Mind Association

Computing Machinery and Intelligence Author(s): A. M. Turing Reviewed work(s): Source: Mind, New Series, Vol. 59, No. 236 (Oct., 1950), pp. 433-460 Published by: Oxford University Press on behalf of the Mind Association Stable URL: http://www.jstor.org/stable/2251299 . Accessed: 26/03/2012 01:15
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Oxford University Press and Mind Association are collaborating with JSTOR to digitize, preserve and extend access to Mind.

http://www.jstor.org

VOL.

LIX.

No. 236.]

1950 [October,

M IND
A QUARTERLY
OF

REVIEW

PSYCHOLOGY

AND PHILOSOPHY
AND

MACHINERY I.-COMPUTING INTELLIGENCE

BY A. M. TURING

Game. 1. The Imitation I PROPOSE tO considerthe question, 'Can machinesthink? of This shouldbeginwith definitions the meaningof the terms be 'machine' and 'think'. The definitions so might framed as to reflect far as possiblethe normaluse of the words,but this so attitudeis dangerous. If the meaningof the words'machine' and 'think'are to be found examining theyare commonly how by used it is difficult escape the conclusionthat the meaning to to and the answer tlie question,' Can machines think? ' is to be such as a Gallup poll. But thisis soughtin a statistical survey sucha definitionshallreplace I absurd. Insteadofattempting the whichis closely relatedto it and is expressed question another, by in relatively words. unambiguous The new formof the problemcan be describedin termsof a game whichwe call the 'imitationgame'. It is playedwith a threepeople, man (A), a woman(B), and an interrogator who (C) may be of eithersex. The interrogator stays in a room apart from other two. The objectofthe game forthe interrogator the is to determine whichof the othertwo is the man and whichis the woman. He knowsthemby labels X and Y, and at the end ' ofthe gamehe says either X is A and Y is B ' or ' X is B and Y is A'. The interrogator allowedto put questions A and B is to thus: C: Will X pleasetell me the length his or herhair 2 of Now suppose X is aetually A, then A must answer. It is A's
28 433

434

A. M. TURING:

objectin the game to tryand cause C to make the wrong identification. His answermight therefore be 'My hair is shingled, and the longeststrandsare about nine incheslong.' In orderthat tones of voice may not help the interrogator the answersshouldbe written, betterstill,typewritten.The or ideal arrangeimaent have a teleprinter is to communicating between the two rooms. Alternatively questionand answerscan be the repeated an intermediary.The objectofthegamefor third by the player(B) is to help the interrogator.The best strategy her for is probablyto give truthful answers. She can add such things as I am the woman,don't listento him! ' to heranswers, but it willavail nothing the man can makesimilar as remarks. We now ask the question,'What willhappenwhena machine takes thepartof A in thisgame ? ' Will the interrogator decide wrongly oftenwhen the game is played like this as he does as whenthe game is played between man and a woman? These a questions replaceour original, 'Can machines think? 2. Critique theNew Problem. of As well as asking,'What is the answerto thisnewform the of question', one may ask, 'Is this new questiona worthy one to investigate ' This latter question we investigatewithout ? further shortan infinite ado, thereby cutting regress. The new problem the advantageof drawing fairly has a sharp line between physical and theintellectual the capacitiesofa man. No engineer chemistclaims to be able to producea material or fromthe human skin. It is possible whichis indistinguishable be that at sometimethismight done,but evensupposing thisinventionavailable we shouldfeeltherewas littlepointin trying it to make a 'thinking machine' morehumanby dressing up in in have set the problem suchartificial flesh. The form whichwe, whichprevents interrogator the reflects factin the condition this or fromseeingor touching the othercompetitors, hearingtheir voices. Some otheradvantagesofthe proposed criterion may be shownup by specimen questionsand answers. Thus: Q: Please write me a sonnet on the subject of the Forth Bridge. A: Countme out on thisone. I nevercouldwrite poetry. Q: Add 34957to 70764 A: (Pause about 30 secondsand thengive as answer)105621. Q: Do you play chess? A: Yes..

COMPUTING

MACHINERY

AND

INTELLIGENCE

435

Q: I have K at my Ki, and no otherpieces. You have only K at K6 and R at RI. It is yourmove. What do you play ? A: (After pause of 15 seconds)R-R8 mate. a The question and answer method seems to be suitable for introducing almostany one ofthe fields humanendeavour of that we wish to include. We do not wish to penalisethe machine for its inability shinein beautycompetitions, to penalise to nor a man forlosingin a race againstan aeroplane. The conditions of our game make thesedisabilities irrelevant. The 'witnesses' can brag, if they consider advisable,as much as theyplease it about their charms,strength heroism, or but the interrogator cannotdemandpracticaldemonstrations. The game may perhapsbe criticised the groundthat the on odds are weighted heavilyagainstthe machine. If the man too wereto tryand pretend be the machine wouldclearly to he make a verypoorshowing. He wouldbe given awayat oncebyslowness and inaccuracy arithmetic. May notmachines in carry someout thingwhichoughtto be described thinking whichis very as but different whata man does ? This objection a verystrong from is one, but at least we can say that if,nevertheless, machinecan a be constructed play the imitation to game satisfactorily, need we not be troubled thisobjection. by It mightbe urged that when playingthe 'imitationgame' the best strategy the machinemay possiblybe something for other thanimitation thebehaviour a man. Thismaybe,but of of I thinkit is unlikely that thereis any greateffect this kind. of In any case thereis no intention investigate to herethe theory of the game, and it will be assumed that the best strategy is to tryto provide be answers thatwouldnaturally givenby a man. 3. The Mlachines in concerned theGame. The questionwhichwe put in ? 1 will not be quite definite untilwe have specified what we mean by the word 'machine'. wish permit It is natural thatweshould to kindofengineering every to technique be used in ourmachines. We also wishto allowthe or than an engineer team of engineers possibility may construct a machinewhichworks, but whosemannerof operation cannot be satisfactorily described its constructors becausetheyhave by applied a methodwhich is largelyexperimental.Finally, we menbornin theusual manner. wishto excludefrom machines the It is difficult frame definitions as to satisfy the so to thesethree conditions. One might for instance insist that the team of

436

A. M. TURING:

engineers should be all of one sex, but this would not really be satisfactory, it is probablypossible to rear a complete for individualfrom singlecell of the skin (say) of a man. To do a so would be a feat of biologicaltechniquedeserving the very of highestpraise,but we would not be inclinedto regardit as a a case of 'constructing thinking us machine'. This prompts to abandon the requirement that everykind of techniqueshould be permitted. We are the moreready to do so in view of the in fact that the presentinterest 'thinkingmachines' has been aroused by a particularkind of machine,usually called an ' electroniccomputer' or 'digital computer'. Followingthis suggestion onlypermit we digitalcomputers take part in our to game. This restriction appears at first sightto be a verydrasticone. I shall attempt showthat it is not so in reality. To do this to necessitates shortaccountofthe natureand properties these a of computers. of It may also be said that this identification machineswith like our criterion for 'thinking', will only digital computers, if be unsatisfactory (contrary my belief),it turnsout that to are in digitalcomputers unableto give a good showing the game. in There are already a numberof digital computers working and it maybe asked,'Why nottrythe experiment order, straight of the away ? It would be easy to satisfy conditions the game. of could be used, and statistics A number interrogators compiled the identification given.' Thewas short to showhow often right all answeris that we are not askingwhether digitalcomputers the at woulddo wellin the game norwhether computers present thereare imaginablecomavailable would do well,but whether answer. puterswhichwould do well. But this is onlythe short' in We shall see thisquestion a diflerent lightlater. 4. DigitalComputers. The idea behinddigitalcomputers maybe explained saying by that these machinesare intendedto carryout any operations whichcouldbe donebya human computer. Thehumancomputer fixedrules; he has no authority is supposedto be following themin any detail. We may supposethat these to deviatefrom rulesare suppliedin a book,whichis alteredwhenever is put he on to a new job. He has also an unlimited supplyof paper on which doeshiscalculations. He mayalso do hismultiplications he on and additions a 'desk machine',but thisis not important. as If we use the above explanation a definition shall be in we

COMPUTING

MACHINERY

AND INTELLIGENCE

437

danger of circularity argument. We avoid this by giving of an outlineof the means by whichthe desiredeffect achieved. is A digitalcomputer usuallybe regarded consisting three can as of parts: (i) Store. (ii) Executiveunit. (iii) Control. The storeis a storeofinformation, corresponds thehuman and to computer's paper,whether is the paper on whichhe does his this calculations that on whichhis book of rulesis printed. In so or faras the humancomputer does calculations his head a partof in the storewillcorrespond his memory. to The executiveunit is the part whichcarriesout the various individual operationsinvolved in a calculation. What these individualoperationsare will vary frommachineto machine. Usually fairly lengthy operations be done such as 'Multiply can 3540675445by 7076345687' but in some machinesonly very simpleonessuchas 'Write down0 ' are possible. We have mentioned that the ' book of rules-'suppliedto the is computer replacedin the machineby a part of the store. It is then called the 'table of instructions'. It is the dutyof the control see that theseinstructions obeyedcorrectly in to are and theright order. The control so constructed thisnecessarily is that happens. The information the storeis usuallybroken intopackets in up of moderately small size. In one machine, instance, packet for a of mightconsist ten decimaldigits. Numbers assigned the are to parts of the storein whichthe various packets of information are stored,in some systematic manner. A typical instruction mightsay'Add the number storedin position6809 to that in 4302 and put the resultback into the latterstorageposition'. Needless to say it would not occurin the machineexpressed in English. It would more likelybe coded in a formsuch as 6809430217. Here 17 says whichof variouspossibleoperations is to be performed the two numbers. In this case the operaon tion is that describedabove, viz. 'Add the number.. . .' It will be noticedthat the instruction takes up 10 digits and so forms packetofinformation, conveniently.The control one very will normally take the instructions be obeyedin the orderof to the positionsin whichthey are stored,but occasionallyan instruction such as

438

A. M. TURING:

' Now obey the instruction storedin position5606, and continuefrom there' or may be encountered, again stored 'If position4505 contains0 obey next the instruction in 6707,otherwise on.' continue straight becausethey Instructions theselattertypesare veryimportant of make it possiblefora sequenceof operations be repeatedover to is but and over again until some condition fulfilled, in doingso on instructions each repetition, the same but to obey,not fresh onesoverand overagain. To take a domestic analogy. Suppose on MotherwantsTommyto call at the cobbler'severymorning his way to schoolto see if her shoes are done, she can ask him she afresheverymorning. Alternatively can stick up a notice once and forall in the hall whichhe will see whenhe leaves for the and also to destroy schooland which tellshimto call for shoes, the noticewhenhe comesback ifhe has the shoeswithhim. can The readermustacceptit as a factthat digitalcomputers be constructed, and indeed have been constructed, according and that they can in fact we to the principles have described, veryclosely. mimic actionsof a humancomputer the we our The bookofruleswhich have described humancomputer as usingis of coursea convenient fiction. Actual human comremember puters really whattheyhavegotto do. If onewantsto make a machinemimicthe behaviourof the human computer in somecomplex operation has to ask himhowit is done,and one then translate the answerinto the formof an instruction table. Constructing instruction tables is usually describedas 'proto gramming'. To 'programme machine carryout the operaa table into the tion A' meansto put the appropriate instruction machineso thatit willdo A. is An interesting varianton the idea of a digital computer a element'.Thesehaveinstructions 'digitalcomputer witha random involving the throwing a die or some equivalentelectronic of for be,' the process; onesuchinstruction might instance Throw die such and put the resulting number into store1000'. Sometimes a machineis described havingfreewill (thoughI would not as use thisphrasemyself). It is not normally possibleto determine fromobserving machinewhether has a random element, it a fora similareffect be producedby such devicesas making can the choicesdependon the digitsof the decimalforw. have onlya finite store. There Most actual digitalcomputers is no theoretical in withan undifficulty the idea of a computer part can have been used limitedstore. Of courseonly a finite amountcan have been at any one time. Likewiseonly a finite

COMPUTING

MACHINERY

AND INTELLIGENCE

439

constructed, we can imaginemoreand morebeingadded as but required. Such computers have special theoretical interest and will be called infinitive capacitycomputers. The idea ofa digitalcomputer an old one. Charles is Babbage, Lucasian Professor Mathematics Cambiidgefrom1828 to of at 1839, planned such a machine,called the AnalyticalEngine, but it was never completed. AlthoughBabbage had all the essentialideas, his machinewas not at that time such a very attractive prospect. The speedwhichwouldhave been available wouldbe definitely faster than a humancomputer something but like 100 timesslowerthan the Manchester machine, itselfone of the slower of the modernmachines. The storage was to be purelymechanical, usingwheelsand cards. The factthat Babbage's AnalyticalEnginewas to be entirely will mechanical helpus to ridourselves a supers,tition. of Importance is often attachedto the factthat modern digitalcomputers are electrical, thatthenervous and system is electrical. Sinc', also Babbage's machinewas not electrical, and sinceall digitalcomputersare in a senseequivalent, see thatthisuse ofelectricity we cannotbe oftheoretical importance. Ofcourseelectricity usually is comes in wherefast signalling concerned, that it is not so surprising that we find it in both these connections. In the nervous systemchemicalphenomenaare at least as important as electrical. In certaincomputers storagesystem mainly the is acoustic. The featureof using electricity thus seen to be is only a very superficial similarity. If we wish to find such we similarities should look rather mathematical for analogiesof function. 5. Universality DigitalComputers. of in considered the last sectionmay be The digital computers '. state machines These are the classified the ' discrete amongst one quite machineswhichmove by suddenjumps or clicksfrom for different definite stateto another. Thesestatesare sufficiently between themto be ignored. Strictly the possibility confusion of speakingthereare no such machines. Everything reallymoves continuously.But thereare many kindsof machinewhichcan be profitably thought as being discretestate machines. For of for the instancein considering switches a lighting systemit is on a convenient fiction that each switchmust be definitely or off. There must be intermediate definitely positions,but for about them. As an example of a most purposeswe can forget discretestate machinewe mightconsidera wheel whichclicks

440

A.

M.

TURING:

round through120? once a second,but may be stopped by a leverwhichcan be operatedfrom outside; in additiona lamp is to lightin one ofthe positions the wheel. This machine of could be describedabstractlyas follows. The internalstate of the machine(whichis described the position thewheel)maybe by of of ql, q2 or q3. Thereis an inputsignali0 or i1 (position lever). The internal stateat any moment determined the last state is by and inputsignalaccording the table to Last State
q1 i0 q2 q2 q3 q2 q3 q1 q3

Input

?i

q1

The output signals, the only externallyvisible indicationof the internal state (the light)are described the table by

State q1
Output
00

q2 0?

q3

01

This exampleis typicalof discrete state machines. They can be number described suchtablesprovided by theyhave onlya finite ofpossiblestates. It will seem that given the initial state of the machineand all the inputsignalsit is alwayspossibleto predict future states. This is reminiscent Laplace's view that fromthe complete of state of the universe one moment time,as described the at of by it positionsand velocitiesof all particles, should be possibleto we predict future all which are considering states. The predictioil is, however, rather than that considered nearerto practicability by Laplace. The systemof the 'universe as a whole' is such that quite small errorsin the initial conditionscan have an of overwhelming effectat a later time. The displacement a single electronby a billionthof a centimetre one moment at mightmake the difference betweena man being killed by an avalanche a year later,or escaping. It is an essentialproperty of the mechanicalsystems whichwe have called 'discrete state machines' that thisphenomenon does notoccur. Even whenwe considerthe actual physicalmachinesinstead of the idealised machines,reasonablyaccurate knowledgeof the state at one momentyields reasonablyaccurate knowledgeany numberof stepslater.

COMPUTING

MACHINERY

AND

INTELLIGENCE

441

fall the class As we have mentioned, digitalcomputers within of discretestate machines. But the numberof states of which large. For such a machine is capable is usually enormously at instance, number themachine the for nowworking Manchester this it about 2165,000, i.e. about 1505?.??. Compare withourexample of the clicking wheel describedabove, which had three states. to of It is not difficult see whythe number states shouldbe so includesa store corresponding the to immense. The computer paper used by a humancomputer. It mustbe possibleto write of into the store any one of the combinations symbolswhich on suppose mighthave been written the paper. For simplicity 0 in that onlydigitsfrom to 9 are used as symbols. Variations are is handwriting ignored. Suppose the computer allowed 100 50 sheetsof paper each containing lines each withroomfor30 digits. Then the numberof states is 10100X5OX30, i.e. 10150,000. of This is about the number statesof threeManchester machines to put together. The logarithm the base two of the number of statesis usuallycalled the 'storage capacity' of the machine. machinehas a storagecapacity of about Thus the Manchester 165,000 and the wheel machineof our example about 1-6. If their capacitiesmust be added two machinesare put together to obtain the capacityof the resultant machine. This leads to of such as 'The Manchester machine the possibility statements contains64 magnetic trackseach witha capacityof 2560, eight storage electronic tubes with a capacityof 1280. Miscellaneous a to amounts about 300 making total of 174,380.' to Given the table corresponding a discretestate machineit is possibleto predictwhat it will do. There is no reason why this calculationshouldnot be carriedout by means of a digital quickly computer. Providedit could be carriedout sufficiently could mimicthe behaviour any discrete of the digitalcomputer the statemachine. The imitation gamecouldthenbe playedwith machinein question(as B) and the mimicking digitalcomputer wouldbe unable to distinguish them. (as A) and the interrogator must have an adequate storage Of coursethe digital computer fast. Moreover, must it capacity as well as working sufficiently for afresh each new machinewhichit is desired be programmed to mimic. This special propertyof digital computers, that they can mimic any discrete state machine, is described by saying machines. The existenceof machines that they are universal has the important with this property consequencethat, consiof to derations speed apart,it is unnecessary designvariousnew machinesto do various computing processes. They can all. be

442

A. M. TURING:

done with one digital computer, suitablyprogrammed each for case. It willbe seenthatas a consequence thisall digitalcomof putersare in a senseequivalent. We may now consideragain the pointraisedat the end of ?3. It was suggested tentatively that the question,'Can machines think? ' should be replaced by 'Are there imaginabledigital computerswhich would do well in the imitationgame ? ' If we wish we can make this superficially more generaland ask 'Are there discrete state machines which would do well? ' But in view of the universality property see that eitherof we these questionsis equivalentto this, 'Let us fix our attention on oneparticular digitalcomputer Is it truethatbymodifying C. this computer have an adequate storage, to its suitably increasing speedofaction, and providing withan appropriate it programme, C can be madeto play satisfactorily partofA in theimitation the game,the partof B beingtakenby a man ? 6. Contrary Viewson theMain Question. We may now consider groundto have been clearedanldwe the are readyto proceed thedebateon ourquestion, to 'Can machines think? ' and thevariantofit quotedat the endofthelast section. We cannotaltogether abandonthe original form the problem, of for opinionswill differ to the appropriateness the substituas of tion and we must at least listento what has to be said in this connexion. It willsimplify matters the readerif I explainfirst own for my in beliefs thematter. Consider themoreaccurate form the first of question. I believe thatin a-bout years'timeit willbe possible fifty to programme computers, witha storagecapacityof about 109, to make themplay the imitation game so well that an average interrogator nothave more will than70 percent.chanceofmaking the rightidentification afterfiveminutesof questioning. The originalquestion,'Can machinesthink ' I believe to be too ? to I meaningless deserve discussion. Nevertheless believethat at theendofthecentury useofwords general the and educated opinion willhavealtered much so thatonewillbe able to speakofmachines without thinking expecting be contradicted. I believefurther to that no useful purpose is served by concealingthese beliefs. The popularview that scientists proceedinexorably fromwellestablishedfact to well-established fact,neverbeing influenced is by any unprovedconjecture, quite mistaken. Providedit is made clear whichare proved facts and whichare conjectures, no harm can result. Conjectures of greatimportance are since linesof research. theysuggest useful

COMPUTING

MACHINERY

AND INTELLIGENCE

443

I now proceedto consider opinions opposedto my own. (1) The Theological Objection. Thinking a function man's is of immortal soul. Godhas givenan immortal to every soul manand woman,but not to any otheranimal or to machines. Hence no animal or machine think. can I am unable to accept any part of this, but will attemptto reply in theological terms. I should findthe argument more if convincing animalswereclassedwithmen,forthereis a greater to difference, my mind, betweenthe typical animate and the inanimatethan there is betweenman and the other animals. The arbitrary character the orthodox of view becomesclearerif we consider how it mightappear to a memberof some other religious community.How do Christians regard Moslemview the that womenhave no souls? But let us leave this point aside and return the main argument. It appears to me that the to argument quotedabove impliesa seriousrestriction the omniof potenceof the Almighty. It is admittedthat thereare certain thingsthat He cannotdo such as makingone equal to two,but should we not believethat He has freedom confer soul on a to an elephantif He sees fit? We mightexpectthat He would only exercise this powerin conjunction with a mutationwhich providedthe elephantwith an appropriately improvedbrainto minister theneedsofthissoul. An argument exactlysimilar to of form maybe madefor case ofmachines. It mayseemdifferent the because it is moredifficult " swallow". But this reallyonly to means that we think- would be less likelythat He would conit sider the circumstances suitableforconferring soul. The cira in cumstances questionare discussedin the rest of this paper. In attempting constructsuch machineswe should not be to irreverently usurping His powerofcreating souls,anymorethan we are in the procreation children: ratherwe a-re, either of in case, instruments His willproviding of mansions the souls that for He creates. However,this is mere speculation. I am not veryimpressed withtheological arguments whatever theymaybe usedto support. Such arguments in have often beenfound unsatisfactory the past. In the time of Galileo it was arguedthat the texts, " And the sun stood still . . . and hasted not to go down about a whole day " (Joshuax. 13) and " He laid the foundations the earth, of
1Possibly thisviewis heretical.St. Thomas Aquinas(Summa Theologica, quoted by Bertrand Russell,p. 480) statesthat God cannotmake a man to have no soul. But thismaynotbe a realrestriction His powers, on but only a resultof the fact that men's souls are immortal, and therefore indestructible.

444

A. M. TURING:

that it should not move at any time" (Psalm cv. 5) were an theory. Withour present of adequate refutation the Copernican appears futile. When that knowsuch an argument knowledge impression. ledgewas notavailableit made a quitedifferent (2) The 'Heads in theSand' Objection. " The consequences would be too dreadful. Let us hope and of machinesthinking believethattheycannotdo so." is This argument seldomexpressedquite so openlyas in the mostof us who thinkabout it at all. form above. But it affects to We liketo believethatMan is in somesubtleway superior the rest of creation. It is best if he can be shownto be necessarily his of is for superior, thenthere no danger himlosing commanding is argument clearly position. The popalarityof the theological in withthisfeeling. It is likelyto be quite strong inconnected more tellectualpeople, since they value the power of thinking highlythan others,and are more inclinedto base theirbelief of in the superiority Man on thispower. substantial is I do not thinkthat this argument sufficiently would be moreappropriate: to requirerefutation. Consolation of in perhapsthisshouldbe sought the transmigration souls. of Objection. Thereare a number results (3) The Mathematical of mathematical logic which can be used to show that there machines. The to are limitations the powersof discrete-state and best knownof these resultsis knownas G6del's theorem,' statements logicalsystem powerful showsthat in any sufficiently be whichcan neither provednor disproved can be formulated is itself inconsistent. the unlesspossibly system within system, the resultsdue to Church, similar, in Thereare other, some respects Kleene,Rosser,and Turing. The latterresultis the most conwhereas to directly machines, sinceit refers venientto consider, indirect argument: can the others onlybe used in a comparatively is if forinstance G6del'stheorem to be used we need in addition logical systemsin termsof to have some means of describing in of and in machines, machines terms logicalsystems. The result a whichis essentially digital refers a typeof machine to question capacity. It states that there are computerwith an infinite cannotdo. If it is rigged to up thatsucha machine certain things game,therewill be as to giveanswers questions in the imitation answer,or fail givea wrong it to somequestions which willeither muchtimeis allowedfora reply. to givean answer all however at Theremay, of course,be many such questions,and questions whichcannotbe answered one machinemay be satisfactorily by
1Author's to namesin italicsrefer the Bibliography.

COMPUTING

MACHINEkY

AND INTELLIGENCE

445

for answered another. We are ofcoursesupposing thepresent by that the questionsare of the kind to whichan answer 'Yes ' rather than questionssuch as 'What do or 'No' is appropriate, you think of Picasso ?-' The questions that we know the the machinesmust fail on are of this type," Consider machine specified follows.. . . Will this machineever answer'Yes' as to any question? " The dots are to be replaced by a deswhich could be criptionof some machinein a standardform, like that used in ? 5. When the machinedescribed something simple relationto the machine bears a certaincomparatively it whichis underinterrogation, can be shownthat the answer This is the mathematical is eitherwrongor not forthcoming. to result: it is arguedthat it proves a disabilityof machines is whichthe humanintellect not subject. The short answer to this argumentis that although it is establishedthat there are limitationsto the powers of any particularmachine,it has only been stated, withoutany sort of proof, apply to the human intellect. thatno such limitations quite so lightly. But I do not thinkthis view can be dismissed Whenever one of these machines is asked the appropriate we answer, knowthat this and givesa definite criticalquestion, answer must be wrong,and this gives us a certainfeelingof superiority.Is this feelingillusory? It is no doubt quite genuine,buitI do not thinktoo much importanceshould be attached to it. We too oftengive wronganswersto questions in ourselves be justified beingverypleased at suchevidenceof to our fallibility the part ofthe machines. Further, superiority on to can onlybe felt suchan occasionin relation theonemachine on over whichwe have scored our petty triumph. There would over all machines. be no questionof triumphing simultaneously In short,then,there mightbe men clevererthan any given but thenagain theremightbe othermachinescleverer machine, again, and so on. I wouLld,think, argument Those whoholdto the mathematical game as a basis for mostlybe willingto accept the imitation discussion. Those who believe in the two previousobjections in would probablynot be interested any criteria. is Consciousness.Thisargument very (4) The Aryumentfrom Jefferson's Lister Orationfor 1949, well expressedin Professor from whichI quote. " Not untila machinecan writea sonnet or composea concertobecause of thoughtsand emotionsfelt, and not by the chance fall of symbols,could we agree that machineequals brain-that is, not onlywriteit but Inow that could feel (and not merely it. No mechanism it had written

446

A. M. TURING:

artificially signal,an easy contrivance) pleasureat its successes, griefwhen its valves fuse, be warmed by flattery, made be miserableby its mistakes, charmedby sex, be angryor debe pressedwhenit cannot -getwhat it wants." This argument appears to be a denial of the validityof our test. According the mostextreme to form thisview the only of way by whichone could be surethat a machinethinksis to be the machineand to feel oneself thinking. One could then describe these feelings the world,but of course no one would to be justified taking any notice. Likewise accordingto this in view the only way to know that a man thinksis to be that particular man. It is in factthe solipsist pointofview. It may be the mostlogicalview to hold but it makes communication of ideas difficult. is liable to believe 'A thinksbut B does not' A whilst believes' B thinks A does not '. Insteadofarguing B but continually over this point it is usual to have the polite conventionthat everyone thinks. Jefferson does not wish to adopt I am sure that Professor and solipsist the extreme pointof view. Probablyhe would be to quite willing accept the imitation game as a test. The game (with the player B omitted) is frequently used in practice someone really underthe name of viva voceto discover whether or understands something has ' learntit parrotfashion'. Let us listenin to a partofsucha vivavoce: Interrogator: In the firstline of your sonnet which reads ' Shall I comparethee to a summer's day ', would not ' a spring day ' do as well or better? Witness: It wouldn'tscan. Interrogator: How about ' a winter'sday' That would scan all right. to Witness: Yes, but nobodywantsto be compared a winter's day. Would you say Mr. Pickwickremindedyou of Interroga.tor: ? Christmas Witness: In a way. is Interrogator: Yet Christmas a winter'sday, and I do not thinkMr. Pickwickwould mindthe comparison. Witness: I don't thinkyou're serious. By a winter'sday one than a special one like means a typicalwinter's day, rather Christmas. Jefferson ifthe sonnetAnd so on. What wouldProfessor say ? likethis in the vivavoce I was able to answer machine writing he do not knowwhether would regardthe machineas 'merely

COMPUTING

MACHINERY

AND

INTELLIGENCE

447

' but if the answerswere as artificially signalling theseanswers, and sustainedas in the above passage I do not satisfactory think he would describeit as 'an easy contrivance'. This intended coversuchdevicesas theinclusion to phraseis, I think, in the machineof a recordof someonereadinga sonnet,with to timeto time. appropriate switching turnit on from In shortthen,I thinkthat most of those who supportthe could be persuadedto abandon it from consciousness argument rather thanbe forced intothe solipsist position. They willthen probably willing accept ourtest. be to I do not wishto give the impression that I thinkthereis no about consciousness. Thereis, forinstance, mystery something of a paradox connected withany attemptto localiseit. But I do notthink thesemysteries need to be solved before necessarily we can answerthe question with which we are concernedin this paper. (5) Argumentsfrom Disabilities. Thesearguments Various take " the form, I grantyou that you can make machines all the do but thingsyou have mentioned you will neverbe able to make one to do X". NumerousfeaturesX are suggestedin this a connexion. I offer selection: Be kind, resourceful, beautiful, friendly 448),have initiative, (p. have a sense of humour, tell rightfromwrong, make mistakes (p. 448), fallin love,enjoystrawberries cream(p. 448),make and some one fall in love with it, learnfrom experience (pp. 456 f.), use words properly, the subjectof its own thought(p. 449), be have as much diversity behaviouras a man, do something of reallynew (p. 450). (Someofthesedisabilities givenspecial are as consideration indicatedby the page numbers.) No supportis usually offered these statements. I believe for founded the principle scientific on theyare mostly of induction. A manhas seenthousands machines hislifetime. Fromwhat of in he sees ofthemhe drawsa number general of conclusions. They are ugly, each is designedfor a very limitedpurpose,when requiredfora minutely different purposethey are useless,the varietyof behaviour any one of themis verysmall,etc., etc. of Naturallyhe concludesthat these are necessaryproperties of machinesin general. Many of these limitations associated are withthe verysmall storagecapacityof most machines. (I am assumingthat the idea of storagecapacityis extendedin some way to cover machines other than discrete-state machines.

448

A. M. TURING:

accuracy as does The exactdefinition notmatter no mathematical discussion.) A few years ago, when is claimed in the piresent it verylittlehad been heard of digitalcomputers, was possible them,if one mentioned concerning to elicit much incredulity their propertieswithout describingtheir construction.That due to a similarapplicationof the principle was presumably are induction. These applicationsof the principle of scientific of course largelyunconscious. When a burnt child fears the fireand showsthat he fearsit by avoidingit, I shouldsay that induction. (I could of course also he was applyingscientific describehis behaviourin many otherways.) The worksand customsof mankinddo not seem to be very suitable material induction. A very large part of to which to apply scientific if space-timemust be investigated, reliable resultsare to be we obtained. Otherwise may (as most English childrendo) speaks English, and that it is silly to decide that everybody learnFrench. to special remarks be made about many Thereare, however, to that have been mentioned. The inability of the disabilities and cream may have struckthe reader as enjoy strawberries frivolous. Possibly a machine might be made to enjoy this deliciousdish, but any attemptto make one do so would be is about this disability that it conidiotic. What is important of e.g. disabilities, to thedifficulty the to tributes someoftheother as man and machine between occurring same kindoffriendliness betweenwhiteman and whiteman, or betweenblack man and black man. " The claim that " machinescannot make mistakes seems a " to curiousone. One is tempted retort, Aretheyany theworse and attitude, forthat ? " But let us adopt a moresympathetic can be try to see what is really meant. I thinkthis criticism of gamne. It is claimedthatthe explainedin terms the imitation the the could distinguish machinefrom man simply interrogator by settingthem a numberof problemsin arithmetic. The machine would be unmaskedbecause of its deadly accuracy. for The reply to this is simple. The machine (programmed answers playingthe game) would not attemptto give the right problems. It would deliberatelyintroduce to the arithmetic the calculatedto confuse interrogator.A mistakesin a manner an show itselfthrough unsuitfaultwouldprobablymechanical able decision as to what sort of a mistaketo make in tbe of arithmetic. Even this interpretation the criticismis not the sympathetic. But we cannotafford space to go sufficiently depends intoit muchfurther. It seemsto me thatthiscriticism

COMPUTING

MACHINERY

AND INTELLIGENCE

449

betweentwo kinds of mistake. We may call on a confusion ' '. of of them' errors functioningand ' errors conclusion Errors are of functioning due to some mechanicalor electricalfault which causes the machine to behave otherwisethan it was designed to do. In philosophical discussions one likes to discussing of ignore the possibility such errors; one is therefore are 'abstract machines'. Theseabstractmachines mathematical they are fictions ratherthan physical objects. By definition of incapable of errors functioning.In this sense we can truly say that 'machinescan nevermake mistakes'. Errorsof conclusion can only arise when some meaningis attached to the output signals from the machine. The machine might,for instance,type out mathematicalequations, or sentences in English. When a false propositionis typed we say that the of an machinehas committed error conclusion. Thereis clearly no reason at all for saying that a machinecannot make this but type out repeatedly kind of mistake. It mightdo nothing ' 0 - 1 '. To take a less perverse example,it mighthave some induction. We method for drawingconclusionsby scientific must expect such a methodto lead occasionallyto erroneous results. The claim that a machinecannot be the subject of its own if thoughtcan of courseonlybe answered it can be shownthat withsomesubjectmatter. Neverthe machine some has thought ' theless, 'the subject matterof a machine's operations does seemto meansomething, least to the peoplewho deal withit. at If, for instance,the machinewas tryingto finda solutionof the equationx2 - 40x - 11 0 one woald be temptedto describethis equationas part of the machine'ssubject matterat can that moment. In this sortof sense a machineundoubtedly be its own subject matter. It may be used to help in making in of the effect alterations or up its own programmes, to predict of the its ownstructure. By observing results its ownbehaviour so its it can modify ownprogrammes as to achievesomepurpose These are possibilitiesof the near future, more effectively. ratherthan Utopiandreams. -cannot have much diversity The criticism that a machine of behaviouris just a way of sayingthat it cannothave much storage capacity. Until fairlyrecentlya storage capacity of even a thousanddigitswas veryrare. hereare often The criticisms disguised that we are considering consciousness. Usuallyifone mainfrom forms the argument of the and describes can tainsthat a machine do one ofthesethings, kind of methodthat the machinecould use, one will not make 29

450

A. M. TURING:

muchofan impression. It is thought thatthemethod(whatever it may be, for it must be mechanical)is really ratherbase. the in Compare parenthesis Jefferson's statement quotedon p. 21. (6) Lady Lovelace's Objection. Our mostdetailedinformation of Babbage's Analytical -Enginecomesfrom memoirby Lady a Lovelace. In it she states," The AnalyticalEngine has no pretensionsto originate anything. It can do whatever know we how " toorderit to perform (her italics). This statement quoted is by Hartree(p. 70) who adds: " This does not imply that it may not be possible to constructelectronic equipmentwhich will 'thinkforitself or in which, biological ', in terms, one could set up a conditioned reflex, whichwould serve as a basis for 'learning'. Whether this is possiblein principleor not is a stimulating and excitingquestion,suggested some of these by recentdevelopments. But it did not seem that the machines or constructed projectedat the timehad thisproperty ". I am in thorough agreement withHartreeover this. It will be noticedthat he does not assertthat the machines question in had not got the property, rather but that the evidenceavailable to Lady Lovelacedidnotencourage to believethattheyhad it. her It is quite possiblethat the machinesin questionhad in a sense gotthisproperty. For supposethatsomediscrete-state machine has the property. The Analytical Engine was a universal digitalcomputer, that,if its storagecapacityand speed were so adequate, it could by suitableprogramming made to mimic be the machine in question. Probably this argumentdid not occurto the Countessor to Babbage. In anycase therewas no on obligation themto claim all that could be claimed. This wholequestion willbe considered again undertheheading of learning machines. A variantof Lady Lovelace's objectionstatesthat a machine can ' neverdo anything reallynew'. This may be parried a for momentwith the saw, 'There is nothingnew under the sun'. Who can be certainthat 'original work' that he has done was not simplythe growth the seed plantedin him by teaching, of or the effectof following well-known general principles. A bettervariantof the objectionsays that a machinecan never 'take us by surprise'. This statement a moredirectchallenge is and can be met directly. Machinestake me by surprise with greatfrequency. This is largelybecause I do not do sufficient calculation decidewhatto expectthem do,orrather to to because, I I although do a calculation, do it in a hurried, slipshod fashion, ' takingrisks. PerhapsI say to myself, I suppose the voltage hereoughtto be the same as there: anywaylet's assumeit is '.

COMPUTING

MACHINERY

AND INTELLIGENCE

451

for is and wrong, theresult a surprise me for I Naturally am often have been is by the timethe experiment done theseassumptions forgotten.These admissionslay me open to lectureson the subjectof my viciousways,but do not throwany doubt on my I to whenI testify the surprises experience. credibility I do not expectthis replyto silencemy critic. He will promentalact are bablysay that suchsurprises due to somecreative no on mypart,and reflect crediton the machine. This leads us the and consciousness, farfrom idea from back to the argument we of surprise. It is a line of argument must considerclosed, of that remarking the appreciation somebut it is perhapsworth requiresas much of a ' creativementalact ' thingas surprising froma man, a book, a event originates the surprising, whether else. machineor anything is The view that machinescannotgive rise to surprises due, and I believe,to a fallacyto whichphilosophers mathematicians that as soon as subject. This is the assumption are particularly a fact is presentedto a mind all consequencesof that fact withit. It is a veryusespringinto the mind simultaneously but ful assumptionunder many circumstances, one too easily of forgets thatit is false. A naturalconsequence doingso is that one then assumesthat thereis no virtuein the mere working principles. data and general from out of consequences in from Contitnuity the Nervous System. The (7) Argument machine. A nervous systemis certainlynot a discrete-state impulse about the size ofa nervous in smallerror theinformation to may make a large difference the size on impinging a neuron, of the outgoing impulse. It may be arguedthat,this beingso, one cannot expect to be able to mimicthe behaviourof the system. witha discrete-state nervoussystem from machinemustbe different It is truethat a discrete-state of machine. But ifwe adhereto the conditions the a continuous will imitationgame, the interrogator not be able to take any can be made clearer advantageofthisdifference.The situation if we consider some other simpler continuousmachine. A analyser differential analyserwill do very well. (A differential type used is a certainkind of machinenot of the discrete-state for some kinds of calculation.) Some of these provide their answersin a typed form,and so are suitable for taking part in the game. It would not be possible for a digital computer analyser to predict exactly what answers the differential would give to a problem,but it would be quite capable of if givingthe rightsort of answer. For instance, asked to give the value of r (actually about 3.1416) it would be reasonable

452

A. M. TURING:

to choose at randombetweenthe values 3*12,3*13,3*14,3-15, 0.55, 019, 006 (say). of 3 16 with the probabilities 005, 0O15, for it Under these circumstances would be very difficult the analyser fromthe the differential to interrogator distinguish digital computer. of fromInformality Behaviour. It is not (8) The Argument to possible to producea set of rules purporting describewhat a man should do in every conceivableset of circumstances. One mightforinstancehave a rule that one is to stop when light,and to go if one sees a greenone, one sees a red traffic ? but what if by some fault both appear together One may perhaps decide that it is safest to stop. But some further to thisdecisionlater. To attempt may difficulty well arisefrom even those providerulesof conductto cover everyeventuality, lights,appears to be impossible. With all arisingfromtraffic this I agree. From this it is arguedthat we cannot be machines. I shall but I fear I shall hardlydo it the argument, try to reproduce like this. 'If each man justice. It seems to run something his set had a definite of rulesof conductby whichhe regulated life he would be no betterthan a machine. But thereare no such rules, so men cannot be machines.' The undistributed is middleis glaring. I do notthink argument everput quite the used nevertheless. like this, but I believe this is the argument between'rules of There may howeverbe a certainconfusion ' conduct' and ' laws ofbehaviour to cloudtheissue. By ' rules suchas ' Stop ifyou see redlights', of conduct' I meanprecepts on whichone can act, and of whichone can be conscious. By 'laws of behaviour' I mean laws ofnatureas appliedto a man's bodysuchas ' ifyoupinchhimhe willsqueak '. If we substitute whichregulatehis life' for'laws of conduct 'laws of behaviour his quoted the unby whichhe regulates life' in the argument middle is no longer insuperable. For we believe distributed by thatit is not onlytruethat beingregulated laws ofbehaviour a impliesbeing some sort of machine (thoughnot necessarily but discrete-state beingsucha machine machine), thatconversely implies being regulatedby such laws. However,we cannot laws of so easily convinceourselvesof the absence of complete behaviouras of completerules of conduct. The only way we such laws is scientific and we observation, know of forfinding underwhichwe could say, know of no circumstances certainly 'We have searchedenough. Thereare no such laws.' moreforcibly that any such statement We can demonstrate would be unjustified. For supposewe could be sure of;finding

COMPUTING

MACHINERY

AND INTELLIGENCE

453

machine such laws if theyexisted. Then givena discrete-s,,ate sufficent to by be it shouldcertainly possible discover observation a and its behaviour, thiswithin reasonabout it to predict future able time,say a thousandyears. But this does not seemto be a computer small the case. I have set up on the Manchester the whereby machine usingonly1000unitsofstorage, programme supplied with one sixteenfigurenumberreplieswith another these withintwo seconds. I would defyanyoneto learn from to about the programme be able to predictany repliessufficient repliesto untriedvalues. Perception. I assume fromExtra-Sensory (9) The Argument perthat the readeris familiarwith the idea of extra-sensory ception,and the meaningof the fouritemsof it,viz.telepathy, and psycho-kinesis.These disturbprecognition clairvoyance, ideas. ing phenomenaseem to deny all our usual scientific the How we should like to discreditthem! Unfortunately It is at evidence, least fortelepathy, overwhelming. is statistical one's ideas so as to fitthesenew facts to verydifficult rearrange in. Onceone has acceptedthemit does not seema verybig step to believein ghostsand bogies. The idea that our bodiesmove with simplyaccordingto the knownlaws of physics,together some othersnot yet discoveredbut somewhatsimilar,would be one of the first go. to one. One can say This argument to mymindquite a strong is in replythat many scientific theoriesseem to remainworkable in practice,in spite of clashingwith E.S.P.; that in fact one about it. This is rather can get alongverynicelyif one forgets is and one fearsthat thinking just the kind of cold comfort, relevant. whereE.S.P. may be especially phenomenon run based on E.S.P. might as follows: A morespecific argument a game,usingas witnesses man who "Let us play the imitation and a digitalcomputer. The is good as a telepathicreceiver, can interrogator ask such questionsas ' What suitdoes the card or in my right hand belongto ? ' The man by telepathy clairvoyancegivesthe rightanswer130 timesout of400 cards. The machinecan onlyguess at random,and perhapsgets 104 right, identification."Thereis an makesthe right so the interrogator whichopenshere. Supposethedigitalcompossibility interesting puter containsa random numbergenerator. Then it will be naturalto use thisto decidewhatanswerto give. But thenthe will be subjectto the psycho-kinetic randomnumber generator might powersof the interrogator.Perhaps this psycho-kinesis cause the machineto guess rightmore oftenthan would be calculation,so that the interrogator expectedon a probability

454

A. M. TURING:

mightstill be unable to make the rightidentification.On the without anyquestionbe other hand,he might able to guessright may happen. ing, by clairvoyance. With E.S.P. anything If telepathyis admittedit will be necessaryto tightenour test up. The situationcould be regardedas analogousto that weretalkingto himself which would occur if the interrogator withhis ear to the wall. was and one ofthe competitors listening room' would into a 'telep-thy-proof To put the competitors satisfyall requirements. 7. LearningMachines. thatI have no veryconvincing The readerwillhave anticipated my of arguments a positivenatureto support views. If I had I should not have taken such pains to point out the fallaciesin contrary views. Such evidenceas I have I shall now give. Let us returnfor a momentto Lady Lovelace's objection, whichstatedthat the machinecan onlydo whatwe tell it to do. One could say that a man can ' inject' an idea intothe machine, and that it will respond to a certain extent and then drop by struck a hammer. Another like intoquiescence, a piano string simile would be an atomic pile of less than criticalsize : an the pile to injected idea is to correspond a neutronentering will disturbance from without. Each suchneutron causea certain the dies away. If, however, size of the pile is whicheventually causedbysuchan incoming the sufficiently increased, disturbance untilthe whole neutron will verylikelygo on and on increasing for phenomenon pile is destroyed. Is there a corresponding ? minds,and is there one for machines There does seem to of be one forthe humanmind. The majority themseem to be ' sub-critical',i.e. to correspond this analogyto piles of subin to criticalsize. An idea presented such a mindwill on average proportion give rise to less than one idea in reply. A smallish are super-critical.An idea presented such a mindmay give to and tertiary of rise to a whole 'theory' consisting secondary, moreremoteideas. Animalsmindsseem to be very definitely to sub-critical. Adhering this analogywe ask, 'Can a machine ?' be made to be super-critical The 'skin ofan onion' analogyis also helpful. In considering the functions the mindor the brainwe findcertainoperations of terms. This we say whichwe can explainin purelymechanical does not correspond the real mind: it is a sortof skinwhich to we muststripoff we are to findthe real mind. But then in if off, whatremains find further to be stripped and so on. we skin a

COMPUTING

MACHINERY

AND INTELLIGENCE

455

in Proceeding thisway do we ever come to the 'real' mind,or do we eventually cometo the skinwhichhas nothing it ? In in the latter case the whole mind is mechanical. (It would not be a discrete-state machinehowever. We have discussedthis.) These last two paragraphsdo not claim to be convincing arguments. They should rather be describedas 'recitations tendingto producebelief'. The onlyreallysatisfactory support that can be givenforthe view expressed the beginningr ? 6, will be thatprovided at of by for and waiting the end ofthe century thendoingthe experiment described. But what can we say -in the meantime What ? steps should be taken now if the experimentis to be successful ? is the one As I have explained, problem mainly ofprogramming. will Advances' engineering have to be made too, but it seems in unlikelythat these will not be adeqaate forthe requirements. Estimatesof the storagecapacity of the brain vary from1010 to 1015binarydigits. I inclineto the lowervalues and believe is that only a very small fraction used forthe highertypes of used forthe retention visual thinking. Most of it is probably of if impressions. I shouldbe surprised morethan 109was required of forsatisfactory playing theimitation game,at anyrateagainst a blind man. (Note-The capacity of the Encyclopaedia is Britannica,11thedition, 2 X 109.) A storagecapacityof 107 even by presenttechwould be a very practicablepossibility niques. It is probablynot necessary increasethe speed of to operationsof the machinesat all. Parts of modernmachines which can be regardedas analogues of nerve cells workaboat a thousandtimesfasterthan the latter. This shouldprovidea ' ' marginof sa.fety which could cover losses of speed arising in manyways. Ourproblem thenis to find howto programme out thesemachines playthegame. At mypresent to rateofworking I produceabout a thousanddigitsof programme day, so that a about sixty workers, working steadilythroughthe fifty years the wentinto the waste-paper mightaccomplish job, if nothing basket. Some more expeditiousmethod seems desirable. In the processof trying imitatean adult hiuman to mindwe are bound to thinka good deal about the process whichhas broughtit to the state that it is in. We may notice three components, (a) The initialstate ofthe mind,say at birth, to (b) The education whichit has been subjected, not to be describedas education,to (e) Other experience, whichit has been subjected.

456

A. M. TURING:

to Instead of tryingto produce a programme simulatethe try adult mind,why not rather to produceone whichsimulates the child's? If this were then subjected to an appropriate course of education one would obtain the adult brain. Prelike a note-bookas one is sumablythe child-brain something and lots the stationers. Rather littlemechanism, buys it from our pointof are and writing from of blank sheets. (Mechanism view almost synonymous.) Our hope is that thereis so little like that something it can be easily mechanism the child-brain in programmed. The amount of work in the education we can to approximation, be muchthe same as forthe assume,as a first human child. We have thus divided our problem into two parts. The and child-programme the educationprocess. These two remain verycloselyconnected. We cannotexpectto finda good childwith attempt. Onemustexperiment teaching the machine-at first one such machineand see how well it learns. One can thentry anotherand see if it is betteror worse. There is an obvious by betweenthis processand evolution, the identificonnection cations material Structure the childmachine= Hereditary of = Mutations ,, Changes ,, = of Judgment the experimenter Natural selection willbe moreexpeditious thatthisprocess Onemayhope,however, is than evolution. The survivalof the fittest a slow methodfor by measuringadvantages. The experimenter, the exerciseof is shouldbe able to speedit up. Equally important intelligence, to the fact that he is not restricted randommutations. If he thinkofthe he can tracea cause forsomeweakness can probably whichwillimprove kindofmutation it. It will not be possible to apply exactly the same teaching processto the machineas to a normalchild. It willnot, for instance,be providedwith legs, so that it could not be asked to go out and fillthe coal scuttle. Possiblyit mightnot have mightbe overcome eyes. But howeverwell these deficiencies to one by cleverengineering, could not send the creature school makingexcessivefunof it. It must without the otherchildren be given some tuition. We need not be too concernedabout the legs, eyes, etc. The example of Miss Helen Keller shows that education can take place provided that communication in both directions betweenteacherand pupil can take place by some means or other.

COMPUTING

AMACHINERY AND INTELLIGENCE

457

We normallyassociate punishments and rewardswith the teaching process. Some simple child-machines can be constructed programmed this sortof principle. The machine or on has to be so constructed that eventswhichshortly preceded the occurrence a punishment-signal unlikelyto be repeated, of are whereas a reward-signal increasedthe probability repetition of of the eventswhichled up to it. These definitions not predo suppose any feelings the part of the machine. I have done on some experiments with one such child-machine, succeeded and in teachingit a few things,but the teachingmethodwas too unorthodox the experiment be considered for to reallysuccessful. The use of punishments rewardscan at best be a part of and the teachingprocess. Roughlyspeaking,if the teacherhas no other means of communicating the pupil, the amount of to information whichcan reachhimdoes not exceedthe total number of rewardsand punishments applied. By the time a child has learntto repeat 'Casabianca' he would probablyfeel very sore indeed,if the text could only be discovered a 'Twenty by Questions' technique,every 'NO' takingthe formof a blow. It is necessarytherefore have some other 'unemotional' to channelsof communication.If theseare available it is possible to teach a machineby punishments rewards obey orders and to givenin somelanguage,e.g. a symbolic language. These orders are to be transmitted throughthe 'unemotional' channels. The use of this language will diminish greatlythe numberof and punishments rewardsrequired. Opinionsmay vary as to the complexity whichis suitablein the child machine. One mighttry'to make it as simple as possibleconsistently withthe generalprinciples. Alternatively ' one might havea complete system logicalinferencebuiltin '., of In the lattercase the storewould be largelyoccupiedwithdefinitions propositions. The propositions and wouldhave various kinds of status,e.g. well-established facts,conjectures, mathematicallyprovedtheorems, statements given by an authority, expressions havingthe logicalform proposition not beliefof but value. Certain propositions may be described 'imperatives'. as The machineshould be so constructed that as soon as an im' perativeis classed as 'well-established the appropriate action automatically takesplace. To illustrate this,supposetheteacher says to the machine,'Do your homework now'. This may cause " Teachersays 'Do yourhomework now ' " to be included amongst well-established the facts. Another suchfactmight be,
1Or rather 'programmed forourchild-machine be programmed in' will in a digitalcomputer. But the logicalsystem nothave to be learnt. will

458

A. M. TURING:

"Everythingthat teachersays is true". Combining thesemay eventuallylead to the imperative, 'Do yourhomework now', being included amongst the well-established facts, and this, by the construction themachine, of willmeanthatthehomework actually gets started,but the effect very satisfactory.The is processesof inference used by the machineneed not be such as would satisfy the most exactinglogicians. There mightfor instancebe no hierarchy types. But thisneed not mean that of type fallacieswill occur, any more than we are bound to fall over unfencedcliffs. Suitable imperatives(expressed within the systems, forming not part of the rules of the system)such as ' Do not use a class unless it is a subclass of one whichhas been mentioned teacher' can have a similar by effect 'Do not to go too near the edge'. The imperatives that can be obeyed by a machinethat has no limbsare boundto be of a rather intellectual character, in as the example(doing homework) givenabove. Important amongst such imperatives be ones whichregulatethe orderin which will the rules of the logical systemconcernedare to be applied. For at each stage whenone is usinga logical system, thereis a very large numberof alternativesteps, any of which one is to permitted apply,so faras obedienceto the rulesofthe logical is system concerned. Thesechoicesmakethe difference between a brilliant and a footling reasoner, the difference not betweena soundand a fallacious one. Propositions leadingto imperatives of this kind mightbe " When Socrates is mentioned, use the syllogism Barbara " or " If one method in has been provedto be ". Some quickerthan another,do not use the slowermethod of these may be 'given by authority',but othersmay be produced by the machineitself, by scientific e.g. induction. The idea of a learningmachinemay appear paradoxical to some readers. How can the rules of operationof the machine how the machine change? They should describe completely will react whateverits historymight be, whatever changes it might undergo. The rules are thus quite time-invariant. This is quite true. The explanationof the paradox is that the rules whichget changedin the learning processare of a rather less pretentious kind,claiming onlyan ephemeral validity. The readermay draw a parallelwiththe Constitution the United of States. feature a learning machineis that its teacher Alnimportant of will oftenbe very largelyignorantof quite what is going on inside,althoughhe may still be able to some extentto predict his pupil's behaviour. This should apply most strongly the to

COMPUTING

MACHINERY

AND INTELLIGENCE

459

later educationof a machinearisingfroma child-machine of well-trieddesign (or programme). This is in clear contrast whenusinga machine do computations: withnormal procedure to one's object is thento have a clearmentalpicture the state of of in the machineat each moment the computation. This object can onlybe achievedwith struggle. The viewthat'themachine a can only do what we know how to orderit to do ',1 appears whichwe can strangein face of this. Most of the programmes put into the machinewill resultin its doingsomething that we cannot make sense of at all, or whichwe regardas completely random behaviour. Intelligentbehaviourpresumably consists in a departurefromthe completely disciplinedbehaviourinvolved in computation, a ratherslightone, whichdoes not but loops. give rise to randombehaviour,or to pointlessrepetitive our for Another important resultofpreparing machine its partin is the imitation and learning that game by a processof teaching ' 'human fallibility is likelyto be omittedin a rathernatural way, i.e. without special'coaching'. (The readershouldreconcile this withthe point of view on pp. 24, 25.) Processesthat are learnt do not producea hundredper cent. certainty result; of iftheydid theycould notbe unlearnt. wise to includea randomelement a learning It is probably in machine(see p. 438). A randomelement ratherusefulwhen is we are searching a solutionof some problem. Suppose for for instancewe wantedto finda number between50 and 200 which was equal to the square of the sum of its digits,we might start at 51 thentry52 and go on untilwe got a number that worked. we Alternatively might choosenumbers random at untilwe got a a good one. Thismethod theadvantagethatit is unnecessary has to keep track of the valaes that have been tried,but the disadvantagethat one may trythe same one twice,but this is not if very important there are several solutions. The systematic methodhas the disadvantagethat theremay be an enormous block withoutany solutionsin the regionwhichhas to be investigated first. Now the learningprocess may be regarded as a searchfora form behaviour of whichwillsatisfy teacher the (or some othercriterion). Since thereis probablya verylarge number of satisfactory solutions the random method seems to be better thanthe systematic. It shouldbe noticedthat it is used in the analogous process of evolution. But there the systematic methodis not possible. How could one keep track
1 Compare Lady Lovelace's statement 450),whichdoes not contain (p. the word'only'.

460

A. M. TURING:

COMPUTING

MACHINERY

AND INTELLIGENCE

of the different geneticalcombinations that had been tried,so as to avoid trying themagain ? We mayhopethat machines will eventually compete withmen in all purelyintellectual fields. But whichare the best ones to start with? Even this is a difficult decision. Many people thinkthat a very abstract activity,like the playing of chess, would be best. It can also be maintainedthat it is best to providethe machinewiththe best sense organsthat moneycan buy, and then teach it to understand and speak English. This process could followthe normalteaching of a child. Things would be pointed out and named, etc. Again I do not know what the rightansweris, but I thinkbothapproachesshouldbe tried. We can onlysee a shortdistanceahead, but we can see plenty therethat needs to be done.
BIBLIOGRAPHY Samuel Butler, Erewhon, London, 1865. Chapters 23, 24, 25, The Book of the Machines. Alonzo Church," An Unsolvable Problem of ElementaryNumber Theory ", American J. of Math., 58 (1936), 345-363. K. G6del, " Vber formalunentscheidbareSatze der Principia Mathematica und verwandter Systeme, I ", Monatsheftefuir Math. und Phys., (1931), 173-189. D. R. Hartree, Calculating Instrumentsand Machines, New York, 1949. S. C. Kleene, " General Recursive Functions of Natural Numbers ", American J. of Math., 57 (1935), 153-173 and 219-244. " G. Jefferson, The Mind of Mechanical Man ". Lister Oration for 1949. BritishMedical Journal, vol. i (1949), 1105-1121. Countess of Lovelace, ' Translator's notes to an article on Babbage's Analytical Engire', Scientific Memoirs (ed. by R. Taylor), vol. 3 (1842), 691-731. Philosophy,London, 1940. Bertrand Russell, History of Western A. M. Turing, " On Computable Numbers, with an Application to the Entscheidungsproblem", Proc. London Math. Soc. (2), 42 (1937), 230-265.

V7ictoria UniversittyManchester. of