No, the author is ultimately unclear why qualia in itself is useful, but by reasoning about the case studies I listed, his argument that qualia is in fact related to recursive internal feedback loops is ultimately a bit stronger than just “these things all feel like the same things so they must be related”.
Humphrey first argues through his case studies that activity in the neocortex seems to generate conscious experience, while activity in the midbrain does not. Further, midbrain activity is sophisticated and can do a lot of visual and other perceptual processing independent of the neocortex, so all of that seems to be dissociable from consciousness. What remains is the generation of sensations in the neocortex. From that we can understand sensations ([in certain parts of?] the neocortex) are separable from perceptions (in the midbrain). So while that doesn’t tell us what consciousness is needed for, it does tell us what it is not used for, including at least some perception and midbrain processes like RL. The vulnerability in this argument IMO is mainly that self report about conscious experience might not be entirely reliable.
Then he advances his “ipsundrum” hypothesis about how recursive (or recurrent, perhaps) sensory feedback loops could have evolved and gives strong arguments about why this was evolutionarily useful. The argument includes the idea that recurrent sensory feedback allows complex integrative processes such as self reflection and theory of mind, and those have strong evolutionary advantages. The development of warm blood might have both facilitated those feedback loops by speeding up neural processing, and necessitated them by requiring a more sophisticated homeostatic apparatus. So an evolutionary feedback loop created the sensory feedback loop.
At this point, I guess we still can’t be confident these processes are inseparable from consciousness, but they seem more closely related to consciousness than other clearly separable processes. That seems valuable for at least lowering our expectations of sentience in species that don’t have the cognitive processes whose relationship to consciousness we haven’t ruled out.
No, the author is ultimately unclear why qualia in itself is useful, but by reasoning about the case studies I listed, his argument that qualia is in fact related to recursive internal feedback loops is ultimately a bit stronger than just “these things all feel like the same things so they must be related”.
Humphrey first argues through his case studies that activity in the neocortex seems to generate conscious experience, while activity in the midbrain does not. Further, midbrain activity is sophisticated and can do a lot of visual and other perceptual processing independent of the neocortex, so all of that seems to be dissociable from consciousness. What remains is the generation of sensations in the neocortex. From that we can understand sensations ([in certain parts of?] the neocortex) are separable from perceptions (in the midbrain). So while that doesn’t tell us what consciousness is needed for, it does tell us what it is not used for, including at least some perception and midbrain processes like RL. The vulnerability in this argument IMO is mainly that self report about conscious experience might not be entirely reliable.
Then he advances his “ipsundrum” hypothesis about how recursive (or recurrent, perhaps) sensory feedback loops could have evolved and gives strong arguments about why this was evolutionarily useful. The argument includes the idea that recurrent sensory feedback allows complex integrative processes such as self reflection and theory of mind, and those have strong evolutionary advantages. The development of warm blood might have both facilitated those feedback loops by speeding up neural processing, and necessitated them by requiring a more sophisticated homeostatic apparatus. So an evolutionary feedback loop created the sensory feedback loop.
At this point, I guess we still can’t be confident these processes are inseparable from consciousness, but they seem more closely related to consciousness than other clearly separable processes. That seems valuable for at least lowering our expectations of sentience in species that don’t have the cognitive processes whose relationship to consciousness we haven’t ruled out.