> Before I add my thoughts to the theoretical discussion about real dogs vs. realistic robotic dogs, let me weigh in with the opinion that we'll never achieve what we're positing here, that a robotic dog can be created that simulates a real dog convincingly. I know, never say never. But I agree with what someone else said -- we are so hard-wired to recognize life (or at least life forms sufficiently like ourselves) that we're not even conscious of all the little tells we perceive. I think that we're in the middle of a huge leap in technology that will enable us to get pretty close to an indistinguishable simulation of an animal -- but close, in this case, is still a world away. I think that, at best, we'll reach a point where we have robotic dogs that are close enough to real dogs that we can't identify what the differences are, and yet people in general will *still* be able to tell the difference somehow. Researchers will have to conduct formal, protracted studies to figure out what people are picking up on to discern the difference, and maybe these studies will eventually yield some pointers on how to tweak our robots, but I'm doubtful that technology will be able to catch up to our own innate abilities to recognize life. Ok, maybe robots will get good enough to fool people for a minute or an hour or even a day, but not once we start talking about people nurturing relationships with these robots over the long term.
>

You've voiced this opinion before, and I'm still not sure I understand your reasoning. I've resisted arguing the point much so far because I can't come up with any counterpoints that are any better than your arguments. Your argument boils down to "I think we'd be able to detect fake dogs somehow" and mine to "I don't think we'd be as good at it as you think we would."

The only somewhat related argument I can come up with has to do with the issue of detecting trickery. Humans are absolutely *not* good at detecting trickery. Professional magicians make their living on that fact. They make you think you saw things you did not see, they make you see what *they* want you to see, not what really occured. You may know you got tricked, if only (for some people) because the magician *tells* you it's a trick and not "real magic", but you were totally taken in by the illusion. There is truth to the fact that if you see the same trick multiple times, you can sometimes detect how it's done. I know a card trick that I can comfortably do for someone only once--more than that, and I fear the ruse would be obvious to many. But for many tricks done by stage magicians, it usually doesn't matter how many times you see it, you still won't see "the trick" unless you know what to look for.

Magic tricks aren't the only example. I saw a program about the Loch Ness monster once where the people doing the show did a "test" of sorts (unscientific to say the least, but telling none the less.) They talked to a group of people by the Loch about Nessie, and then a plant in the crowd pointed out something in the water. It was literally just a log the producers had planted there, with a stick up in the air to simulate a "neck", placed far enough away from shore so it wasn't immediately obvious what it was. I believe it was also being manipulated from shore somehow to make it move around in the water some. After oohing and ahing over it for awhile, the people in the crowd were asked to draw pictures of what they had seen and describe it. Well, very few people drew pictures of a log with a stick on it. Nearly all of them drew (as best they could) a completely traditional looking lake monster. Some of them even swore they saw humps undulating in the water and other such complete fabrications.

The point is, the people didn't see what was there, they saw what they *thought* was there, or perhaps what they *wanted* to see. True, this was a short term thing, and part of your argument is that over the long term, people would be able to tell. And that may be partially true. I agree that simulating something as complex as a dog is a huge undertaking that would take many, many tries to get "right". I totally agree that it would probably take many generations of robodogs before people could totally buy into them as "real pets". But I don't buy that we can't be fooled. In fact, if anything, we're really EASY to fool. We're also really good at fooling *ourselves*, both unintentionally and intentionally. So I think there'd be a big subset of the population that'd totally go in for the robodogs way before they were "perfect" anyway, either because they were fooled enough by them to accept them as loving pets, or because they *wanted* to be fooled by them.

-- Dave