I don't know what I know: Evidence of preserved semantic knowledge but impaired metalinguistic knowledge in adults with probable Alzheimer's disease

Trevor A. Harley*, Lesley J. Jessiman, Siobhan MacAndrew, Arlene Astell

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)

Abstract

Background: At what point does "not telling" become "not knowing"? Previous research has shown that Alzheimer's disease (AD) affects people's ability to define words - primarily, it has been thought, because AD destroys the semantic representations of words that the patients can no longer define. We investigate an alternative hypothesis, that AD also affects metalinguistic ability, which in turn affects people's ability to produce good definitions.Aims: Does AD affect metalinguistic abilities? Are definitions poor because people lose semantic information, or because they lose the knowledge of what constitutes a good definition? Methods & Procedures: We established what constitutes good definitions of a set of words denoting animate and inanimate concepts. We then asked elderly people with AD to define these words. As expected, their definitions were very poor. However, we then asked them forced and open-choice questions about the information that they omitted from their definitions. Outcomes & Results: People with AD can access semantic information that they appear to have lost. The AD group performed significantly worse than control participants on a word definition task, but importantly, some of the information they did not provide spontaneously was provided after questioning. The format of their definitions was also different from the controls; they made particular use of autobiographical information. Conclusions: We conclude that although our participants with AD have lost some semantic knowledge, at least sometimes they do not provide information that they do still know, because of a metalinguistic impairment. In particular, the participants with AD no longer understand what constitutes a good definition. We argue this metalinguistic impairment results in part from frontal atrophy. Our results have the important consequence that just because a person with AD does not offer information, even when asked, does not mean that they do not know it.

Original languageEnglish
Pages (from-to)321-335
Number of pages15
JournalAphasiology
Volume22
Issue number3
DOIs
Publication statusPublished - Mar 2008
Externally publishedYes

Cite this