John Hind
Magonia 2, Winter 1979-1980
In this paper an attempt is made to adapt, extend and formalise techniques developed in the scientific study of folk tales to the study of high-information UFO texts. The first part of the paper introduces the concept of the motif, and suggests a formal technique for the comparison of the motif content of UFO account texts. In the second part of the paper, a standard for a motif-based catalogue of UFO accounts Textcat is proposed.
In the past, most attempts at objective analysis of UFO data have concentrated on applying parametric statistical techniques. A small number of attributes are selected and the distribution of these attributes over a large number of individual reports is investigated. While this approach has yielded and will continue to yield useful results, it seems unable to deal with the complexities of the data. Furthermore, there are several serious pitfalls which many researchers have been unable to avoid. The first is that parametric statistics assume the reliability of the original data - a very unwise assumption in ufology. Statistics, used on low grade data, is merely a technique for turning low class rubbish into high class rubbish: The second problem is that it is often very difficult to relate statistical results to the original data statistics has a distancing effect.
Because the type of data used in research has a feedback effect on the type of data that is collected in information, the predominant ethos has been to prefer quantity to quality. Statistics thrive on quantity, and can disguise low quality. Most investigators have concentrated on collecting data in a small number of defined categories on as many cases as possible. Other investigators have interested themselves in the pursuit of the 'ideal case'; the case that will once and for all prove the 'existence' of UFOs. The Holy Grail of ufology seems as far out of reach today as it ever was.
A recent development in ufology has been a new dedication by individuals and groups of ufologists to the detailed investigation and documentation of high strangeness reports. One such group, the UFO Investigators Network (l) in Britain, has on file several tens of such highly detailed reports, and in the United States, the Centre for UFO Studies has published a smaller number. Flying Saucer Review has been publishing medium-cletail high strangeness reports for many years. Although this data is of variable quality and is often distanced from the original source material by 'storification' aimed at making it more readable, it constitutes the largest body of such data in existance.
In the study of any such large body of data, the problem of how to compare, analyse ann catalogue arises. An attempt at a straightforwarn. listing of abstracts of Type-One (Vallee's classification) reports has been made by Peter Rogerson (2) Rogerson's INTCAT will continue to be a vital reference source, but catalogues of this type are of limited usefulness as research tools. INTCAT is classified chronologically, and it is thus quite impossible to locate a sample of reports having a particular attribute, for example. So the first problem we face is how to catalogue high information content UFO reports without making assumptions, and without limiting future research. A parallel problem arises when one wishes to compare UFO accounts in an objective manner.
A very similar problem arose almost a century ago in a different emerging scientific discipline. Throughout the nineteenth century, a large body of folktales was collected from all over the world by a dedicated body of men and women, who called themselves folklorists. Many theories were proposed to account for the fact that folk tales collected from widely divergent societies showed close similarities. The sheer volume of the data, and the seeming impossibility of classifying it adequately was a formidable obstacle to the testing of these theories and the development of theoretical structures in the science. The folklorists were faced with a body of data closely similar in nature to present-day UFO accounts. This data consisted of written accounts, almost always from verbal originals, and passages of prose. Often several versions of essentially the same story occurred, the data was sometimes of dubious reliability, and it had been collected by people of different capabilities and motivations.
At the risk of alienating some readers, it should be pointed out that there are some similarities of content as well. The most obvious is that both genre are predominantly concerned with interactions between humans and intelligent non-humans. Nineteenth century folklore interacted with contemporary romantic literature in much the same way that modern UFO lore interacts with science fiction. In both cases the stories appear to originate with, and be sustained by, the hierarchically lower elements in society. It is important to realise, however, that one does not have to accept that UFO reports can validly be considered a form of modern folklore to learn methodology from the folklorists.(3)
Around the turn of the century, three main methods of classifying folklore collections were being developed. These were classification by type, classification by key-word, and classification by motif. Although it was the third system that was the most successful, it is instructive to look at the other two as well.
Classification by type depends on the ability to isolate a particular characteristic of the tale, which separates it from other tales. Foe example, the well-known fairy-tale Cinderella has a type of its own in the Arnee/Thompson type index (4). Assuming everyone knows this tale, its type is given simply as 'Cinderella'. It is given type number 510A, which puts it in the range 500-559 which is reserved for tales involving 'supernatural helpers'. Thus the cataloguer has decided that the important point about the Cinderella story is that the heroine is helped by a fairy. This type of system, in which each unit can appear only once in the classification is the sort of system suggested by Vallee (5), Haines (6), and Hynek (7), etc. In each case the units are classified according to a particular attribute shape, date of occurrence, proximity to witness, whether 'seen' on radar, at night or by day, etc.
Classification by key-word means that the cataloguer compiles a list of important words in the stories, arranges them alphabetically and references each word to the tales in which it appears. This process is basically the one used in, for example, the index to a book. Classification by key words gradually evolved through various systems which classified 'key phrases', to the formal method of motif classification pioneered by Professor Stith Thompson (8).
A motif is the smallest structural entity in a story a plot element or detail. The concept is best illustrated by example, and two examples of motif listings are given. The first (Table One) is from folklore, a reduction by a Miss Cox of Grimm's version of Cinderella's cited by Thompson (9,10) We can see that some of the motifs are general, and crop up in many stories (Happy marriage; Help at grave; Ill treated heroine, etc) and some (Heroine hides) are specific to the story and unlikely to turn up elsewhere.
The second listing is of a part of the report on the Aveley Abduction by Andrew Collins (11). The listing is by the author, and it will be noticed he uses a slightly different punctuation protocol. (See Table Two)
The motif listings form a kind of shorthand method of describing an account of a UFO experience. In the listings one has all the essential information obtainable from the original, with several important exceptions. Most of the verbal padding and complexities of tense have been removed. It is an executive decision to what extent detail is lost. For example the very full description of the two types of entity given in the Aveley report is vastly abbreviated in the motif listing, just as in the Cinderella listing we lose all the detail about how excruciatingly nasty the step-sisters are! More fundamental is that specific information such as names, ages and sexes of the participants is eliminated. In a complete system, provision must be made separately to code this information.
More important than the 'boiling down' effect of the motif listing process is its quantising effect. The process has reduced a continuous passage of prose to a number of discrete elements, each of which only contains one idea. These discrete units of text can now be processed independently of the original. This fact will be used later to develop a proposal for a new catalogue based on the classification of motifs, rather than of 'text-units' (Saunders)(12) or 'events' (Rogerson) (11) However, there is a less ambitious use of the motif concept in ufology.
There are two basic types of motif: 'action' motifs and 'descriptive' motifs. From the Aveley reduction, "UFO follows witness in vehicle" is a typical action motif. Action motifs have 'subject' and 'object' ('UFO' and 'witness' respectively) and action' (follows vehicle). A typical descriptive motif is "UFOs have underwater bases". These motifs have 'subject' (UFOs) and 'specifics' (have underwater bases). To take further examples: "Hypnotic regression releases further details" is an action motif, since both since both subject (witness) and object (investigator) are implied. "Entities perform organ and system transplants" is a descriptive motif (because it describes the state of alien technology, rather than a specific event). In UFO motifs, all subjects and objects can be thought of as in the categories outlined in Table Three.
Later, these categories will be used as the basic units of classification with further subdivisions to ensure that any particular motif is easy to find, and that similar motifs are classified in close proximity. However it is possible to derive other classification systems more useful for comparative analysis. An example of a non-exclusive (where a motif is not limited to one category) classification or characteristic set is given in Table Four. The list of characteristics can be structured to suit a particular project, but the tabulated scheme will be used as an example.
A 'characteristic-matrix' can be constructed for a motif listing with respect to a 'characteristic-set'. A table is prepared with motif-numbers vertical, and characteristic numbers horizontal. When a particular motif has a particular characteristic, a '1' is placed at the appropriate intersection. This characteristic-matrix is used to construct a 'profile-vector' as follows:
DERIVATION OF THE PROFILE VECTOR FOR THE AVELEY LISTING
(WITH RESPECT TO THE CHARACTERISTIC LIST GIVEN AS TABLE ONE)
The columns of the matrix are added up, forming a vector. The elements of the vector are added and divided by the number of elements in the vector (the number of characteristics). This process produces a 'scaler' a single number scale factor. If each element in the vector is now divided by this scaler, the profile vector for the motif-list with respect to the characteristic list has been found.
The profile-vector can be directly and objectively compared with a similarly derived vector representing another text unit, derived with respect to the same characteristic list. If it is desired to compare two groups of profile vectors, mean and variance vectors can be used. Thus, given large samples of text units in two different categories (UFO reports/bigfoot reports; normal investigations/hypnotic investigations; 'real' encounters/'imaginary' UFO encounters, etc.) one could objectively argue that the two samples showed more internal consistency than similarity (unlikely to be causally similar) or vice-versa. Again, one could highlight the similarities and differences between two different accounts of the same event, or between two different events.
Motif profile analysis is one application of the concept of the motif to UFO research. It is the contention of the author that developments of this and other similar techniques are one way forward in ufology. They offer the hope that substantive as opposed to reductionist analysis can be applied to detailed investigations of high-strangeness UFO events. Because of this, they should also offer some encouragement to those involved in the dedicated and often thankless task of investigating and documenting such reports.
In the first part of this paper, the concept of breaking down a piece of text related to an alleged UFO experience into motifs was introduced. In this part, the concept of the motif catalogue itself will be discussed. In the previous section the idea of the motif was placed in the folklore context where it originated, and the idea was extended into a research method.
The idea of motif cataloguing originates with Thompson as well (14) The basic idea is that rather than cataloguing alleged 'events' or 'accounts of events' we catalogue the individual motifs present in the accounts. This system of cataloguing is better able to deal with the complexities of UFO accounts because we are not obliged to choose one single characteristic which decides an item's position in the classification. Since the motif is the simplest unit of a story, it is much easier and more valid to slot it into a unique position in a classification system.
Prof. Thompson's motif-catalogue of folk literature consists of a classified list of motifs, each of which has a list of sources in which that motif occurs. The catalogue is thus useful both as an index - if one knows any part of a story one can go directly to a reference to that story - and as a research tool of an advanced nature. In its use as a research tool, one can investigate the distribution of occurrence of particular motifs or groups of motifs; but because of the organisation of the catalogue it is difficult to investigate the association of different motifs in stories, to discern broad similarities between stories, or to recover motif listings of individual stories. These limitations are primarily technological. When Prof. Thompson began his catalogue all the work had to be done by hand a tedious and repetitive task of sorting and re-writing, and the catalogues format was limited by the pen-and-ink technology of the time.
Because of these considerations, Text-Cat has been defined as a computer based system from the start. While this may be an obstacle to its initial implementation, it is felt to be essential in the long-term. As well as significantly cutting down the necessary work, the computerisation of the project enables its scope to be considerable increased. The author feels that, as well as being of great usefulness to ufology, a Textcat based project would be a useful piece of computer technology research, with many potential spin-offs.
A basic review of the functions of the computer made use of by Textcat, with some clarification of terminology, is in order. In a sense the term 'computer' is a misnomer. Modern electronic data-processors spend most of their time and complexity, storing, retrieving and sorting information. The fact that such information is usually in the form of coded digits is neither here nor there, the information can just as easily be coded alphabetical characters. It is usually convenient to talk of the information stored in a computer as being in 'files'. A file is an open-ended (has no fixed length) repository for information. usually in the form of magnetic tapes or discs. For the purpose of Textcat, the computer can be thought of as a super-efficient secretary with a filing cabinet. The cabinet represents the total storage capacity of the computer, and the individual files within the cabinet represent the above computer-files. The 'secretary' portion of the computer keeps the files in order, recognises them when required, and searches for specific pieces of information from the file, presenting the information in a suitable form. The main advantage of computerisation is that large amounts of data can be got at quickly and easily.
In the following outline scheme for Textcat, I will present the system as it would appear to the operator. The actual means by which these functions would be implemented on the computer will be discussed only briefly they would be very dependant on the actual computer system that is used.
1. Inputting new data into the system: The operator will firstly select the actual piece of text to be coded and will input an exact reference to that text. The system will then prompt the operator to input information about the date and location of the experience described, the witnesses involved, etc. Then the operator will input a list of the motifs in that piece of text.
2. Classifying the motifs: When a stock of several hundred motifs has been built up, the system and operator will work on classifying. The actual classification system will be decided only in outline in the first instance (a scheme is suggested later on) with the details being filled in by adding new subdivisions as necessary. Periodically the system will be reviewed and rationalised, probably be committee. The basic classification cycle is that the computer presents a motif and the operator is asked to input the classification. The operator may be helped by a facility to output all motifs previously classified under a particular classification. Again, periodically the motif list is rationalised by going through it in classification order and deleting closely similar motifs so that each motif is stored only once. If the classification scheme is kept in order similar motifs ought to appear in close proximity to each other.
3. 'Housekeeping' functions: From time to time it will be necessary for the system to ask for operator assistance in organising its internal files. The most obvious area in which this will be necessary is in identifying text units in which the same case is described. For this function the computer, on the basis of similarities between the data on location, time and date, and witness name, will present candidates for consideration to the operator who must decide whether they do in fact describe the same event.
Internally, the computer will organise all this information into three primary files as follows:
1. Case File: Case reference number; reference number of all TEXT UNIT FILES referring to this case; (Date; Location; witness names; Sources in which this information is agreed). Subsidiary case files,identical except for the information in brackets, are created for each disagreeing source.
2. Text-Unit File: Text-unit reference number; Source details (source, page, paragraph numbers, word numbers of each coded block of text); Motif listing (a list of all numbers of all motifs in the text unit in the order in which they occur); Comment field (should contain name of operator and any comments)
3. Motif File: Reference number and text of all motifs, reference numbers being arranged according to the classification system. There will be provision for several MOTIF FILES using different systems of classification.
Using this system, the continuous build-up of an extensive data base on high information content UFO reports is made possible.
The main use of the system will be in the sorting, indexing and abstracting of UFO reports on the basis of their motif content. By developing subsidiary catalogues of motifs according to specific desired (description of entities, symbolic or psychological criteria, etc.) it will be possible to organise and investigate the data in quite sophisticated ways. A sample of the questions which the system should be capable of providing answers to would include:
Is there a variation from mystical to pseudo-scientific motifs over the years, or from place to place? In their treatment of the same case, do different authors stress motifs of different character? Is there a variation of the types of motifs reported by witnesses of different age, sex, etc. in the same culture? In what ways have the motifs reported changed over the years? Are there any discernible differences in motifs reported under hypnosis as against motifs reported in the waking state?
Obviously there is almost endless scope here, and they system is versatile enough for us to have reasonable confidence that new and creative uses will be found for it after implementation.
While the primary motif classification system should be based on actual samples of motifs, an initial framework for this evolution will be necessary. One starting point is to realise the limited range of subjects and objects that feature in UFO reports. These have already been listed as Table Three, with some of the specifics which each heading would include Thus it would be possible to classify action motifs according to type of subject and type of object, and descriptive motifs according to type of subject. Finer divisions can be derived, based on experience, but for example, within 'Descriptions of entities' we might like to distinguish 'monster-like entities', 'human-like entities, and 'robot-like entities'.
While the Textcat proposal is much wider in scope than the research procedural suggestions outlined, the potential payoffs are also greater. Implementation of the Textcat proposal would require firstly a secure financial basis (perhaps as low as a few thousand pounds annually) and secondly a small team of committed ufologists and computer professionals on at least a part time basis. Set against this, it could reasonably be expected that such a project would lead to useful pay-offs in the fields of computer science as well as in ufology.
While text based techniques should be of some use to the physicalistic ufologists in their investigation of the possible physical origins of UFO reports, the main application will be in the more recent area of humanistic ufology. This should not be surprising, since the attitude of the folklorists was, and is, for the most part humanistic. Few folklorists interest themselves in alleged proofs of the existence of the fairy-folk, leaving such folly to well-known writers of detective fiction ! Humanistic ufology is however in much need of objective techniques for testing the largely intuitive theories and speculations that its proponents generate. Unless such techniques can be applied the 'New Ufology' is in danger of degenerating into a theoretical fashion house seducing the intellect with black satin theories, characterised by their plunging necklines and high frontal slits!
Notes:
- UFO Investigators' Network. [Now defunct]
- Published in MUFOB/MAGONIA
- See article by the author on Folklore in the ncyclopaedia of UFOs, Edited by Ronalcl Story, to be published by Doubleday (NY) and New English Library
- Aarne, Antti, The Types of the Folktale, translated and enlarged by Stith Thompson. Second revision, Folklore Fellows Communications, no. 148, Helsinki (1961)
- Vallee, Jacques, Anatomy of a Phenomenon, Spearman, 1966
- Haines, Richard F., 'UFO Appearance Recognition and Identification Test Proceedures', BUDC Bulletin, Nay 1977
- Hynek, J. A. The UFO Experience.
- Thompson, Stith, The Folktale, University of California Press, 1977.
- Thompson, Stith, Motif-Index of Folk Literature, 6 vols. Folklore Fellows, Helsinki 1932-36, and Indiana University Press, Bloomington, 1932-36
- Thompson, Stith, (Ibid 1977) p.415
- Collins, Andrew, The Aveley Abduction, UFOIN, 1978.
- Ufocat, the computerised catalogue compiled by Dr David Saunders of the Center for UFO Studies, has one entry per source, per event. This a single event may have more than one entry.
- Conversely, INTCAT has only one entry per event, and each entry may reference more than one source.
- Thompson, Stith, (Ibid 1932-36)