More Thoughts on Love Without a Body
Inspired by the movie “Her,” I recently wrote an article entitled, “Can a Computer Fall in Love if It Doesn’t Have a Body?” The movie’s premise is that an artificial intelligence can feel love; but research on embodied cognition suggests this might not be such a straightforward affair. Love is not abstract; it’s very much shaped by body and biology, and computers are largely disembodied.
Arriving after the article’s publication were these comments from Eiling Yee, a cognitive scientist at the Basque Center on Cognition, Brain and Language (and author of this really neat paper on how manual experience shapes the way we think about things):
Q: Could a disembodied human brain could understand physical love?
EY: With respect to the first question, I would guess that a body isn’t necessary to achieve something like the kinds of experiences that embodied brains experience. The research line that seems related to this is the stuff on mirror neurons (e.g., Rizzolatti & colleagues — you’re probably familiar with this) — the main related finding is that when one monkey sees a second monkey reach towards, e.g., some food, neurons fire in the first monkey’s motor cortex that would control the kind of reaching motion that he is watching. The critical question, which this research doesn’t address, is whether a monkey that had never reached for anything (e.g., a paralyzed monkey) would also activate these same motor cortex neurons. (Sian Bielock’s work on professional athletes and fans may be semi-related too.) And then even if those same neurons *were* activated in the paralyzed monkey, what would that monkey be experiencing? That is, we know that the brain is plastic enough that congenitally blind people activate visual cortex when reading braille — and we wouldn’t interpret this as them experiencing what a sighted person experiences when activating the same parts of visual cortex.
Q: Could an artificial brain love?
EY: Something that seems related is the behaviorist perspective on psychology — if a computer acts exactly in a way that appears to be love, how is that different than love? We have the intuition that it is, but what really makes it different? This is also indirectly related to the Turing test — if a computer could completely convince its correspondent into thinking that it’s human (and in this example, that it is in love) would it be? Would the programming necessary to create that “illusion” constitute love? How can we know that our partner loves us other than through their behavior?
I thought my friend Paul Allopenna [a cognitive scientist at Brown University] would find this interesting, so I passed along your question and he had this interesting comment:
"So, what I would want to know is, could an embodied brain really be "embodied" if it didn’t need to survive? That is, how much of a functional role to our biological (particularly metabolic) predicates play in motivating/organizing/ and executing our behaviors (including our emotive behaviors)?"
So as I understand it, Paul’s question essentially comes down to: if, despite having a body, a brain didn’t need that body to provide it with anything (e.g,. fuel) to survive, what kind of behaviors would it instruct that body to perform? That is, would it demonstrate love (or anything else we normally demonstrate)?