Let me state the strong AI problem in other words, so that it might be easy to understand why there are valid arguments about the assertion a prolog program "does not understand" the same way that a human does. Please note that I am not saying anything about the "quality of understanding" if there is such a thing, i.e. nothing is being said about whether the human model of understanding, as suggested by Steven Harnad is the only way something can be understood.
All human thoughts appear to be expressed in the form of one or more of the senses. In other words, no thought, and consequently understanding, can exist without it's manifestation in some form of real or imaginary sensory perception. Even when you are adding two numbers, there has to be some sensory representation in your mind of the process. Some people might visualize the numbers two and three, others might hear them, and blind people might visualize them in the form of touch. But the point is that all the thoughts are grounded. Understanding, according to the symbol grounding proponents can only occur if the concepts are expressed in terms of actual sensory perceptions. For a robot, the sensory perception could be a video camera and a microphone. A program written solely in terms of logical assertions is not grounded, therefore it cannot claim to have understood anything.
You can search for the 'chinese room argument' for more insight on this issue.
On the other hand, they can 'tout' it even more by citing themselves as an example of why they had to kill american jobs because of the driving down the economics of the software business by open source.
So when open source has completely taken over the market, we expect the average programmer income to rise inspite of the fact that everyone is free to swap their code with everyone else, thus reducing the income of the organization which pays these programmers to zilch?
I fail to see how the 'open source model' for the economy would support the high level of incomes in teh United States.
Considering that americans themselves have a history of arming their erstwhile allies in Afghanistan and Iraq, which then turn into your terrorists, you should not have too much to worry about from american jobs being outsourced to non-americans. They can't possibly fare any worse.
If John Kerry or anyone else wants to 'plug' outsourcing to prevent the income of hard earning Americans from going overseas, they should go and explain to the WTO why American goods should be allowed to be sold in Asia and let the Asians' hard earned income go overseas. They want to live like the americans now, and mind you, they are as good, and much more in numbers.
On an unrelated note, all the so called 'Aid' is used to buy favors from weaker nations for paltry sums of money, while trying to perpetuate economic imperialism.
If you want a "natural" language for computers then it would have to be necessarily of Chomsky-0 type. Thus Turing-complete. And therefore not decidable which implies that a computer cannot parse it.
If a grammar is turing complete, it doesn't follow that a computer cannot parse it. Undecidability means that it's impossible to write a program which takes as inputs an arbitrary turing complete grammar, and an infinitely large arbitrary set of strings (in this case, programs), and decides if this set is generated by the given grammar.
Given a particular program, turing complete or not, it's always possible to parse it and decide whether it's generated by the given grammar.
Furthermore his main point is frankly rubbish: it's well known that the grammar for all human languages follows the same basic rules (Chomsky's hypothesis)
I would imagine this point cannot be rigorously proven.
Slashdot Mirror
Let me state the strong AI problem in other words, so that it might be easy to understand why there are valid arguments about the assertion a prolog program "does not understand" the same way that a human does. Please note that I am not saying anything about the "quality of understanding" if there is such a thing, i.e. nothing is being said about whether the human model of understanding, as suggested by Steven Harnad is the only way something can be understood.
All human thoughts appear to be expressed in the form of one or more of the senses. In other words, no thought, and consequently understanding, can exist without it's manifestation in some form of real or imaginary sensory perception. Even when you are adding two numbers, there has to be some sensory representation in your mind of the process. Some people might visualize the numbers two and three, others might hear them, and blind people might visualize them in the form of touch. But the point is that all the thoughts are grounded. Understanding, according to the symbol grounding proponents can only occur if the concepts are expressed in terms of actual sensory perceptions. For a robot, the sensory perception could be a video camera and a microphone. A program written solely in terms of logical assertions is not grounded, therefore it cannot claim to have understood anything.
You can search for the 'chinese room argument' for more insight on this issue.
On the other hand, they can 'tout' it even more by citing themselves as an example of why they had to kill american jobs because of the driving down the economics of the software business by open source.
So when open source has completely taken over the market, we expect the average programmer income to rise inspite of the fact that everyone is free to swap their code with everyone else, thus reducing the income of the organization which pays these programmers to zilch? I fail to see how the 'open source model' for the economy would support the high level of incomes in teh United States.
And what would explain the higher average GRE scores by the "1/3" of world which is Asian?
Considering that americans themselves have a history of arming their erstwhile allies in Afghanistan and Iraq, which then turn into your terrorists, you should not have too much to worry about from american jobs being outsourced to non-americans. They can't possibly fare any worse.
If John Kerry or anyone else wants to 'plug' outsourcing to prevent the income of hard earning Americans from going overseas, they should go and explain to the WTO why American goods should be allowed to be sold in Asia and let the Asians' hard earned income go overseas. They want to live like the americans now, and mind you, they are as good, and much more in numbers. On an unrelated note, all the so called 'Aid' is used to buy favors from weaker nations for paltry sums of money, while trying to perpetuate economic imperialism.
Only in your continent. MSN and Yahoo messengers were the first and remain the only IM's worth talking about in the rest of the world.
If you want a "natural" language for computers then it would have to be necessarily of Chomsky-0 type. Thus Turing-complete. And therefore not decidable which implies that a computer cannot parse it.
If a grammar is turing complete, it doesn't follow that a computer cannot parse it. Undecidability means that it's impossible to write a program which takes as inputs an arbitrary turing complete grammar, and an infinitely large arbitrary set of strings (in this case, programs), and decides if this set is generated by the given grammar.
Given a particular program, turing complete or not, it's always possible to parse it and decide whether it's generated by the given grammar.
Furthermore his main point is frankly rubbish: it's well known that the grammar for all human languages follows the same basic rules (Chomsky's hypothesis)
I would imagine this point cannot be rigorously proven.