SEMIOTIC MACHINES. Theory, Implementation, Semiotic Relevance
(Final Version)
(Last changes Dec-4 1995)
EVENT: 8th International Semiotic Congress of the German and the Netherlands Semiotic Societies
SECTION: Logic and Philosophy of Science
LOCATION: Amsterdam, University
DATE: August 5-9 1996
ORGANIZATION: Dr. Gerd Döben-Henisch

OBJECTIVES

This workshop is considered as a first exploration of the new paradigm of semiotic machines. Using new tools provided by the philosophy of science approach and by the computer sciences we want to discuss the design and the construction of semiotic machines.





Possible Theoretical Frameworks

From a semiotic point of view there are at least the following basic types of theoretical frameworks conceivable:
  1. An empirical framework is given if the theory-builder uses as source of his/her data only those which are intersubjectively reproducible. His/her formal theory functions then as a kind of representation and explanation of the general structures implicitly given with the empirical data.
  2. A mathematical framework is given if the theory-builder uses as source of his/her data no special data 'outside of the theory'. The formal theory and their properties are the data themselves.
  3. A philosophical framework is given if the theory-builder uses as source of his/her data his/her consciousness without any restrictions; his/her formal theory functions then as a kind of representation and explanation of the general structures implicit in his/her subjective experiences.
Which one is chosen depends on your special objective. From a metatheoretical point of view empirical and mathematical theories are special cases of philosophical theories because everything which is part of a mathematical or an empirical theory can be part of a philosophical theory, but not vice versa.



Semiotic Machines as Simulations of Theories

Independent of the concrete 'nature' of semiotic machines we assume, that semiotic machines are always simulations of theories. The kind of material out of which semiotic machines are realized, doesn't matter. For the theoretical discussion counts only the theory, on which such a machine is based, as well as the relationship between the theory and the concrete machine. The machines function solely as a kind of 'illustration' of the theory. Complex dynamical sign-processes should be made perceivable by semiotic machines and thereby improving the mediation of the theory to the public. You will be able to criticize theories although you are not an expert of the theory itself.



Elements of a 'Semiotic Machine'

To speak about semiotic machines we need some kind of an empirical/ mathematical/ philosophical theory. We don`t have such theories yet, we want to develop one (or more than one). Therefore we cannot give the outline of such a theory here. But what we can do is to give a list of some properties of which we think are essential for such a targeted theory of semiotic machines. It will be part of the discussion whether this list is suitable or not; perhaps one can develop some taxonomy of properties and thereby outlining implicitly a whole hierarchy of semiotic machines.

The methodological status of the properties listed below is therefore pre-theoretical; we are giving these properties as a 'guide' to determine in advance possible classes of semiotic machines which then have to be elaborated more thoroughly.

  1. 'Learning' and 'be situated'

    Whatever a semiotic machine will be able to do in the future, what it should be able to do in any case is (1) to learn any kind of sign-systems, which humans are able to learn, and (2) to be able to apply a learnt sign-system with respect to the actual situation, which it actually shares, and with respect to some knowledge it has accumulated during its time of operation.

  2. 'Sign' and 'Signfunctions' as a starting point

    To define such a system, which humans are able to learn, one needs a suitable concept of a 'sign' and of a 'sign function', i.e. we assume, that the concept of semiotic machines should be centered around the semiotic concept of 'sign' and 'sign function'.

    An object x shall be called a sign, if it is the element of a syntactical defined system of expressions E and there exists a mapping S from the power-set pow(E) of these expressions into the power-set of content-structures pow(C), thus the range of the mapping S is a subset M of pow(C) and this subset M is called the set of meanings of E under S in C. The mapping S:E ---> M is called a sign-function. An element of E is called an expression or a sign vehicle. An element of M is called a sign value or a meaning (This model has strong relationsships to the triadic sign-relation of Peirce, the dual-sided concept of Saussure, and the structural approach of Hjelmslev).

  3. A 'Learning function' for sign functions

    We assume that sign functions are dynamic functions, i.e. they can be changed several times. To enable this we presuppose another function L which is able to do this changes. The function L maps sign functions and some additional values V into sign functions: L: S x V ---> S. Therefore L behaves like a learning function which can be influenced by V (hence the meaning of an expression is 'arbitrary' with respect to the function L).

  4. Meaning has to be distinguished from actual world experience; the 'birth' of the semiotic machine

    To relate meanings to actual states of the world one has to distinguish between meaning and world: that a complex expression x of pow(E) generates some complex meaning S(x) of M does not necessarily tell us anything about the actual states of the world W which we assume as an environment of L and S. Whether S(x) corresponds by some criterion A to some actual states of W has to be decided independently of S. Thus we need a special correspondence function A, e.g. A: rn(S) x RAS x IS ---> [0,1], i.e. A maps the range of S together with the representations of actual states RAS of W and perhaps some internal states IS related to S and L into the real interval [0,1]. The value of A tells us, to which degree there is some kind of correspondence between RAS and S(x).

    One way to speak of an environment W relative to S and L, is to assume semiotic entities called semiotic machines SM, which have an internal structure containing at least the functions S and L (A semiotic machine can be seen as the interpreter in the pragmatic theory of Charles Morris).

    If one wants to describe a 'communication' of states and state-changes of W with a semiotic machine one needs one more mapping from W onto sensorical states SS of SM which in turn can then be used from the semiotic machine for further 'processing': Stim: pow(W) ---> pow(SS) with 'Stim' as the stimulus function of W.

    If one wants, in the reverse case, to communicate states and state-changes of the SM with the world W, one needs another mapping from those states of SM, which shall be communicated, into some states of W, e.g. Reac: pow(BS) ---> pow(W) with 'Reac' as a reaction function of W and BS as a set of body states of the SM which 'induce' by Reac some changes in W.

    How to build up the set of representations of actual states RAS of W 'within' a semiotic machine SM is an open question: there are arguments to suggest that the representation RAS of actual states AS as those states of W, which induce certain sensory states in a semiotic machine SM, have to be distinguished from the sensory states SS; probably one has to assume a special representation function which generates a complex representation, stimulated by the sensory states, but enriched with additional information from other 'sources' within the semiotic machine, e.g. Stim: pow(W) ---> pow(SS), Xgen: pow(SS) x IS ---> pow(ST) and ELEM(RAS, pow(ST)).




IMPORTANT DATES

    February 15, 1996: Submissions deadline for title, abstract, and outline of paper
    March 1, 1996: Notification of acceptance or rejection
    September 30, 1996: Printable versions received


This schedule makes it possible that the final version of the article can be finished AFTER the discussions of the workshop! This gives the opportunity to incorporate interesting critics and suggestions from the discussions.



SUBMISSION DETAILS

Abstract (max. 300 words) and outline of paper (2-5 pages) have to be written in English. The abstract can be accompanied by 2-5 bibliographical references and has to be accompanied by the address of the authors. The papers can be transmitted as ASCII-Files on a PC-/Mac-Disk or as EMail.

Final papers have to be be written in English. They can have up to 25 pages including pictures and references. The papers should be transmitted as ASCII-Files on a PC-/Mac-Disk or as EMail and must have the PostScript-Format


FIELDS

  1. Formal theories describing complete semiotic machines, probably highly experimental and not complete, or only subproblems within an outlined general framework.
  2. Examples of concrete semiotic machines.
  3. Contributions to the mapping problem between formal theories and concrete machines.
  4. Contributions related to the relevance of semiotic machines for semiotics as a science as well as to classical positions in the history of semiotics.
  5. Others.


POSSIBLE TOPICS

Generally: all topics mentioned below can be worked out either as behavioral (including physiological) models, as phenomenological, or as purely mathematical ones. The necessary condition to be accepted is only that they must meet the objective to explain structures/ processes in the context of sign processes similar to the sign processes outlined above, and that they are presented as theories; these theories should be presented in a way that scholars of different disciplines have a chance to understand the main idea.
  • The philosophy of science point of view
  • The epistemological point of view
  • Concept formation triggered by perception
  • Generation of meaning structures triggered by sign expressions
  • Meaning structures and their relationship to the experience of the world
  • Meaning structures and planned behavior
  • The role of drives/ desires/ emotions in the generation of meaning structures
  • The role of memory within meaning
  • Role of the actual situation with regard to memory, planning and meaning.
  • The concept of space within meaning and world experience
  • The concept of time within meaning and world experience
  • Reflection as an overall implicit structure of semiotic machines
  • Types of inference processes within world experience, meaning, planning and language learning
  • A formal model of the learning function L
  • Software architecture of semiotic machines
  • Semiotic machines distributed in networks
  • The world interface of semiotic machines as constraint for possible worldexperience
  • Formal mechanisms to represent formal theories for semiotic machines
  • Formal mechanisms to realize concrete semiotic machines
  • Relationship between formal theories and concrete machines
  • Relationship between traditional concepts of sign/ sign function and the concept of semiotic machines
  • Semiotic machines and pragmatics
  • Possible implications of semiotic machines for semiotics as a disciplin
  • Semiotic machines related to empirical theories
  • Semiotic machines related to philosophical theories
  • others




PUBLICATIONS

We will have the opportunity to publish the papers printed in the series INM-Reports with an ISBN-number.



LOCATION

The workshop will be held as part of the 8th International Semiotic Congress of the German an Netherlands Semiotic Societies, which will take place at the University of Amsterdam (Netherlands).



ORGANIZATION of the WORKSHOP

Institut für Neue Medien
Dr. Gerd Döben-Henisch
Daimlerstr.32
60314 Frankfurt
TEL: 069-941 963 -0
TEL-Dir: 069-941 963 - 10
FAX: 069-941 963 - 22
EMail: doeb@inm.de
http://www.inm.de/kip/kip.html ---> Semiotic Machines




ORGANIZATION of the CONGRESS

Universiteit van Amsterdam
Film- en Televisiewetenshap
Dr.Jürgen E.Müller
Nieuwe Doelenstraat 16
NL-1012 Amsterdam CP
TEL: 020-525 2995/ 2980
FAX:020-525 2938


INM

Daimlerstrasse 32, 60314 Frankfurt am Main, Deutschland. Tel +49- (0)69-941963-0, Tel-Gerd: +49- (0)69-941963-10