3 From gesture to prototype

Perhaps the most common application of the levels metaphor is to gesture loosely at the relationship between different fields of scientific research, levels of science.[3] In neuroscience, for example, some researchers work at “the molecular level,” doing things such as sequencing channel proteins, studying enzyme kinetics, or manipulating genes. Others work at the cellular level, doing things such as staining cells, recording action potentials, or studying neural migration. Others study brain regions, characterizing their anatomical features or, studying the propagation of neural signals within them. Still others work at the level of systems, using functional magnetic resonance imaging (fMRI), transcranial magnetic stimulation (TMS), and cognitive tasks to find large-scale cognitive systems in the mind-brain. One could perhaps insinuate other levels between these, and one could certainly extend the hierarchy further down or higher up. But the central idea is that the scientific fields can be ordered as higher or lower than one another.

Scientific fields are individuated in part by their theories (Darden 1992). The gestural sense of levels, then, can seem to carry the implication that scientific theories are or will someday be ordered more or less clearly into levels. Oppenheim & Putnam’s (1958) influential view of the unity of science is based on a rough correspondence between levels of science, levels of theory, and levels of parts and wholes (see Table 1). They divide the world into six ontological strata (societies, organisms, cells, molecules, atoms, and elementary particles). These strata are defined by mereological relationships among types: elementary particles are the parts of atoms, atoms are the parts of molecules, molecules are parts of cells, and so on. Each of these strata is assigned a distinct science: economics and the social sciences at the top, particle physics at the bottom. Each science develops its theory more or less autonomously from the others, so the theories developed by these sciences can themselves be ordered, like the layers of a cake, from top to bottom. The unity of science, for Oppenheim and Putnam, is to be achieved by explaining phenomena in the domain of a higher-level science, as described in the theory of that science, in terms of the items in the domain of the more fundamental science, as described in the theories of that science. (Levels of mechanisms, as defined below, involve a kind of part-whole relation as well but without any commitment to the idea that such type-level part-whole relationships correspond in even a rough way to the structure of the sciences or to the structures of their theories.)

Wimsatt’s detailed and influential exploration of the levels metaphor confronts Oppenheim and Putnam with the complexity of the levels found in many areas of contemporary science (Wimsatt 1976). Against Oppenheim and Putnam’s six-layer model, Wimsatt’s “Reductionist Illustrative” (Figure 1) represents multiple branches of levels fanning out from the lowest level in subatomic particles to cosmological objects, the sociocultural world (e.g., economic and political phenomena), and the socioecological world (e.g., evolution).

Figure 1: Wimsatt’s tree of levels branches to preserve compositional relationships among levels (1976).

Wimsatt’s tree diagram, however, represents only one aspect of his prototype account of levels that encompasses many more features than Oppenheim and Putnam’s layer-cake mapping in Table 1. The core features in Wimsatt’s prototype are:

• Size. Higher-level items are larger than lower-level items.

• Composition. Higher-level items are made up of lower-level objects and processes.

• Laws. Laws of nature hold only or mostly between items at the same level.

• Forces. Distinct forces operate at different levels.

• Predictability. Levels are local maxima of regularity and predictability that appear at different size scales.

• Detection. Items at a given level tend to be detected or detectable primarily by other items at that level.

• Causes. Causal relationships hold only or mostly between items at the same level.

• Theories. Scientific theories describe phenomena exclusively or mostly at a single level.

• Techniques. Different techniques and instruments detect items at different levels.

• Disciplines. Different disciplines of science direct their attention at different levels.

Wimsatt’s view is a prototype view in the sense that it characterizes the levels metaphor in terms of a core set of features, not all of which must be present in order for the metaphor to apply. Insofar as Wimsatt embraces a prototype model, he can be seen as embracing descriptive pluralism while, at the same time, holding that there is a sufficiently strong family resemblance among the plurality of applications of the levels metaphor to warrant their inclusion in a single prototype.

Is the levels metaphor sufficiently unified across these different applications to warrant a single prototype? My remarks on the relata, relation, and placement questions should already indicate that it is not—that different features in Wimsatt’s list are at best indirectly related and so fail to map to one another in any tidy way. While the prototype approach usefully highlights the complexity of the levels metaphor, it also obscures the extent to which the different features in the prototype are features of different applications of that metaphor.[4]

Table 1: Oppenheim and Putnam’s layer-cake sketch of the levels of mereology (left), sciences (middle), and theories (right).