In this perspective, we ask the question whether the apparently lower solubility of specific proteins in amyloid disease is a cause or consequence of the protein deposition seen in such diseases. We focus on Alzheimer's disease and start by reviewing the experimental evidence of disease-associated reduction in the measured concentration of amyloid β peptide, Aβ42, in cerebrospinal fluid. We propose a series of possible physicochemical explanations for these observations. These include a reduced solubility, a reduced apparent solubility, as well as a long-lived metastable state manifested in healthy individuals as a free concentration of Aβ42 in the solution phase above the solubility limit. For each scenario, we discuss whether it is most likely a cause or a consequence of the observed protein deposition in the disease.