Three problems in the marriage of neuroscience and education.

Whether neuroscience can be informative to educational theory and practice is not debatabledit has been. For example, behavioral data were not decisive in determining whether dyslexia was primarily a visual perceptual disorder, or whether phonology was the more fundamental problem (for a review, see McCardle et al., 2001). Brain imaging data (e.g., Rumsey et al., 1992) showed reduced activation in left temporoparietal cortex, a region known from other studies to support phonology, thus strongly supporting the phonological theory. The question is just how enthused and optimistic should researchers, educators, and parents be? It is wise to remember that there have been at least two other brief pulses of excitement over a possible marriage between these two fieldsdone in the mid 1960s (e.g., Gaddes, 1968) and the other in the early 1990s (e.g., Vellutino et al., 1991). Will the recent reigniting of the romance lead to a long-term commitment, or will it be another passing infatuation? Here I highlight three problems in bringing neuroscientific data and theory to bear on educational practice. Before I describe those three problems it will be useful to draw a distinction between natural and artificial sciences (Simon, 1996). Natural sciences, like neuroscience, are descriptive; the aim is to discover principles that describe neural structure and function and in so doing to bring order and comprehensibility to data. Artificial sciences are normative. Their aim is not the description of the natural world as it exists, but the creation of an artifact, designed to serve a specified goal, within a particular environment. Examples of artificial sciences include urban planning, economics, engineering, and education. The artifact to be created in education is a set of pedagogic strategies and materials. The goal is usually some version of ‘‘instilling skills and knowledge in children,’’ although the goal changes over time and across cultures.