This sci-fi adventure gets points for creativity. A small crew from earth investigates an alien ship somewhere out in the Oort cloud. The stimulating questions include, how would you communicate with aliens, and, what are the aliens like?

Usually in sci-fi, the alien is an unimaginative variant on a human, like E.T., or an insect or mollusk, or some combination of those. Not very alien at all.

Watts’ aliens are alien. How about non-biological entities that feed off electromagnetic energy? I have to admit I had a little trouble picturing them and their behavior. If you’re going to have really alien aliens, you need to put a ton of effort into describing them.

The more interesting question is how would you communicate with aliens. Where do you even start? Carl Sagan was pretty sure any alien would recognize the spectrographic signature of hydrogen and his team engraved a representation of that on the Voyager spacecraft, presuming that any alien would have visual, pictorial perception, would be committed to elementalism in analysis, and other assumptions.

Watts assumes that any aliens will be susceptible to tribalism, so you can make one alien suffer while the other observes, then take note of the communication between them, on the assumption that they’re talking about the ordeal. Maybe, maybe not.  I’m skeptical that tribal feeling is universal. But points for originality.

A central concept of the novel is the Chinese Room, a thought experiment proposed by philosopher John Searle in the 1980’s to illustrate why strong AI is not possible. People outside a room feed questions written in Chinese into a slot. Inside the room, a person who cannot read any Chinese takes the slips of paper and pattern-matches the figures in a big book, and copies out answers found there. He slips the answers out the slot.

Observers marvel at the intelligence of the Room for giving such wise answers. Searle’s point is that there has been zero comprehension of any questions or answers, as the man in the room cannot read Chinese. He has simply performed a mechanical matching task. The same would be true for any computer that purported to pass the Turing test. Despite the performance, there would be no actual intelligence in play.

In this novel, the on-board linguist suspects that the aliens operate like the Chinese room. Despite elaborate and meaningful conversation, she believes the alien entity understands nothing and may not even be sentient. Fascinating idea.

Taking the idea even further, the protagonist and first-person narrator is a man who has had half his brain removed so he completely lacks empathy (you have to accept that empathy is “located” in one half of the brain). Consequently, he understands next to nothing of social intentionality but has an uncanny knack for detailed observation and can deduce people’s meanings from their behavior. He is essentially a Chinese Room.

He is also the main reason why the novel doesn’t work as a story. A narrator who understands nothing is a very tough challenge, and Watts doesn’t overcome it. In fact all of the characters are trans-human, modified in various ways to optimize special abilities, but for the reader, it’s like trying to empathize with six garden tools. For example, the captain of the ship is a genetically modified vampire (you have to accept that vampires are real). Again, points for creativity, but none for character engagement or storytelling.

Plenty of stimulating ideas in biology, physics, neuroscience, cognition, and perceptual psychology fly about, but as coruscation, contributing little to development of story or character.

Ultimately, even the framing theme of the Chinese Room doesn’t hold water if you think about it. There is a homunculus inside the room right from the start, the genuinely intelligent, Chinese-literate  human being who wrote the big lookup-book that enabled the pattern matching. Just because that person’s intelligent cognition was recorded for deferred execution does not make it disappear. Many people, even philosophers, are confused by that simple misdirection.

The real challenge for building a genuine AI is specifying in the first place what intelligence is.  A similar argument can be mounted concerning another central theme, that of blindsight (a blind person who believes he or she can see). We need to ask, what is seeing?

As a story, Blindsight meanders, nor is it successful as a character study, nor does it present new understanding of AI. But in terms of sheer technical creativity, I’d say it’s worth a read.

Watts, Peter (2006). Blindsight. New York: Tom Doherty Associates (384 pp).