As a society we face a range of complex, systemic issues. Often, we use language to understand and clarify both problems and solutions. Take current debates on taxation. The morality of reducing marginal tax rates remains open for debate, but the phrase “reducing marginal tax rates” lacks any moral color—it is relatively neutral. On the other hand, the phrase “tax relief” suggests positive values of “compassion,” “care,” or even “charity.” The very language that we use to describe these issues shapes our moral perceptions.

Different groups emphasize different moral foundations.

With recent developments in data analysis tools, we can access such moral language at large scales. Linguistic analyses of massive collections of text have proven useful for discovering both what people talk about [1] and what they feel about what they talk about (e.g. implied sentiments [2] and various implied personality traits [3]).

Recently, moral psychologists [4] have applied linguistic analysis to the problem of uncovering the differing “Moral Foundations” that different cultural groups rely upon when constructing persuasive messages (the moral color of the message). Moral Foundations Theory [cf. 5] posits six cross-cultural themes: 1. Security, 2. Justice, 3. Autonomy, 4. Community, 5. Authority, and 6. Purity. Table 1 lists some of the core concepts entailed by each foundation. Different cultures and subcultures, at different times, emphasize particular moral foundations over others. For example, Graham et al. [4] found that “liberal” speakers color their messages with the language of security and justice; the remaining foundations dominate the “conservative” palette.

Table 1. Examples of core concepts entailed by six moral foundations

Foundation Valence Examples of core concepts
Security Observation: Empathy, kindness, gentleness, nurturance
Violation: Violence, abuse, irascibility, meanness
Justice Observation: Equity, proportionality, truthfulness, friendliness
Violation: Partiality, prejudice, fraudulence, hostility
Autonomy Observation: Liberty, independence, consent, self-determination
Violation: Oppression, subservience, coercion, subjection
Community Observation: Loyalty, dutifulness, patriotism, self-sacrifice
Violation: Betrayal, dereliction, treachery, egocentrism
Authority Observation: Deference, respect, legitimacy, hegemony
Violation: Recalcitrance, affront, subversion, heterodoxy
Purity Observation: Sanctity, temperance, innocence, healthfulness
Violation: Profanity, indulgence, corruption, infirmity

 

At HI, we care most about the messages delivered in entertainment, “the stuff between commercials.” Amid the proliferation of media choices, network television remains the loudest voice in the lives of many Americans: according to the Bureau of Labor Statistics, television viewing accounts for more than half of daily leisure time activity for Americans over the age of 15 [6]. Fifty-plus years of television research has repeatedly shown that television viewing can influence beliefs about the prevalence of various phenomena—crime, affluence, marital discord, etc.—and various occupations—doctors, lawyers, police, etc. [e.g. 7]. For example, when making intuitive judgements, heavy viewers often overestimate the prevalence of crime in the real world, apparently because the prevalence of on-screen crime increases the availability of criminal incidents in memory. At HI, we wondered which moral foundations prevail on network television: how much does the moral palette of television resemble (the various subgroups of) American society, and how much does the moral palette of today’s television resemble the entertainment of yesteryear?

The next several posts address these questions. Our findings should serve to pump intuitions about how television programming may affect or reflect the Moral Foundations on which Americans rely. These intuitions can drive further research. In this post, we will render the moral palette of the 2011-12 network television season.

Does the moral palette of television resemble American society?

We categorized the language used on television through quantitative content analysis [8]. We started with the moral lexicon developed by Graham et al. [4], which categorized 324 words and word stems that either support or violate each of the Moral foundations. For example, the word stem peace* (capturing peace, peaceful, peaceable, etc.) supports the Security foundation, and the word stem kill* (capturing kill, killer, killing, etc.) violates the Security foundation.

Pilot tests on moralizing texts convinced us that we needed a richer lexicon (many more words and phrases); also, the publicly available dictionary lacks a lexicon for the Autonomy foundation. So we expanded the lexicon following similar procedures to Graham et al. [4]. For example, we generated the Autonomy lexicon by iteratively looking up synonyms, antonyms, and associated phrases for the core concepts [9]. Three judges reduced the resulting 10K words and phrases to 1,069. The judges deleted words and phrases when primary dictionary [10] definitions and usage examples departed too far from the Autonomy foundation. Our final moral lexicon, including 6722 words, word stems, and phrases, is available upon request (send an email to info AT harmony-institute.org).

We applied [11] this expanded moral lexicon to all available transcripts of 85 primetime, narrative series (i.e. not sports or “reality” programming) that aired on five major networks—ABC, CBS, CW, FOX, and NBC—during the 2011-2012 television season [12]. Figure 1 (below) presents a summary of our analysis.

Figure 1. The moral palette of network television: percentages of words & phrases that observe or violate each of six moral foundations.

Series.MoralsStacked

On average, the moral lexicon accounts for 14.2% of all words spoken during the 2011-12 network television season [13]. This is what we mean by moral color: the content of television (the other 85.8% of words) could refer to any number of neutral topics (think, “marginal tax rates”); the six bars reveal how the makers of television chose to color in those otherwise gray topics. Specifically, the six bars represent the average percentages of words and phrases that relate to each of the six moral foundations. The blue portion of each bar represents the percentage of words that generally support each foundation; red represents the percentage of words that generally violate each foundation [14].

Like “liberal” speakers in Graham et al. [4], the language of Security (5.8% of words) featured prominently in dialogue during the 2011-12 network television season. The dialogue focused slightly more on abuses (Security violations, 3.4% of words) rather than kindnesses (Security observances, 2.4% of words), but the Foundation of Security overall was generally important. Like “conservative” speakers in Graham et al. [4], the language of Authority (4.6% of words) and Community (3.8% of words) also colored dialogue during the 2011-12 network television season. When relying on these foundations, the dialogue emphasized respect (Authority observances, 3.9% of words) and loyalty (Community observances, 2.8% of words) over affronts (Authority violations, 0.7% of words) and betrayals (Community violations, 0.9% of words).

Altogether, one can sum up the overall moral message of network television during the 2011-12 season as “be unhurtful (perhaps helpful) and respectful to those in your in-group.”

Disclaimer and Caveats

Readers should accept this interpretation with caution. We analyzed viewer-generated transcripts of a subset of series from five networks during a single television season; the actual scripts of all series from all networks over multiple seasons would much increase reliability. Moreover, our analysis relied upon a newly-developed moral lexicon; a reliable lexicon will require validation beyond reference books and three judges. Interested readers may request technical details about lexicon development, the data, and our analysis. We welcome any ideas or suggestions about how to proceed with further research.

Next Up!

We compare the overall moral palette of network television to the moral priorities of respondents to the YourMorals.org survey.

References & Notes

1. Blei, D. M., Ng, A. Y., and Jordan, M. I. (2003). Latent dirichlet allocation. The Journal of Machine Learning Research, 3(4-5):993-1022.

2. Pang, B. and Lee, L. (2008). Opinion mining and sentiment analysis. Foundations and Trends in Information Retrieval, 2(1-2):1-135.

3. Pennebaker, J. W. and King, L. A. (1999). Linguistic styles: Language use as an individual difference. Journal of Personality and Social Psychology, 77(6):1296-1312.

4. Graham, J., Haidt, J., and Nosek, B. A. (2009). Liberals and conservatives rely on different sets of moral foundations. Journal of Personality and Social Psychology, 96(5):1029-1046.

5. Haidt, J. and Joseph, C. (2004). Intuitive ethics: how innately prepared intuitions generate culturally variable virtues. Daedalus, 133(4):55-66.

6. American Time Use Survey: Charts by Topic: Leisure and sports activities. (n.d.). Retrieved June 18, 2013, from http://www.bls.gov/tus/charts/leisure.htm

7. Shrum, L. J., Wyer, R. S., and O’Guinn, T. C. (1998). The effects of television consumption on social perceptions: The use of priming procedures to investigate psychological processes. Journal of Consumer Research, 24(4):447-458.

8. Silverman, D. (1993). Interpreting qualitative data: Methods for analyzing talk, text and interaction. SAGE Publications, London.

9. Using Roget’s Thesaurus (thesaurus.com)

10. Specifically the Oxford English Dictionary (oed.com), the Merriam-Webster Dictionary (merriam-webster.com), and the WordNet Lexical Database for English (wordnet.princeton.edu/)

11. Using the Linguistic Inquiry and Word Count program to automate the counting; see Pennebaker, J. W., Booth, R. J., and Francis, M. E. (2007). Linguistic Inquiry and Word Count: LIWC2007, Operator’s manual. LIWC.net, Austin, TX.

12. Opensubtitles.org supplied the transcripts. The user community, in turn, supplies the transcripts to opensubtitles.org. As of [6/18/2013], no one has transcribed any episodes of House, M.D., Law & Order: Special Victims Unit, or The Office.

13. We calculated these percentages as the frequency of foundational words and phrases, divided by the total frequency of all words (the denominator excluded the frequency of articles and pronouns).

14. The italics serve as a nod to those whose observance of a particular moral foundation requires what others may see as a violation of that foundation. For example, among some people in the U.S. and Greece, “monarchy” always violates legitimate Authority; in fact, for some Greeks, “regicide” is a virtue.

 

This is the first post in a three part series.

Part Two: Comparing the moral language of television to that of its viewers

Part Three: Moral resonance across entertainment history

Top image: Flickr user Luca Rossato | Creative Commons