One of the main topics of the book is the comparison between the human and machine translation process, including information on translation theories and models of translation. These are illustrated by so-called Think-Aloud Protocols. The longest part of the book is devoted to the evaluation of machine translation, as well as the comparison of the programs themselves. Using different text types (grammar test, business English, user’s manual, children’s story, scientific text), various evaluation methods (four-scale rating, benchmark test, test suite, error counting, backtranslation) are compared. The final chapter looks at the future perspectives of machine translation with the possible inclusion of expert systems, artificial intelligence, fuzzy logic, neural networks and chaostheory.
Frankfurt/M., Berlin, Bern, Bruxelles, New York, Oxford, Wien, 2001. 323 pp., num. fig. and tables
Contents: Comparison between human and machine translation – Possible standards for machine translation evaluation
– Comparison of machine translation programs as well as the evaluation methods themselves – Future perspective of machine