Semionics: A Proposal for the Semiotic Modeling of Organizations

1
Semionics A Proposal for the Semiotic Modeling
of Organizations

• Ricardo Ribeiro Gudwin
• DCA-FEEC-UNICAMP

2
Semiotics and Semionics

• Semiotics
• Science which studies the phenomena of
  signification, meaning and communication in
  natural and artificial systems
• Main artifact the sign
• Tries to model any kind of phenomena as being a
  sign process
• Natural Systems
• Semiotic Analysis
• Artificial Systems
• Semiotic Analysis
• Semiotic Synthesis
• Semionics
• One particular proposal for semiotic synthesis

3
Diadic Model of the Sign
4
Triadic Model of the Sign
5
Semiotics x Semionics
Sign
Interpretant
Object
6
Semiotics x Semionics
Interpreter (Semionic Agent)
Sign (Signlet)
Interpretant (Signlet)
R1 (e.g. symbolic)
R2 (e.g. iconic)
Object
7
Exosemiotics and Endosemiotics
Exosemiotic View
Interpreter (Semionic Agent)
Sign (Signlet)
Interpretant (Signlet)
Internally
Endosemiotic View
8
Endosemiotic Process Modeling

• From the point of view of Semiotic Synthesis
• Endosemiotic understanding of the interpreter is
  very much important !
• Exosemiosic Process
• Composed of many intrincate endosemiosic
  processes
• Complex network of semiosic processes occurring
  in parallel and in real time
• If we want to model (and build) such an
  endosemiotic system
• We need a modeling artifact able to support these
  requisites
• Discrete event dynamics
• Concurrent processes
• Petri Nets

9
Endosemiotic Process Models

• Petri Nets are not enough !
• Tokens are unstructured and transitions have no
  processing capabilities
• Coloured Petri Nets (Object-based Petri Nets)
• Tokens are structured
• Transitions have (some) processing capabilities
• Coloured Petri Nets (Object-based PN) are not
  enough !
• Do not differentiate among tokens
• Tokens which are interpreters
• Tokens which are signs
• Solution
• Create a new extension of a Petri Net
• Semionic Networks

10
Semionic Network Action
Semionic Agent (micro-interpreter)
Signlet (sign)
Signlet (interpretant)
11
Semionic Network Decision
??
??
12
Semionic Agent

• Two Tasks
• Decision
• Choose which sign it is going to interpret
• Decide what is going to happen to it (preserved
  or not)
• Action
• Turn it into an interpretant
• Decision
• Evaluation Phase
• Attribution Phase
• Action
• Assimilation Phase
• Generation Phase

13
Signlets

• Split into compartments
• Organized into classes, according to compartment
  types

Signlet
Data or Function
14
Semionic Agents are Signlets

• Compartments
• Sensors
• Effectors
• Internal states
• Mediated Transformation Functions
• Evaluation
• Transformation

S1
E1
I1
S2
E2
I2
I3
F1
F2
eval
perform
eval
perform
15
Evaluation Phase

• Evaluation Phase
• Starts when a given semionic agent sets up to
  which signlets it is going to interact to
• The semionic agent must evaluate each available
  signlet and decide what it is going to happen to
  it after the interaction
• For each transformation function available at the
  semionic agent
• A set of interacting signlets of the right kind
  is determined
• The semionic agent tests all possible
  combinations of available signlets which can be
  compatible to the inputs of its transformation
  functions

16
Evaluation Phase

• Enabling Scope
• Each possible combination which is compatible to
  a given transformation function
• List of signlets potentially available for
  interaction
• Evaluated by means of an evaluation function
• Should determinate if signlets are to be
  modified, returned to their original places or
  destroyed
• The Phase ends when
• The semionic agent evaluates all available
  enabling scopes and attributes to each one an
  interest value and a pretended access mode
• The pretended access mode describes the semionic
  agents intentions to each input signlet. It
  should inform if the semionic agent pretends the
  sharing of the signlet with other semionic agents
  and if it intends to destroy the signlet after
  the interaction

17
Evaluation Phase
??
Signlets
DESTROY ?
??
Semionic Agent
??
??
F1 ??
??
??
F2 ??
??
Fn ??
??
SHARE ?
WHICH F ?
18
Attribution Phase

• Attribution Phase
• A central supervisor algorithm gets the
  intentions of each active semionic agent and
  attributes to each of them an enabling scope
• This attribution should avoid any kind of
  conflict with the wishes of other semionic agents
• Many different algorithms can be used in this
  phase
• For test purposes, our group developped an
  algorithm (Guerrero et. al. 1999), which we
  called BMSA (Best Matching Search Algorithm),
• Attributes a signlet to the the semionic agent
  that best rated it, respecting the pretended
  access modes of each semionic agent

19
Assimilation Phase

• Depending on the Access Mode
• Read Get a reference to a Signlet, so it can
  have access to its internal content
• In this case, the semionic agent is supposed not
  to change the internals of the signlet
• Get Fully assimilate the input signlets,
  becoming the owner of it
• In this case, the semionic agent is allowed to
  further process it
• After assimilating the necessary information
• Leave the signlet in its original place
• Destroy it permanently (consume it)
• Take it from its original place in order to
  process it

20
Generation Phase

• Generation Phase
• Get available information
• The information collected from input signlets is
  used to generate a new signlet or to modify an
  assimilated signlet
• Process it
• Any kind of transformation function can be
  applied in order to generate new information
• Send it to outputs
• Signlets are sent to their corresponding outputs

21
Special Cases

• Sources
• In this case, the internal functions dont have
  inputs, only outputs
• The result is that signlets are constantly being
  generated and being inserted into the semionic
  network
• Sinks
• In this case, the internal functions dont have
  outputs, just inputs
• These semionic agents are used to take signlets
  from the network and destroy them
• Sources and Sinks can be used to link a semionic
  network to external systems

22
SNToolkit The Semionic Networks Toolkit
23
SNToolkit The Semionic Networks Toolkit
24
Organizational Processes

• Organization
• Network of Resource Processing Devices performing
  a purposeful role
• Resources
• Abstract concept that can be applied to many
  different domains of knowledge
• May have an associated value or cost, which
  can be used on the models being developped
• Kinds of Resources
• Passive Resources (materials or information)
• Active Resources (processual resources)

25
Organizational Processes

• Passive Resources
• Information
• Texts, documents, diagrams, data, sheets, tables,
  etc
• Materials
• Objects, parts, products, raw-materials, money,
  etc..
• Active Resources (Processual Resources)
• Execute activities of resource processing
• Mechanic (Without Decision-making)
• Intelligent (With Decision-making)
• Examples
• Machines, Human Resources (Workers), etc

26
Organizational Processes and Semionic Networks

• Organizational Processes
• Can be described in terms of sign processes
• Organizational Semiotics
• Resources
• Can be modeled in terms of signlets and semionic
  agents
• Passive Resources signlets
• Active Resources semionic agents
• Networks of Resource Processing
• Can be modeled in terms of Semionic Networks
• Both Intelligent and Mechanical Active Resources
• Can be modeled in terms of semionic agents

27
Organizational Processes and Semionic Networks

• The Interesting Case Intelligent Active
  Resources
• Mechanical Processes can be easily modeled by
  standard Petri Nets
• From Peircean Semiotics
• Notions of Abduction, Deduction and Induction
• Abduction
• Generation of newer knowledge structures
• Deduction
• Extraction of explicit knowledge structures from
  implicit knowledge structures
• Induction
• Evaluation of a given knowledge structure in
  terms of the system purposes

28
Organizational Processes and Semionic Networks

• Semionic Agents
• Are able to perform decision-based actions
• Coordination Between Evaluation and
  Transformation Functions
• Allows a semionic agent to perform the three main
  semiosic steps abduction, deduction and
  induction
• The coordinated work of many semionic agents
• May allow the representation of full semiotic
  processes
• In this sense
• We say that the actions performed by semionic
  agents are mediated actions the transformation
  function is mediated by the evaluation function

29
Example Pizza Delivery Organization
30
What Can we Possibly Do ?

• Modeling and Simulation of Organizations
• Multiples levels of abstraction
• Focusing on the resources processed and on the
  deliverables created
• Test and Simulate Multiple Configurations
• Simulated re-engineering of organizations
• Formal Model in order to better understand the
  dynamics of an organization
• Build Information Systems
• Better suited to the organizational structure,
  and which better represent the control demands of
  organizations

31
Conclusions

• Semionic Networks
• Are a potentially interesting tool for the
  semiotic modeling of organizations
• There is still a lot to do !
• Better integration of semionic networks to other
  approaches used in the study of organizations and
  workflows
• Workflow Management Coalition Standards
• Enterprise Distributed Object Computing
  OMG-EDOC
• Other models of business processes
• Study case of complex real organizations
• Only demos have been generated until now
• Real study-cases may suggest new features to be
  included on the tool
• Better understanding of the semiotic
  contributions to this kind of modeling