Language and Computers

Markus Dickinson is Assistant Professor at the Department of Linguistics, Indiana University and currently the director of the Computational Linguistics program. His research focuses on improving linguistic annotation for natural language processing technology and automatically analyzing the language of second language learners. Chris Brew is a Senior Research Scientist with the Educational Testing Service in Princeton, where he is currently the scientific lead for the c-rater project on automated short answer grading. He has been active in Natural Language Processing for over 20 years, first in the UK, then as Associate Professor of Linguistics and Computer Science at The Ohio State University, where he co-directed the Speech and Language Technologies Laboratory, as well as the Computational Linguistics Program. Detmar Meurers is Professor of Computational Linguistics and head of the Theoretical Computational Linguistics group at the University of Tübingen. He has a longstanding commitment to teaching Computational Linguistics and Linguistics in a way that combines current technology and research issues with the fundamentals of the field. His research emphasizes the role of linguistic insight and linguistic models in Computational Linguistics. His most recent research adds a focus on theory and applications related to second language acquisition.

• What This Book Is About xiOverview for Instructors xiiiAcknowledgments xvii1 Prologue : Encoding Language on Computers 11.1 Where do we start? 11.1.1 Encoding language 21.2 Writing systems used for human languages 21.2.1 Alphabetic systems 31.2.2 Syllabic systems 61.2.3 Logographic writing systems 81.2.4 Systems with unusual realization 111.2.5 Relation to language 111.3 Encoding written language 121.3.1 Storing information on a computer 121.3.2 Using bytes to store characters 141.4 Encoding spoken language 171.4.1 The nature of speech 171.4.2 Articulatory properties 181.4.3 Acoustic properties 181.4.4 Measuring speech 20Under the Hood 1: Reading a spectrogram 211.4.5 Relating written and spoken language 24Under the Hood 2: Language modeling for automatic speech recognition 262 Writers’ Aids 332.1 Introduction 332.2 Kinds of spelling errors 342.2.1 Nonword errors 352.2.2 Real-word errors 372.3 Spell checkers 382.3.1 Nonword error detection 392.3.2 Isolated-word spelling correction 41Under the Hood 3: Dynamic programming 442.4 Word correction in context 492.4.1 What is grammar? 50Under the Hood 4: Complexity of languages 562.4.2 Techniques for correcting words in context 58Under the Hood 5: Spell checking for web queries 622.5 Style checkers 643 Language Tutoring Systems 693.1 Learning a language 693.2 Computer-assisted language learning 713.3 Why make CALL tools aware of language? 733.4 What is involved in adding linguistic analysis? 763.4.1 Tokenization 763.4.2 Part-of-speech tagging 783.4.3 Beyond words 803.5 An example ICALL system: TAGARELA 813.6 Modeling the learner 834 Searching 914.1 Introduction 914.2 Searching through structured data 934.3 Searching through unstructured data 954.3.1 Information need 954.3.2 Evaluating search results 964.3.3 Example: Searching the web 974.3.4 How search engines work 100Under the Hood 6: A brief tour of HTML 1034.4 Searching semi-structured data with regular expressions 1074.4.1 Syntax of regular expressions 1084.4.2 Grep: An example of using regular expressions 110Under the Hood 7: Finite-state automata 1124.5 Searching text corpora 1154.5.1 Why corpora? 1164.5.2 Annotated language corpora 117Under the Hood 8: Searching for linguistic patterns on the web 1185 Classifying Documents : From Junk Mail Detection to Sentiment Classification 1275.1 Automatic document classification 1275.2 How computers “learn ” 1295.2.1 Supervised learning 1305.2.2 Unsupervised learning 1315.3 Features and evidence 1315.4 Application: Spam filtering 1335.4.1 Base rates 1355.4.2 Payoffs 1395.4.3 Back to documents 1395.5 Some types of document classifiers 1405.5.1 The Naive Bayes classifier 140Under the Hood 9: Naive Bayes 1425.5.2 The perceptron 1455.5.3 Which classifier to use 1485.6 From classification algorithms to context of use 1496 Dialog Systems 1536.1 Computers that “converse”? 1536.2 Why dialogs happen 1556.3 Automating dialog 1566.3.1 Getting started 1566.3.2 Establishing a goal 1576.3.3 Accepting the user ’ s goal 1576.3.4 The caller plays her role 1586.3.5 Giving the answer 1586.3.6 Negotiating the end of the conversation 1596.4 Conventions and framing expectations 1596.4.1 Some framing expectations for games and sports 1606.4.2 The framing expectations for dialogs 1606.5 Properties of dialog 1616.5.1 Dialog moves 1616.5.2 Speech acts 1626.5.3 Conversational maxims 1646.6 Dialog systems and their tasks 1666.7 Eliza 167Under the Hood 10: How Eliza works 1726.8 Spoken dialogs 1746.9 How to evaluate a dialog system 1756.10 Why is dialog important? 1767 Machine Translation Systems 1817.1 Computers that “translate”? 1817.2 Applications of translation 1837.2.1 Translation needs 1837.2.2 What is machine translation really for? 1847.3 Translating Shakespeare 1857.4 The translation triangle 1887.5 Translation and meaning 1917.6 Words and meanings 1937.6.1 Words and other languages 1937.6.2 Synonyms and translation equivalents 1947.7 Word alignment 1947.8 IBM Model 1 198Under the Hood 11: The noisy channel model 200Under the Hood 12: Phrase-based statistical translation 2047.9 Commercial automatic translation 2057.9.1 Translating weather reports 2057.9.2 Translation in the European Union 2077.9.3 Prospects for translators 2088 Epilogue : Impact of Language Technology 215References 221Concept Index 227

“…a broad introduction is also needed by Bachelor’s students in linguistics, and by those studying to become language consultants, translators, and so on. This book would then be the natural choice.”  (Computational Linguistics, 2013)“In general, I can certainly recommend “Language and Computers” as a broad introduction to language technology aimed at the general audience and at students who are curious about modern language technology.”  (Mach Translat, 1 March 2013)