Page 50 ---------------------------------------------------------------------------- ####### ######## ######## ########### ### ### ## ### ## # ### # Interpersonal Computing and ### ### ## ### ## ### Technology: ### ### ## ### ### An Electronic Journal for ### ######## ### ### the 21st Century ### ### ### ### ### ### ### ## ### ISSN: 1064-4326 ### ### ### ## ### January, 1994 ####### ### ######## ### Volume 2, Number 1, pp. 50-63 --------------------------------------------------------------------------- Published by the Center for Teaching and Technology, Academic Computer Center, Georgetown University, Washington, DC 20057 Additional support provided by the Center for Academic Computing, The Pennsylvania State University, University Park, PA 16802 This article is archived as STODOLOSK IPCTV2N1 on LISTSERV@GUVM ---------------------------------------------------------------- TELEMATIC JOURNALS AND ORGANIZATIONAL CONTROL: INTEGRITY, AUTHORITY, AND SELF-REGULATION David S. Stodolsky, Institute of Political Science, University of Copenhagen Abstract presented at the European Group for Organizational Studies 7th Colloquium, Saltsjoebaden, Sweden, June 12-14, 1985. INTRODUCTION In the middle of the 1980's, the tendency of computer- mediated communication (CMC) systems to undermine the power relations, both within and between firms, was seen as a major problem. Threats could come from above (e. g., supervisory authorities), laterally (e. g., competitors), and below (e. g., employees sharing proprietary information across organizational boundaries). The loss of management control (in traditional organizational forms) caused by CMC utilization was seen in primarily negative terms. Traditional management methods, such as centralized control of information and tight supervision, were seen as inadequate in the face of these new threats. CMC gave the employee greater freedom, and the challenge was how to ensure that this freedom was used to the firm's benefit. One option was to integrate a control strategy from an environment that functioned effectively without direct supervision into the firm. Recent developments in cryptography had made possible an idealized form of peer review, that could be used to structure a firm's CMC system. The idea was to reorganize access to information within the organization, thereby restructuring the power relationships within the firm. The purpose of this paper Page 51 is to introduce a new model for computer-mediated conferencing. First, the relationship among the key concepts will be outlined and then the concepts will be defined more precisely. The structure called a telematic journal will then be characterized in greater detail. The advantages of this structure will be specified in terms of benefits to its users and their organization, and of threats to the organization. The discussion is based upon the assumption that decentralized control of information processing resources can be used to better achieve democracy and to give members of an organization opportunities to achieve their full potential. A telematic journal is a type of on-line electronic publication based on a conferencing system (1). It is distinguished by the capability of the person's own computer to serve as a confidential agent. These computers are integrated through a telecommunication network. The term telematic indicates the integration of computer and telecommunication technologies. Quality of messages is enhanced by a process of peer review, that is, users evaluate each others' messages. The review is blind, neither authors nor referees know actual identities. One function of the person's computer is to maintain confidentiality through the construction of unique and unforgeable pseudonyms. The speed and security of information exchange permitted produces a system qualitatively different from the conventional scientific journal. Thus a telematic journal can perform functions within a business not otherwise practical. Relationship among key concepts A non-hierarchical structure based on the peer review process of scientific journals can evolve on a computer-based conferencing system. Because each participant has, in principle, independent information processing resources, such a structure can enhance self-regulation. These resources also permit distribution of tasks and improved security in the exchange of messages, thereby reducing sources of bias and enhancing integrity. Enhanced integrity and self-regulation lead to outputs from the structure to be judged as authoritative. The organization operating this conferencing system thereby enhances its control over both its own members, the participants, and its environment through the availability of reliable and prompt information. Definitions Authority has traditionally been defined as a relationship of domination between two individuals (Adams & Romney, 1959). By domination we mean the probability that a command will be Page 52 obeyed by a given group of persons. When legitimated, domination is called authority. The basis for authority to be discussed here derives from the rational-legal form of legitimation. Authority is typically invested in an individual who occupies a certain role within the organization (Emerson, 1981). In the rational-legal systems of authority, relationships are defined between roles. However, interactions occur between persons. This contradiction is a source of a major threat to organizational integrity and authority, as explained below. Telematics, the science of communication with the support of automated information systems, offers one avenue towards increased self-regulation, thereby meeting challenges to organizational authority. Self-regulation is defined as self- or democratic-management in this paper. This discussion will not deal with sanctions or rewards, control in the traditional sense, but only dominance relations among information sources. The implicit command in this case is "believe this message." If organizational authority can be defined in terms of dominance relationships among sources of information, then a central contradiction in organizational control has been removed. That is, the dissonance generated by the specification of relationships as among organizational functions, when it is actually relationships among people which are to be managed, is eliminated. Reduction in such dissonance permits better agreement between formal and informal organizational structures, thus improving the comprehensibility of the organization to its members. The scientific journal will be taken as a model and elaborated in the organizational context, a context subject to increasing automation, to illustrate how threats to organizational authority can be met by increased self- regulation. A good part of this paper will be concerned with how a telematic journal will operate in the office automation environment. Rules and procedures will be specified in order to give some substance to the concept. This emphasis on rules and procedures is appropriate for organizational systems which derive legitimacy from rational-legal sources. To clarify, we can quote Eric A. Weiss (1985) who was at one time editor of Computing Reviews: I thought the trouble was the mechanics of the system, the work involved, and the negotiations among authors, editors, and printers. (I later came to understand that the problem was more basic; it was the deep academic concern with the effect of publication on tenure, a subject then as now arcane to me as the tribal rites of the Australian bushpeople.) Page 53 Increasing office automation will make possible the replacement of bureaucratic mediation with automatic mediation. Automatic mediation is defined as the transmission of information among persons without use of human intermediaries (Stodolsky, 1981a; 1985). Thus, when judgement is not required, agreed upon editorial rules or algorithms will be executed by machines. These machine executable rules will increase the speed, accuracy, and security in the exchange of information among participants. More importantly they allow creation of a mechanism which has been an ideal in legal theory, the self-enforcing regulation. Its most important functions will be to administrate fairly and impartially, and provide protection to various participants in the organizational system. One extremely important form of protection is anonymity provided by the review process. Recent advances in cryptography make possible enhancements of this type of protection to a standard well above that possible even with the best manual techniques. Alternative models to the scientific journal for publication and distribution are news reporting and reference networks. News reporting is distinguished in that the sources may never become known and the information is typically second hand. In reference networks (Flood, 1965; Flood & Kochen, 1965), information is distributed among people who already know each other and each other's interests. The scientific journal, by contrast, is a mechanism which introduces previously unknown persons to each other's interests through a protective procedure. Anonymity can be extended to both contributors of messages and to their reviewers to varying degrees as is necessary. The total anonymity of the secret ballot will rarely be necessary in the procedures we are considering. However, less protective forms of anonymity (where authorship of unevaluated contributors is not revealed) must be available even though they may not be used in the review system. In conventional balloting, we ensure by anonymity that each input has equal influence. This occurs when all inputs are received before any votes are revealed and by permanent separation of names from messages (In this case the votes are anonymous messages). In the review process, the equal influence (2) is guaranteed by withholding the author's name until after the publication decision. In this case the submitted paper is the message. The objective in both cases is to ensure equal treatment of messages from all potential sources. In voting this is ensured by counting each message as equal and not releasing any information until counting is completed. In scientific publication the anonymity of authors guarantees that submissions will be processed based upon their content, not their sources. Page 54 In the conventional journal review process, the name of the author and the content of the article are simultaneously submitted to the editor of a scientific publication who plays the role of mediator. The editor typically removes the name from the article and directs the content of the message to various referees for review. When the evaluations of the referees are returned, the editor may return the contribution to the author, in the case of unfavorable evaluation by the referees. In the case of a favorable evaluation, the message with the attached name of the author is placed in the public domain. The major objective of this limited anonymity, which is terminated either by the rejection of a contribution or its publication, is the impartial review of the content. In the situations we have considered thus far, an editor plays the mediating role. The integrity of the entire review process and the protections it offers are completely dependent upon that individual. With automatic mediation, however, integrity of the entire structure need not rest upon one person. Automated systems can directly administer rules and protect individual expression in information exchange environments. These include real-time audio teleconferencing (Stodolsky, 1981a) and organizational self-management (Stodolsky, 1985). Results from research on audio teleconferencing illustrate the effect of a protective mechanism (Stodolsky, 1981b) and its favorable influence on the quality of group performance (Stodolsky, 1987). In these experiments the protective mechanism was equivalent to that extended to contributors of scientific publications. That is, names were only revealed if persons were selected to present their message. Structure and development of a telematic journal Let us consider now, how to introduce such a protective mechanism into an asynchronous conferencing system using the scientific journal as a model. For a journal to function effectively it must have an input mechanism which selects only material appropriate for presentation in that journal. Next, it may classify the subject matter so an appropriate group of reviewers may be selected. Next, a subset of the reviewers must evaluate the contribution. Next, the reviewers' evaluations must be aggregated in some manner in reaching a decision about whether to accept the article, and finally the article must be disseminated or published in some way. Each of these functions may be performed by a different class of individuals and may be based upon different subsets of information. For instance, the selection of a contribution as appropriate for the journal may be based upon key words or an abstract alone. The classification function requires a greater degree of familiarity with the subject matter and perhaps a more thorough evaluation of the body of the contribution. Often at Page 55 the completion of this process the reviewers will be identifiable because of the highly specialized nature of the subject matter. If not, then the number of unfinished referees' reports or, if nothing else, random selection could be the basis for assignment. Finally, aggregation of the referees' evaluations would be necessary to reach a publication decision. The automatic mechanism in this case would function very much like a ballot box. The selected referees would be the voters for that article, and a certain fraction would have to indicate in favor of publication before the processing could take place. Thereafter, dissemination of the publication, that would be reassembled with the author's name, is direct and uncomplicated using telematic technology. This simplified view of the scientific publication process is clearly mechanizable in a way which eliminates dependence upon a single individual. More complicated functions associated with the publication process, however, are also amenable to this type of treatment. As yet we have not discussed how this entire structure might develop. In most cases, development of a journal is a problem of selection, not conflict resolution. Editorial personnel, reviewers, and readers must be solicited. In the case of a telematic journal operating within an organization and serving a legitimating function, a mechanism for choosing among individuals competing to carry out the various tasks might be required. Participants can be classified according to their tasks with the journal. The broadest class is the readership, next come the contributors, then even a smaller group of individuals serve as referees. Within the referee function, there might be those qualified to select the material as appropriate to the journal, others qualified in addition to classify the material according to its subject matter, and finally those who could perform all other functions and also evaluate a submitted contribution. This assumes differential competence among the participants, but does not assume a career path in a hierarchical organization. Rather, we have a hierarchical structure in the information exchange mechanism and individuals may have different tasks in different areas of expertise. In the development process, individuals could move or graduate from one class to another, first readers could become contributors, and then a contributor could be selected for the role of reviewer, and move through the classes of review operations, selection, classification, and finally, evaluation. When more reviewers were needed, they could be selected from among the contributors with the greatest number of accepted contributions. The number of persons involved in an evaluation could depend upon the historical information available indicating the confidence level calculable for that judgement based upon rater reliability (Stodolsky, 1984). Enhanced integrity is always desirable, but must be balanced against the expenditure of scarce expertise. Barring other consideration, we Page 56 would want to balance the reviewer hours spent versus savings of reader hours otherwise wasted reading erroneous material. The above discussion is most easily understood in relation to review processes operating at the highest level of formality, equivalent to that of a large, mature scientific journal. A very informal computer-based review system is a read-only conference. This is a type of conference into which only specified persons can write messages. These messages can then be read by anyone who has at least read- only access permission. Any message appearing must be judged to be of general interest by one of the reviewers, those persons permitted to write to that conference. This technique, which uses access controls to manually implement review functions, would be automatically supported through the application of reputation information with a telematic journal. Additional support could be in the form of secure procedures for exchange of messages. Nothing in the system described requires that messages themselves be enciphered. Thus, in order to establish precedence, it might be best that each submitted message be time stamped, signed, and entered into an archive. These entries could be enciphered so that the clear message was not available until an agreed later time. Threats and benefits The new information technology exacerbates threats to the organization from below, from organizations on the same level, and from above. One threat from below is due to the ease of communication using new data networks. Often, for instance, operators of a given type of computer will maintain close links with other operators of the same type of machine. This is important for resolving problems, but may also lead to the unauthorized flow of information across organizational boundaries. This is particularly true in hierarchical organizations, where the social distance between top management and the operator is very great compared to the social distance between operators in various companies. When the operators take full advantage of the communication capabilities of the networks, such as the ability to execute programs on remote systems, the organizational boundaries can effectively disappear. Only strong loyalty to the company can prevent what management often sees as threatening leakage of information through the boundaries of the organization. When the company is large communication among the divisions not explicitly authorized by top management may create corresponding problems of control. Lateral threats to the organization are exacerbated by the ease of information leakage discussed above and by the vastly increased vulnerability to the loss of valuable informational assets when company records are computerized. Perhaps more important is that information officially released by the Page 59 organization be accepted as authoritative. Otherwise, a degree of control over the organization may be lost to competing authority structures on the same level. For a firm these competing organizations include firms in the same industry, suppliers, customers, unions, and professional societies. Finally, the computerization of information handling in the organization exacerbates the threat from above posed by governmental agencies and other supervisory organs. Maximal benefits from computerization flow from a comprehensive integration of information technology into decision making in the organization. However, every detail of internal operations can be exposed if computerized records are examined by a supervisory agency. The effect could be inhibition of any challenge to superordinate authority, even when justified. The failure to protect internal information effectively may also lead to suboptimal use of the technology as users avoid entry of sensitive information into the provided channels. This is just one possible way information technologies protective capabilities may interact with the threat from above. Detailed discussion of the threat from above is beyond the scope of this paper. Major threats to the integrity, authority, and self- regulation of a journal are typically internal. For instance, dependence on one person in operating the editorial function has often led to the loss of a journal's authority (Ingerman, 1975). The problems of executive succession can disrupt an organization, even if they are well planned and executed with adequate notice. When such a change occurs abruptly, the consequences to the organization can be truly disastrous. However, as the authority of a journal increases, the threat of penetration from an outside source increases. By penetration, here, I mean circumventing of the normal review process for placing an article in a publication, most often with collaboration of responsible insiders. This apparently occurred with an article, "Crime Deterrent Transponder System" (Meyer, 1971), in the IEEE Transactions on Aerospace and Electronic Systems. It described a system including unremovable wrist radios for facilitating continuous surveillance of persons by the police. The article was preceded by an editorial introduction which assured the reader that the controversial material which followed had been unanimously approved of by the editors and referees, even though it was certain that some portion of the readership would be offended by "...this digression from pure mathematical analytical paper...." One of those offended was the editor in whose area the paper's subject fell, and who was not even provided with a review opportunity. He commented, "The applications of technology to social problems are legitimate concerns of the technical community and its journals. However, Page 58 permitting the presentation of a political philosophy as part of a technical paper turns the journal into an organ of propaganda (Meyers, 1971)." The author of this controversial paper was later involved in an attempt to suppress the free publication of information on cryptographic systems, and identified as an employee of the United States National Security Agency (Shapely & Kolata, 1977; Bamford, 1983:444). Such external threats may make necessary the use of cryptographic techniques to ensure that submissions are authentic and that reviews and editorial comment are valid (Diffie & Hellman, 1976; Simmons, 1979). Recently developed public-key (3) cryptographic systems make such secure communication techniques routinely available to office automation users. Related techniques, such as secret sharing (Shamir, 1979), can meet threats from above by creating a irreducible division of responsibility that can be used in evaluation processes. In centralized decision making the secret ballot plays this role. The integration of secure pseudonym-based communication (Chaum 1985; Chaum & Evertse, 1987) and secret sharing methods with machine executable rules makes possible organizational mechanisms of unprecedented integrity. The telematic journal then, operating within an organization, as a self-enhancing structure and as an authoritative source of information, presents a challenge to informal networks containing unreliable information. The anonymity extended to participants as a protective measure by the mechanism also presents a powerful challenge to lateral threats from organizational competitors. Two examples can be cited here. One concerns lateral threat created by dual membership in organizations on the same level. When union members meet to organize a response to new technology being introduced in an industry, full sharing of information is necessary for maximum effectiveness. However, the members from one company may not wish to disclose such information to those of another if they feel its release would jeopardize a competitive advantage. In this type of situation, where goals may conflict, the organization commanding the most authoritative information source is more likely to control the outcome. The availability of a protected channel for release of information could play a crucial role. The organization sponsoring the protected channel could claim greater access to information and thereby the most authoritative view. A second case in which lateral threat is manifested is when decision makers at organizational boundaries are negotiating. For instance, when a buyer and seller of a product agree upon a price, conflicting information on the "true" costs of production are often available. The buyer would prefer to pay a Page 60 marginal-return based price, while the seller prefers a price based on full cost recovery (Langholm, 1969). Each organization benefits if its price estimate is accepted as authoritative. Claims about product reliability and quality are even harder to substantiate then those about cost, but can also can play an extremely important role in the purchase agreement. Authoritative information may also prevent personal relationships among negotiators from influencing the outcome. Such "collusion" can be seen as the formation of an incipient informal network or of an informal organization which spans the boundaries of the formal ones, just as a union does in the previous example. Probably the major practical benefit from the telematic journal, in most cases, would be the automatically mediated review capability. This capability could be used to sort electronically posted messages and thereby eliminate the tremendous information overload to which most users of conferencing systems are subject. This type of filtering of messages can be distinguished from other approaches to the problem of information overload in that it is explicitly based upon aggregation of source reputations as specified by a model of social organization. The technique is not only adaptive, but also jointly constructive of individual reputation and organizational authority. It provides quality control and security mechanisms that support the development of reputation within a social system protecting individual and organizational integrity. The combination of quality control and security is likely to be a powerful impetus to the development and spread of these systems in the future. SUMMARY I have explained what a (improved over normal peer review) telematic journal is, and then suggested how it would counter the tendency of CMC to undermine the power relations in the firm. Cryptography is necessary to implement a telematic journal, because we can't use an editor to hide identities. The firm typically can not commit the resources for an independent editor, nor does the traditional peer review system function with adequate speed or security to be relevant in this setting. Significant benefits can accrue to the both traditional journals and firms that adopt the telematic journal as model for structuring CMC systems. Acknowledgments Preparation of this document was supported in part by a contract, "Rule-mediated Telephone Meeting System", from the Folksam Insurance Group and the Swedish Co-operative Institute. The author is grateful to following persons for their assistance in Page 61 preparing this article: Bengt Stymne, Yohanan Stryjan, Marvin Stodolsky, Peter Naur, Kristo Ivanov, Bernward Joerges, Barbara Czarniawska, David Chaum, Hedwig Miller, and anonymous referees. REFERENCES Adams, J.S. & Romney, A.K. (1959). 'Operant conditioning or functional theory of authority'. Psychological Review, 66: 234-251. Reprinted in W. S. Sahakian (ed.), (1972). Social psychology: Experimentation, theory, research. Scranton: Intext Educational. Bamford, J. (1983). The puzzle palace. Harmondsworth, Middlesex, England: Penguin. Chaum, D. (1985). 'Security without identification: Transaction systems to make big brother obsolete'. Communications of the ACM, 28-10: 1030-1044. Chaum, D. & Evertse, J-E. (1987). 'A secure and privacy- protecting protocol for transmitting personal information between organizations'. Proceedings of Crypto '86. New York: Springer-Verlag. (Lecture Notes in Computer Science) Diffie, W. & Hellman, M.E. (1976). 'New direction in cryptography'. IEEE Transactions on Information Theory, T-IT76 (Nov.): 644-654. Emerson, R. M. (1981). 'Social exchange theory' in Social psychology. M. Rosenberg and R. H. Turner (eds.), 30-65. New York: Basic Books. Flood, Merril M. (1965). 'A stochastic adaptive sequential information dissemination system - SAIDS' in Some problems in information science. M. Kochen (ed.), 276-278. New York: Scarecrow. Flood, M. M. & Kochen, M. (1965). 'Some bibliographic and sociological devices to improve maintenance of current awareness in literature' in Some problems in information science. M. Kochen (ed.), 271- 275. New York: Scarecrow. Ingerman, P. Z. (1985). 'Comments on CR'. Computing Reviews, 25-1: 10. Langholm, O. (1969). Full cost and optimal price. A study in the dynamics of multiple production. Bergen: Universitets Forlaget. Meyer, J. A. (1971). 'Crime Deterrent Transponder System'. IEEE Transactions on Aerospace and Electronic Systems, AES-7-1: 2-22. Page 62 Meyers, Stephen, J. (1971). 'Comments on "Crime Deterrent Transponder System"'. IEEE Transactions on Aerospace and Electronic Systems, AES- 7-4: 697. Shapely, D. & Kolata, G. B. (1977). 'Cryptology: Scientists puzzle over threat to open research publication'. Science, 197-4311: 1345-1349. Simmons, G. J. (1979). 'Symmetric and Asymmetric Encryption'. Computing Surveys, 11-4: 305-330. Shamir, A. (1979). 'How to share a secret'. Communications of the ACM, 22-11: 612-613. Stodolsky, D. (1981a). 'Automatic mediation in group problem solving'. Behavior Research Methods and Instrumentation, 13: 235-242. Stodolsky, D. (1981b). 'Protected actions in dialog'. In W. J. Reckmeyer (ed.), Proceedings of the Twenty-fifth Annual North American Meeting of the Society for General Systems Research. Louisville, Kentucky: Society for General Systems Research. Stodolsky, D. S. (1984, December). Self-management of criticism in dialog: Dynamic regulation through automatic mediation. Paper presented at the symposium Communicating and Contracts between people in the Computerized Society, Gothenburg University, Sweden. Stodolsky, D. S. (1985). 'Information systems for self-management'. Human Systems Management, 5: 39-45. Stodolsky, D. S. (1987). 'Dialogue management program for the Apple II computer'. Behavior Research Methods, Instruments, & Computers, 19, 483484. Weiss, Eric A. (1985). 'Sixteen months on the masthead'. Computing Reviews, 24-1: 8. Notes 1) Conferencing systems are computer systems programmed to permit the posting and reading of messages with a remotely located computer or computer terminal. Simple systems are also called electronic bulletin boards. A conference refers, in this paper, to a set of messages grouped under a topic. 2) This model assumes that an author's reputation has no effect on the actions of the editor who knows the author's identity. Page 63 3) A public-key cryptographic system is one which has two separate keys, one for encoding and one for decoding. The public-key can be widely distributed without risk of revealing the private-key which is used to decode messages and sign documents. This system makes key distribution practical when there are large numbers of users. ---------------------------------------------------------------- BIOGRAPHICAL NOTE: David S. Stodolsky Internet:david@adromeda.rutgers.edu Peder Lykkes Vej 8, 4 tv. david@arch.ping.dk DK-2300 Copenhagen S Tel.: + 45 31 59 76 44 Denmark Fax: + 45 35 32 33 99 --------------------------------------------------------------------------- Interpersonal Computing and Technology: An Electronic Journal for the 21st Century Copyright 1994 Georgetown University. Copyright of individual articles in this publication is retained by the individual authors. Copyright of the compilation as a whole is held by Georgetown University. It is asked that any republication of this article state that the article was first published in IPCT-J. Contributions to IPCT-J can be submitted by electronic mail in APA style to: Gerald Phillips, Editor IPCT-J GMP3@PSUVM.PSU.EDU