It sounds outlandish, but two researchers from the University of Toronto beg to differ, as Maclean's Anne Kingston reported this week:
The sleuth-work in this case was revealed last month at the Rotman Research Institute Conference held at Toronto’s Baycrest, a leading research centre on aging. There, Ian Lancashire, an English professor at the University of Toronto, and his colleague Graeme Hirst, from the computer sciences faculty, presented an abstract of their ongoing investigation of linguistic patterns in Christie’s novels. They conclude the later novels reveal signs of encroaching Alzheimer’s. More significantly, their study confirmed British neurologist Peter Garrard’s 2004 analysis of three novels by Iris Murdoch, who was afflicted with Alzheimer’s, which argued dementia’s onset could be detected in written text “before anyone has the remotest suspicion of any untoward intellectual decline.”
Lancashire initiated the project with Hirst last fall with literary, not medical, motives: he needed data on linguistic patterns of older writers for a book about how authorship is revealed in texts. He believed Christie’s oeuvre, comprised of 85 meticulously plotted novels and plays written over 53 years, offered great scope for investigation. Though the author was never diagnosed with dementia, it’s speculated it could have been responsible for her decline in her late 70s and the muddled, meandering plots of her later novels. Also, the fact editors were reluctant to interfere, with the exception of her last book, meant the texts contained her imprint only.
Sixteen novels Christie wrote between the ages of 28 and 82 were analyzed by being put through a “computational-linguistic” screen to analyze vocabulary and phrase repetition. What they found, according to Kingston, is that "use of indefinite nouns and the indefinite article “thing” increased significantly over time, they found, as did phrase repetition, while vocabulary declined by 15 to 30 per cent. The most precipitous change occurs in Christie’s penultimate novel, Elephants Can Remember, written when she was 81; it contained a 30 per cent drop in vocabulary compared to her writing at age 63, 18 per cent more repeated phrases, and a nearly threefold increase in indefinite nouns."
Some, understandably, remain skeptical that textual analysis will be helpful on a large scale, but Hirst believes the Internet might one day make that possible: “In time, everyone will have 30 years of emails and blogs on their hard disk; we’ll be able to do this kind of textual analysis for everyone.” Freeman remains doubtful. “The linguistic quality of emails is often lacking,” he says. In any case, it's worth checking out the original poster (PDF alert) for further information.
It may not be the way to go for a diagnosis yet, nevertheless these are interesting facts. A vocabulary dropping by 30 % is quite significant in my opinion.
Posted by: Dorte H | April 02, 2009 at 10:05 AM
Alzheimers isn't the only possible explanation for a decline like that, though -- depression or the side effects of medication can also make the mind wander, and interfere with verbal ability. Early medications for high blood pressure, in particular, caused depression in many users, and confusion and dizziness can also be side effects of the blood thinner coumadin.
Posted by: Clair Lamb | April 02, 2009 at 10:16 AM
I wonder what their computer would say about RB Parker.
Posted by: David J. Montgomery | April 02, 2009 at 11:28 AM
Oh, boy I'm disinclined to put anything out there then. My children can track my dementia and confer from month to month.
Posted by: Patti Abbott | April 02, 2009 at 01:55 PM
The first thing that came to my mind was Christie's "mysterious disappearance" in 1926. She was only 36 at the time, and biographer and former doctor Andrew Norman believes that Christie was "in the grip of a rare but increasingly acknowledged mental condition known as a 'fugue state', or a period of out-of-body amnesia induced by stress. In effect, the writer was in a kind of trance for several days."
Either way, it's clear that something was going on with her mental state of well being.
Posted by: Pam Ripling | April 02, 2009 at 02:02 PM
I have noticed the loss of language in my loved one with AD, and on the other hand, he is using words I never ever knew were in his vocabulary! A very strange disease indeed.
Posted by: Virginia Hermes | April 02, 2009 at 05:09 PM
And yet, still better than a lot of the crap they have today.
Posted by: Phillip | April 03, 2009 at 02:40 AM
The next step would be to conduct similar tests on a large sample of writers who did have some kind of progressive amnesia with certainty, writers who may have had it, and writers who definitely did not. That would throw some light on any such correlations. Interesting study.
Posted by: Cine Cynic | April 03, 2009 at 04:44 AM
By age 85 one half of everyone suffer from dementia so it's not long odds here.
Posted by: Eric C | April 03, 2009 at 10:14 AM
Interesting theory, but I'm dubious, largely because I think the claim that "dementia’s onset could be detected in written text 'before anyone has the remotest suspicion of any untoward intellectual decline'" is blown out of the water by Terry Pratchett, whose recent novels (I'm thinking of "Making Money," "Going Postal" and the Tiffany Aching books with the possible exception of "Wintersmith") represent if anything a sort of renaissance for his work. Of course, Pratchett as his weakest is far more brilliant than the majority of writers, and perhaps the filter would reveal surprsing results, but I think you'd be hard put to detect the onset of Alzheimer's based on his language use.
Then again, early onset Alzheimer's may be a different beast from senile dementia, which (as mentioned Clair mentioned above) can be caused by depression, or a multitude of other causes related to advancing age.
Posted by: Rebecca | April 03, 2009 at 03:33 PM