Agency and Adaptation in Human–Technology Configurations: A Fifty-Year Analysis of Computers & Education

Educational-technology histories are often organized around successive tools, obscuring how action, decision-making, regulation, and adaptive work are distributed among human and technological actors. Drawing on 5,387 papers published in Computers & Education over fifty years, this study examines how Agency and Adaptation were configured and redistributed across learners, educators, communities, institutions, and technological systems. A purposive calibration corpus of 48 papers containing explicit Agency discourse was coded using eight Agency–Adaptation indicators comprising 63 category levels. Applying an 18-topic structural topic model as semantic infrastructure, category probabilities were projected to the full corpus through a theta-weighted linear mixture with Dirichlet-style smoothing. Clustering identified five recurring profiles: system agency–programmed adaptation, teacher/institutional agency–orchestrated adaptation, community agency–collaborative adaptation, learner agency–customization-based adaptation, and system agency–algorithmic adaptation. Although period-level distributions of prominent individual categories changed modestly, the relative prevalence of the five profiles shifted substantially, while all five continued to coexist. Historical change involved the recombination and redistribution of human and technological responsibilities, not a uniform movement from human to technological control. An exploratory analysis identified 71 GenAI-explicit papers published during 2023–2026. These papers appeared across all five profiles during this period but were concentrated in system-programmed configurations and were comparatively uncommon in learner-customization configurations relative to contemporaneous non-GenAI papers. The findings show that technological innovation alone does not specify where agency resides or how adaptation is enacted. Understanding educational-technology development requires examining complete arrangements of actors, authority, regulation, and adaptive work, including whether increasingly adaptive systems preserve meaningful opportunities for human direction, evaluation, and recalibration.