Models that consider human activity are particularly difficult to ‘close’ because of their consideration of ‘interactive’ kinds. Ian Hacking highlights the distinction between the classification of ‘interactive’ and ‘indifferent’ kinds. Different kinds of people are ‘interactive kinds’ because people are aware and can respond to how they are being classified. Hacking contrasts the interactive kinds that are often studied in the social sciences with the indifferent kinds of the natural sciences. Indifferent kinds – such as trees, rocks, or fish – are not aware that they are being classified by an observer. This indifference to classification means their behaviour does not change because of it [but see my point at the end of this post].
The representation of interactive kinds potentially results in a ‘looping effect’ that has implications for model closure and validation – socioecological simulation models have the potential to feedback into, and therefore transform, the systems they represent via the conscious awareness of local stakeholders using the model or its results (or participating in the modelling process). If this transformation occurs it is likely that the model will be a less accurate representation of the empirical system than previously. Such a situation implies that a simulation model of a socioecological system may never truly represent that system (if it used by those it represents). Therefore in the case where a model is to be used by those being represented (for decision-making for example) I’d suggest that an iterative modelling process would be most appropriate to ensure continued utility.
[If anyone has any thoughts on how Hacking’s kinds relate to the whole Schrödinger’s Cat problem I’m all ears – interactive or indifferent?]