Wednesday 20 May 2015

Semantic Dementia and the Organization of Conceptual Knowledge

Joseph McCaffrey
In honour of Dementia Awareness Week 2015 (17th-23rd May), we have a post by Joseph McCaffrey, a graduate student in the University of Pittsburgh's Department of History and Philosophy of Science. Here Joseph summarises his recent article 'Reconceiving Conceptual Vehicles: Lessons from Semantic Dementia', published in Philosophical Psychology.

We take our concepts for granted. When you explore the world, you automatically categorize the objects around you, tapping into a bewildering array of information. You see (or hear) a sheep and instantaneously know it is is a mammal, an animal, a provider-of-wool, a white fluffy thing that bleats, and much more. As a philosopher of cognitive science, I am interested in how the mind stores, accesses, and manipulates this conceptual knowledge.

In semantic dementia, a rare variant of frontotemporal dementia, patients lose concept knowledge in a progressive and debilitating fashion. Early on, caused by damage to a brain region called the anterior temporal lobes, patients experience striking semantic deficits (i.e. problems with word and object meaning) while other cognitive abilities, including speech production and autobiographical memory, remain fairly intact. At first, a patient with semantic dementia may be unable to recognize a picture of a duck, saying 'it is some kind of bird'. Later, the same patient may only know that the picture depicts some sort of animal.

My paper explores what semantic dementia means for debates about the 'vehicles' of conceptual knowledge. An old philosophical debate concerns whether concepts are reactivated sensory experiences. The British empiricists of the 17th and 18th centuries, such as John Locke and David Hume, thought of concepts as simulations of past perceptual experiences. Thinking about a hammer might involve simulating what a hammer looks like, how to swing one, etc. On the other hand, some philosophers believe that concepts are distinct from percepts. Descartes famously argued that you can know what a chiliagon (a geometric figure with 1,000 sides) is even though it is impossible to picture one. That knowledge must be something different than a perceptual simulation.

Recent work in cognitive science has revived this old debate. According to 'neo-empiricists', including philosopher Jesse Prinz and psychologist Lawrence Barsalou, concepts are perceptual representations. By contrast, amodal theorists such as Edouard Machery hold that concepts are stored in a distinct way. The same information can be represented quite differently. For example, the word 'sheep' conveys much the same information as a picture of a sheep, albeit in a different format. Our perceptual systems represent the world in characteristic ways (visual representations are images, auditory representations are sounds). The question is: Are concepts the same kinds of representations used by our perceptual systems?

For neo-empiricists, brain data says 'yes'. One brain imaging study found that reading specific action verbs (e.g. 'kick') activates the motor regions used to perform those actions (e.g. kicking). Neo-empiricists also appeal to neuropsychology (the study of patients with brain damage) to support their case. Many studies report that damage in different perceptual systems result in category specific semantic deficits. For example, damage to visual cortex leads to semantic deficits for living things while damage to the motor cortex leads to semantic deficits for tools and other inanimate objects. This leads to a view where concepts are stored in distinct brain systems dedicated to specific perceptual modalities.

I argue that semantic dementia raises trouble for the neo-empiricist picture of where (and how) the brain stores concept knowledge (see also my 2012 paper with Edouard Machery). Semantic dementia patients experience a 'modality general, item-specific' impairment pattern. First, damage to one brain system (the anterior temporal lobes) causes category general deficits. A patient might fail to recognize or name some tools, some animals, etc. Second, for each item affected (e.g. sheep), patience experience semantic deficits in many perceptual modalities (e.g. touch, smell, vision). A patient who cannot identify a sheep visually probably will not recognize its bleat. This pattern confounds the neo-empiricist picture, which predicts that brain lesions will affect some categories, some features, etc. This pattern may suggest that the anterior temporal lobes house a unitary, amodal semantic store.

My paper examines what semantic dementia might reveal about the brain's conceptual system. But keep in mind that semantic dementia (like other dementias such as Alzheimer's) leads to a profound loss of memory, identity, and ultimately life. Frontotemporal dementia is the leading cause of dementia in middle age, and there is no known cure. However, researchers like Dr. Karalyn Patterson, a Cambridge neuroscientist, are making major strides in our understanding of and potential to treat frontotemporal dementia. For Dementia Awareness Week I would like to recognize their work. You can learn more about dementia here.

No comments:

Post a Comment

Comments are moderated.