2:45 PM - 3:00 PM
[MZZ41-05] Discourse on Idealization in Philosophy of Science and a Case Study from Petrology: Examination of the Partial Melting Model
Keywords:scientific model, idealization, petrology, partial melting model
Idealization is an assumption that introduces false aspects into scientific models. In the context of the philosophy of science, the effectiveness of models that include distortion in discussing real phenomena has been a central topic. According to the conventional view, idealizations have seemed to be arbitrarily removable (Knuuttila and Morgan, 2019). Weisberg (2007) classified idealization into three categories: Galilean idealization is used to make a model more tractable. Minimal idealization extracts the causal factors of interest from a complex phenomenon. Multiple model idealization explains a phenomenon using models that are inconsistent with each other. However, this view of decomposable idealization has recently been criticized. Rice (2018) argued that idealizations cannot be added and removed from a model without changing it, due to their crucial and decomposable nature. Carillo and Knuuttila (2022) extended this discussion and proposed that models should be considered as artefacts that already include idealizations in model construction, rather than representations equipped with changes to real target systems.
While the view that idealization is an additional and desirably decomposable component of scientific models is criticized, some philosophers argue that idealization plays a leading role in promoting modelling. Mäki (2020) analyzed Weber’s idea of the “ideal type” and pointed out that idealization functions as a benchmark for starting modelling and reasoning. According to Mäki, a model equipped with some idealizations should not be applied to the real target system, but rather compared with it. This process would allow us to find a new contribution of unconsidered phenomena by comparison the idealized model and the real target system. He has also pointed out the importance of models that are inconsistent with each other. Morrison (2011) also discussed this point in detail. She has argued that real phenomena composed of several causal factors are often not able to be represented by a single model due to its causal complexity, so we need multiple complementary idealization to apply models to various conditions and cases. As discussed above, the condition to apply a model to the real target system would be constrained by the idealization introduced in the model. As discussed above, idealization may be a crucial factor in selecting the condition of model’s application. Therefore, we need to study more cases of how idealization works in scientific models.
We examine the function of idealization in modelling by analyzing the “partial melting model (Shaw, 1970)” used in petrological research. Despite its traditional origin, this model is widely used due to its simplicity and tractability. The partial melting model is a kind of mass balance models based on partition of trace elements between solid and liquid It is classified into two types based on whether the melt remains in the system. The batch melting model assumes a closed system, while the fractional melting model assumes complete fractionation of the melt. These two models are applied to different natural targets, although both represent in common the phenomenon of “partial melting”. Additionally, these two idealized “benchmarks” are often compared with natural targets to estimate which system is dominant in forming the target rocks: a closed or fractional open system. These cases are consistent with Morrison (2011) and Mäki (2020), indicating that models with contradictory idealizations work as benchmarks to position various cases of a phenomenon among them. In addition, idealized assumptions themselves are the target of estimating a phenomenon in partial melting models, supporting the idea that idealizations play a crucial and undecomposable role in models (Carillo and Knuuttila, 2022).
While the view that idealization is an additional and desirably decomposable component of scientific models is criticized, some philosophers argue that idealization plays a leading role in promoting modelling. Mäki (2020) analyzed Weber’s idea of the “ideal type” and pointed out that idealization functions as a benchmark for starting modelling and reasoning. According to Mäki, a model equipped with some idealizations should not be applied to the real target system, but rather compared with it. This process would allow us to find a new contribution of unconsidered phenomena by comparison the idealized model and the real target system. He has also pointed out the importance of models that are inconsistent with each other. Morrison (2011) also discussed this point in detail. She has argued that real phenomena composed of several causal factors are often not able to be represented by a single model due to its causal complexity, so we need multiple complementary idealization to apply models to various conditions and cases. As discussed above, the condition to apply a model to the real target system would be constrained by the idealization introduced in the model. As discussed above, idealization may be a crucial factor in selecting the condition of model’s application. Therefore, we need to study more cases of how idealization works in scientific models.
We examine the function of idealization in modelling by analyzing the “partial melting model (Shaw, 1970)” used in petrological research. Despite its traditional origin, this model is widely used due to its simplicity and tractability. The partial melting model is a kind of mass balance models based on partition of trace elements between solid and liquid It is classified into two types based on whether the melt remains in the system. The batch melting model assumes a closed system, while the fractional melting model assumes complete fractionation of the melt. These two models are applied to different natural targets, although both represent in common the phenomenon of “partial melting”. Additionally, these two idealized “benchmarks” are often compared with natural targets to estimate which system is dominant in forming the target rocks: a closed or fractional open system. These cases are consistent with Morrison (2011) and Mäki (2020), indicating that models with contradictory idealizations work as benchmarks to position various cases of a phenomenon among them. In addition, idealized assumptions themselves are the target of estimating a phenomenon in partial melting models, supporting the idea that idealizations play a crucial and undecomposable role in models (Carillo and Knuuttila, 2022).