400 Edited by Henry Quastler 



of information theory was demonstrated by the apparent harmony that prevailed 

 at this three-day meeting despite the varied professional disciplines represented 

 by the papers and the participants. But the devil's advocate will argue that 

 of potential meeting grounds there are many, and the fact that information 

 theory actually has provided one does not necessarily imply that it was an 

 essential ingredient of success. We must not follow the lure of each and every 

 interdisciplinary beacon; most actual results are still obtained within the safe 

 limits of the estabhshed discipHnes. 



More important is the problem of system sciences in general — that is, 

 of all sciences that deal with the whole rather than with parts, with general 

 principles rather than detailed specifications, with patterns rather than specific 

 mechanisms. This is a many-faceted problem. There are, first of all, the situations 

 of the 'forest-and-the-trees' type; for instance, in physics: a complete description 

 of every particle in a gas would contain implicitly all thermodynamic para- 

 meters — but in this form the information would be useless. Or, to use an 

 example in biology: if we knew the chemical constitution of all substances 

 in all cells, together with all details of distribution, chemical kinetics, in brief, 

 if we had reached the biochemical millenium — then we still would not necessarily 

 know which of all these details are significant on the next higher level of organi- 

 zation, although presumably this information must be impHcitly contained 

 in the known details. Are we simply up against a psychological limitation? 

 We seem able to think only of so much detail within any single train of thought. 

 Faced with amounts of detail considerably beyond our mental capacity we 

 begin to select: in the course of such selection important features are eliminated 

 almost as readily as unimportant ones. Knowledge is not usable for human 

 minds unless it is organized in blocks with not too much detail in each. 



There is more behind the desire to look at the whole rather than the parts 

 than just an awareness of psychological limitations. There are relations within 

 a whole which cannot be expressed in terms of parts alone. This is the fact 

 expressed in the proposition 'the whole is equal to more than the sum of its 

 parts'. The very mixed reactions which this proposition elicits are probably 

 due to a failure to state exactly in what way the whole is more than the sum of 

 its parts. There should be little disagreement if the 'more' refers to propositions 

 concerning the whole which are qualitatively different from any proposition 

 which can be made about any of its parts. Information theory provides a 

 very convenient formahsm to state this situation : the total information content 

 of the whole is exactly equal to the sum of the information contents of the 

 parts — where the description of each part includes all possibiHties of connection 

 with other parts; the mutual dependency of parts being organized into a whole 

 causes a mutual reduction of uncertainty; therefore, the amount of non- 

 redundant information associated with the whole is less than the sum of the 

 information contents of the parts; the difference is exactly the infonnation 

 content of the constraints, or all those propositions which apply only to the 

 whole and not to its parts in isolation. This formulation may be of some help 

 in making clear a puzzling aspect of the whole-parts relation. 



The preference for dealing directly with wholes rather than with parts 

 is greatly supported by contemplation of the way living things are organized. 

 The most striking feature is the existence of the organizational pattern with 



