New Developments in Restructuration Theory and Understanding Complex Systems Through Agent-Based Restructurations

• In Event: Restructuring Concepts and Tools Through a Complexity Perspective

Abstract

Wilensky and Papert (2010) developed a theory of restructuration that examines the relationship between knowledge of a concept and the representations used to learn, express and use the concept. One example would be that knowledge of arithmetic can be expressed and learned through Roman numeral representations (the Roman structuration of arithmetic) or Hindu-Arabic numeral representations (the Hindu-Arabic structuration of arithmetic). Each representation has different affordances for learning and use. The Hindu-Arabic restructuration has greater properties for power and learnability than the Roman structuration. In Europe, around the turn of the first millennium, the Hindu-Arabic restructuration of arithmetic greatly lowered the threshold for people to be able to do arithmetic – using Roman, very few people could do multiplication or division, once it became infrastructural, the vast majority of humans could do so. Another example taken from diSessa’s account of Gallileo’s dialogues (diSessa,2000), shows the increased power and learnability of the algebra structuration of kinematics versus the natural language structuration used by Galileo. Restructuration theory builds on a long history of psychological and historical research that has argued that representational forms shape human knowledge and understanding, both at the individual and societal level (e.g.,Goody,1977;Olson,1994;diSessa,2001).
The advent of powerful computers has enabled new computational restructurations of many fields of knowledge. Increasingly scientists are making use of computational restructurations to represent knowledge across many disciplines, ranging from physics, chemistry, biology to materials science, psychology, sociology, linguistics, etc. (Wilensky&Rand,2015). Agent-based modeling is a form of computational representations and practices that have been increasingly employed by scientists, and, more importantly for the AERA context, has powerful affordances for learning. Just as multiplication and division were difficult for Europeans before Hindu-Arabic, there are concepts and practices that today are difficult for us. Research has shown the many difficulties people have with understanding complex systems (Chi,2005;Jacobson&Wilensky,2006). A key concept in understanding complex systems is emergence, how many distributed individual elements can interact and come together in aggregate patterns. Science is replete with emergent phenomena, from the swarming behavior of birds and insects, to the interactions of atoms and molecules to form different states of matter, and the interactions of people to form human institutions. Again, emergence has been shown to be difficult for learners to grasp (Chi,2015;Wilensky&Resnick,1999).
In the past two decades, researchers have developed agent-based modeling languages designed to enable restructurations of complex and emergent phenomena (Resnick&Wilensky,1993;Wilensky,2001;Wilensky&Resnick,1999). These languages have been used to created restructurated curricula that use agent-based representations. NetLogo (Wilensky,1999) has been used extensively to enable learners at many different levels to understand complex phenomena, including in predator-prey biology (Wilensky&Reisman,2006); behavior of gases (Wilensky,1999b), chemical reactions (Levy&Wilensky,2009), crystallization of materials (Blikstein&Wilensky,2009), flow of electrical current (Sengupta&Wilensky,2009); evolution by natural selection (Horn.et.al.,2014), economics of buyers and sellers and wealth distribution (Maroulis&Wilensky,2014; Guo&Wilensky,2018), development of cities (Hjorth&Wilensky,2018) among many others. This paper describes and evaluates developments of restructuration theory since 2010 and illustrates with examples used widely in schools.

Author

• Uri J. Wilensky, Northwestern University