization | Kauffman, St. | Dennett I 303 Self-Organization/Kauffman/Dennett: Thesis: Evolution itself undergoes evolution. It develops because it is a forced move in the design game. Finding the right path is surprisingly easy - laws of design, not of form - inevitabilities of metatechnics - epistasis: interaction between genes: - aptitude landscape strongly determines development: successful results are sacrificed. Kauffman I 30 Kauffman's thesis: If the band of life were played again, the individual branches of the family tree of life might look different, but the patterns of branches, which initially diverge strongly and then become more and more a refining of details, probably follow a deeper regularity. Self-Organization/Kauffman: Thesis: these structures occur at all levels: from ecosystems to economic systems undergoing technological evolution. Thesis: All complex adaptive systems in the biosphere - from protozoa to economies - strive for a natural state between order and chaos. Great compromise between structure and chance. I 49 Thesis: The best compromises are apparently achieved in the phase transition between order and chaos. I 51 Chaos Edge/Kauffman: great similarity with the theory of "self-organized criticality": thesis: Per Bak, Chao Tang, Kurt Wiesenfeld. I 349 Self-Organization/Kauffman: Bak, Chao, Wiesenfeld, 1988: new theory: thesis: self-organized criticality. For example, a heap of sand on a table that is constantly getting bigger. I 350 Potency Law/Kauffman: many small and little large avalanches. For avalanches there is no typical size at all! It is also independent of the size of the triggering grain of sand. Catastrophe/Chaos/Kauffman: Equilibrium systems do not need massive triggers to start moving massively. I 366 Economy/Organization/Self-Organization/Kauffman: new researches (Emily Dickinson): Thesis: flatter organizations are more successful, split into fields, each striving to improve their own benefit. The trick is how to select the fields. (NK model). Fields can detect peaks. "Simulated annealing": Finding a good approximation method. ("Temperature" see below) I 415 Thesis: we can consider goods and services as strings that interact with other strings. |
Dennett I D. Dennett Darwin’s Dangerous Idea, New York 1995 German Edition: Darwins gefährliches Erbe Hamburg 1997 Dennett II D. Dennett Kinds of Minds, New York 1996 German Edition: Spielarten des Geistes Gütersloh 1999 Dennett III Daniel Dennett "COG: Steps towards consciousness in robots" In Bewusstein, Thomas Metzinger Paderborn/München/Wien/Zürich 1996 Dennett IV Daniel Dennett "Animal Consciousness. What Matters and Why?", in: D. C. Dennett, Brainchildren. Essays on Designing Minds, Cambridge/MA 1998, pp. 337-350 In Der Geist der Tiere, D Perler/M. Wild Frankfurt/M. 2005 Kau II Stuart Kauffman At Home in the Universe: The Search for the Laws of Self-Organization and Complexity New York 1995 Kauffman I St. Kauffman At Home in the Universe, New York 1995 German Edition: Der Öltropfen im Wasser. Chaos, Komplexität, Selbstorganisation in Natur und Gesellschaft München 1998 |
| Neuronal Networks | Pinker, St. | I 144 Rumelhart Thesis: simple neural networks can already handle much of human intelligence. Def Connectionism/Pinker: Thesis: the mind is a large network in which many hidden levels propagate backwards or perhaps a battery of similar or identical networks. The only difference between us and rats is our larger number of nets. Rules and symbols may be useful for a preliminary approximation to the processes in the net, because a psychology cannot follow the many millions of activation streams. |
|
| History | Rorty, R. | VI 355 History of Philosophy/interpretation/past/change of meaning/change of concepts/Rorty: Dilemma: either, a) we impose on the dead, in an anachronistic manner, sufficient problems and vocabulary of our time to make them interlocutors, or b) we limit our interpretive activity to making their false sentences less foolish by placing them in the context of the backward times in which these sentences were uttered. This is not a dilemma at all: we can do both, but we must strictly separate them! >Interpretation/Bennett VI 361 Interpretation/Rorty: Thesis: with such attempts at approximation, of course, anachronistic methods are used. But if this happens consciously, there is nothing wrong with it. VI 375 Interpretation/Reconstruction/History/Philosophy/Rorty/RortyVsCanon: Thesis: RortyVsHistory of Philosophy as Doxography: triggers despair in us. There is no possibility to carry out a real rational reconstruction through hierarchy and the eternally same "important philosophers" and "important problems". The thinkers are being robbed of their salaries. Problems are imposed. Mistakes: to present Leibniz and Hegel, Mill and Nietzsche, Descartes and Carnap as if they were talking about the same topics. Mistakes: half-hearted attempt to tell development only from today's perspective. VI 395 Jonathan Rée/Rorty: Thesis: "No convincing reasons have been given for the importance of a historical consciousness for the philosopher". VI 395 Margaret Wilson/Rorty: Thesis: "It could well be that the real opinions of the great philosophers, if one compares them with those they have in mind, turn out to be rather meager." Rorty pro. |
|
| Convergence | Willams, B. | Rorty IV 32 Convergence / Bernard Williams thesis in scientific investigations, there should be in the ideal case, a converging approximation to an answer. |
Rorty I Richard Rorty Philosophy and the Mirror of Nature, Princeton/NJ 1979 German Edition: Der Spiegel der Natur Frankfurt 1997 Rorty VI Richard Rorty Truth and Progress, Cambridge/MA 1998 German Edition: Wahrheit und Fortschritt Frankfurt 2000 |
| |||