In discussing my plans to write this book with friends, colleagues and acquaintances, the reactions I received ranged from some who thought it was a joke that shouldn’t be taken seriously, to others who looked forward eagerly to seeing it in print. Few people understood, or understand today, that the path that led me to the origin of the dragon idea did not come from addressing this issue directly, but rather from seeing how people reacted toward others who pointed at rainbows.
It began in the fall of 1980, when I was employed as a professor at the University of Leiden in the Netherlands. In accordance with a bilateral agreement between Holland and the Republic of Indonesia, five teachers from Indonesian teachers’ training colleges were chosen annually to spend nine months in the Netherlands, where they focused on one area in the broad field of linguistics (the scientific study of language) as a way of upgrading their professional qualifications before returning to Indonesia as college teachers. The selection process began with a larger group of teachers that received basic instruction in-country, followed by a vetting process and the selection of the lucky five who would go for more in-depth training in Europe.
As part of one of these cycles I was sent to a small community called Tugu, among the scenic tea plantations in the mountains of west Java, where I was tasked from September 30 to November 1, 1980 with teaching the fundamentals of historical linguistics (the study of language change over time) to a group of about 30 Indonesian teachers, drawn from all parts of that large and diverse island nation. My lectures (in Indonesian) took up about six hours of each day, punctuated by a lunch break, ending in the late afternoon or evening with a spirited game of badminton if the weather permitted. At one lunch break during the month I was there a number of us were queued up outdoors waiting for our food in a light drizzle, and a brilliant rainbow suddenly arched through the sky from mountain top to mountain top. Impulsively, I pointed it out to someone standing in front of me. In the most courteous way possible he told me that where he had been raised (somewhere in Sumatra) people think it is not a good idea to point at rainbows. Before I had a chance to ask him why, someone behind me who was from another, rather distant part of Indonesia, said “What—you have that belief too?”. Both then stated that as children they had learned that the consequence of such untoward behavior would be that the offending finger would be permanently bent into the shape of the rainbow. I don’t know whether I gaped, but I was transfixed by this correspondence of fantastic beliefs from widely separated islands. It clearly wasn’t true, so why should people share the same colorful falsehood?
As a first attempt at understanding, I assumed that this was an arbitrary belief that had arisen at a time when the language groups in question were a single community. After all, this was what I was teaching in our classes about the nature of language change: communities split up and over time the language of each changes in its own way, but for a period of up to several millennia the common origin of this set of linguistic relatives can still be detected by systematic comparison. I put the matter aside, we got our lunch, and the rest of my month in Indonesia went smoothly.
That could have been the end of the matter. However, this experience stayed in my mind, and on returning to Holland I began to dig into ethnographies of linguistically unrelated groups, looking for anything I could find about how the rainbow was viewed by traditional tribal peoples. It did not take long for the shock to set in. Within a very short time I discovered very similar beliefs among the tribes of the Naga Hills in Assam, eastern India—people who are linguistically unrelated to those in Indonesia, but who share a similar traditional culture. This meant that I could no longer harbor the comfortable belief that this was an arbitrary invention of one prehistoric culture that had been passed on to its descendants—most of the peoples of Indonesia speak languages that belong to the vast Austronesian (‘Southern Islands’) language family, while the languages of the Naga Hills belong to the Sino-Tibetan language family, which includes the various forms of Chinese.
At this point I had no choice but to assume that there was something about the human reaction to rainbows that had triggered a parallel response to pointing at them in historically unconnected societies. This was a small beginning, but within the next few years my efforts to expand the documentation of this belief (which I had begun to call ‘the Rainbow Taboo’, or RT) showed that it is a global belief, reported in Europe, Africa, East Asia, Southeast Asia, aboriginal Australia, New Guinea, and North, Central and South America. Individual writers had previously recognized that the RT is found across a number of cultures in geographically restricted areas, as North America, but no one had previously seen that it is in fact a culture universal, that is, a feature of culture with a distribution that cannot plausibly be explained by common history or borrowing, and so implies some kind of universal psychological motivation.
The RT has already been extensively documented (Blust 2021), but what is to the point here is how my pursuit of this belief led me to dragons. In my effort to collect as much data as possible concerning attitudes toward the rainbow, I prepared and sent questionnaires to colleagues who were teaching in Papua New Guinea or doing linguistic or anthropological fieldwork in several Pacific Island and African countries, and to heads of missionary field stations connected with the Summer Institute of Linguistics (SIL) and other Bible translation groups around the world, with a request that they be completed by workers in the field, a technique similar to that pioneered in the nineteenth century by the American anthropologist Lewis Henry Morgan in collecting data on kinship systems. This provided by far the richest source of data on the ethnology of the rainbow, and should show that Morgan’s method of data collection remains useful even today for certain types of information.
These questionnaires asked not only about the RT as such, but also about traditional views of the rainbow in general. In effect, this method of data collection allowed me to indirectly interrogate hundreds of indigenous peoples, many of whom still retained large parts of their traditional animistic belief systems. From the completed questionnaires, most of which were obtained between 1981 and 1985, together with published sources from other writers, some of which proved to be extraordinarily helpful (although their authors probably would never have suspected it), it soon became apparent that most traditional societies share strikingly similar beliefs about the rainbow that differ radically from the Western conception in nearly every respect. In pursuing the ethnology of the rainbow, then, I was led metaphorically into the den of the dragon, for it soon became all but impossible to distinguish the two.
Chapters 5 and 7 of this book are very data-heavy, since they provide the basic documentation of the globally distributed traits of dragons and rainbows, respectively. Documentation requires broad cross-cultural sampling, and as most anthropologists have recognized for some time “language, although admittedly not always in agreement with the realities of cultural identification and cultural dynamics, nevertheless offers the only consistent and complete basis for the selection and arrangement of units” (Lebar, Hickey and Musgrave 1964:v). Consequently, the sample units used for purposes of documentation are ethnolinguistic groups (people who share a similar culture and a common language). As much as possible each informant from whom questionnaire data was obtained is identified by name, language group represented, language family affiliation, and in many cases major branch of the family, as well as country of residence. The classification of North American languages follows Goddard (1996), and for most others I follow Simons and Fennig (2018). Because the genetic classification of the non-Austronesian languages of New Guinea and neighboring areas remains unsettled, and because many language names in this area are quite variable, I have had no choice in a few cases but to fall back on the default label ‘Papuan’, although it should be made clear that this term means nothing more than ‘non-Austronesian.’
I wish I could thank those two Indonesian teachers whose mutual surprise in learning that they shared the RT started me on a long journey that has led to this book, nearly forty years later. Alas, their names and identities have long since been forgotten, but there are many others whose help I can still acknowledge. The Acknowledgements provides a partial list, with apologies to anyone who I may have inadvertently overlooked.
Many of the traditional belief systems represented by the material in this book are rapidly disappearing as a result of globalization and the end of animism. The beginning of the Industrial Revolution prodded Jakob and Wilhelm Grimm to record the folklore of Europe before it ceased to be transmitted through an oral tradition. In much the same way this book aims to document concepts of the rainbow as these were expressed in the vast majority of the world’s cultures, before modernization and conversion to a world religion put these ancient beliefs on a path to eventual extinction, leaving only a mysterious trace of their former existence in the form of the dragon itself.
A NOTE ON SPELLING. Because the Rainbow Serpent in aboriginal Australia is commonly treated as a personified being, its name is usually written with upper-case ‘R’ and ‘S’. Not all writers follow this convention, but enough do that I have chosen to adopt it, and since I use it for the Australian version of this creature, consistency has called for me to use it in other areas as well. The dragon, however, is not treated in this way in any source available to me, and as a result ‘dragon’ is consistently written with a lowercase ‘d’. Similarly, when used specifically as the name of the planet we live on, I write ‘Earth’, but in more general references to ground or soil I use the lowercase alternative ‘earth’. Conventions regarding the use of upper and lower case symbols that are used in quoted material are naturally left as found in the original. Finally, both nāga and Shintō properly have long vowels. These appear correctly in de Visser (1913), and I follow his practice here, but vowel length is ignored in the less specialized sources I have cited, and since they are quoted verbatim, I have left the original spelling unchanged in these cases, despite the inconsistency this creates. Finally, the linguistic classification of the 378 ethnolinguistic groups cited in the text is given in the Appendix.