136 PAUL WEISS 



the era ahead is the will to drive the analysis of developmental phe- 

 nomena far beyond the mark where we used to stop to coin a term, 

 and the determination to become as exact, specific, objective, and realis- 

 tic in the formulation of our problems and in the description of our 

 observations and results as has become the rule in the physical sciences. 

 This is what we mean by reorientation. It is just as important as is 

 the gathering of more data. With this in mind, we turn to the question 

 which forms the core of my assignment: What is differentiation? 



What Is Differentiation? 



By the standards just advocated, this question can immediately be 

 recognized as far too indefinite to be answered by anything but another 

 generality. For "differentiation" refers not to a single circumscribed 

 natural phenomenon, but is a summary name for a heterogeneous col- 

 lection of notions about a variety of events associated with develop- 

 ment. Consequently, the various statements and generalizations which 

 have been made in the past concerning "differentiation" and "dedif- 

 ferentiation" are equally heterogeneous and full of contradictions. 

 Only a few authors have deemed it necessary to be explicit about terms 

 of such general currency. In order not to add to the confusion, I shall 

 begin by developing my own definition. Then whatever I shall have to 

 say about "differentiation" will refer to the phenomenon as here defined, 

 and not to anybody else's "differentiation." 



A definition of "differentiation" is partly a matter of objective de- 

 scription, partly of plain logic. We shall deal with the logical portion 

 first, because it has been more persistently ignored. 



Characters vs. Processes 



One source of confusion is the tradition of confounding the process 

 of differentiation with its criteria. We usually tell differentiation by its 

 more conspicuous products; e.g. melanin granules in a pigment cell, 

 myofibrils in a muscle fiber, secretion granules in a gland cell. Our 

 criteria are predominantly morphological and we judge mostly by ap- 

 pearances. In the examples just mentioned, our judgment happens to 

 be well founded. It becomes dubious when based merely on superficial 

 differences of shape, arrangement, or size, and it misleads us completely 

 when we try to reverse the argument and take the absence of obvious 

 differences for evidence of lack of differentiation. To appraise a cell 

 correctly we must take into account all of its properties, not just the 



