ABSTRACT
This chapter aims at developing a holistic view about the model–world relationship, by way of looking closely to scientific practice wherein the very common method of testing biological models (i.e., the maximum likelihood estimation [MLE] method that was roughly described in the previous chapter) is examined at length. The holistic view holds that, for a good model, the model–world relationship constitutes a holistic fit, where holistic fit refers to the degree to which the structure of the model (or output, pattern) resembles its counterpart in its target system or, put differently, refers to the distance between two structures (or outputs, patterns). MLE practice is reexamined in depth for the purpose of deriving philosophical implications pertaining to fleshing out the holistic view. To pave the way for reexamining MLE in Section 6.3, Section 6.2 first scrutinizes a simpler estimation method also commonly used in practice, namely, the least squares estimation (LSE) method. After this, Section 6.4 tentatively suggests that, although one of the goals of this book is to develop a holistic view about biological models, the philosophical implications derived from discussing biological models can be generalized to nonbiological models. To this end, a concrete model, that is, the San Francisco Bay model, is discussed. Finally, given that the notion of structure is also employed in the semantic view of models, a deflationary account of model structures is developed in Section 6.5.
