Thoughts on Personhood

I've found myself in the midst of a debate on abortion over in freerange_snark's LJ. It started when corwyn_ap said:

Now, establishing when personhood occurs is admittedly going to be contentious, but at least the debate will be to the point. I know of no one ... who thinks sperm (or an egg) is a person, nor anyone who thinks a baby free from its mother isn't.

and I responded with

[raises hand] Yes, you do. I think that personhood does not exist without "ability to communicate". Typically, not before 1-2 years after birth. It's admittedly still a fuzzy line, but that's where I draw it.

I don't believe that there's any political possibility of getting that sort of definition widely accepted, but it is what *I* think.

I'd like to back away, for a moment, from the debate which followed, and explain how I arrived at that position. My original thought process had nothing at all to do with the abortion issue. It had to do with AI Rights.

I grew up reading copious amounts of SF, and thinking that I would probably be an Artificial Intelligence programmer when I got through college. Naturally, Isaac Asimov's "Robot" stories were a huge influence on me. They were cool, but it bothered me deeply that these robots, who seemed like "people" to me, had no civil rights. I had no desire to help create a slave race. And it's not just artificial beings I'm concerned about. What about aliens? Are they "people"? What about "uplifted" creatures that weren't formerly intelligent?

Eventually, the issue seemed to bother Asimov as well; his story "The Bicentennial Man" is entirely about a robot's fight for recognition as "human", with all the rights pertaining thereunto. Said robot is eventually successful, but only at great cost. The judge eventually reached a determination of (roughly, from memory) "Any being that is able to *ask* to be considered human, *should* be considered human." That seemed like at least part of the right answer to me.

OK, let's start from there, and try and poke holes in it. "Any being that asks to be considered sentient, *is* sentient." First objection: Parrots. Second objection: Tape recorders. Clearly, just saying the words is insufficient. The crux appears to be the definition of "asks" in this context. It seems like it needs to include intentionality. But isn't "intentionality" what we are trying to determine in the first place? How do we determine the difference between a meaningless pattern of words and a meaningful pattern, when the words themselves are the same?

Very early in my education about AI, I ran into the concept of The Turing Test. Briefly (and not 100% accurately to what Turing originally came up with), this is a test for machine sentience: if the machine can fool a human observer into thinking that it (the machine) is a sentient being, then it *is* a sentient being.

One of the interesting features of this test, which might not have been immediately apparent, is how inherently subjective it is. There is no "objective Turing Test" -- indeed, there *cannot* be such a thing. The Turing Test is adminstered by beings (who presume themselves to be sentient), in an attempt to determine whether or not other beings are sentient --relative to the ones administering the test. I say "feature", because I do think this is a positive element, not a "bug". "Personhood", "sentience", whatever you want to call it, can (in my opinion) only ever be determined relative to a given population. Without such a population, the question doesn't even make sense.

The Turing Test is basically my standard for sentience now -- whether for AIs, aliens,... or other (biological) humans. If it can convince me that it's a person, I will treat it as such. One of the inevitable consequences of that view is that I do not consider very young children (or those with severe mental handicaps) to be "people", as such. [I don't advocate killing or mistreatment of such beings -- but my arguments would be based on "cruelty to animals" principles, not on "human rights".]

Thoughts on Personhood

I've found myself in the midst of a debate on abortion over in freerange_snark's LJ. It started when corwyn_ap said:

Now, establishing when personhood occurs is admittedly going to be contentious, but at least the debate will be to the point. I know of no one ... who thinks sperm (or an egg) is a person, nor anyone who thinks a baby free from its mother isn't.

and I responded with

[raises hand] Yes, you do. I think that personhood does not exist without "ability to communicate". Typically, not before 1-2 years after birth. It's admittedly still a fuzzy line, but that's where I draw it.

I don't believe that there's any political possibility of getting that sort of definition widely accepted, but it is what *I* think.

I'd like to back away, for a moment, from the debate which followed, and explain how I arrived at that position. My original thought process had nothing at all to do with the abortion issue. It had to do with AI Rights.

I grew up reading copious amounts of SF, and thinking that I would probably be an Artificial Intelligence programmer when I got through college. Naturally, Isaac Asimov's "Robot" stories were a huge influence on me. They were cool, but it bothered me deeply that these robots, who seemed like "people" to me, had no civil rights. I had no desire to help create a slave race. And it's not just artificial beings I'm concerned about. What about aliens? Are they "people"? What about "uplifted" creatures that weren't formerly intelligent?

Eventually, the issue seemed to bother Asimov as well; his story "The Bicentennial Man" is entirely about a robot's fight for recognition as "human", with all the rights pertaining thereunto. Said robot is eventually successful, but only at great cost. The judge eventually reached a determination of (roughly, from memory) "Any being that is able to *ask* to be considered human, *should* be considered human." That seemed like at least part of the right answer to me.

OK, let's start from there, and try and poke holes in it. "Any being that asks to be considered sentient, *is* sentient." First objection: Parrots. Second objection: Tape recorders. Clearly, just saying the words is insufficient. The crux appears to be the definition of "asks" in this context. It seems like it needs to include intentionality. But isn't "intentionality" what we are trying to determine in the first place? How do we determine the difference between a meaningless pattern of words and a meaningful pattern, when the words themselves are the same?

Very early in my education about AI, I ran into the concept of The Turing Test. Briefly (and not 100% accurately to what Turing originally came up with), this is a test for machine sentience: if the machine can fool a human observer into thinking that it (the machine) is a sentient being, then it *is* a sentient being.

One of the interesting features of this test, which might not have been immediately apparent, is how inherently subjective it is. There is no "objective Turing Test" -- indeed, there *cannot* be such a thing. The Turing Test is adminstered by beings (who presume themselves to be sentient), in an attempt to determine whether or not other beings are sentient --relative to the ones administering the test. I say "feature", because I do think this is a positive element, not a "bug". "Personhood", "sentience", whatever you want to call it, can (in my opinion) only ever be determined relative to a given population. Without such a population, the question doesn't even make sense.

The Turing Test is basically my standard for sentience now -- whether for AIs, aliens,... or other (biological) humans. If it can convince me that it's a person, I will treat it as such. One of the inevitable consequences of that view is that I do not consider very young children (or those with severe mental handicaps) to be "people", as such. [I don't advocate killing or mistreatment of such beings -- but my arguments would be based on "cruelty to animals" principles, not on "human rights".]