Searle's Chinese Room

Transcript

1 Indiana Undergraduate Journal of Cognitive Science 3 (2008) 51-57 Copyright © 2008 IUJCS. All rights reserved Searle's Chinese Room Argument and its Replies: A Constructive Re-Warming and the Future of Artificial Intelligence Mariano de Dompablo Cordio Philosophy University of Wisconsin, La Crosse __________________________________________ ________________________ Abstract R. Searle in his article "Minds, This paper treats the philosophy of John earle's Chinese room argument (CR), that Brains, and Programs". It shows, using S what Searle calls strong artificial intelligence (AI), the thesis that minds are to er hardware, is not only false, but also brains as computer software is to comput CR does this by arguing in e ffect that there can be no that it must be false. The translation between Chinese and John's English understanding, and likewise neither can computers understand any ques tion put to them because any question addressed to them is like the Chinese to John. On the other hand, ask John in as any native speaker English, which he understands as well , 'Do you understand Chinese?' and he will answer 'No.' Wh at is the difference between John and Chinese and does understand English, computers? While John does not understand ecause any programming language to computers understand nothing. And b computers is like Chinese to John, human-like behavior of a computer charged with running a robot, for example, proves nothing in the way of human understanding on the part of computers. Because the CR does such a good job of proving the falsity of strong AI, a fundamentally different approach to the creation of AI is necessary. But, this is aves strong AI definitively behind. problematic for strong AI, namely, it le 1. Introduction When we ask, 'Is artificial intelligence (A I) possible?' we really ask 'Can we create consciousness in com puters?' This, as I see it, has been established by the philosophy of AI discussion so far. So, can we? And, if we can, what would we be

2 M. Dompablo Cordio / IUJCS 3 (2008) 52 doing with the computer (if manipulating its inner structure or otherwise) in order for consciousness to be created with it? Thes e questions put us in a place to define the AI project in two goals, a primary a nd a secondary. The secondary goal is a step toward the primary goal, consciousne ss in a computer. The secondary goal can be either of two possible alternatives as defined by the AI philosophy discussion so far: 1. writing a program, which when im plemented in a hardware, results in a nipulating the inner observation OR 2. ma conscious computer according to test by structure of a computer in order to in effect reproduce within the computer the context for consciousness in a brain, providing the context necessary for consciousness but doing so in a computer. Both alternatives have an important y goal. 1. takes consciousness (and a background as they try for the primar a significant extent, and says that if materialistic, biological account) for granted to behavior of the disputed conscious entity (the computer) can be mistaken for a , that entity understands (Turing 212). human being given the proper circumstances at programs for human-like behavior), and 1. says, 'Write the right program (one th you have a conscious computer.' 2. assume s that consciousness is a property of the context provided by brains (and allows fo r a materialistic, biological account of consciousness). In order to prevent any fudging on th e definition of what strong AI (the thesis that minds are to br ains as computer software is to computer hardware) supporters have as the goal of their projects, what I defined above as the primary goal, it is important to understand that str ong AI projects and strong AI itself seem to be rooted in large part in science ficti on. This happens in two ways. First, in so far as strong AI seems to borrow ideas for its projects like that of consciousness in innocuous. Ideas from science fiction computers from science fiction, it is d. Second and more im portant is the very examples of AI are not all that is borrowe idea, the computer model of the mind, that hardware performing computation according to software-indicated directions could be intelligence, which may also be a holdover of too much science fiction, is damaging to a more biological account. But a deeper investigation into the origin of models of the mind over the course of history will have to be left for another time. To better pin down the goal of the projects of computer science from another direction, refer to th e technical language used to name and describe computer scientists' AI proj ects. Some examples are: genetic programming, memory, master/slave programs, teaching programs, and programs that learn. All of these exampl es exude something of a human sense, a personification of their projects. In thes e examples, I have as sumed that we see what computer scientists believe their proj ects can or will be able to do. But there is a problem with their usage of that language. What computer scientists are trying to do with their projects is evident,

3 M. Dompablo Cordio / IUJCS 3 (2008) 53 illustrated by their language. The problem realize the primary goal, and that is seems to be that computer scientists do not understand the significance of the ure of the mental an incomplete pict language that they use, which is telling of or qualitative, subjective language e.g. consciousness properties involved in their states of awareness and intentional states. So, to make a proper attempt at realizing the mind or its property consciousness in their computer would be almost happenstance because they do not understand the goals implicit to their language as teaching or learning. This taking for when, for example, they describe programs granted of one of these goals, consciousness, in the attempt to create AI reaches yet cognitive science, of which computer farther into the interdisciplinary field, of cognitive science as well often leaves science is a part. The philosopher branch out or inadequately treats in its philoso phy of mind the What is it like? or the qualitative experience had by consciousness (Nagel 321). Conscious understanding and subjective intentional states are not , however, left out by the philosophy of John R. Searle in "Minds, brains, and pr ograms" (235). His wonderful article is where this paper will begin. 2. Chinese Room Argument Before continuing to my adapted rende ring of the Chinese room argument appearing in Searle's article, the reader should understand that the Chinese room signed to be identical in principle to any that Searle describes in his argument is de e room can or cannot do parallels all computer. Thus, anything that the Chines relevant computer capac ities. A person, John, is in a room. John does not understand Chinese symbols nor is he capable of recognizing Chinese symbols such that he can distinguish Chinese sy mbols from Japanese symbols nor is there anything "To [keep John from believing th at] Chinese writing is just so many meaningless squiggles" (Searle 236). Also in the room are two windows and two boxes. In Box 1 are directions written in English and divided into sets. Each particular set of English directions corre spond to an attached Chinese symbol also inside Box 1. In Window 1 comes a Ch inese symbol which John receives. John proceeds to Box 1, matches the Chinese symbol received to the same type Chinese symbol to which are attached a particular set of English directions. John follows these directions, which he understands as well "as any other native speaker of English", by finding in Box 2 the direction-indicated symbol (Searle 236). John then proceeds to Window 2 and puts this symbol out the window so that John effectively correlates one set of formal symbols wi th another set of formal symbols, and all that 'formal' means here is that [John] can identify the symbols entirely by their shapes . . . [ans wering] by manipulating uninterpreted

4 M. Dompablo Cordio / IUJCS 3 (2008) 54 formal symbols. (Searle 237) And, all of this constitutes John simp ly behaving as a computer, performing "computational operations on formally specified elements" (Searle 237). So, John successfully performs computation without understanding Chinese. After all, his responses to this point have been indicat ed by the English directions. But what if we gave John questions written in English? Suppose that we ask in English 'John, do you understand Chinese?' and indicate that he should pass his answer out Window 2. John would answer 'No.' What is the difference between John's res ponse to the question in English and his essed in Chinese? Ask yourself what it response to the same question instead addr would be like to be John, compari ng both experiences; John experiences of the English question and his answer. He understands nothing of understanding e. Nor does he understand his Chinese the same question addressed in Chines answer as anything more than the pushing of a shape out of a window that reads "Window 2" over it. What does Searle provide us in this thought experiment? A distinction between human understanding of a familiar language versus a language one does not speak. It is also important to notice that computers and humans share an engagement with the shape of writte n words, but unlike John who understands the English and not the Chinese, computers cannot understand any language even though they work in them. These consider ations provide us with an appropriate setting for a few of the many replies to the Chinese room argument. 3. Objections to the Chinese room argument ese room argument to be treated will The first of the objections to the Chin The systems reply concedes that the man be the systems reply as named by Searle. inside the Chinese room does not understa nd the Chinese version of the question put to him 'Do you understand Chinese?' Howeve r, John is but a part of the system, and the whole system, the Chinese room and any digital computer by way of its having at its disposal everything accessi ble to a digital computer, understands the question. Searle replies, let John memorize every system element, the English directions and attached Chinese symbols in Box 1 and the Chinese symbols in Box 2 so that the composite's aspects comprise all aspects of the entire system. John still does not understand Chinese. There si mply can be no translation between John's English understanding and the Chin ese symbols, no matter John's Turing test mistaken understanding. And, "a fortio ri neither does the system [understand], because" anything in the system is part of John (Searle 240). So, strong AI is false, the systems reply fails, and the Turing test has counterexamples, so it is

5 M. Dompablo Cordio / IUJCS 3 (2008) 55 s reply, Searle makes some additional ineffective. With regard to the system remarks, which I think outline what ha s been the state of these affairs. someone who was not in the grip of It is not easy for me to imagine how an ideology would find the idea at all plausible. Still, I think many people of strong AI will in the end be who are committed to the ideology inclined to say something very much like this; so let us pursue it a bit further. (240) a formal symbol manipulation system' The systems reply replies: "'the man as really does understand Chinese ." (Searle 240) In this reply, the systems reply begs h of its claims without argumentation in the question, that is, it insists the trut the systems reply is false. There are addition to its original argument. So, gard to objections to the Chinese room additional comments made by Searle with re that do a great job of outlining the central errors of strong AI supporters. to the Chinese room argument. It The robot reply is the second objection asks us to think about a new program. This computer, with its program written not only for the taking in and putting out of symbols, would perform the function of operator of the robot in which it is pl aced. The computer would operate the robot in such a way that its behaviors are similar and can be confused for something with human-level understanding. The idea is that this computer instead of the original room argument would have understanding. digital computer treated by the Chinese What should be noticed about the robot reply? The robot reply indirectly claims that cognition is about being-in-the-world causal force in the world instead of just with certain realities of what it is to be a formal symbol manipulation (Searle 243). The obvious reply to the robot reply is robot does not get rid of the original that putting another computer inside of a problems outlined by the Chinese room argument; innovation in programming whatever cannot improve upon the problems wh ich have thus far been outlined and are essential to all algorithm based approaches to the creation of AI, strong AI. In order to illustrate this criticis robot is John instead of the m, imagine that inside that extra computer with its new program. Th e presence of John in the room for the purpose of carrying out the computation needed for the robot's operation is in principle equivalent to the needed com putation otherwise being carried out by the computer. After having replaced the comput er with John, follow through with the original Chinese room argument, that is , to the conclusion that there can be no translation between John's English unde rstanding and Chinese, and one will understand that the robot reply is false. The implicit notion underlying the robot reply (as in the systems reply) that 'If it behaves like it, it must be it

6 M. Dompablo Cordio / IUJCS 3 (2008) 56 upon the essential state of John in the (understanding)' does nothing to improve room and strong AI by the same force of the Chinese room argument. Given strong e creation of AI will be necessary since AI's falsity, an adjusted approach to th strong AI must be false. 4. Further inquiry into the possibility of AI A fundamentally different approach to th e task of creating AI may be that of a computer in order manipulating the orientation of the firings in the hardware of to reproduce the necessary physical contex t for consciousness in brains but doing may not reproduce the physical context so in a computer. Though this approach d only reproduce the electrical portion necessary for consciousness because it woul chemical reality in the brain. However, of what would presumably be an electro- e in our pursuit of the reproduction of the electrical approach seems plausibl conscious subjective states necessary in order to in some way produce the precise primary goal of AI, intelligence. understanding necessary for what is the strong AI, namely, it leaves strong AI But this approach is problematic for definitively behind. Strong AI takes it as its assumption that hardware can realize certain desired properties of the brain w ithout a neurophysiological account of the brain to reach the same goal. This adjusted approach leaves behind the use of just computer hardware in order to provide a physical context for consciousness, which Without consciousness, however, intentional is not just computer hardware proper. AI seems plausible with the appropriate states would be had by nothing. So, while context for consciousness, neurophysiology is required before it can be created. Strong AI cannot be true. 5. Conclusion The original assumption of strong AI is that certain desired mental properties can be achieved using hardware in a comput er instead of a brain, that brains are effectively hardware. But, this view, st rong AI, has been proven false on multiple occasions by John R. Sear le's Chinese room argu ment. The Chinese room argument has shown that there can be no translation between John's English understanding and Chinese, making unders tanding impossible at either John's level, that of the central processing unit (C PU), or at the level of the entire system, that of the CPU, the wires, and any a dditional hardware used in computation. Likewise, understanding is impossible when placing a computer inside of a robot so that it may move around in the world, for if John were in the room performing the same computational tasks, he c ould not understand Chinese, and moving around in the world in a human-like fashion does not improve upon the state of computers. The falsity of this last exam ple, the robot reply, and the notion that mistaken behavior is a sure indicator of understanding casts serious doubt on the

7 M. Dompablo Cordio / IUJCS 3 (2008) 57 cessary for the successful creation of AI. Turing test. So, an adjusted approach is ne a fundamentally different The approach described in this paper, however, is manipulating the orientation of the firings approach to that of creating strong AI, as er leaves the original assu mption of strong AI behind, in the hardware of a comput thus leaving strong AI definitively behind. Indeed, the re-orientation of firings in ogically defined orie hardware on the model of a neurophysiol ntation of neuron firings in order to provide the appropriate context for consciousness while using a materialistic, biological account of the brain is not hardware proper as required by the strong AI thesis. Rather, it is a synt hesis of hardware and neurophysiology. This approach may not even reprodu ce the necessary physical context for consciousness because it only reproduces th e electrical portion of what would presumably be an electro-ch emical reality in the brain. So, AI (and not strong AI) is possible in principle, but it is depende nt on a materialistic, biological account of sness as defined by a consciousness as a the physical context for property consciou ain-consciousness relation. AI on the property of the brain hypothesis of the br basis of the non-biological account of the brain in the strong AI thesis cannot be possible. References Nagel, Thomas. "What Is It Like to Be a Bat?" Philosophy of Mind: A Guide and Anthology . Ed. John Heil. Stamford, CT: Wadsworth, 2000. Searle, John R. "Minds, Brains, and Progr ams." Philosophy of Mind: A Guide and Anthology . Ed. John Heil. New York: Oxford University Press, 2004. Turing, Alan M. "Computing Machinery and Intelligence." Philosophy of Mind: A Guide and Anthology . Ed. John Heil. New York: Oxford University Press, 2004.

Related documents

At the Dawn of Belt and Road: China in the Developing World

At the Dawn of Belt and Road: China in the Developing World

At the Dawn of Belt and Road China in the Developing World Andrew Scobell, Bonny Lin, Howard J. Shatz, Michael Johnson, Larry Hanauer, Michael S. Chase, Astrid Stuth Cevallos, Ivan W. Rasmussen, Arthu...

More info »
China 2017 WTO Report

China 2017 WTO Report

201 7 Report to Congress On China’s WTO Compliance United States Trade Representative January 2018

More info »
kinarapport a4

kinarapport a4

CHINA AND THE END OF POVERTY IN AFRICA – TOWARDS MUTUAL BENEFIT? Penny Davies CHINA AND THE END OF POVERTY IN AFRICA – towards mutual benefi t? PENNY DAVIES This report was produced by the Swedish dev...

More info »
pisa 2012 results volume I

pisa 2012 results volume I

PISA 2012 Results: What Students Know and Can Do tICS, themA StuDent PeRfoRmAnCe In mA ReADIng AnD SCIenCe Volume I rogramme for ssessment A tudent S nternational I P

More info »
Microsoft Word   20190227 Chinas Impact on the US Education System

Microsoft Word 20190227 Chinas Impact on the US Education System

United States Senate PERMANENT SUBCOMMITTEE ON INVESTIGATIONS Committee on Homeland Security and Governmental Affairs Rob Portman, Chairman Tom Carper, Ranking Member CHINA’S IMPACT ON THE U.S. EDUCAT...

More info »
"Not an Idea We Have to Shun": Chinese Overseas Basing Requirements in the 21st Century | China Strategic Perspectives 7

"Not an Idea We Have to Shun": Chinese Overseas Basing Requirements in the 21st Century | China Strategic Perspectives 7

ategiC 7 tiveS PeRSPeC China StR “Not an Idea We Have to Shun”: Chinese Overseas st Basing Requirements in the 21 Century by Christopher D. Yung and Ross Rustici with Scott Devary and Jenny Lin Center...

More info »
301 Report Update

301 Report Update

OFFICE UNITED STATES TRADE REPRESENTATIVE of the EXECUTIVE OFFICE OF THE PRESIDENT UPDATE CONCERNING CH INA’S ACTS, POLICIES AND PRACTICES RELATE D TO TECHNOLOGY TRANSFER, INTELLECTUAL PROPERT Y, AND ...

More info »
Lives for Sale

Lives for Sale

L IVES FOR S ALE Personal Accounts of Women Fleeing North Korea to China Committee for Human Rights in North Korea

More info »
2019 CHINA MILITARY POWER REPORT

2019 CHINA MILITARY POWER REPORT

OFFICE OF THE SECRETARY OF DEFENSE People’s Republic of China Annual Report to Congress: Military and Security Developments Involving the ANNUAL REPORT TO CONGRESS Military and Security Developments I...

More info »
Defeat, Not Merely Compete: China's View of Its Military Aerospace Goals and Requirements in Relation to the United States

Defeat, Not Merely Compete: China's View of Its Military Aerospace Goals and Requirements in Relation to the United States

P N O I T C O R A O R Defeat, Not Merely Compete China’s View of Its Military Aerospace Goals and Requirements in Relation to the United States Scott W. Harold

More info »
low performing students 5jrp0jd15gtc

low performing students 5jrp0jd15gtc

PISA PISA Low-Performing Students WHY THEY FALL BEHIND AND HOW TO HELP THEM SUCCEED Low-Performing Students PISA Low-Performing Students More than one in four 15-year-old students in OECD countries ha...

More info »
diux chinatechnologytransferstudy jan 2018 (1)

diux chinatechnologytransferstudy jan 2018 (1)

Defense Innovation Unit Experimental (DIUx) China’s Technology Transfer Strategy: How Chinese Investments in Emerging Technology Enable A Strategic Competitor to Access the Crown Jewels of U.S. Innova...

More info »
12015

12015

HPG Working Paper Exploring the links between Chinese foreign policy and humanitarian action Multiple interests, processes and actors Miwa Hirono January 2018 HPG Humanitarian Policy Group

More info »
FINAL China Technology Report 6.18.18 PDF

FINAL China Technology Report 6.18.18 PDF

★ ★ ★ How China’s Economic Aggression Threatens the Technologies and Intellectual Property of the United States and the World White House Office of Trade and Manufacturing Policy June 2018

More info »
Asking the Tiger for His Skin: Rights Activism in China

Asking the Tiger for His Skin: Rights Activism in China

Fordham International Law Journal Article Issue 4 2006 30, 6 Volume Asking the Tiger for His Skin: Rights Activism in China ∗ Eva Pils ∗ c Copyright © 2006 by the authors. Fordham International Law Jo...

More info »
AndersBehringBreivikManifesto

AndersBehringBreivikManifesto

2011 , London – By Andrew Berwick

More info »
NAAS16 post election report

NAAS16 post election report

This survey was supported by a major grant from the National Science Foundation, with support for supplemental data collections from the Ford Foundation, California Immigration Research Initiative, an...

More info »