In a recent paper Walton (3) has described the diagnostic and predictive accuracy of the Modified Word Learning Test (Walton and Black (4)) in relation to psychiatric disorders of the senium. Walton assessed the predictive value of the test in terms of changes in diagnosis over a two-year period. However, as Payne (2) has pointed out, a diagnosis, as such, can only be regarded as a kind of label, of interest not in its own right, but because of its implications. These, Payne has argued, are commonly fourfold, being related to (a) aetiology, (b) description, (c) the effects of treatment and (d) prognosis.