Are meat alternatives a moral concern? A comparison of English and Japanese tweets

Concerns are growing about the negative impact of increased meat consumption on human health, the environment, and animal welfare, and meat alternatives—primarily plant-based products designed to mimic meat—are being considered as potential solutions. Despite promotion by researchers and governments, consumers have been reluctant due to perceived inferior taste, higher cost, and limited convenience or availability. However, market investment, product accessibility (e.g., availability in restaurants and supermarkets), and broader sustainability initiatives (e.g., UN SDGs) appear to be shifting attitudes. In Japan, although soy-based foods are familiar, adoption of modern meat alternatives that closely mimic meat has been slower than in Western countries, with limited public awareness; consumers with stronger scientific interest are more open to purchasing. By contrast, the U.S. has led the market, including approval for cultured meat sales in 2023, and alternative options are widespread in common restaurants, though there is pushback (e.g., labeling regulations). Earlier research emphasized taste, health, and convenience over moral motivations (environment, animal welfare). Yet broader social changes suggest a shift in consumer interests: while in the mid-2000s few considered reducing meat for social reasons, by the 2010s more consumers supported meat reduction for sustainability and policy measures favoring sustainable products. Social media provides a rich, time-stamped, multilingual source to observe evolving consumer discussions and preferences without recall bias. We leverage Twitter data to examine temporal and cross-cultural shifts in interest and moralization around meat alternatives, focusing on English (likely reflecting North American views) and Japanese tweets, given differing market maturity and food cultures. The purpose is to reveal changes over time and across cultures, moving beyond taste-focused promotion to include health and ethical considerations, to inform strategies that may help popularize meat alternatives and address food-related social problems.

Prior work has generally found that non-moral factors—especially taste, health benefits, and convenience—drive acceptance of meat alternatives, with consumers more positive toward plant-based options than cultured or insect-based products. Studies stressed the importance of making alternatives resemble meat in taste, smell, texture, and product form to encourage adoption. However, evidence from the 2000s to the 2010s indicates changing social attitudes: initial low acknowledgment of plant-based diets’ effectiveness in addressing social issues shifted toward greater willingness to reduce meat for sustainability and support for related policies. Social media analyses have also captured growing eco-friendly attitudes and stable links between food preferences and broader health, political, and economic preferences, as well as distinctions between “vegan” and “plant-based” consumer identities. These strands suggest evolving consumer values that may elevate moral motivations (environmental protection, animal welfare) alongside or even above sensory and health considerations.

Data source and collection: Tweets were collected via Twitter Academic API v2. The English dataset comprised all tweets from 2006–2021 mentioning English terms related to meat alternatives (e.g., “alternative meat”, “fake meat”, “plant-based meat”, “soy-based meat”), totaling 770,049 tweets. The Japanese dataset comprised 368,262 tweets from 2008–2021 using Japanese counterparts (代替肉, フェイクミート, 植物肉, 大豆ミート). Preprocessing: To focus on spontaneous user posts (“organic tweets”), retweets and repeated posts by the same user with identical content were removed. Language filters excluded non-English tweets from the English dataset and non-Japanese from the Japanese dataset. Final datasets contained 253,860 English tweets from 156,754 users and 134,091 Japanese tweets from 66,342 users. Text processing: Tweets were tokenized into morphemes; symbols, emojis, URLs, and mentions were removed; English words were lowercased; Japanese characters were normalized (Japanese to full-width, English to half-width). Morphological analysis used NLTK (English) and MeCab with mecab-ipadic-NEologd (Japanese). Measuring consumer interest: A list of motivational words reflecting food choice motivations (derived from Onwezen et al., 2019) was compiled into categories (e.g., sensory appeal, mood, familiarity, weight control, health, animal welfare, environmental protection, natural context, political values, price, convenience). Co-occurrence of these words with alternative meat mentions was counted in each tweet (with wildcard handling, e.g., “animal*” matching “animals”) to infer interest trends over time. Measuring moral foundations: To assess moralization, the study used the Moral Foundations Dictionary (MFD) and the Japanese MFD (J-MFD) to count co-occurrences of moral words (across Care/Harm, Fairness/Cheating, In-group/Betrayal, Authority/Subversion, Purity/Degradation; both virtue and vice terms) with alternative meat tweets. Temporal trends in moral word usage rates were analyzed overall and conditionally within key motivational categories (e.g., animal welfare, environment). Analyses included proportions of category mentions over time, normalized trend plots (base year 2011), word-level frequencies (2012, 2015, 2018, 2021), differences over 2012–2021, and conditional probabilities of moral word usage in tweets mentioning selected categories.

• Tweet volumes: Organic tweets about meat alternatives rose sharply in both languages, with a surge in 2019 coinciding with major market events (Beyond Meat IPO; fast-food launches).
• Diversification of interests: Discussions initially focused on sensory appeal (taste), but by the late 2010s broadened to health, animal welfare, environment, and price.
• Declining sensory emphasis: Sensory-related words decreased annually in both languages.
• Rising animal, environment, and health: Both English and Japanese showed increased mentions of animal welfare and environmental protection; health remained prominent but showed relative decline.
• Environment grew fastest: Normalized trends show environmental terms had the most rapid growth (2011–2021) in both languages.
• Price sensitivity increased: Mentions of price rose notably, especially in English (tipping point in 2019; in Japanese in 2021), indicating increasing attention to affordability as consumers consider adoption.
• Word-level shifts (English, 2012→2021): health* decreased 36%→26% (−10pp); taste 30%→23% (−7pp); animal* 11%→18% (+7pp); environment* 3%→9% (+6pp); price 1%→7% (+6pp); tasty 8%→3% (−5pp); texture 3%→5% (+2pp); weight 4%→1% (−3pp).
• Word-level shifts (Japanese, 2012→2021): 美味しい(tasty) 46%→36% (−10pp); 味(taste*) 15%→21% (+6pp); 動物(animal*) 4%→10% (+6pp); 環境(environment*) 0%→9% (+9pp); 健康(health*) 10%→8% (−2pp); 食感(texture) 8%→5% (−3pp); 見た目(appearance) 5%→2% (−3pp); 価格(price) 1%→2% (+1pp).
• Moralization over time: The proportion of alternative meat tweets containing moral words increased in both languages; levels were generally higher and rose earlier in English, with a 1–2 year lag in Japanese.
• Conditional moral language: Tweets mentioning “animal” or “environment” used moral words at roughly twice the rate of all alternative meat tweets and more than tweets referencing “sensory appeal” or “health”.
• Dominant moral foundation: In tweets containing “animal” or “environment,” the Care/Harm foundation accounted for nearly half of moral references, more pronounced in English, indicating focus on protection, safety, harm, and suffering.

The study demonstrates a shift from predominantly sensory-focused conversations about meat alternatives toward broader concerns—especially environmental protection and animal welfare—accompanied by increasing moral language. This moralization suggests consumers increasingly frame alternative meat choices in terms of protecting the environment and avoiding harm to animals. English-language discussions show stronger and earlier diversification and moralization than Japanese-language discussions, possibly reflecting differences in market maturity and food culture. While taste remains salient—especially in Japanese—attention to health, environment, animals, and price has grown, with environmental concern showing the steepest increase. The elevated presence of Care/Harm language in environment- and animal-related tweets underscores a moral framing aligned with preventing harm and promoting safety. These findings refine earlier work that emphasized taste and health as primary drivers by highlighting a growing moral dimension that may be leveraged in communication and marketing. However, whether moral concerns translate into actual purchasing behavior likely depends on trade-offs with price and functionality (taste, texture), warranting further empirical study.

This paper provides longitudinal, cross-linguistic evidence that public discourse about meat alternatives has diversified and become increasingly moralized, particularly around environmental protection and animal welfare. Moral language—dominated by the Care/Harm foundation—has grown over time and is more pronounced in English than in Japanese. Practical implications include complementing taste- and health-focused messaging with ethical framing to better resonate with contemporary consumer concerns. Future research should clarify the causal impact of moral appeals on actual purchasing and consumption, examine cross-country heterogeneity, and differentiate between types of meat alternatives (plant-based, cultured, insect-based) to tailor interventions and policy.

• Behavioral proxy: Twitter posts are proxies for interest and do not directly reflect purchasing or consumption. The causal link between moral appeals and increased consumption remains to be established.
• Demographic and platform bias: Twitter user demographics may skew younger or more engaged, limiting generalizability to broader populations.
• Context detection limits: Dictionary-based moral analysis (MFD/J-MFD) does not adequately capture sarcasm, negation, or nuanced context; the study tracks presence of moral foundations rather than valence/quality of judgments.
• Geographic granularity: Tweets were not classified by country, constraining interpretations of regional differences.
• Category aggregation: The study did not distinguish among subtypes of meat alternatives (plant-based, cultured, insect-based), which may have distinct consumer perceptions.
• Market and price dynamics: Price remains a barrier; taste/texture still under development, potentially limiting translation of moral concern into habitual consumption.