1 Introduction

CQPweb, as the fourth generation corpus retrieval platform, has been widely used in language research and teaching due to its powerful functions, rich corpus, and free and open source. Many universities and research institutions around the world have adopted CQPweb, whose corpus resources are diverse and mostly publicly available. It is used in research fields such as digital humanities, critical discourse analysis, and sociolinguistics, as well as in teaching applications such as data-driven learning and textbook development. The search function of CQPweb includes simple queries and complex search modes, the latter using CQP syntax, supporting regular expressions and Boolean operations, which is its core function. However, due to the complex syntax concept model of CQP, insufficient relevant literature, and installation dependence on Linux systems, its use is not widespread. With the development of computer virtualization technology, the installation threshold of CQPweb has been lowered, and it is expected that its users will grow, and the use of CQP syntax will also be more widespread. This article aims to simplify the CQP grammar retrieval model, analyze retrieval difficulties, and demonstrate its retrieval function from the perspectives of vocabulary, phrases, and grammar, in order to promote the in-depth application of CQPweb in teaching and research.

2 CQP syntax and CQPweb

CQP syntax plays a core role in the vocabulary generation, index analysis, word collocation calculation, and topic analysis functions of CQPweb, especially in index analysis, where CQPweb automatically converts simple queries into complex retrieval modes supported by CQP syntax. CQP syntax supports regular expressions and Boolean operations, initially customized for corpus workbenches and later widely adopted. Although CQP grammar has a long history, Chinese materials are relatively lacking. The CQP syntax of CQPweb is based on an open-source corpus workbench, but for security reasons, it only supports query syntax and does not include command verbs. Difficulty in obtaining learning materials for CQP grammar, potential conflicts in manual content, and complex retrieval models can lead to difficulties in practical applications. This article aims to simplify the CQP syntax retrieval model and reduce the difficulty of learning and application. CQPweb defaults to simple query mode, while complex search mode requires manual selection of CQP syntax. Users can switch and observe the conversion between simple queries and CQP syntax through the Query history interface to familiarize themselves with the similarities and differences between the two. CQP syntax is more flexible and comprehensive than simple queries, and can complete more retrieval tasks. The relevant images were taken from the BFSU CQPweb platform.

3 CQP Grammar Retrieval Model and Related Concepts

The CQP syntax retrieval model is understood through two paths: bottom-up induction and extraction, and top-down starting from the data model. This section simplifies the CQP grammar retrieval model and related concepts, and Section 4 will provide examples of linguistic level retrieval. The CQP syntax retrieval file is in a vertical format, and the data model includes u003ctextu003e tags, structural attributes (s-attributes), and positional attributes (p-attributes). The location attribute is the retrieval subject, including morphological and part of speech attributes. Word retrieval can use regular expressions and Boolean operations, such as [word=". at"] and [word=". at" u0026 pos="VVD"]. Multi word search is a horizontal extension of word search that supports regular expressions and Boolean operations.

4 Search Examples

Detailed exploration of the application of CQP grammar in phenomenon retrieval at different linguistic levels, including word retrieval, multi word retrieval, grammar retrieval, special retrieval forms, and common retrieval problems. In the word retrieval section, methods for searching by word form, part of speech, other annotated attributes, as well as fuzzy retrieval and joint retrieval were introduced. Special attention was paid to the precautions of CQP syntax in terms of case sensitivity, accuracy of part of speech encoding, and use of regular expressions. The multi word retrieval section discussed the retrieval techniques for continuous phrases, non continuous phrases, and non continuous and positional variant phrases. The grammar retrieval section demonstrates how to simulate grammar category retrieval through part of speech encoding sequences and how to optimize retrieval expressions to improve accuracy. The section on special search forms introduces advanced features such as zero width conditional markers, language component boundary markers, backtracking function labels, and custom macros. Finally, common types of errors when using CQP syntax were summarized, and suggestions for avoiding these errors were provided.

5 Conclusion

CQPweb supports multilingual retrieval, can handle large-scale corpora, is easy to install, and helps popularize CQP syntax. The article simplifies the CQP grammar model, identifies common errors, and demonstrates its retrieval capabilities at the lexical, phrase, and grammatical levels. Domestic universities have developed school-based search manuals based on CQP grammar, and there will be more easily understandable documents in the future. The CQP grammar concept has been incorporated into international standards, and the new language model Ziggurat will expand its retrieval capabilities with broad application prospects. It also has reference value for Sketch Engine and RcppCWB users who use similar retrieval syntax.