Despite repeated denials that transmissional (syntactic) information can be used to estimate semantic information, a new and working method is offered giving the possibility of estimating quantitatively the semantic content of any text-segment.
The method is based on prediction of speech elements and enables us to measure the meaning which the reader extracts from the text, and on the basis of which he predicts the further meaning of the same text.
The prediction is carried out on the syntactical level: good prediction decreases the syntactical information of a text more than a bad prediction. If one knows two words from the beginning of a text, one can predict better than if he knows one word.
Get full access to this article
View all access options for this article.
References
1.
Bar-Hillel, Z. and Carnap, R. ( 1953). Semantic information. Brit. F. Philos. Sci., 4.
2.
Bogodist, V.I., Georgiev, H.C., Pestunova, V.N. and Raitar, S.V. ( 1971). Izmerenie semantičeskoj informacii teksta . In Piotrovskij, R. (ed.), Avtomatičeskaja Pererabotka Teksta Metodami Prikladnoj Lingvistiki, Materiali Vsesojuznoj Konferencii (Kishinev).
3.
Georgiev, H.C. ( 1973). Entropia na Bolgarskia jezik, Ban, Sofia. In Slavističen Sbornic.
