On Klatzky and Creswell (2014): saving social priming effects but losing science as we know it?

Klatzky and Creswell (2014) offer an interpretation of the unreliability of social priming effects by analogizing them to what is known about the complexity of cross-modal transfer effects in perception. The complexity of these transfer effects arises because they are both multiply determined and stochastic. In this commentary, I argue that Klatzky and Creswell's thoughtful contribution raises the possibility that there might be deep and substantive limits to both the replicability and the generalizability of many of the phenomena that most interest psychologists, including social priming effects. Psychological phenomena largely governed by what Fodor (1983) called the "central system" may resist both replication and generalization by their very nature and not because of weak and underpowered experimental methods. With such phenomena, science might give us very good tools for explanation, but not for prediction (replication).