While the field is in its relative infancy, what is known about language and AD is rapidly developing. Here, we focus on studies which look at natural language–conversations, story telling, etc. over formal language elicitation tasks (e.g.How many animals you can name in 60 seconds?). We’re interested in how language is functionally used in real-life interactional contexts.
Several recent studies have explored changes in individuals speaking and writing with healthy aging adults who progress to clinical or possible AD diagnoses. These studies tend to focus on individuals for whom there is a lot of publicly available language data (e.g. the writings of author Iris Murdoch, political speeches by Ronald Reagan). Broadly, they show that language patterns begin to be distinctive far before clinical dementia diagnoses are provided. However, it is not yet clear how unique these changes are to the specific individuals studied (versus as measures which could potentially be applied as early indicators in the general population), and how successful such measures may be in contexts with less long-term data available.
The most well-known predictive findings related to language and AD come from the Nun Study, which compared early-life writing samples of Nuns in two American convents to their late-life dementia onset and autopsy-based AD diagnosis confirmation. Broadly, these suggest that there are early life indicators of AD evident in writing. However, their findings are not conclusive enough to use in the clinical sphere to predict AD, for a number of reasons. Our research has replicated these findings, using late life spoken narratives.
In terms of language background, being bi-/multilingual has a negative correlation to dementia onset. That is, multilingualism does not ‘block’ one from developing dementia, but the onset of dementia-related symptoms are delayed, typically by four years. This phenomenon is often referred to in research as ‘cognitive reserve’: in one interpretation the claim is that the differential wiring of the brain, as bi/multilingualism develops, acts as a protective force against the early neuro-degeneration processes associated with AD.
There is also very interesting research on formulaic language and dementia (e.g. idioms, which are always framed in the same way, like barking up the wrong tree, but also including word combinations which are frequently used and potentially stored as a ‘chunk,’ e.g. I don’t know). The basic argument from this research is that there is an increased reliance/use of formulaic language chunks with dementia progression.This may contribute to the global interpretation of AD language as grammatically reasonable (not ungrammatical), but progressively more ’empty.’