| In scientific research, data is often considered as a qualitative or quantitative description of the facts which is the basis or evidence derived to scientific theories, The concept of data is usually associated with the concept of information. For most of people, the two words can replace each other, but it is not. Philosophy of information puts information to be the same with matter and energy, as the one of the most basic elements in the universe as, and the data is generally given to an established empirical understanding. Of course, the above is only as one attitude for data ontological. Thus, further philosophical research on data would be particularly necessary. The introduction part summarizes the three major ontological approaches of data. Based on the work of ontology, we can explore the significance of data from different perspective in scientific epistemology.After the logical empiricism, they left two major problems between while is the dichotomy principle of observations and theoretical, and the bridge principle of theory and experience. P. Suppes, as the leader of he semantic approach on theory put forward that the process from phenomena to the data model, and then the theory is a model system. The first chapter describes the role of data models in the form of semantic position, to highlight its potential empirical characteristics. At the same time, the formation of the data model still can not get rid of the dependence on empirical observation. So this approach lacks of sufficient ontological interpretation of the phenomenon. Since the last century, there were lost of scholars in the University of Pittsburgh’s began the philosophy research on data, which mainly covers the scientific epistemology, data mining and causality and knowledge representation. The second chapter mainly covers the philosophy of data from J. Bogen and J. Woodward, who were the representative on data epistemological problems. They present data-phenomena-theories(D-P-T) model to completely replace dichotomy distinction between observation and theory, and thought data is the basis for the production of scientific knowledge. Surprised, this thought actually is the core concept of big data science currently. However, D-P-T model still has many problems, such as the data being the pure objective verification of delusions, in fact, there must be the theoretical background when the data come out. Thereby we can see that D-P-T model is incomplete in the epistemology position of data.In order to adapt to the new scientific model of big data scientific paradigm, the third and fourth chapters present integration in epistemology of empiricism and fundamentalism of data, and conclude that the production of scientific theories is a dynamic construction between data, phenomenon and theory. On the one hand, construction of scientific theories is not merely a linear program from phenomenon to theory or from data to theory, is actually a concurrent and dynamic process. On the other hand, the phenomenon of ontological commitment is critical, which is used as the basis of scientific realism. The construction of the theory would take into account the phenomenon of "salvation", so the structure of theory can be mapped to the structure of the phenomenon, which corresponding to structure of real world. Therefore, the interaction between data, phenomena and theories drawn out the most robust theory that needs to "save" the ontology of phenomenon in epistemological level. D-P-T dynamic construction model avoids empiricism trap in semantic approach, meanwhile it modify the weakness of linear models of D-P-T triple distinction, so that this model adapted to the changing technological environment of big data and new scientific paradigm shift. |
