Natural language

Natural language
"Natural Language: A human language whose rules have evolved from current usage, as opposed to an artificial language whose rules are prescribed prior to its construction and use, as in the case of a computer language. In database searching, a natural language search allows the user to type words as input in the same way that a person normally speaks them." (Holliday, 2001).

"The term natural language is used to distinguish languages spoken and signed (by hand signals and facial expressions) by humans for general-purpose communication from constructs such as writing, computer-programming languages or the "languages" used in the study of formal logic, especially mathematical logic. In the philosophy of language, the term ordinary language is sometimes used as synonymous with natural (as opposed to mathematical or logical) language. Natural language is also considered a field of weak artificial intelligence. The term has been adopted to describe computer input terms and language modeled after or based on natural human languages rather than the artificial syntax and terms of computer languages, particularly in the areas of search engines or search functions." (Wikipedia, 2005).

In Library and Information Science (LIS) has the term "natural language" in particular been used in opposition to "controlled vocabulary". Natural language is thus the language of the documents represented in a given system, while controlled vocabularies are classification codes, descriptors or other symbols taken from a classification system, a thesaurus or other kind of system established before a specific document is being indexed. In other words: natural languages are the opposite of the artificial languages established by the LIS community itself.

When searching a given database the searcher may choose to search "natural language fields" such as titles, abstracts or full text, or to search "controlled fields" such as classification codes or descriptors (or, of course to combine them).

The term natural language is also often used about natural language processing, natural language query and natural language interface.

The expression "natural language" is of course relative. In a legal text may the abstract be written in "natural language", "natural" in relation to, for example, a classification code like UDC. Legal language is, however, a language for special purposes (LSP), which is not "natural" compared to general languages like English or Danish (Although some linguists argue that structurally are LSPs not languages at all).

Svenonius (1986) is skeptical towards the tendency to use standards for controlled vocabularies rather than to examine the natural language in a given discipline and to base the construction of controlled vocabularies on such investigations. Bhattacharyya (1974) represent an attempt to explore different LSPs for this purpose.

The most ambitious attempt to control a LSPs is the US National Library of Medicine's Uniform Medical Language System (UMLS).

Literature:

Bhattacharyya, K. (1974). Effectiveness of Natural Language in Science Indexing and Retrieval. Journal of Documentation, vol. 30(3), 235-254.

Holliday, W. (2001). The Native American Research Tutorial. http://www.sjc.cc.nm.us/pages/1763.asp

Lancaster, W. (1989). Natural language versus controlled language; a thirty year review of the literature of information science. IN: Perspectives in information management 1. Ed. by C. Oppenheim et al. London: Butterworth. (Reprinted in Lancaster, W.: Indexing and Abstracting in Theory and Practice, 1991, pp. 193-218).

Smeaton, A. F. (1992). Progress in the Application of Natural Language Processing to Information Retrieval Tasks. The Computer Journal, 36(3), 268-278.

Svenonius, E. (1986). Unanswered Questions in the Design of Controlled Vocabularies. Journal of the American Society for Information Science, 37(5), 331-340.

Wikipedia, the free encyclopedia. (2005). Natural language. http://en.wikipedia.org/wiki/Natural_language

Birger Hjørland

Last edited: 25-03-2006

Home