The binding problem requires a solution at the level of individual neurons, but no definite mechanism has yet be given. Therefore, the neuronal level is as yet inadequate for modeling cognitive processes in which binding plays a crucial role. Moreover, the neuronal level involves too many details that are unlikely to be essential for understanding cognition. A general model of cognitive brain functioning is described in which cognitive tasks are represented in a network of cell assemblies. In the network, binding is functionally defined in a way that is compatible with the neuronal level. A computer simulation of the model clarifies how the binding of location and identity of a set of simultaneously presented letters takes place and how questions about the location and identity of the letters are answered. From the simulation of the task three predictions on the logistics of neural processes are derived: 1. When the cell assembly representing a letter participates in more than one temporary excitation loop, it will reach its critical threshold faster. At the behavioral level this means that as the number of identical letters in the display increases, responses will be faster. 2. In order to answer questions about the location and identity of presented letters cell assemblies representing the target location and the target identity have to become bound to their appropriate values. As a consequence the facilitatory effect of identical letters will be stronger if they involve the target location or the target identity than when identical non-targets are involved. 3. Negative identifications are more dependent on the presentation time of the letters than positive identifications because the excitation loops involved take more time to reach the critical threshold. Therefore, the facilitatory effect of identical letters is stronger when the external activation is relatively strong, i.e., when presentation time of the letters is sufficiently long. The reaction times obtained in three behavioral experiments support these hypotheses. Effects of binding can therefore be predicted on the basis of the general logistics of neural processes, without assumptions about a specific binding mechanism at the neuronal level.