Annabelle Jeon, Seamus Kim, Ani Abramian, Liam Fink
We’ve all celebrated Thanksgiving or some other special occasion when people from opposite sides of the aisle come together to celebrate an obligatory meal in the name of family—and chaos ensues. The estranged uncle is spouting conspiracy theories and someone’s aunt is convinced that raising taxes will be the downfall of the economy. As amusing as this hypothetical scenario is, growing polarization is a very real issue that contributes to increased social fragmentation and wider political turmoil. This study aimed to investigate the source of polarization—specifically as a result of differing interpretations of the same information due to different positions of power. Our sample consisted of students and police officers at the University of California, Los Angeles (UCLA), all of whom were instructed to watch the same video of a protest in Belarus and interviewed to gauge and compare their linguistic behavior. Our results suggest that differences in power do indeed contribute to different interpretations and subsequently different linguistic behavior as those in power, the police officers, showed a markedly neutral and indifferent response in contrast to those without, the students, who demonstrated a clearly supportive response when given the same information.
Introduction and Background
There is a common sentiment that Americans are becoming increasingly polarized (Figure 1, Pew, 2023). Across the political spectrum, many identify the foremost culprit as: misinformation. While there is a growing body of research that points to dangerous levels of misinformation in our current digital age, to attribute polarization solely to misinformation is an oversimplification (Vasist et al., 2023). As information systems researcher Stahl points out, there is no singular “truth” (2006). We each carry with us implicit biases that paint the way we process information. Influenced by his research, we opted to study how divergence arises from people interpreting the same information differently due to identity-based idiosyncrasies.
Our design was influenced by Kahan et al. (2012), which asked subjects of various political beliefs whether the police acted lawfully or not after watching a video of an ambiguous protest to investigate the potential influence of ‘cultural cognition’ on perceptions of legally consequential facts. Since positions of power have been shown to play a significant role in shaping linguistic behavior in social settings, we decided to use relative authority as a factor of interpretative divergence—and—linguistic behavior as a measure for participants’ interpretations (Danescu-Niculescu-Mizil et al. 2012).
Thus, we investigated how UCLA students and UCLA police officers interpreted an ambiguous protest scenario, hypothesizing that officers would use more critical language while students would use more positive terms, despite receiving the same information.
Methods
We employed a qualitative interview design to investigate how relative authority (being a student versus being a police officer) might correlate with differences in interpretation of the same video, as evidenced by linguistic behavior. We interviewed 6 participants–3 UCLA students and 3 UCPD officers–in-person. First, we asked about their age, ethnicity, gender, rank/year, and political leaning to account for potential both inter- and intra-group differences due to factors beyond position of authority. Then, the interviewer provided a single sentence of context for the video, describing it as a factory strike over stagnant wages in Belarus. The participants then viewed a ten second clip of the strike and were asked at least four questions:
- How would you characterize the people striking?
- How would you characterize the overall atmosphere of the protest?
- What do you believe is the primary motivation behind the protest?
- What impact do you think the protest has on the local community?
Interviews were to be 10-15 minutes long to provide enough data for rigorous linguistic analysis. In case responses ran under, interviewers were instructed to ask follow-up questions for each primary question, including:
- What from the video are you basing this perspective on?
- How do you think people with opposite political beliefs from you would respond to this question?
The questions were designed to elicit elaboration beyond the immediate facts of the video (there are people; they’re carrying signs; etc.), without insinuating any particular interpretation.
We analyzed transcriptions of their responses for diction (buzzwords, specific phrasings, etc.) to identify any notable patterns in linguistic behavior, which were then compared to reveal potential intra- and inter-group differences in the interpretation of the scenario. Finally, the findings were situated within a larger societal context to contribute to ongoing discussions of the relationship between information and perspective.
Results and Analysis
We engaged in two levels of analysis, starting with an examination of each transcript for keywords and phrases, patterns in descriptions, and other notable features. Below is a short excerpt from one of the responses collected.
Our data revealed that all participants shared a relatively positive view of protests. Though the difference in results were not as significant and explicit as expected, there were still clear linguistic differences. In contrast to students, who emphasized overt and immediate solidarity from the start, stating that the protestors were “supportive of each other” and “perseverant”, police officers were more hesitant in their responses, describing the protests as “passionate… but pretty peaceful” and “not very aggressive” before expressing mild approval.
The officers also mentioned that the protest could “cause disturbances” and “be a little disruptive” when asked about the possible impact on the community. Meanwhile, students continued to express support, characterizing the protests as a communal movement for the greater good. One student even expressed personal willingness to join, citing reasons such as fair wages, drawing from their own protest experiences. To summarize, all three students demonstrated clear patterns of empathetic identification with the protestors, while the officers exhibited a more distant, moderate sense of support.
Discussion and Conclusions
Our data reveals notable linguistic patterns amongst the police officers that differ from those of the students, strongly suggesting differences in the interpretation of the same scenario, which moderately aligns with our hypothesis. It’s important to note, however, that while the students used clearly positive diction to express support, the police officers were implicitly critical rather than using outright critical diction in describing the protestors. In sum, the discursive behavior of the students in comparison to the police officers suggested that the students by default held the protest in a positive regard, while the officers determined that the protest was generally acceptable but only because it met certain criteria (non-aggressive, minimal disruption, within scope of rights, etc.). Furthermore, our results align with that of Karjus’ study, which demonstrated a strong correlation between linguistic differences and American political polarization as each political faction constructs their own linguistic echo chamber of shared narratives of certain words and phrases, reinforcing their beliefs and causing further division (2023).
An unexpected challenge in gathering and analyzing our data was the consistently brief responses from officers, which were about a quarter to half as long as those from students. This reduced data volume may decrease confidence in our findings regarding police perspectives. However, the brevity could also suggest neutrality or indifference, reflecting a lack of strong opinion. Another plausible explanation: because of their stronger authority relative to the student, the officers felt less inclined to answer as thoroughly as the students. In this way, our positionality as students may have also affected the results of this study, a reminder of the elusive nature of the ecological validity that we anthropologists continue to seek.
This study contributes to research on communicative divergence and polarization by highlighting the role of relative authority/power in shaping differing views, even when information is controlled for. We urge future studies to explore additional factors such as socioeconomic status, ethnic/cultural background, and region, which may also influence divergent interpretations. Understanding these factors is crucial for modifying approaches to cooperation and peaceful resolution to combat polarization. Rather than focusing solely on ‘the facts,’ advocates must acknowledge and align with the underlying principles driving diverse interpretations.
References
Danescu-Niculescu-Mizil, C., Lee, L., Pang, B., & Kleinberg, J. (2012, April). Echoes of Power: Language Effects and Power Differences in Social Interaction PDF. Ithaca; Cornell University Department of Computer Science. https://www.cs.cornell.edu/~cristian/Echoes_of_power_files/echoes_of_power.pdf
Kahan, D. M., Hoffman, D. A., Braman, D., Evans, D., & Rachlinski, J. J. (2012). “They Saw a Protest”: Cognitive Illiberalism and the Speechconduct Distinction. Stanford Law Review, 64(4), 851–906. http://www.jstor.org/stable/41511108
Karjus, A., & Cuskley, C. (2023, September 4). Evolving linguistic divergence on Polarizing Social Media. arXiv.org. https://arxiv.org/abs/2309.01659
Pew Research Center. (2023). Americans’ feelings about politics, polarization and the tone of political discourse. https://www.pewresearch.org/politics/2023/09/19/americans-feelings-about-politics-polarization-and-the-tone-of-political-discourse/
Stahl, B. C. (2006). On the difference or equality of information, misinformation, and disinformation: A critical research perspective. Informing Science, 9, 83.
Vasist, P. N., Chatterjee, D., & Krishnan, S. (2023). The Polarizing Impact of Political Disinformation and Hate Speech: A Cross-country Configural Narrative. Information systems frontiers: a journal of research and innovation, 1–26. Advance online publication. https://doi.org/10.1007/s10796-023-10390-w