Background Recent US proposals suggest defining Alzheimer's disease (AD) based on β‐amyloidosis alone. This sparked debates that echoed historical ones about the significance of brain lesions and clinical phenotype. Methods This review covers debates on AD nosology through three key periods: AD's discovery in German‐speaking countries in the early 20th century, its redefinition in Anglo‐Saxon countries in the 1960s–1980s, and current debates on the biological or clinicobiological definitions of AD. Key players' opinions are focused on. Results At the beginning of the 20th century, AD was defined as a clinicopathological entity. Debates arose around the pathological anchor, which included extended neurofibrillary tangles versus neuritic plaques (Alzheimer vs. Fischer) and its association with senile dementia (Kraepelin). In the 1960s–1980s, the debate shifted towards whether AD could be diagnosed using qualitative or quantitative neuropathological features and whether it was a unique process (Terry and Katzman) or had subtypes (Roth). The current definition proposed by the US Alzheimer's Association is based purely on biological β‐amyloid abnormalities and represents a double break: from the historical clinicopathological definition of AD and from the historical emphasis on tau or combined tau and β‐amyloid high levels of pathology. Conversely, the clinicobiological proposal of the International Working Group remains aligned with historical concepts of AD. Conclusions This historical perspective illustrates the unresolved questions surrounding AD pathogenesis, role of lesions, and the clinical phenotype, especially for sporadic cases. The intense nosological debates throughout the history of AD also illustrate the diversity of theoretical frameworks for defining disease in medicine.