HOME
PREVIOUS PAGE
NEXT PAGE

Preface

We set out to write this book with two purposes in mind. Firstly, we wanted to communicate in English some of the results arrived at through NORIS, a research program conducted at the Norwegian Research Center for Computers and Law by our colleagues and ourselves for the last four years. Secondly, we wanted to give a general introduction to the field of legal informatics, and especially to the sector of this field which may be called legal reference retrieval. It seems only fitting that in a book involving information systems we should mention the problem we ourselves faced in collecting material for the book. The problem will be well known to all who work in a field undergoing rapid development in a number of different countries simultaneously. We approached the problem partly by utilizing the excellent library in computers and law maintained by Oslo University Library at the Center. In addition we enlisted the help of friends and colleagues in the United States, Canada, the United Kingdom, and continental Europe by asking them to glance through relevant sections of the manuscript. For their gracious cooperation we want to thank each one of them. The result of our efforts, however, is bound to be incomplete and inadequate in many respects, and we welcome our readers to communicate their comments to us at the address of the Center given in this book. We have tried to bring our description of retrieval systems up to date as of 31 December 1975. Development in 1976 has in several instances been mentioned, but no systematic updating has been done.

The book has been written in close cooperation between the authors. The actual responsibility for chapters 1-7 and 12 rests with Jon Bing, while the responsibility for chapters 8-11 rests with Trygve Harvold. This book could not have been written without the constant exchange of ideas which characterize the work at the Center. We should like to thank all those who have contributed to the projects on which the book

[Page 5 ]

is based. We want especially to thank Ms. Valborg Wasmuth, who has sorted out and typed the manuscript and who has been a constant help to us during the preparation of this book. We are indebted to the Norwegian Research Council for Science and the Humanities, which granted funds for much of the research within the NORIS program, as well as for the writing and production of this book. In 1976 the work has also been supported by the Norwegian Ministry of Justice.

I LEGAL DECISIONS

[Page 13 ]

1 A model of the legal decision process

1.1 PERFORMANCE OF LEGAL DECISION PROCESSES

This book focuses on legal information retrieval. "Information retrieval" is used as a label for one of the elements in a legal decision process. We think that the "information retrieval" element cannot be properly understood if dissociated from the total process. Our aim in this introductory chapter is consequently to describe information retrieval as part of the legal decision process.

We will use the term "decision" in a rather broad sense so as to include forms of legal reasoning that do not conclude in a decision, but in an opinion or in advice, cfr. Eckhoff (1971:23), Ekman (1970:15). Slayton (1974:3) uses the expression "the lawyering process"; we prefer the expression "decision process" as it indicates a typical and important characteristic of the process. Kilian (1974:11-12) restricts his concept "juristischen Entscheidigung" to the typical decisions of, for instance, a court.

In our discussion of legal information retrieval, our interest lies in finding ways of making the retrieval more effective. But in order to determine what is "effective", we have to consider information retrieval as a sub-process of the decision process. Improvement of an information retrieval system only has value to the extent that the decision process as a whole is improved. This has been stressed by Slayton (1974:21), who suggests "that electronic legal retrieval systems designed to assist in, or substitute for, a key part of the legal thought process have been developed with little understanding of what the process is, and what the consequences of changing it will be". Without taking a stand on the justification of Slayton's criticism, we agree that he outlines the only adequate perspective for describing legal information retrieval.

For a comment on Slayton's view, see Mackaay (1974) and Rubin (1974), as well as Slayton's response to these comments, Slayton (1974b).

As the information retrieval process has the same goal as the decision process itself, it becomes important to identify this goal. This will be done

[Page 15 ]

through the description of the decision process in section 1.2. At this point we would, however, like to stress that the question is not easily answered. The goal of the legal decision process may be specified at many levels, and at each level it is probably of a rather composite nature. At the ultimate level the goal may be identified with ideals such as "the rule of law", "justice", etc. These ideals are not controversial, but in the process of giving a more detailed description of them, controversial issues are bound to arise. (Cfr. the analysis of "justice" by Eckhoff 1971b.) On a practical level, the goal may be identified with for instance "speedy" or "cheap" decisions. And on an intermediate level, the goals may be found in the norms governing the decision process itself.

For the better part of this book, we have chosen to identify the goals of the legal decision process with the norms governing this process. (The nature of these norms will be described in the next paragraph.) In chap. 8 we will, however, try to relate information systems and their part in legal decision processes to more ambitious goals.

1.2 A MODEL OF THE LEGAL DECISION PROCESS

1.2.1 Introduction

The purpose of presenting this model of the legal decision process is to specify and relate the different elements of a legal decision process. It is intended to reflect a rather uncontroversial picture of how a legal decision is formed, based upon contemporary Norwegian legal theory. But there is, of course, a rather strong reservation in the term "model"; it is not an attempt to describe the psychological process going on in a lawyer's mind when working on a problem. It is a description in principle of the different elements of a decision process, as they may be sorted out after a decision has been formed.

In reality many of these elements will be represented by intuitive leaps of the mind, and it may consequently seem somewhat non-realistic to attempt to portray them in a model. Also, the model grossly understates the interaction between the different elements when a problem is being worked out in the mind. The content of any element may not be determined until the decision is made - only then is a balance established which makes it possible to analyze the decision. But - for representational reasons - the process will be described as a sequential process with one beginning and one end.

It is necessary to stress the limitations of our approach. On the other

[Page 16 ]

hand, we think this approach is useful when discussing elements of the decision process, as for instance the retrieval process. The model elucidates the relationship between this and other elements in the total decision process. It makes it possible to discuss terms which otherwise may become vague and imprecise.

The model is based on an earlier article (Bing 1975), which was inspired by the works of Eckhoff (1971:25, 30) and Andenæs/Kvamme (1969:19). Valuable inspiration was found in the German discussion on "Rechtsinformatik", especially Kilian 1974, and Gartska; and also in the survey made by Slayton (1974:3-6). Predecessors of this model may be found in Bing/Harvold 1973:2-10 and Bing 1974:100-102.

In presenting the model, we will bear the purpose of this book in mind. Consequently we shall explore the anatomy of the decision process with our attention centered on the retrieval process. Other elements of the decision process will be sketched, but not in detail. For a more exhaustive description, cfr. Bing 1975.

1.2.2 Initiation

A legal problem is something that emerges out of interaction between humans; it is part of society and exists by itself before it is brought to the notice of a lawyer. We are, of course, only concerned with the legal problems brought to the notice of a lawyer, either by a client, who experiences the problem, or by the activity of the lawyer himself.

The problem involves one or more persons whom we shall term "the client". In order to activate the legal decision process, the client must contact a lawyer. This is in its own right an observation of some relevance, as the initiative by the client presupposes that he is aware of the legal nature of his problem.

For the purpose of this discussion, we use the term "lawyer" to denote all persons working with legal problems, regardless of their formal education or degree, cfr. Stadler (1973:75), Luhmann (1966:10). For the sake of simplicity, we also restrict our description of "problems in reality" and the lawyer to the typical client-lawyer relation.

The lawyer himself is not part of the problem, but a sort of spectator giving advice to his client. His contact with the problem is indirect - the private lawyer through his client, the civil servant through a citizen. A judge also has a "client" - who is not, of course, one of the contesting parties before the court, but rather the case which these parties jointly present.

[Page 17 ]

1.2.3 Legal problems

The problem, as experienced by the client, is not primarily classified as "legal" or otherwise. The first step of the lawyer is to determine whether the problem - as presented to him through his client - is legal or partially legal. This is not as trivial as it may appear. If a client complains about health, housing, his family or his economy, it is not evident that the best solution is an invalidity pension, housing grants, and a divorce. The legal problems may be part of a more complex problem situation - or, indeed, symptoms of other problems. The client might perhaps be better aided by medical care, retraining and contact with a social adviser than by extensive legal assistance.

As a lawyer, one is nevertheless restricted to extracting the legal problems out of the totality. This presupposes that the lawyer can identify a legal problem. This will in most cases be second nature to a lawyer. He is an expert with background knowledge of the legal system - and is consequently able to grasp the legal problems that are part of the totality. As a characteristic, one may say that a legal problem is a problem to whose solution legal argumentation may contribute. This is a pragmatic characterization, but it will suffice for our purpose.

Looking a bit closer at the nature of a "legal problem", one may specify three typical examples:

(1) There exist valid norms prescribing that problems of a certain type may be solved by legal reasoning. Typical examples are decisions made by the public administration - statutes governing these activities will imply that decisions are the results of legal decision processes.

(2) The problem is a dispute that may be brought to court. The possibility of a trial will throw a shadow across the problem and make a legal decision the normal solution even when settling out of court.

(3) The parties (or the environment) accept that a legal decision process is a valid way of settling the problem. This acceptance will depend on several factors, for instance the social prestige of the lawyer or lawyers in general, or the effectiveness of a legal decision process (which is a rather effective way of arriving at a solution).

1.2.4 The probable facts of the case

When the lawyer has settled which are the legal issues of his client's problem, he sets out to describe the facts of the case. Having no direct knowledge of the case, he has to rely on evidence - primarily discussions with his client, but also with other persons concerned with the problem, examination of documents, etc.

Some facts will be evident (like the identity of his client). But here a few

[Page 18 ]

meta-norms exist prescribing how the lawyer is to arrive at the probable facts of the case.

The term "meta-norm" is, for the purpose of this book, used for all norms governing a legal decision process. This is in contrast to norms of substantive law, which prescribe the possible solutions to the initial problems. Meta-norms will consequently define or determine the outline of the legal decision process. We are aware of the fact that meta-norms themselves may be of different categories, but feel that this terminology will suffice for our purpose.

The meta-norms relevant for determining the probable facts of the case, are relative to the position the lawyer finds himself in. In Norwegian law, a judge is obliged to ascertain that the case is sufficiently documented before he makes a decision (cfr. rl. § 190 and strpl. § 331, 5. section). A similar obligation is prescribed for the civil servant (cfr. fvl. § 13), and if the case is not adequately documented, the decision may be deemed void (cfr. Frihagen 1966:104- 118).

There also exist meta-norms governing what is allowed as evidence. In Norwegian law these norms are very modest, and restricted to a few topics like the question of whether evidence secured by illegal means may be allowed in court (cfr. Andenæs 1962:336-341). In England and USA, however, the law of evidence plays a more prominent role in the legal decision process.

In the main the lawyer builds his assessment of the evidence on general human knowledge. When in doubt, he will choose the set of facts most likely to be true.

But also normative attitudes of a different order may be introduced at this stage - for instance meta-norms of client loyalty. The lawyer may feel inclined to accept the version of the facts presented by his client even though another version may appear more likely to be true. Also, client loyalty may influence the lawyer's assessment of what is probable. The sort of influence exemplified by client loyalty may also be viewed as a feedback loop from later stages of the decision process, indicating the dynamic aspect of the process, which our model does not wholly justify.

1.2.5 The proven facts of the case

In most legal systems there exist meta-norms that introduce special qualifications regarding probability. A set of facts is not regarded as proven unless it is qualified as more probable than any other sets of facts. These meta-norms govern the burden of proof, and they are of course

[Page 19 ]

relative to the legal system. A general example is the prosecutor's burden of proof in a criminal case - you are "innocent until proven guilty".

This group of meta-norms is, however, not too well defined. It has for instance been proposed in Norwegian legal theory that norms governing the burden of proof may be generated by the special facts of the case itself, for instance built on an assessment of the consequences of alternative decision, cfr. Bratholm/Hov 1973:300-301. Once again we see that this set of meta-norms may be influenced by feedback from later stages in the decision process.

Through the meta-norms governing the burden of proof, the lawyer arrives at a set of facts which are proven. These may differ from the probable facts of the case to the extent that qualified probability is neccessary for a certain fact to be proven.

The case - as described by the proven facts - is the lawyer's starting point when looking for relevant norms. (To us it is important to stress that our model takes the facts of the case, rather than the legal norms, as the point of departure - cfr. Kilian 1974:42.) He has used his legal background knowledge in order to extract the proven facts from the totality of the problem. Also he may already have used feedback information from later elements in our model of the decision process. But generally speaking we may say that up to this point substantive law has not entered into the process.

In order to arrive at a decision, the lawyer has of course to select the relevant norms from the legal system. The retrieval process is part of - but not identical to - this selection.

1.2.6 Legal sources

For the purpose of this book, we understand by the term "norm" a certain content founded in certain statements (cfr. Sundby 1974:17). A "norm" is of a semantic nature, but is based on certain statements of a syntactic nature.

In order to identify legal norms, it is therefore necessary to define what statements qualify as a foundation when arguing that a certain norm exists. This definition is given by certain meta-norms, and the statements qualified according to by these meta-norms are called "legal sources".

The distinction between "legal sources" and (legal) "norms" is fundamental. In regard to computers and law, the distinction is also of a practical nature. Legal sources are typically of a written nature (statutes, regulatory law, court decisions, etc.). The text of these sources may be

[Page 20 ]

processed by the computer, and the assistance of the computer may be used to sort out documents containing certain words or combinations of words. But when this selection is brought to an end, the lawyer is only left with a set of legal sources. The legal norms that may be founded on this set of legal sources must still be formulated in the mind of the lawyer.

It would fall outside the scope of this book to discuss problems related to legal norms or their nature. On the other hand, norms are fundamental to all legal thought and a clear understanding of their nature is necessary even in a book with our limited scope. We have taken recent legal theory as our frame of reference, cfr. especially the detailed discussions by Sundby (1974).

In our terminology, norms are of a semantic nature. This does not imply that we disregard the efforts made to represent legal norms in a more unambiguous way than is possible through natural language. Both deontic logic and certain projects with computer-based systems have indicated ways of representing norms more adequately. This does not, however, alter the fact that the norms are still of a semantic nature, while the representations - though unambiguous - are of a syntactic nature.

By our definition, a legal source is a text, a statement, or an opinion,¹ the content of which the lawyer must, ought to or may take into account when arguing the existence of a specific legal norm (cfr. Eckhoff 1971:20-21). A group of legal sources originating from the same "source" is termed a type of legal sources.

¹ An opinion voiced for instance by the local chamber of commerce on the sales practices within a certain trade.

Only arguments that derive from legal sources may be used when advancing reasons for the existence of a specific legal norm. Consequently the qualification of a statement as a legal source becomes a matter of some importance. This qualification is based on legal meta-norms relative to the legal system. They are rarely of an explicit nature. On the other hand, they are not, at least in relation to the Norwegian legal system, controversial. Disagreement on a legal point rarely centers on whether the statement basic to the argument is a legal source or not. In Norwegian legal theory there have, however, been a few instances where the status of certain types of legal sources has been discussed - cfr. for instance Fleischer (1965:152) on the status of decisions by first instance courts and Augdahl (1961:105-110) on the status of travaux préparatoires.

Through the meta-norms a volume of statements is defined. This volume represents the total number of legal sources of a legal system. The lawyer looking for norms relevant to his case must direct his attention towards these.

[Page 21 ]

1.2.7 Availability factors

Usually the lawyer does not have access to the total volume of legal sources. He turns his attention to the part of the total volume that is available to him.

The availability is determined by practical factors, and one may distinguish between two main groups: physical and psychological factors.

The physical factors are trivial, but of great consequence. It may be illustrated by the part that a simple thing like distance plays in determining the use of legal sources. Evidently a lawyer will be more inclined to use legal sources available at his own desk than the sources available at a library some distance away from his office. In principle he may of course order a volume of case reports which may contain relevant sources - but this requires a certain extra activity on the part of the lawyer: writing an order slip, mailing it, waiting till the book arrives, etc. User research in Norwegian public administration has proved that even legal sources available rather close to the office of the lawyer (for instance a precedents file on the next floor) are not used owing to the effort involved. Cfr. Føyen/Harboe/Lie (1973:92) and Bing/Harvold (1973:276).

Another example is the research by Eckhoff (1971:74) into the use of travaux préparatoires when arguing for a certain interpretation of the statute by the courts. In the decisions of courts situated outside Oslo hardly any travaux préparatoires were cited, in contrast to the decisions of courts situated in Oslo. Eckhoff hints at the better libraries in Oslo as an explanation - i.e. the better availability of the type travaux préparatoires in Oslo.

An example of a physical availability factor of another kind is given by Føyen/Harboe/Lie (1973:49). A publication of case reports, administrative decisions, etc. pertaining to Norwegian tax law has an important place in the total information system within this sector. In a certain office, money for binding the separate issues into volumes had not been granted for some time. The issues were kept on a shelf tied together with string. A lawyer had to untie the string and shuffle through the stack in order to find the issues of interest. The extra effort involved was sufficient cause for not consulting the issues.

As an example of psychological availability factors, we may mention case reports printed with Gothic types. This makes the reports difficult to read - and consequently demands greater effort on the part of the user. Other examples may be the language of a written source.

The retrieval system - in a broad sense, indicating the total means a lawyer has of retrieving possible relevant documents from a collection of legal sources - may be viewed as an availability factor in its own right. We prefer to discuss it out of that context. The retrieval is directed toward the accessible collection of legal sources - a base primarily defined by factors of availability other than the retrieval system. But there is a close

[Page 22 ]

relationship between a retrieval system and availability factors. A better retrieval system will make it possible to search a collection of sources in less time. The time gained may be used in an effort to conquer other availability factors, which may ultimately result in a bigger accessible base - from the user's point of view.

Availability factors define an accessible collection of legal sources in regard to the user. This collection will rarely be uniquely defined to, for instance, the content of the library of the user. The user may always expand the collection into new libraries or document files. But the expansion to new collections outside the user's "home collection" will entail a cost measured in time, money or otherwise, i.e. the effort required to conquer the availability factors. This "cost" will in practice set a limit to what sources the lawyer can afford to look into. In determining what he can "afford", the lawyer will, however, take into account several pragmatic factors, as for instance the importance of the problem he is confronted with. A common reason for increased effort is the existence of an influential or wealthy client. (Cfr. Operation Compulex 1972:7-8.)

In a computer-based retrieval system, the obvious availibility factor is the form of the legal source. Only machine-readable texts may be made part of the search file and available through the system. Consequently, the accessible collection of legal sources will - in regard to a certain computer-based system - be well defined.

Coverage is defined as the ratio between the accessible volume and the total volume of sources which may be relevant to the lawyer (Bing/Harvold 1974:103). What may be relevant to the user is in principle defined by the legal meta-norms. As these norms rarely are explicit, it is difficult to determine the volume of possible relevant sources. Even in relation to a certain legal area (criminal law, tax law, etc.), this will be difficult. But if coverage is related to a certain type of legal sources, for instance decisions of the Supreme Court, it will be possible to quantify the total. When also the accessible collection of sources is well defined - as in a computer-based system - an estimate of coverage in this limited sense may be given.

Our definition of coverage - though vague - is more objective than the one usually given, cfr. for instance McCarn/Stein 1967:111, who define coverage as "the extent to which the system includes a data base required by the user". In this case, the definition of the total volume of data is referred to the subjective requirements of the user. In our case, we propose that the total volume be determined by legal meta-norms, which may be vague, but which are defined apart from the user. Also, this makes it possible - as for instance in relation to types of sources - to give rather objective estimates of the coverage of the system. We believe

[Page 23 ]

The importance of the accessible collection of sources in any legal information system can hardly be overstated, cfr. Tapper 1973:78-80. It represents an upper limit of the quality of the service that this information system can offer to the user - the simple reason being that a source which is not part of the collection cannot be found, regardless of how closely you look. In a conventional system coverage is important - in a computerbased system it becomes essential. In a conventional system - as discussed earlier - the accessible collection is not well defined, and deficiencies in coverage may be more or less hidden from the user. A computer-based system declares exactly what its accessible collection of sources is - and consequently coverage deficiencies become more apparent. User research has also illustrated the importance placed on coverage by the user confronting a computer-based system, cfr. Brukerforsøk 1975:6-7.

1.2.8 The retrieval process

At this point, the lawyer has described his case through a set of proven facts. His problem is to find the legal sources from which he may arrive at legal norms relevant to his case. He has to go through a retrieval process. The sources he may retrieve are contained in the collection of accessible legal sources, which are more or less explicitly defined by the availability factors. To facilitate his retrieval, he may employ a retrieval system - which may be a computer-based system, but may also be a conventional system of indexes, reference works, etc.

A legal norm may be pictured as composed of two distinct segments: one segment defines the conditions for its use, the other segment describes the consequences of its use. This picture is clearly - and this point ought to be stressed - much too simple for a discussion of the nature of legal norms. It is, however, sufficient for our needs at the moment, as it demonstrates an important feature of norms: the conditions for their use correspond with the facts of possible cases. Whether a legal norm is brought into play or not is determined by the conditional segment of the norm, which qualifies a certain constellation of facts as a valid activation of the norm.

The legal sources contain the norms, and do consequently also contain the facts that activate the norms. The treatment of facts in the sources represents the bridge between the case at hand and the accessible collection of sources.

[Page 24 ]

In order to retrieve possible relevant sources, the lawyer must formulate a query. The query must be based on the proven facts of the case - these being the only part of the legal problem which the lawyer in principle knows at this stage of the decision process.

This query, which is a representation of the proven facts of the case, is matched with the accessible collection of legal sources. Where the facts mentioned in a source match the facts specified in the query, the source may be considered retrieved.

This matching process is often rather intuitive on the part of the lawyer; his memory effortlessly presents to him the possible relevant statutes or precedents. But usually a search is necessary. This may take the form of browsing through textbooks on the subject etc. - and in this case the query may remain vague and inexplicit, more or less as a question in the lawyer's mind. But more often the lawyer employs some sort of retrieval system which demands of him a more explicit formulation of his question. He may have to choose index terms at the back of a textbook, choose a systematic number to get access to a classified compilation of case law, or look up the commentary on a section of a statute. In all these examples the lawyer must formulate an explicit query which adequately represents his question within the limits of the retrieval system.

We stress this point just to point out that all retrieval systems imply the formulation of a query, regardless of whether the system are computerbased or manual. The accessible collection of legal sources is - as discussed earlier - of a syntactic nature. The questions in the lawyer's mind are of a semantic nature. In order to employ a retrieval system, a bridge has to be built from the semantic to the syntactic level. The query - which is a representation of the question in the terms allowed by the retrieval system - represents this bridge.

In a computer-based system the query does play an essential part. Such systems usually offer more possibilities when formulating queries than conventional manual system, and retrieval strategies - how to formulate adequate queries - become correspondingly more important, reflecting the possibilities of choice.

We do not, however, find any differences in principle between the use of a computer-based retrieval system and a conventional library system (i.e. an index). Both presuppose that questions are transformed into queries - the flexibility of the query language will vary from one system to another, but this will only account for a difference in degree. The principal difference is between systems where a syntactic query formulation is necessary and systems which do not demand this of the user. Regrettably few systems of the latter category exist

[Page 25 ]

As an illustration of the restrictions imposed on the user in formulating his query by a conventional retrieval system, we may take the precedents files in public administration. User research has shown that a number of these files are organized in accordance with the sections of the statutes. The user cannot in this case use the proven facts of the case at hand when formulating his query - he must formulate the query as a section of a statute. If he is mistaken in his choice of sections, he will not be able to find an identical case included in the precedents file (cfr. Bing 1974:114 with further references).

Matching the query with the accessible collection of legal sources, the lawyer makes a preliminary selection of legal sources. Once more we stress that the result is a set of sources - typically a set of written documents - not a set of legal norms. But by interpreting these sources, the lawyer may construct the legal norms (which are of a semantic nature). This interpretation may only be a trivial process like reading through the text. But in our terminology any understanding of a legal source (through reading or listening), presupposes an interpretation. (A "decoding" in the psychological sense, cfr. Rommetveit 1972:45, 71-89.) A closer characterization of "interpretation" will be made below under section 1.2.9.

During the retrieval process, the lawyer makes a preliminary interpretation of the legal sources in respect of the facts. Taking the description of facts found in the sources, the lawyer compares them to the proven facts of his case. Where a correspondence is believed to be found, the legal source in question is put aside as relevant.

This examination of the legal sources will take place regardless of the information system employed by the lawyer. In an interactive, computer-based system, this phase will correspond to the browsing on a display. The text of the retrieved sources may be displayed, and the lawyer may rapidly browse through them in search of possible relevant documents. Special features of the system - for instance highlighting or

[Page 26 ]

focusing, cfr. below in section 10.5.4 - facilitates the browsing. In this phase, the computer functions as a "reading glass".

In going through a set of legal sources, the lawyer will gain insight into the legal problems connected with his case. The browsing stage is also a learning stage. User research has illustrated that the lawyer may spend quite a bit of time at the terminal, and the explanation given is that the time was spent mapping the general legal background on which the case at hand was to be judged (Brukerforsøk 1975:25).

The feedback given by the system during the browsing stage may alter the lawyer's understanding of his problem. He may want to rephrase his query in order to reflect this change.² The new query may retrieve additional sources, which once again broaden the lawyer's understanding and lead to another rephrasing of the query.

² It may become necessary to go back to even earlier stages in the decision process - this is one of the dynamic aspects understated by our model.

In our model, the retrieval process is in principle an iterative process; the lawyer will pursue this loop until he is satisfied that all (or a sufficent number of) sources are retrieved.

The iterative nature of the retrieval process ought to be reflected by the design of a computer-based retrieval system, offering the possibility of modifying earlier queries and using them as part of a new query.

Whether the iterative nature of the retrieval process is revealed by a practical example will depend on a number of pragmatic factors. For instance, a simple legal problem may have such an evident solution that the lawyer simply retrieves the decisive legal source (for instance a section of the statute), and does not bother to map the borderlines of the norms he employs. This is often the case in public administration, where a great number of trivial legal problems are decided as a matter of routine. Also, the situation in which the lawyer finds himself is of great importance: again, public administration offers an illustration of a situation where the user is often faced with a time pressure which makes him settle for a sufficient set of reasons for his decision rather than a fully satisfactory set (cfr. Føyen/Harboe/Lie 1973:103-104).

It is important to be explicit about this aspect of the retrieval process, as the pragmatic factors exemplified above really may determine what is a satisfactory performance of the retrieval system. But these pragmatic factors may be opposed to the legal meta-norms which govern the whole decision process, and which demand that the lawyer in principle should examine all possibly relevant sources before a decision is made. In fact,

[Page 27 ]

a decision by the public administration may be declared void if the retrieval process has not been satisfactory. Also a private lawyer may have to pay damages to his client if his retrieval has not been satisfactory - cfr. for instance the Swedish case reported in NJA 1957:89, where a lawyer failed to find an important decision by the Supreme Court.

These examples illustrate that legal meta-norms play some part in establishing standards for what is considered satisfactory performance by a retrieval system. It is well known that the user situation - as determined by pragmatic factors - determines standards for the performance of a retrieval system, cfr. for instance Meadow 1973:173-175. But the normative factors are not usually discussed. This may be due to the fact that the normative character is lacking in most non-legal decision processes.

The relation between legal meta-norms determining what performance is satisfactory in a retrieval system, and the available retrieval systems, is complex. Probably there exists an interdependence between meta-norms and retrieval systems. The meta-norms determine what performance is satisfactory, but what is "satisfactory" is in turn partially determined by the systems available. This may be illustrated by the case reported in NJA 1957:89. A lawyer had given unsound advice, which resulted in his client suffering financial loss. The advice was founded on a section of a statute which the court found ambiguous. But this ambiguity had been decided by a case in 1904. The court points out that this case is not cited in the current compilation of statutes in force, but is mentioned in a well-known textbook. Therefore the court concludes that the lawyer ought to have had knowledge of the case. In regard to our discussion, it is of interest to note that the court, when deciding the culpability of the lawyer, discusses the retrieval systems available: the compilation of statutes and the text-book. The relationship between the meta-norms and the retrieval systems is thus uncovered. One may wonder what impact better retrieval systems may have on the legal decision process in this respect. A similar case, an appeal case in British Colombia (Regina v. James Bazzlo, 1971), is cited in Operation Compulex 1972:25 as an example of the lack of quality in legal research.

The lawyer, when deciding that one fact among his proven facts is relevant, bases his opinion on a legal norm which includes this fact. But norms may only be formed on the bases of legal sources, and to find these the lawyer must construct a query in which he selects facts of legal relevance. At this stage he does not know whether the selected facts really are relevant - but must rely on his background knowledge. Consequently this background knowledge is essential. It is necessary both in order to select facts for constructing the query and for formulating the query, i.e. selecting the words that are most likely to represent the facts in the accessible documents (or an index to these documents). A lack of background knowledge will only accidentally be remedied by an effective

[Page 28 ]

retrieval system. Therefore a computer-based system too will serve the knowledgeable lawyer better than the ignorant one. And laymen, lacking legal expertise, will be more or less helpless when confronted with such a system. An information system designed to give legal information to non-lawyers must be constructed on other principles, including for instance a sub-system capable of problem analysis, a process which in our model is pictured as manual. Cfr. McCoy/Chatterton 1968, Seipel 1975:268-269, Bing/Harvold 1973:11-15.

One may note that this analysis of query construction point to one of the major problems in computer-based text retrieval as experienced by the user - the problem of specificity. When constructing the query, the lawyer uses the proven facts of the case. But in the legal sources, the facts may be represented by words different from those one would be inclined to select when describing the facts of the case. If the case concerns a cow, a legal source of a general nature (for instance a statute), most probably will not use the word "cow", but rather a more general term like "domestic animal". In a legal source of a casuistic nature (for instance case law), the fact may be represented by the word "cow", but equally well by terms denoting other domestic animals, like "horse" or "ox". This illustrates that background knowledge not only of the possible relevant legal norms, but also of the legal language is of importance. We shall return to this problem below in section 10.4 where retrieval strategies for overcoming this difficulty will also be discussed.

Subsumption. In Norwegian law, the term "subsumption" is often used to describe the classification of a certain set of facts by a certain section of the penal code. In our terminology, subsumption corresponds to a certain stage of the decision model: the stage where the lawyer has selected the relevant legal sources and interpreted these with regard to the legally relevant facts. This terminology is not completely identical to that tradionally given - the difference mainly derives from our strict distinction between legal sources (a section of the penal code) and a legal norm (the legal norm based on the understanding of that section in its full context).

We have now described the retrieval process - including query construction, the utilization of feedback information, etc. When our lawyer is through with this process, he is left on the brink of the "real" decision - which is mostly a mental process. In this mental process he processes the information of the facts and of the norms, arriving at a decision. We shall describe the rest of this process, and since our aim will be only to outline the process of which the retrieval process is a part, we shall not go into every detail.

[Page 29 ]

1.2.9 Interpretation and harmonization

During the retrieval process, the lawyer interprets his legal sources in relation to the facts of his case. This "partial interpretation" may be viewed as a partial development of the legal norms latent in the retrieved sources. The task of the lawyer is now to develop these further into fully interpreted legal norms relevant to his case.

The interpretation is governed by meta-norms. The lawyer is taught as part of his education, what sort of argumentation is considered valid when interpreting legal sources. Some of these meta-norms may have found an explicit formulation - for instance the maxim of ratio decidendi in relation to case law - but most remain rather implicit.

We will not try here to describe these meta-norms, but will give an example of two different methods used by the lawyer when interpreting a single word (or expression) occurring in a formulation of a legal norm, for instance in the text of a statute.

A standard formulation of a legal norm is the formula "If A, then B" - A giving the conditions for the legal norm being relevant, B describing the consequences of the norm being relevant. Let us take an imaginary example of copyright law, where a norm prescribes that "If book, then B".

"Book" is a word with a core of well-defined content, but with vague limits. It may - at least in Norwegian - refer to rather different objects. An ordinarily bound novel is of course a "book". So is a paper-bound edition or a tome in a series. But where should one draw the line between "magazine" or "a collection of loose leaves" and "book"?

The problems of interpretation raised by words like "book" may be traced back to the nature of language itself. Our language does not as a rule use words as well-defined semantic elements, but rather as nodes in a semantic-associative network. The interpretation of the word is influenced by the total context as well as the background knowledge of the interpreter. A well-known example illustrating this aspect of words is the different meaning one will tend to read into the word "man" when combined with another word (cfr. Rommetveit 1972:64):

• "man" and "animal"
• "man" and "woman"
• "man" and "boy"
• "man" and "son"

The lawyer will, when interpreting this word, mainly follow the metanorms for interpretation of language generally accepted also outside the legal profession.

[Page 30 ]

This method may be contrasted with a formulation of a legal norm containing a word like "culpable", sampled from an imaginary statute on torts: "If culpable, then B". The word "culpable" represents a node in the semantic-associative network of the lawyer, and it is possible to interpret this word following more or less the same meta-norms as described above. But this would be incorrect according to the legal meta-norms governing the interpretation. To a lawyer, the word "culpable" (at least in the Norwegian legal system) is a reference to a set of legal norms qualifying certain acts as culpable. These norms are of a fragmentary nature, and may for instance be pieced together from a great number of different cases.

This is evidently a different matter than following the meta-norms determining the interpretation of a word like "book", as the interpretation is not determined by common use of the word, but by legal norms based on the interpretation of legal sources apart from the one containing the original formulation of the norm. The norm cannot be interpreted by itself, but only as part of a legal system.

These two examples are of some relevance also in regard to legal information systems. When one is retrieving a legal source containing a formulation of a norm, the words in this formulation do not by themselves tell the user whether they are to be interpreted according to the common use of the word, or whether they are just departure points for further forays into the twin worlds of legal sources and legal norms. Even trained lawyers may discover this only after working with the problem for some time. And to a layman, the distinction would be problematic indeed.

Actually the interpretation process may itself contain problems justifying information retrieval. The query is then constructed of the problematic words, and sources clarifying the interpretation of these words are retrieved. In our model this aspect is understated as secondary to (or part of) the process where queries are constructed on the basis of the proven facts of the case at hand. A distinction in principle between the two causes of igniting a retrieval process is hardly possible. We are satisfied to keep our simple model, but would like to point out that the interpretation process itself may be expected to be the driving force behind the iterative process of information retrieval in many instances.

The discussion of interpretation above is, of course, insufficient. As for the interpretation of words as part of a communication, see Rommetveit 1972:71-89, who describes the

[Page 31 ]

semantic-associative network. The example of the word "culpable" igniting a process involving a further set of legal norms is in itself just one example of what has been described as the "fragmentary structure of legal norms". In this structure, the type of norms called "guidelines" have an improtant place - cfr. Sundby 1974:190-306. We feel, however, that for our purpose we may leave out a discussion of the nature of norms themselves. Our example does only illustrate the normative aspect of interpretation in a legal decision process, which may easily cause the lawyer to arrive at a content of a word distinctly different from the content the same word has in common use.

Through the interpretation, arguments are selected from the legal sources in order to arrive at the legal norms relevant to the case. During this process, the lawyer may discover that two legal sources may contain arguments for diverging legal norms. There are, for instance, two cases that seem to disagree on the interpretation of a statute. In this instance the legal sources (the statute combined with one of the two cases) may serve as the basis for two diverging norms. In such an instance, harmonization of the legal sources is necessary.

Another situation may also demand harmonization of a different order. The lawyer has terminated his interpretation and discovers that two of the relevant norms that he has arrived at, are in conflict. In this case, harmonization of the legal norms is necessary.

It is possible - and probably appropriate in many cases - to distinguish between these two types of harmonization. But for our purpose it does not seem to be essential. Our point is that as part of the interpretation, or following the interpretation, it becomes necessary to harmonize the legal norms. Without harmonization, the norms relevant to the case at hand are conflicting (or at least diverging); consequently the lawyer will have difficulties in arriving at a decision.

A curious example may be found in the rather complex Norwegian legislation governing the sale of alcoholic beverages. The sections 14 and 21 of the Norwegian "Spirits Acts" (5.4.1927) state that alcoholic beverages are not to be sold to persons below the age of 18. An older statute, however, of 31.5.1900 No. 5 states in its section 23 that an exception may be made when the beverage is served as refreshment with a meal of during travelling. Two conflicting legal norms may be derived from these two sections. The conflict is resolved by using the principle "lex posterior derogat priori", the norm derived from the most recent of the statutes is given priority (cfr. letter from the Ministry of Justice, 26.7.1974, ref. 1925/74 E TS/AV).

There exist meta-norms governing this sort of harmonization. Some of them are commonly known as maxims, as for instance the "lex posterior" principle exemplified above, and the other classical principles of "lex superior" and "lex specialise But even these are rather vague, and the meta-norms governing harmonization have as a whole not been very well

[Page 32 ]

analyzed - at least not in Norwegian theory (cfr. Aarbakke 1966:506).

Most of the harmonization norms are founded on a ranking of types of legal sources. A legal norm derived from a source of higher rank is given priority in a case of conflict. Such a ranking of types of legal sources will usually place the Constitution at the top, and proceed through statutes enacted by the parliament down to case law and regulatory law. The details of the ranking will certainly be relative to the legal system, and even in regard to one legal system, the ranking may be relative to different user groups.

We have several times used the expression "conflict of norms" without describing or defining this expression. What really is a conflict of norms, depends to a great extent on the nature of norms. For our purpose it is sufficient to use the expression "conflict of norms" as a characteristic; a discussion may be found in Eckhoff 1971:270-305 and Sundby 1974:278-281.

In the process of harmonizing the interpretation of legal sources and diverging legal norms, the relative rank of the different types of legal sources is essential. Actually this relative rank is a controversial matter - at any rate in the Norwegian legal system. The relative rank of, for instance, decisions by a court of first instance is a disputed matter.

It is important to stress that the matter of determining relative ranks of the types of legal sources has two aspects.

Firstly there is the normative aspect: what sort of rank should a certain type of legal sources have according to the meta-norms of a certain legal system? This question is complex, as a single type of legal sources may have different rank relative to the user group or the overall decision situation. "Administrative opinions" do, for instance, certainly have a different rank for lawyers within the administration compared to the rank given by a court. As these meta-norms are rather complex, and also to some degree neglected in the literature, they remain vague and controversial.

Secondly there is the empirical aspect: what relative rank do lawyers (or different user groups) give to a type of legal sources? This may in practice be discovered through empirical research, for instance a questionnaire. This has been done in Western Germany - where a table of 17 possible "information sources" were presented to a number of lawyers.³ The "information sources" were given weights between 0 and

³ "Information Sources" - "Informationsquellen" - do not correspond to "type of legal sources", but rather to "publication". The lack of distinction between "legal sources" and "publications containing one or more legal sources" does, in our opinion, to some degree reduce the general usefulness of the survey. This survey is described by Jungjohann/Seidel/Sorgel/Uhlig 1974, and a summary is given by Uhlig 1975:343-344.

[Page 33 ]

5, 5 being the highest score. Actually legal journals ("Fachzeitschriften") scored highest (4.42), followed by "commentaries to statutes" ("Kommentare" - 4.26), and updating of compilations of statutes ("Ergänzungslieferungen zu Gesetzsammlungen" - 4.25). Great differences between user groups were revealed - monographs were given rank 4 by professors (after legal journals and textbooks), while lawyers in the tax administration etc. gave this rank 15.

Clearly there is an interdependence between the normative and the empirical aspect. Firstly, the fact that a type of legal sources is given high rank by the meta-norms will certainly make lawyers inclined to attach great importance to this type in practice. But the vagueness and complexity of these meta-norms may also make them sensitive to feedback - the fact that a lawyer finds a type of legal source useful, will help to mould the meta-norms. The interdependence can only be seen when changes occur - and they are most likely to occur in the empirical aspect, for instance where a new series of case reports makes a type of legal sources more easily available than before. This type will become more useful, and this higher degree of usefulness may cause an upgrading of the relative rank given to this source by the meta-norms.

Our own research has given us reason to believe that the information systems available (in a broad sense, comprising all publications, routines, etc.) are important in determining the practical value of a type of legal source. The German survey is in fact an indication of this, being related to publications rather than types of legal sources.

A better information system will probably cause the types of legal sources benefiting from this improvement to become more useful. They will be cited more frequently by the lawyer - and we suspect that the meta-norms of the legal system will change, upgrading the relative rank of these types. Computer-based retrieval systems will represent a drastic improvement with regard to the legal sources included in the data base. Consequently this may influence the legal meta-norms themselves governing the harmonization of the interpretation of legal sources and diverging legal norms. The possibility of impacts on the legal systems of this kind should be considered when assessing the desirability of computerbased retrieval systems. (In part IV we will revert to this and related questions.)

We have now discussed some aspects of the process of interpreting legal sources, the harmonization of these sources as part of the interpretation, and the harmonization of possible divergencies in the norms arrived at

[Page 34 ]

through the interpretation. The aspects we have discussed, are mainly aspects of interest in relation to legal informatics. But even with this limitation, we think the outline of the process has emerged.

The process leaves the lawyer with a set of legal norms which are not in any way in conflict with each other. These legal norms are relevant to the case at hand - and we might think that a simple combination of the norms with the facts of the case would provide the lawyer with his decision. This would also be an adequate way of portraying it in regard to simple decisions.

However, we think this description may be too simple. The final legal norms are defined by the sources found relevant, interpreted and harmonized according to legal meta-norms. As we have stressed a number of times, these meta-norms are vague. Also, the legal norms themselves may be vague - norms allowing "discretion" may often give room for different solutions (even when the lawyer is through with his interpretation and harmonization). We find it generally accepted that the result may not be well-defined, but may rather be represented as a normative interval - or, as Stone (1968:192 and 320) puts it: there still exists discretion within "the leeways left by the guides of law".

Some main causes for the existence of such a "leeway" may be listed: (1) Reasonable disagreement on what sources are to be qualified as legal sources. (2) Reasonable disagreement on the interpretation of legal sources, causing reasonable disagreement on which norms are relevant. (3) Reasonable disagreement on the priority of diverging norms. Instances of such "reasonable disagreement" may explain why two or more reputable lawyers may arrive at different decisions even though agreeing on the proven facts of the case.

1.2.10 The selection of norms within the normative interval

The lawyer, when confronted with a normative interval, cannot arrive at a decision without selecting one of the alternative norms in the interval left by the "guides of law". This selection is, of course, not a random process. It is also governed by norms and meta-norms, but of an extra-legal character. An important aspect of the legal decision process is, in our belief, that it integrates the use of extra-legal norms. This aspect is usually trivial, but in controversial questions, where little support can be found in traditional legal sources and, consequently, the normative interval is so broad that it includes distinct alternatives, this aspect may attract the attention of the public.

It falls outside the scope of our book to examine this selective process

[Page 35 ]

in any detail. We will just sketch a few methods by which the lawyer may select his norms.

(1) As part of the extra-legal norms there may be meta-norms of a nature closely related to those governing the harmonization process. Such norms may enable the lawyer to give one of the alternative norms priority over the others, and this norm is the one selected. An example of such an extra-legal meta-norm would probably be "client loyalty" - the lawyer selects the norm most favorable to his client. Another example may be the known political intention behind a certain statute, following this the lawyer would give priority to the alternative most closely corresponding to the intention.

(2) As part of the extra-legal norms, there may exist norms of a nature closely related to the legal norms, i.e. norms with a segment stating the conditions for their use, and another segment stating the consequences of their use. The conditions are specified, as in the case of legal norms, as a set of facts. The lawyer consults the facts of his case to see if any of these extra-legal norms are relevant. This may cause him to select additional facts not pointed out as relevant by any of the legal norms. Using such extra-legal norms, the lawyer will be able to constrict his normative interval - maybe to such a degree that only one alternative is left, which consequently is selected.

The extra-legal norms employed by a lawyer in this manner may be of a political, moral, or ethical character. It ought, however, to be stressed that the lawyer is not free to choose whatever prejudice or ideology he personally believes in. There are legal meta-norms governing what sort of extra-legal norms the lawyer is allowed to take into consideration, even when he is selecting just one alternative within the normative interval (diverging opinion, Kilian 1974:228).

The use of extra-legal norms as part of a legal decision process is of interest with respect to information systems. The lawyer finds his legal norms through an interpretation of legal sources, and often employs a retrieval system in order to find these sources. What about the extra-legal norms - are they found in a similar way, based on a set of extra-legal sources? We let these questions remain posed, but not answered, partly because of our conviction that the extra-legal norms are not arrived at through a process governed by meta-norms similar to the legal decision process, but are more intuitive and also more subjective.

1.2.11 The result - and feedback from the result

The lawyer has now arrived at the relevant norms, which, combined with the facts of the case at hand, give the result.

[Page 36 ]

Our model has not allowed for the effects of the result playing any part in the reasoning of the lawyer up to this point (excluding the indirect way in which this may have determined for instance the burden of proof etc.). The effect of the result - or the reasonableness of the decision - is, however, a legal source in its own right. In our model, this legal source is represented as a feedback loop. The lawyer evaluates the result according to his extra-legal value norms. His evaluation is then taken into consideration as a legal source, and as such will to some degree determine the normative interval. This feedback may cause a revision of the normative interval and a corresponding revision of the result. An iterative process is initiated, which only comes to a stop when the feedback cannot further influence the normative interval.

It may be noted that the "evaluation of the result" is a legal source of a qualitatively different nature from the other legal sources. We have several times stressed that the majority of the other legal sources are of a syntactic nature. The "evaluation of the result" is of a semantic nature, being the evaluation in the mind of the lawyer. It has, however, one aspect in common with the other legal sources: its influence on the normative interval.

Another typical aspect of the "evaluation of the result" may also be noted - which represents a typical, though not principal, difference with regard to the other sources: the evaluation can only be carried out in respect of a given case with a given result. It does not exist apart from the case, but is created by the case itself. Consequently it cannot be "retrieved" from any data base established prior to the case.

We believe the feedback loop represented by the "evaluation of the result" to be of great interest, also in regard to legal informatics. It is one of the major illustrations of the iterative nature of the legal decision process - an iterative nature which most automated decision processes have not been able to represent or have just ignored.

It may be appropriate once more to stress that our model is just a model of the stages in a legal decision process, describing the relation between these stages - but not the psychological process itself. Certainly a lawyer will have selected his result at a far earlier stage of the process than represented in our model. He will select what Soelberg (1967:23 and 26) has named a "choice candidate", and in practice the lawyer's activities may be wholly concerned with justifying that his choice candidate may be arrived at without violating the meta-norms governing the legal decision process. Cfr. also Eckhoff 1971:29, who discusses what comes first in the mind of a judge: the result or the reasoning justifying the result.

[Page 37 ]

The legal decision process is a formal process; it is governed by metanorms to a greater extent than decision processes within other areas. The meta-norms prescribe how a lawyer is allowed to argue for a certain result. These meta-norms are admittedly vague and leave room for disagreement even between lawyers, but they are nevertheless binding. If a lawyer violates these meta-norms - or rather, if such a violation is found to have taken place - his decision may be declared void. The meta-norms mostly demand that a case be decided on what hashappened, and according to legal norms that were in existence at that time. It is a "retrospective process" in which the lawyer most of the time concentrates on a situation from the past.

The "evaluation of the result" is an escape from this retrospective perspective. Legal meta-norms allow the lawyer at this stage of the process to look at the present and even into the future, asking, "What will be the effect of my decision?"

Actually this represents a safety valve in the legal decision process. Through the "evaluation of the result", the lawyer may make his decision more oriented toward its consequences.

An illustration of this was noticed in one of our surveys of the Social Security Administration (Bing/Harvold 1973:228). The Social Security Administration is mostly staffed with civil servants without formal legal education. They have more often been trained in a medical or social environment, and are used to thinking in terms of the health and welfare of their clients. A medical decision, for instance, is usually oriented toward its consequences: if the patient gets better, the ordained cure was "correct" - even when selected by intuition alone and in disagreement with the opinion of an authority. Not so in a legal decision, even if the decision makes the client happy, the decision is invalid if in conflict with a statute.

The bridge between the "medical" and the "legal" decision models is the "evaluation of the results". We found that this legal source was given higher rank within the Social Security Administration relative to arguments deriving from sources like regulatory law. When the rank of the "evaluation of the result" was upgraded, the gap between "medical" and "legal" reasoning was reduced.

This observation may have some relevance to legal informatics. A better legal information system may - as we have mentioned earlier - result in a displacement of the established relative ranks of legal sources. If a better legal retrieval system was established within the Social

[Page 38 ]

Security Administration, this might result in the upgrading of conventional types of legal sources, like administrative decisions. It would correspond to a reduced rank for the type "evaluation of the result". As this type at present seems to serve as a sort of bridge between two dissimilar decision processes, i.e. the "medical" and the "legal", the reduced rank might break the bridge. The possible consequences of launching a better legal information system, and through this creating a dynamic situation with the implications sketched above, should not be underestimated.

1.3 A NOTE ON RELEVANCE

Throughout the previous sections, we have used the term "relevance" without giving a definition of the term. With respect to legal retrieval system, the meaning of "relevance" holds a key position, as the standard measures of system performance, recall and precision, rely on a relevance assessment. We shall revert to this below in section 9.4. Here we will only make a few comments which tie in with our model of legal decisions.

In relation to the model, a "relevant legal source" may be readily described. In our model, a decision is based on legal norms. When arguing for the existence of a certain valid legal norm, the lawyer must base his argumentation on legal sources. The legal sources used as a basis for his argumentation in the case at hand represent the relevant legal sources in that case.

This description makes the term "relevant" relative to both the lawyer and the case at hand. It corresponds to definitions set out by Saracevic (1970b:671): "any document which on the basis of the information it conveys, is considered to be related to your (user's) question even if the information is outdated or familiar" and by Lancaster (1969:643): "a 'relevant' document is nothing more nor less than a document of some value to the user in relation to the information need that prompted his request".

The user-relativity of the term "relevance" is, however, somewhat reduced in a legal decision system by the meta-norms governing that system. As we have shown by examples (the Swedish case discussed above in section 1.2.8), the legal meta-norms do themselves point out certain sources as "relevant". Even though the user has neglected to base his argumentation on a certain document, a court or a superior instance may qualify that document as relevant. Tendencies toward similar

[Page 39 ]

meta-norms may be found in most decision processes - for instance a thesis within the field of psychology may be criticized because the author has neglected to take into account certain experiments. But within legal decision systems, these meta-norms are rather less vague, and are in fact enforced through courts, etc.

A definition of relevance in a legal decision process should therefore take these meta-norms as its point of departure. On the other hand, it should leave room for the discretion of the lawyer. We think an adequate formulation would be:

"A legal source is relevant if -

• (1) The decision would be different if the source were not taken into account, or
• (2) Legal meta-norms demand that the user consider whether the source belongs to group (1), or
• (3) The user himself deems it necessary to consider whether the source belongs to group (1)."

This definition does to a certain extent reduce the subjectivity of the term "relative". For instance, two lawyers may disagree on whether a case belongs to group (1) - one maintains that this is decisive in respect of some point in the case at hand, the other strongly denies this. This is a disagreement on whether the case belongs to group (1), but they may agree that the case belongs to group (2) or (3). Consequently they may agree upon the relevance of the case - though not on its relationship to the result.

It may be noted that our definition diverges from our description. A case considered by a lawyer, but which in the end is put aside, is relevant according to our definition. But in our model this source would not be used as a basis for the legal norms employed by the lawyer in order to arrive at his decision. The definition allows for a more lenient understanding of the term "relevance".

We think this is necessary for our purpose, as we are mainly concerned with retrieval systems. If a legal source is retrieved, which the lawyer then spends some time pondering over only to arrive at the conclusion that it is of no consequence in regard to the case at hand - we should like to say that this is a relevant source. The system should furnish the lawyer with material bordering on what is necessary - the lawyer is the one to draw the line, not the system. Cfr. Saracevic's definiton, where it is stated that a document is relevant also when "outdated or familiar".

Experiments have demonstrated that lawyers often disagree on the

[Page 40 ]

"relevance" of a source law. In the Joint American Bar Foundation and IBM Project the panelists - four experienced lawyers - disagreed upon the relevance in 65.7 per cent of the instances, in 11 per cent they even disagreed on whether the source was "on point" or "irrelevant" (cfr. Eldridge 1968:43). Similar results were arrived at by Tapper (1973:121). He had a number of lawyers describe a set of cases in accordance with a defined set of indexing terms. There appeared to be such a lack of consistency between different lawyers indexing the same case, that he had to drop this approach. This lack of consistency is not restricted to lawyers, cfr. Saracevic 1968:116-129.

Relevance assessment may also vary over time. Tapper has demonstrated that the same case might be described differently by the same lawyer when presented in another context (1973:121). Saracevic (1968:119) found intra-indexer consistency not much better than interindexer consistency. We refer to these results in order to stress the relativity of "relevance". Also, using our definition, relevance will be something relative to both user and the case at hand. The assessment of a certain source, may also vary with time.

But even admitting the relativity of relevance, it may be pointed out that the disagreement uncovered by experiments may in part be due to lack of distinction between two different concepts - the concept of priority and the concept of relevance as introduced through our model.

According to our model, a legal source can only be relevant or not relevant. A legal source is of a syntactic nature. If it is relevant then this is due to the fact that the lawyer can use it when arguing for the existence of a certain legal norm. Whether this norm - or the contribution to this norm originating in the source - is decisive, is another question. A document should not be confused with the content of this document. A norm may through the complexities of the interpretation process be constructed by arguments originating in a number of sources. The priority of the resulting norms should not be confused with the relevance of the sources used as a fundament.

Consequently, a legal source may be found relevant with respect to the argumentation - the priority or weight of this argumentation is something else. This dichotomy is well known from literature, but often overlooked.

Evidence of the fact that it is overlooked, is the relevance scales used in the American Bar Association's Case Law Retrieval Project ("on point", "relevant", "related" and "irrelevant") and Saracevic's project

[Page 41 ]

("relevant", "partially relevant" and "nonrelevant"). Such a scale presupposes a grading of relevance. This seems to indicate that the understanding of the term "relevance" in these experiments also included what we have called "priority" (or the weight of the argumentation based on the document).

We believe it is much more difficult - even within the structure of a formal decision process - to agree upon priority, than to agree upon relevance. This may also partially explain the high degree of disagreement found in relevance assessment or indexing in these experiments.

[Page 42 ]

2 Typical user situations and information systems

In this chapter, we shall very briefly consider some aspects of typical user situations related to the choice of information system. We will restrict our discussion to user situations found within legal decision systems - which, as we have stessed above in chapter 1, are to a great extent governed by meta-norms. We will relate user situations to user groups - but this will of course only provide examples, as all lawyers will to a certain extent be in a complex user situation where elements of all three typical user situations are present.

2.1 USERS ORIENTED TOWARD NORMS

The traditional image of the lawyer depicts him against a background of legal textbooks and case reports. He is presented with a problem - and he solves this problem by applying the law, by applying the legal norms relevant to the problem.

This image does correspond to the typical situation of the lawyer in private practice who is consulted by a client with a legal problem. The lawyer advises his client by describing the relevant legal norms and their consequences when applied to the client's problem. In this situation, the lawyer's research problem is primarily directed toward the legal sources, through which he may gain insight in to the legal norms.

The image also corresponds to the typical situation of a judge, who is presented with conflicting views of the facts of the case as well as the applicable legal norms by the parties. The judge has to decide the case by applying the law, as he sees it, to the facts of the case, as they appear to him. His research problem will primarily be directed toward the legal norms - he has to examine the legal sources himself in order to control the interpretation and to ensure that important sources of law are not overlooked.

[Page 43 ]

The legislator is not usually considered to be in this situation - nevertheless, when discussing the need for a reform or when drafting new legislation, he will often have research problems directed toward the legal sources. He will want to know the current legal situation, and he may also want to get a picture of the legal norms surrounding the one he is about to revise. His need for information will, however, be of a more general nature. This really means that the legislator is not typically restricted by all the facts of a specific case, but has a more generalized factual situation in mind. Such a situation is not, however, unfamiliar to the private lawyer who is hired by his client to give a general picture of the law within a certain field, for instance as planning information in a new business venture.

The civil servant will often be in a situation similar to that of the lawyer. He will also have the same need of finding the relevant legal norms. But his situation is typically different - characterized by a great number of similar cases in contrast to the lawyer's more personal relationship with individual clients.

Users in this situation are oriented toward legal norms. As we have discussed in chapter 1, legal norms are constructed on the basis of the information found in legal sources. The research problem of the lawyer is consequently directed toward legal sources.

Legal sources are mostly composed of texts - statutes, cases, regulatory law, etc. When selecting an information system, the lawyer must choose a system capable of coping with the text.

Another typical feature of the user situation must also be dealt with: the way queries are constructed on the basis of the known facts of the case. The lawyer will be confronted with endless variations and unexpected circumstances. The information system must be flexible enough to cope with this - the lawyer will want to be able to construct general as well as specific queries.

The aim is to find legal sources relevant to the problem. The lawyer does not hope to find an answer to the problem in the sources, but only the ore which he can refine into legal norms.

This type of retrieval situation may be termed reference retrieval: through the query the lawyer obtains reference to a text dealing with the problem, and by understanding the text he comes somewhat closer to solving his problem.

[Page 44 ]

2.2 USERS ORIENTED TOWARD FACTS

As illustrated by the model of the legal decision process, the lawyer has to uncover the facts of the case. Establishing what has happened, what the probable or provable facts of the case really are, will in some instances be a major enterprise.

The lack of facts will, however, have different consequences for different user categories. If the client furnishes the private lawyer with insufficient facts, the lawyer may refuse to take on the case - or he may attack the problem at his client's risk. In a civil court case, the parties may agree upon certain circumstances and through their agreement confine the problem to certain disputable issues. (They may disagree on whether a certain sum of money has been borrowed, but agree that if it was borrowed, the claim is not prescribed.)

Other user categories have a more accentuated need for facts. A typical example is the civil servant dealing with a claim by a citizen. In order to settle the case, he will have to take certain facts into consideration. When settling a claim for a disability pension, the civil servant must assemble a number of different facts: the claimant's income, medical information, information on dependent persons, education, retraining possibilities, etc.

The civil servant is not at liberty to dismiss the case (though he may, in extreme situations, put the case aside because of the unwillingness of the citizen to provide information). The civil servant himself has to uncover a great deal of the relevant facts - through contact with other authorities, employers, etc. The legal norms regulating his decisions will usually require him to show zeal and initiative.

The private lawyer too will have a great need for access to certain fact retrieval systems, for instance real property registers - cfr. Operation Compulex 1972:31.

The judge may also be in a similar situation - though the contradictory principle and the more elaborate procedure will usually guarantee that the parties themselves provide the necessary evidence.

Users in this situation are oriented toward facts. They are usually able to describe the nature of the facts in rather precise terms, like "annual income", "number of dependents", "nature of present job", etc.

When a comparison is made with the user situation described above in section 2.1, we find typical differences. The user's ability to give a rather well-defined description of his information need is one such difference. The user's possibility of assessing the relevance of the result of his

[Page 45 ]

research is another: often he has decided on the relevance before starting his research. If he is looking for information on the annual income, and he finds the figure representing this income, the relevance assessment is trivial or non-existent. This is in contrast to reference retrieval, where a retrieved text has to be understood - often through a complex interpretation - before its value can be assessed. In fact retrieval, a negative result may also be of more use: in investigating whether a person is married, the civil servant may look up this information in the Central Personal Register (a matter of routine in many administrative systems). Finding no mention of a wife or a husband, he may safely conclude that his client is not married. In reference retrieval this sort of antithetical conclusion is rarely justified: the fact that the user has not found any statute law dealing with a certain topic may just as well be explained by an insufficient query as by the non-existence of such statutes.

Fact-oriented retrieval may also be directed toward legal sources, though this will hardly be typical. Examples are queries concerned with syntactical qualities of the sources - for instance sections of statutes with an explicit reference to a certain statute, or all documents of a certain type of legal sources, for instance all decisions by the Supreme Court. This last example illustrates how fact retrieval may be used to cut out segments of a greater data base preceding reference retrieval.

Our example of fact retrieval directed toward syntactical features of the legal sources should not be confused with reference retrieval, where syntactical features (usually words) are used as indications of the probable existence of a certain content.

The differences in fact and reference retrieval will be of some consequence when designing information systems. As queries in a fact retrieval situation will usually be more well defined, the information system may also reflect this through a more well-defined data structure.

A more pragmatic difference may also be pointed out: fact retrieval will in a conventional user situation represent a sort of "retrieval sub-routine": the user will grab his phone to get information on the annual income, write a letter to a hospital requesting a diagnosis, etc. The introduction of a new information system in this environment will be facilitated by the established retrieval sub-routines: instead of making a telephone call, the user will pose his query through a terminal to a computer-based system. A user experiment in the Norwegian Social Security Court illustrated just this point. The clerks of the Court as a matter of routine changed from telephone inquiries to terminal queries when collecting information on the income of the applicants. In reference retrieval, sub-routines are usually not established in the same way. Though the user may use retrieval systems such as indexes or textbooks, he does not recognize them as such. A computer-based retrieval system stresses the importance of the query. The problems of learning a search language and, more important, the limitations of a computer-based system become obstacles in the way of rapid change.

[Page 46 ]

2.3 USERS ORIENTED TOWARD FORMS

The modern welfare state leans rather heavily on bureaucratic administration, which handles a great number of cases. In order to obtain an effective administration, specialized branches dealing with the different sectors of society are established. Norway, being a typical "welfare state", may serve as an example. Major administrative bodies are established to handle tax cases (mainly assessment of the tax returns), social security cases, social welfare cases, there are public employment offices, etc.

The majority of the cases processed by these bodies are of an uncomplicated nature. There are no difficulties in establishing the facts of the case, and the legal norms are well known to the experienced civil servant handling the case.

A large volume of stereotyped cases may lead to the establishment of forms, thus easing the work of the civil servants. The form makes fact collection easier: if the citizen himself fills in the form, it will show him what facts are needed to settle his case. The civil servant can limit his effort to checking dubious facts. The forms also structure the facts according to the legal norms relevant to the type of service requested by the citizen.

Seen from this angle, a form represents some sort of information system. It conveys information on legal norms to both the citizen and the lawyer. On the one hand, it specifies what facts are needed, on the other, it relates these facts to the relevant, legal norms, and in some cases the forms assign weight to the facts, making the decision more or less the result of a summation of pros and cons. Actually a form may be - and in several instances has been - developed into a decision table.

A form does as a rule partially serve as a legal information system. When the client approaches him, the lawyer picks out the appropriate form. The form is selected not on the basis of a thorough analysis of the problem as described above in section 1.2.2.-5, but rather on the client's own preliminary claim or characterization of his problem. This is possible because the client's problem is typical. The mass of similar cases has made it worthwhile to develop a form, which sets out what facts are required to decide the case according to an analysis and a structuring of the possible relevant legal norms. Consequently, the lawyer does not have to consult the legal sources themselves, but may trust the form to guide him to the correct decision. The great number of similar cases will gradually provide the lawyer with considerable expertise in handling the cases and the forms, and this expertise (or background knowledge) may guarantee

[Page 47 ]

a sensible application of the forms - borderline cases or special cases will, one hopes, be treated in an appropriate and non-standardized manner.

In our opinion, the use of forms within the administration in legal decision processes is rather fascinating. In this section we are only concerned with the form as a legal information system, but other aspects are also of great interest. We may point out that a form - to a certain extent - is a representation of the legal norms. It is constructed through an interpretation of legal sources. This interpretation is then represented as a form. But the form itself is of a syntactic nature, and must be interpreted once more before it can be used. As the form is secondary in relation to the legal sources, this new interpretation may bring about a false understanding of the legal norms. The forms are a mirror reflecting the legal sources, but there may be flaws in the mirror. This problem will be relative to the user of the forms: if the user is a lawyer, the interpretation of the forms can be checked by the background knowledge of the lawyer. But if the user is the client himself, the brief text of a form may often induce doubt and misapprehension. An example of how the norm may be distorted by a form may be given. In Norway, social security benefits to single supporters are granted only when the supporter does not share residence with the other parent of the child. "Sharing residence" is probably to be interpreted as "living with", though the forms use a phrase denoting "living in the same house". Consequently social security will not be granted to single supporters living in the same house as the other parent of the child, even when the house happens to be a 13-floor apartment building, and the persons in question occupy separate flats.

We stated above that a decision table would be one extreme result of a form. A more refined example would be a "computer-based form". The computer could, of course, store a form more complex than can conveniently be represented on paper. Facts could in part be collected through the application of the client, but might also be collected from other files - by automatic routines - for instance the files of the central tax administration.

Examples of dynamic, computer-based "forms" are emerging, and may rather be viewed as automated decision systems ("deontic systems") than as information systems. It is, however, of interest to note this change from legal information systems into what has been called "radical computer use in law" (Slayton 1974c).

In Norway the foremost example of such computer systems is the housing grant system - where the applicant through a form activates the system, which then collects data from a number of registers and completely automatically combines them to make a decision. For a discussion of this system and some of its implications, see Blymke 1974 and NORIS (13).

These systems also exist, at least in principle, in more sophisticated versions. An early example is Lombardi's (1969) "Rules of the Road"-system, a more advanced example is Stamper's (1973) LEGOL project, yet another McCarty's TAXMAN (Slayton

[Page 48 ]

Systems of this type may function as a "reversed legal information system". As stated above, they (or rather, some of them) may be considered as complex, computer-stored forms. These "forms" may be filled in by the client himself, the system showing him alternatives and drawing conclusions. The best example of this type of system seems to be the Computer Assisted Legal Services Project, cfr. McCoy/Chatterton 1968 and Computer Assisted Legal Services, Progress Report 1969.

[Page 49 ]

3 Focusing on reference retrieval

In this book, we will turn our attention to reference retrieval. This is mainly due to the fact that our research so far has been directed toward this type of retrieval - and to some extent, this book is a summing up of earlier projects within the NORIS program. But the choice is not wholly pragmatic. Fact retrieval is not a situation primarily associated with lawyers as users - though, as illustrated above in section 2.2, lawyers too may be oriented toward facts. We believe, however, that the legal decision process is not the most adequate context for discussing fact retrieval.

The discussion of forms as an information system and - ultimately - as a decision system is, however, closely associated with the legal decision process. But the model developed for the legal decision process above in chapter 1, is hardly sufficient for discussing such systems. In particular one would have to discuss in greater detail the nature of norms, the nature of "discretion", the iterative nature of the process, etc. Consequently, we think that a discussion of the use of forms would fall outside the scope of our book.

We are then left with reference retrieval - and in our opinion this is important. The basically significant fact is that reference retrieval corresponds to the traditional legal research carried out by lawyers as an integrated part of their work. Retrieval systems - in the form of manuals, indexes, etc. - have for a long time been of assistance to the lawyer in his research. The computer-based systems represent a technical revolution in this respect, and the nature and consequences of this revolution ought to be of interest. Another reason is the existence of computer-based reference retrieval systems, both commercial and public, both specialized and general. The value of these innovations should be discussed. Furthermore, a basis for understanding and evaluating the different alternatives and systems should be established.

The task indicated above surely provides more than sufficient grounds

[Page 50 ]

for discussing reference retrieval systems separately. In this book we cannot hope to supply an exhaustive discussion, but we do hope that we shall be able to establish some sort of departure point.

As stated above, reference retrieval corresponds to the traditional legal research. Computer-based reference retrieval systems represent an attempt to make this research more efficient. But opinions may differ with respect to what "more efficient" should imply.

(1) The time aspect. The simplest interpretation would be "efficient" in the sense that the user is able to carry out the same amount of research in less time. The justification for introducing a computer-based system is then founded on the time-saving element: the time saved is balanced against the cost of the system.

It should be evident from our model of the legal decision process that the retrieval sub-process is just part of the total process. Whether this sub-process is time-consuming, depends upon several pragmatic factors: the nature of the problem, the lawyer's place in the decision system, organizational questions, work load, etc. Also we want to stress the importance of the availability factors, which represent resistance to legal research in the information system. Whether the lawyer is motivated to overcome this resistance is also dependent upon pragmatic factors (for instance the nature of the problem or the importance of the client). Overcoming availability factors requires user effort, typically measured in time.

Against this background it is hardly surprising that empirical studies show that the time spent on legal research is just a fraction of the total working hours of the lawyer. A Canadian study of the practising lawyer indicates that lawyers on an average spend 21 per cent of their time on legal research. Legal research is split into two categories - "finding law" amounts to 32 per cent of the time spent on legal research, while "analysis to determine relevance" amounts to 68 per cent (cfr. Operation Compulex 1972:65). The justification for such a division may be questioned, however - our model of the legal decision process indicates that interpretation of retrieved legal sources is an integrated part of the retrieval process, and one of the main causes of the iterative nature of that process. Experience from a Norwegian user experiment indicates that when the user has access to legal sources through a terminal, a great part of the terminal session is used to study the texts on the display (Brukerforsøk 1975:13-14).

The Canadian survey revealed great differences between users - even within the category of "practising lawyers" - regarding time spent on

[Page 51 ]

legal research. While 14 per cent of the lawyers spent 15 per cent or less of their time on research, 14 per cent spent more than 50 per cent of their time on such work. This illustrates the importance of the pragmatic factors in the legal decision process - and even greater variation might have come to light if the survey had included other user categories. This may be illustrated by the German survey (Jungjohann/Seidel/Sörgel/Uhlig 1974:44), where lack of time ("Zeitmangel") is pointed out as the most important single factor leading to an unsatisfactory result of the legal research by several user categories (for instance civil servants) while it is ranked rather lower by other user categories (for instance professors and students). The variation among user categories makes it difficult to formulate any general statements on the possibility of saving time by introducing better retrieval systems - such statements must be relative to the user's situation. But there is reason to believe that only a few instances can be found where introduction of a computer-based retrieval system can be wholly justified by the saving of time alone.

The German survey cited above does, however, point in another direction that should not be ignored: lack of time as an availability factor. When a lawyer is pressed for time - as is very often the case within public administration - his lack of time becomes one of the pragmatic factors limiting his legal research, cfr. Jungjohann/Seidel/Sörgel/Uhling 1974:44, 49. In order to change this situation, it is necessary to change some of the pragmatic factors. One solution would be to ease the workload by increasing the number of lawyers - another would be to change the information system. The effects of this would be difficult to see from a survey concentrating on the time aspect - and because a modified information system would change the work situation of the lawyer, the conditions of the empirical survey would no longer hold true.

Against this background, the reasoning in the Canadian survey (Operation Compulex 1972:26) seems precarious to us - taking average percentage of time spent on legal research today measured in minutes as the maximum time which it is possible to save through a better retrieval system (18 minutes a day). As stated above, a better information system would represent a change in the availability factors, and the consequences of this change may be difficult to measure in terms of time saved.

(2) The quality aspect. The discussion above does indicate that there are limited possibilities of justifying computer-based retrieval systems by arguing that the user will be able to do the same amount of research in less time. The flaw in this argument is inherent in the phrase "same amount of research". The introduction of a better retrieval system will in

[Page 52 ]

itself imply a change in the "amount of research carried out". This is a qualitative change - and this change may in itself be the justification for introducing the new information system.

If we go back to our model of the legal decision process, we maintained that there exist meta-norms determining whether the retrieval carried out by the lawyer was satisfactory. Lack of time is one availability factor that reduces the possibilities the lawyer has of carrying out extensive research. Another such factor is the lack of an adequate retrieval system, making the retrieval process relatively time-consuming.

In special decision systems pragmatic factors tend to make it a practical impossibility for the lawyers to conduct what would be "satisfactory research" according to the legal meta-norms. Especially within public administration, critics have pointed to situations where they find satisfactory research a practical impossibility. This argument has actually been formulated into a slogan: the legal information crisis - which was the title of the analysis presented by Simitis in 1970 ("Informationskriese des Rechts"). He argues that modern society has caused a legal information explosion. Legal norms are used as a tool for implementing social reform and gaining political control, resulting in a deluge of statutes, regulatory law, etc. ("Normenflut"). At the same time the complexities of modern society give rise to more frequent conflicts - which in turn has lead to the establishment of an increasing number of agencies for solving conflicts (administrative courts, revision boards, etc.), causing a deluge of legal decisions and precedents ("Entscheidungsflut").

This legal information crisis undermines the "rule of law" in its traditional sense - Simitis argues that the rule of law can be strengthened only by extensive use of computer-based systems (cfr. also Bing 1973). And this is also the main justification given for the massive effort by the German Ministry of Justice to create a national, legal information system (JURIS, Germany) - cfr. the foreword of the Minister of Justice, Gerhard Jahn, to the initial report of 1972 (Das Juristische Informationssystem 1972:3-4), where he quotes Simitis (cfr. also Fabry 1973:299-300). A similar argument can be found in the Australian report produced by the committee on Computerization of Legal Data (1974:10-12).

This improvement in the quality of the legal decisions is usually emphasized as a justification for introducing computer-based legal information systems. The Canadian report discusses this aspect, and concludes that the practising lawyer does not have a great need for this sort of improvement, Operation Compulex 1972:24-25.

[Page 53 ]

We feel, however, that the Canadian report fails to take into consideration that an improved information system represents a basic change, which will hardly make a simple extrapolation of today's situation valid. Consequently the statement that a better retrieval system will represent improvement "for some 20% ... of a typical lawyer's clients" seems precarious. It seems for instance a likely guess that lawyers - when they are in a position to choose - tend to select cases involving little legal research, at any rate in cases of clients with small resources. This tendency seems to hold true in respect of social security cases in Norway. An improved information system may consequently cause changes in the composition of the lawyer's clientele.

As is illustrated by the German example, the "information crisis "may be most closely associated with public administration. Our own research into public administration has also led us to believe that the justification of a computer-based system must be founded on the possible improvement of the decisions, cfr. Brukerforsøk 1975:17-18.

The discussion of the user experiment in the Norwegian Social Security Court may serve as an illustration. In the court two typical attitudes toward decision-making were witnessed; one was oriented toward efficiency (in a narrow sense), the other toward the role of the Court in the administrative system. The first attitude underlined the time factor, and motivated a search for tools to increase the turnover of the Court (simple procedure, more lawyers, etc.). This group was at first attracted to the retrieval system, as one hoped this might reduce the time used on each case. It turned out, however, that the new retrieval system did not save time. Whereas the lawyers using the traditional system had been restricted to sampling a few precedents selected from the manual file, they were now able to browse through a great number of precedents at the terminal. Consequently their research habits began to change; their interest was directed toward the legal background of the case at hand. If they had restricted themselves to the same sort of research as was earlier done by a manual indexing system, time might have been saved. But the type of computerbased system available (the IBM-system STAIRS) encouraged a different retrieval strategy and a different type of research. Although this was a disappointment to the efficiencyoriented users, it appealed to the users concerned with the legal and organizational role of the Court: they argued that better decisions would in time lead to a general change of quality all through the administration. And the possibility of greater consistency would also strengthen the predictability of the decisions of the Court, and in time reduce the number of decisions appealed to the Court. Cfr. Brukerforsøk 1975:11-16.

In this section we have outlined some of the relevant arguments involved in the decision to introduce a computer-based reference retrieval system. We think that this outline may also serve as our justification for focusing on reference retrieval: it corresponds to the traditional form of legal research. Legal research comprises problems which are important to lawyers and may best be understood from the lawyer's point of view. This holds true for both the question of how to design a high performance reference retrieval system and the question of whether the introduction of a computer-based system may be justified.

[Page 54 ]

HOME
PREVIOUS PAGE
NEXT PAGE