Socially Sensitive Computing A necessary Paradigm Shift for

Socially Sensitive Computing: A necessary Paradigm Shift for Computer Science “Rules are for the obedience of fools and the guidance of wise men” Tom Addis Computer Science & Software Engineering Douglas Bader (1910 – 1982) University of Portsmouth Dave Billinge Creative Technology University of Portsmouth Bart-Floris Visscher Computer Science & Software Engineering 30 th March 2004 GC 7 Non-Classical Computingof Portsmouth University

Formal Semantics n Church-Turing Thesis: n Equivalent Representation Systems n n Computer program (Turing Machines) Functional Statements (Functional Machines) Tractatus Propositions (Predicate Machines) It follows that: n n Everything is potentially unambiguously describable All sets are rational (countable) Set membership is always specifiable and context independent or has an explicit context Fuzzy (ordinal) and probabilistic sets (ratio of integers) are countable 30 th March 2004 GC 7 Non-Classical Computing 2

Consequences of Formal Model n Practical n n n Any set of names can be used in a program to represent a proposition. There is an infinite but bounded set of possible organisations of a program. There is a such a thing as a ‘minimum program’ Programs can only have one interpretation. Social n Rules can be constructed that can describe unambiguously any situation. Thus: n n Rules can bypass human judgement. There is only one correct way to see the world 30 th March 2004 GC 7 Non-Classical Computing 3

Problem 1 n Programs n can only have a single interpretation. But programs have at least two interpretations The Computer State n The Problem Domain n Problem Domain Program Computer States (bits) 30 th March 2004 GC 7 Non-Classical Computing 4

Formal Interpretation n Mapping onto Wittgenstein’s Objects. Independent The only rational n Atomic interpretation of a n Exist program in all possible worlds n Immaterial n Indescribable Computer States (bits) n Self governed n Program The bit has all these properties n Such objects are rarely in the Problem Domain n 30 th March 2004 GC 7 Non-Classical Computing 5

Problem 2 n Everything is NOT potentially unambiguously describable There also irrational sets (not countable). Chair. Specification: A Chair n Some sets depend upon human usage and Designedspecificallytotobe besat satupon, upon context. Standson onits itsown n Jean-Francois Dupris Hasfourlegs n Examples: Hasaaback n Games, chairs and life Sitter’s Feet touches floor 30 th March 2004 GC 7 Non-Classical Computing 6

Current Status of Computer Science n We have computer programs with a semantics based upon computer bits. and n We create programs that cannot rationally be assigned meaning to the very problem domain for which we write them. 30 th March 2004 GC 7 Non-Classical Computing 7

Computing with Irrational Sets 1 n Programs must remain in the domain of rational sets. n We have the freedom to use the program’s accidental properties: The choice of names n The choice of program organisation n n The choices are used to provide a semantic link with the problem domain. 30 th March 2004 GC 7 Non-Classical Computing 8

Practical Consequences n n n Choice of program names and organisations must be flexible. The dynamics of program names, organisations and assignment of meaning must be linked to a social system. A method for a ‘minimum program’ is helpful. The problems underlying such issues as natural language understanding will be significantly reduced. There will be the possibility of having a truly creative machine. 30 th March 2004 GC 7 Non-Classical Computing 9

Computing with Irrational Sets 2 Problem Domain Contexts allows the use of rational sets Names & Organisation The problem domain contains irrational Social sets. Minimum Program sensitive So we cannot use denotational semantics feedback Computer States (bits) 30 th March 2004 GC 7 Non-Classical Computing 10

Social Consequences n Since rules cannot be devised that are unambiguous n there will always be a need for human judgement. n There will always be a wide range of perceptions of a problem domain n n that cannot be predicted and hence planned for. 30 th March 2004 GC 7 Non-Classical Computing 11

Two Semantics of Irrational Sets n Wittgenstein: is? h Use family resemblance insteaddof sets. t o u o y o d w o n Use word usage (and structures) instead of h s i e g n e l l a reference. The Ch n n Lakoff (and Johnson)1: Use prototypes (paradigms) n Use metaphor instead of reference n [1] Lakoff and Johnson (1980) – Metaphors we live by Lakoff (1986) – Women, Fire and Dangerous Things 30 th March 2004 GC 7 Non-Classical Computing 12