Aesthetics. Sentience. Extinction.
And that brings us to the final question, lurking way down in the anoxic zone: the question of what consciousness costs. Compared to nonconscious processing, self-awareness is slow and expensive [112]. (The premise of a separate, faster entity lurking at the base of our brains to take over in emergencies is based on studies by, among others, Joe LeDoux of New York University [117, 118]). By way of comparison, consider the complex, lightning-fast calculations of savantes; those abilities are noncognitive [119], and there is evidence that they owe their superfunctionality not to any overarching integration of mental processes but due to relative neurological fragmentation[4]. Even if sentient and nonsentient processes were equally efficient, the conscious awareness of visceral stimuli—by its very nature— distracts the individual from other threats and opportunities in its environment. (I was quite proud of myself for that insight. You'll understand how peeved I was to discover that Wegner had already made a similar point back in 1994 [120].) The cost of high intelligence has even been demonstrated by experiments in which smart fruit flies lose out to dumb ones when competing for food [121], possibly because the metabolic demands of learning and memory leave less energy for foraging. No, I haven't forgotten that I've just spent a whole book arguing that intelligence and sentience are different things. But this is still a relevant experiment, because one thing both attributes do have in common is that they are metabolically expensive. (The difference is, in at least some cases intelligence is worth the price. What's the survival value of obsessing on a sunset?)
While a number of people have pointed out the various costs and drawbacks of sentience, few if any have taken the next step and wondered out loud if the whole damn thing isn't more trouble than it's worth. Of course it is, people assume; otherwise natural selection would have weeded it out long ago. And they're probably right. I hope they are. Blindsight is a thought experiment, a game of Just suppose and What if. Nothing more.
On the other hand, the dodos and the Steller sea cows could have used exactly the same argument to prove their own superiority, a thousand years ago: if we're so unfit, why haven't we gone extinct? Why? Because natural selection takes time, and luck plays a role. The biggest boys on the block at any given time aren't necessarily the fittest, or the most efficient, and the game isn't over. The game is never over; there's no finish line this side of heat death. And so, neither can there be any winners. There are only those who haven't yet lost.
Cunningham's stats about self-recognition in primates: those too are real. Chimpanzees have a higher brain-to-body ratio than orangutans [122], yet orangs consistently recognise themselves in mirrors while chimps do so only half the time [123]. Similarly, those nonhuman species with the most sophisticated language skills are a variety of birds and monkeys—not the presumably "more sentient" great apes who are our closest relatives [81, 124]. If you squint, facts like these suggest that sentience might almost be a phase, something that orangutans haven't yet grown out of but which their more-advanced chimpanzee cousins are beginning to. (Gorillas don't self-recognise in mirrors. Perhaps they've already grown out of sentience, or perhaps they never grew into it.)
Of course, Humans don't fit this pattern. If it even is a pattern. We're outliers: that's one of the points I'm making.
I bet vampires would fit it, though. That's the other one.
Finally, some very timely experimental support for this unpleasant premise came out just as Blindsight was being copy edited: it turns out that the unconscious mind is better at making complex decisions than is the conscious mind [125]. The conscious mind just can't handle as many variables, apparently. Quoth one of the researchers: “At some point in our evolution, we started to make decisions consciously, and we're not very good at it.” [126]
The child Siri Keeton was not unique: we've been treating certain severe epilepsies by radical hemispherectomy for over fifty years now [127]. Surprisingly, the removal of half a brain doesn't seem to impact IQ or motor skills all that much (although most of hemispherectomy patients, unlike Keeton, have low IQs to begin with) [128]. I'm still not entirely sure why they remove the hemisphere; why not just split the corpus callosum, if all you're trying to do is prevent a feedback loop between halves? Do they scoop out one half to prevent alien hand syndrome—and if so, doesn't that imply that they're knowingly destroying a sentient personality?
The maternal-response opioids that Helen Keeton used to kickstart mother-love in her damaged son was inspired by recent work on attachment-deficit disorders in mice [129]. The iron-scavenging clouds that appear in the wake of the Firefall are based on those reported by Plane et al. [130]. I trawled The Gang of Four's linguistic jargon from a variety of sources [81, 131, 132, 133]. The multilingual speech patterns of Theseus crew (described but never quoted, thank God) were inspired by the musings of Graddol [134], who suggests that science must remain conversant in multiple grammars because language leads thought, and a single «universal» scientific language would constrain the ways in which we view the world.
The antecedent of Szpindel's and Cunningham's extended phenotypes exists today, in the form of one Matthew Nagel [135]. The spliced prosthetics that allow them to synesthetically perceive output from their lab equipment hails from the remarkable plasticity of the brain's sensory cortices: you can turn an auditory cortex into a visual one by simply splicing the optic nerve into the auditory pathways (if you do it early enough) [136, 137]. Bates' carboplatinum augments have their roots in the recent development of metal musculature [138, 139]. Sascha's ironic denigration of TwenCen psychiatry hails not only from (limited) personal experience, but from a pair of papers [140, 141] that strip away the mystique from cases of so-called multiple personality disorder. (Not that there's anything wrong with the concept; merely with its diagnosis.) The fibrodysplasia variant that kills Chelsea was based on symptoms described by Kaplan et al. [142].
And believe it or not, those screaming faces Sarasti used near the end of the book represent a very real form of statistical analysis: Chernoff Faces [143], which are more effective than the usual graphs and statistical tables at conveying the essential characteristics of a data set [144].