Problem Statement – Background

Posted on February 2, 2010


(still working on this)
There is, at the present time, a good deal of debate about the meaning of the word design, but for the purpose of this dissertation, i will rely on the very general definition provided by the great polymath Herb Simon, who defines it as
..a person or people trying to devise a course of action aimed at changing an existing situation into a preferred one [Simon, 1996: 111]
Sociotechnical System
The term Socio-Technical Systems has been used since the 1950’s to refer to the object of study of various research agendas also considered to be part of organizational behavior [Trist&Bamforth, 1951], Human-Computer Interaction (HCI) [Hutchins, 1995] Computer Supported Collaborative Work (CSCW) [Sutcliffe, 2005] and Social Informatics [Davenport, 2008].  In a 2008 NSF workshop, the term “Science of Socio-Technical Systems” has also been used.  The term Social Shaping of Technology [Williams, 1996] has also been used within studies of sociology, which shares some similarities with socio-technical systems (STS).
General Definition of a Sociotechnical System
The original novel contribution of the early exploration of STS – stemming from the early work of Trist and Bamforth in 1951 – was a reaction to previously deterministic, hierarchical and non-democratic theories of organizational design – in which “technology” included both the organizational structures as well as the tools used within those structures.  Different from previous approaches, STS focused on adaptive, decentralized methods for understanding and designing systems so that
“..each member of the work group should have an optimal level of variety; learning opportunities; scope for making decisions; organizational support such as training and good supervision; a job recognized as important by the outside world; and the potential for making progress in the future.” (Mumford, 2006, quoting Fred Emery)
As the scope of the STS research agenda expands beyond the design of sociotechnical organizational structures, the definition of the term may require additional consideration.  For the purpose of this dissertation, I will define a sociotechnical system as:
A set of human and technical components together with the relations connecting them to form a whole unity. (adapted from [Krippendorff, 2001])
While each of these words – “human”, “technical”, “relations”, “whole” and “unity” – in this definition has been the subject of debate in other non-STS areas  – and we will not be able to here solve those debates – I will here offer a slightly more robust series of propositions in order to better define sociotechnical systems for purposes of this paper.
A socio-technical system:
  1. Is Social.  The human portion of its behavior cannot be understood or described at the level of the individual human actor.
  2. Is Technical. The system must have at least one man-made (technical) component.
  3. Is sociotechnical. The system’s human (more than one) and technical (one or more) components are actors which each affect the other.
  4. Is Systemic.  It exhibits the basic characteristics of systems (partially derived from [Ropohl, 1999])
    a. Its meaning is greater than the sum of its elements
    b. Its function is at least in part determined by its structure
    c. It cannot be described on just one level of hierarchy
  5. Is Dynamic. The meaning of a system only begins to emerge once there is an interaction between at least two of its components.
This series of propositions has some important implications for our further consideration of sociotechnical systems.  The first is that the meaning of a sociotechnical system necessarily emerges out of the interactions of its elements.  In other words, it is impossible to have a priori knowledge of the meaning of a sociotechnical system.  Another implication is that the meaning of a system will continue to emerge over time.  As the elements continue to interact, meaning continues to emerge.   A final implication is that there is no socio-technical system in which only the human or only the technical components is completely responsible for its behavior.
The word evolution is used quite a bit to refer to change to a given organism or system over time, as is the case when we say that “this organization has evolved into a market powerhouse.”  While this is not an incorrect use of the word, it does not get at the more technical definition of evolution that will be the subject of this dissertation, which has to do with the change in a class of things over time as one generation of things passes some sort of information to successive generations of the thing.  This information is subject to change, though, as a result of the processes of
  1. selection
  2. variation (which can happen both genetically and epigenetically)
The Problem
Buried in all of this technical terminology is a big, but perhaps very subtle problem: if i am a designer who wants to design a system which will involve both people and technology, and that will benefit from its ability to evolve beyond my initial control, then i understand at the start that i can neither fully predict nor fully control the meaning that the system creates in the world.  However, i must still design the system in a way that allows it to move itself and its situation toward some preferred state:
  1. Where do i start?
  2. How much of the system can i control?
  3. How much of the system should i control?
  4. How much variation do i want in the successive generations of the system?
  5. Is it possible to control the amount of variation in successive generations?
  6. What are the situations in which this type of design work best? How do i know?
  7. How do i communicate my design with other stakeholders in a way that allows us all to understand the system?



This problem exists all around us.  In business, the concept of the franchise, the ponzi scheme, the chaordic organization [Hock, 2000] are all examples of a design of sociotechnical systems that are meant to evolve, to some extent, beyond the full control of the original designer.  In software, open source software is an example of a sociotechnical system that is meant to evolve, to some extent, beyond the full control of the original designer.  More recently, sociotechnical systems like Twitter, Facebook, Salesforce, etc. have used API’s as a mechanism for enabling evolution beyond the full control of the original designer.  The designers of these systems, however, have a huge task before them.

If they get the initial design right, they enable evolution of their product in one or more successive generations of “offspring” in the form of applications, which, though different from their parent, still maintain some of the characteristics of their shared parent.  Twitter, for example, spawned a second generation of applications which, based on the initial design decisions of the Twitter organization, had a certain amount of variation in the population.

Potential Starting Point

One potential starting point for solving this problem is to develop a language with which to talk about systems with these characteristics.  In previous personal and professional projects, i have used various visual adaptations of other theories as a way to think through and to (try to) communicate with other stakeholders about such projects.  The most problematic part of these projects was the lack of a shared language or shared conception of the salient elements of the project.  Here are a few examples of my initial attempts:

Single Level of Sociotechnical System (leaning heavily on Krippendorff)

Two Levels of a Sociotechnical System (again leaning on Krippendorff)

Levels of a Sociotechnical system (incorporating Bourdieu's notion of a habitus)

Levels of a Sociotechnical System (Bordieu, Giddens, Bryant)

Levels of a Sociotechnical System (Briggs)