As a widespread phenomenon sound-archives have not been in existence for very many years compared to other types of archives and collections from man’s past. We have only just started to realize some of the problems which other forms of collections have had for many years; for instance the problem of preservation, and problems in connection with the rapid growth of collections in modern times. Suddenly, the collections may - for different reasons - be considered too large, awkward to handle and too difficult to use.
We all know - or have a clear suspicion - that audiovisual materials are much more difficult to preserve than documents. The cost of preservation may in the long term be so large that the necessary resources cannot be raised. Utilizing material in the archives, in research, in broadcasts or reused in the form of sound- or video-cassettes will for many subject-groups become a problem, because of the sheer size of the collections. It is not only limited space that causes us to declare the collections too large - audiovisual material is very, very slow to work with, and one can fear that some research-disciplines will give up using audiovisual material to answer traditional problems and questions if unreasonable amounts of time have to be used referring to audiovisual material compared to other forms of source-material. This problem should not be underestimated. The type of questions which the users of the archive expect an answer to after investment of a reasonable amount of work, will degenerate into simplification and end up in becoming irrelevant for serious research, if the collections are too large or confusing.
Up until now we have only seen the beginning of a selection-debate within the realm of sound-archives - and have been spared the merciless necessity of radical selection until now. This is the beginning of a debate which will last for many years and which will be constantly reformulated in connection with technological developments, changing types of users and economic development. The debate will be intensified and we will hear echoes of it every year at IASA conferences and many articles about the problems in the Phonographic Bulletin.
The problem of selection has many aspects and it can be discussed in several ways. It is of course impossible to solve the whole problem in one short session. This paper will, therefore, concentrate on a somewhat over-looked aspect; namely how selection is conceived of philosophically and methodologically. What kind of philosophical conception offers the best interpretation of the general structure in the selection process, and what type of methodological conception can, in the most fruitful way, be used in analysing the details of the selection process?
In the first place it is necessary to create a clear image of a selection process, because it is the image, which is the starting point of the analysis. It is not unimportant what kind of selection process you imagine. Some archives have very rigid selection-typologies - so rigid that it is almost meaningless to speak of selection - and other archives take everything they can lay their hand on. It is clearly unfruitful for our analysis to take such procedures as a starting point. It is necessary to build up our image from elements of a selection process which is reasonably free from rigid selection-typologies and at the same time considers it absolutely necessary to make rather thoroughgoing selection. When one is investigating philosophical problems it is important not to get lost in practical problems - we want to analyse the structure of the selection problem, and to that end it is necessary to construct a kind of ideal-typical image of the selection process - ideal-typical in the sense that the image consists only of elements in the process which are relevant for the analysis. The ideal-typical cannot be found in any existing archive but, nevertheless, our construction of the image has to originate from a knowledge of existing archives. I will, therefore, describe very briefly the selection process in my own archive and thereafter, you must supply this image with elements from other archives, which are known to you. In short: we want to create an image of a ‘clean’ selection-process, and in this creation or construction we only use the elements which characterize and accentuate the structure in the process.
The archive used as an example is Radio Denmark’s archive for spoken word and non-commercial music. The only principle we need to take into consideration in our selection of material is ‘possible re-use in radio programme production’, and as radio programmes are being made about almost everything - or at least one can imagine so - we are relatively free from rigid typologies. But we are restricted by the fact that if the collections become too large, they will at the same time become unfit for programme production - a problem which is normally greater for journalists than for researchers. 95% of the material is actively selected by archive staff and the selection is based on short descriptions of content which are sent to the archive from the programme production departments, including television when the sound track is sometimes relevant for our collections. Every day we receive these short descriptions of content covering about 50 hours of production and we select on average 5% from the large number of complex problems being treated in a nation-wide non-commercial broadcasting station. The intention of selection is that the material shall contain the essence of the subject and its treatment; and that, of course, cannot be done by selecting some ‘objective’ percentage from news and cultural relevant programmes - in fact it is a question open to debate as to whether ‘objective’, automatic percentage-selection deserves the name ‘selection’ - rather it should be called ‘sorting’. The selector is supposed to possess an educated problem-consciousness - an education which not only derives from knowledge of the radio-medium itself and its specific way of treating the problems, but also from independent and thoroughgoing knowledge of the problems essential and characteristic of his own time and our knowledge of man.
In Radio Denmark the selection procedure is mainly founded on one staff member who examines all the radio and TV programmes each day. After this first examination the selected material is secured by other persons such as technical staff and the programme departments themselves - but the important thing here is the principle - one person/all the programmes - which forms the corner-stone of selection in Radio Denmark. One of the advantages of this principle is that it makes it possible to build up the collections systematically around essential themes and developments in the medium. For example, all the small but perhaps important elements from news and magazine programmes are examined in the selection procedure. Separately they are perhaps not relevant for the archive, but if systematically pieced together they can be used in the future description of many important problems, some while ago in Copenhagen there were episodes in connection with young people who illegally moved into empty houses. Several programme departments, both television and radio, reported these episodes, which were rather interesting in bringing into focus otherwise unrealised problems. But the reports where only given as short items in magazine and news programmes, and they would have been lost if selection was limited to the large, unique programmes. This is often the case if the programme department has a monopoly of selecting material for the archive. One problematic point in the procedure is that the basis for selection is short written summaries. To what extent is it possible to use written abstracts as a basis for selection of audiovisual material? Another facet of the selection process, which cannot be covered in this paper.
To sum up, the process of selection can be described as follows. Each day short abstracts of about 50 hours of transmission are received representing about ten pages of written abstracts. On average you must select five per cent and it must happen in such a way that the selected material does not give a passive reflection of the broadcast material. The selection ought to be active and creative, so that new problems can be treated and researched in the selected material.
The next step in the analysis is to move from this newly created image of the selection process to a precise formulation of the selection problem itself
Formulation of the problem is of great importance to the kind of answer arrived at in the end. The formulation of the problem depends on the purpose of the analysis, and the purpose is to reveal some essential logical structures in the selection process. We are now moving into a broad and difficult field within theory of science, it is the field concerned with the relationship between general or ‘nomological’ knowledge and concrete, ‘ontological’ knowledge. This relationship between general, nomological knowledge and concrete, ontological knowledge is basic to understanding that the sciences are composed of a systematic, ‘law-seeking’ part, and when trying to use a science to investigate specific, practical problems and events. This can be applied to most of the sciences; theoretical physics being used on practical problems, for instance, Newtonian mechanics can be used to calculate the time for a specific solar eclipse, or other general physical laws can be used in specific weather forecasts, or theoretical economics can be used to explain specific economic events; or within our own archival realm, a generally formulated selection-typology or subject-typology can be used to decide whether a specific recording is to be selected or erased!
But a subject-typology is not scientific ‘knowledge’; and there are other differences. However, the structure of the problem of the relationship between a generally formulated subject-typology and a specific, concrete decision to select or erase is the same as in the just mentioned sciences. Of course this does not mean that a decision to select or erase has the same degree of precision as an astronomical calculation of the orbit of Halley’s comet; a comparison between selection-decisions and weather-forecasts might be more proper. Selection is not, and cannot become, a science in the classical sense of the word - but the structure of the problem of the relationship between general and specific knowledge is identical in the sciences and in selection-work. When the structure is identical, then we are permitted to use the available philosophical and methodological scientific literature in so far as this literature is concerned with the structure of problems and not for instance with the nature of the objects to be analysed. Before taking up these matters it is important to solve the formulation problem, or rather the two interconnected problems, which appear in the selection process. As it is in the sciences, the case in our archival interest of knowledge is split in two different ways of thinking. One of these is systematic and general and is concerned with the formulation of the archive’s selection-typology or subject-typology, and the other is concerned with the problems of the specific decisions of selection. The selection problem can be formulated as follows: the problem of selection is the problem about the interconnection there is or ought to be between the more or less explicit general subject-typology and the specific selection-decisions. The subject typology need not be explicitly formulated-in a way it is just our general conception of what ‘ought’ to be selected.
An example of a short subject-typology could be:
Now, one can ask, what is the condition of possibility for creating such a typology? What must necessarily be known before the typology can be formulated? The answer to these questions is of course that it is necessary to have a more or less conscious or clear conception of what it is essential to know about man, combined with a conception of the way in which the audiovisual material in the archive can be expected to contribute. Some of the most important aspects of the problem of selection are to be found in the thinking connected with formulation of the subject-typology. The distinction between essentials and inessentials, which in the typology is formulated in a general way and in relation to the archive’s purpose, rests on criteria derived from conceptions of a metaphysical character such as conceptions of ‘man’, of ‘society’, of ‘culture’ etc. It is in these concepts, which in our knowledge and actions work in a heuristic and not a deterministic way, that our personal ‘philosophy of history’ or our personal ‘metaphysic’ are to be found. All these concepts are often very vague and only partly consciously analysed, but nevertheless they are necessary and form the basis for our image of world. They are a kind of a priori, torso-like, but absolutely necessary total-image of our life-world, and in reality it is this flickering and unanalysed image, which is fundamental for our thinking knowledge, when we act, and when we select. It is the flickering, metaphysical image, which is behind both the formulation of the general subject-typology and the specific selection decisions. In the typology the image has been formulated in a general way - in the specific selection the image has the character of a heuristic basis for a decision. Everyone who has made decisions about essential/unessential has to find his arguments in these metaphysical conceptions if his selection decision is being criticized and therefore needs some kind of argument. It is an eventual critique, which makes clear the metaphysical character of selection work - and metaphysics does not mean pseudo-religious conceptions, but the common principles and concepts, which create coherence and integration in our thinking. Coherence and integration are the key words!
It can be seen therefore that the problem of selection is founded one kind of metaphysics; but it also has other philosophical aspects. One of these lies within theory of knowledge. One could say, that the metaphysical part of the problem is concerned with the answers to the question: what are the conditions of possibility for selections - what is the a priori in selection? But the theory of knowledge problem is concerned with the answers to the question: why must this material be selected and that erased? It is the structure or logic of thinking in practical selection work we are looking to find. How does the selector motivate a specific selection?
At this point the discussion can take advantage of the previously mentioned debate about the relationship between general and specific knowledge, between nomological and ontological knowledge. In the space available we can mention only a few elements in the dominant conception of the problem and some of the points of critique. The dominating school is positivism, and it is the positivistic conception which underlies what many Americans call ‘hard’ science’.
The literature about positivism is very large, but a classical representative is Carl G Hempel, whose articles The Function of General Laws in History, (1942) and Studies in the Logic of Explanation (1948) are worthy of reading. In short, the positivistic ideal is, that when you must explain an event, then it is necessary to split it up into several parts, so that every part can be explained by means of the general laws formulated within secure sciences. If, for instance, you must explain a specific traffic accident, the first thing to do is to establish the experiential data; influence of alcohol, braking distance, the specific traffic situation, mental condition of the driver, etc., and then you will investigate these data by means of the relevant sciences, which in the case will be: medical science, technical science, traffic-sociology, psychology and so on. The general knowledge from these sciences, which can be used, is called ‘covering laws’ because they are considered to cover the established facts. After the analysis of the facts by means of the respective sciences, the positivist nominates the sum of these separate explanations to be the explanation.
The same kind of procedure is involved when a selector motivates a specific selection by referring to it being ‘covered’ by one or often several categories in the subject-typology of the archive. But this kind of explanation suffers from great weaknesses and the critique of positivism is about to uncover some of them. A brief critique of positivistic theory of knowledge can be found in Maurice Mandelbaum’s article: The problem of covering laws, in History and theory, (1961), and a very profound critique can be found in one of the most exciting, modern philosophical works, Bernard Lonergan’s Insight. A study of human understanding, (1958).
The starting point for Lonergan’s theory of knowledge is the analysis of the event, which we call an insight. He is not interested in insight as a psychological phenomenon, but analyses the character of the increase of our knowledge, which is involved when we get an insight. For Lonergan it is, so to speak, the insight, which is the atoms of knowledge – for the positivists the atoms are verbally, formulated, elementary statements about what is considered the reality. There are of course many problems involved in this, but the point is, that the positivistic summing up of their many separate explanations involves an insight, an epistemological operation, which is overlooked by the positivists. The separate explanations of the mentioned traffic accident are: influence of alcohol, bad brakes, bad traffic conditions and a depressed driver, and the positivist considers it unproblematic to put a circle around all these separate explanations and to nominate the sum: The explanation!
What is overlooked is that the circle or summing up involves an increase of our knowledge, and therefore it must be motivated. The positivistic theory of knowledge is unable to motivate the summing up! There are perhaps secure sciences which can be used in explaining the separate parts of the event - but there is no science which can explain how the specific parts of the event are integrated again; remember the key words are coherence and integration! The positivistic theory of knowledge will only be able to create explanations in the form a long string of possible separate explanations. In the same way, a positivistic selector can only motivate his specific selections by referring to this or that point in the explicit subject-typology of the archive - that is the reason why a positivistic selector always has an explicit and very detailed subject-typology: he simply cannot live without it.
In short, the problem for the positivist is, that his theory of knowledge does not include principles by means of which coherence and integration can be created. In practical work there are no differences between a positivist and a non-positivist. The difference is in their philosophical and methodological understanding of what is involved in their practical work. To the positivist, the integrating principles of coherence are ‘unscientific’ or ‘subjective’ and are, therefore, not included in his theory of knowledge. But the selector with a fully developed self-awareness will recognise the integrating insights, which are involved as a necessary part of his work and these integrating principles are exactly the same metaphysical principles, which have been mentioned before. It is, therefore, a self-delusion to believe that a reference to some selection-typology can be used as a proper motivation for selection - unless of course the typology is so restricted that the problem of integration does not exist. Then the positivistic understanding can be used, because no integrations of separate explanations or motivations are involved. If, for instance, the archive must select every kind of material on a certain person, then it is obvious that the motivation problem does not exist. However, in that case the question remains as to whether selection is involved at all.
The problem of selection has only been solved in a very general way in the formulation of the archive’s subject-typology. The kind of general solution of the problem in the subject-typology cannot be directly used to solve specific problem of selection. Often you will have a feeling that the selection problem is buried in the typology. For instance, the typology contains the category ‘contemporary history’, but within this very category there are essentials and inessentials. And furthermore, the category contemporary history cannot be defined in relation to history as a science. It is defined in accordance with the nature of audiovisual material or the needs of the programme producing departments. This means that the criteria traditionally used in the science history cannot be directly applied. A typology can only in a very restricted sense be used to solve specific ad hoc selection problems. It can be used to give potential customers a general idea of what can be found in the archive, and it can be used to define in a formal way the working-purpose of the archive. It can also be used in the selection of material, which may be unambiguously described as in a particular category.
Application of a general typology presupposes an integrated insight into the elements of the specific problem of selection. Otherwise you will not be able to discover that the typologies or categories are involved in the specific problem - and in the same way you are unable to discover the specific integration of the categories. In short, the application of a general typology needs interpretation in every specific selection decision. The typology is only one out of several factors involved in specific selection work, and it must be stressed that the typology itself does not contain principles by means of which the integration can be motivated. The principle of integration is to be found in ourselves in the form of the aforementioned heuristic, pre-scientific, but necessary total-image, which is activated in the meeting between the selector and his specific selection problem.
Therefore selection is a kind of metaphysical achievement resulting in a creative, integrating insight and decision, and it is founded on a complicated combination of our more or less conscious and educated total-image, explicitly formulated general subject-typology and understanding of the specific selection-problem involved. Selection is not in the category objective, secure, hard science, but in the category meaningful, subjective action. For that reason a selection decision will always have to live with the possibility of being questioned. The selection decision is not a science but an action, which is open, creative and perhaps even playful - and long may it continue to be so. Objectivity in selection, for instance in the form of percentage selection or the impossible (in the long run) total selection will transform archives into stores and the creative, educated selection into automatic, mechanical sorting.
Therefore, in reply to the first of the two questions raised in the paper’s introduction, I am of the opinion that the problem about the philosophical conception of selection is best solved by conceiving selection as meaningful subjective action, always open to debate, and as a philosophical method I would recommend self-awareness in relation to the practical selection work based upon the epistemological phenomenon insight as understood by Bernard Lonergan.
The second question, which was raised in the introduction, concerned the methodological conception of selection. The problem has only been touched upon and space decrees only a very brief mention of the research-discipline, which will be staid the selector to make his practical work transparent. What I have in mind is the decision-making theory.
To recapitulate, I tried at the beginning to show the presence of a kind of metaphysical heuristic total-image, our subjective-reality, as a foundation for selection-work. This image was formulated in the general subject-typology as a description of what categories of material the archive must select to fulfil its purpose. The general subject-typology has the character of a normative description, however specific selections are not description, but decisions! The typology and the specific selection are two different ways of thinking about the same problem and that is why the problem of essential/unessential seems buried in the typology when you try to motivate specific selections by means of the typology. The typology is not a set of rules to be followed blindly; rather it is a kind of framework for creativity. The specific selection is a creative decision, and it is of course unwise not to use the large literature about this subject to reach a more profound understanding of what is involved in decisions.
Some of the fundamental works within the discipline are to be found in the general action theory, but the most directly usable works are strangely enough to be found in the special elaboration of the theory for analysis of political decision-making. Richard Snyder, Graham T Allison and John D Steinbruner ought to be mentioned and if you are interested in analysis of the concept meaningful subjective action, Alfred Schutz ought to be consulted. Space forbids me to enter into details of decision-making theory, but the self-understanding of the selector will be much improved by studying the decision-making theory’s analysis of the many very complicated factors involved in decisions, from the cultural blind spots, through the organisational forms of the archive itself and on to the individual psychology of the selector.
Understanding these factors cannot, of course, make selection objective but it can make the selection decisions less arbitrary - and the selection debate can profit from the definitions and the very differentiated concepts within the decision-theory. Another problem, which can be illuminated by decision-making theory, is the problem of differentiating between types of selection. Earlier in the paper I outlined the type of selection in my own archive, but there are of course a lot of other types - one of them I have referred to elsewhere as sorting. In the decision-theory there are at least six different paradigms or types, each of them characterizing a type of decision. For instance, there is an analytical paradigm, a cognitive paradigm and a cybernetic paradigm - the last one can, with advantage, be used to characterize the very simplified and automatic form of selection, which I have called sorting. But these differentiations and refinements will come in connection with the intensified selection debate, which we can expect in the years to come.
Poul von Linstow is the Radio Archivist of Denmark Radio, Copenhagen.
This paper was given at the Brussels Conference in 1982.