Abstract
The present paper employs a synergetic approach to exploring the relationship between the length and frequency among English multiword formulaic sequences. It attempts to test whether Zipf’s assumption on the inverse relationship between word length and word frequency, as well as the synergetic model constructed at lexical level, can be extended to multiword formulaic sequences. A corpus-driven approach is adopted to acquire sufficient data. Results show partial applicability of the assumption to a small part of formulaic sequences: sequences of words with whole structure and complete meaning. However, the majority of formulaic sequences, lexical bundles, are proved to be an exception to the conventional rule. The paper tries to offer explanation by exploring further the features of lexical bundles. It concludes that the synergetic system of length and frequency among lexical bundles is operated under the compromise between the requirements of minimization of production efforts and the requirements of maximization of information. Based on the analysis of synergetic relationship between length and frequency, this research shall prove that language is a self-regulating system, and that observed language phenomena can be understood and explained by laws and models. Besides, this study also includes register differences and offers a new perspective to register analysis.
Notes
1. ‘That the magnitude of words tends, on the whole, to stand in an inverse (not necessarily proportionate) relationship to the number of occurrences’ (Zipf, Citation1935, p. 25).
2. FLOB is a one-million-word corpus covering 15 text categories. It equals twice the size of LLC (500,000 words). Therefore, only half of FLOB (500,000 words) was used in this study to keep balance between written and spoken texts. The cutoff was made based on text categories: texts in FLOB are divided into 15 register categories and marked from A01to R09. The first half texts in each text category were selected as the database for this study.