06.09.2021 Views

Mind, Body, World- Foundations of Cognitive Science, 2013a

Mind, Body, World- Foundations of Cognitive Science, 2013a

Mind, Body, World- Foundations of Cognitive Science, 2013a

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

that fail to materialize (Dreyfus, 1992, p. 85): “Despite predictions, press releases,<br />

films, and warnings, artificial intelligence is a promise and not an accomplished<br />

fact.” Connectionist cognitive science argues that this pattern <strong>of</strong> failure is due to the<br />

fundamental assumptions <strong>of</strong> the classical approach that fail to capture the basic<br />

principles <strong>of</strong> human cognition.<br />

Connectionists propose a very different theory <strong>of</strong> information processing—<br />

a potential paradigm shift (Schneider, 1987)—to remedy this situation. Even<br />

staunch critics <strong>of</strong> artificial intelligence research have indicated a certain sympathy<br />

with the connectionist view <strong>of</strong> information processing (Dreyfus & Dreyfus, 1988;<br />

Searle, 1992). “The fan club includes the most unlikely collection <strong>of</strong> people. . . .<br />

Almost everyone who is discontent with contemporary cognitive psychology and<br />

current ‘information processing’ models <strong>of</strong> the mind has rushed to embrace the<br />

‘connectionist alternative’” (Fodor & Pylyshyn, 1988, p. 4).<br />

What are the key problems that connectionists see in classical models? Classical<br />

models invoke serial processes, which make them far too slow to run on sluggish<br />

componentry (Feldman & Ballard, 1982). They involve explicit, local, and digital<br />

representations <strong>of</strong> both rules and symbols, making these models too brittle. “If in a<br />

digital system <strong>of</strong> notations a single pulse is missing, absolute perversion <strong>of</strong> meaning,<br />

i.e., nonsense, may result” (von Neumann, 1958, p. 78). Because <strong>of</strong> this brittleness,<br />

the behaviour <strong>of</strong> classical models does not degrade gracefully when presented with<br />

noisy inputs, and such models are not damage resistant. All <strong>of</strong> these issues arise<br />

from one underlying theme: classical algorithms reflect the kind <strong>of</strong> information<br />

processing carried out by electronic computers, not the kind that characterizes the<br />

brain. In short, classical theories are not biologically plausible.<br />

Connectionist cognitive science “<strong>of</strong>fers a radically different conception <strong>of</strong> the<br />

basic processing system <strong>of</strong> the mind-brain, one inspired by our knowledge <strong>of</strong> the<br />

nervous system” (Bechtel & Abrahamsen, 2002, p. 2). The basic medium <strong>of</strong> connectionism<br />

is a type <strong>of</strong> model called an artificial neural network, or a parallel distributed<br />

processing (PDP) network (McClelland & Rumelhart, 1986; Rumelhart<br />

& McClelland, 1986c). Artificial neural networks consist <strong>of</strong> a number <strong>of</strong> simple processors<br />

that perform basic calculations and communicate the results to other processors<br />

by sending signals through weighted connections. The processors operate<br />

in parallel, permitting fast computing even when slow componentry is involved.<br />

They exploit implicit, distributed, and redundant representations, making these<br />

networks not brittle. Because networks are not brittle, their behaviour degrades<br />

gracefully when presented with noisy inputs, and such models are damage resistant.<br />

These advantages accrue because artificial neural networks are intentionally<br />

biologically plausible or neuronally inspired.<br />

Classical cognitive science develops models that are purely symbolic and<br />

which can be described as asserting propositions or performing logic. In contrast,<br />

128 Chapter 4

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!