Phantoms

So I've just finished reading Phantoms in the Brain, V.S. Ramachandran's first book. It's a bit old, from 1998, but still a classic. The BBC did a related miniseries, if you're more into videos.

It's a good pop neuroscience book. The first two hundred pages or so are very similar in style and content to Oliver Sacks' classics, The Man Who Mistook His Wife for a Hat and An Anthropologist on Mars, which Ramachandran references several times. The topic flits through case studies of seemingly unrelated neurological conditions, including phantom limbs, blindsight, Charles Bonnet syndrome, neglect and anosognosia, Capgras' delusion, and pseudocyesis. All of it was fun and fascinating reading, but for me the most interesting part of the book was Chapter 12, on qualia and consciousness.

Most people interested in strong AI (or the Hard Problem of consciousness) will already be familiar with qualia. It's a philosophical topic, so the arguments often get quite complex, but the basic idea is that qualia are the subjective aspects of personal experience that can't be easily communicated to others. The canonical example of a "quale" is "What is it like to see red?"

Now, the reason that qualia are related to strong AI is that they're often brought up as an example of an essentially human experience that machines can never emulate. A machine might be able to listen to and identify a piece of classical music, but will it experience the music? Qualia are generally seen as one of the major philosophical objections to reductionism, at least within the field of cognitive science. At least to some people, qualia represent one of the last bastions of mind-body dualism.

So I rather like Ramachandran's opinion on qualia:

"I'd like to argue that there is in fact no such barrier, no great vertical divide in nature between mind and matter, substance and spirit. Indeed, I believe that this barrier is only apparent and that it arises as a result of language. This sort of obstacle emerges when there is any translation from one language to another.
How does this idea apply to the brain and the study of consciousness? I submit that we are dealing here with two mutually unintelligible languages. One is the language of nerve impulses - the spatial and temporal patterns of neuronal activity that allow us to see red, for example. The second language ... is a natural spoken tongue like English or German or Japanese ..."
p. 231

In other words, he says, the problem with qualia is a language barrier. If we were to bypass spoken/written language and directly connect the neural pathways from my color-processing areas to yours, there would be no intermediate language translation. At least in theory, you would see what my eyes saw, and know what it's like for me to see red.

"The key idea here is that the qualia problem is not unique to the mind-body problem. It is no different in kind from problems that arise from any translation, and thus there is no need to invoke a great division in nature between the world of qualia and the material world. There is only one world with lots of translation barriers. If you can overcome them, the problems vanish."
p. 232

His conclusions are comforting to strong AI proponents, although I doubt he had them in mind. Ramachandran lists 3 critical features of qualia (earlier developed in Three Laws of Qualia), which provide testable qualifications for a perceptual representation system comparable to our own. His theory doesn't even need to be accurate, simply possible, to be relevant to my interests. But I'll be interested to see if his opinion on the nature of qualia has changed over the last decade.

Anyway, for an 11-year-old work on popular neuroscience, Phantoms in the Brain is still fresh and engaging reading.