This is entire test of a real TOEFL Writing Section, extracted from a new real TOEFL test. While you're working on Task 1, please check back the reading passage and lecture transcript below.
Remember to SUBSCRIBE with us for more tests and support.
----------------
READING PASSAGE (3 minutes)
Do computers think? It isn’t a new question. In fact, Alan Turing, a British mathematician, proposed an experiment to answer the question in 1950, and the test, known as the Turing Test, is still used today. In the experiment, a group of people are asked to interact with something in another room through a computer terminal. They don’t know whether it is another person or a computer that they are interacting with. They can ask any questions that they want. They can type their questions onto a computer screen, or they can ask their questions by speaking into a microphone. In response, they see the answers on a computer screen or they hear them played back by a voice synthesizer. At the end of the test, the people have to decide whether they have been talking to a person or to a computer. If they judge the computer to be a person, or if they can’t determine the difference, then the machine has passed the Turing Test.
Since 1950, a number of contests have been organized in which machines are challenged to the Turing Test. In 1990, Hugh Loebner sponsored a prize to be awarded by the Cambridge Center for Behavioral Studies—a gold medal and a cash award of $100,000 to the designer of the computer that could pass the Turing Test; however, so far, no computer has passed the test.
----------------
LECTURE TRANSCRIPT
Professor:
Philosopher John Searle has challenged the validity of the Turing Test because it’s premised on behavior rather than on thought. To prove his argument, he’s suggested a paradox, which he refers to as the Chinese Room. If a monolingual English-speaking person receives questions on a computer terminal from a Chinese person in another room, naturally the English-speaking person won’t understand the questions. However, if there’s a large reference that can be accessed, and if the reference is detailed and comprehensible, then the English speaker could, conceivably, break the code. For example, if a sequence of Chinese characters are received, the reference could indicate which sequence of Chinese characters would be expected in response. In other words, the behavior would be correct, although the English speaker wouldn't be thinking at a level that included meaning. The person would be manipulating symbols without understanding them, or, as Searle suggests, the person would be acting intelligent without being intelligent, which is exactly what a computer could be programmed to do.
Therefore, at least theoretically, a computer could be designed with complex input that would allow it to provide adequate behavioral output without being aware of what it's doing. If so. then it could pass the Turing Test. But the test itself would be meaningless because it doesn’t really answer the most basic question about artificial intelligence, which is, can the computer think?
