Otherness and suspiciousness: a comparative study of public opinions between the Confucius Institute and Goethe-Institut in developing countries

The Confucius Institute, a Chinese government-funded program promoting language and culture, faces challenges in shaping international public opinion. This is partly due to the "spiral of silence," the blockage of semiotic hegemony and linguistic alliances, and the perceived lack of Chinese capacity for global promotion. While developing countries generally hold more favorable views of China compared to developed nations (Pew Research Center, 2019), understanding how the Confucius Institute is linguistically portrayed in developing countries' social media is crucial for overcoming current communication obstacles. Previous studies on public perception of Confucius Institutes suffer from methodological gaps and lack comparative analyses. This study addresses these limitations by including a control group (the Goethe-Institut) and employing Hoey's lexical priming theory enhanced by artificial intelligence (AI) techniques. The research questions explore the differences in lexical priming features between news reports on Confucius Institutes and Goethe-Instituts, their influence on audience attitudes and stances, and how these features shape public opinion images of both institutions. The selection of the Goethe-Institut as a control group is strategic; it's a well-established institution with a positive international image, offering a robust comparison point to understand the different communicative challenges faced by the Confucius Institute.

Existing research on Confucius Institutes (CIs) is diverse, focusing on cross-cultural communication, strategic analysis, and teaching. Studies highlighting CIs' advantages in language education foster positive perceptions (Starr, 2009; Selezneva, 2021). Strategic analyses often view CIs more critically, emphasizing their political implications as diplomatic tools (Lien et al., 2012; Huang et al., 2019) or economic instruments (Li et al., 2009; Lien and Co, 2013). Research on cultural communication is divided; some argue CIs significantly enhance China's soft power (Gill and Huang, 2006; Kluver, 2014; Brazys and Dukalskis, 2019), while others question their effectiveness (Paradise, 2009; Selezneva, 2021; Xie and Page, 2013; Zhou and Luk, 2016) and even highlight concerns about interference with academic freedom (Acquaye, 2020; Yeh et al., 2021). Studies on public opinion towards CIs range from single-nation analyses (An and Xu, 2015; Liu and Zeng, 2017; Zhang, 2021; Peng and Yu, 2016; Min, 2012; Zhang and He, 2016) to global perspectives (Yan, 2018). Methodologies have advanced from qualitative to quantitative analyses and from content-based to discourse analyses (Ye, 2015; Liu and Zeng, 2017; Zhang, 2021). However, objective quantitative comparisons and discourse analyses often heavily rely on researcher interpretation, and previous analyses often focused solely on high-frequency words, neglecting the nuances of contextual meaning. Hoey's lexical priming theory, focusing on collocations, colligations, and semantic associations, addresses these limitations. This theory, combined with AI-powered tools, allows for a more comprehensive analysis of public perception.

This study employs a quantitative corpus-based approach grounded in Hoey's lexical priming theory, analyzing collocations, semantic associations, and colligations. The data source is the NOW Corpus, a large English language corpus containing online news sources. The keywords "Confucius Institute" and "Goethe-Institut" were used to collect data from January 1, 2014, to April 1, 2023, focusing on developing countries (India, Sri Lanka, Pakistan, Bangladesh, Malaysia, Philippines, South Africa, Nigeria, Ghana, Kenya, and Jamaica). Two subcorpora were created: CIDC (Confucius Institute Developing Country Corpus) and GIDC (Goethe-Institut Developing Country Corpus). Data processing involved cleaning, tokenization, and the application of several AI techniques. Pointwise mutual information (PMI) using MI3 (Mutual Information Cubed) formula was calculated to assess the strength of collocations. Word2vec (Skip-gram and CBOW models) generated word vectors to capture semantic similarities. Stanford CoreNLP performed dependency parsing for colligation analysis. Latent Dirichlet Allocation (LDA) extracted topics and their associated words to analyze semantic associations. Cosine similarity was used to measure word vector similarity and Euclidean distance to calculate the distance between target words and topic centers. Statistical analyses compared lexical priming features across CIDC and GIDC, examining concordance patterns, PMI and word vector similarities, semantic associations, and colligations (active/passive voice, subject counts, attribute counts, etc.) to determine differences in linguistic features and usage patterns between the corpora.

The analysis revealed significant differences in lexical priming features between the CIDC and GIDC corpora. Concordance pattern analysis showed a stronger association of "China" and "Chinese" with the Confucius Institute, highlighting its national identity, while the Goethe-Institut's associations were more diverse and geographically spread. PMI analysis revealed that general collocations for the Confucius Institute primarily related to its operational aspects (language teaching, cultural exchange), while specific collocations focused on its establishment and expansion, sometimes conveying a more assertive or even aggressive image. Conversely, the Goethe-Institut exhibited a higher overlap between general and specific collocations, suggesting a more consistent and established image. Semantic association analysis categorized contexts related to the Confucius Institute into events, locations, relationships, and functions. Many positive semantic associations were found surrounding its language-teaching and cultural promotion activities but alongside events that portrayed a more state-controlled image. The Goethe-Institut's semantic associations were more varied, spanning diverse projects, artistic activities, and collaborations with other institutions, suggesting a less government-affiliated image. Colligation analysis demonstrated a greater use of passive voice in reports on the Goethe-Institut compared to the Confucius Institute, where active voice is more prominent. This suggests a difference in how the two institutions are perceived: the Goethe-Institut as a supportive collaborator and the Confucius Institute as an active and possibly assertive entity. Topic modeling (LDA) revealed distinct thematic distributions. The Confucius Institute's topics frequently included establishment events, operational details, and occasional negative news items, while the Goethe-Institut's topics were primarily focused on artistic activities and international collaborations. The research also found that the English name “Confucius Institute” itself may contribute to negative perceptions due to the lack of clear explanation and deliberate rhetoric associating it with impediments to academic freedom.

The findings indicate that the Goethe-Institut possesses a more stable and positive public image in developing countries, primarily associated with cultural exchange and collaboration, while the Confucius Institute's image remains more fluid, strongly linked to its national origin and government sponsorship, sometimes with negative connotations. This difference is attributed to several factors. First, the political context influences public opinion, and while developing nations are less reliant on developed countries, they remain cautious of potential interference with their academic freedom, leading to critical assessment of both institutes, but more so for Confucius Institutes. Second, audience preferences shape media coverage, which often caters to prevailing sentiments. Finally, the ambiguous English name "Confucius Institute" and its association with university campuses, coupled with negative narratives about interference with academic freedom, have contributed to the perception of "otherness" and suspicion. The research highlights the importance of nuanced communication strategies for the Confucius Institute to address these challenges and improve its global image.

This study demonstrates the contrasting public perceptions of the Confucius Institute and Goethe-Institut in developing countries, highlighting the importance of lexical priming in shaping institutional image. The Goethe-Institut presents a stable, positive image emphasizing collaboration and cultural exchange, while the Confucius Institute's image remains closely tied to China, leading to a greater sense of "otherness" and susceptibility to negative narratives. Future research should investigate diverse linguistic contexts, explore alternative collocation metrics, and conduct more comprehensive quotation analyses to further refine our understanding of public perception and inform effective communication strategies. The findings highlight the need for the Confucius Institute to clarify its role and proactively address concerns surrounding academic freedom and political influence to improve its global standing.

This study's limitations include the geographical scope of the NOW Corpus, which may not fully represent the diverse viewpoints of all developing countries. Furthermore, the analysis of certain words, like "said," could benefit from a more in-depth qualitative analysis, and alternative collocation and semantic association measures could provide additional insights. The study primarily uses quantitative methods, potentially overlooking subtle nuances in qualitative aspects of the discourse.