「出口轉內銷」：中國外交部發言人推特文轉入官媒新聞宣傳研究

Abstract

「講好中國的故事」是習近平作為對外宣傳的口號，然而近年來中國對外的言論中，流露戰狼外交的激進形象，利用中國境內屏蔽的西方社群媒體推特，進行對外宣傳。而中國內部，網路新聞媒體受到政府監管，中國官媒採用外交部發言人推特文，對內傳達發言人在推特上的言論。本研究欲探討什麼樣的外交發言人推特文會被轉入中國官媒，藉由比較有被轉入官媒的推特文和沒有被轉入的推文；本研究的主要問題：1.外交發言人推特文的主題有哪些？轉貼率高的主題為何？2.外交發言人推特文分為有轉組和沒轉組。兩組推文的字詞使用差異為何？

本研究藉由文字探勘的方法，使用非負矩陣分解模型（NMF）和K-means模型進行文本分群，並利用文字散布圖工具（Scattertext）找出兩組推文字詞頻率的差異。研究發現，中國外交部發言人的推文轉貼率，以「美國爭議」相關的主題被轉到國內的機會最高，內銷推文呈現美國負面形象、美中關係問題、美國內部社會種族問題；其次，與「內政發展」相關的推特文會被轉入地方官媒新聞中，主要呈現地方觀光景觀和傳統文化，讓地方人民知曉外交發言人向世界推銷地方特色。

In recent years, the “tell China stories well” diplomatic strategy adopted by Chinese president Xi Jinping has leaned towards an aggressive posture. Part of this so-called “wolf warrior diplomacy” utilizes the Western social media platform Twitter as its tool for external propaganda. At the same time, the Chinese Communist Party (CCP) continues its domestic censorship, which ironically bans the use of Twitter by the broader Chinese public. Interestingly, it can be observed that some of the external propaganda materials initially published on Twitter are translated and republished in the state-run media for domestic consumption, while others are not.

In this study, we examined tweets posted by the two then-Chinese Ministry of Foreign Affairs spokespersons, Hua Chunying and Zhao Lijian. Some of those tweets were only targeting externally (hereinafter referred to as Group Ext.), while some of the tweets were also transferred, in effect, for domestic republication (hereinafter referred to as Group Dom.). We compared these two groups of tweets to determine how they were distinguished in terms of the topics covered and the frequency with certain phrases were used. Based on machine learning methods, the study used non-negative matrix factorization (NMF) and K-means to identify tweets by topic-based clusters. We also analyzed term frequency with Scattertext in both Group Ext. and Group Dom.

The study found that the tweets of Chinese Ministry of Foreign Affairs spokespersons had the highest chance of being transferred to the domestic audience when they were related to "US controversies". Domestic promotional tweets presented negative images of the United States, issues related to US-China relations, and social and racial issues within the United States. Second, tweets related to "internal development" were more likely to be transferred to local government-run media, which mainly presented local tourism and traditional culture, making local people aware of the diplomatic spokesperson promoting local characteristics to the world.

This study presents a data science approach to Chinese domestic propaganda and contributes to an understanding of China’s public diplomacy.