Constructing Old Age: A Historical Study of Framing, Tone, and Genre in Aging Discourse

Background and Purpose: Aging is not merely a biological process but a socially constructed phenomenon shaped by dominant cultural narratives, institutional norms, and policy discourse. In the realm of social welfare, representations of aging have long been embedded in questions of who is deemed deserving of care and public support—questions often mediated by policy frameworks such as medicalization and dependency. These frameworks not only guide how services are allocated but also shape how older adults are portrayed and understood in public discourse. Although scholars increasingly recognize the sociopolitical construction of aging, little is known about how aging narratives have evolved over time, particularly in institutional and public media such as newspapers, magazines, nonfiction literature, and academic texts. This study addresses that gap by asking: (1) how has the tone of aging narratives in public texts changed over the past two centuries? (2) to what extent can these changes be explained by dominant policy-related frames? and (3) do narrative trajectories differ between genres?

Methods: This study used the Corpus of Historical American English (COHA), a diachronic corpus of over 100,000 texts published between 1820 and 2010 across fiction, nonfiction, newspapers, and magazines. We extracted 314 aging-related collocates—words appearing within four words of six common synonyms for “older adult” (e.g., elderly, aged). Each collocate was scored on three dimensions: narrative sentiment tone (1=most negative to 5=most positive), indicating the affective portrayal of older adults; medicalization (0=none, 0.5=moderate, 1=highly medical), reflecting biomedical framing; and dependency (1=autonomous to 5=highly dependent), capturing portrayals of frailty or passivity. Ratings were generated using three large language models—ChatGPT 4.5, Gemini 2.5, and Copilot—based on the rationale that LLMs, trained on large historical and contemporary corpora, reflect contextualized language use. A shared prompt was used to ensure consistency, and inter-rater reliability (Cronbach’s α) ranged from 0.79 to 0.94. We applied generalized additive models to assess how tone evolved and was influenced by medicalization, dependency, and genre, weighting models by collocate frequency.

Results: Findings revealed a significant nonlinear decline in narrative tone over the last 200 years (F=16.06, p<.001), with tone becoming more negative after a slight rise in the mid-19th century and a steep drop beginning in the 1950s. Fiction consistently demonstrated a more positive tone than nonfiction. Increased use of medicalized (β=−0.89, p<.001) and dependency-framed (β=−0.83, p<.001) language was strongly associated with more negative portrayals of aging. High-frequency negatively framed terms associated with medicalization included invalid, feeble, retarded, and handicapped, while those linked to dependency included needy, poverty, homeless, and alone. In contrast, positively framed collocates across dimensions included dignified, respectable, wise, and active, reflecting more empowering portrayals of older adults.

Conclusions and Implications: Rather than viewing age as a neutral biological stage, the findings affirm aging as a site of ideological construction—where discourse influences not only perception but also social inclusion, legitimacy, and moral worth. Understanding these patterns invites critical reflection on how we come to know aging and whose experiences and identities are amplified or diminished through narrative framing.