One can not study science and not be aware of the existence of underlying themes that weave their way throughout the whole fabric of science. The strong hold that certain themes have on scientists' minds helps to explain the stubborn faith with which some scientists cling to a theory in the face of supposedly contradicting evidence. Historically, these themes have neither evolved directly from observation nor could they be incorporated as part of the logical, mathematical structure of science. One example from physics is the theme of conservation laws, such as the conservation of momentum. The conservation theme has remained a guide, even when its language has been forced to change.
Conservation laws are not the only themes. One can argue that the 2500-year course of science has been guided by such underlying themes as transformation, inevitability, reductionism, and unity. From at least Thales on, the belief that nothing springs from nothing, but all exists as transformations of some fundamental entity in a multiplicity of ways, has been reaffirmed again and again. Equal to the theme of transformation is the belief in the inevitability of scientific thinking to comprehend the physical world. The theme of inevitability operates on all levels in science from the formation of scientific laws to the course of science itself. The reductionism theme is that all of the natural world can be reduced to the nature of smaller and smaller units. For example, we can understand stars in terms of the atoms that compose them, and the atoms in terms of nuclei and electrons. The unity theme is the belief that a few underlying principles manipulate the elements of existence to produce the mulitplicity of form and behavior observed in the natural world. The belief that a single theory will eventually be found that encompasses all aspects of human physical existence has been a most persuasive theme in the growth and diversification of science.
In order to understand how science grows, we need a mental picture of the relational structure within the scientific enterprise. Various relational models have been asserted--pyramid models, circular models--but the one we think that most accurately represents science is to liken science to a tapestry in which the threads are the various fields of astronomy, physics, chemistry, and mathematics. These threads do not interact with each other at just one point, but they are interwoven so that they interact at numerous points. Trying to be selective in what one accepts as valid in science is to misunderstand the interlocking structure of science. Thus any attempt to remove one thread would unravel, like an old sweater, the whole fabric of science.