DEVELOPMENTAL THEORIES APPLIED TO CROSS‐CULTURAL COGNITIVE RESEARCH *

Almost from the outset, psychologists engaged in cross-cultural research seemed to realize that their work posed methodological problems different from, and probably in addition to, those that faced their colleagues in other branches of their science. It has been generally understood that it is one thing to observe a difference in behavior across cultural groups and quite another t o interpret it. This realization is reflected in the continuous concern of cross-cultural psychologists with problems of methodology, dating from Rivers’ and Titchener2 t o contemporary investigators such as C a m ~ b e l l , ~ Berry,’ Goodnow,’ Glick,’ ’ and others. We, too, have been concerned with questions of method and the special difficulties of inference from observation t o psychological process that are endemic t o the cross-cultural enterprise. Some of our work has been concerned with the problems of specifying the culturally determined independent variables that relate t o the dependent variables we study (Cole, Gay, Glick and Sharp5). Following the lead of Campbell and many others, we have sought t o use the opportunities offered by different cultural settings t o deconfound theoretically promising causal factors that are ordinarily “packaged” in modern, technological societies (Whiting2 6 ) . This work has engaged us in a companion issue that has been of great concern to us: what significance can we attach t o our dependent variables? Here we enter the perennial debate between anthropologists and psychologists as t o the proper methods for studying cognitive behavior, a debate that has centered around deciding what inferences about psychological processes of individuals are warranted on the basis of experimental and naturalistic observations (c.f. Cole and Scribner7 and Scribner. ’ ) Like most anthropologists, we are committed t o the view that observations of intelligent behavior in everyday life are an important source of information about culture and cognitive processes. But we also believe experiments t o be important and probably necessary tools for disentangling the complex relationships among culturally determined experiences and specific intellectual skills. To use Scribner’s2 ’ term, this position requires us to “situate” the psychological experiment as one of many contexts in which to sample behavior. This approach to “behavior-in-context’’ leads us t o question the generality of inferences from experiments that are not corroborated by nonexperimental data. At the same