Computers and Social Change by Judith A. Perrolle

[ Main Contents | Chapter 1 Contents ] Revised April, 1999

Part One. Information Technology and Society

Chapter 1. Basic Concepts

To understand the way computers and communications technologies are contributing to the transformation of the postmodern world, we must understand the social context in which they are created and used. In order to do this, we will have to begin with some basic concepts of information, society, and technology.


In a book comparing many scientific disciplines' approaches to the study of information, Fritz Machlup observed:
Information is not just one thing. It means different things to those who expound its characteristics, properties, elements, techniques, functions, dimensions, and connections. Evidently, there should be something that all the things called information have in common, but it is not easy to find out whether it is much more than the name. If we have failed and are still at sea, it may be our fault: Explorers do not always succeed in learning the language of the natives and their habits of thought (Machlup and Mansfield, 1983:4-5 ref ).
In the following discussion, that something called information is defined in a way that is useful for understanding the social consequences of computers and communications technology.

1.1.1 Data, Information, and Knowledge

The term information means many things in ordinary language. We may mean facts about the world represented by numbers, words or pictures (for example, the word "trees", a photograph of trees, or the phrase "four trees"). We also use the word information when we mean relationships among facts -- for example, the sentences: "There are more trees in Idaho than in Rhode Island" or "The maple is a common urban tree". Finally, in our evaluation of or understanding of the world, we speak of information like "Oak makes better furniture than pine" and "Trees help maintain humidity and prevent erosion on hillsides". To distinguish among these meanings, we can use the words data, information, and knowledge. Data
Data are specific numerical or symbolic representations of facts about the world. Data are the elements to be input, stored, and manipulated by the computer. During data processing, computers are simply used to transform facts from one medium to another, for example, printing a payroll check from a list of employee salaries. Validity refers to the question of whether or not our data adequately describe the reality they were meant to. Data are not valid just because we have copied them correctly from a data base to a computer program. They are valid if they accurately represent the real world. Invalid data can be either due to a correctable error (as when we type the wrong number at our keyboard) or due to a conceptual mistake (as, for example, if we used people's voter registration as Democrats or Republicans to represent which party they voted for in an election actually won by the minority party). Information
Information is a useful organization and selection of facts, not the number of facts available. Information is how data are organized. It involves relationships among the represented facts. Computers process information when they store, retrieve, or rearrange relationships among data. For example, a telephone book contains data representing the names, addresses, and telephone numbers of people in a region. The information in the phone book is the relationship between name and phone number and the alphabetical order of the names. If you know someone's name, you can use that information to find his or her phone number. There is less information to help you find the name of someone whose phone number you have -- you might have to look at each item of numerical data before you could use the information relating name to number. A phone book for the whole world would contain more data than a city directory, but it would not necessarily have more information. Indeed, it might contain too much data for you to find the information you want. If you are trying to reach the only Ohara family in your town, it would not be helpful to see how many Japanese or Irish people with that name have telephones.

A computerized phone book (such as those used by directory assistance operators) displays a page when a person's last name is typed in. When the operator selects the desired number, the computer plays a recording of the digits. Information is transmitted to the caller, but no new information is produced. In order to make information, new connections must be made among data. This can be done by physically rearranging the data, as when a person or a computer sorts records into some kind of meaningful order. As another example, if we put data on people's smoking habits together with data on lung cancer, we can produce information about the risks of smoking. Knowledge
Knowledge involves the evaluation and understanding of information. It refers to the meaning of information with respect to human interests and purposes. You could have a great deal of information available to you, as for example in an encyclopedia, without understanding what it meant or how to apply it to your own situation. In the telephone book example, your knowledge includes an understanding that the names refer to people and the number is a code enabling you to operate a machine to speak with them. Your understanding of how to use a telephone is a form of knowledge. So is the scientific knowledge of how telephone circuits operate and the cultural knowledge that you should call your mother on Mother's Day.

The expression garbage in; garbage out applies to computerized information in several ways. If the data are not valid, no amount of careful organization can make them represent reality. If the data are valid, but our arrangement is at fault, then we will not have accurate information. Finally, if we have valid data organized to provide excellent information about things of no interest or value to us, we have not contributed positively to our knowledge. Formal and Informal Knowledge
Written information is usually part of formal knowledge. This is knowledge that is consciously known and communicatable as a set of procedures. Informal, or tacit knowledge is usually acquired through experience, is often unconscious, and is frequently difficult to put into words. To appreciate the difference, try writing a formal description of how to ride a bicycle. My attempt to do this started with "Try to keep your center of gravity directly above the center of gravity of the bicycle." Then I got stuck trying to describe how you locate your own center of gravity. To appreciate the importance of tacit knowledge in information technology, dial into your least favorite voice mail menu or online help system and ask yourself what it is that people who answer telephones know that didn't get captured in the design. Wisdom
Wisdom is the ability to know when and how to apply knowledge. It is the ability to imagine the consequences of our actions for ourselves and for others. It is a capacity for good judgment that we acquire through experience. Just as we are only now beginning to understand the effects that our industrial society has had on the world's environment, we sometimes acquire wisdom through a long period of painful mistakes. Wise use of computers involves a careful study of their context and consequences, not only their inner workings. Yet, because our present knowledge about the effects of computers is incomplete, we may find in the future that some of our choices about computerization were unwise. Acquiring wisdom doesn't mean that we stop making errors, only that we never stop learning from them. Information Hierarchies
Sometimes information is treated as data, for example each item in a library catalog might be a whole book of information on some subject. A library's data retrieval system actually contains information about information. Higher-level information or meta information, as information about information is called, is a way of structuring data. Humans understand and interact with the world using hierarchies of concepts that are also structured in this way. In this case higher means more abstract, as in the sequence of concepts:
    this dog       (a particular object)
    all dogs       (a class of similar objects)
    the word dog   (a symbol representing a set of properties that
                    defines a class of objects)
    concrete nouns (a set of words defining properties of classes of
High-level computer languages allow programmers to refer to complex sequences of machine instructions or arrangements of data with commands that make sense in human conceptual terms. Advanced information processing systems allow users to create useful structures of relationships among different levels of information. At higher levels of abstraction much of the detail of data is lost, but important information is preserved. A road map, for example, preserves relationships among routes and distances between towns. It does not contain all the detail of an aerial photograph, but is much easier to read.

Another form of information hierarchy is a metalanguage, such as SGML (Standard Generalized Markup Language ). Metalanguages provide a set of standards for the formal description of other languages. SGML defines a descriptive markup language for electronic documents. HTML (HyperText Markup Language) is an application of SGML used to write web pages like this one. For an example, use the view file source option on your browser to look at this page. Data Base Architecture
When computer users complain that they cannot get the information they need from their data bases, it is often because their information retrieval system is little better than a data retrieval system. It does not let them select the relevant and avoid unwanted detail. Data base architecture is the study of how to match computer data structures to way humans organize and use information. In the 1960's data bases were little more than lists and tables. The computer user had to know a great deal about where data was physically located in the machine in order to use it. By the 1980's relational data bases allowed people to work more easily with information based on relationships among data. The growth of computer networking led to distributed data bases and virtual storage that made it generally unnecessary for users to know where data is kept in the machine.

Hypertext and hypermedia were data base architectures of growing importance in the 1990's. These are organized to connect text, graphical images, film, and sound data in ways that permit computer users to navigate through complex networks of information. An example of hypertext is the Electric Cadaver data base in use at Stanford University's School of Medicine. Students can browse through x-rays and slides, look up related text, and even simulate the manipulation of body parts.

The explosive growth of the World Wide Web in the late 1990's has created the potential for linking most of the world's information. It represents a serious challenge to the designers of search engines and data bases to move beyond keyword and page metaphors for information storage and retrieval. The web also represents a social challenge to our ways of creating, sharing, and controlling information.

1.1.2 Cultural Information

Unlike many animal species, most human knowledge of how to act is learned rather than genetically built-in. In complex societies, individuals do not carry all necessary information in their brains. Instead, they rely on the external storage of information in the form of culture. Though the word is popularly used to describe classical music and great literature, culture is a broad term defining the entire way of life shared by a people. Thus the English language, several religions, a two- party political system, soap operas, computers, fast food restaurants and garbage cans are all part of U.S. culture. While it includes our tools and artifacts, much of human culture consists of information. Language, ideas, beliefs about the meaning of life, and information representing our shared impressions of the physical world are all part of our culture. Symbols
Symbols are the units of cultural information. A symbol is a meaningful representation of some object or abstract concept. Cultural knowledge is expressed and shared through our linguistic and non-verbal communications. Although it appears that we have the genetic capacity to use symbols, we must be taught their meanings. Symbols can represent human emotion, as when we draw a heart to stand for love. Mathematical symbols, like the unknown x of algebra, represent abstract quantitative relationships. A crucifix and the American flag are examples of religious and political symbols. Money is used as a symbol for the economic value of goods and services. Words, which are themselves symbols, can be used to express visual and emotional concepts, as in poetry:
For all the history of grief
An empty doorway and a maple leaf
(MacLeish, 1962:51ref. Full text text )
In his capacity as Librarian of Congress, Archibald MacLeish expanded its collections to include film records of the Great Depression. These photographs of impoverished farm families, some of which can be seen in James Agee and Walker Evans' Let Us Now Praise Famous Men (1960), are powerful examples of visual symbols.

Computers can now make it easier for people to preserve culturally important images. Work in the field of computer vision has expanded the capacity of computers to process video input. Developments in pattern recognition programs can now recognize letters, numbers, geometrical shapes, and even fingerprints. But computers have not been very successful at handling abstract symbols, especially those referring to the emotional qualities of human experience. Sacred and Secular Knowledge
Historically, our culture has conceptually divided reality into the realms of the secular and the sacred. The secular is the ordinary reality of science and everyday life; the sacred is the realm of religion, magic, and the supernatural. As scientific understanding developed, the domain of the secular expanded to include first astronomy and physics, then chemistry and medicine. With each expansion, there was a social struggle to replace mysterious explanations with rational ones.

During the history of scientific knowledge, the secular and the sacred were not so differentiated. Myths traditionally explained the relationship of human beings to time, life, death, and the sacred (Campbell, 19**). According to Giorgio de Santillana and Hertha von Dechend (1969), myths were the form in which our earliest scientific knowledge was passed from generation to generation. Although a myth has the form of a story, it is a high-level symbolic expression of the workings of the universe.

Myths today still have the power to explain the human condition in symbolic terms. An example is the Greek myth of Prometheus, who was punished for bringing technology to humanity. Although we do not believe the story to be true in the scientific sense, the Promethean theme has been used by scholars to symbolize the unforseen consequences of technological change. In applying it to computers, Patricia Warrick (1980) argues that the myth is a warning that nature has placed limits on humanity's ability to create and control. The Two Cultures
C.P. Snow (1963) introduced the term the two cultures to describe the twentieth century split between people seeking knowledge through scientific inquiry and those interested in knowledge of religion, the arts, and the humanities. Although cultural information is not neatly divided between the two, there can be misunderstanding between technically and humanistically oriented people, as if they came from different societies rather than sharing the same culture. Humanistic knowledge is a way of understanding what the world means. It is a source of wisdom based on our society's whole range of experience. Scientific knowledge is an understanding of how the world works, validated through a careful process of experimentation. We expect it to change over time. Although cultural symbols also change, we tend to experience meaning as absolute truth validated by inner faith.

Although no one can know their entire culture, people who learn only one way of understanding are ignoring an entire dimension of the world. If they knew only how things worked, their lives would be without meaning. If they were entirely ignorant of technological culture, they would find modern society full of mysterious phenomena and incomprehensible machines. When the two cultures are merged, we find people who are intrigued by both the how and the why of the universe. Some modern scientists and poets have bridged the gap between the two cultures (Hoare, 1987). Today in the computer field there are philosophers, musicians and historians. There are also a growing number of applications of computers to the work of writers, artists, and performers. Computer science, by adding a new dimension to the question "What does it mean to think?" is making a new contribution to an age old philosophical question as well as expanding the realm of scientific knowledge. Computers and Cultural Values
The selection criteria that define what information is relevant to human purposes are part of our culture's values. Values are the oughts and shoulds of society. American values include respect for individuals, freedom of speech, property, and equal opportunity for all. We also value things like automobiles, health, money, and fresh air. Sometimes, as in the case that we ought to be able to drive automobiles and we ought to be able to breathe unpolluted air, cultural values are contradictory.

Cultural analysts agree that computers themselves are highly valued in American society, and that this evaluation will have consequences for the rest of culture. Sherry Turkle (1984) predicts that the experience of using computers will cause us to devalue calculation and logical reasoning. In other words, the ability to calculate and reason logically will become less important for people as it is done more and more by machines. Instead, she finds computer users placing higher value on emotion and feelings to define what it means to be human. Daniel Bell (1980b) believes that information will become more highly valued, with the ability to use it becoming our most important skill. Joseph Weizenbaum (1976) has suggested that computer-based data will become so important that we will neglect our cultural traditions and fail to explore new nontechnological areas of human experience. Echoing the theme of the Promethean myth, he fears our fascination with the power of the computer to let us design and control imaginary worlds will lead us to tragedy in the real world of social cooperation and conflict.

1.1.3 The Scientific Study of Information

Traditional Western culture makes a sharp distinction between ideas and material objects. The notion that an idea or other intangible form of information can be studied scientifically is still at odds with cultural beliefs about the difference between between "real" objects and "unreal" abstractions. For people who believe that information is ideas and that scientists study things, measuring information seems impossible. However, to the physicist, our familiar world of solid objects is a representation of a complex universe consisting energy and empty space. Since Einstein, we have known that matter itself is a form of energy. Yet scientists are like everyone else when it comes to treating tables or trucks as solid objects. At one level of understanding they know that a truck isn't "really" solid. At another, they understand why being hit by one would feel solid. Information, although it is abstract, is "real". It can be studied and measured. Measuring Information
Some people object to measuring information because they do not want to reduce symbolic expression of human relationships to mere data. The philosophical concept of reification describes such situations. For example, if I treat relationships (such as my friend's affection for me) as if they were merely things (the box of chocolates he gave me), I have lost something important about human love. If I evaluate a book by counting its pages, I will be mistaking data for knowledge. Because human understanding has many levels, a computer analysis of dancers' motions does not reduce the symbolic meaning of a ballet to numeric data. Instead, it offers an additional way to understand dance. The scientific study of information becomes reductionistic only if we mistake one approach to knowledge for the only way to understand.

Information measurement in the computer field owes a great deal to Norbert Weiner (1948) and Claude Shannon (1948). They defined the quantity of information in a system as a statistical measure of its organization. Shannon defined information as the probability of a message being transmitted. Improbable signals contain more information than highly probable ones. For example, if someone tells you something they've told you many times before (and that you expect them to keep on telling you over and over), there's not much information in the message. From this perspective, a new artistic expression has more information than a repetition of a traditional cultural form.

Shannon's approach to measuring information is now most commonly used in the fields of telecommunications and electronics. The signal to noise ratio can be used as a measure of information. Noise is the highly probable, randomly-generated part of the transmission (for example, the static during a telephone conversation). The signal is the non-random, information-bearing part of the transmission (for example, the voice you are listening to in a noisy room). In studies of information transmission, the focus is on the speed and accuracy with which data can be communicated through a variety of electronic media. This approach is invaluable for the development of computer hardware and communications software. It does not, however, really consider the meaning of information to humans. Thus linguists have criticized Shannon's definition by pointing out that a sentence like "Fred is a dog" contains more information than the less frequently heard sentence "Fred is a mammal", where Shannon's theory predicts it should contain less information. However, his theory was intended to deal with message transmission, not with symbolic meaning. Nor is it suggested in Shannon's theory that more data transmission through computer networks will automatically add to our cultural knowledge. Cybernetics
From Weiner's perspective, information was the key to the way machines or living organisms modify their behavior to take into account the outcome of their previous actions. He defined the field of cybernetics as the study of communication and control mechanisms. Feedback is the information process that allows an organism or a machine to be self- regulating, as in the way a thermostat works to maintain room temperature. Information about how hot the room is causes the thermostat to turn the furnace on or off, changing the temperature of the room in the desired direction. Another example is when you reach for a moving object. You see where your hand is in relation to the object, then use that information to correct the motion of your hand. This sort of feedback is an essential part of the way humans survive in their environments. The application of the feedback principle to computers allows us to build robots that can correct their own behavior. It also allows us to design "self-regulating" computer controlled machinery. Entropy
Entropy is a measure of the natural tendency of physical systems to become disordered. For example, most of us comb our hair every morning; by afternoon it has become disordered "all by itself". Throughout our lives we put energy into activities like combing hair or straightening up our rooms. By random processes of the wind blowing and our putting things down anywhere, our hair and our rooms get messy. Of all the possible ways our heads and houses could be arranged in space, only a few are "neat"; the vast majority are not. Thus the untidy state is much more probable than the neat state. Weiner related information to negative entropy by showing that it had the same mathematical properties. Information is analogous to negative entropy because we use it to create improbable, organized systems.

When we pick up a room, we scan the situation, locating objects in space and comparing their distribution to our mental pattern for "clean room". We select each object that is out of place and put it where it belongs. As we identify, select, and relocate objects, we are using information to identify objects and feedback to observe and control our cleaning activity. As we work, the arrangement of objects in the room gets closer to our mental goal. To appreciate the way you use information to create order, try cleaning a room in the dark. Unless you are blind and used to identifying objects by touch and sound, you may find it quite difficult in the absence of visual information.

Computer novices often have similar problem keeping track of their files. Without the visual feedback they are used to from books and papers, they have trouble imagining "where things are" in the computer. Even experienced programmers find it helpful to draw "pictures" of their data structures. This illustrates the indispensability of mental concepts. We cannot create order unless we have in our minds a set of criteria for selecting and arranging the objects we are trying to organize. These non- random mental criteria for identification, selection, and action are themselves a form of information stored in the biochemical processes of our brains. In the computer, they can be made part of information processing software or hardware. In writing software or building hardware, they are an essential element of design. The entropy concept also underlies the need for computer hardware and software maintenance. Computer systems will become disorganized unless we continue to use energy and information to keep them functioning properly. Although maintenance jobs are sometimes viewed as unexciting, they are a large part of computer system costs and an essential ingredient in their success (Couger, 1985). Information and Social Science
Social scientists measure cultural information in ways like asking people what they believe or observing how they communicate. Historical research gives us information about cultures of the past. Studies of cognition and education provide us with knowledge of how people learn. Like cybernetics, sociology is concerned with how information affects behavior. Only sociologists study social facts rather than physical phenomena.

A social fact is a cultural phenomenon that has consequences for human behavior. Cultural values, for example, are social facts. So is the common sense wisdom of "what everybody knows". Social facts may be true in the scientific sense, for example if students majored in computer science rather than history because they believed that starting salaries were lower for historians. Often, however, social facts are not based on accurate knowledge. As an example, so many people believed that AIDS could be spread by casual contact that individuals with the disease were fired, evicted, and banned from school. Our scientific information indicates that AIDS is spread only through sexual or direct blood product contact, yet the social fact of erroneous medical knowledge produced very real patterns of fear and discrimination. When we explain something using social facts we refer to the way human society is organized and what people believe, rather than looking at physical phenomena.

Although the astronomers and navigators of Columbus' time had good scientific evidence that the earth was a globe, his expedition suffered from public fear that ships would fall off the edge of a flat earth. Among the social facts about computers are:

  1. They never make mistakes.
  2. If something went wrong, it was a computer error.
Like cultural values, social facts are sometimes contradictory. Measuring the Value of Information Products
In order make and sell information products like computer software or data bases, businesses must be able to measure the costs of producing them and their value to consumers. We often think of an amount of information in terms of its physical form -- a book of 354 pages or a disk file of 12 tracks. The cost of producing information, however, is often not directly related to its physical size. For example, if your job were to make a mailing list of parents who might be interested in buying ACME Expensive Baby Toys, you could start by computerizing records of recent births in your sales region. It would be expensive to travel to your state's Department of Vital Statistics, select the appropriate birth certificates, and enter the data into your computer. Your final product, while large, would contain many names of people not interested in your Expensive line of toys. A potentially cheaper way to acquire a better list might be to hold a contest with Expensive Toys as prizes. The list of contestants would be your information product. In this case, much of the work of identifying who might be interested in baby toys would be done by the prospective customers, not by ACME clerical employees. Even cheaper might be to sort and copy part of the National Toy Company's computerized records of who recently ordered baby toys. If National were unwilling to let ACME access the data, a bribe to one of their employees or an attempt to break into National's computer might obtain the list anyway.

Besides the ethical and legal problems involved in the above example, it should be clear that the cost of making an information product is not proportional to its size, but depends upon how much effort and expense is involved in locating, selecting, and arranging the data. Once made, an information product can be copied at little cost. Size is usually directly related to the costs of reproducing information products, and it is often exponentially related to the time it takes to search for information in a library or data base. It is not, however, a good measure of the original cost of making the product.

The value of information products to consumers depends upon what they want to know and how difficult it would be for them to get the information elsewhere. Because people expect some kinds of information to be freely available, they may resist having to pay for it. If the costs of information products is high, people will be tempted to copy them, feeling that "stealing" information doesn't really harm the original. As we will see in Chapter 8, the difficulties with protecting information property present a new challenge to our legal system.

1.1.4 The Data Explosion and the Information Lag

The growth in the production of information (as estimated by the number of periodicals or books published each year or by the approximate number of words communicated by electronic media) has been called "the information explosion". Since the 1960's the volume of printed media has grown quite slowly while electronic media (including television) have expanded even more rapidly than before. However, the "information explosion" should probably be called the "data explosion". This is because we are being confronted with facts faster than we are able to integrate them into useful information. Information Overload
The human capacity to read or hear this increased volume of words has not expanded appreciably. Nor is it clear that all the words being produced are particularly relevant to us. We are facing, according to many analysts, an information glut that threatens to overwhelm us. Results from a research project on how Americans process the news indicate that people have several effective strategies for handling too much information. According to Doris Graber (1984) we pay full attention to only about 20% of news we read and less than 10% of TV news. Our selection schemes ignore redundant information, accepting what is important to our own way of thinking. While details are remembered haphazardly, our fund of general knowledge grows. By organizing data into information and information into knowledge, we are able to reduce the volume of facts with which we are confronted. The Crisis in Data Processing
According to James Martin, the computer industry is facing its own information overload. We are often unable to get the information we need from our computerized data. There is an information lag between our relatively slow progress in information reorganization capability and our rapid progress in data base capacity. The information lag is the more common experience of managers and others who want information relevant to their decision-making processes. A student might experience the information lag by locating 483 publications with a computerized library search -- on the topic of a paper due the following week. The problem of how to select the "best" references is beyond the capacity of most automated library systems. A trip to a bookstore in search of an introduction to programming a home computer provides a similar experience. One computer professional suggests the following solution to dealing with the estimated 75 technical and 175 popular computer magazines: when the pile of unread publications gets two feet high, throw them all away because the technology they describe will be obsolete (Turner, 1984). More seriously, Hiltz and Turoff (1985) point out that computer systems can be designed to reduce information overload by automatically filtering unwanted input. For instance, if you received your mail through a computer, you could program it to throw away junk mail.

1.1.5 Intelligent Information Processing

Computer science is grappling with how to get computers to apply human selection criteria to data. Many experts believe that, if the "information lag" is to be overcome, we must develop software that can help us reorganize data and information. One solution is to find better ways to computerize graphs and pictures organized so that the observer can easily grasp the general concept displayed. Visual displays of quantitative information have a long history (Tufte, 1984) and are often preferred for business or educational data presentation.

The field of computer science known as artificial intelligence, or AI, involves the design of computer programs and automated equipment, such as industrial robots, with a limited capacity to behave in ways that at least resemble human thought processes (for a technical survey, see Barr and Feigenbaum, 1982, Hayes-Roth, 1983, or Coombs, 1984; for a sympathetic popular history, see McCorduck, 1979). Information from the outside world can be sought, interpreted, and used as the basis for "heuristic" decisions which in humans would be called "best guesses." The programs can, within the narrow range of the world to which they are applied, draw inferences, suggest solutions to previously unsolved problems, select relevant information according to their own internal criteria, and modify their own behavior as a result of the outcomes of their previous actions.

Automated programming, industrial planning by machine, and mechanization of the professions were topics on the agenda of a 1958 international conference on the emerging field of artificial intelligence (National Physical Laboratory, 1959). In addition to saving labor, managerial control and profitablity were among the reasons advanced for why AI should be supported. During the next twenty-five years, artificial intelligence was transformed from academic research projects to widely publicized commercial applications (Feigenbaum and McCorduck, 1983; Hayes- Roth, 1984). Knowledge Engineering
Expert systems are a type of AI software developed by knowledge engineers. They promise that their software will "capture" the knowledge of experts in programs that enable a less skilled person to achieve expert results:
Knowledge is a scarce resource whose refinement and reproduction creates wealth. Traditionally the transmission of knowledge from human expert to trainee has required education and internship years long. Extracting knowledge from humans and putting it in compatible forms can greatly reduce the costs of knowledge reproduction and exploitation...skill means having the right knowledge and using it effectively. Knowledge engineering addresses the problem of building skilled computer systems, aimed first extracting the expert's knowledge and then organizing it in an effective implementation (Hayes-Roth, Waterman, and Lenat, 1983:5,13)
The theoretical possibility of representing human knowledge and decision-making processes in computer programs has been fiercely debated on both scientific and moral grounds, with the strongest objections coming from the philosopher Hubert Dreyfus in What Computers Can't Do (1972) and the artificial intelligence expert Joseph Weizenbaum in Computer Power and Human Reason (1976). One important issue is the degree to which human decision-making is believed to be rational and logical. Intelligent software has been most successful or those applications in which the knowledge of human experts is very well understood and rather routine. Critics of knowledge engineering doubt that computers can actually be designed to handle any but the simplest symbolic meanings.

While the debate between those who argue that machines can think and those who argue that they can't continues (Boden, 1977; Haugeland, 1981), the practical success of "intelligent" programs which play chess, infer chemical structures from molecular data, and diagnose illnesses indicate quite clearly that artificial intelligence is being "put to work" at industrial and professional tasks, despite the reservations of many theorists.

The most ambitious practical proposals of the 1980's involving expert systems were those for the new 5th generation "supercomputers" (Feigenbaum and McCorduck, 1983; "Supercomputers: The High-Stakes Race To Build a Machine that Thinks," 1983). Promising higher industrial productivity and greater national security, the proposals called for many areas of military and civilian expert decision-making to be turned over to the faster, soon-to-be smarter machines. In his critique of the fifth-generation idea, Weizenbaum questions Feigenbaum's assertion that computers will produce the future knowledge of the world, asking how are we to understand just what information the computer produces and how (Weizenbaum, 1983). But if information itself is seen as a product made for profit by efficiently organized employees, then information can be produced by the computer in the same way that products were made by the factory machinery of the first industrial revolution. Search Engines on the World Wide Web
In the 1990's the challenge of intelligent information processing turned to problems of searching networked databases, especially the World Wide Web. Powerful search engines like AltaVista made use of high speed computer processors and emerging standards for indexing and searching SGLML and HTML documents. Libraries, governement agencies, universities, and corporations began to build huge databases of searchable documents even though serious social issues of property rights, privacy, authenticity, and security have not been solved. As individuals made their own work available through web homepages, search engines like Yahoo! configurable search engines for personal use.

The development of computer programs called intelligent agents represent a use of expert systems to empower information seekers. Much of their use by corporations is in the areas of marketing and entertainment. It remains to be seen how these programs will be applied in the long run to the work of librarians and other information experts. Intelligent information processing can mean the automation of intellectual work as well as the enhancement of information retrieval.


In order to understand social change, we must have a conceptual model of society. In other words, we have to understand the "something" that is changing. Throughout history, models have been proposed as convenient ways of thinking about society. Society as a family with ruling "parents" or society as a self-contained city are two traditional models. During the past few centuries, models of society as an evolving biological organism or as a machine became popular with theorists. Today the model of society as a system is attractive to many social scientists, particularly those using (functionalism) or social networks theory.

1.2.1 System Characteristics

A system is a model of specific components interconnected by well- defined relationships. A system has a boundary that separates what is being analyzed from the rest of the world. Input/output characteristics describe the way the system interacts with its environment. An example of a simple system is a lightbulb, a switch, and a battery connected by wires. The bulb, switch, and battery are components; the wires form the interconnections between them; everything else is considered outside the system. Light is a possible output; the hand that moves the switch can be considered an input.

Matter, energy or information may flow through a system. System processes describe the way components act on one another and on the material flowing through. In the above example, energy flows through the system; the bulb processes electricity to produce light. The function of a system is a description of what it does; our example provides light. The function of a component is what it does within the system; the switch controls the flow of electricity and "remembers" if the light is on or off.

The state of a system is one possible arrangement of parts, with each part in a particular condition. For example, the bulb, switch and battery system has six states:

  1. Bulb burning, switch "on", and battery supplying electricity.
  2. Bulb not burning, switch "off", battery not supplying electricity.
  3. Bulb not burning, switch "on", battery supplying electricity.
  4. Bulb not burning, switch "on", battery not supplying electricity.
  5. Bulb burning, switch "off", battery supplying electricity.
  6. Bulb not burning, switch "off", battery supplying electricity.
When the system is functioning as intended, there are only two states, "on" and "off", associated with whether or not the light is burning and the switch is set to accomplish that purpose. If we have a defective switch, a loose connection, a short-circuit, a dead battery, or a burnt-out light bulb, other states occur. An investigation of the way a system changes from one state to another can often help us understand how it functions (or doesn't). In investigating the relationship between the battery and the bulb we would discover that electric current must flow through the system if the bulb is to burn, and would have made an empirical generalization about the little world of our circuit.

Once we understand the functions of a system, we can begin to predict its behavior, as when we expect that the bulb will light when we close the switch. However, even in the simplest of systems like the one above, we cannot predict all of its behavior (such as the conditions under which we can expect short circuits and burnt out bulbs). This is because our conceptual model oversimplifies the system we are observing and fails to take "everything" into account. This is especially so when we have ignored inputs to our system from the "outside" (how did the battery get charged?) or fail to understand lower-level processes (how does the electricity come from the battery and move in the wires?). Dead batteries and short circuits are only understandable if we know more about the situation than we have modeled here.

In large systems there are so many possible states and transitions between them that we cannot predict, except in probabilistic terms. For some large systems that can be formulated mathematically, we may build computer simulations or solve mathematical equations to make predictions. For less well-defined systems, like weather patterns or societies, the mathematics of systems analysis is very difficult to apply. For these, the concept of a system remains a useful aid to thinking, but does not often provide a method of quantitative analysis.

1.2.2 Social Interaction

As part of identifying a shared way of life, cultural information defines appropriate behavior that individuals are expected to engage in. These behaviors occur as informal social interactions. A social interaction is a situation in which two or more people communicate and modify one another's actions. Social interaction is the fundamental process in a social system. Conflict and cooperation are two of its forms.

The social constructionist area of (symbolic interaction theory) describes and predicts how social structures are created out of the interactions of individuals acting within the social system. This is in contrast to other theories which see human behavior as determined by the external constraints of their social environments. Information and Power in Social Interactions
From the individual's perspective power is both the ability to affect the physical environment and the ability to make other people do what you want them to do, even against their will. From a social perspective power is the ability of groups to interact successfully with one another and the physical world. The capacity to motivate individuals for cooperative purposes is an essential part of a society's ability to survive. Although we may think of our power in terms of military strength or superior technology, our abilities to investigate problems and negotiate solutions are just as important. Power is more than the ability to control others; it is also the capacity to organize effective action.

The exercise of power requires information. We cannot influence people unless we can communicate with them. We cannot offer them material rewards to do our bidding unless we can come to an understanding about the exchange. We cannot even forcibly move people or objects without knowledge of where they are vulnerable to our efforts. Planning long-term actions requires procedures to gather new information, evaluate it in terms of shared goals, and use it in choosing a course of action.

Although coercive power, or force, occurs in social interactions where the will of one person or group is imposed on the unwilling, most power in interactions is of other sorts. We may be influenced by others because we like or respect them. Or we may do what they ask because we think it is legitimate (right or legal) for them to give us orders. This is called normative power, named after norms (the unwritten -- often even unspoken -- rules for how to behave in specific situations). Wearing clothes in public, pausing in a conversation so that someone else may speak, and not eating one another are all examples of norms. We tend to think of norms as human nature, but children have to be taught to dress, not to interrupt, and not to bite. Norms are based on the more general moral and ethical principles, cultural values. Computers and Social Interaction
Since the meaning communicated through information exchange is the basis of social interaction, one important impact of computer technology is the way it changes the process of interpersonal communication. For example, microcomputer users are beginning to form electronically-based communities organized around the exchange of information about common interests. However, in some computerized workplaces, employees report increased isolation and a lack of human contact on the job. At present, the computer's effects upon human social interaction appear contradictory, as will be explored in Part Two of this book.

Computers are sources of power for those who use them to manage information. In some cases individuals or groups can use computers to increase their power at the expense of others. In other cases, the use of computers can make it easier for people to negotiate and reach decisions. The uses to which computer power can be put is the subject of the remaining sections of this book, especially in the concluding chapter where their effect on social decision-making is explored.

1.2.3 Social Structure

Repeated social interactions help define social groups such as families, teams, or networks of friends. Relatively stable patterns of interaction among components form the social structure of a group. When social scientists speak about the structure of a society (or smaller social group) they refer abstractly to the concept of relationships among people performing particular activities. For example, the structure of a softball team involves nine positions to be filled by real human beings, as well as a set of expected behaviors associated with each position. Since some social positions are similar, we can group them into categories like "infielders" or "outfielders" to make a simpler model of social structure. Social Roles
Positions in society, along with their associated norms of expected behavior are called social roles. Roles often occur in pairs -- doctor and patient, parent and child, teacher and student, pitcher and batter. Socialization is the process through which we learn roles, norms, and values. From a systems perspective, socialization produces new components whose properties fit into the existing social structure. From an individual's perspective, socialization is the process of learning how to behave from families, schools, friends, and co-workers.

Some norms for roles are formal rules. In the baseball example: "Third base players may not pitch to the batter." Others, such as: "Shortstops cover second base when the second base player goes after an infield ground ball," are informal rules. In the case of formal rules, special social positions often exist (for example baseball umpires) to make judgements and enforce expectations. In the case of informal rules, people apply social pressures (for example, dirty looks, praise, or a shove) to keep others behaving properly. Norms are essential to cooperative human activity. Social interactions to enforce norms are one part of the process of social control. Cooperative forms of social control (such as making sure a church congregation behaves reverently) are generally based on symbolic communication more than on force or economic power. Status
Occupational roles -- programmer, engineer, administrator, data entry clerk, and so on -- are an important part of how people define their "place" in an industrial society. When some of these roles are believed to be more honorable, more powerful, or more important than others, sociologists describe them as being ranked by status. For example, a softball player will be evaluated by his or her teammates not only according to how well he or she can play, but also according to the status of his or her position. Although players bring personal characteristics to their positions, right fielders are presumed to be less able than center fielders. Sometimes characteristics people are born with and cannot change (like sex or race) are used to assign them to positions, or to make judgements about their expected performance. For example, when a female comes up to bat, many male outfielders move in on the assumption that she is a poor hitter. In professional baseball, black men were presumed unqualified until Jackie Robinson began his career. Part Three of this book examines the changing status of jobs in the computer field and looks at the evidence of whether of not they are open to everyone on the basis of merit.

1.2.4 Institutions

The larger elements of social structure are fairly durable arrangements of social roles called social institutions. Schools, churches, and the family are all social institutions. These involve complex patterns of expected behavior on the part of people in such roles as student or professor, minister or member of the congregation, parent or child. The impact of computers on social institutions occurs as the technology changes the ways people are expected to behave. The use of computer-aided instruction, for example, can radically alter the mutual expectations of student and teacher.

Institutions like business, government, and the military make up the economic and political structure of society. Law, government, and the other institutions supporting a democratic political process in society perform several functions. They are the way we make decisions affecting all of us, the way we allocate our public resource, and the way we establish official agents of social control. As discussed in Part Four of this book, these institutions are changing as we introduce new information technologies.

Economic institutions produce and distribute society's material goods and services. Computers and communications technology is being used to redefine the tasks expected of employees. Computer applications have begun to alter business management, product design and marketing, and financial record keeping. Computer technology is also being used to alter the basis of our economy -- property. Property
Property appears to be a relationship between people and things. For example, if you own land, a car, or a dog, you may think you have absolute control over what happens to it. However, social custom and law set limits on your rights to use and dispose of property. You can't plant marijuana or block public access to a beach without penalty. What you can do with (or in) your car is limited by regulations and standards of acceptable public behavior. Surveillance by your neighbors and the Society for the Prevention of Cruelty to Animals discourage you from beating your dog. Thus property is also a relationship between people. Cultural information defines social controls over property as well as ownership rights.

Computers affect property relationships in two ways. First, information production is changing the kind of industries we have and the sorts of jobs available. Because some individuals and companies are better able to take advantage of these new economic opportunities, there will be some changes in society's distribution of wealth. The issue of how information is to be used is a second way computers affect property relationships. The democratic social values of privacy and freedom of information are often in direct conflict with our concepts of personal and corporate intellectual property. Stratification
Stratification is an institution based upon the inheritance of unequally distributed property, power, and status. As with other systems, stratification is defined by specifying the relationships between components. Here the relationships are ranks based on resources, power, and influence. Individuals, families, or large groups with the same economic status (called social classes) are the components of a stratification system. When studying stratification, sociologists usually look at occupational roles or at classes. Processes within the stratification system include social interactions involving the exercise of influence and economic or political power, especially those that pass on advantages to our children.

In societies where individuals can choose their jobs, their religious and other group memberships, and can raise or lower their social rank through education and effort, a person's place in the stratification system is only partly inherited. During the social mobility process people rise or fall from the status they received at birth. Where mobility is possible, the institutions of family and education are where people acquire the skill and training to be "successful" or are judged "failures".

The computerization of work will probably be the major mechanism by which the computer alters social stratification. Because so much of a person's social status in modern societies depends upon his or her occupation, changes in the types of work people do (especially if there are corresponding changes in wages and salaries), can drastically change the stratification system. If many new jobs are created at the "top" of the social structure, more individuals will have opportunities for success and status. If, on the other hand, new jobs are created at the low-wage "bottom", it will be more difficult for individuals to gain social status. If computers are seen as appropriate for use mainly by men, status opportunities for women could be restricted. If educational institutions provide computer science education mostly to middle class children, poor and minority children could experience even greater barriers to occupational success.


Computer technology represents the introduction of new tools for communication, information processing, and remote control. Technology, however, is more than tools. It involves the social processes which produce tools, the social behaviors involved in using tools, and the socially defined meanings of tools. Information about tools, whether their designs or the techniques for using them, is an essential component of technology.

1.3.1 Tools

In its most general sense, a tool can be defined as an object or agent through which human activity is directed towards some goal. Thus I can speak of my computer as a tool for writing this book. If I called my research assistant a tool for writing this book, it would be insulting as well as inaccurate. She has her own goals and is self directing; tools are only instruments. In social relationships involving power over others, people are used as tools for someone else's purposes. In cooperative social interactions, people use the power of influence to arrive at common purposes.

The purpose of tools can be as specific as the zax (used for punching holes in slate roof tiles) or as general as a rope (with thousands of uses, from walking a dog to putting up a flag). Tools are often used as extensions of the human body to gather information about and to manipulate the physical world. Microscopes and telescopes extend our vision; hammers and space probes extend the reach of our hands. Tools like cameras or tape recorders store sensory information; tools for writing and painting allow us to make a durable record of our inner ideas and visions that can be shared with others. Information storage media, from stone carvings to data bases, facilitate the communication of information from person to person and from generation to generation.

Computers can be very specific tools (for example, to regulate an engine's performance) or very general-purpose tools such as the programmable digital computer. Although some people still consider the computer to be useful only for computation, computers are tools for communication and control of all types of information. Analog computers handle non-digital processes (like monitoring an electric current or the temperature of a room); multimedia capabilities enable us to process visual and audio information; peripheral devices such as remote sensors can process air pressure, chemical composition of the atmosphere, and a host of other data.

As an information processing tool, the computer's major characteristic is the speed with which it processes extremely large quantities of data organized in complex ways. Although computers are popularly noted for their perfect accuracy, all large and interesting computer systems are prone to error. Hardware and software bugs, human errors in data entry, and the built-in possibilities for less than perfect performance (such as the ability to "guess" or "forget" that is a feature of the heuristic programs used in artificial intelligence) mean that computer technology is not the way to perfection. For many applications, however, computers offer more efficient means of performing tasks than previous methods.

The control over geographically dispersed information is an extremely important feature in business and military applications, as well as in the communications industry. Computer technology provides us with remote controlled extensions of our bodies. Remote sensors used in satellites and space probes extend our ability to gather information on subjects as diverse as the vegetation of Africa or the rocks of Mars. With telecommunications equipment we can hear from any part of the earth and far into space. Via robotics, we can work from a safe distance on the ocean floor or with hazardous chemicals. Also, and more dangerously, computerized weapons have vastly extended our ability to throw deadly objects at one another.

Computer-based decision-making is at the heart of the integrated software systems now being designed for industrial and military uses. These systems coordinate decisions from the purchase of raw materials through automated plant operation to customer billing. Although the expression "Computers only do what you tell them to do." has become almost a folk saying, decision-making by computer is becoming increasingly sophisticated.

Perhaps the most striking characteristic of the computer is its extension of the human mind; both our memories and our abilities to calculate and reorganize information have been enhanced by computers. Edward Feigenbaum (reference) believes that we will enter into a "partnership" in which computers perform calculation and memory functions, while humans exercise their analytic capacities. The danger in this, expressed by Joseph Weizenbaum (reference), is that we will neglect those areas of human judgement and reason which cannot easily be computed. He fears that the new doors opened by computer extensions of the mind will close other, more important doors of human thinking.

1.3.2 Technique

In order for any tool to be used successfully, the technique for using it must be understood. A technique can be thought of as a method for performing a task, without necessarily including a full scientific or social explanation of what is being done. The technique for driving a car can be learned without any understanding of how an internal combustion or diesel engine works. No knowledge of traffic rules or the consequences of driving head-on into a truck is necessary to put in the key, start the engine, put the car in gear, and go. The techniques of safe driving include a broader understanding of the social consequences of making a car run. The techniques of automotive design include a much fuller scientific knowledge of the principles behind the car's motion. The techniques of social impact assessment would include an understanding of the consequences of the automobile for such phenomena as the patterns of urban residence, air pollution, energy resource use, and the industrial structure of the economy. "Computer Literacy" as a Technique
The techniques for using the computer can be as simple as a set of instructions for turning on a machine and entering data into a packaged program or as complex as the knowledge of how to design hardware or software. Most computer users, even those within the profession, know how to use the computer in a limited fashion. Very few of us could build one, especially if we had to create our own chips from materials we dug out of the earth. This is because contemporary industrial society depends upon an enormously complicated and interrelated set of techniques for making and using tools. Social theorists like Ellul have even suggested that our modern techniques are so complex that modern technology can be understood by individuals only in fragments. Despite a great deal of recent discussion about "computer literacy", most individuals, even in the most computerized of futures, will need little knowledge about computers to go about their daily lives. Like "driver's education" for motorists, computer literacy can be the teaching of simple techniques. As automobile technology became more complicated, fewer drivers knew how to repair their own machines; fewer still could design and build them. Automobiles became, with the introduction of automatic transmission, more "user-friendly". Most of us use cars; few of us understand them or appreciate the social changes resulting from their use. If we are to understand computers better than we do automobiles, we will have to look beyond the techniques for operating them. In the meantime, many Americans are struggling to learn computer techniques. Because there are relatively few skilled individuals to help them, many people learn through a frustrating process of trial and error. Although part of the fascination of hacking is to discover some new (and usually unintended) way of using a computer system, the majority of new users like to be "shown". One of the major frustrations of computer use is when poor documentation makes a simple technique seem mysterious. Even for experienced users, a face-to-face human demonstration communicates technique more effectively than an instruction manual. Technique and Ritual
Rituals, group activities in which people act out symbolic meanings, are a very old form of transmitting cultural techniques. Although today we think of rituals as religious activity, in many non-industrial societies there was no clearcut separation between the sacred and the ordinary. For example, in one Southeast Asian society, agricultural rituals taught techniques of planting and harvesting as well as the sacred meaning of agriculture. Even when participants fail to grasp the meaning of what they do, the habit of ritual ensures that people will continue to behave in symbolically appropriate ways.

Some people learn to use computers in ritual ways. Without necessarily understanding what they are doing, they go through a sequence of steps to make a computer "magically" respond. Computer technology appears to them as one of the mysterious forces of the universe. Although it is still possible to transmit technique through ritual, there are more effective ways to learn to use a computer. Also, these private computer user rituals lack the social dimension of shared meaning that make public ritual a continuing element of human culture.

1.3.3 Design

Whether by accident, through casual play, or "on purpose" according to an imagined mental plan, humans invent tools and develop the techniques to use them. In the case of the programmable digital computer, the original design (by Charles Babbage in the 1830's) preceded practical implementation by many decades. Babbage's design was itself inspired by earlier tools for operating looms and for aiding mathematical computations. Today, computer scientists and engineers typically develop designs based upon mathematical principles and plan techniques well in advance of the actual construction of hardware or software. Creativity and Design
Industrial and artistic design can be considered two different processes (Jervis, 1984 ref), but both are part of a creative process that connects imagination to rational judgment. Whether for a computer program, for a building, or for a painting, a design is an imaginative vision connected to a practical implementation plan. Design involves more than rational calculation known facts. Creative people report that unconscious mental processes and play are important elements in the ability to reach past the boundaries of conventional knowledge and integrate formerly disparate information (Ghiselin, 1952 ref).

Popular explanations of creativity often equate it with the free expression of unconscious impulses, with mysticism, or even with insanity (Becker, 1978 ref). Misunderstandings of brain function lead some people to erroneously assume that creative people use the right half of their brain (the part that usually controls the left side of the body) while analytical people are "left-brained" (Calvin, 1983 ref). Instead, creative people seem to be able to use their whole brains effectively. Creative designers imagine ghestalts -- whole, complex patterns -- that can be translated into real- world materials and shared cultural symbols. The artist Michaelangelo wrote that he "saw" his sculptures in the stone and only had to take away the extra material around them. Karl Marx said that the difference between an architect's building and a bee's hive is this human ability to build in the imagination. Computer-Aided Design
Computer-aided design (CAD) can be described as a partnership between human and tool in which the computer performs the drudgery and the person provides the creativity. In programming, techniques like CAD are called "software tools". While a tool is usually thought of as a material object and a technique as an abstract method for using tools, programs to help us design, write, and test other programs are tools. If the "things" with which we work are plans, designs, concepts and data -- in other words, if we are doing mental work -- then some information products serve us as logical tools. We use them to create other information products like programs or data bases.

The effects of computer-aided design are controversial. Proponents argue that CAD frees designers from time-consuming drafting and calculating chores, enabling them to try out more imaginative designs. CAD critics question whether such programs really encourage human creativity or, like lego sets and coloring books, they limit the range of possible plans. The effects of CAD on software design are similar to the use of standard parts for craftsmen. If a cabinet maker's standard parts include screws and nails of certain sizes and boards of different thicknesses, it may have no negative impact on his or her ability to design furniture (and it saves the tedious labor of cutting trees and forming metal parts). If the cabinet maker's standard parts are preformed cabinet pieces that merely have to be assembled, however, little original design is possible.

1.3.4 Technology and Reification

By building tools and developing techniques we make our imagined designs real. But sometimes, as we embody our ideas in machines, we reify social relationships as well. In other words, we imagine that our relationships to one another are "in" the technology. An example of this is lie detection technology -- polygraphs or the newer computerized voice stress analyzers.

Lying is a very human phenomenon. We often present ourselves to others as nicer, smarter, more attractive, or more competent than we secretly feel. People deliberately distort information for their own advantage or to try to avoiding hurting others' feelings. Although we have strong norms against lying to gain power over others, we expect "white lies" in polite conversation. We often say: "I'm fine, thanks. How are you?" when we feel terrible.

Lie detectors, according to a review by the Office of Technology Assessment (Saxe, 1986 ref), do not detect lies. They detect the physiological changes that occur when we are emotionally stressed. There are many sources of such stress besides guilt or fear of being caught at lying. If someone lies without guilt or fear, the technology detects nothing. An honest answer to an embarrassing or disturbing question will show evidence of stress. Yet some people treat lie detectors as if they were a technology to revel the truth in others' minds without our having to go through the social interaction processes that establish trust in one another. Trust based on social interaction has been replaced by trust in technology.

At its best, the use of computer technology will give us new power to cooperate and realize common purposes. At its worst, the relationships between people and computers will be substituted for social ones. But before going on to examine in detail the effects of the human/computer interface, we must take a closer look at the process of social change and the question of why people began to use computers at all.

[ Main Contents | Chapter Contents | Next Chapter ]