Information Analysis and Decision Making

Even in the fully realized Metalibrary, universal accessibility of information does not in itself solve any practical problems. Finding solutions is a multistage process resembling the scientific method that leads from raw data first to knowledge and understanding and then to decisions. Moreover, discovering knowledge and making decisions are not necessarily on the same path, but may often be nearly independent of each other.

The full Metalibrary, like present-day paper libraries, provides material for the first step in the process, by organizing raw data by category and giving users tools to relate the data to other categories, to analyze it, and to record conclusions or argue with those drawn by someone else. All this is done now in scientific journals, though not very efficiently. However it takes place, a community or collective consideration of data is necessary before information can be derived from it. Indeed, it has been in the high density strongly interactive population concentrations that great new ideas took root and flowered in the past, and there is a sense in which the Metalibrary makes the entire population of the Earth into a single city.

When the subject matter can be described in quantitative terms, (chemistry, cell biology, economics, and demographics,) the first step is to establish what are the facts, that is, what data are valid and what information they convey. Expert forums operate through today's journals (and tomorrow's Metalibrary) to achieve consensus on what the facts are, given the data available.

However, managers and other administrators often need to make decisions long before there is a widespread consensus about what the pertinent information means. Ostensibly, such a manager makes decisions on the basis of available information about past history and probable consequences for the future. On the other hand, two people may easily make a different decision under the same circumstances. Here are a few examples:



1. The task is to implement a trial version of a universal data base. The problem is to decide on who should have access to what information. The company hires a lawyer and a computer scientist to advise. Their recommendations flatly contradict each other; one wants a completely open system, and the other desires tight controls.

2. A bank determines that it requires a new computerized billing system for its expanded safety-deposit department. Extensive studies are run, and software is chosen to control the data base. But many compatible machines can run this software, including brand I, brand C, and brand X. Systems from all three vendors are tested and the results charted. Brand C comes out on top in price and performance, with brand I second and X third. The branch manager then overrules the selection committee's recommendation and decides to buy brand I, because the mainframe presently in her office is from the same company and she values brand loyalty higher than price or performance.

3. The Fraser Valley Library acting on recommendations from the Ministry of Human Resources, has decided to build a branch to serve a slum neighborhood of Aldergrove. The only available property is an old park adjacent to a heritage building, formerly the residence of a certain well-known author. Psychologists, social workers, and government officials claim that dramatic improvements in similar slums have always resulted when a library was built. They insist that the house and park ought to be sacrificed. Historians and some local community leaders point to community pride for one of theirs who made good, as well as to the benefits of the park for their children. They do not deny the potential value of the library, but hold that the value of the house and park are greater, intangible though that value may be. The recommendations are again contradictory.

4. During World War II, the British scored a intelligence coup by breaking the German coding scheme and routinely translating military messages from the opposite side. One night, the decoded message contained instructions to bomb the city of Coventry at a particular time. The British prime minister, Winston Churchill, knew that many lives could be saved if he evacuated Coventry. He also knew that this move would reveal to the German High Command that their secrets had been breached, and the codes would immediately be changed. By the time the British could decipher the new ones, many more lives could be lost on the battlefront than could be saved at Coventry. A utilitarian, Churchill did not warn the city; the bombs came, and civilian lives were lost. Clearly, an act-oriented ethics would have dictated the opposite course of action.



The point of these four examples is that the mental filters through which both history and consequences are passed often have more influence on a decision than the facts and probabilities themselves. People do not make decisions on facts; they make them for other reasons. The decision in example two hinges not at all on the data--in fact collecting it turns out to be a wasted effort, for the manager makes the choice irrationally, basing it on emotional familiarity rather than on facts. Business people commonly do decide things emotionally, particularly when it comes to technology--this explains why inferior systems can become commonly used. Such scenarios are normal in any situation where the people making the decision are not personally familiar with the technology; they do not understand the data, or they trust advertising more, so they ignore fact and embrace emotion.

The human element is critical to the outcome of the decision-making process, and the world view (including the ethical view) of the decision maker may well determine the outcome quite apart from (or in contradiction to) the facts. Above the individual's world view, and creating its context, is that person's group culture. Depending on education, peer group, social status, local ideas, organizational outlook, and national goals or prejudices, each person shapes a world view in some degree of conformity with others sharing the same culture. Membership in a given subculture of society will determine whether a person even sees certain data, much less understands it enough to make informed decisions. Thus, even extending decision making to the entire populace would not guarantee that better decisions will be made, or that they will last long in the face of the fickleness of popular opinion.

The full Metalibrary could help with some of these problems, would exacerbate others, and would create some new ones. For instance, it could be used to enforce a requirement that some expertise be demonstrated before participating in a decision. In matters such as the building of libraries or parks, a simple test on the facts of the case could be required to gain voting status. Those living in the affected area would read a selection of the arguments for each course of action, and answer some simple questions to show that the issue is understood. The decision would then be made by the informed and affected people.

Decisions with wider effects and more profound consequences might require a different voting structure, in which the degree of knowledge about the problem would determine each person's share of the vote. If a dam is proposed on the Columbia River, economic benefits would have to be weighed against environmental effects. It might be too much to expect everyone affected to become sufficiently knowledgeable about the proposal and its effects to cast an informed vote--there is too much technical information for non-experts to digest. Moreover, those with the best engineering expertise are not necessarily those most knowledgeable about costs and benefits or about environmental effects. Perhaps a formula could be devised to weigh the votes of those with greater (or multiple) expertise more heavily than those who qualify with less knowledge but are still affected. This would give those with a strong interest a powerful motivation to do some research, and might make it more likely that a consensus on the decision could be reached. However, this particular issue is complicated by the fact that the river in question crosses the Canada U.S. border. How could the relative interests of two entire countries be weighed when one is larger in area and has more environment to affect, and the other has more people and a bigger economy?

The premise behind such electronic participatory democracy schemes is that everyone sufficiently informed would be more likely to come to the same conclusion. This would be a major (and in many cases unjustified) assumption; as has already been indicated, good information is not the only factor in decision making. Such systems would also be a substantial modification of current democratic practice; whether they would be found acceptable or not is another question--especially if they vary in any way the accepted democratic idea of "one person-one vote." It is also important to note that the mere technological enabling of weighted voting is not in itself a reason for implementing such a scheme. Moreover, such a scheme does away with the cherished idea of "one person one vote" and may not therefore be seen as an improvement.

Yet another common supposition is that the existence of comprehensive communications and information facilities such as the Metalibrary would tend to reduce or eliminate differences in culture and world view and thereby promote unanimity in decision making. This would continue a process begun by books, radio, and television and fostered by modern-day population mobility. However, the world of the 1990s was still far from the global village envisioned by some in the 1960s, even though its peoples share far more in common now than for thousands of years. Indeed, though there might gradually be fewer sharply distinct cultures based on geography, and fewer international boundaries as well, there are some very basic conflicts of world views that are unlikely ever to be eliminated. If the fall of the former Soviet Union has taught us anything it is that centuries-old ethnic hatreds such as the ones it brutally suppressed can still survive for generations and be readily be called upon to create new bloodbaths when that repression is removed. In such cases, the availability of more technology merely means that people are killed at a faster rate than before.

Moreover, along with its new kinds of information filters, the Metalibrary could well create new culture and world view conflicts, for not only will people still be able to perceive information differently, but they would also be able to personalize their view of the information to the extent that they will not even have to look at the same data.

An ivy-leaguer with great pride in her type of institution might accept information connection threads only from people at similar schools and choose not to see the threads attached by anyone from smaller or less prestigious universities. Prejudices over spiritual ideas would remain, with some religious people refusing to read certain scientific works and some scientists refusing to read certain religious works. The same applies to those of differing political persuasions. Except when a person placed a foot into the other camp or crossed over altogether, people on one side of any debate could pretend that the other side did not even exist, much as happens today.

When a new link did cause a thread to trail over the self-imposed borders, it would at once be obvious. Denial of recognition would remove the threat and the troubled mind could once again be safely closed. New ideas and the related data would not reach people unless allowed to do so. As is done today with existing information techniques, such denials perpetuate an already well-formed group thinking pattern and increase the possibility that decisions made by such people would be bad ones, because they are not fully informed. Once again we see that automating a bad process (here it is decision making) does not make it better. Rather, it merely produces the bad results faster.

On the other hand, as the pornography issue illustrates, not all information is either useful or beneficial, and it may be a good thing to prevent some of it from entering one's home, or, in such extreme cases, to prevent it from even being available. With a fully implemented Metalibrary, the former may be rather easy, but the latter likely very difficult.

One could suppose that a kind of natural selection (good decisions and advice are more efficient and useful than bad) would gradually reduce the influence of the close-minded as the poor quality of their decisions became evident. However, there is no guarantee that a particular discipline or specialty would not become as rigid and unbending as can now happen using the medium of journal articles. A control belief group has the power to reject new ideas by collectively refusing to look at them. Denial of recognition by the cultural leaders--who may still be termed reviewers or editors--would guarantee that new ideas would not be read. That is, for all its promise as an information utility, the Metalibrary might make it even harder to challenge the control beliefs of a society, for each sub-culture using it would still have unlimited ability to effect intolerance of competing views. A possible way around this difficulty would be to have the Metalibrary rules allow universal visibility to new information links regardless of who makes them, at least until such time as a person reads the item and expressly denies the link. Another possibility would be to create ombuds-reviewers who can make connections that every user will see for at least a certain period of time after the person first reads the new material.

There is no completely satisfactory solution to the problem of intellectual intolerance, however. Everyone filters what they will read and who they will talk to. They must, for there is too much for one person to assimilate. The filters in the Metalibrary will in some senses be more tangible, but they too are necessary. Although the narrowness of specialization may be greatly reduced because much less knowledge will need to be memorized--looking it up will be better--specialties will still remain, and their practitioners will still have difficulty communicating cross-culturally.

Once again, it becomes evident upon some thought that ideas, like goods and services, (whether cultural or academic) are accepted or rejected by society as a whole in the short run on their perceived merits, not on absolute standards. In their own generation, the guardians of the control ideas and beliefs can always refuse to acknowledge anything else, or even suppress competition. It is only in the longer (historical) run that they come to be evaluated with more global measuring sticks.

Neither will language barriers necessarily be broken down, for eventually the Metalibrary would communicate with users in their own languages. There would therefore be no incentive to learn another tongue, and meeting other people personally might even become more difficult. Spoken communication could suffer and isolationist tendencies increase, balancing off the improvements in written communication.

The full Metalibrary will also be sophisticated enough to allow the use of cultural, religious, or personal values to assist in filtering information and making decisions. Since it will record every data search and every decision, it could record how each person's filters operate and suggest solutions to problems consistent with one's stated values and past decisions. Again, though there would be benefits to this, there would be no incentive to re-examine one's presuppositions periodically, for the Metalibrary could be set to reinforce them.

In any event, the advent of the full Metalibrary would make it clear that everyone has a world view. Each person would construct a reflection of that world view in the process of learning the system, developing the filters, and making decisions at both the information and interpretation levels. Since there would still be "superstars" of each discipline even in this new medium, there would be a demand for the ability to adopt other people's world views (or sets of connecting threads.)

So, in addition to being able to modify one's own personal set of connecting threads to recognize any other person's links, it ought to be possible to rent another's. This is different from incorporating in one's own set the links with that person's name on them, for that does not also add the connections the second person has recognized from other people. Borrowing a whole world view would allow people literally to see things as others do. In this scenario, world views would be a commodity for rent or sale, and would be mergeable with one's own. A person could keep several independent world views on hand and switch between them or revert to an older version of a connection set.

Many scientists who are also writers have remarked that although they can travel in several academic cultures, they seem almost to become different people when they do so. This is an routine phenomenon also, for everyone has different mindsets and vocabularies (called "registers") to communicate with different people. (One vocabulary subset and thinking pattern for the children, another for clients, and other for co-workers, and so on.) The Metalibrary would allow someone to be (intellectually) as many different people as desired, though it is likely that most users would integrate their interests into a single collection. Some people will undoubtedly make their world views available as a public service for anyone to use, others might make a tidy profit selling theirs, much as they now do from books.

All of this would allow for decision making that is potentially more factually informed and that enables the participants to better consider each other's points of view and how these were formed. This does not mean that making decisions will be any easier than it is now or that most (or any) will be unanimous. It does imply broader participation and less bureaucracy, as well as the possibility of more satisfaction with the results. There would still be differences of opinion and there would be more opinions expressed than ever. All opinions could be considered, even though all surely would not be. To put this another way, being better informed may be good in the ethical sense of the word, but it is not clear to what extent that "good" will be sought after. It is even less clear how well it could be enforced.

Yet another problem with an information-based society is the potential to rely too much on machines for the decision making process. When this is done, it is easy to forget that information is more than whatever is stored in or processed by computers. To have meaning, it must be communicable. Assigning and communicating meaning, judging value, and taking action based on informed decisions are all part of the unique province of human activity, and there is as yet no indication that any of these can be automated. It is easy to rely on the neat rows and columns of figures in a spreadsheet, but unless the assumptions behind the formulas used to produce the output are known, the reader cannot make informed human judgement on the information content. There are value judgements behind the process of data collection in the first place; there are value judgements involved in organizing it; and there are value judgements involved in deciding on what meaning to assign to (and what action to take upon) the material in the end. Thus, who decides, and out of what value system, turns out to be what gives information its ultimate quality and meaning. Humans can think about and evaluate their thinking process; machines cannot. This appears to provide an answer for the (ethical) question: who ought to decide for humans--themselves, or machines?

The availability of instant information also creates pressure to make instant decisions. For instance, because it is easy to do so, and the means are at hand, many people respond to electronic mail messages and Internet news postings as soon as they receive them. As users of such systems are well aware, this results in a large volume of intemperate, ill-considered, and impolite mail traffic and news. (Such messages are called "flames".) Likewise, if thought processes and analytical techniques are unsound or if decision makers are so culturally conditioned as to be incapable of considering alternatives, the Metalibrary facility will not help. Computerizing a bad decision-making process does not produce good decisions, it only causes the bad ones to be arrived at in milliseconds instead of days.

As mentioned, prejudice will also remain. That is, irrational dislike of others and refusal to consider things from another person's world view will be as likely then as now. Perhaps the greatest contribution to decision making of instant and universal information availability could be the recognition of legitimate differences among world views as people realize (in the process of automation) how they have been making their decisions. Perceptual and decision-making filters would be obvious instead of hidden; their existence could no longer be denied or ignored. This has the potential to blur the boundaries between sub-cultures, promote communication, broaden specialties, make learning easier, and promote the possibility of sounder decisions. On the other hand, prejudice has stood the test of time as a stronger force in human affairs than any of these potential benefits.

Thus, as for all technologies, the impact of electronic media on knowledge and decision making will be mixed. Great benefits will be available, great abuses will be possible, and for many people there will just be a transfer of their old ways of thinking to a new medium.