Social Knowledge and Social Action: Heterogeneity in Practice

Th is paper is about some of the social aspects of knowledge and action relevant to th ink ing in A I , and in par t icu lar the basic experience of mu l t i p l e perspectives and integrat ing different kinds of local knowledge. It discusses ways of re th ink ing a number of fami l ia r concepts inc lud ing facts, in teract ion, knowledge, and organ iza t ion , rais ing questions about how well we can current ly capture their social dimensions conceptual ly, representationally, and computat ional ly . It suggests several approaches to developing more complete computat iona l models of these phenomena. 1 I n t r o d u c t i o n Th is paper is about some of the social aspects of knowledge and act ion relevant to th ink ing in A I , and I want to expla in a b i t about how and why I these ideas came to interest me so deeply. The crux of the mat ter is that in my own inte l lectual l i fe, I have been continuously involved in more than one " f ie ld" at once. Two of the p r imary arenas in which I've found myself have been comput ing—speci f ica l ly exper imental computer science and A I — a n d sociology-specif ical ly symbolic interact ionism and studies of the integrat ion of technology and work. The or ig ina l idea I had was that these seemingly qui te different arenas actual ly had a lot in common, and it wou ld be interest ing to t ry to art iculate what I found fascinat ing in each to people who were more focused on the o the r—to be a boundary-spanner. For example, when I f irst seriously encountered comput ing in the m i d 1970s, the problemat ic of large-scale software development was central , and it certainly seemed that large-scale software 1) was developed by large groups of people, (see, e.g., [Scacchi 84]) and 2) might have many of the s t ruc tu r ing characteristics of large-scale organizat ions of other types, when software was seen as collect ions of processes (cf., [Durfee, et ai 87; Fox 79; Fox 81 ; Gasser, et ai 87]). S imi lar ly , some AI researchers were beginning to deal w i t h questions of scale and concurrency. They turned their a t tent ion to mul t iagent and d is t r ibu ted systems, ostensibly to cope w i t h increasing prob lem complex i ty and to reflect what they saw as basic characteristics of the " rea l -wor ld" [Bond and Gasser 88]. These movements w i t h i n AI seemed to me to offer a very happy marr iage w i th what I was reading about certain branches of sociology—most especially symbol ic interact ionism (the very name itself seemed jus t r ight) and later w i t h "science studies" the sociology of science and scientific knowledge. In the midst of this ferment, 1 quickly found that whi le the discourses in c o m p u t i n g / A I and in sociology had s t imu la t ing and encouraging s imi lar i t ies, there were certa in under ly ing assumptions that made them, in practice, v i r tua l l y incommensurate. A n d , in a very real way, my personal experiences in the several worlds as a boundary-spanner mir rored the incompat ib i l i t ies in the world-views. Th is experience of l i v ing in several worlds at once, then, is both the p r imary mot i va t ion for and the main point of the present paper: to help explain what i t means for several world-views not to j i ve , and to provide some new ways of seeing ( look ing into) some basic issues of AI and comput ing through glasses of in commensurate perspective—-to get at the pragmatics of deeply-experienced heterogeneity. The basic experience of mu l t ip le perspectives, what might be called a fundamental po in t , is s imply these two observations: 1) experiences vary over place and t ime, and 2) experiences in different places and t imes interact, necessitating some sort of in tegrat ion. Elsewhere, I have discussed several not ions tha t derive f rom these observations [Gasser 86; Gasser 9 1 ; Gasser 92]. They include the existence of a fundamenta l tension between local and global knowledge and act ion; certain results concerning the impossib i l i ty of common semantics or shared representations; the nature of and strength behind commi tment as a b ind ing force in interagent relations, and the causal direct ion of goals, social laws, rules, programs, etc. in shaping behavior. Taken together, it seems to me, the impacts of a perspective shift tha t first takes heterogeneity, not commonal i ty , as given, and that second, observes the j o i n t and interlocked nature of so much human and human-computer interact ion , are great. How can this pat terned, ongoing, j o i n t , interlocked behavior be explained and designed, w i thou t recourse to commonal i ty? The outcomes of longterm inquir ies in to these issues seem interesting to me, because they help to address questions such as what are the differences between ma-