1 Introduction

The parallel binomial in English is a common language form composed of words of the same word class, which enhances the accuracy, fluency, and interest of language expression. Studying the constraining factors of parallel binomial word order is of great significance for non-native English speakers to improve their language expression ability. Existing research has not delved into the differences in constraining factors between parallel binomial words of different word classes. This study focuses on the English parallel binomial connected by the conjunction 'and', exploring the constraints of different factors on word order, and whether there are differences in the constraint ranking of parallel binomial in different parts of speech (nouns and adjectives).

2. Research background and research questions

Reviewing the history of parallel binomial word order research, it is pointed out that Pinker u0026 Birdsong's (1979) psychological experiment found that phonological factors affect word order selection, but do not reflect the real language processing process. Cooper U0026 Ross (1975) proposed the "I first principle" to explain psychological markers, while Kelly et al. (1986) found that cognitive preference for "I" is one of the reasons for word order differences. Fenk Oczlon (1989) found that word frequency is a determining factor, but the research methods are limited. Mollin (2014) reanalyzed the data and emphasized the importance of semantic factors. This study aims to explore the impact of different constraining factors on word order, whether the same factor has the same effect on noun and adjective word order, and whether the determining factor is universal.

3 Theoretical Framework and Research Methods

This chapter introduces the corpus based paradigm of variation linguistics research, emphasizing the heterogeneity of language, dynamic changes, and the connection between social environment. The study hypothesizes that language variation is the only form of language existence and advocates for the collection and analysis of natural language materials, particularly oral expressions. Adopting a probabilistic grammar perspective, it is believed that the selection of language variants is influenced by multiple factors. The research method is multi factor analysis, which identifies significant factors and their interactive effects that affect the parallel binomial word order in English through large-scale corpus retrieval and manual annotation, providing a basis for theoretical explanation.

4 Research Design

This chapter introduces the method of using the British National Corpus (BNC) to study the parallel binomial word order in English. Researchers extracted over 700000 parallel binomial terms using Python and regular expressions, and selected the top 600 noun and adjective parallel binomial terms in frequency ranking. The study considered the ratio of written and spoken language, excluding the parallel binomial of repetition, coordinating phrases, and multi word components. The researchers referred to previous studies and selected 10 potential limiting factors that may affect word order, including semantics, prosodic speech, non prosodic speech, and other factors, and annotated them using binary and ternary methods. The study used the reversibility curve and reversibility score to analyze the reversibility of the parallel binomial, where the reversibility score reflects the relative proportion of dominant and disadvantaged word orders in the corpus and is used to evaluate the degree of word order solidification.

5 Data Analysis and Discussion

This chapter cross validated the constraining factors of English parallel binomial word order through success rate ranking and random forest analysis. Research has found that the number of syllables is the main factor affecting the binomial word order of nouns, while the brightness of the tail sound has a significant impact on the binomial word order of adjectives. The parallel binomial of nouns is influenced by multiple factors, among which the success rate of syllable quantity, tail sound brightness, similarity, power, first sound brightness, and nucleus vowel length is higher. In the binomial system of adjectives, tail tone brightness, psychological markers, syllable count, and word frequency are the main influencing factors. The results of random forest analysis are consistent with the success rate ranking, further confirming the importance of these factors. However, there are also some special cases where psychological markers and resemblance may take precedence over other factors in some parallel binomial models. In addition, the completely irreversible parallel binomial may also experience inversion in certain situations. These findings provide a new perspective for understanding the word order selection of the parallel binomial system in English and offer references for language standardization.

6 Conclusion

This study utilized the BNC corpus and R software to investigate the reversibility and limiting factors of the binomial relationship between nouns and adjectives through random forest analysis. Research has found that the number of syllables has a significant impact on noun word order, while tail sound brightness has a significant impact on adjective word order. There are limitations to the research, such as a small amount of corpus and a lack of universality in semantic factor annotation. It is recommended that future research use large corpora for validation and distinguish between written and spoken language, as well as parallel binomial systems in different language domains.