Tools of the Scribe

People all around the world now carry out nearly synchronous conversations using text. This puts a premium on efficient writing, something that is easier in some writing systems than others and for some individuals than for others. Fast production of text, however, is not a new problem, and has its roots in typesetting, stenography and assistive technologies. Some of these areas of technological innovation were hugely successful in the West, but were less successful in other parts of the world, such as Asia, where differences in scripts and writing systems made simple solutions to fast text production elusive. Many of these same problems remain today, but the existence of very large text corpora and the advances in AI that this has enabled now permit the use of natural language technology that makes text production faster and more accurate.

This book presents writing technology past and present with a broad focus, discussing both widely used technology as well as technology serving communities of writers with special needs. For example, text is the principal communication modality for many with severe motor disabilities such as cerebral palsy: How does one type if one cannot easily or reliably point to a specific key on a keyboard? Cross cutting the discussion are several themes: How does one’s language and script influence the technology and its use? How does the technology interact with the user’s motor abilities? How does text input differ when one is writing in one’s own language, writing in several languages, or writing in a second language in which one may not be fully competent? And if one’s immediate goal is not efficiency but learning, does technology that aids efficient writing also support efficient learning?

This book is the first treatment of writing technology that considers the process of writing from such a broad range of perspectives.

Brian Roark is a research scientist at Google. He received his PhD from Brown University in 2001, and joined the Speech Algorithms Department of AT&T Labs – Research, where his research focused on syntactic processing techniques and statistical language modeling approaches. In 2004, he joined the Center for Spoken Language Understanding of the Oregon Health & Science University, where he expanded his research program to include biomedical applications such as augmentative and alternative communication, brain computer interfaces and automated neuropsychological assessment. He was PI on numerous NSF, NIH and DARPA grants, and published more than 50 papers in journals and major conferences, receiving several best paper awards. He joined Google as a research scientist in 2013, where he continues to research problems in speech, NLP and text entry on mobile devices.

Richard Sproat is a research scientist at Sakana.ai, working on artificial intelligence in language processing, agentic systems and image understanding. He received his PhD in Linguistics from MIT in 1985. He has published in various areas of linguistics and computational linguistics, and he has a particular interest in writing and symbol systems. His prior relevant books in this area include A Computational Theory of Writing Systems (2000), Language, Technology, and Society (2010) and Symbols: An Evolutionary History from the Stone Age to the Future (2023). He was an invited speaker at various international venues, such as the “Signs of Writing” conference (Chicago, 2014; Beijing, 2015), and a keynote speaker at “Grapholinguistics in the 21st Century” (Paris, 2022). Contributor to the Routledge Handbook of the English Writing System (2016), he wrote a chapter (with Amalia Gnanadesikan) on writing systems in the Oxford Bibliographies (2018), and a chapter on writing systems to the Oxford History of Phonology (2022). He is on the editorial board of Written Language and Literacy.

Su-Youn Yoon is a manager of the automated scoring team at EduLab, Inc., Japan. She received her
PhD in Linguistics from University of Illinois at Urbana Champaign in 2009 and joined the NLP and
Speech group at Educational Testing Service (ETS). Her early research centered on scoring non-native speakers’ oral proficiency. In the automated speech scoring field, grammar and vocabulary scoring presented notable challenges. Yoon, as one of the pioneering researchers, addressed this by leveraging shallow parsing and vocabulary profiling. Yoon has actively contributed to the improvement of the assessment landscape, particularly in test security. Her research delved into potential risks associated with automated scoring systems, such as test takers’ attempts to manipulate the system. In 2019, she left ETS and joined EduLab, Inc, expanding her focus on automated learning solutions for non-native writers. She is developing an automated system to identify priority issues of language learners, offering personalized and actionable feedback. She has published more than 50 papers in journals and major conferences, and authored chapters in Automated Speaking Assessment: Using Language Technologies to Score Spontaneous Speech, published by Routledge (2020).