HOME
PREVIOUS PAGE
NEXT PAGE

IV THE RELATIONSHIP BETWEEN LEGAL DECISIONS AND INFORMATION SYSTEMS

[Page 223]

12 The "rule of law"

12.1.1 The "rule of law" as a justification for improvement in legal information systems

Computerbased legal information systems represent investments in time and money. When justifying the change in technology, one might argue that the new information system is more efficient - making legal research less time-consuming and consequently cheaper in the long run. This is obviously not the main motivation behind the creation of better information systems. We have several times pointed out that the new technology represents a basic change influencing the research habits of lawyers. Comparing the state of research before and after the introduction of the new technology, one will find a difference in quality. A different type of research is conducted; the lawyers do not confine themselves to doing what they were doing before the change; they do more or something else. These changes in the quality of legal research may not adequately be translated into quantitative terms (time or money). And it is precisely these changes in quality that are very often pointed out as the chief justification for introducing the new technology.

In our historical survey we have given several examples of this sort of argument. In the United States, a major concern was the failure of conventional information systems to cope with the information growth (cfr. Lawlor 1962:30-0-302). In regard to the RIRA and the JURIS systems (above at sections 4.4.3 and 4.4.4), the principles of the rule of law were emphazised: the new information systems were to act as guarantees for equal decisions in equal cases, coordination of the stands taken by the government, etc. Actually these systems represented measures taken by the public administration in order to reconfirm that the "rule of law" was their major concern, and that the information growth had not undermined the system. A similar tendency may be traced in France, where just the lack of publication of a major type of legal sources - the

[Page 225 ]

decisions of the appeal courts - was stressed when introducing the documentation system of IRETIJ and other centers, cfr. above at section 5.3.3.

But this line of reasoning has become most apparant in Western Germany, where the slogan "Informationskriese des Rechts" was launched by Professor Simitis (1970). The plan of a total, national, legal information system (JURIS) has in many ways been introduced as a therapy to an ailing conventional system. Critics maintained that the status of Western Germany as "Rechtsstaat" was threatened by the information crisis. The JURIS system should once more bring the capabilities of the information system up to the standards demanded by the volume of information, thus reestablishing the "old order".

These examples must suffice to demonstrate that the label "rule of law" has served as a major justification for introducing legal information systems. "Rule of law" is, however, a very vague term. It characterizes an ideal common to most legal systems. But in order to relate "rule of law" arguments to the legal information systems, it is necessary to break it down into less abstract components - which we shall do in the next section.

Before proceeding to this process of specification, we ought, however, to point out the relationship between social change and legal information systems. The "rule of law" is a political or ideological value. In relating changes in the information system to consequences for the "rule of law", we look into one of the aspects of the relation between social change and legal information. This is not, however, the only aspect. The legal information system is a vital part of the infrastructure of a society, legislation being a major technology for governing the behavior of the citizens. Consequently, changes in this infrastructure will have political effects that are not visible from the "rule of law" point of view.

Our interest in the "rule of law" has several reasons. The "rule of law" has traditionally - at least in Norway - been discussed in legal theory, the strengthening or impairing of the "rule of law" being a central issue in administrative reforms - this in contrast with other social consequences of changes in the legal system. Also, the "rule of law" is closely associated with the model of the legal decision process presented in the first chapter. It may be claimed that the norms governing the legal decision process should be designed to optimalize the "rule of law". The "quality" of a legal decision is - according to this view - measured in how well the metanorms fulfil the requirements of the "rule of law". And the legal

[Page 226 ]

information retrieval process is part of the decision process and may not be assessed apart from this. This intimate relationship between information retrieval, legal decisions, and the "rule of law" justifies our attempts to trace some of the major factors linking them together.

This close relationship also explains why there will be some repetitions of arguments made earlier in respect of the decision model. To some degree we will cover the same ground, though in the first chapter the terrain was mapped from the viewpoint of the legal decision, while in this chapter we will do the mapping from the viewpoint of the "rule of law".

The "rule of law" is a term less satisfactory than the German "Rechtssicherheit" or the Norwegian "rettssikkerhet". Our use of the term "rule of law" may include elements more commonly associated with the term "due process". The "rule of law" is, however, only employed as a common label on a set of more precisely characterized elements; we will therefore not dwell on this terminological problem.

12.1.2. Elements of the "rule of law"

The "rule of law" is, as we stated above, too abstract a concept to be handled with ease. It is necessary to split this concept into various elements and relate these elements to the legal information system.

As our starting point, we take the work of Eckhoff (1966), in which the "rule of law" is split into 5 elements:

• predictability
• fair and just process
• objectivity in application of norms and discretion
• the principle of equality
• democratic control

The context in which this specification was made (efficiency and the "rule of law" in public administration) focused attention on special problems. For our purpose, we have decided to concentrate on two of the elements, predictability and the principle of equality. In addition we will discuss some possible effects on legal meta-norms of changes in the legal information system, and we will follow up the discussion of "time versus quality" initiated above in section 3.

Certainly this specification of the "rule of law" is not exhaustive. On the other hand, our two main elements - predictability and equality - will be broken down further and probably include elements which otherwise and in different contexts would be considered as major elements of their own.

[Page 227 ]

12.1.3. The legal information system

The "rule of law" point of view directs our attention to the social consequences of legal information systems, a macro level in contrast to the micro level to which our perspective has so far been restricted. This makes it necessary also to outline a macro picture of the legal information system.

A legal information system may be considered to consist of three main components.

(1) The producers of legal sources.

Within a legal system, there exist qualification norms defining what are to count as legal sources. Such a "definition" usually designates some person or body performing a certain role as a "producer" of legal sources, and specifies a certain "production procedure" which has to be followed if the product is to count as a valid legal source. In this way we recognize statutes, case law, etc. to be valid legal sources within the terms of our legal system. By adding up all legal sources produced since the legal system came into existence, we find the total volume of legal sources of that system.

Some of these sources may date from before the creation of the legal system, as parts of preceding systems may be reaffirmed when the new system is established. In Norway the Constitution of 1814 did not invalidate earlier Danish statute law, but instituted a new procedure of replacing it through the Norwegian parliament.

(2) Distribution routines.

This corresponds to what in relation to the theory of information systems is termed "channels". The distribution routines may be of a varying nature; they may for instance be selective with respect to material included. Distribution routines have two major functions.

(a) The alerting function. When a change takes place in the legal system, the distribution routine should alert all concerned. Which group this will be, obviously depends on the nature of the changed norms. Often the whole population is the target group, and the distribution routines have then to be designed to bring information to all citizens - i.e. through newspaper advertisements, radio, or television.

(b) The updating function. With the recipients, the routines will accumulate in the form of a local collection of documents, a subset of the total volume of legal sources. When a change in the legal system takes place, this accumulation should be updated. When the user then consults this collection in order to apply the current law, his collection will be up to

[Page 228 ]

date and include the change. As a further supplement, some distribution routines include a "deleting function", offering the user the means of deleting outdated information. This may be done through a loose-leaf system in which new pages replace old, or by the periodical replacement of a bound compilation of, for instance, statutes in force. In computerbased systems the updating of the data base will usually also involve deletion of outdated material.

These two functions may certainly be combined in one distribution service: a lawyer receives a law gazette, which alerts him to new legislation and at the same time updates his compilation of statutes in force. But in some instances - such as a newspaper notice announcing a new statute - the separation of the two functions are evident: the notice may alert a person to a relevant change, but he would be badly served by a scrapbook of such notices when taking a stand on a legal issue.

(3) Retrieval systems.

The volume of legal sources is too big for anybody to handle without searching aids. These are what we in this contex call "retrieval systems". They may be directed toward the user's local collection of legal sources: for instance the annual register in a case reporter or the index cards the user himself has made up for his own personal needs. Or they may be directed toward a bigger subset of the total volume of legal sources: for instance all Supreme Court cases, a legal bibliography for a certain area of law, etc.

It is the information system - composed of producers, distribution routines, and retrieval systems - which we will discuss with respect to the "rule of law". We cannot, however, take into consideration the total system. We will focus on a few elements of this information system: the retrieval systems and the updating function of the distribution routines, especially on the accumulation of local document collections.

12.2 PREDICTABILITY

12.2.1 Introduction

If a legal problem is presented to a lawyer, he may predict the outcome for his client if the problem is brought before the court. If the rate of predictability is high, the client may choose to negotiate and find a solution that at least is cheaper measured in time and effort.

Also, when a private person considers his choice of alternatives he will take the legal norms into account. If the predictability is high, the person

[Page 229 ]

will know which of his actions will bring him into conflict with the norms, and the outcome of such a conflict.

Predictability in a legal system is a prerequisite for enabling this system to fulfil its basic functions of solving conflicts and controlling behavior. Reduction of predictability will foster conflicts and result in appeals, and will make legislation a less effective means of manipulating the behavior of the citizens.

Predictability may be reduced by many factors. One such factor is the nature of the norms - if they leave room for discretion, they also leave room for alternative solutions to the same conflict, and, in turn, they reduce the predictability of the system. Competing legal ideologies may be another cause of reduced predictability: two opposing schools of thought maintain for instance different theories as to the interpretation of statutes, one interpreting them "in the spirit of the legislator", one in a strictly philological manner. This lack of unity in accepted legal method may cause reduced predictability where different methods may justify different results.

We are here, however, concerned with the relationship between predictability and the legal information system. Predictability presupposes knowledge of the legal norms, which can only be a result of interpretation of legal sources. Whether these legal sources are available or accessible therefore becomes a central issue in regard to the predictability.

12.2.2 The alerting function

Above, in section 12.1.3, we have outlined the legal information system and pointed to the "alerting function" of the distribution routines. Through the legal information system people should be made aware of their duties and rights. They should acquire a background knowledge of the legal norms relevant for their actions. Only with such a background knowledge will they be able to recognize that there may be legal aspects of a problem, and acting on this insight, seek professional aid. This turns our attention to the alerting function of the legal information system, which in this way becomes basis to the predictability.

If the alerting function is not satisfactory, then the background knowledge of the citizen will be the result of casual reading or personal resources. People with greater resources (education, money, etc.) will more easily become aware of the possible existence of relevant legal norms, and will consequently look into the matter. They will also usually have a greater general background knowledge of society (and also of the legal system as part of the functioning of society). This flaw in the legal distribution routines will manifest itself as a

[Page 230 ]

We will not pursue this argument here. As stated above in section 12.1.3, we shall not discuss the alerting function of the distribution routines, but concentrate on other parts of the information systems more closely related to the discussion in previous chapters. We feel it is necessary, however, to stress that the alerting function is the basic function. As the solving of conflicts in itself produces new legal sources (case law, administrative decisions - even new legislation), the whole structure of the legal information available may become assymetrical because of a lack of concern for the alerting function. This is even more important, since this function - at least in Norway - has not received as much attention as it deserves; the emphasis in the information system has been on the interface with the professional users. The discussion of legal aid, free lawyers, etc. has brought to public attention the lack of insight into the legal system on the part of low-status citizens. One of the causes is, we feel, unsatisfactory routines for the alerting function, and one of the remedies could be - along with a free legal advice service - a reformation of this part of the legal information system.

12.2.3 Coverage discrimination

As stated a number of times, predictability depends upon the insight into norms as created through the interpretation of legal sources. The volume of legal sources consulted by a lawyer will restrict the possibility of his gaining insight into the system of legal norms. The quality of the services rendered by the lawyer will to some extent be relative to the extensiveness of his research. This will, of course, incorporate subjective elements. The client who follows the advice of a careless lawyer may be in for some surprises. But this lack of predictability, depending on the personal qualifications and dedication of the individual lawyer, is not our concern. We will focus our attention on the availability factors in the legal information system. These may in fact create conditions in which even the most careful and dedicated lawyer is not able to gain the necessary insight into the system of norms - because he does not have access to the relevant legal sources.

As we have discussed above in section 1.2.7, the accessible data base of a lawyer is defined by availability factors. These factors are themselves

[Page 231 ]

part of the pragmatic situation of the lawyer: his time-pressure, his distance from a library, etc. Consequently no two lawyers will be in an identical situation, and the accessible data base will in fact vary from one lawyer to another. This relativity is not due to the qualifications or dedication of the lawyer, but to a complex pragmatic situation.

The distribution routines of the information system are, however, one such factor which the lawyer himself is not always master of. Let us take as an example the decisions of a particular court which constitute a type of legal source of some importance to our lawyer. These decisions are pronounced and recorded by the court. But if there is no distribution routine to which our lawyer may subscribe, how then is he to use this material?

The lack of appropriate distribution routines of this kind is probably rather common, not as the result of ill will, but as the result of a lack of attention to the problems created.

An example was uncovered in our analysis of the legal information system within the Norwegian social security administration (Bing/Harvold 1973: 250-261, Bing 1974: 111-112). This is a typically hierarchical administration with a central agency - the National Insurance Institution - in addition to regional and local agencies. In respect of disability pension, the applicant's case is prepared by the local agency and submitted to the regional agency where the decision is made. The applicant may appeal his case through the NII to the Social Security Court. The decisions of this court should then be adhered to by the regional and local agencies. They represent a type of legal source of quite high rank within the social security administration.

The Social Security Court has created its own precedents file, which the lawyers of the Court consult when preparing a case. The precedents file is just a subset of the total number of cases decided, including only cases of principal interest. A number of the cases are also published and distributed among the civil servants within the social security administration. At the time of our survey, however, the number of published cases totalled only about 20 per cent of the numer of cases included in the Court's precedents file. Consequently the accessible data base for the lawyers of the Court and the civil servants in a regional agency was very different.

The effect of this difference in coverage is obvious. Let us imagine that a case in which a precedent of the Court is relevant is brought before the regional agency. It would then probably be decided without knowledge of the relevant precedent. If appealed, the case would come before the Court, the relevant precedents would be retrieved as a matter of routine, and the decision of the regional agency would be revised according to the new information.

This sort of situation is caused by coverage discrimination between the Court and the regional agency. The Court's revised decision will possibly reduce the client's confidence in the regional agency and propagate appeals, as well as inducing frustration in the regional agency itself.

[Page 232 ]

Coverage discrimination reduces the possibility a user group has of gaining adequate knowledge of the norms, since part of the relevant material is absent in their data base. It would generate frustration and reduce predictability.

Actually this coverage discrimination may be seen as a parallel to another tendency: that of intensive use of legal sources (Bing/Harvold 1973: 222-223, Bing 1974: 110). When the number of sources becomes small, the user still has to find solutions to his problems. In our survey of the Norwegian social security administration, we traced a tendency toward intensive use of legal sources in the agencies with low coverage. When in doubt, the user would reread his source. This is in contrast with the agencies higher up in the hierarchy and with better coverage: when in doubt, the user would look into additional sources.

This tendency of intensive use of legal sources where the coverage is low, and extensive use where it is high, may be two sides of the same coin. It may, however, also have some connection with the positions of the agencies in the administrative hierarchy, and with the professional background and education of the users in the different agencies.

It ought, however, to be pointed out that extensive use of legal sources is the ideal. A legal problem should be solved in the light of as many sources as necessary. Because of the information system, extensive use of legal sources is very difficult or impossible at the lowest levels owing to the simple fact that the sources are not available in their data bases.

In our example there exists a publication routine, although inadequate. A distribution routine is not always believed to be necessary by the administration itself, especially where all decisions within a field of administrative law are made by the agency itself. The agency might create a precedents file - like the Social Security Court - but would feel no need to create a publication routine. In this instance the coverage discrimination would be severely felt by all other users penetrating that area of law, as the agency has in fact a monopoly of the important type of legal source constituted by its own decisions.

12.2.4 Transparency

The term "transparency" or "Transparenz" (Simitis 1970: 29) has been used to characterize the situation where users have become the victims of an information crisis; where faults in the information system make it difficult to ascertain what is actually the law. We shall use the term to characterize a special problem of coverage discrimination, the difference in the information situation between a user outside and one inside the public administration.

[Page 233 ]

Today, the public administration decides a great many legal cases. Special distribution routines have been developed in order to make the relevant legal sources available to the civil servants within the administration. This may be in the form of compilations of annotated statutes and regulatory law, examples, precedents, etc. Also special retrieval systems are available, such as the index to a precedents file.

As we have discussed in the last section, coverage discrimination may reduce the predictability also within such a hierarchical organization. In this section we want, however, to emphasize the difference in the information situation between a user within this organization and a private lawyer outside the organization.

(1) Pragmatic factors. The private lawyer will usually not be at the receiving end of the distribution routines. The reasons for this may be several, let us point to three:

(a) To subscribe to a distribution routine would presuppose a certain frequency of cases to which the distributed material would be relevant. The type of trivial questions usually settled by public administration are atypical for the private lawyer today.

(b) The material distributed is meant for civil servants who often have no formal legal education. A lawyer may find its attention to detail, its casuistic nature, and pedagogical approach almost annoying.

(c) Just the detailed nature of the distributed material makes it voluminous. For the reasons mentioned above it will seem out of proportion (and unnecessarily expensive) to accumulate and update such a volume of detailed material.

What the lawyer probably would have found useful, is a textbook or a commentary on the same questions. At least in Norway this is recognized as a deficiency, and the Ministry of Justice is considering taking the initiative in having commentaries to statutes produced, cfr. Forundersøkelse 1974:36.

Actually the establishment of specially designed distribution routines within the administration may be a symptom of the non-existence of satisfactory, more general distribution routines. The difference between the tax authority and the social security administration in Norway may be taken as an example. Tax law is served by several distribution routines created and maintained by institutions outside the tax administration. The tax administration makes use of these routines for solving the information problems of their civil servants. In this case users both within and outside the administration are the recipients of identical distribution routines. Social security law is, however, not served by such independent routines, and the administration has had to develop its own system, which is mainly designed to meet the special needs of users in local and regional agencies, cfr. Bing/Harvold 1973: 274. One may reflect on the basic causes of the different situations in these two areas

[Page 234 ]

When a client seeks legal advice from his lawyer, he will often learn that the lawyer lacks background knowledge in the sector of law covered by the public administration, and that he does not have the means of retrieving relevant sources - for the reasons indicated above. This is not, however, merely a fault in the information system. The distribution routines are in existence, and the lawyer may subscribe to them. In these cases one may say that pragmatic factors related to the information system cause lack of transparency, but the information system itself is hardly to blame.

One may, however, point out pragmatic factors causing a lack of transparency which can hardly be blamed on the user outside the administration. The indexes to precedents files may serve as an example. Such a retrieval system is usually only available in physical proximity to the file itself, often integrated into the file as indexing cards. To use such an index, the user has to overcome the trivial availability factor connected with the distance between his office and the file. The civil servants of the public administration will generally be at the same location and close to the file. User research indicates that distance, though trivial, is an availability factor which it takes quite an effort to overcome. This effort will always be more onerous for users outside the administration. In many cases, for instance with respect to out-of-town lawyers, this will practically exclude them from the services of the retrieval system (the index). Obviously a terminal network would represent a solution to problems of this sort, but they could also be solved by more conventional methods - for instance by publication of the index. In respect of precedents, however, nomative factors also tend to reduce the transparency.

(2) Normative factors. In some instances the lawyer is denied the possibility of subscribing to the distribution routines. This may be for practical reasons (the distribution center is not geared to handling paying subscribers). The reason given for the few examples in the Norwegian legal information system where a subscription routine excludes private subscribers is, however, a normative one. One maintains that the distribution routine is an "internal matter" available only to the persons who can demonstrate a close relationship to the branch of public administration in question.

[Page 235 ]

Cfr. Hallan 1973:33, 37 and 39 - the distribution routines in question relate to prison and police administration, and to defence. Cfr. also de Capua/Feydt/Mathiesen 1973, who use the term "secret legislation" in respect of the regulatory law relating to prison administration.

Another reason for reduced transparency is the nature of the legal sources within public administration. Precedents often play an important role - as demonstrated in the example from the Norwegian social security administration above in section 12.2.3. These precedents are per se cases containing personal information. The social security administration cannot disclose these precedents without breaking the secrecy imposed by law. Consequently the norms of secrecy (privacy) act as an availability factor that reduces transparency.

Although incorporating the principle of public access to governmental files, Norwegian administrative law includes other norms which prevent access to precedents files, indexes, and certain journals. These norms are usually justified by privacy or secrecy considerations, and the effects on the legal information system have not - it would seem - been considered. Such norms may no doubt be found in most legal systems.

In conclusion, we have in this section concentrated on the systematic coverage discrimination against users outside public administration. The lack of transparency does reduce the predictability of the legal system, as it will be practically impossible for a citizen - even through the assistance of a private lawyer - to explore fully his own legal situation. The lack of possibly relevant legal sources will always inject a dosage of uncertainty into the situation.

We are of course not maintaining that the lack of transparency is the major characteristic of administrative law. But to us it does seem reasonable that one should pay considerable attention to ways of increasing the transparency. A lack of transparency will inevitably breed mistrust in the administration, and in fact make the control of behavior - which is central to many tasks of public administration - less efficient.

12.2.5 Control of representativity

Lack of transparency is often due to coverage discrimination. Related to these factors is the possibility of controlling whether a legal source is representative of the totality of legal sources.

An example may clarify the problem. In a case brought before the court, one of the parties cites an earlier decision by a court of the same instance as the present. In Norway, decisions at first instance are

[Page 236 ]

considered a legal source, though of a rather low rank. So the judge, when presented with this decision, will have to take it into consideration. The publication of first instance decisions is, however, arbitrary (Forundersøkelse 1974:19) - and the judge has no means of controlling whether the decision before him is representative of the majority of decisions relevant to the case at hand.

This dilemma may be solved by maintaining that because of the lack of control, first instance decisions should not count as a legal source (cfr. Fleischer 1965:152). Interviews with judges have, however, revealed that they are inclined to take the decision cited into account, but feel uncomfortable when their background knowledge appears to be insufficient for forming an opinion as to the representativity of the case, cfr. Forundersøkelse 1974:20.

It is easy to see how this may reduce the predictability of the legal system. An experienced lawyer may have a rather large file of first instance cases to choose from, and when going before the court he will certainly not present the cases that are not in favor of his client. Consequently he may make use of deficiencies in the information system to load the scales in favor of his client.

This is not restricted to the conflicts arising between private parties. As discussed above in section 12.2.4, the public administration often has information systems tailormade to their needs. In a tax case, for instance, the lawyer representing the administration will have access through specialized retrieval systems to an extensive precedents file, from which he can retrieve precedents in favor of the stand taken by the administration on the issue at hand. The judge will have to rely on counsel's willingness also to produce precedents in disfavor of his argumentation. Or rather, he will have to rely on the other party to produce this material. Actually, this may function satisfactorily in relation to tax law, as the legal system in this area is comparatively transparent, the expertise of the administration being countered by corresponding expertise on the part of private lawyers. (Cfr. Bing 1974b).

Lack of possibilities of controlling the representativity is mainly a lack of satisfactory retrieval systems. Even though the local accumulation of legal sources has a small coverage, the user should - once he has been presented with a relevant legal source - be able to control whether similar legal sources are in existence. Going back to our outline of the legal information system, we see that this is a demand for retrieval systems acting on the total volume of produced legal sources - or, on the total output by one producer (i.e. first instance courts). Usually retrieval

[Page 237 ]

systems are part of a distribution system, taking the form of indexes to publications. Where the distribution routines are unsatisfactory, there will usually also be a lack of a possibility of controlling the representativity.

We related this question to the position of the judge, and think that this is an important link. The principle of contradiction is often thought of as an element of the "rule of law". This principle presupposes a balanced information situation, wherein one party may challenge the representativity of the arguments presented by the other party. This possibility is reduced by deficiencies in the information system, like lack of transparency or coverage discrimination.

12.2.6 Conclusion

In this section we have discussed the relationship between predictability and legal information systems. It has not been a comprehensive discussion, but rather a highlighting of some central issues.

It will have become apparent that the core of the matter is the uneven distribution of knowledge among lawyers. Some of the causes of this inequality are beyond our control, like the competence of the lawyer. Some causes are of a pragmatic nature, but do not in principle reduce the opportunity all lawyers have to gain equal access to the information (for instance the reluctance of lawyers to subscribe to certain distribution routines).

We are left with two major causes. Firstly, legal norms that constrict access to certain legal sources to certain user groups - typically the access to administrative precedents. The double nature of these decisions as personal information and legal sources may not have been thoroughly considered, and the possibility of making them accessible in an anonymous form or in special situations ought to be deliberated. Secondly, we have the instances of badly designed distribution routines or the nonexistence of such routines, lack of adequate retrieval systems, etc. This is closely related to the legal "information crisis" - a term which probably would be an exaggerated characterization of the Norwegian situation today. However, through our discussion we hope to have demonstrated that a malfunctioning of the information system will reduce the predictability of the legal system and rock the foundations of the "rule of law".

[Page 238 ]

12.3 THE PRINCIPLE OF EQUALITY

12.3.1 Introduction

The principle of equality is related to the predictability of a legal system, but is not identical with it. It is possible to have equality (in a certain sense of the word) without predictability: the criteria by which two cases may be considered equal are not known to the public. (Cfr. the situation in Roman Law before the slave Gresus Slavia stole the formulas of the Pontifical College.) And there may be predictability without equality where the criteria for deeming two cases to be equal are irrelevant (all red-haired persons are found guilty). Usually the principle of equality has a counterpart in the principle that different cases should be decided differently. Whether this is a logical consequence, may be doubtful. But a realistic picture of the principle of equality should include the balancing of these two considerations: a new case should be decided in the same way as older, similar cases and all cases are to a certain extent unique.

The principle of equality presupposes that there is at least one older case. In a situation where no earlier decisions have been made - for instance in relation to the first case decided according to a new statute - the principle of equality has nothing to act upon. In this we see a difference between the principle of equality and other legal meta-norms governing the rank of legal sources. Other legal sources are given weight in accordance with a certain scheme in order to establish predictability in the system. The principle of equality is a further demand, probably rooted in considerations of non-discrimination or "justice". It may have something close to a double nature: on one hand, it is a legal meta-norm demanding that the lawyer should decide a new case in harmony with earlier cases, consequently qualifying these earlier cases as important legal sources. On the other hand, it is a principle of justice, demanding equal distribution of rights and obligations.

Owing to this second nature of the principle of equality, it will be more important with regard to public administration than, for instance, courts of law. The public administration decides a great number of cases as a matter of routine, and the "justice" of these decisions is to a large extent measured by how well the principle of equality is enforced. But with respect to a decision by a court of law, the measure will not primarily be a comparison with earlier, similar cases - but how well the decision corresponds to statutes, regulatory law, and other legal sources of higher rank.

[Page 239 ]

In Norway an exception may be found in the decisions of the Supreme Court, which to some extent is bound by its own decisions, and where special rules of procedure are provided for the occasion when there is a possibility of a new sentence deviating from the ratio decidendi of an old one. This is not, however, founded on the principle of equality, but rather on the fact that the decisions of the Supreme Court are a legal source of high rank, and that such deviations may reduce the predictability of the legal system. In the Anglo-American legal system, governed by the stare decisis principle, the distinction between predictability and the principle of equality may be more complex, but it is probably similar to that outlined in respect of Norwegian Supreme Court decisions.

Our arguments are certainly relative to our national legal system, but there are indications of the same reasoning in the United States, cfr. the justification given for the JURIS and RIRA systems above in sections 4.4.3 and 4.4.4. Legal information systems serving a public administration must be designed to strengthen the principle of equality; consequently these systems must meet standards of a different nature to those applying to "private" legal information systems.

We shall not try to describe in detail what is meant by two cases being "identical", but may characterize this by turning back to our model of the legal decision process. In this model, we found that the lawyer selected a number of facts as relevant. We think that two cases may be said to be equal when all the relevant facts are identical. Such identity will rarely occur, and one is left with the problem of assessing the similarity between an old and a new case.

The Anglo-American theory on interpretation of case law has, of course, explored the difficulties inherent in the problem posed above. The possibility of qualifying a new case will, for instance, nearly always open the possibility of claiming that the new case is not similar to the old one.

12.3.2 Representative coverage

The principle of equality acts upon a single type of legal sources, i.e. the precedents of the public agency in question. The total volume of such precedents may be readily defined as the total production of this agency.

In order to ensure full equality, the total volume of precedents should be consulted before deciding a new case. This is unrealistic for several reasons.

Firstly, the total number of precedents may not be as easy to define as it may appear. The agency may be a successor to other institutions

[Page 240 ]

deciding similar cases: these decisions may be relevant as legal sources, but not relevant with respect to the principle of equality. Also, the normative background may have been modified by amendments of statutes, new regulatory law, etc.

Secondly, the sheer volume of precedents may make it impractical to handle. Actually, this may be seen as one aspect of the legal "information crisis" - the number of decisions made by public administration is too great for the principle of equality to survive.

Thirdly, "one agency" may be represented by a rather complex organization, including central, regional, and local offices. To ensure full equality would imply that all offices could easily get access to all precedents in other offices. This sort of communication is hardly possible in Norway today. New technology - for instance an effective data transmission network - might make it more of a practical possibility.

Pragmatic factors will therefore exclude the alternative of obtaining total coverage with respect to precedents. The solution usually opted for by public administration, is the creation of precedents files. This precedents file is designed to include a representative selection of the total volume of precedents. The selection criteria are, however, usually vague and often not explicit. Decisions of "general interest" or deciding points of a "principal nature" are to be included, cfr. Hallan 1973: 71-72. The quality of the selection is probably relative to the person responsible for the maintenance of the precedents file. Whether this file really is representative of the total volume of precedents is probably often a matter of dispute. Cfr. also below in section 12.5.2.

The consequences of the precedents file not being representative is, of course, that the principle of equality is not ensured.

Certain precedents "similar" to the case at hand cannot be retrieved, as they are not part of the accessible data base. Our own user research has led us to believe that the problems of "editing" precedents files are often underestimated.

Though precedents files are the typical example of information systems created to ensure the principle of equality, they are by no means the only example. For instance, important decisions may be synthesized into regulatory law, explanatory notes, or summarized in textbooks. The users often find this form of preprocessed information more convenient to use, cfr. Bing/Harvold 1973:280. It does, however, insert a second element between the user and the precedents. Not only must the selection of precedents made available in this way be representative, but the summaries must represent adequate interpretations of the original documents. We will return to these problems below in section 12.5.2-3.

[Page 241 ]

12.3.3 The retrieval system

In our characterization of the principle of equality (above in section 12.3.1), we maintained that two cases were similar to the degree to which the relevant facts of the cases were identical. In order to retrieve an earlier and similar precedent, the user has to look for a case in which the facts are more or less identical with those in the one at hand. When formulating his query, he has to use the facts of the case at hand as criteria. This presupposes that the retrieval system available permits queries formulated in this way.

Precedents files constitute the typical information system available to the user in this situation. Such files have a rather simple organization. Usually precedents are filed under the sections of the statute they are related to (cfr. the Uniform Issue List described for the RIRA system above at section 4.4.3). A supplementary index constructed of terms characterizing the facts of the filed case is usually lacking, or it is composed in an arbitrary and less satisfactory manner.

The result is that the user cannot construct a query of the facts of the case at hand, but must formulate a query corresponding to the section in the statute which he thinks is related to the case. This query he can then use to look up possible precedents (cfr. Bing 1974: 113-114).

From our point of view it is important to emphasize that in relating the case to a section of the statute, the user has already taken an important step toward his final decision. And if he selects an incorrect section, he will not be able to retrieve a precedent, although the facts in the earlier case are nearly identical with those of the one at hand.

In many instances, this lack of flexibility in the query language has small consequences; there may be small problems in relating cases to the statute, or the users may have great insight into the problems they are dealing with. It is, however, interesting to note that the lack of a sufficiently flexible query language is in itself a flaw impairing the principle of equality.

One of the reasons for this lack of adequate retrieval systems is, of course, the great problems related to designing and maintaining an indexing system. The simpler methods offered by computer-based text retrieval systems will probably represent a revolution in relation to precedents files. It is no coincidence that there are so many examples of administrative bodies computerizing their information systems. And the principle of equality probably does play an important part as a justification for these innovations.

[Page 242 ]

12.3.4 Legal unity

The principle of equality implies that a case should be decided in the same way regardless of the residence of the person in question.

This aspect of the principle of equality is related to the principle that "all should be equal before the law". This last principle, however, is mainly a standard for legal norms, requiring that these should not give certain groups of persons privileges or be of a discriminatory nature. Actually there may be good reasons for taking the whereabouts of a person's residence into consideration. An example may once more be found in Norwegian social security law. When a person's disablement is being assessed, this is related to the person's possibilities of acquiring a job. In Norway the labor market varies from one region to another, a factor that will be reflected in the assessment of the person's disablement.

This may be rephrased as a principle of legal unity - within a legal system there should not (with certain modifications) be local norms deviating from the norms adopted through the legislative process.

The modifications indicate, of course, that local customs, trade customs, etc. are qualified as a legal source by the meta-norms of most legal systems, and that when these sources are found relevant, norms may be found which are valid for part of the realm only. These modifications are usually restricted to certain areas of law, especially the law of contract and the law of property.

Actually legal unity is a more problematic issue today than it used to be - even (or especially) in a country like Norway, which is small and where the situation has not been complicated by a federal structure (as in the United States or Western Germany). This is due to the increased activity of the public administration. The administration interfaces in principle with the whole population. In many branches of the administration this is reflected by a hierarchical organization, reaching out to the local communities. At present there is also a tendency to further decentralization, both in the sense of geographical decentralization and in the sense of delegation of decision power to local offices.¹ In addition, we have the nature of the norms on which the administration bases its decisions; there is a tendency to leave room for discretion, especially in norms regulating the benefits due to a citizen of a modern welfare state.

The decentralized administration brings the lawyers of this administration into closer contact with the local communities. This is a good thing in many ways, making it easier to collect relevant facts, easier for the citizen to seek out the person in charge of his case, etc. But it also opens the way to some dangers. The discretion left to the lawyer must be based on value judgements (in our model of the legal decision process, this

¹ For a general discussion of the relationship between the legal information system and the decentralization issue, see Karlsen 1973.

[Page 243 ]

would be the extra-legal value norms). Such values are not created in a social vacuum, and the lawyers may be unduly influenced by the mores of the local community. At least in Norway, it is not difficult to find differences from one region to another, for instance in the attitude toward unwed mothers. Harboe (1974: 18-19) cites has several examples from Norwegian tax law.

There are several strategies to choose from when trying to avoid the creation of "local law". One such strategy would be to normalize the discretion through regulatory law. A discretionary norm formulated in a statute would be reformulated into a complex, but strict, rule through regulatory law - leaving little room for discretion.

Another strategy could be to isolate the controversial issues and extract these cases for decision by a central agency - a strategy which in fact has been adopted in the Norwegian social security administration. (Cfr. Karlsen 1975:14.)

A third strategy would be to concentrate on the information system. Usually quite some effort is spent in decentralized branches of the administration in co-ordinating their decisions. This is done through courses, meetings, and the information distributed throughout the administration.

In our opinion, two of these strategies may be compared. One is the transformation of discretionary norms into strict rules. This is, in fact, a way of centralizing decision power while keeping a geographically decentralized organization. In many ways such a strategy would be opposed to the wish for flexible norms. The other strategy is a better legal information system. Through better access to a file composed of a representative selection of precedents, a co-ordinating effect may be achieved without centralizing decision power. Such an information system would function as a communication network for the administration, integrating a geographically decentralized organization.

Also, the transparency of the system might be improved if the citizen were given access to the information system - a step that may reduce the number of appealed cases and have other beneficial side-effects.

12.4 DISCRETION VERSUS PREDICTABILITY

12.4.1 Introduction

We have in sections 12.2 and 12.3 tried to relate aspects of the legal information system to the principles of predictability and equality. Both

[Page 244 ]

these lines of reasoning favor a development of the present legal information system, making more legal sources available to a greater number of users. This will improve the predictability in the legal system and ensure the realization of the principle of equality to a greater degree than is the case today.

It ought, however, to be stressed that such a development of a legal information system may also be negative with respect to other standards. One such standard is the desirability of discretion in the legal decision process. This is related to that element of the "rule of law" which Eckhoff (above in section 12.1.2) calls "fair and just process".

Legal decisions should reflect the unique nature of each case and not be preprogrammed by earlier decisions. An improved information system may strangle the development of new legal norms, unduly strengthen the conservative nature of law, and lead to a state of rigor mortis in the legal system infected by the improvements.

In this section we will look more closely at this conflict, which is - we think - an issue of some importance. It is also a problematic issue; as advocates of a new technology our prejudices may seduce us to decide in favor of changes.

12.4.2 The effects of more frequent use of a type of legal source

If a legal information system is improved, this will be in order to facilitate access to one or more types of legal sources. An information system may be introduced which gives Norwegian lawyers systematic access to a representative selection of first instance decisions, or to precedents within the administration. The most important types of legal sources are today served by the most satisfactory information systems. Such types of legal sources (statutes, regulatory law, Supreme Court decisions) may also be included in the system, but the relative improvement will be greatest for sources which are of a low rank.

If access is facilitated, the consequences will be more frequent use. Lawyers will be able to retrieve relevant sources not accessible (or only accessible through great cost) in the old system.

According to our model of the legal decision process, a relevant legal source will influence the normative interval. This influence will typically be a reduction leading to a corresponding reduction in the liberty of the lawyer to select alternatives. This will also typically represent a reduction of discretion.

As stated, the legal sources of a higher rank will usually be served by

[Page 245 ]

better information systems. The reduction of discretion will therefore not be felt in cases in which the solution might be found in these high-ranking sources (statutes, regulatory law, etc.). The reduced discretion will therefore mainly affect the importance of other factors - according to our model these are most likely to be the extra-legal value norms and assessments of the result. Finding that this change takes place is in itself neither good nor bad. An evaluation must be based on a comparison of the roles played by discretion before and after the change.

The critics seem to presuppose that the amount of discretion in the present system is desirable or optional. This is by no means true.

If the discretion is due to the nature of the norms themselves, and a goal aimed at by the legislators, then a reduction would be problematic. But if the discretion is due to arbitrary access to legal sources, lack of knowledge of the valid norms, then a reduction of the discretion would be desirable. Even accepting that the typical change would be a reduction of discretion, this cannot be assessed out of context, but would have to be based on a concrete evaluation of the legal information system in the actual sector of law. As the traditional legal information system all too often is a result of historical coincidences rather than of planned development, it would be surprising indeed if they promoted the optimal balance between predictability and discretion.

We stated that a typical change - according to our model - would be a restriction of the normative interval. However, this may not be the actual fact. Let us take the example of a lawyer citing to a judge a number of first instance decisions as a basis for his argumentation in favor of his client. The judge, acknowledging these as legal sources, would impose a restriction on his own alternatives. The present information system gives him no possibility of checking the representativity of these decisions, or of doing a bit of legal research of his own into alternative solutions. If he therefore wants to base his decision on a norm in conflict with those apparently applied in the documented decisions, he has to stress the uniqueness of the case at hand, stress the importance of a good result or something else.

In the case of an improved information system, the judge could have checked the representativity of the documented decisions. And he might, in fact, have found other decisions which might serve as a basis for counter-arguments, he might discover certain distinctions made in earlier cases not easily apparent from the documented decisions, etc.

One may argue that access to a great material of legal sources -

[Page 246 ]

especially of a casuistic nature like case law or precedents - will furnish the user with an arsenal of potential arguments which may be applied to let the uniqueness of the case at hand come into full play.

To illustrate this further, one may picture a situation where the lawyer has to consult a precedents file in order to ensure the enforcement of the equality principle. In a conventional system this file would have to be of a manageable volume; too many references for each entry would lead to a state of over-recall and make the system impractical. Consequently rather few cases are selected to represent the totality and the analogous use of precedents would be encouraged. A better information system would reduce the problem of over-recall - in a text retrieval system it would be on a completely different scale. Consequently a greater number of cases might be included, thereby discouraging analogous use of precedents, and encouraging the lawyers to take into account the unique traits of each case.

An example from the user experiment in the Norwegian Social Security Court may also be mentioned. The clerks prepare the cases for the judges. They themselves have a formal legal education, and do often have strong opinions as to what the law should be on the disputed issue. It was admitted by one that he had used the text retrieval system to isolate precedents which strengthened the opinion he held to be correct, cfr. Brukerforsøk 1975: 26 - this example may indicate that a better information system will not reduce the discretion of the lawyer, but will make it easier for him to justify his decision. This is, of course, also a rather problematic effect.

12.4.3 The effects of increased rank

When discussing our model of the legal decision process, we maintained that there was an interdependence between the legal information system and the meta-norms governing the decision process. The meta-norms assigning rank to the different types of legal sources were described as decisive when resolving conflicts between norms. When an improved information system propagates the more frequent use of a type of legal sources, this will not in principle cause this type to gain higher rank. But the interdependence will be a dynamic force when changes are made in the information system. One may therefore foresee that improved access to a type of legal sources causes higher rank to be assigned to this type.

It is very difficult to say whether this is a probable effect or not. It would be practically difficult to distinguish between the two effects discussed in this and the former section. Probably a shift of rank will be subtle, and can only be ascertained over a rather long period.

[Page 247 ]

Changes of this sort in the information system do not seem to be desirable, however. The legal meta-norms should not be the result of pragmatic factors (though this probably often is the case), but of a conscious attitude among lawyers and others. One of the aims of legal literature must be to make lawyers conscious of the legal meta-norms, revealing different alternatives and potential problems. For this purpose some sort of conventions have now been created for how statutes should be interpreted, conflicts of norms solved, etc.

Modern technology has made possible major changes in the tools available to the lawyer. These changes will also represent changes in the pragmatic factors in connection with which our present meta-norms have been forged. The legal profession is not accustomed to such drastic changes, therefore it may not be prepared to counteract possible undesirable effects of the change.

In our opinion, the change in an information system should be paralleled by an effort to make the users more aware of the meta-norms governing the use of the sources made available through the system. Only in such a way - through the awareness of the user - may one guarantee against undesirable side-effects, like shifts in rank.

Actually, we should like to add that a greater awareness among the users would probably also give rise to criticism of the existing information systems, which in many respects are not designed to strengthen the commonly accepted meta-norms.

12.4.4 Conclusion

We have in this section looked upon two factors that may unduly reduce the degree of discretion in legal decisions through the introduction of an improved information system: the more frequent use of a low-ranked type of legal sources, and the higher ranking of such a type.

In appraising this shift, we maintain that one must first determine whether the degree of discretion in the present system is desired, or whether it is the result of arbitrary factors. If the latter is the case, increased predictability is clearly desirable. If the former is the case, one should consider whether reduced discretion is indeed the probable effect. If this is found to be the case, it must be taken into account when planning (or keeping) the information system.

The role of the user may have been understated in the above sections. The user is not, of course, the passive mouse in the paws of a feline information system. Where the user has a background in education and

[Page 248 ]

tradition, we trust that he will make a conscious effort to follow the meta-norms he holds valid. To a certain extent the critics who are negative to improved information systems reveal a mistrust of the user, which is often the mistrust of a formally educated lawyer toward the civil servant (or other decision-makers) lacking formal legal education.

Actually we think this change in perspective is symptomatic. A better information system will give each user a greater potential, a potential that will make the personal qualifications of the user more important, cfr. Blanck 1975: 119. Thus a better information system may become a challenge to civil servants and the management of public administration. But this is no real argument in disfavor of a better information system: management through information deprivation does in our view come close to mismanagement.

12.5 OBJECTIVITY

12.5.1 Introduction

In two sections above we have touched on problems related to the objectivity of the legal system. In section 12.2.5 we surveyed the problem of checking the representativity of documented sources of law - where this possibility is lacking, it may reduce the predictability of the legal system. And in section 12.3.2 we discussed the problem of having a representative coverage in order to ensure that similar cases really are decided in the same way.

These two sections may be viewed as parts of a greater issue related to the "rule of law" and legal information systems: the principle of objectivity. The legal information system should be designed in such a way that it does not promote the interest of a single group or a single ideology. This problem is, of course, not merely related to computerized information systems. But in conventional systems, there is usually no real general plan - the system is as a rule composed of a number of rather independent publication routines, indexes, etc. It is believed that the plurality of the system does give some sort of guarantee for the objectivity of the system. The question of objectivity will become more acute when a general plan is launched for a national legal information system. One may then feel that the grand design has an integrated selectiveness or other attributes that will reduce the objectivity of the information system.

Actually, a conventional and privately operated legal information system may achive a practical monopoly. As mentioned in chapter 4, the

[Page 249 ]

systems of FLITE and JURIS have license arrangements with West Publishing Company. These arrangements have been questioned, and the Antitrust Division of the Department of Justice has opened a formal investigation into the anticompetive and legal implications of the arrangements between West, the Air Force, and the Department of Justice itself, cfr. Rubin 1976: 9.

As far as we know, discussion on this issue has been mostly restricted to Germany, where the JURIS project represents such a grand design, cfr. Simitis 1974. It seems that the group responsible for the JURIS project has recognized this danger, cfr. Fabry 1973: 6:

"A comprehensive public system would, owing to access through certain channels and a certain monopolistic position, necessarily gain importance as a means to influence opinion to a high degree. It would, therefore, be intolerable if such a system were to reduce the plurality of stored legal points of view or to produce a distorted picture of their quantitative representation."

Though the problem is recognized by the designers, there may still be disagreement as to how well the problem has been tackled in practice. However, our major concern here is not the JURIS system, but the problem of objectivity in general.

It ought to be pointed out that conventional systems too may lack objectivity. For instance, a number of the publication routines constituting the present Norwegian system are maintained by private organizations, and most of these routines incorporate selective mechanisms. Also, the private organizations often represent one of the parties in the potential conflicts (cfr. Auto-Trafikkforeningens Domssamling). Such traits would not seem to strengthen the objectivity of a legal information system. Also, the problem of objectivity is related to the discrepancy between "the law of the books" and "the law in action". Often academic lawyers and legislators lack systematic knowledge of "the law in action" at first instance courts and administrative agencies - as there are no distribution routines making information available. They may therefore remain in the dark about the "law in action" - which may be quite different from the "law of the books", especially where the judges have discretionary powers. This ignorance may project a distorted picture of the actual situation, cfr. below in section 12.5.4.

Though we are here interested mainly in the information system, one should note that other factors too may reduce the objectivity. An example: tax law cases are often connected with considerable financial interests, and consequently clients who really have something to gain

[Page 250 ]

12.5.2 Coverage

The primary problem of objectivity is the problem of coverage. We have argued earlier that the principle of equality will only be satisfied by a representative coverage. The principle of equality is founded on the comparison between a pending case and a case already decided. The principle of objectivity has a wider scope, it will only be satisfied when the coverage is representative in relation to all relevant interests. It may be argued that in reality this implies that any coverage less than total is biased (cfr. Simitis 1974: 15, 27-28). Any selective mechanism must choose some standards as a basis for the selection, and any human judgement will reflect some sort of subjective - and consequently biased - norms.

It is very useful to focus the attention on the selective mechanisms of both conventional and computer-based information systems. In computer-based systems, the availability of the stored sources is very high - and the frequency of use will probably correspond to this. Therefore, any selection mechanism ought to be scrutinized.

We do not think, however, that the Norwegian legal system can do without selective mechanisms. The number of legal sources is so great that it would not be realistic to make all of them available to all lawyers through today's information technology. This may be different in a greater country like Germany, but we strongly suspect that it will become very difficult indeed to make the totality of all sources of any potential interest available. Also, the information system would then be drenched with trivial decisions of no obvious value to the user, making over-recall an acute problem and demanding a very high quality retrieval system - and a correspondingly high degree of knowledge of the use of retrieval strategies from the user.

It would seem that we are left with the dilemma of selection mechanisms for the forseeable future. One must therefore try to find alternatives that will secure a high degree of objectivity.

One possibility would be to have a decentralized, dynamic selection mechanism. This might be designed in many ways, but we should like to sketch one alternative proposed in Forundersøkelse 1974:48-58. The proposed system (PASAR) is one for the distribution of

[Page 251 ]

12.5.3 The problem of document representation

The problem of representativity repeats itself in regard to document representation. As we have seen, the documents are in some systems represented by a document surrogate. It is recognized that the indexer will project an interpretation on the original document when composing the document surrogate. Consequently a certain distortion will take place - a distortion that will reduce the objectivity of the system. This may be taken as a strong argument in favor of a full-text system.

The problems of document surrogates biasing a legal information system have also been touched upon in our description of the Italgiure system, cfr. above at section 5.4.2 and the critical comments of Ciampi 1974:711-713, 721-725. For a Norwegian example, cfr. Feyen/Harboe/Lie 1973: 65.

Simitis (1974: 32) claims that although there are technical alternatives to full-text representation, these are not politically acceptable. In earlier sections we have indicated that controlled experiments in text retrieval indicate that systems based on full-text have better retrieval capabilities than systems based on document surrogates. Retrieval performance is therefore not an argument in disfavor of a full-text solution. Cost may be one, however. But if the text can be captured at the source, a full-text solution may be as cheap as a solution based on document surrogates - the somewhat higher storage costs being offset by the effort saved in composing the surrogate. Capturing of the text at the source is not, however, always possible. Retrospective documentation would for instance presuppose re- registration. In this case - especially where a document surrogate is already available (for instance as a headnote to a

[Page 252 ]

published decision) - the economic arguments may have greater weight than the fact that the full text ensures better objectivity (and retrieval performance).

Therefore one should also consider which alternative method is best suited for producing unbiased document surrogates. In the case of retrospective documentation one is often left without the possibility of determining the procedure, as the abstract has been produced outside the system. But when designing new routines (as in our example of PASAR in section 12.5.2), one will have a choice. Again, the possibility of a decentralized production of abstracts would seem to guarantee diversity and, consequently, some sort of objectivity.

Moreover, we would think that a non-normalized terminology would also ensure higher objectivity than a well-defined indexing language. Simitis (1974: 26) has a very convincing argument: the use of a defined indexing language will increase the conservative nature of a legal system. Law is by nature oriented toward the past; social changes will generally only be reflected in the legal norms (for instance as new legislation) some time after they have been politically recognized. The use of a normalized vocabulary would increase this inertia, as not only the legal norms would have to reflect the change, but also the definitions in the indexing language.

A dynamic aspect would be more difficult to introduce into a document surrogate, though the system should, of course, provide for the possibility of revisions or annotations. The use of citation structures might be a partial solution to this problem. The indexes of new monographs, text-books, etc. could be incorporated into the data base, giving the user a reference to sources commenting upon the document that is the original of the retrieved surrogate.

The possibilities offered by hybrid systems like RIRA or DARIUS may also be considered. Though these systems include only document surrogates in machine-readable form, the full text is readily available in microform - giving the user an easy way to check the interpretation of the surrogate. The arguments of surrogates biasing the information system are strongest when the surrogate replaces the original document (as in Italgiure). Where the full text too is readily available, the retrieval system based on the surrogates will only function as an index to the original documents. Even so, the index may be biased and in fact exclude the possibility of employing certain retrieval criteria or strategies (cfr. Simitis 1974: 25-26).

12.5.4 The influence on the law in action

Lastly we should like to emphasize once more an aspect pointed out above under section 3, and mentioned in the introduction to this section. The types of legal issues brought to a decision may in turn be partially

[Page 253 ]

determined by the infra-structure of the legal information system. We would think that problems which demanded extensive legal research usually would represent less profitable cases than problems where the lawyer may rely on his background knowledge or be assisted by adequate information systems. At least the need for extensive legal research would make the handling of such a case more time-consuming and expensive - making the prospective clients less motivated for seeking legal advice. Our material does not provide us with anything but questions in this respect - and further probing would come close to being speculation. All the same one would think that the introduction of an improved legal information system would have some effects on what sort of conflicts find a legal solution. This would be a long-term effect - not affecting the quality of the isolated decision, but affecting the quality of the total legal decision system and its social function.

Actually, it may be argued that the existing legal information system is reflected in a certain specialization among lawyers. Some lawyers - especially academics - are employed in producing an extensive statement of disputed legal problems of a complicated nature. They are consulted, no doubt, because of their personal qualifications and experience. But to some degree we believe that they are also consulted because of their proximity to the well-stocked university libraries and other important information systems, as well as their training in the use of such aids. This may therefore be an effect of coverage discrimination, which might be reduced by an improved information system - giving the ordinary lawyer better possibilities for conducting this kind of research.

12.5.5 Conclusion

Objectivity is a recognized ideal of a legal information system, related to the "rule of law". It can in principle be attained only through total coverage in full text. Even then factors outside the information system (systematic discrimination of issues brought before the court, etc.) make the "objectivity" problematic.

It will be realistic to expect that the computer-based systems of our near future will have only partial coverage, and also have some type of legal sources represented as document surrogates. In this situation the selection mechanisms and procedures for producing surrogates will be of importance, as they ought not to reduce the objectivity unduly. We have not made a thorough survey of the strategies available for ensuring optimal objectivity in this situation, but have indicated that a practical alternative might involve decentralized and dynamic schemes for selection and surrogate production.

[Page 254 ]

12.6 TIME VERSUS QUALITY

12.6.1 Introduction

Above, in section 3, we outlined the two possible justifications for introducing better retrieval systems: the possible profit of reducing time (or effort) spent through a more efficient technology for legal research, and the possibly improved quality of legal decisions. In the same section, we discussed two reports on user behavior (Operation Compulex 1972 and Jungjohann/Seidel/Sörgel/Uhlig 1974), which both seemed to indicate that the time consumed by legal research was a relatively small part of the total time spent on a case by the lawyer. Against the background of our model of the legal decision process, this was hardly surprising. We did, however, argue that there was no easy way of extrapolating this status into the future, claiming that these surveys "prove" that the new technology is not viable, from an economic point of view. This was due to the fact, we said, that the new technology would introduce a change in the pragmatic factors defining the work situation of the user, and that the relationship between this change and the other factors is complex, making the effects of such a change less predictable. We have, however, accepted the argument so far as to maintain that the justification of introducing computer-based reference retrieval systems cannot chiefly be founded on the possibility of saving costs, but on the improvement of the quality of legal decisions.

For a further criticism of user research in this field, cfr. the remarks of Simitis (1974:14-24), which are very much to the point. He emphasizes that user research is necessary, but cannot do more than give a diagnosis of the present situation. It will not by itself be a sufficient basis for well-founded prognoses. He also suggests that user research may be used as a legitimation by the government when introducing a legal information system of a certain design.

In this section we shall carry the discussion a bit further, in order to stress some of the dynamic forces expected to be brought into play when the information technology changes. This section may be seen as the conclusion to the introduction presented in section 3.

12.6.2 Research time - balancing turnaround time

We have several times maintained that the user's situation is defined by pragmatic factors. An analysis of these factors would fall outside the scope of this book. But we may characterize the situation of the civil

[Page 255 ]

servant in a modern public mass administration - as the situation emerges through the user research of the NORIS program (cfr. Bing/Harvold 1973:202-305, Feyen/Harboe/Lie 1973, Brukerforsøk 1975). These studies have discussed the information situation in the tax and the social security administration.

The overshadowing factor in the situation of the civil servant is the number of decisions that have to be made, this being - of course - a reflex of the number of cases injected into the system. The time spent on each case may be considered as a simple calculation in which the number of cases is divided between the number of manhours available. This is - of course - not an entirely realistic calculation, but it will indicate the margins of time involved. In fact, Føyen/Harboe/Lie found (1973: 104) in one instance that the average time spent on a tax return was 12 minutes - which certainly does not leave much time for extensive legal research.

The number of civil servants will often be related to the anticipated volume of work - in the tax and the social security administration the staff of a local agency is to some extent a function of the number of inhabitants in the local community. This is, however, only an approximation leaving quite a lot of room for variations.

The number of cases will be determined mainly by (1) the legal norms, and (2) by socio-demographic factors. In the tax administration, the first factor is dominant (all Norwegians fill in a tax return which is assessed by local tax agencies), in the social security administration socio-demographic factors are more prominent (labour market, distributions of age, type of work, etc. in the local population).

We will not go into detail at this point, but feel that this background will suffice for claiming that the workload will be determined by factors outside the organization itself. It is also well documented that these administrations are understaffed, implying that they feel they are not able to give each case as much attention as they would like.

When a case is being decided, certain routines will have to be executed - like checking the summations on a tax return, filling out the appropriate forms, etc. The time consumed by such "necessary tasks" may not be reduced by the user. Legal research, however, is something rather flexible. If he feels a certain time pressure, he may react by cutting down on his research. Our user research indicates that this is not an uncommon reaction. The user will not aim at "satisfactory", but at "sufficient" legal research, cfr. Føyen/Lie/Harboe 1973: 103-104.

In this situation, availability factors demanding any effort at all from

[Page 256 ]

the user in terms of time will drastically reduce the use of the legal sources barred by such factors. Examples may be a precedents file on another floor (Bing/Harvold 1973: 276) or unbound volumes of case reporters tied together by string (Føyen/Harboe/Lie 1973: 49).

Just this situation may charcterize - we think - most public administrations. Legal research time is used to balance the time spent on each case.

And just this situation will be the environment of a reference retrieval system. Introducing this system will not primarily reduce the time spent on legal research. This time - already reduced by scarcity of time - comes close to a minimum determined by the workload. It will, however, give the possibility of more satisfactory research - pre-supposing that the computer-based system has a more satisfactory coverage than the coverage defined by the availability factors earlier surrounding the user.

Therefore, one may not expect a reduction in turnaround time when introducing the new technology. But it may be hoped that a well-planned new system will improve the quality of the decisions without increasing turnaround time. And as the quality of the decisons - measured by the standards of the "rule of law" outlined earlier in this chapter - may be less than satisfactory or even acceptable, this improvement will by itself be more than sufficient justification.

This improvement might be reached through other strategies, the most obvious would be to increase the number of staff, thus reducing the amount of work resting on each civil servant. This is, of course, an alternative - but in the battle between a computer-based solution and a solution demanding more employees, the former will probably win. This will be the result of a political decision, ushering in the classical discussion of men versus machines, which we shall not pursue here.

Another strategy might be to give the civil servants better education, hoping that a widening of their background knowledge would improve the quality of the decisons. Actually the introduction of a text retrieval system might be viewed as a tool for education - the retrieval process will also be a learning process, cfr. above at section 1.2.8. If the text retrieval system is introduced in order to facilitate the decentralization of decision power, this decentralization will also presuppose increased qualifications on the part of the user, cfr. Blanck 1975: 119.

As for most other strategies available for solving the problem outlined in this section, better staffing and education are not exclusive alternatives to text retrieval systems, but may be viewed as elements of an overall solution.

12.6.3 Transparency and workload

In our description of the workload, we left out one factor: the frequency of appeals. In the Norwegian administrative system all decisions may be appealed to a higher administrative level. The total workload of the

[Page 257 ]

administration is consequently also relative to the frequency of appeals. One strategy aimed at reducing the workload would be to reduce the number of appeals.

The information system may be related to the number of appeals in at least two respects.

Firstly, we have the quality of first instance decisions. Our user research has indicated that the possibility of appeal may serve as an excuse for dismissing the case without undertaking satisfactory research. The scant justification for such a procedure is that the client - if dissatisfied - will appeal the decision, and that there will be more time for legal research in the next round. This attitude is hardly satisfactory as regards the "rule of law" (it reduces predictability and threatens the principle of equality). It will also lead to a higher frequency of appeals and a greater total workload. Furthermore the appeals would probably be systematically distributed; resourceful clients would appeal relatively more frequently.

Introduction of a more efficient reference retrieval system would encourage higher quality in the decisions, thus reducing the frequency of appeals and the total workload.

The second - and probably more important - relation is the one between improved predictability and the frequency of appeals. This is a long-term benefit that may be expected to flow from a better information system.

One of the reasons for appeals is the lack of predictability due to coverage discrimination within the administration (cfr. above at section 12.2.3) which may breed distrust among the clients, who will find that appeals are granted to the puzzlement of the first instance agency. Especially where there are no costs on the part of the client involved in an appeal, the client may feel that he will have "everything to gain and nothing to lose". A better information system, eliminating internal coverage discrimination, would gradually strengthen the prestige of the first level agency - and consequently reduce the number of casual appeals.

Even more effective would be the improved transparency. The better information situation would make it easier for the client or his legal advisor to examine his own legal position. At least in cases where it is obvious that an appeal would be dismissed, the client will not bother to appeal.

The dynamic aspect of this situation should not, we feel, be underestimated. With an ailing information system, the situation is rather

[Page 258 ]

murky both for the first level agency and the client. The introduction of a better information system would act as a clarifying agent.

Actually, a tendency in this direction was noticed in the user experiment at the Norwegian Social Security Court. The text retrieval system was used by one of the judges to retrieve and examine all decisions relevant for "alcoholics applying for invalid pension". On this background a general norm was laid down in a new decision.

Not only would the existing legal norms be more apparent, but it will become easier to elucidate the situation through new decisions or regulatory law - as the problematic issues will be more in evidence for all concerned. The predictability and transparency of the system would steadily increase until it reached a new state of balance - and all through this process the frequency of appeals would be expected to diminish.

[Page 259 ]

HOME
PREVIOUS PAGE
NEXT PAGE