“Exploration of Buddhist Texts: Traditional Methods in a Digital World” Lewis Lancaster Professor Emeritus, University of California, Berkeley Adjunct Professor, University of the West It is hard for me to realize that my participation in digital projects related to the Buddhist canon started more than four decades ago. When I tell the current generation that I set up an input project for the Goryeo canon at the Institute of Chemical Research in Shanghai in 1989, and they hear that it was because the Institute had a computer with “a”, that is one, megabyte of memory, they can’t help laughing. Today, I have a very modest cloud-based storage account that allows me to use three terabytes of memory, three million times the amount on the Shanghai computer. In those early days, desktops had 32 Bit Intel Chips that could miraculously do 5 million instructions in a second. Today, my laptop can process nearly 500 million instructions in the same time. Such growth in memory and speed leaves me somewhat dazed and hardly able to comprehend what has happened in a relatively short period of time. But then remember that I was once dazzled with the idea of working on a computer with one megabyte. As an “immigrant” into the digital world, I have had to struggle with the implications and opportunities available through the so called “new” technology. It is thanks to a number of compassionate engineers who have been willing to patiently lead me through the maze of software, coding, and analytics, that I have been able to do anything at all with the computer aided research. ven) Two of these extraordinary men are Howie Lan from University of California, Berkeley who started helping me in the mid-90s and more recently Alex Ames from Google. At Fo Guang Shan, I have received a great deal of support from computer literate monastics such as Vens Youzai, Miao Guang, and Zhiyue. They have stood behind me as we attempt to make use of the great power given to us by digital developments. When the digital era came into being, our first task in the Humanities and Buddhist Studies was to input enough data for the computer to be of any use. Today with the ever-growing size of memory power and functionality of high speed, so called “super” computers, the need for data is even more critical than it was at the beginning of technological advances. One digitized Buddhist canon such as the Taisho version from the early 20th century, occupies a tiny segment in storage units with terabytes of space. Any limitation to our input and collection of information based on the capacity of available memory is unwarranted. The probability of creating enough Buddhist data to exceed the current limits of computer memory is so distant that it should not be a factor is our planning. How much less likely is it that space will be a problem in the future. The new methods of storage, just emerging from computer research and development laboratories, hold promise of being able to work in a hand-held device with data the size of ten national libraries. Even such capacity, which looks imposing to us in the present, will undoubtedly bring a condescending smile from generations yet to be born. We are aware that the mere presence of data, no matter how large, will have significance only to the degree that there is a companion set of tools with which to explore, analyze, search, retrieve, and share. Work that involves the multiple canons of Buddhism, small as they may be in comparison with the data output of the collider in Switzerland, will have complexity that is beyond many of our existing tools. This complexity magnifies when our research involves texts appearing in a number of languages, scripts, formats, and relative developments that span a thousand years or more. Our scholarship in the study of the Buddhist canons cannot be limited to quantity. It also involves interpretation and contextualization. Computers have no trouble doing numbers, but difficulties arise when our research involves qualitative approaches. How can we use this large memory, rapid quantification and analysis to study text structures, language features, meaning and interpretation? What purpose can the computer play when we deal with ideas that are embedded in a text and whose function and meaning can only be spotted through inference. The problems of textual work from our canons becomes even more difficult when there are metaphors, allusions, and similes contained in complex sentences. Computer programs have yet to fully rise to the level of the challenge created when segments of texts deal with causality, sequence, and comparison, and when all this is discussed as an event structured as a flashback in time. The traditional methods of doing critical text study, seeking the original text that existed before multiple copying resulted in inadvertent errors, additions, and omissions, still are valuable and being widely used. Computer tools should be directed toward making these tasks easier and faster. At the same time, the digital capacity should also be used to analyze our methods as well as the data that results from such work. Do we need to revise our methodology? Or do we need to adopt a completely new approach based on the techniques that can be employed using applications that complete processes of inquiry in a few seconds? I have had one experience which was both upsetting as well as thrilling. As mentioned, I am an “immigrant’ to the digital world. When my research and writing for the doctoral degree was done more than fifty years ago, there was no computer, no digital canon, no way to do quantifying analysis in a few seconds. Searching for word occurrence in seven Chinese translations of a Perfection of Wisdom text, took nearly two years to complete and quantify. After the first digital Chinese canon appeared from the input of the Goryeo version in Korea, I sought to imitate my two years of labor using the power of the computer. It took me 19 minutes to reach the same level of results. Now I have available to me analytic tools that can do many of the same tasks in two seconds. The expansion of speed from performing a task that took two years to two seconds is beyond any expectations that I had when tool building and computerization of the Canon began. While speed and size of data has been spectacular over the last two decades, the question remains for scholars to decide “what is it that we want to be done?” Is it enough to ask the engineers to make sure that we can do our traditional tasks quicker and with greater accuracy? Can I be content with tools that can show me comparisons of texts over time so I can chart the history of how the document was created and used? Is it enough for the software to report on differences that have appeared in a text history over time so that I can within seconds be looking at a family tree or stemma of my text? What am I to do with the results of such constructions of data development, especially when it indicates that my Buddhist text seems not to have resulted from a single ancient source but a multitude of sources? If I can’t find or reconstruct a beginning point for a text but am shown that it started with many traditions that over time become intertwined in a complex matrix, how will this affect my approach to the text? Will the computer help me to deal with the unknown people who assembled some version of the document in front of me? Can I examine a Sanskrit manuscript and receive a report on how many scribes made the copy based on handwriting analysis? What can I do with such information? What will it mean to search horizontally, that is to search multi versions at the same time rather than searching source by source in silos that require me to leave one net page and go to another for what might be called vertical searching of one text at a time? How am I to use search results that are images rather than a listing of the data itself? This means: what do I do with graphs, maps, timelines, density imagery? And what do I do when the image introduces new complexities? Can I use an image to search for data, rather than words? What results when I use a picture of an architectural feature and look for all other pictures that have the same design? As we determine the way in which our canonic research will develop, it is well to remember that one aspect will be the use that is made of our material in education. There has been a marked shift in education from knowledge acquisition to learning as a social discourse. In this regard, the Buddhist canons can be seen as a multitude of “voices”. These voices in languages, as distinct as Sanskrit, Pali, Tibetan, Tangut, or Khotanese, form a pattern of overlapping accounts of the Dharma. Our study of the variety of language canons must also include the understanding that they not only contain passages that are similar and thus overlapping, but in addition that they indicate interactions with one another. The Chinese canon contains translations from Indic texts, these represent various historical formations of the teachings of Buddhism. In the process of being translated, Chinese and Indian cultures and thoughts have interacted. That is why we can never merely study a Chinese text as purely cognitive utterances which have materialized from the inner experience of meditation and spiritual insight. This large collection of texts requires some form of study that recognizes the glue holding them together, recognizes the “thread” that binds ideas into the completeness of a “text”. A significant part of the glue or the thread is the social and cultural features of history and human encounter with others. The “voices” we find in Buddhist texts are echoes of the social discourse from centuries ago. Can the power and sophistication of our computers help us determine the structures of how a sutra in written form has come down to us in the 21st century? Can the computer make clear the cohesion of this “chain” of utterances and retrieve a vision of how this page in front of us reached the completeness we see, input, store, and search? In a more simple question: how did the unnamed ancient monastics and lay people make the copy of sutras we see in our received publications? Assuming that many of the expressions in the sutras extend back to the time just after the Nirvana of the Buddha when the First Council tried to remember what had been taught, how have these expressions been put together to make a cohesive document on palm leaves? Our digital resources can certainly help us see the family tree of the expressions over time, in various languages, and in different cultural spheres. It is also possible to do automated part-of-speech tagging: is the expression a verb, noun, subject, object, modifier? With dictionaries, we are able to trace the words and their shifting formations over the whole of these multiple canons. Of greater difficulty is the coding algorithm for estimating relevance of a word in one or more canons. Can the computer “read between the lines” to capture inferred meanings that are put into figurative language and thus are hard to detect? Since many texts in the canon contain information that is abstract, intricate, and theoretical, what is a method for being able to state the meaning that is revealed in the totality of the text? Most reading on the internet is done by looking at snippets that are listed for us. What we receive is based on the search words and the functionality of the search engine. This type of reading, while providing a great deal of information, often fails to provide the reader with the major theme of the complete text. Can software ever show us the ideas that dominate when a book is seen in its totality? There is another realm in the computer that is opening up for scholars in the Humanities: the ability to “Augment” “Reality”. Using constructed visuals is still not widely done in our canonic research. With the help of Howie Lan, I have been exploring the use of “Blue Dots” to create a better method of seeing patterns of word occurrences. We changed the text of the Goryeo version so that in place of characters, we see only blue dots. It is deceptive because the “Blue Dot” links to a wealth of information. The software knows what character is being covered up, what line and sequence in the line is held by the character, the text title, translator, time of translation etc. Sarah Kenderdine placed our “Blue Dot” version into Virtual Reality at City University in Hong Kong. She made it possible to convert the dots back to the original characters for reading. Users can stroll among the characters of the texts, show search results with a red dot replacing the blue ones. One can set up multiple pages in the surround and can “turn” not just one page but one or two hundred pages at a time. I don’t yet know how to make use of this application. There is a negative view of spending time on something that you are not sure is useful. However, there are times when we need to do critical inquiry without knowing the immediate usability of the results. This is often the only way to find new and helpful information. I think Buddhist scholars need to think more about critical inquiry, the basic science, and not simply be constrained by the thought that “everything I do must have a discernable use.” My VR attempt may not lead me to a specific product or software but there is no question about the need for Buddhist Studies scholars to begin thinking about this technology. In another project, I have been working on an Atlas of Maritime Buddhism. Thanks to the support of my partners Sarah Kenderdine now at the Polytechnic Institute in Lausanne and Jeffrey Shaw at City University of Hong Kong, we have constructed a VR form of the Atlas with 3-D images from nearly 70 sites. Ven Ruchang the Director of the Buddha Museum here at Fo Guang Shan has been brave enough to schedule an installation of the Atlas starting next April . We are pioneering a new look at Buddhist history and that includes a different approach to the way in which the Buddhist canons moved from India to China. I invite you back to Fo Guang Shan for the opening April 30th of next year. We will find out whether people find such an installation of value when compared to reading a book filled with static pictures. As you can hear from this lecture, the study of the Buddhist canons reflects all the computer developments and social issues related to digital technology. We are challenged to find out the best use of our content as well as applications of it. What are the ways in which the traditional patterns of textual research can be enhanced, and what will result from the search for new methods and data use? Our content may not constitute the largest data banks in the world but the complexity of dealing with multiple canons in different languages with different lists of included texts, makes Buddhist Canon studies among the most complicated and challenging. I hope that this conference can help to bring about an empirical revival of text studies, in part by moving beyond the focus on our material texts and entering a computing environment that may allow us to better “see the past” and through our text criticism of the digital versions understand more completely how the canons were transmitted and transformed. This will require that we come to grips with the mechanism and data representations from which our received versions have been constructed. The tasks before us are difficult to define and hard to fulfill. It is our karma to be in this world at the beginning of the digital age, and we have a unique opportunity to perform pioneering feats in relation to critical inquiry, data representations, and proto-types for canonic study. Thank you for participating in this conference and I look forward to our discussions and believe through sharing we can find inspiration for the future.