{"id":290,"date":"2013-10-01T22:40:00","date_gmt":"2013-10-01T19:40:00","guid":{"rendered":"http:\/\/yarimada.gen.tr\/?p=290"},"modified":"2013-10-30T23:13:46","modified_gmt":"2013-10-30T21:13:46","slug":"ag-sadece-ag-degildir","status":"publish","type":"post","link":"https:\/\/yarimada.gen.tr\/?p=290","title":{"rendered":"A\u011f, Sadece A\u011f De\u011fildir"},"content":{"rendered":"

Richard Stallman’\u0131n GNU projesini duyurmas\u0131ndan bu yana otuz y\u0131l ge\u00e7ti. Stallman’\u0131n \u201c\u00f6zg\u00fcr bir i\u015fletim sistemi\u201d geli\u015ftirme fikri uzunca bir s\u00fcre bo\u015f bir hayal olarak g\u00f6r\u00fcld\u00fc. \u0130nsanlar, metala\u015fman\u0131n \u00f6n\u00fcne ge\u00e7ilemeyece\u011fini d\u00fc\u015f\u00fcnd\u00fcler. Art\u0131k bug\u00fcn cep telefonlar\u0131ndan web sunucular\u0131na, masa\u00fcst\u00fc bilgisayarlardan CERN’e kadar her yerde \u00f6zg\u00fcr yaz\u0131l\u0131m var. Ancak \u0130nternet olmasayd\u0131 \u00f6zg\u00fcr yaz\u0131l\u0131m belki \u00fcniversitelerde ve ara\u015ft\u0131rma laboratuvarlar\u0131nda olurdu ama evlerimizde ve telefonlar\u0131m\u0131zda olamazd\u0131. Linus Torvalds, Linux \u00e7ekirde\u011fini geli\u015ftirmek i\u00e7in \u0130nternet \u00fczerinden di\u011fer insanlardan yard\u0131m isterken eyleminin kazara bir devrime yol a\u00e7aca\u011f\u0131n\u0131 bilemezdi. Sadece biraz e\u011flenmek istemi\u015fti!<\/p>\n

\u00dcretimde \u00f6zg\u00fcrl\u00fck, t\u00fcketimde e\u015fitlik olarak tan\u0131mlayabilece\u011fimiz \u00f6zg\u00fcr yaz\u0131l\u0131m, \u0130nternet’in geli\u015fmesi ve yayg\u0131nla\u015fmas\u0131yla beraber \u00f6zg\u00fcrl\u00fc\u011f\u00fc de toplumun geneline yayabildi. Bu s\u00fcre\u00e7te, insanlar\u0131n a\u011flar \u00fczerindeki kolektif eylemi ve \u00fcretimi, bilim ve teknoloji merakl\u0131lar\u0131 d\u0131\u015f\u0131nda pek ilgi \u00e7ekmedi. Fakat son y\u0131llarda \u0130nternet’teki sosyal a\u011flar\u0131n yayg\u0131nla\u015fmas\u0131, kamuoyunun a\u011flara olan ilgisini de artt\u0131rd\u0131. Oysa \u0130nternet d\u0131\u015f\u0131nda da a\u011flar hep vard\u0131, sadece biz g\u00f6remiyorduk. A\u011flar\u0131 fark etti\u011fimizde ise daha \u00f6nce fark edemedi\u011fimiz olgular\u0131 g\u00f6rmeye ba\u015flad\u0131k. A\u011f biliminin tan\u0131nm\u0131\u015f isimlerinden Barab\u00e1si (2010), b\u00fct\u00fcn\u00fc anlamak i\u00e7in \u00f6nce bile\u015fenlerinin de\u015fifre edilmesi olarak adland\u0131rd\u0131\u011f\u0131 indirgemecili\u011fin 20. yy bilimsel ara\u015ft\u0131rmalar\u0131n\u0131n \u00e7o\u011funun itici g\u00fcc\u00fc oldu\u011funu, fakat bu yakla\u015f\u0131m\u0131n b\u00fct\u00fcn\u00fc g\u00f6rmeyi de zorla\u015ft\u0131rd\u0131\u011f\u0131n\u0131 belirtiyor. Barab\u00e1si’ye (2010) g\u00f6re bu yakla\u015f\u0131m, do\u011fadaki karma\u015f\u0131kl\u0131\u011f\u0131n anla\u015f\u0131lmas\u0131n\u0131 engelliyor. Ya\u015fam\u0131 bir a\u011f olarak kavramsalla\u015ft\u0131rmak olgular\u0131 ili\u015fkisel bir b\u00fct\u00fcn olarak g\u00f6rmemizi sa\u011flayabilir.<\/p>\n

Bu yaz\u0131da, \u0130nternet’teki sosyal a\u011flar yerine genel olarak a\u011f kavram\u0131, a\u011f bilimi ve a\u011f \u00e7al\u0131\u015fmalar\u0131n\u0131n tarihsel geli\u015fimi anlat\u0131lacak. Sosyal a\u011flar, a\u011flar\u0131n sadece bir t\u00fcr\u00fc. Kara yollar\u0131, hava yollar\u0131, telefon, elektrik \u015febekeleri, molek\u00fcller, sinir h\u00fccreleri… Hepsi birer a\u011fd\u0131r. B\u00fcy\u00fck matematik\u00e7i Leonhard Euler’in g\u00f6sterdi\u011fi gibi temel nokta, bizim olgulara nas\u0131l bakt\u0131\u011f\u0131m\u0131zla ilgilidir.<\/p>\n

Lewis (2008), a\u011f biliminin geli\u015fimini \u00fc\u00e7 d\u00f6neme ay\u0131r\u0131r:<\/p>\n

    \n
  1. A\u011f \u00d6ncesi D\u00f6nem (1736-1966)<\/li>\n
  2. Ara d\u00f6nem (1967-1998)<\/li>\n
  3. Modern D\u00f6nem (1998-)<\/li>\n<\/ol>\n

    A\u011f \u00f6ncesi d\u00f6nem, dizge (graph) teorisinin temellerinin at\u0131ld\u0131\u011f\u0131 ve \u00e7al\u0131\u015fmalar\u0131n daha \u00e7ok matematik\u00e7iler taraf\u0131ndan ger\u00e7ekle\u015ftirildi\u011fi bir d\u00f6nemdir. Bu d\u00f6nemde, a\u011f biliminin olu\u015fum s\u00fcrecini ve ilk teorilerin ortaya \u00e7\u0131k\u0131\u015f\u0131n\u0131 g\u00f6r\u00fcr\u00fcz.<\/p>\n

    Ara d\u00f6nemde, a\u011f \u00e7al\u0131\u015fmalar\u0131 matematik d\u0131\u015f\u0131ndaki alanlara y\u00f6nelir. Daha do\u011frusu, farkl\u0131 disiplinleri bir araya getirir. Milgram’\u0131n k\u00fc\u00e7\u00fck d\u00fcnyalar deneyi, a\u011flar\u0131 sosyal bilimlere sokar. Daha sonra Watts ve Strogatz, Milgram’\u0131n bulgular\u0131n\u0131 fizik ve biyoloji alanlar\u0131na ta\u015f\u0131r.<\/p>\n

    Modern d\u00f6nem, \u0130nternet’in yayg\u0131nla\u015fmas\u0131 ve bili\u015fim teknolojilerindeki geli\u015fmeler ile ba\u011flant\u0131l\u0131d\u0131r. Bu geli\u015fmeler, a\u011f \u00e7al\u0131\u015fmalar\u0131n\u0131 nicel ve nitel olarak etkiler. G\u00fcn\u00fcm\u00fczde a\u011f bilimi, bir \u00e7ok disiplinle etkile\u015fim halinde olan son derece canl\u0131 bir \u00e7al\u0131\u015fma alan\u0131d\u0131r.<\/p>\n

    1- A\u011f \u00d6ncesi D\u00f6nem (1736-1966) : K\u00f6nigsberg’in Yedi K\u00f6pr\u00fcs\u00fc<\/h2>\n

    K\u00f6nigsberg’in 18. yy’daki haritas\u0131 a\u015fa\u011f\u0131daki gibidir:<\/p>\n

    \"18.<\/a><\/p>\n

    Kent, nehirlerle ayr\u0131lmaktad\u0131r ve bu nehirler \u00fczerinde yedi k\u00f6pr\u00fc vard\u0131r. Bilmecelere merakl\u0131 olan K\u00f6nigsberg halk\u0131n\u0131n akl\u0131nda \u015fu soru vard\u0131r: Her k\u00f6pr\u00fcden yaln\u0131z bir kez ge\u00e7ilerek yedi k\u00f6pr\u00fcn\u00fcn t\u00fcm\u00fcnden ge\u00e7ilebilir mi?<\/em><\/p>\n

    Euler i\u015fe soruyu basitle\u015ftirerek ba\u015flar. Normal bir insan\u0131n sadece k\u00f6pr\u00fc ve nehir g\u00f6rd\u00fc\u011f\u00fc yukar\u0131daki haritay\u0131 a\u015fa\u011f\u0131daki gibi basitle\u015ftirir ve sorunu g\u00fcn\u00fcm\u00fczde \u00e7izge (graph) olarak adland\u0131r\u0131lan bir \u015fekle sokar (bkz. http:\/\/tr.wikipedia.org\/wiki\/K\u00f6nigsberg’in_yedi_k\u00f6pr\u00fcs\u00fc<\/a>\u200e):<\/p>\n\n\n\n
    \"Konigsberg2\"<\/a><\/td>\n\"graph\"<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    Kara par\u00e7alar\u0131n\u0131 d\u00fc\u011f\u00fcm<\/strong> (A, B, C, D), k\u00f6pr\u00fcleri ba\u011flant\u0131<\/strong>, d\u00fc\u011f\u00fcme ba\u011fl\u0131 olan ba\u011flant\u0131 say\u0131lar\u0131n\u0131 d\u00fc\u011f\u00fcm derecesi<\/strong> olarak adland\u0131r\u0131rsak bilmece \u015f\u00f6yle olur: Herhangi bir d\u00fc\u011f\u00fcmden ba\u015flayarak, her ba\u011flant\u0131dan yaln\u0131zca bir kez ge\u00e7erek t\u00fcm ba\u011flant\u0131lar\u0131 dola\u015fabilir miyiz?<\/p>\n

    Euler, bunun m\u00fcmk\u00fcn olmad\u0131\u011f\u0131n\u0131 ispatlar ve \u015f\u00f6yle der: \u201cbunun i\u00e7in tek dereceli d\u00fc\u011f\u00fcmlerinin say\u0131s\u0131, e\u011fer varsa, iki olmal\u0131d\u0131r. Tek dereceli d\u00fc\u011f\u00fcmler, dola\u015fman\u0131n ba\u015flang\u0131\u00e7 ve biti\u015f d\u00fc\u011f\u00fcmleri olacakt\u0131r. \u00c7izgede b\u00f6yle d\u00fc\u011f\u00fcmler yoksa dola\u015fmaya herhangi bir d\u00fc\u011f\u00fcmden ba\u015flanabilir.\u201d<\/p>\n

    \u00c7\u00fcnk\u00fc bir d\u00fc\u011f\u00fcm ba\u015flang\u0131\u00e7 ya da biti\u015f d\u00fc\u011f\u00fcm\u00fc de\u011filse, o d\u00fc\u011f\u00fcme gelen ki\u015finin d\u00fc\u011f\u00fcm\u00fc terk etmesi gerekir. \u00d6rne\u011fin, 3. k\u00f6pr\u00fcy\u00fc kullanarak A’dan D’ye geldiniz ve 7. k\u00f6pr\u00fcden \u00e7\u0131kt\u0131n\u0131z. E\u011fer D yolun sonu de\u011filse 4. k\u00f6pr\u00fcden ge\u00e7tikten sonra bir \u00e7\u0131k\u0131\u015f\u0131n\u0131z\u0131n olmas\u0131 gerekir. Varsayal\u0131m ki D ba\u015flang\u0131\u00e7 noktas\u0131. Ama di\u011fer k\u00f6pr\u00fcler de tek dereceli oldu\u011fundan b\u00f6yle bir gezinti yapmak olanakl\u0131 de\u011fildir.<\/p>\n

    A ba\u015flang\u0131\u00e7, C biti\u015f noktam\u0131z olsun. B ve D aras\u0131na, 7. k\u00f6pr\u00fcye ek olarak 8. k\u00f6pr\u00fcy\u00fc yerle\u015ftirirsek arzulad\u0131\u011f\u0131m\u0131z gezintiyi ger\u00e7ekle\u015ftirebiliriz!<\/p>\n

    G\u00f6r\u00fcld\u00fc\u011f\u00fc gibi a\u011f, d\u00fc\u011f\u00fcmlerden ve d\u00fc\u011f\u00fcmler aras\u0131ndaki ba\u011flant\u0131lardan olu\u015fan bir ili\u015fkiler k\u00fcmesidir. A\u011flar, hayat\u0131n i\u00e7indedir ve ke\u015ffedilmeyi beklerler. Bili\u015fimciler i\u00e7i a\u011f y\u00f6nlendiriciler (router) ve anahtarlar (switch), sosyologlar i\u00e7in insanlar aras\u0131ndaki etkile\u015fim, biyologlar i\u00e7in salg\u0131nlar\u0131n ve vir\u00fcslerin yay\u0131lmas\u0131, iktisat\u00e7\u0131lar i\u00e7in yeniliklerin ve icatlar\u0131n nas\u0131l yay\u0131ld\u0131\u011f\u0131d\u0131r…<\/p>\n

    Peki a\u011flar\u0131 matematik\u00e7ilerin \u00e7izgesinden ay\u0131ran nedir?<\/p>\n

    NRC’ye g\u00f6re a\u011f, g\u00f6zlemlenebilir bir ger\u00e7ekli\u011fin temsilidir, ger\u00e7ekli\u011fin kendisi de\u011fildir (National Research Council, 2005). Dolay\u0131s\u0131yla a\u011f bir soyutlamad\u0131r. Ama \u00e7izgeler gibi salt matematiksel ili\u015fkiler olmay\u0131p ger\u00e7ekle ili\u015fkilidirler. A\u011f bilimi ise d\u00fc\u011f\u00fcmlerden ve ba\u011flant\u0131lardan olu\u015fan a\u011f\u0131n yap\u0131sal ve davran\u0131\u015fsal \u00f6zelliklerini inceler. Son y\u0131llarda \u00f6zellikle a\u011flar\u0131n davran\u0131\u015fsal \u00f6zellikleri \u00f6nem kazanm\u0131\u015ft\u0131r. Farkl\u0131 disiplinlerden bilimciler, a\u011flardaki de\u011fi\u015fim ve bu de\u011fi\u015fimin nedenleri konusunda \u00e7al\u0131\u015fmalar yapmaktad\u0131r.<\/p>\n

    Tekrar a\u011flar\u0131n ilk d\u00f6nemine d\u00f6nersek…<\/p>\n

    Euler’in \u00e7izge teorisinin temellerini atmas\u0131ndan sonra matematik\u00e7iler \u00e7izgeler \u00fczerine \u00e7al\u0131\u015fmaya ba\u015flarlar. Bu \u00e7al\u0131\u015fmalar, bilgisayar biliminde ve elektrik m\u00fchendisli\u011finde \u00f6nemli bir yere sahiptir.<\/p>\n

    20. yy’n\u0131n ortalar\u0131nda yine iki \u00fcnl\u00fc matematik\u00e7i Erd\u0151s ve R\u00e9nyi a\u011flar\u0131 tart\u0131\u015fmaktad\u0131r (Barab\u00e1si, 2010): A\u011flar\u0131n g\u00f6r\u00fcn\u00fc\u015flerini ve yap\u0131lar\u0131n\u0131 belirleyen yasalar nelerdir?<\/p>\n

    Tabi a\u011flara matematik\u00e7i gibi, d\u00fc\u011f\u00fcmlerin ve ba\u011flant\u0131lar\u0131n ger\u00e7ek hayattaki somut varl\u0131klar\u0131yla ilgilenmeksizin yakla\u015f\u0131rlar. Erd\u0151s ve R\u00e9nyi i\u00e7in toplumsal a\u011flarla telefon a\u011flar\u0131 aras\u0131nda bir fark yoktur: B\u00fct\u00fcn farkl\u0131 sistemler i\u00e7in genel bir a\u011f yasas\u0131 olu\u015fturulabilir mi?<\/p>\n

    Erd\u0151s ve R\u00e9nyi, \u00f6nc\u00fcllerinin tersine d\u00fczenli \u00e7izgeler yerine belirsizli\u011fin ge\u00e7erli oldu\u011fu a\u011flar\u0131 incelemeyi tercih ederler. Bunun i\u00e7in de d\u00fc\u011f\u00fcmlerin rastgele ba\u011fland\u0131\u011f\u0131 bir a\u011f olu\u015ftururlar. Algoritmalar\u0131 do\u011frultusunda n tane d\u00fc\u011f\u00fcm aras\u0131ndan rastgele iki tanesini se\u00e7erek zar atarlar. Alt\u0131 \u00e7\u0131kt\u0131\u011f\u0131nda d\u00fc\u011f\u00fcmleri ba\u011flarlar ve daha sonra rastgele iki d\u00fc\u011f\u00fcm daha se\u00e7erler, ayn\u0131 i\u015flemi m tane ba\u011flant\u0131 olu\u015fana kadar tekrarlarlar. Algoritman\u0131n uygulanmas\u0131 sonucunda, ba\u011flant\u0131 say\u0131s\u0131n\u0131n artmas\u0131yla, t\u00fcm d\u00fc\u011f\u00fcmlerin kat\u0131ld\u0131\u011f\u0131 dev bir k\u00fcme olu\u015fmaya ba\u015flar ve kritik bir e\u015fikte a\u011f ili\u015fkilerinin d\u0131\u015f\u0131nda kalan d\u00fc\u011f\u00fcm say\u0131s\u0131 da \u00fcsl\u00fc olarak azal\u0131r.<\/p>\n

    Erd\u0151s ve R\u00e9nyi’nin rastgele bir a\u011f olu\u015fturmak i\u00e7in kulland\u0131\u011f\u0131 algoritma kusursuz de\u011fildir. Bazen ayr\u0131k d\u00fc\u011f\u00fcmler ya da m\u00fckerrer ba\u011flant\u0131lar olu\u015fturabilmektedir. Buna ra\u011fmen \u00f6\u011freticidir.<\/p>\n

    Rastgele olu\u015fturulan bu a\u011f son derece demokratiktir. D\u00fc\u011f\u00fcmlerin sahip oldu\u011fu ba\u011flant\u0131 say\u0131lar\u0131 birbirine yak\u0131nd\u0131r. Ortalama e\u011filimden sapan \u00e7ok az d\u00fc\u011f\u00fcm vard\u0131r. Ger\u00e7ek hayattaki a\u011flarla \u00e7eli\u015fen bir durumdur ve a\u011flar\u0131n geli\u015fig\u00fczel olu\u015fturulmas\u0131n\u0131n do\u011fal sonucudur. Erd\u0151s ve R\u00e9nyi, karma\u015f\u0131kl\u0131kla geli\u015fig\u00fczelli\u011fi bir tutmak gibi bir hata yapm\u0131\u015flard\u0131r. Asl\u0131nda Erd\u0151s ve R\u00e9nyi, genel ge\u00e7er bir a\u011f teorisi kurmaya \u00e7al\u0131\u015fmak yerine sadece a\u011flar\u0131n matematiksel g\u00fczelli\u011fine kendilerini kapt\u0131rm\u0131\u015flard\u0131r (a.g.e.).<\/p>\n

    2- Ara D\u00f6nem (1967-1998) : Alt\u0131 Ad\u0131m Kural\u0131 ve K\u00fc\u00e7\u00fck D\u00fcnyalar<\/h2>\n

    Ba\u015fka \u00fclkelerde nas\u0131ld\u0131r bilmiyorum ama \u00fclkemizde \u201caraya adam koymak\u201d gibi hayli yayg\u0131n bir deyimimiz var. Yetkili kimselere eri\u015fmek i\u00e7in tan\u0131d\u0131klar\u0131m\u0131z\u0131 kullan\u0131yoruz. \u0130\u015f bulmak i\u00e7in bir tan\u0131d\u0131k buluyoruz. Baz\u0131 yerlerde adam bulamazsan\u0131z i\u015fleriniz ya \u00e7ok yava\u015f ilerliyor ya da hi\u00e7 ilerlemiyor. Tan\u0131d\u0131klar, olmaz\u0131 olur yap\u0131yorlar. D\u00fcnya k\u00fc\u00e7\u00fckt\u00fcr. Bizim olmasa bile bir arkada\u015f\u0131m\u0131z\u0131n tan\u0131d\u0131\u011f\u0131 i\u015fimizi halledebiliyor. Peki tan\u0131d\u0131klar\u0131m\u0131z\u0131n menzili ne kadar? \u00d6rne\u011fin, Obama’ya tan\u0131d\u0131klarla eri\u015febilir misiniz? Ka\u00e7 ad\u0131mda eri\u015firsiniz?<\/p>\n

    Macar yazar Frigyes Karinthy 1929’da bir \u00f6yk\u00fcs\u00fcnde \u015f\u00f6yle yazar: \u201cD\u00fcnyan\u0131n bir bu\u00e7uk milyar sakini aras\u0131ndan herhangi birinin ad\u0131n\u0131 vermemizi istedi ve se\u00e7ilen bu ki\u015fiyle biri \u015fahsen tan\u0131d\u0131\u011f\u0131 olmak \u00fczere, en fazla be\u015f tan\u0131d\u0131k arac\u0131l\u0131\u011f\u0131yla ba\u011flant\u0131 kurabilece\u011fi yolunda bahse girdi.\u201d<\/p>\n

    1967’de sosyolog Stanley Milgram bu iddiay\u0131 ara\u015ft\u0131r\u0131r: Herhangi iki ki\u015fiyi birbirine ba\u011flamak i\u00e7in ka\u00e7 tan\u0131d\u0131\u011fa gerek vard\u0131r? Bu sorunun yan\u0131t\u0131n\u0131 bulmak i\u00e7in bir deney yapmaya karar verir ve rastgele iki ki\u015fi se\u00e7er: Massachusetts, Sharon’da bir lisans\u00fcst\u00fc ilahiyat \u00f6\u011frencisinin kar\u0131s\u0131 ile Boston’dan bir borsac\u0131. Sonu\u00e7 \u015fa\u015f\u0131rt\u0131c\u0131d\u0131r: D\u00fcnya k\u00fc\u00e7\u00fckt\u00fcr. Bir ki\u015fi, di\u011ferine en fazla alt\u0131 ad\u0131mda ula\u015fabilmektedir.<\/p>\n

    \"6adim\"<\/a><\/p>\n

    Granovetter’in 1973’te yay\u0131mlanan Zay\u0131f Ba\u011flar\u0131n G\u00fcc\u00fc<\/em> ba\u015fl\u0131kl\u0131 makalesi k\u00fc\u00e7\u00fck d\u00fcnyalar teorisini zay\u0131f ba\u011flar teorisiyle tamamlar (bkz. http:\/\/sociology.stanford.edu\/people\/mgranovetter\/documents\/granstrengthweakties.pdf<\/a>). Makale, 23000’den fazla al\u0131nt\u0131yla, sosyolojideki en etkili makalelerden biridir. Granovetter, toplumun hem kuvvetli hem de zay\u0131f ba\u011flarla bir birine ba\u011fland\u0131\u011f\u0131n\u0131 s\u00f6yler. \u0130nsanlar dostlar\u0131na kuvvetli ba\u011flarla, tan\u0131d\u0131klar\u0131na zay\u0131f ba\u011flarla ba\u011flanm\u0131\u015ft\u0131r. Granovetter’a g\u00f6re toplum, birbirini \u00e7ok iyi tan\u0131yan insan k\u00fcmelerinden olu\u015fur. Bu \u00e7ok kuvvetli ba\u011flar i\u00e7eren k\u00fcmeler, zay\u0131f ba\u011flarla d\u00fcnyan\u0131n geri kalan\u0131na ba\u011flan\u0131rlar. \u201cAraya koydu\u011fumuz adamlar\u201d \u00e7o\u011funlukla zay\u0131f ba\u011flard\u0131r. \u00c7\u00fcnk\u00fc iki insan aras\u0131ndaki ba\u011flar\u0131n kuvvetli olmas\u0131 nedeniyle arkada\u015f \u00e7evrelerinin \u00f6rt\u00fc\u015fme olas\u0131l\u0131\u011f\u0131 y\u00fcksektir. \u00d6rne\u011fin, i\u015f bulmada zay\u0131f ba\u011flar\u0131m\u0131z daha etkilidir. Farkl\u0131 a\u011flara ancak zay\u0131f ba\u011flarla (tan\u0131d\u0131klarla) eri\u015febiliriz.<\/p>\n

    Milgram’\u0131n deneyi saman alevi gibi parlar. 1970li y\u0131llarda k\u00fc\u00e7\u00fck d\u00fcnyalar kapsam\u0131nda \u00e7ok say\u0131da ara\u015ft\u0131rma yap\u0131l\u0131r. Sonras\u0131nda, yirmi y\u0131l sonra tekrar hat\u0131rlanmak \u00fczere, unutulur.<\/p>\n

    Milgram d\u0131\u015f\u0131nda ba\u015fka sosyologlar da a\u011flar\u0131n potansiyelini fark ederler. Bonacich (1972), insanlar aras\u0131ndaki etki a\u011f\u0131n\u0131 bir ba\u011flant\u0131 matrisine aktar\u0131r. A’n\u0131n B \u00fczerinde etkisi vard\u0131r, B’nin de C \u00fczerinde. Bu etkileme a\u011fda da\u011f\u0131lmakta m\u0131d\u0131r? Obama’ya eri\u015ftik, ama onun \u00fczerinde etkimiz ne olacakt\u0131r? \u201cObama sava\u015f istemiyoruz!\u201d dersek bizi dinler mi? Tan\u0131d\u0131klar\u0131n kendi tan\u0131d\u0131klar\u0131 \u00fczerindeki etkisi ve bunun yay\u0131lmas\u0131 nas\u0131ld\u0131r? Bonacich’in \u00f6nerdi\u011fi basit a\u011f modelinin g\u00fcn\u00fcm\u00fczde hi\u00e7 bir ge\u00e7erlili\u011fi yoktur. Fakat Bonacich’in \u00e7al\u0131\u015fmas\u0131 a\u011f biliminin potansiyelini g\u00f6stermi\u015ftir: Yeni bir \u00fcr\u00fcn\u00fc piyasaya s\u00fcrmek i\u00e7in nereden ba\u015flamak gerekir? Grup i\u00e7inde konsens\u00fcs nas\u0131l olu\u015fmaktad\u0131r? Salg\u0131nlar nas\u0131l yay\u0131lmaktad\u0131r?<\/p>\n

    3- Modern D\u00f6nem (1998 – )<\/h2>\n

    Milgram’\u0131n deneyi modern d\u00f6nemde tekrar g\u00fcndeme gelir. Watts ve Strogatz, a\u011flardaki k\u00fcmele\u015fmeyi g\u00f6zler \u00f6n\u00fcne sererler ve k\u00fc\u00e7\u00fck d\u00fcnya teorisini sosyal a\u011flardan fizik ve matemati\u011fe ta\u015f\u0131rlar (Lewis, 2008).<\/p>\n

    Barab\u00e1si (2010), benzer bir sonucun ba\u015fka a\u011flarda ge\u00e7erli olup olmad\u0131\u011f\u0131n\u0131 sorgular. Bunu ara\u015ft\u0131rman\u0131n en basit yolu web sayfalar\u0131 aras\u0131ndaki ba\u011flard\u0131r. Say\u0131sal bir a\u011f\u0131n bilgisayar programlar\u0131yla say\u0131salla\u015ft\u0131r\u0131l\u0131p analiz edilmesi hem bir \u00e7ok a\u011fa g\u00f6re daha kolayd\u0131r hem de daha kesin sonu\u00e7 verir. \u00d6rne\u011fin, insanlar aras\u0131ndaki sosyal a\u011flar\u0131 analiz ederken tan\u0131d\u0131\u011f\u0131n\u0131z ki\u015filerden hi\u00e7 ummad\u0131\u011f\u0131n\u0131z ba\u011flant\u0131lar \u00e7\u0131kabilece\u011fini fark edemeyebiliriz. Bir otob\u00fcs yolculu\u011funda yan yana gitti\u011finiz ve sohbet etti\u011finiz biri hi\u00e7 umamayaca\u011f\u0131n\u0131z ba\u011flant\u0131lar sa\u011flayabilir. Web sayfalar\u0131nda ise bu tarz bir belirsizlik yoktur. \u0130nsanlar t\u00fcm tan\u0131d\u0131klar\u0131n\u0131 listeleyemez ama sayfalar\u0131n i\u00e7erdi\u011fi t\u00fcm ba\u011flant\u0131lar g\u00f6r\u00fcl\u00fcr.<\/p>\n

    A\u011f analizi sonucunda u\u00e7suz bucaks\u0131z g\u00f6r\u00fcnen Web’in de k\u00fc\u00e7\u00fck bir d\u00fcnya oldu\u011fu \u00e7\u0131kar. \u0130ki web sayfas\u0131 aras\u0131ndaki ortalama ad\u0131m 19’dur: Arama motorlar\u0131na ba\u015fvurmadan, bir sayfadan di\u011ferine ortalama 19 ad\u0131mda gidebilirsiniz. \u0130nternet’in d\u00fc\u011f\u00fcmleri aras\u0131nda ise gerekli ad\u0131m 10’dur. Barab\u00e1si (2010), alt\u0131 ad\u0131m\u0131n \u201cmodern toplumun ve temas halinde olma \u0131srar\u0131m\u0131z\u0131n bir sonucu\u201d oldu\u011funu s\u00f6yler. \u0130nternet’teki sosyal a\u011flar ise alt\u0131 ad\u0131m\u0131 daha aza indirmi\u015ftir. Ancak i\u015fin ilginci, d\u00fc\u011f\u00fcmler aras\u0131nda yak\u0131nl\u0131k bir \u00e7ok a\u011f\u0131n ortak \u00f6zelli\u011fidir.<\/p>\n

    Bu d\u00f6nemin en b\u00fcy\u00fck yeniliklerinden biri Barab\u00e1si’nin \u00f6l\u00e7eksiz a\u011flar\u0131d\u0131r ve Erd\u0151s ve R\u00e9nyi’nin rastgele a\u011flar\u0131ndan bir kopu\u015f ger\u00e7ekle\u015fir. Erd\u0151s ve R\u00e9nyi’nin rastgele olu\u015fturdu\u011fu a\u011f\u0131 hat\u0131rlayal\u0131m. Rastgele bir a\u011f a\u015fa\u011f\u0131daki gibidir:<\/p>\n

    \"random\"<\/a>Ortalamadan fazla ya da az dereceli d\u00fc\u011f\u00fcmlerin say\u0131s\u0131 azd\u0131r, demokratik bir da\u011f\u0131l\u0131m\u0131 vard\u0131r. D\u00fc\u011f\u00fcmler ve sahip olduklar\u0131 ba\u011flant\u0131lar bir \u00e7an e\u011frisidir:<\/p>\n

    \"bellCurve\"<\/a><\/p>\n

    Erd\u0151s ve R\u00e9nyi’nin yakla\u015f\u0131m\u0131nda dikkat \u00e7ekmeyen iki varsay\u0131m vard\u0131r: Birincisi, a\u011f d\u00fc\u011f\u00fcmlerinin say\u0131s\u0131n\u0131n sabitli\u011fi ve statikli\u011fi; ikincisi ise d\u00fc\u011f\u00fcmler aras\u0131nda bir ayr\u0131m\u0131n olmamas\u0131d\u0131r (Barab\u00e1si, 2010).<\/p>\n

    Ger\u00e7ek hayattaki a\u011flar ise \u00e7o\u011funlukla rastgele de\u011fildir, Barab\u00e1si ve Bonabeau’nun (2003) ifade etti\u011fi gibi \u00f6l\u00e7eksizdirler:<\/p>\n

    \"scale-free\"<\/a><\/p>\n

    Ba\u011flant\u0131lar\u0131n da\u011f\u0131l\u0131m grafi\u011fi, kuvvet yasas\u0131 da\u011f\u0131l\u0131m\u0131na denk gelir. D\u00fc\u011f\u00fcmler aras\u0131nda bir e\u015fitlik yoktur. \u00c7o\u011fu d\u00fc\u011f\u00fcm son derece az ba\u011flant\u0131ya sahiptir ve t\u00fcm a\u011f\u0131 \u00e7ok say\u0131da ba\u011flant\u0131s\u0131 olan g\u00f6beklerin tuttu\u011fu g\u00f6r\u00fcl\u00fcr. Dolay\u0131s\u0131yla art\u0131k ortalama ba\u011flant\u0131ya sahip d\u00fc\u011f\u00fcmlerden olu\u015fan bir \u00e7an e\u011frisi yerine a\u015fa\u011f\u0131daki kuvvet yasas\u0131 da\u011f\u0131l\u0131m\u0131n\u0131 elde ederiz:<\/p>\n

    \"kuvvetYasasi\"<\/a><\/p>\n

    Neden baz\u0131 d\u00fc\u011f\u00fcmlerin derecesi daha y\u00fcksektir? Barab\u00e1si ve Bonabeau (2003), ger\u00e7ek hayatta a\u011flar\u0131n birbirine rastgele ba\u011flanmad\u0131\u011f\u0131n\u0131, d\u00fc\u011f\u00fcmlerin ba\u011flanmas\u0131n\u0131n bilin\u00e7li ya da bilin\u00e7siz bir tercihe dayand\u0131\u011f\u0131n\u0131 belirtir. A\u011flar\u0131n evrimi rastgele ve demokratik de\u011fildir. Tercihli ba\u011flanma yasas\u0131 taraf\u0131ndan y\u00f6nlendirilir. \u00d6rne\u011fin, web’de daha \u00e7ok ba\u011flant\u0131s\u0131 olan sayfalar daha \u00e7ok tan\u0131n\u0131r. Biz de sayfalarla ba\u011flant\u0131 kurarken \u00e7o\u011funlukla bunlar\u0131 tercih ederiz. Ya da ba\u011flant\u0131lar\u0131n\u0131z fazlaysa, A ki\u015fisiyle tan\u0131\u015fma (ba\u011flant\u0131 kurma) \u015fans\u0131n\u0131z, \u00e7evresi zay\u0131f birinden daha y\u00fcksektir. Arama motorlar\u0131, web’in hepsini taramaz. \u00c7\u0131kard\u0131\u011f\u0131 sonu\u00e7lar, \u00e7o\u011funlukla ba\u011flant\u0131 say\u0131s\u0131 fazla sayfalard\u0131r. Biz de bunlara eri\u015fti\u011fimiz i\u00e7in bunlara ba\u011flant\u0131 veririz ve bu sayfalar\u0131n derecesini daha da artt\u0131r\u0131r\u0131z: Ba\u011flant\u0131, ba\u011flant\u0131y\u0131 \u00e7eker.<\/p>\n

    A\u011f biliminin ilgilendi\u011fi bir ba\u015fka soru da a\u011flar\u0131n sald\u0131r\u0131lara kar\u015f\u0131 ne kadar dayan\u0131kl\u0131 oldu\u011fudur? Hem rastgele a\u011flarda hem de \u00f6l\u00e7eksiz a\u011flarda d\u00fc\u011f\u00fcmlerin b\u00fcy\u00fck bir b\u00f6l\u00fcm\u00fc kald\u0131r\u0131lsa bile a\u011fdaki d\u00fc\u011f\u00fcmlerin ba\u011fl\u0131l\u0131\u011f\u0131 belirli bir e\u015fik de\u011ferine ula\u015f\u0131ncaya kadar etkilenmez. Fakat \u00f6l\u00e7eksiz a\u011flarda, a\u011fdan kald\u0131r\u0131lacak d\u00fc\u011f\u00fcmleri rastgele se\u00e7mek yerine \u00f6nemli, farkl\u0131 \u00f6bekleri bir birine ba\u011flayan pop\u00fcler d\u00fc\u011f\u00fcmleri se\u00e7erseniz a\u011f\u0131 \u00e7\u00f6kertebilirsiniz. \u00d6rne\u011fin, \u0130nternet’te \u00e7ok say\u0131da sunucuda ar\u0131za olmas\u0131 ya da sunucular\u0131n sald\u0131r\u0131ya u\u011framas\u0131 durumunda bile \u0130nternet \u00e7\u00f6kmez. Ama birka\u00e7 y\u00fcz civar\u0131ndaki \u00f6nemli g\u00f6be\u011fe yap\u0131lan e\u015f zamanl\u0131 bir sald\u0131r\u0131 \u0130nternet’i alt \u00fcst edebilir. Teorik olarak b\u00f6yle olsa da t\u00fcm \u00f6nemli g\u00f6beklerin e\u015f zamanl\u0131 \u00e7\u00f6kmesi\/\u00e7\u00f6kertilmesi \u00e7ok k\u00fc\u00e7\u00fck bir olas\u0131l\u0131kt\u0131r (Barab\u00e1si, 2010).<\/p>\n

    Lewis (2008) a\u011f biliminin bu yeni d\u00f6nemi i\u00e7in baz\u0131 genellemelerde bulunman\u0131n hen\u00fcz erken olabilece\u011fini yazar. Ama Lewis’e g\u00f6re bu d\u00f6nemin a\u011flara bak\u0131\u015f\u0131 a\u015fa\u011f\u0131daki gibi \u00f6zetlenebilir:<\/p>\n