Keywords: Social media, big data, sentiment analysis, and emerging technologies

This work aims to analyze Chinese people’s social media language use and meaning construction for “milk”, the most common consumed drink which reveals ordinary people’s ideological changes.
“Language is not a mirror of reflected thought, but a ‘mode of action’” (Malinowski, 1923). It is also a system of choice in which the use of words and the way of expressing are organized with specific purpose. For Michel Pêcheux (1969, 1982), individual’s use of words and expressions is more determined by the economic, institutional and ideological factors tied to a position occupied by he or her in a social structure. Social communication is related to one’s ideological formations, which determine what can and must be said from a given social or ideological position (Wallis, 1998). According to the Belief Gap Hypothesis, the discord about a specific issue comes rarely from pure scientific disagreement (Hindman, 2009). It is ideology that determinates people’s belief about the reality and their judgment about the same issue to be true of false (Hindman, 2009; Gaziano, 2014).
Ideologies are not unchangeable. Jost et al. suggested that most people are ideologically inclined (Jost, Nosek, & Gosling, 2008). They possess an ideological belief system with a balanced left-right dimension. The choices are made according to some psychological reasons in individual’s needs to reduce uncertainty and threat. Thus the imbalance of this dimension would conduce to people’s ideological shift. Events like the 9/11 (Bonanno, & Jost, 2006), the 2004 Madrid terrorist attack (Echebarria & Fernandez, 2006) and the European migrant crisis (Erel, 2017) had led more people to endorse more conservative.
In China, food can reveal one’s social ideology and interpersonal relationship with others (Ma, 2015). However, due to the outbreak of many food safety scandals (especially the Sanlu Milk Powder scandal in 2008), food safety became people’s major concern since the past years (Qiao, Guo, & Klein 2012). Existing research focuses mostly on obvious consequences of those incidents but rarely on people’s ideological behind their behavior. As different ideological beliefs may form different lifestyle patterns (McIntosh, 1995; Wang, Worsley, & Cunningham, 2009), we hypothesis that: The 2008 Milk scandal has not only harmed Chinese’ people’s confidence in local infant formula, but also left permanent influence on their perception of “milk”. A widening ideological gap has appeared between ordinary people’s perception of “local milk” and “imported milk” as social and culture symbols.
In this case, we collected people’s posts related to “milk”, specially to “local milk” and “imported milk” from Sina Weibo (one of the biggest social media in China) with Python. Referring to Pêcheux’s method of automatic discourse analysis (AAD), we will firstly transpose manually part of social media users’ posts (discourse) into standardized syntactic form, in order to establish a machine training model, which allows us to make inter-discourse comparisons. Relations between different semantic domains of analyzed corpus will be then interpreted in order to result in a semantic network representing an ideological structure.