Why a Lonely Wallet Makes Perfect Sense in Japanese but Not in English

Say “my wallet is lonely” to an English speaker and they will likely pause, puzzled, before working out that you mean your wallet is empty. Say it in Japanese and nobody blinks. A new study published in Frontiers in Language Sciences reveals that this everyday quirk of Japanese speech is not just a translation curiosity—it reflects a deep difference in how the two languages process anthropomorphic language, one that existing theories of metaphor comprehension cannot explain.

Personification is a type of metaphor in which a non-human entity is treated as a person. English speakers do it all the time: “my car is temperamental” works because we can view a car’s engine as akin to a human mind—a complex, hidden controller whose moods produce unpredictable behavior. But Japanese offers personification metaphors like “my wallet is lonely,” where a wallet has no internal system resembling a mind at all. Franklin Chang of Kobe City University of Foreign Studies and Tomoko Tatsumi of the Max Planck Institute for Psycholinguistics set out to determine whether these culturally distinctive forms are actually easy for Japanese readers to understand, or whether they trip up comprehension just as they do for English speakers.

To characterize Japanese personification, the researchers spent years opportunistically collecting examples from television, internet videos, websites, and everyday life—much like early speech error corpora. They gathered 46 spoken and 13 written examples. A kindergarten teacher encourages a boy to eat by speaking of “Poor Carrot-san” who “feels lonely.” A mother consoles her adult daughter about a sold air conditioner by writing that it “was happy being taken care of by you” and hoping it will “do well with its next master.” A YouTuber amused viewers by asking “Tomato-chan and Lettuce-kun, why are you guys such close friends?” about sandwich ingredients stuck together. A veterinarian discussing a sick cat’s constipation remarked, “if Feces-chan gives us the benefit of coming out, then things will be more comfortable.” These examples routinely apply human titles—-san, -chan, -kun, -sama—to simple objects, and many appear in formal business contexts where English anthropomorphic idioms would be unusual.

Crucially, most of these expressions are not conventionalized idioms. Searching the National Diet Library Ngram corpus, which spans 2.3 million books and magazines with roughly 1.7 billion words, the team found that only 17 of their examples appeared with the inanimate argument in the corpus; the remaining 37 forms occurred only with human arguments. The researchers also mapped their dataset onto animacy ratings from VanArsdall and Blunt, who had English speakers judge 1,200 common nouns on features including the ability to think and goal-directedness. The objects in the Japanese anthropomorphic expressions—carrots, bread, belts, lettuce—clustered at the bottom left of the animacy space, rated as having almost no capacity for thought or goals, even compared with an ant or a computer. This is puzzling for standard personification theory, which assumes a mind-like element in the target: a car’s engine plays the role of a mind, but a slice of cheese has nothing comparable.

The authors propose an alternative: a group-based schema. The phrase “Tomato-chan and Lettuce-kun are close friends” activates the concept FRIENDS, a feature of HUMAN GROUP, which can be aligned with sandwich INGREDIENTS, where the corresponding concept is STUCK TOGETHER. This schema requires no mind-like internal element. The carrot is lonely because it is separated from the other foods in its group; the air conditioner is happy because it joins a new household. Japanese relational titles themselves reflect hierarchical group membership—-sama and -san for senior members, -chan and -kun for equals or juniors—so the same cultural logic explains both the emotions and the honorifics.

The team then designed story comprehension experiments based on methods from Nieuwland and Van Berkum’s discourse studies. Participants read five-sentence stories in which the fourth sentence was either literal or anthropomorphic. In one story, a depressed gambler tells a friend he lost all his money in one bet; the expected fourth sentence is “my wallet is empty,” while the unexpected version is “my wallet is crying.” Control stories created an expectation about animacy instead: a man realizes he forgot something in his car, and the expected sentence describes finding his bag while the unexpected one describes finding his dog. Critical sentences were matched in length within two characters, and true/false comprehension questions kept readers attentive. Japanese participants were recruited through Crowdworks and screened with culture and language questions drawn from the Japanese Language Proficiency Exam; English monolinguals came through Prolific.

The results were striking. In Experiment 1 with 64 Japanese and 61 English participants, English readers were significantly slower at the fourth sentence when it was anthropomorphic than when it was literal, but Japanese readers showed no difference at all. Japanese readers did slow down for the control stories, where an animate entity violated an animacy expectation—showing they were generating expectations and were sensitive to surprise. Corpus frequency analyses ruled out a simple explanation: in Japanese, the literal forms were on average more than 365 times as frequent as the anthropomorphic forms, yet Japanese speakers read both equally fast. In English, the anthropomorphic forms had essentially zero frequency in the NGRAMs database, and English readers struggled with them.

Experiment 2 replicated the finding with new participants—67 Japanese and 69 English speakers—and 50 percent more stories, including items from formal settings such as a veterinarian and a kimono dresser. The pattern held: a significant four-way interaction showed that the lack of an expectation difference for anthropomorphic stories was unique to Japanese readers. A combined analysis pooling 261 participants and analyzing only first exposures confirmed the result, and the authors note that the direction of the effect in the Japanese anthropomorphic condition was actually reversed, making a power explanation implausible.

These findings contradict both major accounts. Dependency theories of comprehension, which predict difficulty whenever arguments and predicates mismatch—and which are supported by ERP evidence like the P600 for inanimate subjects of active transitive verbs—predict trouble in both languages. Anthropomorphism theories, grounded in universal biases like elicited agent knowledge and effectance, predict ease in both languages. Instead, the results support a Culture account: Japanese readers appear to possess culturally specific schemas, such as a group-based personification schema, that let them apply human predicates to simple non-moving objects with no processing cost. The authors point to possible cultural background factors, including Shinto beliefs in kami residing within inanimate objects called shintai, the tradition of “married” rocks and trees, and Buddhist kuyō funeral rites for dolls, needles, glasses, and even robot dogs—ceremonies with roots reaching back to a needle funeral hall established in Kyoto by 881. Yet they caution that modern survey data show few Japanese report religious affiliation, and evidence about collectivism is mixed, so the precise cultural mechanisms remain open questions.

What emerges is a vivid demonstration that culture-specific conceptual knowledge can change the very mechanics of sentence comprehension. English speakers can eventually work out what a lonely wallet means, but the processing cost is real and measurable in reading time. Japanese speakers treat the same expression as unremarkable, as ordinary as “my wallet is empty.” The mind, it turns out, does not process metaphor in a cultural vacuum—and somewhere in Japan, a wallet may be lonely without costing anyone a millisecond of surprise.

Subject of Research: Lonely wallets: differences in comprehending anthropomorphic language in Japanese and English

Article Title: Lonely wallets: differences in comprehending anthropomorphic language in Japanese and English

Article References: Chang, F., & Tatsumi, T. (2026). Lonely wallets: differences in comprehending anthropomorphic language in Japanese and English. Frontiers in Language Sciences, 5, Article 1845613. https://doi.org/10.3389/flang.2026.1845613

Image Credits: AI Generated

DOI: 10.3389/flang.2026.1845613

Keywords: Lonely, wallets, differences, comprehending, anthropomorphic, language, Japanese, English, scientific research