行政院國家科學委員會專題研究計畫 成果報告
開放科學運動的理念與挑戰
研究成果報告(精簡版)
計 畫 類 別 ： 個別型 計 畫 編 號 ： NSC 100-2410-H-004-020- 執 行 期 間 ： 100 年 08 月 01 日至 101 年 07 月 31 日 執 行 單 位 ： 國立政治大學法律科際整合研究所 計 畫 主 持 人 ： 李治安 計畫參與人員： 碩士級-專任助理人員：楊雅竹 報 告 附 件 ： 出席國際會議研究心得報告及發表論文 公 開 資 訊 ： 本計畫涉及專利或其他智慧財產權，1 年後可公開查詢中 華 民 國 101 年 10 月 31 日
中 文 摘 要 ： ｢開放科學｣並不是晚近才出現的概念，事實上，自二十世紀 以來，多數科學研究社群都受到開放科學的價值精神影響， 科學研究向來具有很強的開放性、社群性及相互批判性等社 會規範與價值。近年來，由於愈形密切的產學合作關係及政 府規範研發成果的法令，使得科學研究的成果漸趨商業化， 科學社群過去所強調開放與分享的社會規範與價值受到相當 大的挑戰，也因此使得開放科學的呼聲時有所聞，開放科學 的概念也又重新受到重視。究竟應如何解決或舒緩科學研究 過度封閉、排他或商業化之問題，以實踐開放科學的理想， 並維持開放科學的成果，涉及許多法律與政策問題，亟待學 界討論研究。 本研究計畫介紹近年來開放科學運動的推動背景及正當 性，其中著重在探討制度及法律上造成科學研究漸趨封閉的 原因。本研究亦探討相關推動開放科學之作為所可能面臨的 困難與挑戰，由法律、政策、科學研究的本質及科學社群運 作慣例等角度切入，對不同的開放科學推動策略作出可行性 及成功可能性之評估。此外，本研究亦希望發現國內科學社 群對於開放科學議題之態度，及國際上開放科學運動之對國 內科學社群之影響，以為未來制定相關政策時之參考。 中文關鍵詞： 開放科學、科學社群、專利法、產學合作、生醫研究、商品 化、公眾授權
英 文 摘 要 ： ＇Ｏpen science＇ is not a new idea to the scientific communities. The value of open science has rooted in most scientific communities since twentieth century. Science has been recognized as a field of openness, community, and criticism. Nonetheless, in recent years, science has been more and more commercialized and its openness has been partly eroded because of the increasing academic-industrial collaboration and government regulations. Therefore, there has been a call for ＇open science＇ from various scientific communities, especially biomedical ones. How to solve or alleviate the over-closed, over-exclusive, and over-commercialized problem in scientific research and maintain the openness of science are associated with a number of legal and policy issues, which is valuable for further research.
This project introduces the background and justification of the open science movement, especially the laws and policies that cause the increasing secrecy and competition with the
scientific communities. This project also analyzes the obstacles and challenges facing the open science movement via the lenses of law, policy, the nature of science, and the social norms underlying the
scientific communities. Moreover, this project assesses the feasibilities of varios open sceince strategies. Lastly, this project will observe the impact of open science movement on domestic
scientific communities, and provide some relevant information for future policymaking.
英文關鍵詞： open science, scientific community, patent, academic-industrial collaboration, biomedical research,
1 壹、前言 科學的創新，除了端賴於大量有價值資訊的生產外，亦與該等資訊能否被廣 泛的分享有關1。科學社群過去所風行的社會規範與價值是強調開放與分享，但 是近年來隨著學術期刊電子化及研究成果商業化的發展，科學研究的成果已經不 再是可供公眾或科學社群分享的資訊，也因此，開放科學(open science)的呼聲近 年來在科學界時有所聞。事實上，開放科學並非近年才有的概念，依據社會學家 Robert K. Merton的研究，｢開放科學｣係由以下四種概念所構成 2 1. 普遍性(universalism)：科學的發展應對所有的研究者公開。 ： 2. 公有性(communalism)：思想應廣泛被分享，而不應被任何人擁有。 3. 公正性(disinterestedness)：科學家應超越個人主觀。 4. 系統性的懷疑(organized skepticism)：所有的科學發現應接受同儕審查(peer review)。 本文之目的，即在於介紹近年來開放科學運動的推動背景及正當性，亦探討 相關推動作為所可能面臨的困難與挑戰。除了｢開放科學｣以外，亦有論者使用 Science Commons 來表達對於開放科學研發成果之主張與期待，其中創用 CC(Creative Commons)計畫甚至以 Science Commons（以下以 SC 簡稱此一分支 機構）為名，成立一分支機構推動一連串開放科學研究成果的計畫。因此，有時 相關文獻中提到 Science Commons 時，其指涉之對象不一定是創用 CC 計畫之前 述分支機構，而是代表一般開放科學理念之泛稱。然而 SC 所處理的問題，包括 學術期刊之開放近用及研究材料及資訊的分享流通，的確是近年來開放科學運動 所關注的核心主軸，此亦為本文所欲處理之主要對象。由於(1)學術期刊之開放 近用及(2)研究材料及研發成果的分享流通，對於科學社群來說，係不同的兩個 問題，因此處理的模式有相當差異，前者與著作權法相關，後者則與專利法較有 關，因此本文以下論述亦多以學術期刊及研究材料及資訊兩種不同的研究成果區 分討論開放科學運動或 Science Commons 之發展。 貳、Science Commons 的形成背景 一、學術期刊 近年來學術期刊訂閱價格高漲已成為學術界關心的一個重要問題，以美國 為例，自 1986 年至 2004 年之間，學術期刊的訂閱費用一共成長了 220%，而單
1 Robert Cook-Deegan, The Science Commons in Health Research: Structure, Function, and Value, 32 J.
TECH.TRANSFER 133, 135 (2007).
2 R
2 就科技及醫學的期刊而言，自 1982 年至 2002 年之間，其訂閱價格成長了超過 600%，其中化學研究的學術期刊訂閱價格更是成長了 752%，一些學術期刊的單 年訂閱費用，動輒超過兩萬美元以上。如此的趨勢已經使得許多小型圖書館、獨 立研究人員及開發中國家的研究機構無法負擔，但是此等價格調整是否純粹反映 了期刊的製作成本，抑或是出版商提升獲利的手段，其所造成的社會成本是否必 要，均是值得探討的問題。 在過去網際網路及數位科技尚未興盛時，學術期刊出版商主要藉由出售紙本 期刊作為主要獲利來源，但是隨著網際網路及數位科技的進步，期刊出版商開始 將期刊內容數位化，並藉由不同的數位權利管理(digital rights management, DRM) 技術限制使用者對於期刊內容的接近使用，期刊的定價方式也不再只是過去單純 的紙本售價或訂閱價，而是有一套更精密的數位定價系統，出版商可以文章下載 次數、系統使用人數或使用期間等方式作為定價基礎，向使用者或使用機構收取 費用。 藉由著作權及新發展的數位權利管理技術，出版商對於期刊內容的控制力 更甚於以往純粹紙本出版的時代，許多出版商甚至將研究機構所需要及不需要的 不同期刊組合成單一的授權包裹(license package)，要求研究機構照單全收，以收 取更高額的授權費用。這種趨勢造成學術論文使用者取得期刊內容的成本日益提 升，許多小型的圖書館及開發中國家的研究機構都無法負擔高額的整批出售 (bundle)費用。此外，整批出售手法的運用對於學術期刊出版的市場結構亦具有 關鍵影響，由於一般研究機構及圖書館訂閱經費有限，所以許多機構都會選擇能 涵蓋最多期刊或最重要期刊的包裹作為採購對象，此種選擇的結果將會使期刊市 場最終被少數出版商所把持，從而對市場競爭造成負面影響。 事實上，學術期刊的出版市場本來就不是一個完全競爭市場，超過半數的 科技及醫學學術期刊是由全球前二十大的出版商所出版，而某些期刊在專業領域 具有無可取代的學術聲望（如科學研究的 Nature 及 Science 等期刊），學者們多 競相爭取在該等期刊發表的機會，以提升學術聲望或作為獲得升等的有利條件， 學術機構也將該等期刊列為絕對必須的訂閱對象，相對而言，其他期刊甚難取代 這些一流期刊的地位，也因此，這些一流期刊的出版商能夠恣意提高期刊訂閱價 格，或搭配其他期刊一起整批出售。 相較於其他類型的出版市場而言，學術期刊出版市場的特色在於不同作者 之間及作者與讀者間的關係其實是更為密切的，許多學者身兼不同學術期刊的作 者、讀者及編輯委員等不同身分，他們花了相當多的工作時間及資源為學術期刊 撰搞、審稿，並閱讀期刊上其他學者的論著，但是他們的報酬並非由期刊出版商 所支付，而是由其所屬的研究機構給付薪資，就此而言，雇用學者的研究機構藉 由給付學者薪資及投稿費用、獎勵學者為學術期刊撰稿等行為，其實是某種程度 補貼了期刊出版商在學術文章產出上所應花費的成本。有趣的是，支付學者薪資
3 的研究機構通常也是學術期刊訂閱的最大宗客戶，為了使其所雇用的學者能做出 具有水準的研究，許多研究機構每年都支付給期刊出版商鉅額的訂閱費用。 基於上述分析，學術機構（或者是為其提供財務支持的政府、納稅人或贊 助人）其實對學術期刊的出版提供了雙重的財務支援，一方面支付學者薪資，使 其從事研究、為學術期刊提供內容，一方面再支付給期刊出版商高額的訂閱費用； 而出版商除了取得學術機構的雙重補助之外，還取得了期刊論文的著作權，依據 Gadd、Oppenheim、Probets(2003)對英、美兩國八十家學術期刊出版商提供給作 者的合約所作的研究，90%的出版商要求作者將著作權轉讓給出版商，另外 10% 的出版商中，則有一半會要求取得作者的專屬排他授權。但是，經濟上是否有必 要對出版商如此重複補助，卻頗有探討的餘地。既然學術期刊的內容都由學者無 償提供，則相較於其他出版事業而言，學術期刊出版商的內容生產成本應相對較 低，從而訂閱費用亦應維持在一定水準以下，然而，近年來學術期刊的價格走向 卻呈現相反的趨勢，有學者指出許多學術期刊的訂閱費用的總和早已遠超經營該 等期刊所需要的必要費用。 二、研究材料及成果 ｢開放科學｣並不是晚近才出現的概念，事實上，自二十世紀以來，在大學中 的科學研究社群都受到開放科學的價值精神影響3，在學術界的科學研究具有很 強的開放性、社群性及相互批判性等社會規範與價值4。然而，由於近年來愈形 密切的產學合作關係，科學研究的成果漸趨商業化，科學研究的開放性逐漸萎縮5； 而政府對於其所資助達成研發成果之態度，亦是造成開放科學縮減的原因之一， 舉例而言，為了使聯邦政府所資助發展的研究成果能獲得更廣泛的利用，並鼓勵 私人投資者能將科學發明商品化，美國於 1980 年代所通過的拜杜法案(Bayh-Dole Act)鼓勵科學家為研究成果尋求專利保護6。該法為大學及科學家們提供了相當 誘因，將其研發成果專利化與商品化，也因此，縱使是對於包括DNA序列、蛋 白質結構及疾病路徑(disease pathway)等基礎科學或一般研究工具，科學家們也 開始思考應如何就其取得專利權保護7。
3 James Stewart, Comment, The Academic-Industrial Complex: A Warning to Universities, 75 U.C OLO. L.REV.1011, 1026-30 (2004).
4 Cook-Deegan, supra note 1, at 136.
5 Megan Ristau Baca, iBrief, Barriers to Innovation: Intellectual Property Transaction Costs in Scientific Collaboration, 2006 DUKE L.&TECH.REV. 4, ¶4.
6 See e.g. Cook-Deegan, supra note 1, at 138; Arti K. Rai & Rebecca S. Eisenberg, Bayh-Dole Reform and the Progress of Biomedicine, 66 DUKE J.L.&CONTEMP.PROBS. 289, 290 (2003).
4 (一) 研究工具與材料移轉 過去生醫研發之上游階段所發展出來的研究工具(research tool)，常提供予下 游研究者之診斷或治療目的而無償利用8；但是，今日科學研究發展的現況卻非 如此，以基因序列為例，特定基因序列在被分離之時，可能立即取得專利權保護， 專利權人此時甚至對於該序列可成為何種商品還一無所知9，準此，上游研發專 利化的結果，不但增加了下游研發的成本，易減緩整體生醫之產業創新。許多研 究機構縱使不尋求專利權保護，也會限制研究材料及試劑的流通，避免許多可能 帶來商業價值的後續研究10。也因此，有學者用反共有悲劇(tragedy of the anticommons)描述生醫研發的財產權歸屬狀態11。凡此種種，均使得過去原屬於 開放科學的領域逐漸縮減。該等科學研究成果商業化的發展趨勢，亦使得科學社 群及學術研究機構間之關係發生質變，過去科學界習於協力研究，但今日卻更注 重研究進展的秘密性及競爭性12。 生醫研究機構常需使用其他研發機構所發展出的材料（特別是包括基因、細 胞株或其他生物產品等研究工具）作為研發基礎，而為了處理系爭材料所衍生之 智慧財產權問題，許多研究機構都以材料移轉契約(material transfer agreements, MTAs)規範系爭材料在何等範圍內、可供何種使用。為了控制下游研發者對系爭 材料之使用，原材料開發者通常會在材料移轉契約中要求系爭材料之下游使用者 將其研發成果之智慧財產權授權或移轉，或以其他方式控制下游使用者之研發成 果13；對材料授移轉者而言，材料移轉契約亦可為其提供相當保障，避免其在需 要材料時，被要求承諾不合理的條件，或因材料的延遲移轉，造成研究空窗期， 影響研究進度或研究補助申請14。
8 Michael Heller & Rebecca S. Eisenberg, Can Patents Deter Innovation? The Anticommons in Biomedical Research, 280 SCIENCE 698, 698 (1998).
9 Molly A. Holman & Stephen R. Munzer, Intellectual Property Rights in Genes and Gene Fragments,
85 IOWA L.REV.735, 739 (2000).
10 Rai & Eisenberg, supra note 6, at 291. 11 Heller & Eisenberg, supra note 8, at 699.
所謂｢反共有悲劇｣是指數人就單一珍貴資源， 擁有排除他人使用之權利， 因此沒有一個人就系 爭資源有絕對之決定權，進而導致係茲資源使用不足之不效率情況，國內文獻對此理論之介紹 ，參馮震宇，從司法院法學資料庫隱私權問題看政府資訊委外的發展與問題，月旦法學雜誌，1 54期，2008；李素華，基因及基因藥之專利法制發展趨勢，法學新論，4 期，2008。
12 See e.g. Baca, supra note 5, at ¶18, ¶ 21; Cook-Deegan, supra note 1, at 136. 13 Baca, supra note 5, at ¶23.
5 然而，研究發現，縱使在過去曾有合作關係的研究機構間，材料移轉契約條 款的談判都常花費許多時間及費用，該等談判亦會導致研究時程的拖延15。由於 材料移轉契約談費所耗費的可觀交易成本，許多科學家所幸就不將自己研發所得 之材料與其他科學家分享16。由此可見，不論是否簽訂材料移轉契約，現今科學 資訊的流通都受到相當的阻礙。 (二) 人類基因體計畫 引發開放科學風潮的另一原因是科學界為了發現人類基因體序列而引發的 相關爭議，1996 年成立的國際人類基因體定序聯盟（International Human Genome Sequencing Consortium）係由一群來自世界各地、以分享人類基因序列資訊為宗 旨的科學家所組成，該組織成員均依據「百慕達原則」(Bermuda Principles)，主 張相關研究成果應儘速置於公共領域(public domain)17。然而，在 90 年代末期， 某些公司開始探索將人類基因序列商業化之可能性，前述國際人類基因體定序聯 盟的一個成員成立了賽雷拉基因公司(Celera Genomics)，該公司搶先於國際人類 基因體定序聯盟之前，完成了第一組人類基因序列18，並大量就人類基因序列申 請並取得專利19。2001 年時，國際人類基因體定序聯盟及賽雷拉基因公司個別 發表類人類基因組織工作草圖，然而，前者卻主張後者係抄襲前者已公開發表之 資訊，後者卻否認之20，雙方對此等爭議最後並未有一致性的訴訟或仲裁處理， 但是卻使得開放科學的觀點受到更多重視。 (三) 小結 無論是科學材料移轉的限制及人類基因體研究成果的歸屬，都使得開放科學 的問題受到重視，而將專利權保護範圍擴及至所有研發成果，也相當程度的阻礙 了研究機構間及科學家間的合作研發機會，根據 2002 年所作的一項實證研究顯 示，基因科學家間的合作研發已較過去大幅減低，主要原因是科學家們因為彼此 15 Id. at ¶24-25. 16 Id. at ¶25.
17 Andres Guadamuz González, Open Science: Open Source Licensing in Scientific Research, 7 N.C.J.
L.&TECH. 321, 333-34 (2006).
18 John Sulston, Intellectual Property and the Human Genome, in G
LOBAL INTELLECTUAL PROPERTY RIGHTS:KNOWLEDGE, AND DEVELOPMENT 61, 64-65 (Peter Drahos & Ruth Mayne eds., 2002).
19 Guadamuz González, supra note 17, at 334-35.
20 關於雙方之爭議及立場，請參見R. Waterston, E. Lander & J. Sulston, On the Sequencing of the Human Genome, 99 PROC.NAT’L ACAD.SCI. 3712 (2002); E. Myers et al., On the Sequencing and
6 合作而造成的專利侵權風險21。此外，範圍過廣的專利範圍，對市場競爭亦有負 面影響。針對此等問題，開放科學的擁護者亦開始思考，究竟應如何實踐開放科 學的理想，並維持開放開學的成果。 參、推動 Science Commons 之具體作為 一、學術期刊 針對前述學術期刊之近用問題，國際間已形成何謂「開放近用」運動，根據 目前最為人所熟知的布達佩斯開放近用提議(Budapest Open Access Initiative)， 「開放近用」指的是：資料能在公眾網際網路上被自由取用，並允許任何使用者 得閱讀、下載、重製、散布、列印、搜尋或連結至該等文章的全部內容、得為檢 索目的爬梳(crawl)該等文章、得傳遞該等文章至軟體作為資料，或為任何其他合 法目的，在除了與近用網際網路不可分離的限制外，別無財政、法律或科技障礙 之下而能使用該等文章。對於重製及散布的唯一限制，及著作權在上述範圍內的 唯一角色，是必須允許作者得控制他們著作的完整性，和保障作者應被適當表彰 及引用。
而貝斯達開放近用出版宣言(Bethesda Statement on Open Access Publishing)則 對何謂「開放近用出版」給了定義。只要符合以下兩個條件，即可稱為開放近用 出版。 1. 作者及著作權人給予所有使用者自由、不可撤回、全球、永久的得近用其 著作的權利，並授權所有使用者得在任何數位媒介、為任何合法目的、在 適當表彰著作人的條件下重製、使用、散布、傳輸、公開展示、改作其著 作，及得為個人使用的目的少量印刷其著作。 2. 完整的著作內容及所有補充資料，包含上述所允許的著作重製物，必須以 適當的標準化電子格式、在著作初出版時立即被存放在至少一種線上素材 庫(online repository)中，而該等線上素材庫必須是由學術機構、政府組織或 其他設立完善以尋求促進開放近用、不受限制地散布、可互通性及長期典 藏為目的的組織所支持。 延續布達佩斯開放近用提議及貝斯達開放近用出版宣言的精神，於 2003 年 簽署的柏林宣言關於科學及人文學知識的開放近用(Berlin Declaration on Open Access to Knowledge in the Sciences and Humanities)進一步列舉了開放近用內容 (Open Access Contribution)，舉凡原始科學研究結果、原始資料及後設資料、以
數位方式呈現的圖像資料和學術多媒體資料都屬此處所定義的「開放近用內容」；
此外，柏林宣言關於科學及人文學知識的開放近用對於「開放近用內容」的判斷 標準，則幾乎相同於貝斯達開放近用出版宣言對於「開放近用出版」的認定標準。
21 D. Blumenthal et al., Data Withholding in Academic Genetics, 287
7
前述 SC 下即成立一個｢出版｣計畫，處理學術期刊的出版問題，「出版」計畫
的提出主要為響應近年來提倡開放近用的浪潮，期透過標準化學者與出版商間的 契約，來消除以往傳統學術出版模式所帶給文章作者本身及大眾在取用資料上的 不便利。該計畫中，目前發展最完整的應該是開放近用法律計畫(Open Access Law Program,下稱 OAL 計畫)，該計畫主要係為鼓勵法學類期刊的開放近用。OAL 計 畫分別列出幾項原則，讓贊同其理念的法學期刊出版商（或其編輯）及法律學者
可資遵循。供法學期刊出版商（或其編輯）遵守的原則（「開放近用法律原則」，
Open Access Law Principles）要求：(1) 該期刊向期刊文章作者所取得的商業出 版專屬授權必須是合理的、有限的期限，且不得阻礙作者將其文章以和創用 CC 「姓名標示-非商業性」同樣自由的授權釋出；(2)文章獲該期刊刊登的作者則必 須承諾在再版的情況下，會標明該期刊為這篇文章的原始出版人；(3)文章刊登 之後，期刊出版商必須提供作者該篇文章的電子檔；(4) 若該期刊並無採用 OAL 出版合約(Open Access Law: Publication Agreement)，則期刊出版商必須在其網站 上提供其所採用的出版合約內容讓大眾得閱覽，且該出版合約不得違背以上所列 的 4 項原則(Science Commons, n.d.j)。目前許多知名的法學期刊，如杜克大學法 學期刊(Duke Law Journal)、哈佛大學法律與科技期刊(Harvard Journal of Law & Technology)、紐約大學法學期刊(New York Law School Law Review)等都已加入 了這個計畫(Science Commons, n.d.h)。 而要加入 OAL 計畫的法律學者則必須承諾將遵守以下 3 原則：(1) 在該學 者對某法學期刊具有編輯權限的情況下，將採用「開放近用法律原則」；(2) 當 該學者擔任某法學期刊顧問(advisor)時，將鼓勵該期刊的編輯採用「開放近用法 律原則」；(3)該學者將只向支持「開放近用法律原則」的期刊投稿。 除了針對法學期刊出版商（或其編輯）和法律學者分別擬定可資遵循的原則 外，OAL 計畫也制訂了供兩方共同簽署的制式出版合約，讓作者得保有其文章 的所有著作權，且納入 4 種創用 CC 授權的選項：「姓名標示」、「姓名標示-非商 業性」、「姓名標示-非商業性-相同方式分享」、「姓名標示-非商業性-禁止改作」 讓作者自行決定其文章欲以何種創用 CC 授權提供大眾利用。若某法學期刊原先 並非採用 OAL 所擬定的出版合約，OAL 計畫也提出 3 種具修改原先出版合約效 力的附件(addendum to publication agreement)來達到文章得開放近用的效果。「開 放近用創用 CC 1.0 附件」(OpenAccess-Creative Commons 1.0 Addendum)規定， 作者保有得為非商業性目的利用其著作的權利，並得將其著作依類同於創用 CC 「姓名標示-非商業性」效果的授權方式提供予大眾利用。例如，作者得在教學 及研究的過程中重製及散布其著作，亦得在其個人網頁、甚至其他開放近用數位 資料庫上張貼其著作。第 2 種附件，「開放近用 1.0 附件」(OpenAccess-Publish 1.0 Addendum)除規定作者得散布其文章的任何一個版本（含期刊最後刊登的版本）， 更進一步寫明作者得為和其教學、研討會發表、演講、其他學術著作和專業活動
8 相關的目的利用其文章，但不強制規定期刊出版商須授權作者得將其文章以創用 CC 授權；最後一種附件，「開放近用延遲 1.0 附件」(OpenAccess-Delay 1.0 Addendum)內容大致和「開放近用 1.0 附件」相同，但另外規定作者需在文章刊 登 6 個月後才得將文章最終刊登的版本提供大眾取得。無論是上述那一種附件中 皆規定，若期刊出版商在不簽署附件的情形下逕行刊登作者的文章，將視同期刊 出版商同意附件之條款。 再者，由於長久以來，學術機構是傳統學術期刊出版商的主要支持者，為出 版商提供期刊內容來源及訂閱費用，所以在可預見的未來，學術機構及學者間的 集體行動將為傳統出版模式與開放近用出版模式的消長帶來關鍵影響。近年來許 多世界一流學術機構都已經意識到出版商對學術研究成果過度控制所造成的不 良後果，諸如美國加州大學、哈佛大學及康乃爾大學等學術機構都開始抵制高價 的學術期刊及其出版商，要求校內教授考慮避免在該等期刊上發表著作，這些學 術機構也鼓勵教授在接受期刊刊登邀約時，應保留在公共網路上自我典藏該著作 的權利，但是出版商是否會接受，則涉及雙方的議價實力及商業考量，儘管有研 究顯示高達 90%的期刊允許某種形式的自我典藏，然而，實際運作上，學者卻容 易受到不同期刊有不同的政策（如允許著作典藏的格式、著作出版後經過多久的 期間始得公開等）所困擾；基於相同的原因，學術機構也開始思考，如何藉由鼓 勵開放近用的出版模式，重新取得對學術研究成果的控制權。 二、研究工具與材料移轉 基於研發成果專利化與商業化在科學研究上的發展趨勢，近年來針對政府補 助完成研發成果之開放科學呼聲時有所聞，美國國家衛生研究院(National Institute of Health, 以下簡稱NIH)於 2004 年時即提議，所有由NIH補助完成的研 發成果都應及時對其他科學家、醫療服務提供者、學生、教師、甚至所有有此資
訊需求的美國人民公開22。而許多開放科學的支持者已開始思考，是否應參考自
由軟體 (free software)或開放原始碼軟體(open source software)的授權模式，來促 進科學資訊的流通與分享，如此的想法最早由英國聖格研究院（Sanger Institute） 的科學家Tim Hubbard提出，其認為開放原始碼的授權模式非常適合應用於人類 基因體研究成果23，Hubbard雖然自己草擬了基因研究成果的相關公眾授權條款， 但卻未被聖格研究院採用，因為該院認為，所有的科學研究材料及成果，都應釋 出到公共領域，而既然該等成果已在公共領域，則自無以其授權之可能性24
22 Press Release, National Institutes of Health, Enhanced Public Access to NIH Research Information
(Sept. 3, 2004), available at http:// grants.nih.gov/grants/guide/notice-files/NOT-OD-04-064.html. ，該
23 Guadamuz González, supra note 17, at 336. 24 Id.
9 院之觀點反應了部分科學家對於將科學研究成果以公眾授權釋出的懷疑，因為這 些科學家認為，授權代表將研發成果財產權化，但是財產權化其實向來被認為是 有害於科學資訊及材料的分享與流通25。 承前所述，材料移轉契約其實帶來了相當可觀的交易成本，因此並無法達到 促進科學資料分享與交流之理想，因此科學界近年來開始逐漸開始出現簡化科學 資料分享程序之呼聲26，申言之，若科學家們將其研發材料分享予其他科學家時， 無須花費驚人的精力及成本，則其將有更高的意願將系爭材料提供分享。然而， 該等簡化資料分享程序之目標其實並不易達成，過去美國公共衛生學界及產業界 代表曾協力制訂了一份｢標準生物材料移轉契約範本｣(Uniform Biological
Material Transfer Agreement, UBMTA)，其目的即為藉由該標準化的範本，簡化生
物材料分享程序、降低交易成本27，但是該範本實際推行的結果卻是，許多持有 生物研究材料之機構仍站在本位主義之立場，將該範本之條款修改為對自己有利 的條款，對生物研究材料之移轉增加許多原本範本所無之限制，有論者指出，會 有如此發展趨勢之原因，係因為研究機構的技術移轉單位多將收取授權金或相關 費用當作自己之主要工作內容，因此在生物材料移轉契約之制定、談判及簽訂時， 都會盡期所能的增加契約對造取得或使用系爭材料之限制28。然而，將原本標準 化的契約範本過度客制化的結果，不但增加了契約兩造就契約談判之交易成本， 亦使得原本契約範本簡化生物材料分享程序及促進材料分享之目的無法達成。 創用 CC 組織基於其過去發展彈性內容公眾授權條款之成功經驗，近年來亦 開始在其 Science Commons（以下簡稱 SC 計畫）中投入發展標準化的材料移轉 契約。與前述 UBMTA 不同的是，創用 CC 過去發展出的授權條款，在一定程度 內提供了授權人就特定授權方式（如是否供商業使用、是否可改作以及改作作品 是否須依相同方式分享等）的選擇性，且依據該授權條款所授權釋出之內容多可 利用網路搜尋引擎尋得。 對 SC 所要發展的標準化材料移轉契約而言，創用 CC 的成功經驗可提供以 下的啓發：為了降低契約雙方協商的交易成本，應該僅可能將雙方較無爭議或利 益衝突性較低之條款確定，而在雙方較有可能需要花費資源協商之條款中，依照 材料使用受限制之程度，設計出限制程度寬嚴不一的一系列使用條款，而該等設 計係將現今材料移轉契約之設計實況予以類型化。此外，應將材料移轉契約數位
25 Sulston, supra note 18, at 64. 26 Baca, supra note 5, at ¶27.
27 Uniform Biological Material Transfer Agreement: Discussion of Public Comments Received:
Publication of the Final Format of the Agreement, 60 Fed. Reg. 12771, 12771 (Mar. 8, 1995).
10 化，使科學家們能輕易的透過簡單的電腦操作，將其所開發出的材料授權與其他 科學家使用，或搜尋其他科學家有哪些材料、依據何等條款可授權移轉他人使用。 最後，標準化的材料移轉契約應協助科學家在「放棄全部權利」及「保留全部權 利」的兩個極端中，取得「保留部份權利」但仍開放部份資源之可能性。 肆、推動 Science Commons 運動所面臨的挑戰 一、學術期刊 就學術期刊經營者的觀點而言，開放近用經營模式的商業價值仍有待觀察， 就財務的觀點而言，純粹開放近用的經營模式可節省行銷及訂戶管理的費用，但 是卻喪失了傳統經營模式主要的獲利來源：訂閱費用。目前大部分開放近用出版 的經營費用係來自：出版機構的自有經費、外界捐贈、會員加值服務費用、作者 所繳納的稿件處理費用及廣告收益等。此外，就稿件供給的觀點而言，開放近用 出版模式多要求作者繳納相當的稿件處理費用，此等要求可能會影響作者提供稿 件的誘因，該誘因的降低，對於尚未建立學術聲望的期刊而言，係屬不易處理的 課題。 從作者的角度而言，開放近用的出版模式除了可增加讀者的數量以外，另一 個潛在的利益是作者可以得到更多關於其作品的回應，開放近用期刊的出版者可 以透過網際網路及資訊科技，為每一篇文章建立獻上論壇，另外，作者亦可以將 其研究的基礎資訊，以不同的形式（如聲音、影片、數據、實驗筆記及資料庫等） 置於網站上供讀者檢驗參考，如此將可提升研究的準確性及嚴謹度；從讀者的角 度而言，開放近用出版除了降低接近使用期刊的成本以外，更重要的是，讀者可 透過 Google Scholar、Citeseer 或 Citebase 等搜尋引擎找尋其所需求的學術文章， 由於許多研究者都習慣在各自專業領域的資料庫中尋找相關文獻（如法律學者習 慣使用 Westlaw 及 LexisNexis 資料庫），所以對於其他領域的相關主題研究可能 不易得知，但是透過開放近用出版的期刊及網際網路，研究者將可更輕易的取得 其他領域的相關文獻，就此而言，開放近用的出版模式亦有助於跨領域研究。 由於開放近用的出版模式不向讀者收取線上閱讀及下載的訂閱費用，所以對 於此種經營模式最常見的質疑就是其在經濟上是否能達到收支平衡。事實上，為 支付專業期刊的必要開銷（如排版、安排同儕審查、網站管理及印刷等行政費用）， 一些開放近用學術期刊的出版者會向作者收取一定的出版費用，以知名的開放近 用出版組織公共科學圖書館(Public Library of Science, PLoS)為例，其所出版的期 刊要求作者或其所屬的研究機構支付每篇美金 1,250 元到 2,500 元不等的出版費 用，但是如果作者及其所屬機構無力負擔該出版費用，則可申請免除負擔部份或 全部出版費用，此外，如果作者所屬的研究機構支付美金 2,000 元至 100,000 元 的會員年金，則作者在出版費用上可獲得 10%到 75%的折扣，PLoS 的主要經濟
11 來源來自捐款、獎助、紙本的訂閱費用以及上述作者或其所屬機構所支付的出版 費用。另一家英國的開放近用期刊出版商 BioMed Central(BMC)則向作者收取每 篇美金 490 元到 2,340 元的處理費用，但若作者或其所屬研究機構已經支付過 BMC 的會員費用，則無須支付文章處理費用。而近年來一些實證的研究也顯示， 開放近用的學術期刊經營模式，在經濟上是可行的，而開放近用學術期刊的數量， 也在持續增加中。 開放近用期刊能否經營成功，另一個關鍵性的因素是此等期刊能否建立卓著 的學術聲譽。對作者來說，在開放近用的期刊上刊登文章，或許可以增加被閱讀 的次數及被引用的頻率，而此種頻率的增加雖然可以某種程度的提升作者的學術 地位，但是在開放近用期刊建立起一流學術聲譽之前，許多作者可能還是會把最 滿意的著作優先投稿至如 Science 或 Nature 傳統的一流學術期刊，以獲得學術界 普遍認可的肯定，就此而言，開放近用期刊似乎不易在短時間內，挑戰既有的一 流學術期刊所建立的學術聲望。此外，實證研究也發現，許多學者對於專業領域 中的開放近用學術期刊仍不熟悉或缺乏信心，此種情形極可能導致開放近用期刊 獲得較少高水準的稿件，而難以提升其學術水準。 因此，開放近用期刊能否在學術市場全面成功，將端視學術界能否以成功的 集體行動支持開放近用的出版模式，事實上，近年來學術界的集體行動已經逐漸 展開，除了前述一流學術機構對傳統期刊出版模式的反動以外，許多重量級的學 者也開始擔任開放近用學術期刊的編輯委員，以表達對開放近用理念的支持及期 刊內容的肯定，而 PLoS 更是由 1989 年諾貝爾醫學獎得主 Harold Varmus 博士所 發起成立，並獲得多位諾貝爾獎得主的支持，其所出版的期刊內亦不乏諾貝爾獎 得主的著作。 二、研究材料與成果 前述創用 CC 所欲發展之標準化材料移轉契約其實亦受到不少質疑，許多論 者指出，過去創用 CC 的發展經驗並無法完全移植到材料移轉契約，因為創用 CC 授權條款所針對者係受著作權法所保護之著作，但材料移轉契約所欲移轉者 則多為具體的物或不受著作權法保護之資訊；此外，材料移轉所涉及之相關當事 人關係可能較創用 CC 著作權授權之主體更為複雜，許多大型研究機構就研發成 果都受到自身規定、相關法令及提供研究資助者的規範，而如前所述，許多大學 或研發機構的技術移轉部門人員，對於凡是有可能降低授權金收入之條款，都不 易接受。凡此種種，都使得 SC 要改變既存的科學材料移轉實務相當不易。 此外，科學研發的成果與過去適用創用CC授權條款之內容有顯著不同，後 者係受著作權法保護之客體，受到著作權法創作保護主義29 29 著作權法第 10 條：｢著作人於著作完成時享有著作權。但本法另有規定者，從其規定｣。 之規範，一旦創作完
12 成，即可授權釋出，但前者所涉及的智慧財產權卻多為專利權及營業秘密，若是 專利權，則開發者須待專利申請核准後，方有權利將系爭專利授權釋出，然而專 利取得之成本及時間有時相當可觀，有可能會影響開發者授權釋出之權利及意願。 此外，由於取得專利權的成本有時相當龐大，因此許多從事科學研究的機構都期 待藉由專利授權能回收部分研發成本與投資，並希望以專利權作為其和其他研究 機構或企業協議交互授權之談判籌碼30。因此，縱使有完整的授權條款，亦不易 讓研究機構以公眾授權的方式，將研發成果授權釋出。 伍、結語 雖然有許多頂尖的科學家都支持開放科學運動，但是保留甚至創造一塊純淨的 Science Commons 園地卻非常困難，因此開放科學運對所面臨的挑戰其實甚為嚴峻。 這些挑戰除了來自科學研究過程中部不擴張的商業力量外，亦與既有的科學研究組 成結構及國家科技政策有密切關係。 就學術期刊的開放近用而言，雖然已出現了少數經營成功的學術期刊，但是多 數科學社群的成員對於開放近用出版能否有成功、延續的經營模式，仍高度存疑， 本文以為，雖然學術期刊的開放近用對於不同的學術領域的價值及適用方式或有差 異，但是政府機關或政府所屬研究機構所出版的學術期刊，均應採取開放近用的出 版模式，以使得其所出版的研究成果能更廣泛的為研究人員及社會大眾所接近使用； 再者，透過聲譽卓著的學術機構（諸如美國加州大學、哈佛大學及康乃爾大學等） 漸漸出現的抵制高價學術期刊及其出版商的行動，及民間組織如創用 CC 成立 SC 「出版」計畫來推動期刊的開放近用。 而研究材料及成果的開放及分享，在實踐上其實比學術期刊的開放近用更為困 難，蓋其所涉及之專利權對許多研究機構均具有聲譽上及財務上之重要性，以現今 各國以專利權衡量研發成果及促進產學合作之政策而言，研發成果專利化之趨勢早 已蔚為風潮。雖然，不可否認的，財產權對於提升研發績效有相當程度的正面影響， 但是過度財產化所有研究材料及資訊其實對長期科學發展並非益事。未來開放科學 運動的推動成效，實有賴於科學家、科學社群及研發機構間的集體行動，否則單依 少數組織或個人的大聲疾呼，實難改變既有科學研究社群所瀰漫的專利化及秘密化 之競爭結構。
1
F
ILTERINGO
NLINEC
ONTENT:
“C
ODE ISL
AW”
IN THEC
ASE OFC
HINAJyh-An Lee*
I. Introduction ... 1
II. Internet Filtering in China ... 4
III. Code-is-Law in the Context of Internet Filtering ... 9
A. Code-Is-Law Theory ... 9
B. Theory Application ... 11
1. Law v.s. Code as Regulation... 11
2. Fulfilling Policy Goals via Architecture Design ... 14
3. Architecture’s Impact on Human Behavior ... 18
4. Regulating the Intermediaries ... 21
IV. Conclusion ... 23
I. Introduction
In the book “The World Is Flat,” Thomas Friedman states that anyone with an Internet connection has the ability to find almost any information on the web now.1
* Assistant Professor of Law, National Chengchi University, Taiwan.
2
number of commentators also claim that the Internet enables free flow of information and a freer society.2 This may be true in most, but not all, countries in the world.3 In China, the government has built probably the world’s most sophisticated Internet filtering system to block a number of foreign websites,4 which are viewed by the government as a threat to the Chinese State. The blocked websites include those containing information associated with Tibetan Independence, Taiwan Independence, human right, Falun Gong, and other perceived threat to the Communist Party.5
1 T
HOMAS L.FRIEDMAN,THE WORLD IS FLAT:ABRIEF HISTORY OF THE TWENTY-FIRST CENTURY (2006). The government argues that such practice is desirable as it can prevent the Western world from “dumping” information on China. Therefore, there is no surprise that in a public talk, Hu Jintao once said, “[w]hether [the government] can cope with the Internet is a matter that affects the
2 Ronald J. Deibert, Dark Guests and Great Firewalls: The Internet and Chinese Security Policy, 58 J.S OC. ISSUES 143, 143 (2002); Christopher Stevenson, Breaching the Great Firewall: China’s Internet
Censorship and the Quest for Freedom of Expression in a Connected World, 30 B.C.INT’L &COMP.L. REV. 531, 533-34 (2007).
3 Stevenson, supra note 2, at 534; see also Deibert, supra note 2, at 143 (China “is a ‘hard case’ for those
who argue that the Internet cannot be controlled”); Kristen Farrell, The Big Mamas Are Watching: China’s
Censorship of the Internet and the Strain on Freedom of Expression, 15 MICH.ST.J.INT’L L. 577, 590 (2007) (“[t]he Internet [in China] has increasingly become a tool for security agencies to identify, monitor, arrest and imprison potential dissidents.”)
4 See e.g. Stevenson, supra note 2, at 536-37; Lijun Tang & Peidong Yang, Symbolic Power and the Internet: The Power of a “Horse”, 33 MEDIA,CULTURE &SOC’Y 675, 678 (2011). See also YUEZHI ZHAO, COMMUNICATION IN CHINA:POLITICAL, ECONOMY,POWER, AND CONFLICT 32(2008)(stating that “[w]ith the increasing sophistication of firewalls and filtering software, the survival time for offensive content in cyberspace has been progressively reduced.”).
5 See e.g. Stevenson, supra note 2, at 541; Robert Faris & Nart Villeneuve, Measuring Global Internet Filtering, in ACCESS DENIED: THE PRACTICE AND POLICY OF GLOBAL INTERNET FILTERING 5, 12 (Ronald Deibert et al. ed., 2008). See also Farrell, supra note 3, at 587 (“China considers a wide range of topic sensitive and controversial…including the Tiananmen Square uprising, support for a free Tibet, the Falun Gong spiritual movement, criticism of China’s human rights and social justice records, independent news media, and pro-democracy/pro-Western commentary.”).
3
development of socialist culture, the security of information, and the stability of the state.”6
When the Internet was first introduced to China, some researchers had optimistically viewed it as a librating force in China’s democratic development.7 They assume that free flow of information fostered by the Internet would definitely lead to a free society8. Nonetheless, the Chinese government has actually been using network technologies to control online information and grafting its own ideology to the Net. Digital technologies have become the government’s tool to tamp down political threats.9 Technically, the Chinese government has ordered Chinese Internet carriers, like China Telecom, to deploy Cisco’s equipment to block unwanted materials from entering China.10
In fact, the Chinese government has been attempting to control online content via several different targets, including Internet content providers, individual consumers,
This practice has significantly changed the open nature of the Internet. Although the government may choose to use the law to regulate people’s online behavior, controlling access to online information via technical architecture seems to be a much more effective approach.
11
6 Xiao Qiang, The Rise of Online Public Opinion and Its Political Impact, in C
HANGING MEDIA,CHANGING CHINA 202, 207-08 (Susan L. Shirk ed., 2011).
and content on foreign websites. Nonetheless, this Article only focuses on the filtering mechanism that prevents local Chinese Internet users from accessing foreign online content. This Article will use Lawrence Lessig’s pronouncement “code is law” as lenses to understand the Internet filtering system in China. The theory is that technology can
7 See e.g. Yutian Ling, Upholding Free Speech and Privacy Online: A Legal-Based and Market-Based Approach for Internet Companies in China, 27 SANTA CLARA COMPUTER &HIGH TECH.L.J. 175, 215 (2011).
8 See e.g. Stevenson, supra note 2, at 533-34; S
HANTHI KALATHIL &TAYLOR C.BOAS,OPEN NETWORKS, CLOSED REGIMES:THE IMPACT OF THE INTERNET ON AUTHORITARIAN 1-2 (2003).
9 It is not easy to draw the relationship between the Internet and democracy because there are a number of
complicated human experiences, institutions, and other factors in between. See GUOBIN YANG,THE POWER OF THE INTERNET IN CHINA 10 (2009).
10 See e.g. Bambauer, at 379; Stevenson, supra note 2, at 542.
11 See e.g. Gudrun Wacker, The Internet and Censorship in China, in C
HINA AND THE INTERNET:POLITICS OF THE DIGITAL LEAP FORWARD 58, 69-70 (Christopher R. Hughes & Gudrun Wacker ed., 2003); Yang,
4
fulfill a regulatory function or at least have the same effects as regulation. The essential characteristic of code-as-regulator is that “[a] rule is defined, not through a statute, but through the code that governs.” Through the application of Lessig’s theory to the Chinese great firewall, this Article intends to illustrate the theory’s new implications and the government’s policy options in cyberspace. Indeed, the Internet filtering practice in China has drew considerable criticism, especially from the perspectives of democratic development and the value of open Internet.12 Nonetheless, the aim of the Article is not to criticize the Chinese Internet filtering system, but rather to illustrate how a government can regulate and shape human behavior via architecture.
II. Internet Filtering in China
The information and communications technologies (ICT), including the Internet, in China have been growing rapidly because of strong support from the government during the past 15 years.13 The Internet infrastructure in China has experienced extraordinary growth in terms of scale, technology, and quality.14
12 For example, When President Barack Obama visited China in 2009, he said:" I am a big supporters of
non-censorship...in the United States, the fact that we have free Internet--or unrestricted Internet access is a source of strength, and I think [it] should be encouraged." See Robert Mackey, Obama Walks China's
"Great Firewall," N.Y.TIMES, Nov. 16, 2009, at http://thelede.blogs.nytimes.com/2009/11/16/obama-on-chinas-great-firewall/. For other criticisms on Internet filtering, see e.g. Kevin Werbach, The Centripetal
Network: How the Internet Holds Itself Together, and the Forces Tearing It Apart, 42 U.C.DAVIS L.REV. 343, 367 (2008).
In the meantime, the Chinese government has endeavored to control the information flowing on the Internet via various approaches, such as regulations and technologies.
13 See e.g. Wacker, supra note 11, at 58.
14 See e.g. Wei Wu, Great Leap or Long March: Some Policy Issues of the Development of Internet in China, 20 TELECOMM.POL’Y 699, 699-701 (1996); Jonathan J.H. Zhu & Enhai Wang, Diffusion, Use, and
5
“Filter” originally means blocking data from entering or leaving a network by programming the router.15 The aim was to provide Internet service providers (ISPs) with means to control viruses, worms, and spam.16 The same technology has been employed by the government to filter online information,17 and that became the “Internet filtering,” which represents the technical approach to prevent Internet users from accessing specific Internet Protocol (“IP”) addresses.18 The reason to block online information from citizens is that such information is deemed too sensitive by the government.19 A great number of countries have developed their own Internet filtering systems because of political, moral, religious, or security concerns.20 Traditionally, there are two types of Internet filtering technique: the inclusion filter and the exclusion filter.21 The inclusion filter typically uses a “white list” to include websites that are permitted for browsing, whereas the exclusion filter employs a “blacklist,” which specifies websites that users are prohibited from visiting.22 Countries blocking websites usually request Internet service providers (ISPs) to implement the task because it is the cheapest way to filter online information.23
The Chinese government has adopted the exclusion filter by requesting carriers, such as China Telecom, to install Cisco’s apparatus, which can drop information from at least three hundred IP addresses.24
15 H
UMAN RIGHTS WATCH,RACE TO THE BOTTOM:CORPORATE COMPLICITY IN CHINESE INTERNET CENSORSHIP 9 (2006)
The Chinese government provided the carriers with a list of forbidden websites and their addresses, and ordered these carriers to block the sites
16 Id. 17 Id.
18 Marc D. Nawyn, Code Red: Responding to the Moral Hazards Facing U.S. Information Technology Companies in China, 2007 COLUM.BUS.L.REV. 505, 510 (2007).
19 Jonathan Zittrain & John Palfrey, Introduction, in A
CCESS DENIED: THE PRACTICE AND POLICY OF GLOBAL INTERNET FILTERING 1, 1 (Ronald Deibert et al. ed., 2008)
20 Id. at 3; Faris & Villeneuve, supra note 5, at 6, 9. 21 Nawyn, supra note 18, at 510.
22 Id; Ling, supra note 7, at 184.
23 Faris & Villeneuve, supra note 5, at 13-14. 24 Id. J
ACK GOLDSMITH &TIM WU,WHO CONTROLS THE INTERNET:ILLUSIONS OF A BORDERLESS WORLD 93 (2006).
6
through Cisco’s equipment. 25 These sites include Amnesty International’s
(www.amnesty.org), Reporters without Borders (www.rsf.org), the BBC (news.bbc.co.uk),26 the Economist (http://www.economist.com), and the New York Times (http://www.nytimes.com)27
From the government’s perspective, because new websites continuously emerge, the inclusion filter usually includes too few websites, while the exclusion may exclude too few.
. In this way, certain information has been dropped and can never reach the domestic end users.
28
In order to avoid such over-blocking or under-blocking, governments have started to the “content-analysis” technique as a new tool for Internet filtering.29 Content-analysis approach prevents users from accessing any website or URL path containing certain keywords designated by the government.30 One advantage for the government to adopt the content-analysis approach is that it does not have to incessantly update the white list or blacklist. In China, keywords for content analysis may include Tibetan Independence, Taiwan Independence, human right, Falun Gong, and etc.31
The Chinese government has built a complicated technical system into the Internet to filter online information since the digital network was built.32 In 2002, Jonathan Zittrain and Ben Eldman produced a list of foreign websites blocked in China through the help of an end user there.33
25 G
OLDSMITH &WU, supra note
The Chinese government blocked them or has blocked them because they were deemed as a threat to the Chinese state. Of course, China is not the only country that filters away politically sensitive content. Other countries with similar motives include Bahrain, Ethiopia, Libya, Iran, Indonesia, Malaysia, Myanmar, Thailand,
24, at 93-94.
26 A
NDREW MURRAY,INFORMATION TECHNOLOGY LAW 74 (2010); Farrell, supra note 3, at 588.
27 Deibert, supra note 2, at 147. 28 Nawyn, supra note 18, at 510-11. 29 Id. at 511.
30 Id. at 511; Faris & Villeneuve, supra note 5, at 15; Ling, supra note 7, at 184; Susan L. Shirk, Changing Media, Changing China, in CHANGING MEDIA,CHANGING CHINA 1, 14 (Susan L. Shirk ed., 2011)
31 See e.g. G
OLDSMITH &WU, supra note 24, at 96; Stevenson, supra note 2, at 541.
32 Nawyn, supra note 18, at 512; Stevenson, supra note 2, at 540.
33 Jonathan Zittrain & Benjamin Edelman, Internet Filtering in China, 7 IEEEI
NTERNET COMPUTING 70 (2003).
7
Pakistan, Saudi Arabia, Singapore, Syria, Tunisia, Uzbekistan, and Vietnam34. For different purposes such as blocking pornography, some democratic countries, including Australia, Britain, Canada, France, Japan, and New Zealand, filter online content as well.35
But how could the Chinese government control the information flow into the country? It actually built a great firewall via the help of the U.S. hardware vendor Cisco,36 which has made the whole country’s Internet into a huge intranet.37 It is estimated that the company earns USD$500 million each year in China.38 Other companies that provide filtering software to China include Sun Microsystems, Websense, and Bay Networksboth, all of which are U.S. companies as well.39 The filter has been constructed on different layers of China’s Internet, but primarily at the backbone, which is the physical infrastructure that links the domestic Internet to global networks.40
34 Bambauer, at 382; Faris & Villeneuve, supra note 5, at 9-10; Shaojung Sharon Wang & Junhao Hong, Discourse Behind the Forbidden Realm: Internet Surveillance and Its Implications on China’s Blogosphere,
27 TELEMATICS &INFORMATICS 67, 74 (2010).
35 Bambauer, at 382; Derek E. Bambauer, Filtering in OZ: Australia’s Foray into Internet Censorship, 31
U.PA.J.INT’L L. 493, 516-17 (2009).
36 G
OLDSMITH &WU, supra note 24, at 93; Stevenson, supra note 2, at 541-42.
37 Deibert, supra note 2, at 147; Stevenson, supra note 2, at 540-41. 38 Stevenson, supra note 2, at 542.
39 Id; Deibert, supra note 2, at 148; see also Farrell, supra note 3, at 587 (“American engineers aided the
Chinese in censorship by developing special routers, integrators, and a special firewall boxes”); Wacker,
supra note 11, at 69 (“[i]t is ironic, therefore, that while the Western media frequently criticize China for
obstructing the development of the Internet, it is Western firms that are supplying the technological means which enable China to carry out surveillance.”) Some literature focuses on the legality of those U.S. companies’ support of the Chinese filtering regime, especially whether they violate the Global Online Freedom Act of 2006, see e.g. Ling, supra note 7, at 192-94; Nawyn, supra note 18, at 544-554; Stevenson,
supra note 2, at 545-558. In the meantime, human right supporters also sent out appeals criticizing Cisco's
involvement with the Chinese filtering regime. See e.g. Electronic Frontier Foundation, Tell Cisco: Stop
Helping China Abuse Human Rights!, at
https://secure.eff.org/site/Advocacy?cmd=display&page=UserAction&id=504 (last visited Aug. 10, 2011).
8
The metaphor most frequently used in describing the Internet filtering in China is “the great firewall,”41 which is obviously the combination of “the Great Wall” and “firewall.” The Great Wall was originally built to keep foreign barbarians out of ancient China while the great firewall denotes China’s attempt to block undesirable content from its netizens. Different from the firewall established to protect enterprises’ information security, the Chinese great firewall is set around the whole country.42
The country’s Ministry of Information Industry (MII) is authorized to build the network connected to the global Internet, and, thus, has the opportunity to ensure government control over the network.
43
Because online information enters into the country through a limited number of points, the Chinese government is able to control the information via controlling these points.44 Government control over information flow is via several Internet access providers (IAPs), “each of which has at least one connection to a foreign Internet backbone.”45 IAPs peer at three Internet exchange points (IXPs) run by the Chinese government, and these IAPs “grant regional Internet service providers (ISPs) access to backbone connections.”46
41 See e.g. Ling, supra note
Therefore, individual Chinese end users purchase Internet access from several thousand ISPs, and those ISPs are in effect retail sellers of Internet access wholesale of the few IAPs. Different from the decentralized Internet architecture in most countries around the world, most ISPs in China need to connect to the global network through one of the four state-controlled companies operating the IAPs
7, at 177, 180, 184; David Pierson, Brief Web access in China as Great
Firewall Falls, WASH.POST, A09, Jan. 5, 2010, at http://www.washingtonpost.com/wp-dyn/content/article/2010/01/04/AR2010010403599.html; Qiang, supra note 6, at 206.
42 G
OLDSMITH &WU, supra note 24, at 92
43 Farrell, supra note 3, at 585; H
UMAN RIGHTS WATCH, at 9.
44 Y
OCHAI BENKLER,THE WEALTH OF NETWORKS:HOW SOCIAL PRODUCTION TRANSFORMS MARKETS AND FREEDOM 267 (2006); GOLDSMITH &WU, supra note 24, at 93.
45 OpenNet Initiative, Internet Filtering in China: 2006-2007, http://opennet.net/studies/china2007 (last
visited Mar. 26, 2010); see also Deibert, supra note 2, at 147 (“[s]uch funneled access provides the most important outer layer of control and basis for “firewall” technologies to be implemented that ostensibly block controversial or politically undesirable Web sites”); Faris & Villeneuve, supra note 5, at 14
(“[China’s] blocking is done at the International gateway level affecting all users of the network regardless of ISP.”)
46 G
9 and IXPs.47
In other words, the Chinese government had architecture the national’s Internet into two layers. The lower layer is the network where ISPs provide Internet access to consumers, while the upper layer is another set of connections where the lower layer can connect to the network outside the country.
By effective managing the IAPs and IXPs, the Chinese government is able to control information flowing from abroad.
48
It is reported that in the upper layer there are nine gateways connecting the nation’s Internet to the international Internet.49 By controlling a number of key connection points in the upper layer, the government is capable of controlling online information flowing from abroad. Therefore, commentators have described the country’s Internet as a huge Intranet.50
III. Code-is-Law in the Context of Internet Filtering
This section applies the code-is-law theory to the Internet filtering scenario in China. By this application, we can see how code-based regulation is different from the law, and how architecture shape human behavior.
A. Code-Is-Law Theory
The code-is-law theory is most notably illustrated by Professor Lawrence Lessig. He argues that code—software or hardware—can perform regulatory function and can have the same effects as legal regulation.51
47 Ling, supra note
The architecture of the Internet, including the languages and protocols underlying software and hardware, has determined how messages are moved from one place to another and how people perceive them. Therefore, whether and how the Internet is regulated depends on its architecture of code. The code is
7, at 184.
48 See e.g. Stevenson, supra note 2, at 540. 49 Qiang, supra note 6, at 207.
50 Stevenson, supra note 2, at 541. 51 L
AWRENCE LESSIG,CODE AND OTHER LAWS OF CYBERSPACE VERSION 2.0 5 (2006)[hereinafter LESSIG, CODE VERSION 2].
10
law in the sense that it constrains what you may or may not do in cyberspace. It enables certain activities while disabling others. Lessig believes that that the “code” which controls the Internet effectively creates the Internet's architecture and its “laws.”52 In a place like cyberspace, sometimes it is the code, not the law, has the greatest impact on human behavior. In Lessig’s words, “[a] rule is defined, not through a statute, but through the code that governs the space.”53 According to him54
The software and hardware that make cyberspace what it is constitute a set of constraints on how you can behave…The code or software or architecture or protocols set [certain] features, which are selected by code writers. They constrain some behavior by making other behavior possible or impossible. The code embeds certain values or makes certain values impossible. In this sense, it too is regulation.
,
Lessig observes that “We can build, or architect, or code cyberspace to protect values that we believe are fundamental, or we can build, or architect, or code cyberspace to allow those values to disappear.”55 From a policy perspective, Lessig reminds policymakers to ponder on which means, including the law and architecture or code, can best advance their goals.56 Just as code’s functionality define the digital universe where people act, it also defines the range of regulatory options for policymakers. Although Lessig explicitly recognizes the fundamental differences between the law and the code,57 some commentators criticize his theory as a disingenuous representation of the role of technologies in regulation58. 52 Id. at 5-6. 53 Id. at 24. 54 Id. at 124-25. 55 Id. at 6. 56 Id. at 129. 57 Id. at 5.
58 R. Polk Wagner, On Software Regulation, 78 S.C
11
B. Theory Application
By building probably the most complicated Internet filtering architecture,59 the Chinese government has been crafting a new Internet architecture according to the nationalist ideology60. This architecture is deviated far from its counterpart in the Western world,61 which has been characterized as openness and freedom.62 Comparing the differences of these two types of Internet architecture, it is not difficult to understand Lessig’s argument that “some architectures enable better control than others.”63
1. Law v.s. Code as Regulation
The “code is law” theory raises interesting questions regarding the role of code or architecture as an alternative to the law. When policymakers have regulatory options between code and law, they shall take into account the impact of each and the costs and benefits associated with each option.64
59 See e.g. Rebecca MacKinnon, Flattered World and Thicker Walls? Blogs, Censorship and Civil Discourse in China, 134 PUB.CHOICE 31, 32 (2008); see also Farrell, supra note
In China, the government has employed several mechanisms to regulate online information available to its citizens. Such mechanisms
3, at 577 (stating that “[c]ompared to other states, China’s censorship regime is pervasive, sophisticated, and effective”).
60 L
ESSIG,CODE VERSION 2, supra note 51, at 89.
61 Id.
62 Stevenson, supra note 2, at 533-34. 63 L
ESSIG,CODE VERSION 2, at 24.
64 See e.g. Jay P. Kesan & Rajiv C. Shah, Shaping Code, 18 H
12
include laws,65 forcing search engines to remove inappropriate content,66 intensive cyber policing,67 and technologies that filter online content.68
Comparing to being regulated by the law, it is usually more difficult for citizens to sense that they are regulated by the code. When a Chinese end user fails to open a forbidden website, the screen will not show that “the Website is blocked by the Government.”
69
It will only show the signal of “site not found.”70 Some countries, such as Tunisia, Iran, United Arab Emirates, and Saudi Arabi,71 use SmartFilter software, developed by the U.S. company Secure Computing, as a proxy filter, and the software provides “a blockpage that looks like the…browser’s default error page.”72 The software used by China is similar to SmartFilter, but is developed by China itself.73 All the software helps to conceal the fact that blocking is taking place. Uzbekistan’s Internet filtering likewise hides government’s blocking by redirecting users to Microsoft search engine www.live.com.74
65 Deibert, supra note
Therefore, it is quite difficult for the user to know whether this
2, at 148-49; Farrell, supra note 3, at 588-90, 598; Ling, supra note 7, at 180-84; Nawyn, supra note 18, at 509-10; Stevenson, supra note 2, at 537-40; YANG, supra note 9, at 48-49.
66 L
ESSIG,CODE VERSION 2, at 80, 309; Stevenson, supra note 2, at 532; 543-44.
67 According to Professor Yuezhi Zhao, a great number of “cyber police squads…are patrolling Chinese
cyberspace, deleting politically incorrect content in real time, blocking websites, monitoring networking activities of citizens, and tracking down and arresting offending individuals.”) See Zhao, supra note 4, at 20.
68 Stevenson, supra note 2, at 540-41. See also Shirk, supra note 30, at 14 (describing that “human
monitors are proactively censor content on their sites”); Qiang, supra note 6, at 207, 208 (introducing the Internet police in China and noting that “human monitors are employed by both Web sites and the government to manually read and censor content.”)
69G
OLDSMITH &WU, supra note 24, at 94.
70 Id. Bambauer, at 391.
71 Stevenson, supra note 2, at 542. 72 Faris & Villeneuve, supra note 5, at 15. 73
HUMAN RIGHTS WATCH,at10; THE OPENNET INITIATIVE,INTERNET FILTERING IN TUNISIA IN 2005:A COUNTRY STUDY (2005), available at http://
www.opennetinitiative.net/studies/tunisia/ONI_Tunisia_Country_Study.pdf.; see also Wacker, supra note 11, at 69 (stating that Chinese companies have begun to supplied the government with filtering software).
13
problem is because of government intervention or a pure technical problem.75
The invisible feature of Internet filtering proves that Lessig’s concern over code-based regulation is not over-stated. Lessig has warned us that since regulating by code is not as transparent as regulating by the law, the former may weaken the democratic value in a society.
And in this way, code shapes and regulates human behavior more surely and subtly than the law.
76
Put it more clearly, when citizens are regulated by code, rather than the law, they will “experience these controls as nature.”77 This is what’s now happening in China. The Chinese government has never disclosed what it filters.78
Of course, governments implementing filtering system can choose not to disguise the fact that they are blocking website. They may decide to declare what material they block in laws or public announcements.
When citizens are more used to the fact that a great number of websites cannot be viewed via their computers, they will more likely to take such intervention and control for granted.
79
For example, Saudi Arabia has disclosed explicitly the reason for Internet filtering in its government website.80 The country using SmartFilter as well, has decided to provide a blockpage that notifies users the requested content is blocked.81 The blockpage also informs users the way to lift the block.82 However, Saudi Arabia is just one of the few countries willing to disclose blocking information and to provide remedy among those with Internet filtering systems.83
Using the law or the code to regulate may bring about different costs. Law regulates behavior through an ex post approach. Law will not be enforcement until a violation Therefore, when regulating by code, the government has the option of whether to disclose its intent in constraining behavior.
75 Bambauer, at 391; G
OLDSMITH &WU,supra note 24, at 94. 76 L
ESSIG,CODE VERSION 2.0, at 138.
77 Id.
78 Bambauer, at 394. 79 Id.
80 Id. at 390-91.
81 Faris & Villeneuve, supra note 5, at 15. 82 Id.
14
behavior takes place.84 Although law enforcement may threaten potential violations in the future, it may also incur significant costs for the regulator. From the perspective of the Chinese government, sending violators, who use the Internet to disseminate prohibited content, to jail or imposing other punishment may draw considerable attention and negative impression internationally. The associated costs are extraordinarily high given China’s increasing importance and visibility in the global community. In contrast, regulating by code is an ex ante approach. Although adopting the Internet filtering techniques may lead to certain criticisms regarding citizens’ right to information, its costs are respectively low for the government as opposed to regulating by law. At least the Chinese government may explain that such practice can be justified by the fact that a large number of countries word wide is filtering online content.85
Sometimes regulating by code needs to be implemented via laws and policies. When such laws and policies are announced, policymakers may still experience considerable costs because this is an ex post approach, rather than purely regulating by code. For example, in 2009 the Ministry of Information Technology initiated a project requiring all computers made and sold in China be preinstalled the filtering software Green Dam Youth Escort.86 However this project was cancelled because of strong public protest.87 The filtering software is eventually required only for computers in schools and Internet cafes.88 The Green Dam Escort initiative is actually an ex post approach, that made citizens aware of the subject regulation and, thus, able to oppose it before it went into effect. Therefore, comparing with the Green Dam Escort initiative, Internet filtering in the gateway level causes much less costs and is obviously more effective.
2. Fulfilling Policy Goals via Architecture Design
84 L
ESSIG,CODE VERSION 2.0, at 124.
85 Faris & Villeneuve, supra note 5, at 6, 13. 86 Qiang, supra note 6, at 209.
87 Id. Ling, supra note 7, at 184-85. 88 Ling, supra note 7, at 185.
