In corpus linguistics, the negative correlation between word frequency and word length is a well–documented phenomenon referred to as Zipf ’s law. This linguistic universal observed by Zipf, which posits that the length of a word is in an inverse relation to its frequency (but not necessarily proportional to), has also been confirmed by numerous studies, and its implications can be observed in different fields such as language teaching and cognitive language processing. However, there is a gap in research data when it comes to studying this phenomenon from the perspective of loanwords. Even though it has been observed that translation equivalents in Croatian generally exist for loanwords or English words that appear 5000 times or more in the Croatian corpus (ENGRI corpus), the question still remains why speakers of Croatian resort to using English words in such cases where first language (L1) equivalents exist. This paper examines the systematicity of the language universal that shorter words are more frequent when it comes to foreign (primarily English) words in the (Croatian) language, i.e. whether the most frequent English words are shorter than their Croatian equivalents. For the purpose of this research, the Database of English words and their equivalents in Croatian was examined. Results indicate that some degree of systematicity between word length and frequency can be observed, but they also highlight the need for incorporating a semantic component into the analysis. The results contribute to the theoretical discussion on language universals, and explain why Croatian users prefer English words and whether language economy is one of the reasons for the use of English words in Croatian.
A Fri, study studied this question.