2. Knowledge = Skill

Gilbert Ryle, Knowing How and Knowing That.

The terms ‘orthodox cognitivism’ or ‘computationalism’ is taken here in the sense of Fodor, JA (1981) Representations: Philosophical Essays on the Foundations of Cognitive Science. Brighton, Harvester Press.

This is a generic term for approaches to the study of issues pertinent to the design and application of computer-based technologies which differ from what is termed a machine centred approach. The human-centred approach is best explained in a comprehensive report by Dr. KS Gill, ref: Summary of Human-Centred Systems Research in Europe, SEAKE Centre, Brighton University, 1990. The idea of human-centredness has now become part of mainstream thinking but with a focus on ‘technology’ rather than ‘system’.

‘Theoretical knowledge’, in this context, is analogous to ‘knowing that’, as distinct from ‘knowing how’, cf. Ryle (1949). It has been described as mathematically structured knowledge (cf. Göranzon 1992a)

NB. some of these authors, particularly those from the domain of the sociology of scientific knowledge (SSK) critique the descriptions/formal language of science, contrasting them to its practice. They do this by either undertaking apprenticeship in the specific scientific area, and/or by recording its everyday discourse. For further literature, in addition to Woolgar (1988) and Collins (1974, 1975), see for example, Gilbert and Mulkay (1984). Rosenbrock (1988) is critical of the application of the formalist notion of science (as in propositional knowledge) to the domain of engineering at the expense of sensibility and human purpose. This is discussed later, herein.

Rationalism denotes the emphasis on objectivity in terms of the mind-body split., cf. Descartes (1639), Hume (1748), Kant (1929). Toulmin (1991) discusses the ‘seventeenth century Cartesian ideal of intellectual exactitude, … it’s idolization of geometrical proof’, citing the ‘rationalist procedures of Newtonian mathematics’ (ibid. p. 41) and compares this ideal to that of the sixteenth century humanists, cf. Bacon, who believed in a ‘humanly fruitful’ science. The contrast lies between an emphasis on theory and on praxis.

Johannessen (1988a) discusses the higher status in society placed upon analytical and empirical validity for one’s having knowledge, versus the doing of knowledge, i.e. its practice. He traces this back to logical positivism, cf. the Vienna circle, Carnap (1956), Chaps. II and III.

As in scientific law for nature, for example, Galileo: ‘The book of nature is written in mathematical symbols’; Newtonian physics – ‘a vote in favour of theoretical cosmology, not for practical human dividends’, cf. Toulmin (1991). Toulmin describes Leibniz’s characteristica universalis (universal system of characters), published in 1666, as the dream for a universal unambiguous, context independent, language, and thereby knowledge.

For an outline of the development of the dream of exact language see Toulmin, op. cit. He discusses Leibniz’s (op. cit.) project to develop a language based on mathematical symbolism. Johannessen also cites this period and the idea of an exact language in a discussion on the rule and the positivist conception of knowledge, in Johannessen, K (1988a) Rule-Following and Tacit Knowledge, in AI & Society Journal, Vol. 2.4. Springer-Verlag, London.

Leibniz, see footnote 9.

Toulmin, op. cit. p. 36.

Boden, op. cit. p. 32.

See An Enquiry Concerning Human Understanding, first published in 1748.

AJ. Ayer, (1971) Language truth and logic. Pelican.

Johannessen, op. cit. p. 290.

See Boden (1991), p. 10. This is in contrast to developments in parallel processing and connectionism (see for example, Rumelhart and Norman 1986) which have been seen to be a challenge to the conventional model based on the von-Neumann machine (cf. Fodor and Pylyshyn 1988). However, all these various approaches of AI and computational psychology have originated from the same mid-century ideas about the brain’s logical-computational potential. See also Boden (1988) for a comprehensive overview of differing perspectives in AI and cognitive science.

cf. Liebnitz, op. cit., see Toulmin, op. cit.

See footnotes 9 and 10. Of course, quantum mechanics challenges this idea, as it recognises contingency in the world of physics. Building a quantum computer is the next big challenge.

See Johannessen op. cit.; see Ayer op. cit.

Johannesen, op. cit. p. 288.

It is important to note that the logical positivistic view of knowledge has been traced as far back as Plato’s dialogues about expertise and skill and his conception of rationality. NB the historical context for understanding this thinking:

cf. H. L. Dreyfus (1989), Is Socrates to Blame for Cognitivism? p. 219. Johannessen, op. cit; Plato (1975). The Republic. Penguin, 2nd ed. Translation, D. Lee; Toulmin, op. cit.

F.J. Varela (1988) Cognitive Science: A Cartography of Current Ideas. Editions du Seuil, Paris. See also F.J. Varela et al. (1991). Fodor and Pylyshyn (1988) describe ‘classical cognitive science’ as an extended attempt to apply the methods of proof theory to the modelling of thought; Pylyshyn (1984) describes cognition as a type of computation.

Boden, op. cit. p. 30.

See for example, the classical AI paradigm works of Fodor (1976), Newell (1980), Newell and Simon (1972), Pylyshyn (1984), Michie and Johnston (1984), Minsky (1963), Feigenbaum and McCorduck (1983). Also see The Handbook of AI, (eds.), Barr and Feigenbaum (1981).

Classic AI programming languages such as Prolog (Kowalski 1974, 1982), and PopLog (Sloman 1989) are based upon systems of logic. Prolog is based on predicate logic, and PopLog explores computational linguistics.

NB. Newell (1981) discusses the significance of Polya’s (1945) idea of heuristics which greatly influenced the work on problem solving in AI. Polya, himself was a mathematician, not directly involved in AI or cognitive science. AI drew upon his heuristics but omitted those aspects of his theory which were ‘ambiguous’ or intangible, for example, his emphasis upon personal motivation or commitment which he considered to be essential for effective problem solving.

Varela, op. cit. p. 18–21.

Boden, op. cit. p. 20.

It is noted that Boden has questioned whether this model is purely formal-syntactic when she criticises Searle’s assumption of this in his argument about translation and the Chinese box. Searle (1980) argues that a computer cannot understand. Understanding entails intentionality. A computer cannot have intentionality because it is a formal syntactic system. He therefore argues that AI ideas or computational theories based on them cannot describe or explain mental processes. Boden argues, however, that any simple program has some semantic properties and that computational theories are not essentially incapable of explaining meaning. This is because explanation involves assimilation of something to something else which is analogous to it but not identical with it. Searle also argues that the physical nature of the brain produces intentionality and that on intuitive grounds, the metal and silicon cannot. Boden argues that the brain’s ability to generate intentionality is intelligible only through its information processing capability. The computer has such a capability.

This is distinct from current approaches to knowledge engineering which embody a humanistic/interdisciplinary approach to design. TKE is based on the ideas of Feigenbaum and ‘traditional expert systems design’; see Feigenbaum, EA and McCorduck, P (1983) The Fifth Generation. Addison-Wesley, MA.

For example, see cf. Bramer (1988), Ostberg et al. (1988), Whitaker and Ostberg (1988), Hayes-Roth (1984, 1985)

Lipscombe (1989) provides a critique of the development of expert systems in medicine and suggest that approaching the design/purpose of these systems to facilitate problem analysis rather than problem solving will provide for more effective tools. See also Alvey (1983).

Partridge (1987) puts this down to the explanation problem in ES technology which is inherently limited to the relatively static and relatively context free domain of abstract technical expertise. In this article, he strongly recommended ES designers to concentrate on ‘low road’ applications and stay away form complex ‘high-road’ ones, otherwise there will be a software crisis.

See for example, Sprague, RH and Watson, HJ (1986), Jones and Walsham (1992).

Cf. Polanyi (1966) The Tacit Dimension. New York: Doubleday, p. 4.

G. Ryle (1949) The Concept of Mind.

Polanyi, op. cit. p. 10.

Ibid. p. 10.

Polanyi has developed his theory of personal knowing in his work, Personal Knowledge: Towards a Post Critical Philosophy (1962), first published in 1958.

As in the case of control systems design.

Rosenbrock is concerned that the notion of the automated machine allows for the idea of a workerless factory. It rejects the tacit dimension of worker’s knowledge and skill, seeing them as automata, reducing their tasks to ever simpler and more closely definable fragments (cf. Fordism). He cites Needham’s view in the 1920s (cf. 1927) that ‘in science, man is a machine; or if he is not, then he is nothing at all’.

Rosenbrock, H (1988) Engineering as an Art. AI & Society Journal, Vol. 2 No. 4.

Rosenbrock was involved in a European venture upon Human-Centred CIM Systems. [CIM – Computer Integrated Manufacturing]. This project attempted to apply the symbiotic model in the construction of a human centred CIM system.

Cf. Cooley 1987a, p. 12.

Cooley, M. (1987a), Architect or Bee. Hogarth Press [new edition], p. 11. cf. cybernetics, see Wiener (1949).

Ibid. fig. 2.1. ‘The tacit Area’. p. 12.

Feigenbaum and McCorduck, op. cit.

The terms, ‘holistic similarity recognition’, intuition’, or ‘know-how’, have been adopted from Dreyfus and Dreyfus. They use them synonymously to denote, ‘the understanding that effortlessly occurs upon seeing similarities with previous experiences … intuition is the product of deep situational involvement and recognition of similarity’. They, in turn, have drawn upon Polanyi (1966).

Dreyfus and Dreyfus (1986) present this balance between analytical thinking and intuition in their theory of skill acquisition. Here, there are five stages of development from being a novice to gaining expertise. These are (1) novice; (2) advanced beginner; (3) competent; (4) proficient; and (5) expert. A novice follows context-free rules. The relevant components of a situation are defined to enable the novice to recognise them. The novice lacks any coherent sense of the whole task and can judge his/her performance only in terms of how well the learnt rules have been followed. The advanced beginner, through practical experience in concrete situations, learns to recognise situational elements, which cannot be defined as objective context-free features. Recognition occurs through discerning similarities to prior examples. The competent learner can make plans and choices in order to achieve certain outcomes. Being proficient means that one has acquired an intuitive ability to use patterns without reducing them to features. The competent performer still relies on analytical thinking in combination with intuition. An expert, however can discern whole scenes without decomposing them into elements. Because of their experience they can use their intuition to make decisions and cope with uncertainties and critical situations. This framework has come under criticism from Harry Collins (2013) as not representative of other ways of learning that may involve more or less steps as in the case of touch typists.

The hermeneutics of Goranzon and Josefson are based on interpretations of Wittgenstein, Polanyi, and Kuhn, whereas the hermeneutics of Ehn covers interpretations of Heidegger and Marx as well.

Two additional aspects of tacit knowledge, based on an interpretation of the Scandinavian, (cf. Gill 1988, 1990, 1992) are ‘experiential knowledge’ and ‘personal knowledge’. Personal knowledge is that which is specific to the individual personality, gained from our personal life experiences (like family culture, school, friends), and expressed in our social values, beliefs etc.; experiential knowledge is that of direct experience which can cover specific experiences, such as within the context of the workplace, e.g., knowledge gained from interaction with work colleagues, group culture, organisational culture etc.

See Janik (1988) for an overview of the hermeneutic approach adopted in particular by the Scandinavian researchers based at the former Swedish Centre for Working Life.

They cover the fields of ethics, morality, and aesthetics, cf. Nordenstam (1968, 1987, 2013), Johannessen and Nordenstam (1981), and Johannessen (1981).

See Fann (1969) for a discussion of this shift.

Johannessen (1988b) The Concept of Practice in Wittgenstein’s Later Philosophy, in Inquiry.

Johannessen, K (1992) Rule-Following, Intransitive Understanding and Tacit Knowledge. pp. 41–63.

Wittgenstein, L (1958) Philosophical Investigations. para 199, p. 81.

Wittgenstein, L (1977) Remarks on Colour. (ed) G.E.M. Anscombe. para 317.

Johannessen, K (1988b).