Logics of authentication, lieing and obscurity · Logics of AuthenticAtion, Lieing And obscurity...
Transcript of Logics of authentication, lieing and obscurity · Logics of AuthenticAtion, Lieing And obscurity...
Logics of authentication, lieing and obscurity
ORATIENOVEMBER 22Nd 2012
PROf.dR. duskO PAvlOvIc
6093 Oratieboekje Pavlovic.indd 1 13-12-12 12:15
PROf.dR. duskO PaVlOVic
6093 Oratieboekje Pavlovic.indd 2 13-12-12 12:15
3
NOVEMBER 22Nd 2012
iN auguRal lEctuRE giVEN tO MaRk tHE asuMPtiON Of tHE POsitiONas PROfEsOR Of
sEcuRity PROtOcOls
at tHE faculty Of ElEctRical ENgiNEERiNg, MatHEMatics aNd cOMPutER sciENcE at tHE uNiVERsity Of tWENtEON tHuRsday NOVEMBER 22Nd 2012By
PROf.dR. duskO PaVlOVic
logics of authentication, liEiNg and obscurity
6093 Oratieboekje Pavlovic.indd 3 13-12-12 12:15
4
Logics of AuthenticAtion,Lieing And obscurity
cOnTEnTs
1 Introduction...............................................................................5
2 Question.....................................................................................6
3 Protocols....................................................................................9
4 Authentication..........................................................................13
5 Impersonation..........................................................................18
6 Science.....................................................................................22
7 Obscurity,socialcontext,andtheendoflying........................25
8 Lieingbeyondlying...................................................................28
9 Thanks......................................................................................30
6093 Oratieboekje Pavlovic.indd 4 13-12-12 12:15
5
1 introduction
MijnheerdeRector,bestecollega’senstudenten,damesenheren.Het
isvoormeeenbijzonderplezieromhetambtvanhoogleraarjuisthier
inNederlandtemogenaanvaarden.
Hoewel ik tot nu toe bij de universiteiten vanmaar lieftst 7 landen
in dienst ben geweest, en zelfs nu alleenmaar een vijfde vanmijn
tijd inNederlandwerkzaamben,ende restelders—heb ik toch in
Nederland afgestudeerd en gepromoveerd, in Nederland een beetje
logicageleerd,en ikhebookweltweeNederlandsekinderen.Maar,
op ditmoment zijnmijnNederlandse kinderen in LosAngeles, and
myCaliforniansonLukaandmywifeAbbyarehere.Sowhile Istill
feelhomehere,IhopethatyouwillpermitmetocontinueinEnglish.
6093 Oratieboekje Pavlovic.indd 5 13-12-12 12:15
6
2 Question
Myinterestinlying,whichIwantedtotellyouabout,startedfarfrom
here,alongtimeago.IspentmychildhoodinSarajevo,Bosnia,which
usedtobeapartofasocialistcountrycalledYugoslavia.Inaddition
tofreeeducation,freehealthcare,andnottoomanyworkpressures,
socialism gave us an abundance of political life,whichwas carried
outinaparticularpoliticallanguage.Watchingmygrandfatherwatch
theTVnews,listeningtomymotherexplainherviewsofourcountry
to a curious police inspector, I started askingmyself: ”Can a lie be
recognizedbyitsgrammaticalform?”
SoIwentontostudyphilosophy,andquicklylearnedthataliecould
notberecognizedbyitsgrammaticalform.Therearedeepandskilful
philosophicalargumentsforalmostanything.Forinstance,therewas
aGreekphilosopherParmenides,apre-Socratic,whoarguedthatthere
wasnomovementintheUniverse.AndwhileParmenideswasexplaining
thatallmovementwasjustanillusion,amerewrinkleonanimmobile
Truth (αληθεια), his disciple Zeno paced up and down, mobilizedby the logical paradoxes arising from the concept of movement.
SoIwenttostudymathematics.Mathematicsis,ofcourse,alsonot
gearedtowardsrecognizinglies,buttowardsprovingthetruths:ifyou
canproveatheorem,andifitsassumptionsaretrue,thenthetheorem
mustbetrue.Orifyouknowthattheconclusionisfalse,thensomeofthe
assumptionsmustbefalse.Youmaynotbeabletodecidewhichparticular
assumptionisfalse,butyouknowthattheycannotbetruetogether.
Somathematicshelpsustorecognizesomeliesthroughtheirmutual
inconsistencies.Butthismaytakeanyamountoftime,andmoreover
manyliesareperfectlyconsistent,andcannotberecognizedinthisway.
Infact,realmathematiciansareusuallynotsointerestedinwhatistrue
andwhatnot;theyjustlookforhardproblemstosolve.Thedifference
6093 Oratieboekje Pavlovic.indd 6 13-12-12 12:15
7
betweenamathematicianandaprogrammeristhatthemathematician
seeks out hard problems,whereas the programmer avoids them
wheneverhecan,andonlyscalestheobstaclesthathecannotavoid.
SoIgraduallydroppedoutofmath,andbecameaprogrammer.That’s
howIgotstuckwithcomputers.Ifthereisonethingthathappened
duringmy lifetime, that is, of course, computers. Later I ended up
workinginmathsagain,butonlyasmuchasmathsdrivescomputers.
SoforawhileIforgotaboutlyingandobscurity,becauseitseemedthatfor
computers,itdoesn’tmattersomuch.Youcannotreallylietoacomputer.
Acomputerprogramcannotbe trueor false. Itcanbe fastorslow,
simpleorcomplicated,safeorunsafe,correctorincorrect;butitcannot
bealie.Sotheproblemdisappeared,andwealllivedhappilyeverafter.
ButthentheInternethappened,andtheWebontopofit,anditallgot
mixedup again: people and computers andphones and advertising
andpolitics.Itallstartedbrewinginthisbigpotofinformationonthis
globalnetworkofnetworks,whichbecame thenewComputer. The
Computerusedtobeabox,firstbigboxandthensmallerandsmaller,
butinanycaseyoucouldsitsomewherenexttothisthingandprogram
it.Withtheadventofnetworks,thecomputationreallyescapedfrom
thisbox;itspreadaroundtheworld,andtheComputerdisappeared:
thecomputations that I initiateonmy laptopdonothappenonmy
laptop,butsomewhereelse.Theycallothercomputationsthathappen
stillfurtherafield.Soforallpracticalpurposes,thesecomputationscan
bethoughtofashappeningeverywhere,whichisthesameasnowhere
inparticular.TheComputerdisappeared.Andcomputationbecamea
formoflife,spreadingviruses,swarmingintobotnets,robbingcyber
casinosandimpersonatingrichwidowsfromNigeria.Lyinginoneform
or another came to be one of the central technical problems of
computersecurity,andcomputersecuritycametobeoneofthebest
paid jobsevenforamoderatelynerdypersonlikeme.SoIendedin
computersecurity.
6093 Oratieboekje Pavlovic.indd 7 13-12-12 12:15
8
Tocutthestoryshort,afterawhileofdoingthatinCalifornia,Ibecame
ProfessorofSecurityProtocolshereinEnschede.Somepeoplebecome
professors ofMathematics, or of Computer Science; others become
professorsofPhilosophy,orofChemistry,orofQuantumMechanics.
AndIbecameaProfessorofSecurityProtocols.
Iexpect thatmanyofyouarenowtemptedtoask:”What kind of a
science is that — Security Protocols?”
Well,thatsciencehappenstobethislogicsofauthentication,andlying
andobscurity,thatIwanttotellyouabout.
6093 Oratieboekje Pavlovic.indd 8 13-12-12 12:15
9
3 ProtocoLs
To begin, let us look at a security protocol. Here is one that
mostofyouprobablyknow:theonlinebankingprotocol.
IstolethepicturefromourfriendsinNijmegen.Youusethisprotocol
ittopayyourbillsfromhome.Youprobablyusedtodothisbylogging
in into your bank’s web site with a password. This worked until a
coupleofyearsago.Sincethepasswordsareofteneasytoguess,this
turnedouttobeinsecure,sotheyimprovedtoprotocol,andnowin
additiontoyourpassword,youalsoneedyourbankcardtologininto
yourbank’swebsite.Thisiscalledtwo-factor authentication. Thetwo
factorsarethepasswordandthebankcard.Theauthenticationisthis
processwherebythebankmakessurethatthepaymentrequests,that
claimtobefromyou,arewreallycomingfromyou.Thisonlinebanking
protocolauthenticatesyouastheonlypersonwhohasyourbankcard,
andmoreoverknowsthecorrespondingPINandyourpassword.Inthe
timebeforecomputers,thetraditionalbankingprotocolsrequiredthat
youcometoyourbank,andauthenticateyourselfbyshowingaphoto
ID.YouwereauthenticatedastheonlypersonwhohadyourphotoID,
andwholookedlikeyou.
Mostofyouare,ofcourse,familiarwithmanyprotocolslikethis.But
letmeshowyouareallyoldone,goingbacksome6000years.
6093 Oratieboekje Pavlovic.indd 9 13-12-12 12:15
10
Ontherightyouseethe”bankcards”thatwereusedinUruq(Iraq)
inYoungNeolithic,around3700BC.Ontheleftyouseeapictureof
a”bankingprotocol”wheretheseearly”bankcards”wereprobably
used.AlicehasalambandBobhasbuiltasecurevault,perhapswith
multiple security levels, spacious enough to store both Bob’s and
Alice’sassets.This vault canbe thoughtof asanearlybank.When
Alicewould go for a vacation, shewould leave her sheep in Bob’s
bank.Butoneyear,whenshecamebackfromhervacation,andasked
forhersheep,Bobsaid:”Whichsheep?”Luckily,Bobwasonlyjoking,
andheeventuallygavebackAlice’ssheep.
But next year, when Alice needed to go for a vacation again, she
askedBobtogiveheraproofthatshelefthersheepwithhim.Bob
theninventeda security token:hemadealittleclaytablet,anddrewa
pictureofAlice’ssheeponit,orapictogramcorrespondingtoamark
on thesheep.Alice took thisclay token,andwent for thevacation.
Butwhenshecameback,shebroughtwithher3such tokens,and
requested3sheep.Thejokegotsomehowresolvedagain,ornot;but
inanycase,Boblearnedthelesson.
NextyearwhenAlicecametodeposithersheepinBob’sbankagain,
Bob invented a tamper resistant security token. This time, he didn’t
6093 Oratieboekje Pavlovic.indd 10 13-12-12 12:15
11
giveAlicejust3claytokenscorrespondingtohersheep.Hefirsttook
abiggerlumpofclay,madeahollowsphereofit,liketheonethatyou
seeonthepicture.Suchspheresarenowadayscalledbullæ,andkept
inmusea.TheoneonthepictureiffromLouvre.Bobusedthisbullaon
thepictureasanenvelope:heputthetokensinsideit,andthenbaked
it,sothatthetokenscanonlybeaccessedifthebullaisbroken.Now
Alicecannotaddtokensonherown,notwithoutbreakingthebulla.
Theprotocolsaysthat
• onlyBobisallowedtoopenthebulla;
• Alicemustsubmitthebullatogetthesheep;
• Bobmustreleasethesheepwhenthebullaissubmitted;
• anyonewhosubmitsthebullagetsthesheep.
It is interestingtonotethatAlicecould,andonoccasionsurelydid,
passonherbullawiththesheeptokenstoCarol,inexchangeforsome
of Carol’s goods, say an amphora full of olive oil,which could also
bestoredinBob’sbankandsecuredbyatokenenclosedinanother
bulla.Sobytradingtheirbullæ,AliceandCarolcouldtradetheirgoods
without ever moving them from Bob’s bank.Whenever needed, of
course,theycouldwithdrawtheirgoodsfromthebankbysubmittingto
Bobthebullæthattheyownatthatmoment.Carolcouldalsopassher
sheepbullatoDave,andsoon.Thisishowmarketeconomywasborn.
Obviously,asthemarketexpanded,morerobusttokenswereneeded.
Thetokensusedinthebullæevolvedintotheearliestformsofmoney,
andthe inscriptionsonthemledtotheearliestnumeralsystems,as
wellas toSumeriancuneiformscript,whichwasoneof theearliest
alphabets.Authenticationprotocolsthuspredatedliterature,science,
mathematics,andevenmoney.
Infact,theyevenpredatethecivilization,asthereisalotoflyingand
authenticationinnatureevenwithoutus.Hereisareallyoldprotocol.
6093 Oratieboekje Pavlovic.indd 11 13-12-12 12:15
12
Lookatthislittlebirdontheleft.Itisawren.Itsmainpurposeinlife
istoraiseitschicks.Buteveryonceinawhile, insteadofraisingits
ownchicks,itendsupraisinganotherbird’schicks.Ontherightyou
seethelittlewrenfeedingacuckoochick,whichgrewwithinacouple
ofweekstobelargerthanitsunwittingadoptiveparent.Thecuckoo
eggwassecretlylaidintowren’snest.Thepoorwrenwasunableto
tellitsownchicksfromthecuckoochickwhenthechickswerelittle.
Sothewrenfedallchicksthathatched in itsnest.Thecuckoosnot
onlydevelopedthetricktolayeggsinwren’snests;thecuckoochicks
developedthetricktopushwren’schicksoutofthenest,andtakeover.
In response to these tricks, thewrens developed an authentication
protocol.Recentlyithasbeenshownthatsomebabywrenslistento
theirmother’schirpswhilestillintheegg,andchirpliketheirmother
as soon as they hatch. Themother then only feeds the chicks that
chirpinherparticularway.Thecuckoosstilldidn’tdevelopacapability
to chirp like wrens, so they don’t pass this authentication test.
Nature is full of this: animals lie just like we do, and they try to
detect lies, just like we do. We didn’t invent any of that. And to
detect lies, they evolve authentication protocols, just like we do.
6093 Oratieboekje Pavlovic.indd 12 13-12-12 12:15
13
4 AuthenticAtion
Sohowdoauthenticationprotocolsuncoverlies?Firstofall,whatare
lies?
Tobegindefining lies, letussuppose thatwearegivena language,
say English, and we make statements that we all understand.
Suppose furthermore that we have a domain of interpretation for
thesestatements,saythisroom.Soeachstatement in itself iseither
trueorfalse,dependingonthestateofaffairsinthisroom.Forsome
statementswecandirectlyobservewhethertheyaretrueorfalse;for
otherswe cannot. For instance, I couldmake a statement that this
deskisflat,andthatIamanaccomplishedpianist.Fromyourposition,
youcanprobablyseethatthefirststatementistrue;butthesecond
statement might be a lie. It is difficult to establish its truth value,
becausemy capability as a pianist is not directly observable in this
room. So that is something that you might need to authenticate.
Authentication is theprocesswherebyweestablish theconnections
between the observable and the unobservable properties. More
precisely, an authentication protocol allows us to conclude that an
unobservablepropertyissatisfiedfromthefactthatacloselyrelated
propertyhasbeenobserved:
ObservableUnobservable
For instance, intheonlinebankingprotocol, thebankwasonlyable
toobservethemessagesreceivedfromthecustomer,whowashome
at his computer, and could not be directly observed. The protocol
was then designed to guarantee that, if the messages received
by the bank are in a certain form, then thesemessagesmust have
beensentby thecustomer;andmoreover that thecustomer indeed
intendedtorequestthetransactionsexactlyasinterpretedbythebank.
6093 Oratieboekje Pavlovic.indd 13 13-12-12 12:15
14
Buthowisthispossible?Whyisitsoundtodrawconclusionsabout
whatisunobservablefromwhatisobservable?
Suchconclusionsare,ofcourse,notsound ingeneral.But insome
cases,weknowthatsomethingthatweseemusthavebeencaused
bysomethingthatwedonotsee.Asuddenrippleonaperfectlycalm
water surface,without a breath ofwind in the air,must have been
causedbyamovementunderthesurface.Aparticularchirpthatonly
my babywrens can produce,must have been produced by one of
mybabywrens.Amessage thatcouldonlybecomputedbyAlice’s
smartcardenabledbyherPIN,musthavebeenproducedbyAlice.
Someobservablephenomenacanonlyariseundersomeunobservable
conditions.Ifweobservesuchphenomena,thenwecanbesurethat
the unobservable conditions must have been satisfied. That is the
essenceofauthentication.
Hereistheprimordialexampleofauthentication.
On the left, I observe that I think. In the comicbook language, the
statement that I see that I thinkmeans that I can seemy cloud of
thoughts.Moreover,Ialsoknowthatingeneral,acloudofthoughts
cannot ariseon itsown: itmustbeabove someone’shead. Itmust
6093 Oratieboekje Pavlovic.indd 14 13-12-12 12:15
15
beacloudofsomeone’s thoughts.And I alsoknow that ingeneral,
everyone can only observe their own thoughts. They cannot read
anyone else’s thoughts. So these thoughts that I observe must be
someone’sthoughts,andthatsomeonemustbeme.Inthisway,Ican
concludewhatIcannotobserve:sincemythoughtsarethere,Imust
alsobethere,underthiscloudofthoughtsthatIsee.Thisis,roughly,
how Descartes authenticated himself, and in the end reached the
well-knownconclusion:”Cogito, ergo sum”. Ithink,thereforeIexist.
JustastepbeyondRenéDescartes’authanticationofhimselfisBob’s
cryptographicauthenticationofAlice,presentedonthenextpicture.
AlthoughAlicemaybefarawayfromBob,andthereforeunobservable
tohim,ifBobreceivesamessageboundtoAlice’ssignature,thenhe
canbesure that themessagemusthaveoriginatedfromAlice.This
kind of reasoning, supported by cryptography, underlies the online
bankingprotocol,fromwhichwestarted.JustlikeRenésaidtohimself:
”Ithink,thereforeIexist”,BobsaidtoAlice:”Youdecrypt,therefore
youexist.”
6093 Oratieboekje Pavlovic.indd 15 13-12-12 12:16
16
In general, authentication thus derives something unobservable
fromsomethingobservablebytrackingdownaninvisibleactionthat
musthavecaused thevisible reaction, thatweobserve.This logical
derivationisoftenformalizedasinformationflowthroughachannel,
ormorepreciselythroughanauthentic channel.Anauthenticchannel
canbeimaginedasNewton’scradlebutsuchthatyoucanonlyseethe
reactionononeside,andyoucannotdirectlyobservetheactiononthe
otherside.Theactionisattheinputofthechannel,thereactionatthe
output,andyouonlyseetheoutput.Sowhenyouobservethereaction
attheoutputofthechannel,thenyoucanbesurethattheactionmust
havetakenplaceattheinput,althoughyoucannotobserveit.Thatis
authentication.
All examples of authentication that we have seen are based on
authentic channels: René’s introspection into his thinking is an
authenticchannel,AliceandBob’scryptographyprovidesanauthentic
channel,thebankcardsareanotherone,theneolithicclaybullæare
anotherone,babybirds’biometric chirping is yet another authentic
channel.Designingauthenticationprotocolsmainlyconsistsoffinding
andutilizingsuchchannels,andpushingsomeunobservableactionsat
theinputtogetthecorrespondingobservablereactionsontheoutput.
6093 Oratieboekje Pavlovic.indd 16 13-12-12 12:16
17
Authentication protocols allow us to prevent and detect many lies.
Social life isatissueformedbymyriadsofprotocols.Someofthem
are obvious,many are invisible. An inaugural lecture iswovenwith
many authentications. The Professor authenticates her academic
background. The University authenticates its academic roots.Many
authentications are embedded in the redundanciesof the language,
initsgrammatical,stylistic,orthographicconventions,uncoveringour
buriedsecrets,talents,andshortcomings.Mydressstyle,mannerisms,
demeanor are the social projections thrown into the space of
codes that that evolved in order to authenticatemy social position.
Protocols prevent us from lying that we are someone else. But
protocols alsoenableus toprove thatweare someoneelse—when
we manage to defeat a protocol and impersonate someone else.
6093 Oratieboekje Pavlovic.indd 17 13-12-12 12:16
18
5 imPersonAtion
Thereisaninterestinggeneralmethodthatletsyougetawaywithalie
bydefeatingalmostanyauthenticationprotocol.
Forinstance,supposethatyouwanttoauthenticatenotjustwhether
someonebelongs to thisor that socialgroup,butwhether theyare
humanatall.AlanTuringcameupwiththisauthenticationtaskalmost
immediately after he invented the concept of computer. Here is a
pictureofTuring’scomputer,theTuringMachine.
Thismachinewas proposed as an abstract,mathematicalmodel of
computer,butitisfairtosaythatalldigitalcomputersarereallybased
onthismodel.Thearchitectureofourcomputersgoesunderthename
von Neumann architecture,butJohnvonNeumannhimselfattributed
toAlanTuringthefundamentalideaofauniversalcomputer,capable
toperformallpossiblecomputationsbyexecutingdifferentprograms.
Havingprogrammed thisabstractcomputer foracoupleofyears in
theearly1930s,AlanTuringrealizedthat,inprinciple,suchcomputers
couldgeneratecorrect sentences, say inEnglish.Maybe theycould
be programmed to generatemeaningful sentences?Maybe even to
maintainaconversation?Soheraisedthequestionwhetheryoucould
specifyanauthenticationprotocol,areceipthowtorunaconversation
withsomeoneyoudon’t seeorhear, justbyexchangingmessages,
6093 Oratieboekje Pavlovic.indd 18 13-12-12 12:16
19
somethinglikeemail,orinstantmessaging.Nowcouldyourrunthis
conversationinsuchawaytobeabletoeventuallytellwhetheryou
areconversingwithacomputerorwithahuman?Whatwouldyou
askthem?SuchanauthenticationprotocolisknownasaTuring test.
Turing raised this question in 1950, in the articlewith the titleCan
machines think?Therehasbeenalotofphilosophyaboutthis,butin
themeantime,thisbecameapracticalquestion,andyouarenowadays
required to undergo a Turing test whenever you want a free web
service,e.g.toopenanewwebmailaccount.Themostfamiliarform
ofTuringtestareCAPTCHAs.
CAPTCHAsweredesignedatCarnegieMellonUniversityandthename
stands forCarnegieAutomatedPublic Turing test to tell Computers
andHumansApart.Theyare those littlegraphicfiles thatdisplay in
6093 Oratieboekje Pavlovic.indd 19 13-12-12 12:16
20
yourbrowseroftwirlycharacterstrings.Whenallworkswell,ahuman
caneasilytellwhichcharactersaredisplayedinthescreen,whereas
abot(whichisjustacomputerprogramcapableoftravellingonthe
Internet) cannot. So the humans pass the test, and the bots don’t.
CAPTCHAs were developed because the spammers needed a lot
of webmail addresses to send spam, so they wrote bots to open
thousandsof freewebmail addresses every day atHotmail, Yahoo!,
Gmailetc.Theseprovidersdidnot likethis,sotheyneededaTuring
testtorecognizethebots.
But as soon as the CAPTCHAs were developed and deployed, an
unknownspammerdevisedthefollowingmethodtodefeatanyTuring
test.
Hesetupafreepornsite,thesecondrectanglefromtherightonthe
pictureabove.Sinceitisfree,therearesomehumanvisitorsthereat
alltimesofdayandnight.Soabotgoestosetupawebmailaccount
withGmail,orYahoo.Thebot is thesecondrectangle fromthe left,
playedbyAgentSmithfromthemovieMatrix.NowGmailasksAgent
SmithtosolveaCAPTCHA.AgentSmithisjustapieceofcode,sohe
doesn’tseetheCAPTCHA.Butheforwardsthegraphicfiletotheporn
site,andthepornsitedisplaystheCAPTCHAtooneofthevisitors,and
askshimtosolveit,inexchangeforbeingabletocontinuebrowsing
for free. Thevisitor solvesheCAPTCHA, theporn site forwards the
solutiontoAgentSmith,thebotthensubmitsthesolutiontoGmail,
Gmailgiveshimanaccount.AgentSmithhasimpersonatedahuman.
6093 Oratieboekje Pavlovic.indd 20 13-12-12 12:16
21
Such impersonations are often called Man-in-the-Middle attacks,
becausetheattacker,whichisinthiscaseateamconsistingofAgent
Smithandthepornsite,sitsinthemiddlebetweentheauthenticator,
suchasGmail,andtheauthenticatedclient,inthiscaseahuman,and
justforwardsthemessagesbetweenthem.Itiseasytoseethatasimilar
attackcanbesetupinanynetworkconnectinghumansandcomputers.
In a network, it is thus impossible to distinguish between the two.
So we have seen how authentication protocols prevent lying by
findingandutilizingsomeauthenticchannels;andwehaveseenhow
impersonationattacksdefeatauthenticationprotocolsbyfindingand
utilizingsomeunintendedauthenticchannels.
6093 Oratieboekje Pavlovic.indd 21 13-12-12 12:16
22
6 science
Besides authentication protocols, there is another important realm
of human endeavor which mainly consists of finding and utilizing
authenticchannels.ThisrealmofhumanendeavoriscalledScience.
Experimentalscientists,likemywife,spendtheirdaysauthenticating
their results, and testing whether their theories about some
unobservable phenomena are confirmed by their observations.
Galileo’stelescopeandLeeuwenhoek’smicroscopearejustchannels
thatmake visible what used to be invisible. Rembrandt’s Professor
Tulpdissectscadaverstoobservesomeotherwiseunobservableparts
of human body. The Large Hadron Collider provides a channel to
subatomicparticles.Scienceadvances throughprotocoldesign,and
theexperimentprotocolsauthenticatelawsofnature.
Inasense,scienceisaspecialcaseauthentication.Inanothersense,
authentication is also a special case of science. Either way, both
scienceandauthenticationboildowntothesamething:testingsome
6093 Oratieboekje Pavlovic.indd 22 13-12-12 12:16
23
specifiedhypotheses,provingthosethataretrue,anddisprovingthose
thatarefalse.Asabyproduct,werecognizesomelies.
ButlistentowhatRichardFeynman,oneofthegreatestscientistsof
XXcentury,hastosayaboutthis.
Ifwehaveadefinitetheory,fromwhichwecancomputethe
consequenceswhichcanbecomparedwithexperiment,then
inprinciplewecanprovethattheorywrong.
But notice that we can never prove it right.
Supposethatyouinventatheory,calculatetheconsequences,
and discover every time that the consequences agreewith
theexperiment.Thetheoryisthenright?No,itissimplynot
provedwrong.Inthefutureyoucouldcomputeawiderrange
ofconsequences,therecouldbeawiderrangeofexperiments,
andyoumightthendiscoverthatthethingiswrong.
That iswhy laws likeNewton’s laws formotion of planets
last such a long time. He guessed the law of gravitation,
and it took several hundred years before the slight error in
themotion ofMercurywas observed.During all that time,
thetheoryhadnotbeenprovenwrong,andcouldbetaken
temporarilytoberight.—Weneveraredefinitelyright;we
canonlybesurewhenwearewrong.
This is fromFeynman’s lecturesaboutThe Character of the Physical
Law.
6093 Oratieboekje Pavlovic.indd 23 13-12-12 12:16
24
Sohereisthebestkeptsecretofscience:
• Scienceneverprovesthetruelawsofnature.
– Itnevermakesanydefiniteassertionsoftruth.
• Scienceonlydisprovessomefalsehypotheses.
– Itdetectslies.
Detecting lies is not a byproduct of science. It’s the only
thing it does! If you seek certainty, something to rely upon
—science is thewrongplace togo to!Certainty isonlyclaimedby
religion.
Religion says:Thisisthetruthabouttheworld.Youcanrelyuponit.
Art says: Thisisastoryabouttheworld.Youcanrelaxandplaywitit.
Science says:Thisatheoryabouttheworld.Youshouldn’trelyupon
itmorethanyouhaveto.Youshouldn’trelax,butworktoimproveit.
Science is an important foothold in anyone’squest for amethod to
recognizelies.Butinourscience-drivencivilization,itmaybegoodto
alsorememberwhatscienceisnot.ItisnotThePathtotheTruth.Itis
justaquestforeverbettertheories.Justamethodtorecognizelies.
6093 Oratieboekje Pavlovic.indd 24 13-12-12 12:16
25
7 obscurity, sociAL context, And the end of Lying
ButFeynmancontinues:
Another thing that Imust point out is that you cannot prove vague
theory wrong.
Hereisanothersecretofscience: Itgetsstuckatobscurity.Science
bustles with untestable theories: string theory, intelligent design,
many-worlds interpretation, universe before the Big Bang. . . Since
theycannotbedisproved,theseobscureanduntestabletheoriesoften
persist longer than their clear and testable counterparts. For similar
reasons,theobscure,prejudicialdiscoursepatternsalsopersistinmany
areasof social life, frompolitics to raising children.Neither science
orauthenticationdon’treallyhaveanythingtosayaboutthepolitical
languagesthatmygrandfatherwaswatchingontheTV.Socialismis
gone,buttheTVnewsarestillplayed.Theymayseemmoreadvanced,
but Idon’tseeanessentialdifference.Differentbutsimilar logicsof
lyingpersistacrossthewholewideworld,andacrosstheWorldWide
Web.
Therewasatime,inthelate1990s,whenyouwouldenterakeyword
tosearchtheWeb,saytheword”Kilimanjaro”,andthesearchengine
wouldreturnnothingbutporn.Thesearchenginesatthattimeonly
indexed the keywords, and the porn sites stuffed their index pages
withmillionsofkeywords,usuallyhiddeninasinglepixel.Inthisway,
theyspoofedthesearchengines,andattractedwebtraffic.Thiswas
calledspamdexing,orkeywordstuffing.Lyingwaseasyandblatanton
theWeb,andthispollutionobliteratedgenuinecontent,andledtothe
demiseofWeb1.0.
6093 Oratieboekje Pavlovic.indd 25 13-12-12 12:16
26
Googlemadeanendtoallthat,Web2.0flourished,andwearenow
abletofindontheWebalotoftheinformationthatwearelookingfor.
TheyspelledtheEndofLyingontheWeb.Itisilluminatingtorecallhow
thiswasachieved.Inadditiontoindexingthekeywords,modernsearch
enginesalsoindexthehyperlinksleadingfromonewebpagetoanother,
andgathermanyothermetadatawhich allow them tomeasure the
reputationofthewebpages.Thisapproachmadetheinformationflow
ontheWebmorereliable,andchangedthewayweliveourlivesinmany
areas.Theconceptscannowberecognizedasthehighlyconnected
communitiesofnetworknodes,whichpresumablyusethoseconcepts.
6093 Oratieboekje Pavlovic.indd 26 13-12-12 12:16
27
Itisdifficulttooverestimatetheimportanceofthisparadigmshift.It
says:
• Aliecannotberecognizedbyitsgrammaticalform.
• Aliarcanberecognizedbyhissocialcontext.
In a sense the methods of concept analysis, underlying the
modern information technologies, led us back from Fregean formal
logicsofXXcentury,backtoNietzscheanquestion:
Who speaks?—asthemethodtoendlying.
6093 Oratieboekje Pavlovic.indd 27 13-12-12 12:16
28
8 Lieing beyond Lying
The trouble is, though, that not only liars can be recognized by
their social context. All kinds of people can be recognized by their
social context: alcoholics, hackers, professors, football supporters,
people who like classical music, people who are planning to
refurbish their house. You can recognize them all by their social
context and behavior on the web, and you can advertise to them!
You can convince them to buy what they otherwise wouldn’t buy.
You can make them believe what they otherwise wouldn’t believe.
Intheworldwithtoomanyauthenticchannels,wemustlearntolieto
protectourselvesfrombeingliedto.Aliarcanberecognizedbythe
socialcontextonlyifthesocialcontextisnotalie.Otherwise,creating
alieiscomputationallyeasierthandetectingit.Thereliesthesolution.
6093 Oratieboekje Pavlovic.indd 28 13-12-12 12:16
29
Iuse”lieing”, themisspelledversionof theword”lying”, todenote
this movement towards rejecting the total network authentication.
Youmaydisliketheobscurityofthedistinctionbetweenlieingandlying.
Bothwordsarepronouncedthesame,andevenwhenyouseethem
written,theoddsarethatyoueitherwon’tnoticethedifference,orwill
assumethatthewritermadeamistake.Unlessyouknowthewriter
andtheirid-iosyncrasies.Theideaoflieingistouseidiosyncrasiesfor
watermarking.Theword”lieing”isanexampleofitself.
TheonlywaythatIcouldthinkoftoresistthetotalnetwork authentication
is to create multiple authentic identities. I call the general concept
creative privacy.Insteadoftryingtodamtheflowofprivatedata—a
futileeffort,itseems—,weshouldcreateanddisseminateafloodof
privatedata!
Watermarking can be used to distinguish what is authentic from
whatisnot.Yourfriends,thosewhoknowyoualready,willhavethe
informationneededtofindthewatermarks,theotherscannot.Hidden
in thetext, theword lieing isa toyexampleofsuchawatermark of
rejectionofthetotalnetworkauthentication.Onlythosewholookfor
itwillfindit.
ThisistheideaoflieingthatIwantedtotellyouabouttoday.
6093 Oratieboekje Pavlovic.indd 29 13-12-12 12:16
30
9 thAnks
In theend, I canonly try to thank thosewhohelpedme. It is truly
astonishinghow lucky Iwas tomeetallofyou.DirkvanDalenand
HenkBarendregtintroducedmetotheDutchSchoolofLogic,while
IekeMoerdijk andMartin Hyland tried to teachme some Category
Theory, as much as my lack of coordination and background
permitted. InMontreal,MikeBarr and JimLambek taughtme some
Mathematics, and back in London, Samson Abramsky taught me
someComputerScience. InPaloAlto,DougSmith taughtmesome
SoftwareEngineering,andCathyMeadowsandJohnMitchelltaught
mehowtothinkaboutsecurityprotocols,andhowtocollaborate.Bob
CoecketaughtmeabitaboutQuantumMechanics,andmychildren
Stefan,TemraandLukataughtmethatthereisnojusticeonEarth:I
wasateenagerfromhell,evenwhenIwas27,andtheyarelikeangels
(albeitoccasionallygrumpyangels),oratleastlikebutterflies.Andlast
andmostofall,mywifeAbbytaughtmethatitispossibletosettle,
andthatwecansurviveanything.
Butthen,ifyousquintalittleatthis,yourealizethatAbbydidn’treally
makemesettlebutleftherjobatStanfordandstartedmovingwithme
fromcountrytocountry.Inspiteofthebesteffortsofallofmyteachers,
thebestexpertsintheirrespectiveareasofscienceandcollaboration,
Ididn’t learnmuchaboutsecurity,oraboutQuantumMechanics,or
aboutSoftware,ComputerScience,Mathematics,CategoriesorLogic.
Ievencan’tspell.
We liveour lies. It is difficult to tell humans fromcomputers apart,
andit isevenhardertotellapartourliesfromourtruths.Butthisis
notabug,butafeatureofthesoftwareoflogics.Thisfeatureshould
be explored, and used, sincewe really need it.Moreover, the logic
of our stories is at least as interesting as the logic of our sciences
andmathematics.Justlikeouroldbeliefs,manyoftoday’sscientific
6093 Oratieboekje Pavlovic.indd 30 13-12-12 12:16
31
theories will be tomorrow’s religions, and some of today’s dreams
and fantasies will be realized. The distinction between science and
fantasy, between fact and fiction, between the lies and the truths
—isabitofa lie itself.Thedistinctionbetweenournightmaresand
daydreams on one side, and the reality on the other side, is itself
sometimesanightmare,sometimesadaydream,andonlysometimes
itisreal.DidI,likeChuangTzu,dreamthatIwasabutterfly,ordida
butterflydreamthatshewasaprofessor,andthatsheworeablack
toga, and gave an inaugural lecture in Enschede, on the obscure
topics of lieing? And how cartesianwas her introspectionwhereby
sheintheendreachedthewell-knownconclusion:”Ik heb gezegd”?
6093 Oratieboekje Pavlovic.indd 31 13-12-12 12:16
6093 Oratieboekje Pavlovic.indd 32 13-12-12 12:16