PHILOSOPHICAL PERSPECTIVES : ARTIFICIAL INTELLIGENCE
There is little doubt that computers and computer metaphor reformed psychology and philosophy of mind . Philosophical controversy generated by the early computers caused by Tin ING bold prediction that a machine can compete in a conversation with a man, and his assertion that this situation justified the application of psychological predicates to these machines . In his useful introduction to the nine essays collected in the volume rh , 1 Ringle score ili.it the ensuing debate was not really about me his powers or potential ol computers , but on the ordinary meaning of psychological terms . This is no trick to believe ” , ” believe ” or other term , that term also applies to computers and people to solve a meaning. But that would tell us exactly nothing like computers think, believe , etc. , because these concepts are usually included . the point seems to have escaped John McCarthy , the longest essay in this volume for the construction of a notion of faith , not only for people and computers , but also spends thermostats . McCarthy also grant knowledge , but in a ” passive ” books ! (165 ) The bold prediction Turing set the tone for many subsequent declaration asks just around the corner . But Hubert Dreyfus early I970 had the temerity to suggest that the area seemed receeding at about the same rate as the work progressed . In his opinion , it is no coincidence , since the digital computer is a poor model for human thought . Dreyfus has proven invaluable service to both deflate some of the most outrageous predictions of artificial intelligentsia , and highlighting the holistic nature of mental concepts . To see something as recognize a painting , Dreyfus said, can require a lot of knowledge about human practices , habits and geometry will not suffice. But in his contribution to this book , as in his earlier work , the argument Dreyfus is scarce on the essential point : why should we believe that computer simulation ol basically impossible thought? are much emphasis on the importance of Gestalt phenomena and prototypes Roschian in human thought and perception. And he is certainly right in saying that these phenomena have been simulated by a computer. Bui , as far as I can see , the claim that they can not be simulated fully argued . The trial of Thomas Simon points out the weakness of the arguments of the ” impossibility ” , highlighted by Dreyfus and others , and finds surprising connections between these arguments and ethical objections against avian influenza Weizenbaum Computer Power and Human reason .
When waxing philosophical , AI people tend to describe with traditional epistemology , their ongoing work as it investigates how knowledge is represented , updated and used effectively . There is some truth in what we think of epistemology as practiced by Locke and Hume . But in recent epistemology of psychological problems have taken to questions about the justification , skepticism and the definition of ” knowledge ” a back seat. Thus , there is surprisingly little ol contact between AI and contemporary epistemology . But, as Sayre score iu his essay , the concern of all in the representation of common sense knowledge connects to another venerable philosophical activity , viz . conceptual analysis . For a program to be displayed in a way that facilitates the inference and reveals the conceptual links between summarizing writing concepts. Required to answer questions about the concepts in history a computer on a news story and ask invoked And , very close, is the traditional purpose of conceptual analysis . But the rules iu AI are not quite the same as that of philosophy , because the computer is a literal beast infuriating and mind must all listed in the most explicit detail. It is an effective hedge against the philosophical side sway , but it also imposes a kind of myopia , focusing on the trees instead of the forest . Optimistic ( and the one I included here) can provide change in duties with philosophers who serve as scouts and conceptual AI types that detailed conceptual cartogaphy . The paper is Lehnert , McDermott and Schank give something of the flavor of the concept as it is pursued in AI but they are too short to the power and subtlety ol technical analysis to demonstrate . Neither their optimistic tone suggests the enormous obstacles remain overcome.2 The two test trials in volume I saved for last . It is widely discussed “Artificial Intelligence as philosophy and psychology ” Dennett iu which he exhibited his homuncular perspective tnind as composed of interacting sub – intelligent ( or homuiiculi ) , each in turn composed of homunculi . However , the homunculi each level are stupid , more limited , more specialized and less flexible than those directly above “Finally , this nest boxes within boxes lands you with homunculi as Stupki ( whatever they do to level is to not forget to yes or to say no if asked ) that they can be , as they say , “is replaced by a” machine ” ( 73) . The purpose of a homuncular theory is essentially the same as the target of a theory in the AI. Shows how complex tasks can be carried out ” by organizing armies of idiots to do the work ” ( 73 ) It is important to advice Dennett is that the legitimate use of internal representations in psychology for Dennett , psychology was being chased by a argument that went something like this : . internal representations ( that if they cognitive maps , mental images or phrases ) satisfy any explanatory function , unless they can be interpreted , but their interpretation requires an interpreter , a homunculus , and rely on homunculi is circular ( just because they have the capacity that we try to explain ) or leads to infinite regress ( since capacities homunculi are explained using even homunculi ) . Dennett sunk to show how the use of progressively stupid homunculi can avoid . Both circularity and infinite regress argument by Another gem in this book is the means Pylyshyn cognitive psychology to reflect some of the lessons of AI . “Psychology , ” Pylyshyn notes ” exploded with micro – . Models for phenomena that are so narrow that in many cases , they are on a single experimental procedure The discipline is largely oriented paradigm , instead of being led by the major theoretical systems ” ( 26). The remedy , in his opinion, is to build complex and integrated theories ignore around these traditional areas such as psychological learning, memory and perception. Adjacent Many of Pylyshyn ‘s article is devoted to an examination of some of the concepts and lessons learned from the construction of complex systems within Amnesty International. a practitioner of AI , the tone is refreshingly modest and tentative conclusions . But psychologists and philosophers , there is much to think about.