Computers and Social Change
[ Main Index | Chapter Contents ]

Chapter 5. Social Factors: The Computer Impact on Human/human Interactions

Does your family have a computer? Were they used in your schools? Do you play with them? If so, what have your learned from using computers? How have they changed your social relationships? The study of socail factors in computing looks beyond the immediate connection between person and machine to the consequences for the social institutions in which the technology is used. The effect of the computer in these institutions depends on more than the interface between individual and machine. It depends on the way computers are linked together into communications networks, and how computer-based communication alters social interaction among people.

The family and the school are the two most important socializing institutions. Within them, children are introduced to the intricacies of their culture. For older children and for adults, social networks and economic institutions are important agents of socialization, providing opportunities to learn new social roles and behaviors. As the commercial production of information expands, some of the socializing functions of family and school have been taken over by leisure activities and mass entertainment. In place of public interactions among communities of people, we often find private interactions between people and machines. But in the social interactions among people using computer networks, we find new evidence of social integration and community formation.


For some philosophers, the dilemma of modern humanity has been the development of instrumental rationality in science, technology, and business at the expense of the human need for meaning, continuity and belonging derived from direct, face-to-face interaction with one's particular group. The advantage of the rationalization of culture has been the development of universalistic social relationships. We are able to form larger communities and treat strangers with less suspicion. The disadvantages have been individual isolation, reification of social relationships, and a weakening of shared norms and values. Computer-based communication shows evidence of being a continution of the trend towards universalism in social interaction. It also offers some new possibilities for enhancing social integration in organizations and communities.

5.1.1 Loss of Face to Face Contact

Loss of face-to-face communication is reported in many studies of computerized homes and offices. About 20% of the secretaries surveyed by the Diebold Group (1984) reported less face-to-face contact with their peers. 22% of the managers and professionals also reported a decrease in face-to-face contact. A study of women working with computers in their homes found social contact between them and their friends to be reduced, although interaction with their children and spouses increased (McClintock, 1984). Even when face-to-face contact is replaced by more contact via telephone and electronic message, the quality of interaction changes. In other words, people report that talking through a computer network is "just not the same" as face-to-face.

Consider, for example, the difference between a proposal of marriage by computer mail and a proposal in person. Even when two users of an electronic bulletin board fell in love via network, the marriage proposal was made face-to-face ("Love, Computer Style", 1983). The emotional overtones of social interaction are expressed through gestures, tone of voice, and even through smell (with chemicals called phermones). Just as the telephone removed vision and physical presence from human conversations (giving us in return the ability to communicate over vast distances), so most computer message systems remove the perception of sound and sight from interaction.

5.1.2 The Absence of Vision
In a review of the effects of the withdrawal of visual contact on conversational outcomes, Reid (1977) concluded that there was no measurable impact on people's abilities to exchange factual information and cooperate in problem-solving tasks. These results suggest that we may expect interactions via computer network to be as effective for communicating technical information as direct contact. Although experiments found people able to form accurate impressions of others' traits, they tended to rate others less favorably when they couldn't see them. In the absence of vision, they were also less confident in their judgements of personal characteristics. This means that, even in the absence of visual clues, we get accurate information about people from what they say, but are less able to make positive, emotionally based evaluations of them. This can mean the loss of emotional solidarity in computer network communications, since feelings contribute to the strength of group ties.

The absence of vision is an advantage for people who "look" low status or unattractive. It is a well known social psychology research finding that attractive people are preceived to be more competent, while "ugly" people tend to be devalued. The telephone removed visual clues to a person's age, sex, race, ethnicity, and social status. Computer networks remove many of the verbal clues (for example accent and pitch of voice) to status as well. If everyone has equal access to computer networks, we may find ourselves paying more attention to what they are saying than to how important they look and whether they appeal to us emotionally.

5.1.3 Multimedia Interfaces and Emotional Communication.
Multimedia interfaces that let you hear the person you are exchanging information with and let you see them on television offer us the possibility of adding emotional communication to computer netwoks. This could restore some of the social control that now appears to be missing from computer conversations. But this technological innovation will not necessarily strengthen our perception that we are speaking to another human being. What is most important in face-to-face contact is our ability to imagine ourselves in the presence of another self. Because this is an emotional response to another person, the addition of non-verbal elements to computer messages can make it easier to preceive another's humanity. However, if our emotional response to someone's accent, color, sex or status makes it more difficult to imagine them as like ourselves, the quality of our interaction will not be improved by multimedia interfaces. Multimedia interfaces will certainly increase the distance over which we can present ourselves to others; but they will not automatically supply us with the cultural value that we should communicate with them.

5.2 Computer-Aided Conversation and the Problem of Trust

In a collection of essays on communication and social evolution, Jurgen Habermas (1979) points out that communication involves claims on the part of participants about: In conversation we evaluate the truth of what is being said, the motives of the speaker, and the social appropriateness of the communication. We can predict that the use of computer-mediated conversation will increase our attention to validity claims about the world-- in other words, claims that the messages are "true". As it becomes more difficult for us to tell who is speaking, we may have difficulty in evaluating the intentions of the speaker. We may expand our willingness to trust the unseen stranger whose words we read, or we may place our trust in the computer itself as provider of messages. In the latter case, we may think of ourselves as engaged in conversation with an anonymous network, and lose sight of the humans whose programs and messages we are using.

5.2.1 Trust and the Nature of External Reality

Part of the problem of trust is a question of validity. Is what is being said true? Is the information in our computer data based accurate? In the past, we have arranged an elaborate set of social conventions about authorship, official editions, authoritative versions, and so on in order to define which information is reliable (Foucault, 1984). Until we have developed a new set of social norms for validating computer information, there will be some uneasiness about how much we can trust computer networks (Thompson, 1984). For example, new users of an electronic mail system go through a period in which they read everything because they haven't learned to trust its reliability (Hiltz and Turnoff, 1985). In both social interactions and interactions with computers, trust is based on experience. Our experience teaches us that our friend's loyalty can be counted on; we trust that the sun will rise tomorrow because it always has. When I worked on the first time-sharing systems of the 1960's our experience taught us not to trust computers very much. Crashes and catastrophic data losses were chronic. One company I consulted for had lost a million dollars worth of accounts receivable records. Today, although a computer is much more likely to work as it was designed to, there is still an enormous problem with data base validity. Our computerized information can at best approximately reflect external reality. Computer graphics raises a new problem of trust. Research has shown that people trust pictures more than they do printed words, and that they are more likely to trust a television image than a newspaper article (Graber, 1984). Thus we could predict that computer graphics will increase our trust in computer-based information. But computer graphics allow us to manipulate images in realistic-looking ways. As it becomes harder to tell "real" pictures from "retouched" ones, we may be reluctant to accept pictures as evidence of external reality (Brand, et. al., 1985). For example, one bank distributed a picture of its building in downtown Columbus, Ohio. A rival bank's sign had been deleted from the image and a parking lot had been filled with grass (Wall Street Journal, March 27, 1984:37).
5.2.2 Intentions in Conversations with a Computer
ANIMALS was a program distributed with the AppleIIe computer. A version of the game, "Twenty Questions", ANIMALS builds a data base as you play with it. As a player converses with ANIMALS, the program begins by trying to "guess" the animal you are thinking of with the question: DOES IT LIVE IN THE WATER? If you type "yes" you are asked if you are thinking of a frog; if you type "no" you are asked if you are thinking of a moose. If you weren't thinking of a frog or a moose the program "says": ALL RIGHT, I GIVE UP. WHAT ANIMAL WERE YOU THINKING OF? When you type in the name of a new animal, for instance "goat", the program asks: WHAT QUESTION WOULD HAVE THE ANSWER YES FOR GOAT AND NO FOR A MOOSE? If you supply the question "Is it a small ruminant?", the information that goats are small ruminants will go into ANIMALS' data base. In the sense that artificial intelligence researchers use the word "know", the ANIMALS program now knows the difference between goats and mooses. If you used ANIMALS long enough, it would have all the information about animal names and characteristics that you do. You would have made ANIMALS into a very simple expert system for guessing animals according to their traits. The educational function of the game (and presumably the intention of the programmer) is to teach you that you can have fun while putting data and relationships of your choice into a computer. An interesting article in The Humanist Sociologist (Brownstein, 1984) provided an entirely different interpretation of the conversational intentions of the ANIMALS program. The author was a novice computer user who assumed he was having a conversation with the programmer who wrote ANIMALS. Understanding the verb "to know" from a humanistic perspective, the author knew that a computer program can't really know anything. Not being able to know anything, it couldn't guess. So the author guessed that the programmer was making a joke. He missed the programmers' lesson about computer data bases but raised a fundamental philosophical problem. To whom are we speaking when we converse with a computer?

The answer from artificial intelligence researcher Richard Bolt is that we are engaged in a process of mutual self-disclosure: "The information base discloses itself to you as you disclose yourself to it (1984:87)". If this is the case, conversational interfaces not social activity. In place of the intentions of another speaker we have the external "reality" of the data base. This is an example of a reified social relationship. The relationship between the programmer and the program's user appears to be a relationship between the user and the program. As Kiesler reports from her study of social interactions in electronic mail systems, the user's attention gets focused on the message rather than on the people.

5.2.3 Computers and Conversational Norms
Studies of interruption patterns in conversations show that, in America, women are interrupted by men over 80% of the time. Women, who make less than 20% of conversational interruptions, are expected to allow themselves to be interrupted (Pfeifer, 1985). There are similar conversational norms allowing high status people to interrupt low status ones. By interfering with our ability to tell if we are conversing with a man or a woman, the computer interface could change the social norm of interruption of women by men. An alternative less favorable to women and low status people would be the establishment of social hierarchies in the access and priority designs of computer networks. Status differences can be preserved if our computer message systems tell us who is more important than whom. For example, the system on which this book was typed has several communications privilege levels. Faculty members can send files across the system; students are restricted to a subset of resources. The social history of the telephone gives reason for optimism in the prospects for equality in computerized communications. Early advertisements for the telephone depicted it as a device for giving orders to employees or servants; it became a means for two-way conversations between equals instead (Cherry, 1977). There is also reason for optimism in Reid's review of conflict and decision-making in the absence of vision. Although some studies report a reduction in cooperation and a lengthening of decision-making time, others report an increase in participants' willingness to change their opinions. Emotionally-based arguments, which often sway opinions in face-to-face situations, are less likely to influence the outcomes of computer-based discussions. Keisler (1985) reports the negative effect of removing emotion from discussion. Participants in her study were "out of control" and violated the rules of polite conversation. We use emotional expressions of approval or disapproval to exert normative power. With these ordinary mechanisms of social control missing, participants in computer based discussions are free to develop new conversational rules and arrive at unconventional decisions.

What these findings for telephone and computer network decision-making suggest is that computer-aided decision-making will be more rational in terms of the information available to participants, but less strongly controlled by shared (or conflicting) emotional evaluations. For low status participants, and in technical decisions where shared emotions are less relevant to the outcome than questions of fact, this should be an improvement. For decision-making about human goals, however, the absence of face-to-face contact can be a problem. The same non-verbal processes that reinforce feelings of inter-personal solidarity or antagonism also act to affirm shared values or to arrive at new ones. Therefore, it is possible that computer-aided conversation will loosen the "glue" among groups of friends, but also facilitate conversations among strangers.

5.3 Social Integration in Electronic Networks

The negative effects of computer-based communication -- loss of face- to-face contact, strains on trust, and a reduction in normative social control -- are balanced by the possibilities for social integration in communities formed around electronic networks.

Organizational Interfaces

Thomas Malone (1985,ref) defines an organizational interface as the parts of a computer system that connect human users to one another. A text sharing system, for example, is an organizational interface. The computer system I used during the mid 1980's for teaching had common storage areas for each class, so that students could use the same files. Assignments and exam questions were "read-only" to discourage unauthorized modifications by class members. An announcements and current events file was available for everyone to write in. For hackers and people planning careers in computer security there was a protected file called GRAFFITI. At the start of every academic term it contained the single line: "Students may read this, but they can not write in it." Today, the Graffiti file is kept on the GnuClassroom Blackboard.

Other kinds of organizational interfaces are management tools for keeping track of resources and tasks (Fox, et. al., 1983ref ; Kedzierski, 1982ref; Sluizer and Cashman, 1985ref). The design choices for such interfaces are an important macroergonomic issue. If, like telephone conference calls, they are made for cooperative interaction, the organizational interfaces will integrate individual activities and facilitate the development of group goals. If, on the other hand, organizational interfaces are designed to monitor activities and allocate tasks to people, they become the instruments of managerial control. They will supervise and coordinate individual activities without providing the social interactions that produce group solidarity.

In a review of organizational interfaces allowing many-to-many communication, Chandler Harrison (1981) identified several attitudes that inhibit people from using computer networks in social ways: o Feelings of isolation and powerlessness o Inability to understand the information available on the network o The belief that important decisions should be left to expert decision makers o The expectation that information comes in pre-made packages, rather than being exchanged through social interaction o The belief that competition is more advantageous than cooperation These attitudes are relevant to organizational interface design. If users are restricted to small subsets of system resources and given few choices, they will be isolated and powerless. If the interface is not user- friendly, people will have difficulty understanding what is availble to them from other people on the network. If the belief that only experts should make decisions is assumed by the designer, it may not be technically possible to use the system for any other form of decision-making. If information comes in packages supplied by the system, the "nature of external reality" cannot be negotiated by those using it. Finally, if those who build organizational interfaces do not believe that coopertion is worthwhile, it may be difficult to use the interface for cooperative purposes.

5.3.2 Community Interfaces

Our traditional concept of a community has a geographical base, with people in regular face-to-face contact with one another. Yet widely scattered people have maintained their social ties through letters, phone calls, and periodic gatherings. My family of several dozen aunts, uncles and cousins has an annual reunion; the professional community of American sociologists gathers at the end of every summer to go through the ritual of delivering papers and talking to colleagues.

One kind of computer network that facilitates solidarity within groups is the electronic bulletin board. Community bulletin boards like Berkeley, California's Community Memory Project, are experiments in social integration and democratic participation. Terminals in public places provide access for individuals who do not have computer terminals at home or work. Commercial information utilities like the Source or CompuServe offer bulletin board services to a variety of special interest groups (Glossbrenner, 1984). Churches and informal religious groups -- from born- again Christians to mystic pagans -- hold study classes, communicate with parishioners, and exchange information over computer networks ("Churches Move into the Computer Age," 1984).

But computer based social networks, like other communities, can generate conflict between groups. From the Aryan National Liberty Net, identified by the FBI as sponsored by a right-wing terrorist group believed responsible for several murders, you can get a list of America's enemies (New York Times, February 15, 1985:11). Hacker bulletin boards are the target of U.S. software and credit companies, who blame them for promoting credit-card fraud and software piracy (Newsweek, April 15, 1985:17). In their enthusiasm to do away with software pirates, some companies have suggested outlawing all non-commercial bulletin boards. In their defense, many of the bulletin board operators point out that they are encouraging free speech and social interaction, not theft.

5.3.3 The Social Use of Public Space

Even the best interface design for an organizational or community interface has its limitations. As Scragg (1985) points our by comparing his company's electronic mail system to Post-It notes, sometimes we want to put a message in a place, not send it to a person. If the message shown in Figure 12 were put in everyone's electronic mail box, it would not be as effective for group communication as a note on the door. The door is a public place; comments on it are group property. In most interface designs, there is no provision for public places or public commentary. And, without public social interaction, we cannot have communities (Sennett, 1974 ref.)

It is not enough to have public places where common information is stored and retrieved. As Jane Jacobs pointed out in The Life and Death of the Great American Cities, social interaction in public spaces provides social control over what goes on in them. Without norms for public conversation, public messages collect graffiti -- individual expressions of art, obscenity, philosophy, or protest. Figute 5.1 is a public message:


What about YOU and Nancy?

I don't drink coffee -- Nancy

Who is YOU and why don't you turn it off yourself? - B.G.

To everybody, from M. Lee:
would the last person to leave the lab *please* make sure that the last coffee drinker turned off the pot (so we won't have a fire) Thanks

The message about the coffee pot is not just information to be stored and retrieved. It is a public discussion about organizational behavior and responsibility. It is this sort of discussion that distinguishes a social interface from an interface that only coordinates individual human/computer connections.
[ Next Chapter | Main Index ]