The well-known John Searle’s “Chinese room” thought experiment put a man who didn’t know Chinese language inside a room with a hypothetical “rulebook” to be used in deciphering texts submitted though a slot in the door. Using the algorithms in the “rulebook” the man located the answer to the query and produced it correctly without any understanding of Chinese language. This seemed to be a proof of that the man acted like a machine – no difference in how a device would deal with the same task, and thus devices can’t have the real intelligence, or the strong intelligence in the true sense of the word “comprehension”, since they do not have to think to work an output answer to the input query. Not all go along with Searle in his thinking and one of the most well-known replies, called the “Robot Reply”, will be discussed here.
Individuals giving the “Robot Reply” to John Searle’s “Chinese Room” reasoning admit that Searle’s argument is potent in displaying that a variety of devices — including units that could pass the well-known Turing Test — will not always be intelligent and will not actually “comprehend” the words that they communicate. The reason is that a machine placed on a table without any sensory technology and any techniques to facilitate a causal interaction with things in the world is going to be unable to comprehend a language. This kind of a device could work with symbols, including the phase of generating end result that may deceive humans and consequently pass successfully the Turing Test (CCSI, n.d.). On the other hand, the words generated by this type of a device would be short of one essential component: the words would not convey any thoughtful content material and consequently could not eventually be “about” anything.
The issue with a computer resting on a table is that it features no sensory technology as well as no capacity to causally connect with things in the environment. A solitary desktop computer can be designed to produce all kinds of remarkable phrases about pigs. It could perhaps present a report about pigs and even generate expressive poems. On the other hand, one might carry a pig inside the room and place it over the pc’s keyboard and it would discover not a thing about it. Comprehending what the concept ‘pig’ signifies, demands that a speaker’s usage of the term be causally connected to actual pigs. Devoid of that an essential component required for substantive language is absent (CCSI, n.d.).
What can be inferred, in that case, from this train of thought? To begin with, a person providing the “Robot Reply” will follow J.Searle that the Turing Test may not be a dependable test for comprehending a language. The type of conduct displayed in the Turing Test will not likely be enough to indicate linguistic understanding. The place where the “Robot Reply” disagrees with John Searle is in denial of his vision that the “Chinese room” argument is successful in exhibiting that ANY digital computer is similarly disposed to Searle’s point. People who provide the “Robot Reply” consider that the proper type of robot-controlling digital computer could in fact be intelligent and comprehend a language (CCSI, n.d.).
Hence, what does John Searle deals with this point? Does he acknowledge that his argument happens not to be potent against the appropriate type of robot? Not a chance. He proposes that putting the computer within a robot is likely to produce no change in any way (Searle, 1984).Searle isn’t swayed by the “robot reply”. To understand that the inclusion of a robotic body does not produce a change, Searle states that one merely has to prolong the thought experiment by putting the Chinese Room within a robot.
At this point, all the processing that proceeds within the robot is going to be completed in the presently transformed “Chinese Room”. Along with symbols entering the room by means of questions, this time there is likewise going to be symbols entering the room coming from the video cameras which obtain visual data regarding pigs in the farm.
Searle thinks that this thought experiment triumphs over those who consider that a causal link involving an actual pig as well as “pig” utterances is enough for comprehending a language. However, why does Searle consider so? Because of the binary language (CCSI, n.d.). A pc on a table receives strings of symbols in a certain language and delivers them as end result of the equivalent language. If we wish a pc to pass successfully the Turing Test in English, it should have the ability to receive as input questions in this language and provide end answers in English as well. On the other hand, digital computers fail to “identify” (that is , do not execute calculations immediately on) the symbols that constitute the given English words as well as sentences (for example, “P..i..g”). They have to initially transform those symbols into representations of the sole language that pc’s precisely “comprehend” (that is, the sole language whereby they could execute all kinds of functions): the “inhuman” binary code of 0 and 1.
What is more, the computer does not understand a language, be it a computer language, for the computer being a processing machine, working on the basis of a set of programmed commands has no, so to say, “attitude” or individual association, to the symbols it deciphers in the forms of 0 and 1. Operating with numbers and values, and producing a command programmed by a given set of numbers and values, does not mean that the computer could do the same command “independently” with no urge of “cause and effect”. The computer does not understand the language because it cannot use it its own will, since there is no “own will”, there is only input and output.
We could consider a scenario when Kismet robot is employed for the response. This kind of robot could exhibit a broad range of emotions. To accomplish this Kismet was provided with moving facial characteristics, able to communicate common emotional conditions that mimic the ones of a human baby (Dreifus, 2000). Kismet could therefore allow its “parents” to understand if it needed a certain measure of stimulation- a process rendering Kismet to be an intelligent unit with basic “comprehension” of the environment. This strategy of generating AI was grounded on developing behavioral patterns acquired by interactive learning. We could ask Mr. Searle, “Do you believe Kismet understand a language?”. He may be prone to produce a question to our question, “Does it do what it does because it wants to? Does it bow its head and put its eyes and ears down, when it is being shouted at, because it truly understands the meaning of the words or because its sensors are programmed to react in certain patters according to the values of identified external triggers?” Our response would be “Yes, Kismet uses the built-in audio-visual technology and programmed responses to do what it does”. “Thus”, Searle would possibly conclude, “a program is unable to give a pc a “mind”, “comprehension” or “awareness”, irrespective of how intelligently it could make it react. We are not able to refer to what the device performs as “thinking process”. (Searle, 1984). Therefore, I suppose, Searle would argue that due to the fact that it doesn’t think, it fails to possess a “mind” in any typical perception of the word.
I agree with Searle that computers do not understand a language in a way humans can because they do not have an “individual self”, cannot possess an attitude and cannot convey their attitude to the outer world. Computers cannot go in their processing of input data beyond the given limited set of algorithms set up by humans. Computers cannot create their “own” response, and since we all can agree that a human utterance or written response can be presented in multitudes of forms and each particular kind is a matter of our mind’s creativity, computers cannot understand a language, since they don’t have a “mind”, cannot create responses , in other words, can’t “speak their mind”.