Computational Studies in Terminology

Reinhart HerzogCOLING 1969

TL;DR

This paper, presented in 1969, highlights the growing importance of terminology in applied linguistics and argues for the necessity of computers to manage its complexity. It proposes that computational linguistics techniques, originally developed for machine translation, can be effectively applied to automate the creation of bilingual and multilingual technical dictionaries.

What Problem Does This Paper Address?

The paper addresses the challenge of documenting and controlling the ever-expanding and complex world of specialized vocabularies, known as terminologies. As technology and science advance, new terms emerge rapidly, making it difficult for humans to keep pace. This is particularly problematic for international companies that rely on precise and standardized terminology for communication, whether through product documentation or translation. Traditional methods of compiling dictionaries are becoming insufficient to handle the sheer volume and dynamic nature of these specialized vocabularies.

Methods

The paper suggests leveraging computational linguistics techniques for the task of terminology management. Specifically, it points to methods and procedures originally developed for automatic sentence analysis and mechanical translation. These techniques can be adapted to:

  • Isolate relevant lexical items: Computers can be programmed to sift through large volumes of terminological texts and extract key terms.
  • Determine meanings: Algorithms can help analyze texts to identify the various meanings and nuances associated with specific terms.
  • Automate compilation: The core proposal is to use these computational approaches to automate the process of creating bilingual and multilingual technical dictionaries. This involves using computers to assist in research and control over terminological data.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • The growing necessity of computational power: The paper strongly advocates that managing the complexity and growth of terminologies is practically impossible without the aid of computers.
  • Cross-pollination of NLP techniques: It highlights the potential for techniques developed in one area of computational linguistics (like machine translation) to be highly effective in another (like terminology management).
  • Economic drivers for better terminology: The paper emphasizes that industry's heavy reliance on clear and standardized communication makes automated terminology management a valuable and practical pursuit.
  • The future of dictionary creation: It foresees a future where computers play a central role in the laborious and time-consuming task of compiling comprehensive technical dictionaries.

Key Results

Not discussed in the paper. The paper is more of a position statement and a proposal for future research and development rather than a report on experimental results.

Limitations and Future Work

The paper implicitly acknowledges the limitations of manual methods. The future work suggested is the application and further development of computational linguistics methods for terminology management, particularly for the automated compilation of bilingual and multilingual technical dictionaries. The paper serves as a foundational argument for why this direction is necessary and promising.

Summary updated

BibTeX
@inproceedings{herzog1969computational,
  author = {Reinhart Herzog},
  title = {{Computational Studies in Terminology}},
  booktitle = {International Conference on Computational Linguistics},
  year = {1969},
}
Figure