# Causality in Complex Systems: the Role of Pattern Recognition

### Federica Raia, City College of New York of CUNY

Processes of self-organization, adaptation, emergence, characteristics of complex systems, are regulated by causal principles and causal couplings that are not describable by a linear chain of causes and effects and not defined in the deterministic framework. For example simultaneity of causal interactions -where causes are at the same time effects- is of fundamental importance to understand negative and positive feedbacks processes and continually changing boundary conditions.

Results from research I am conducting on student understanding of complexity indicate that students utilize simple linear model of causality (LMC) and establish a one-to-one correspondence between cause and effect which impede a conceptual understanding of complex causal relations. A very interesting and important result was found in the relation between this approach -with the explicit use of only mechanistic causality (a force for example) - and the absence of description of patterns in student discourse. In both my research and classroom experience I saw that students, given a pattern and asked how it has emerged rush to identify "the cause" to justify the observed phenomena while never describing the pattern as an essential part and condition to proceed in their explanation of a shape formation (as for example crystals, convection cells etc) or maintenance/ modification. Similar attitudes have been also reported in physics and chemistry students' reasoning (Rozier and Viennot, 1991; Viennot, 1998; Nicoll 2001; Taber, 2001).

Based on my on-going research I am seeing that it is of fundamental importance to help students recognize time and space distribution of variables as causal determinant of system behavior (formal causality â€“Raia 2008). Distribution of variables as a form of causality is unknown to students but, it is amply utilized in science. Unfortunately it seems that it is not made explicit to students For example in the teaching of natural systems, initial and boundary conditions are most often provided to students as a given and the students are very rarely asked to identify and describe them or considering how the same boundary conditions can change and modify systems' behavior. Students are also rarely asked to describe patterns and variables distributions in space and time as important controls on the system behavior. These represent, as discussed in previous study (Raia 2008), types of causality necessary to integrate in the description and analysis of natural complex phenomena. Of particular importance in complexity is the understanding that from variations of distribution of variable amplification (positive feedback) or a dampening (negative feedback) of a phenomena or systems characteristic can emerge.

In my research I observed that the recognition and utilization of specifically formal causality helps students develop and utilize richer and more adequate repertoire of causal models for the analysis of natural complex phenomena. Based on the above, in my classroom I teach students to describe pattern, distribution of data, their change in space time and, build from their observations possible lines of explanations and investigation.