When computers first started to be created in the early 1950s, Alan Turing published a ground-breaking article that outlined a test which he claimed would prove that machines could think and were conscious. His test is as follows: a judge would sit in a room at a terminal that is connected to two other rooms. In one room, a computer would be set up in one room with a program of some linguistic capabilities. In the other room, a human would be sitting at another terminal. The judge would proceed to have a conversation with both the human and the computer. At the end, the judge would have to determine which of the two other rooms held a person, and which of the rooms held the computer. For logistical reasons, the it can be assumed that all three participants know standard English, and the judge is allowed to ask any questions he/ she wants, as long as it is natural conversation, and not tricky or nonsensical. The test can also be modified slightly, where there would be only one room, and the judge would have to determine whether or not the occupant was a human or a computer. Either way, Turing concludes that if the judge was convinced that the computer was really a human being, the computer would be, in a sense, conscious and able to think.

Karelis summarizes that Turing’s view is that “a ‘machine,’ in the artificially narrow sense of a product of engineering techniques other than bio-engineering techniques, must be able to ‘think,’ in that ordinary sense of the word which connotes problem-solving, strategy-devising, and inference-drawing if (though not only if) it can pass the Turing test.”  The philosophical importance of this test is that Turing was able to avoid the mind-body problem, and ascertain that whatever acts sufficiently intelligent is intelligent. I argue that regardless of whether or not the test would or would not prove that the computer is conscious, the question is meaningless because even in theory, a robot could never pass the Turing Test.

Pinker outlines in his book a few of the attempts of creating a machine that could convincingly pass the Turing test. He mentions several competitions where amateur programmers have submitted their programs for judgment. He stated that, in fact, “after interacting with several programs and human foils for about seven minutes apiece, the judges ranked all the humans as more human-like than any of the computers.”  Pinker gives some examples of conversations that the judges had with the computer:

Judge: It has been a long day! Now I think I’m ready for some fun. The weather isn’t very pleasant today. Do you have any jokes that could lift the clouds?
Computer: Everybody talks about the weather but nobody seems to do anything about it. Why do you need to tell me now you think you’re ready for some fun? Groannnnn......
Judge: You’re trying to make this difficult for me, aren’t you? I just explained I was ready for some fun!
Computer: Would you prefer it if it was not true that I are trying to make this difficult for you aren’t you? I just explained that I was ready for fun! I am, therefore I compute ... What makes you think I am trying to make this difficult for you?

Regardless, these programs are clearly not in the spirit of the Turing test. The programmers approach the problem by creating a program that manipulates words using some structured grammatical algorithm that repeats phrases, and essentially the program only spits back at the judge his own words. If the judge had continued the conversation for more than seven minutes, and tried to talk about a variety of subjects, and even asked the computer its opinions, the programs would fall painfully short of natural conversation.

The point that we can deduce is that for a computer to have an extended convincing conversation with a human being, it would have to have some world knowledge and experience. It is also important to note that the program would have to have knowledge that is applicable to humans. As French argues, “the Turing Test could be passed only by things that have experienced the world as we have experienced it, ... and that the Test provides a guarantee not of intelligence but of culturally-oriented human intelligence. [sic]”  French gives an analogy to the Turing Test, but this one is meant not to test intelligence, but rather the ability to fly. French says to imagine being on a large Nordic island where the only flying animals known are birds. Some philosophers see these birds, and being to talk about the “essence” of flying. They determine that the essence doesn’t fundamentally involve many definitions you might give to it like ‘moving through the air’ really, because a thrown pebble or a released helium balloon do the same. And it doesn’t fundamentally involve having feathers or wings, because penguins have both but don’t fly. So, the philosophers concede that flying has something to do with being in the air, and something to do with having wings and feathers and hollow bones, etc. So, they propose that if a being has all those requirements, it could fly. Now, clearly, airplanes, jets, bats or beetles wouldn’t pass the Seagull Test. But, this doesn’t mean they can’t fly, it only means that they’re not seagulls.

The point is clear. It would not be useful for the computer to have an understanding of a dog’s-behavior, let’s say, or a dog’s world-view, because a human would not relate to or understand what a dog would have to say. Clearly, the computer would have to live a human life, in a sense, and learn from a human existence, to have any ability to converse about the world.

What exactly would this entail? First of all, the computer would have to be able to learn about its environment. It would be impossible to program a computer about every fact in the world, every cultural bias and trend, etc.. The computer would, in a sense, have to have an ability to deduce information and apply it to relevant situations. In other words, to pass the spirit of the Turing test, the computer would have to experience life like a human would.

Realistically, there are some significant practical limitations to this reality. Cog is the closest thing we have at this point to a robot that can perform these functions, and he can’t do all that much. Rodney Brooks from the MIT artificial intelligence lab says that Cog’s “neural networks don’t actually duplicate the way we think because we don’t yet know how we think. The task is enormous.”  As of now, the research team is aiming to program the computer to act like a infant (6 mos. - 1 yr.) AND go through the development stages, because development -- learning about the world -- is central to understanding.

Only recently have projects been undertaken to teach computers behavior, rather than instill it in them by use of a program. In the video “Minds, Machines and Mystery,” there is some footage of a project underway at Brandeis, where robots are programmed to pick up little pucks and put them in the corner. One of the researchers, Maja Mataric notes that “part of the research is understanding how hard it is to learn. What does it take for us as programmers to build in to make the computers learn, just like what did it take for nature to build in how to learn? We’ve had computers who learn to forage, who learn to co-ordinate their behaviors. Learn to behave socially, share information, let each other pass. We build in very little. We give them some behaviors, we call them basis behaviors. But then the computers have to build on them.”

Another feature of this computer would be that it would have to, in a sense, look human or at least have sensory inputs in roughly the same places that we do. A being that had ears on its toes would clearly experience the world differently than we would. Even though this wouldn’t show necessarily that the computer was not intelligent, it would show that the computer was different and for that reason the judge might realize that it was a computer. Regardless, we are a long way off from having a computer that could live and learn just like humans do.

Practical limitations aside, even if we could create a silicone being that looked like a human, could learn like a human, and could interact with the world, the computer still would not be able to pass a rigorous Turing Test. As I specified before, the computer would have to have an understanding of human culture for it to successfully pass the test with flying colors. After all, there is one feature of humanity that a computer could never have that would severely impair its ability to understand human culture -- a silicone computer could never procreate or reproduce.

Fundamentally, our drive to procreate (and survive for that matter) is at the center of our being and our consciousness. This is Darwin’s evolutionary theory. After all, what is consciousness but a tool to help us pursue those goals over anything else.
 Mc Gill at Rutgers University, however, argues the contrary. He states that “nobody knows why consciousness evolved. It’s the old question. The same question could be asked, why didn’t evolution create creatures that were robots, which had no consciousness, wouldn’t they be just as efficient to survive? What’s the exact role of consciousness? I don’t think that anyone knows the answer to that. It’s a difficult question. It looks as though consciousness doesn’t play any role in the evolution of the mind.”

I concede that consciousness does , however, serve as a useful tool to us. A fundamental aspect of an animal’s consciousness is that, in effect, it helps it to survive and negotiate its way around in order to ultimately procreate. Granted, with us, it’s taken on a seemingly new level, where we process information that is always immediately relevant to our survival. But, possibly, you could build a case that it’s all indirectly important to our survival and our quest to procreate. Academics and learning pursuits are essentially about deconstructing and categorizing the world, which we deem essential for our survival. Look at how education has helped humans learn about ourselves and the world around us. This contributes quite obviously to our well-being and our survival. If there are reasons beyond this, you’d be hard pressed to show how that has nothing to do with our survival.

Therefore, without this fundamental feature, a computer could never have the ability to understand human culture or realistically reproduce human behavior. And, because the Turing Test is a test to determine whether or not a robot has culturally oriented human intelligence, the robot could never pass the Turing Test. Indeed, the scope of the test is limited. Although I deem it a sufficient test to determine a object’s intelligence, passing the test is not a necessary condition for intelligence. It only proves a certain kind of intelligence - namely, human intelligence.