We have departed from the traditional practice of combining discussion of retrieval with discussion of the use made of the data that have been retrieved. Having considered the retrieval process in the previous chapter, we now consider the use of what has been retrieved. We regard it as a quite different sort of process.
Information defined
The term "information" has been widely and loosely used. We would prefer, for present purposes, to follow Fairthorne in adopting a strict definition: Information is not stuff but a process. It is the process of becoming informed.. 1 According to this definition, information retrieval systems do not retrieve information: they retrieve physical things, such as signals, data, documents. These physical things may, when perceived by somebody with appropriate prior knowledge and suitable cognitive skills, contribute toward a change or increase in that person's knowledge. Here we are concerned with the process of becoming informed.
The strict definition of "information" as the process of becoming informed is, intellectually, a satisfying one. Unfortunately, the term "information" has also been used very widely to denote what is, or might be, retrieved—the "physical things" referred to above. In consequence, restricting the use of the term "information" to the process of becoming informed would be substantially at variance with customary usage. (The term "regulation" is also ambiguous in denoting either a process or an entity.) For clarity, therefore, we will use the rather cumbersome phrase "becoming informed" to denote the process. We view this as a two-stage process as depicted in figure 9. 1. The first stage in this process is the physiological perception of the signals (can they be seen? ... or heard?).
Fig. 9.1 The process of becoming informed
The second stage is the cognitive process of becoming informed. For this, other factors are involved, notably prior knowledge and cognitive skills:
- Can the person read—as opposed to merely see—the text?
- Is the text in a language that the reader understands?
- Are the concepts involved comprehended by the reader?
If any one or more of these three conditions were not met, then it is unlikely that the individual would become informed.
Some of the paradoxes of becoming informed can be readily explained once variations in the reader's prior knowledge and cognitive skills are taken into account. For example, the very same book can be simultaneously highly informative and not-at-all informative on the same subject area.
A textbook would be uninformative for an expert who is already familiar with the subject matter, but very informative to a novice. A brief note or inscription might be highly informative to the expert who has the necessary background of knowledge and cognitive skills to understand, but not the novice who has not.
The situation is similar in the case of allegories and other literary forms with allusions. The expert reader—one with substantial relevant prior knowledge or insight—will "get more out of" the text than will someone without such knowledge. 2
The vexing problem of choice of criteria for the evaluation of retrieval systems results in part from this aspect of information. Retrieval can be performed more or less effectively, but only in terms of the attributes assigned to the documents. What can be done, therefore, is retrieval which is more or less accurate with respect to chosen attributes. However, the ultimate goal is to be informative. But, as just noted, the extent to which a document is informative depends largely on the reader, i.e., on factors external to the document.
What results is a troublesome paradox: retrieval systems exist in order that users may become beneficially informed. Yet, strictly speaking, retrieval systems cannot be evaluated on their ability to inform beneficially, because becoming informed depends, in part, on factors external to the retrieval system. We, shall return to and resolve this paradox in the final chapter.
Knowledge and opinion as belief
Viewing library services primarily as information services intended to reduce distressing ignorance, risks understating the role of library services in influencing and changing opinions. 3 Influencing opinions and, therefore, attitudes can be a motive in providing library services, as, for example, when sponsored by one government in the territory of another.'4 Yet the distinction between knowledge and opinion is not really a fundamental one for our purposes since both are, at root, a matter of belief. Even when considering scientific knowledge, it is well to remember that library collections in past times contained plenty of scientific information no longer believed by scientists on such topics as the orbiting of the earth by the sun, the origin of species (before Darwin), the laws of physics (before Einstein), and much else besides. Indeed, this scientific knowledge is still made available by libraries in the older books in collections. Presumably, some of the contents of the newest books will not find acceptance, or, if accepted, will cease to be accepted in the future. That some of what is believed in one century is superseded by different beliefs in another is the essence of progress in scholarship in science as in other subjects. New history books not only add new assertions concerning historical events, but refute some older assertions and change the way that others are understood. This is even more marked in other areas more concerned with values and less firmly rooted in factual evidence. In brief, we take the view that both people's knowledge and their opinions are elements of what they believe, and that the process of becoming informed is a process that adds to or changes individuals' beliefs. 5
Barriers to becoming informed
There are different sorts of barriers to becoming informed. 6 In the following discussion, four sorts of barriers will be discussed as they refer to the retrieval of documents.
Let us imagine that someone has an inquiry concerning a specific aspect of Buddhism.
Indicative access
There would be a problem of defining what document, if any, would be likely to contain writing which would probably be helpfully informative with respect to this inquiry. We could well call this indicative access since it pertains to the ability of indexing systems to indicate which of the parked documents is likely to be worth retrieving. (It is common in information retrieval circles to refer to this "indicative access" as "intellectual access," which would seem too broad a term and best used only when other sorts of access that we discuss separately below are also included.)
Physical access
Having identified a document that will probably be useful, physical access is needed to it in order that it can be read. If the only known copy of this document is in Bhutan; if permission is denied for it to be photocopied or sent out of the country on loan, and if access to Bhutan is denied, then the indicative access will not result in any intellectual access—for lack of physical access.
Linguistic access
Even if physical access has been arranged, if it is written in Tibetan, there is likely to be a problem of linguistic access. One would have to learn Tibetan or arrange for a translation in order to overcome what would otherwise be an effective language barrier.
Conceptual access
Even when problems of indicative, physical, and linguistic access have been solved, there may remain a difficulty in conceptual access if comprehension of the text requires an understanding of concepts of Buddhist metaphysics which the reader simply does not have.
These four sorts of accessibility all pertain to retrieving, and being informed by, documents. As Wilson points out, libraries tend to concentrate on the first two (indicative access and physical access) and to ignore the second two (linguistic access and conceptual access). This uneven attention to the problems faced by people who seek to be informed warrants some discussion.
Libraries do provide some assistance. Typically, translations and dictionaries and, commonly, a list of translators are maintained for problems of linguistic access. Introductory texts, directories of experts, and encyclopedias are also stocked so that the would-be reader can, on a self-help basis, work toward an understanding. The selection of books bought emphasizes the language of the library's users even when the language of the original is different. At a more elementary level public libraries have long been concerned with fostering literacy. 7 Similarly, the selection ought also to emphasize the conceptual level expected of the library's users. Generally, however, it is true that the emphasis in libraries is overwhelmingly on the indicative access and the physical access. The wonders of interlibrary loan arrangements are an important feature of both the practice and the folklore of libraries. Linguistic skills are among the qualifications sought in recruiting library staff, but it is very rare for a translation service to be provided except in libraries of corporations.
As for assistance with conceptual access, a high standard of general education is expected of librarians (their professional training in the United States and Canada begins only after a first general or specialized university degree), and librarians are often encouraged to develop a knowledge of the subject matter dealt with in their libraries. However, this knowledge, for good practical reasons, tends to be wide rather than deep and there are severe limits to the extent to which it can be used to assist users with problems of conceptual access.
The two forms of access that libraries emphasize (indicative and physical) and the two they do not (linguistic and conceptual) also differ in another way. Linguistic access and conceptual access have to do with direct communication. There is no nonintellectual barrier between the document and the eye. Indicative access and physical access involve communication indirectly. Until both of them have been resolved there can be no basis for communication. One can hardly read something if one does not know what it is that one should be reading and, even when that problem has been resolved, one still cannot read it if it is not visible. This distinction between direct communication and the indirect facilitating of communication reinforces the view that libraries are primarily concerned with information retrieval rather than with communication generally.
There is, however, an alternative explanation which has to do with the practical probabilities of people needing help with each of these barriers and with the availability of alternative forms of help. This argument might run somewhat as follows: Library users can, within limits, adapt to and cope with linguistic and conceptual problems, but indicative access and physical access are problems which the library service can handle more effectively than can an individual—and they are problems that can be delegated to the library service with less difficulty. People adapt by tending to direct their reading to texts that are in languages and are at conceptual levels that they can handle. A good library service will provide for some latitude in this matter when selecting and can generally arrange access to other collections of advanced material when that is needed. People tend to confine their attentions to languages they can read fluently. What is more, they tend to confine themselves to the material which happens to be at hand and known to them. However, pursuing this line of argument is likely to be circular. People adapt to situations, therefore it is not clear when behaving one way is the preferred way, or the way that one has to in order to cope with prevailing circumstances. It becomes unclear how libraries might order their priorities differently.
For whatever reason, libraries do presently emphasize—and emphasize heavily—activities that are characteristic of information retrieval rather than those which are not, even within the range of activities which have to do with communication and becoming informed. (Different aspects of access will be examined further in chapter 15.)
The historical features of retrieved objects
In the previous chapter, we discussed the essentially historical aspects of indexing or "signaling through time" in Mooers' words. 8 We now extend that discussion to consider the consequences of these essentially historical aspects for the process of becoming informed.
All communication is somewhat historical in at least the trivial sense that the physical communication of signals must take some time. In direct interpersonal dialog, the delay is so short as to be insignificant. However, when it comes to reading documents, the delay becomes more important.
Setting aside, until later, the special cases of intentional and unintentional inaccuracy in writing, we find that the most that can be said of a text is that it records what was alleged to be accurate as of the time it was written. It is usually assumed that the assertions are still accurate at the time of retrieval. However, two points are fundamental to written statements:
- They are only descriptions and should not be confused with what they describe. A written statement to the effect that the moon is made of green cheese does not, in and of itself, guarantee that this is so, nor does it cause the moon to become so.
- Even if the description were accurate at some point in time, it does not follow that the accuracy will continue. Even an accurate statement that, on a given day, the exchange rate for the U.S. dollar was 13.90 Austrian schillings, does not constitute reliable evidence that the exchange rate will still be 13.90 a year later.
The importance of this historical aspect will depend on the changeability of the reality being described. The longitude and latitude of the city of Berkeley do not vary, but the availability of specific houses for sale in Berkeley does. One can, therefore, use an old gazetteer to ascertain geographical coordinates with more confidence than an old list of houses for sale when seeking to buy a house.
The factual data derived from libraries tend to be the reporting of what is printed in the latest edition of whatever book contains data on the matter in question. Crowley's study of the accuracy of public library reference services was criticized because, although the sorts of questions he asked were reasonable ones (e.g., "What is the name of the Secretary of Commerce?"), they did tend to deal with matters that were volatile relative to the currency of the reference works generally found in libraries. 9 All data and documents yielded by retrieval systems necessarily relate to a past state of affairs.
For some special purposes, the administrative fiction is used that the recorded description is to be deemed accurate unless it can be proved incorrect. This is to be expected when ascertaining the actual state of affairs is substantially more difficult than referring to retrievable data. The description is, officially, true. This is usually assumed in airline seat reservations. It is also importantly true of ownership, as for example, the ownership of houses. Proving that official records are in error may be a substantial task.
A further consequence of the time interval in retrieval-based information systems is that use of them must necessarily be an historical activity. For some purposes, especially in the study of the humanities, this may be advantageous. One might very well wish to compare what Austrian writers thought before, as opposed to after, the dissolution of the Hapsburg empire. To repair an elderly elevator, one would probably find an elderly repair manual more helpful than a recent one. In the humanities, the range of uses for which records are needed is extremely extensive. Most obviously, this includes biographical data of all kinds: age, attitudes, activities. Further, there are literatures of many genres which reflect the human experience. A proper understanding of the social and cultural contexts that are distant in time or space is important for the proper comprehension of the human experience of people in those places and/or times. The range of circumstances in which people need to use records to reconstruct the past is very large: stage-set design, restoration of houses, appreciation of past literature, and so on.
Logically distinct but similar in practice is the attempt to examine parts of the past in order to examine and, perhaps, explain how the present came to be—or, indeed, how the present is.
In scientific studies, essential features are the statement of hypotheses and the replicability of experiments. These have to be formally recorded. They become historic records to be stored for future critical examination. Without this "archive" or "corpus" denoting the "body" of science, the cumulative development of science through the scientific method as we know it would not be possible.
For the conscious study of the past, the historical nature of records seems self-evident. What is less self-evident in practice is that all use of retrieval-based information systems is necessarily historical. Records are descriptions of how something is supposed to be. After the recording process, the record describes what was—or, rather, what was supposed to be. In accounting systems, the record describes how income and expenditures were supposed to be at the time the record was made—or, more commonly, for some even earlier point in time since in accounting it is inconvenient to keep absolutely up-to-date. Weather stations record the temperature and other climatic phenomena at particular points in time. A will records the maker's intentions at the particular date with respect to the transmission of his or her property. Even if one consults such records promptly, one cannot know with certainty that the financial situation indicated in the accounts, the temperature at the weather station, or the desired disposition of property is still as described. Indeed, a change within hours is likely in the first two cases. One may choose to assume that there have been no significant changes, but that is different from knowing with certainty that there have been none. One cannot ascertain by inspection of a record alone whether or not what is recorded is, at the time of inspection, an accurate reflection of the object described. It is, however, sensible to assume that the description is accurate if the probability of change is believed to be low and/or the consequences of being misinformed are not expected to be serious.
A special case occurs when the object retrieved is not a document (i.e, not a conscious description). Museum objects, for example, can be highly informative without being descriptions, e.g., fossil or an obsolete farm implement. However, the difference is superficial. The museum object is regarded as informative about the situation whence it derives, just as a document is. It is possible, in both cases, for the resulting information to be misleading since what is retrieved might be an unrecognized fake, might have been assigned incorrect attributes (of provenance, of nature), and, even if neither a fake nor wrongly attributed, might still be misleadingly atypical. Of course, the situation whence it derives is likely to have changed in the meanwhile.
The difference between document and museum artifact would appear to lie more in the assumptions of the beholder than in the signals perceived. That is, we are more likely with documents and data to assume that the information derived describes the present time as well as describing the time of its creation. Indeed, it may sometimes be necessary as a practical matter to make such an assumption. Yet, it remains an assumption of uncertain validity.
Misinformation, harm, and distress
In order that we may later consider the "goodness" and the evaluation of library services, it is desirable to examine two evaluative aspects of becoming informed.
Misinformation
Let us consider the case of misinformation wherein one's knowledge is changed to include "facts" (or rather beliefs) that are false. In simple cases, this is fairly straightforward. If the flight to New York is scheduled to leave San Francisco airport at 9:27 a.m. but the travel agent states that it is scheduled to leave at 9.27 p.m., one is being falsely informed.
If one probes more deeply into the question of truth and falsehood, one rapidly gets into difficult philosophical issues which we prefer to leave to others. We simply note that one can be misinformed (intentionally or otherwise) and that even information believed to be correct at the time (e.g., scientific knowledge) may well be regarded as incorrect at some future time.
Harm and distress
Misinformation is generally associated with harm and distress. One may be distressed or harmed (or both) by missing an airline flight to New York because one was misinformed about the the time of departure. However, although misinformation is commonly distressing and harmful, truth is different from benefit. It can happen that one is distressed by information generally accepted as correct, as when learning that there isn't a single "real" Santa Claus. One can even be harmed by correct information. An example of this is the argument that publishing reliable predictions of earthquakes might cause more disruption than would refraining from publishing them. Hence, the phrase "Ignorance is bliss." (Mis-information and lack of information are not the same, but the distinction does not appear to be important for present purposes.)
On the other hand, it can conceivably happen, even if only exceptionally, that misinformation is beneficial. This would be the case if the flight to New York, which one missed through misinformation, were to crash after take-off.
Our intention here is merely to assert that knowledge is not necessarily always beneficial, nor ignorance (or misinformation) necessarily always harmful or distressing. Yet, in so arguing, we are necessarily asserting social values since beneficial and harmful imply some standard of value. One could make the argument that knowledge is invariably beneficial and lack of correct knowledge is invariably harmful. Yet to do so would imply that one's dominant social value is knowledge. While one could adopt such a social value it is clear that it is not universally accepted. More commonly, happiness in this world or the next is seen as dominant, with knowledge regarded as normally beneficial because instrumental. As soon as one gets away from the dominant notion of knowledge itself being the supreme social value, then there is scope for knowledge to be at odds with social values. This is seen in censorship. The view is commonly found that the retrieval of pornographic materials is, in at least some circumstances, contrary to human happiness. The view is also commonly found that the unrestricted dissemination of some religious, social, or political views—or secrets of national defense—can be harmful. However, our present purpose is to analyze not to evaluate. We seek only to assert that such social values can be held by people who are involved in the offering of library services. It is important to stress that these attitudes to information and misinformation derive from individuals' social values.
A different but related matter is the assignment of priorities to different sorts of knowledge. Even if some sorts of knowledge are not perceived as being harmful (for example, horse racing results), one might feel that the social benefits derived from such knowledge are trivial compared with other sorts of knowledge, such as home improvements, literary masterpieces, or one's own political or religious point of view. Such views of priorities will affect the relative benefit attributed to being informed in different ways. This may sound paternalistic. It is. Library services are normally provided by some for others.
Limitations on the helpfulness of library services
Although library services are, by general consent, intended to be helpful, there are at least two different ways in which the librarian cannot be entirely sure what will be helpful for a given patron because the information is only fully significant in relation to the unknowable inner workings of the user's mind. We have noted above that the process of becoming informed depends on the combination of signals from outside and the cognitive skills and prior knowledge within the user's mind. The librarian cannot fully know what the cognitive skills and prior knowledge of the user are. However, an imprecise estimate may be sufficient for most practical purposes.
A deeper problem has to do with the use of information. In seeking information, it is likely that the user is interested in making a more informed decision in some matter. (The assumption is that becoming better informed will result in a better decision. This is plausible but not certain!) However, decisions are taken on the basis of criteria determined by the decision maker, otherwise the process would be measurement or implementation rather than genuine decision making. The values guiding the decision are necessarily in the mind of the decision maker. Indeed, they may be rooted deep in the subconsciousness of the decision maker who may not be fully conscious of his or her motivating values. It is even less possible for the librarian to know these values; and without a complete knowledge of what the values are, it is not possible to know fully what information would be pertinent.
In brief, the librarian cannot expect to know completely what information is appropriate for decision making even when the decision to be made has been described.
In the last four chapters we have considered why people might want to use libraries, how collections and retrieval systems serve to provide access to data and documents that might be useful, and the process of becoming informed by what has been retrieved. These three processes form a cycle. In the next three chapters we move away from this cycle and look at library services in broader terms. In the next chapter we ask what determines the amount of use that is made of libraries. We then ask how library services come to be provided in the manner and to the extent that they are. In attempting to answer this question we start, in chapter 11, by viewing library services as they are ordinarily provided—as a free public service. In chapter 12 we broaden this discussion to include some consideration of the effects that might result if library services were provided on a commercial basis.
Go to Chapter 10
1 "... information is an attribute of the receiver's knowledge and interpretation of the signal, not of the sender's, nor some external omniscent observer's nor of the signal itself" R. A. Fairthorne, "The Theory of Communication," Aslib Proceedings 6, no. 4 (November 1954): 255-67. [Reprinted in R. A. Fairthorne, Towards Information Retrieval (London: Butterworths, 1961, pp. 64-79)].
2 "The early Church fathers sometimes used a threefold method of interpreting texts, encompassing literal, moral, and spiritual meanings. This was refined and commonly believed to have achieved its final form in the medieval allegorists' 'fourfold theory of interpretation.' This method also began every reading with a search for the literal sense of the passage. It moved up to a level of ideal interpretation in general, which was the allegorical level proper. ... Still higher above the literal and allegorical levels, the reader came to the tropological level, which told him where his moral duty lay. Finally, since Christian thought was apocalyptic and visionary, the fourfold method reached its apogee at the anagogic level, at which the reader was led to meditate on the final cosmic destiny of all Christians and of himself as a Christian hoping for eternal salvation." Encyclopaedia Britannica, 15th ed., 1977 vol. 7, p. 135, s.v. "Fable, Parable, Allegory."
3 "The despotic ideal is twofold: to find out what a man is thinking, and to determine absolutely what he shall think. The first has not yet been achieved. Progress has been made with the second by indirect methods, and it is with these that libraries are concerned—methods which do not prevent thinking, but alter its quality." A. Broadfield, A Philosophy of Librarianship (London: Grafton, 1949), p. 12.
4 See chapter 6, footnote 1.
5 This rather simplified view should suffice for the purposes of this book. For a broader introduction to this matter see F. Machlup, Knowledge and Knowledge Production (Princeton, N.J.: Princeton University Press, 1980), esp. chapters 2 and 5.
6 This has been discussed in P. G. Wilson, Public Knowledge, Private Ignorance: Toward a Library and Information Policy (Westport, Conn.: Greenwood Press, 1977).
7 E. G. Smith, "The Literacy Gap: The Involvement of Public Libraries in Literacy Education," Library and Information Science Research 6, no. I (January-March 1984):75- 94; issue on "Adult Education, Literacy, and Libraries," Library Trends 36, no. 2 (Fall 1986): 183-345.
8 C. N. Mooers, "Mooers' Law, or, Why Some Retrieval Systems are Used and Others Are Not," American Documentation 11 (1960): 204.
9 T. Crowley and T. Childers, Information Service in Public Libraries: Two Studies (Metuchen, N.J.: Scarecrow Press, 197 1). Summarized and discussed in F. W. Lancaster, The Measurement and Evaluation of Library Services (Washington, D.C.: Information Resources Press, 1977), pp. 91 ff.
Copyright © 1988, 1999 Michael K. Buckland.
Document maintained at
http://sunsite.berkeley.edu/Literature/Library/Services/chapter9.html
by the SunSITE Manager.
Last update March 2, 1999.
SunSITE Manager:
manager@sunsite.berkeley.edu
