|
|
|
|
Early Warning and Early Response, by Susanne Schmeidl and Howard Adelman (eds.)
Gavan Duffy
Global Affairs Institute
Maxwell School of Citizenship and Public Affairs
Syracuse University
An earlier version of this essay was presented at a workshop on “The World-Wide Web as a Statistical Producer,” sponsored by ESPKIT, the informationtechnologies program of the Directorate General for Industry of the European Commission, and held under the auspices of the University of Ulster in Derry, Northern Ireland, April, 1997.
| The Secretary–General may bring to the attention of the Security Council any matter which in his opinion may threaten the maintenance of international peace and security. |
| —Article 99 of the UN Charter |
1. Introduction
The Secretary–General maintains a crisis situation center at the United Nations headquarters in New York City. In one windowless room, military attachés from various member states monitor international news reports at computer stations and field telephone reports from UN officers, officials of member states, and humanitarian organizations. Next door, large tables arranged in a square dominate the conference room that overlooks the East River. Before each executive-style chair at the tables sits a telephone and a laptop computer. Two large screens for projecting televised news reports and documents cover one wall. Video cameras are mounted throughout the room to record crisis interactions. Although technically sophisticated and professionally managed, this crisis situation center symbolizes a limitation of the United Nations—a limitation of which UN officials are painfully aware. Much like the institution itself, the crisis center is designed to react to crises after the fact. The UN maintains no facilities for anticipating crises.
If he had such a facility, or “early warning” capacity, the Secretary–General could more effectively serve the alert function envisioned for him in Article 99 of the Charter. Early alerts of potential catastrophes—wars, famines, epidemics, floods, ecological disasters, etc.— would provide the world community time to marshal the resources needed for effective political, humanitarian, or military response. Where human actions produce the potential calamity, the institution and its member states could engage more systematically in “preventive diplomacy”—efforts to ease tensions by intervening diplomatically before tensions erupt. Such a facility could profoundly benefit all humankind. It could save millions of lives, prevent unestimable property losses, and help make this planet safer and healthier for us all.
As part of his “Agenda for Peace,” former Secretary–General Boutros Boutros–Ghali recommended development of an early warning capacity centralized in the Secretariat. On December 18,1992, the General Assembly formally encouraged the effort. 1 Without specifying the scope and parameters of this capacity, the General Assembly invited member states and intergovernmental organizations (IGOs) to participate in its design and construction. Initially, the Secretary–General devoted significant resources to the effort. The UN’s Office for Research and the Collection of Information (ORCI) devised the early plans. But ORCI was disbanded in one of the Secretariat reorganizations during the Boutros–Ghali administration. Participants in the ORCI project were scattered to various offices. Some work continues but on a much smaller scale.
Although the UN’s own efforts waned, substantial interest in UN early warning has emerged among peace scientists, political analysts with area expertise, non-governmental organizations (NGOs), and some state and non-state funding agencies. UN early warning has become a recurring theme at conferences of scholars of world politics. Several workshops on the topic have taken place throughout the 1990s, funded primarily by small grants from the Canadian government and coordinated by the Centre for Refugee Studies at York University outside Toronto. These have recently included participation from representatives of the UN Secretariat. At these venues scholars and practitioners have proposed various system designs, discussed practical issues concerning information-gathering and warning transmission, and considered methodological issues pertaining to the analysis of that information.
Before we travel too far down the road toward UN early warning, I think it important to sound a note of caution—an early warning on early warning, if you will. UN early warning, as presently conceived, is wholly infeasible. My argument against the UN conception does not revolve around obvious institutional problems, such as the financial shape of the world body or its well-known bureaucratic dysfunctions. These difficulties are easily resolvable. Rather, it concerns the unsustainability of a political consensus for any early warning system centralized in the United Nations. But my warning extends beyond the UN conception. I will argue further that any centralized system of global early warning is politically unwise.
However persuasive my arguments might be, global early warning can potentially yield such enormous human benefits that we ought to consider ways to make it feasible and to create it wisely. I contend that we would be wise if we were to develop a decentralized global early warning system, for which the World–Wide Web offers an appropriate and convenient venue. But arguing that this is the wise approach does not make it a feasible one. The decentralized, web-based conception of early warning poses tremendous technical challenges. Moreover, imposing political obstacles stand in the way of its creation. I will outline the challenges and identify the obstacles that are apparent to me now and suggest ways to surmount them. I hope thereby to engender a discussion concerning the prospects of sustainable system of global early warning and the steps we can take to realize it.
2. The Infeasibility of UN Early Warning
States traditionally consider early warning part of their intelligence function. For states, early warning typically concerns alerts of impending attack. The UN, however, would generate early warnings of impending human catastrophes. Of particular interest to the UN, as reflected in Boutros–Ghali’s 1995 annual report, 2 are warnings of interstate and intrastate conflicts with the potential of generating large-scale human calamities. This is a distinction without a difference, as both state-based and UN-based early warning involve the collection, analysis, and distribution of information—in a word, intelligence.
The United Nations engages regularly in “monitoring,” “fact-finding,” and “observing” missions that could be construed as intelligence-gathering activities. The UN, long been sensitive to potential charges that it spies on member states, prefers to refer to theses activities as “information-gathering.” Traditional intelligence doctrines distinguish between information-gathering—the collection of raw data—and intelligence-gathering—the analysis of those data. 3 The Agenda for Peace proposal to operate an early warning system within the Secretariat steps over this somewhat fuzzy conceptual line. In order for his warnings to be at all informative to the Security Council, the Secretary–General must analyze the raw data collected.
The Secretary–General might produce intelligence unproblematically on a variety of matters. Intelligence on global climate change, agricultural production, or trends in population should threaten no powerful interest sufficiently to motivate them to stymie the effort. Trouble will arise, however, if and when a UN early warning system analyzes political and security information. The question will arise: Should the Secretary–General operate a political intelligence apparatus?
Some of the great powers will see a UN intelligence operation as a potential threat to the information advantage that helps them maintain their positions of global power. Since these states provide the bulk of the UN’s financial support, they would find it a relatively straightforward matter to quash UN early warning before it starts. The United States has already shown that, simply by withholding its contributions to the world body, it can prevent the reelection of a Secretary–General and prompt the wholesale reconfiguration of the Secretariat. Conservative political elements in the United States would doubtless perceive an intelligence operation in the UN as a direct threat to American sovereignty. Not only would they seek to quash UN early warning, they would do so with patriotic fervor.
Powerful political forces may opt simply to dominate UN early warning, not kill it. In that eventuality, the frameworks and assumptions selected for analyses would cohere with those of the dominating powers. Political trends deemed worthy of notice and questions deemed worthy of asking would be selected from within those frameworks and with those assumptions. Warning messages output from the system would consequently tend to serve the interests of these powerful states.
The rest of the world would rightly find this an abuse of the world body. American domination of UN early warning would lend credence to contemporary third-world critics who view the UN as an American instrument. Under Article 18 of the UN Charter, the General Assembly must approve the budget of the organization. A combination of lesser powers could terminate the system (by a 2/3 vote under Article 19) by rescinding its budget, should they believe that had been captured by one or a few powerful states.
This is the basic case for the infeasibility of an early warning system centralized in the United Nations. Due to the intrinsic intelligence nature of early warning and due to the location of the United Nations within the structure of the Westphalian system of sovereign states, a UN early warning system is politically unsustainable.
But a more general argument can be made. Let us suppose that somehow, magically, powerful states do not notice or do not care whether the Secretary–General acquires an intelligence operation. Even under this fanciful scenario, any centralized UN early warning system would prove politically untenable. Assuming that the UN cannot allocate unlimited resources to early warning, the Secretary–General’s early warning bureau would naturally select a restricted range of principles and frameworks with which to guide choices regarding the appropriate information to gather, the appropriate interpretation of that information, and the appropriate warnings to convey to the Secretary–General. But which frameworks and which principles will it rely upon? Or, more to the point, whose frameworks and whose principles will it use?
The world is composed of many peoples with various experiences and different cultural traditions. They share many commonalities, but they also diverge one from another in a myriad of ways. They see the world—particularly the social world—quite differently. Where some see private loans of capital as good economic sense, others see usury. Where some see liberty in free markets, others see economic domination. Where some find prosperity in the exploitation of natural resources, others find the rape of nature. Where some believe governments relate to citizens as parent to child, others believe governments relate to citizens as servant to master. In short, people believe in and act upon different principles and different frameworks. Whose principles and whose frameworks will guide global early warning?
3. Science to the Rescue?
Well, the answer is easy, some might say. Early warning should be founded on the scientific attitude. It should rely upon scientific principles and scientific frameworks. Reason and reasoned criticism should guide early warning. Although I am personally quite friendly to this answer, and to the scientific attitude generally, I want to argue that it cannot save global early warning. There are two problems:
Perhaps even more fundamentally, the scientific attitude itself is not value-neutral. It embraces certain values, like parsimony or coherence. But what seems coherent from one set of experiences and in one culture may not seem so coherent from another set of experiences in another culture. At base, judgments of value in science, as in any domain of human experience, rest upon a theory of the good. 4 And theories of the good vary across human cultures and traditions.
None of this implies that truths, social truths or any other, are relative to human communities. Different human communities cherish different values, construct different concepts, hold different principles, operate under different frameworks, and arrive at different theories. But all these relativisms do not sum to render truth relative. To eliminate a confusion that arises whenever I express the heterodox views I express here, I do not subscribe to what Karl Popper once called the myth of the framework—the idea that, from the observation that cognitive constructs are relative across human communities, leaps to the conclusions that truth is relative across them, that communication is incommensurable across them, and that the goal of consensual scientific knowledge is pure chimera. The myth of the framework is a myth, Popper argued, because it falsely presumes that we cannot hold our frameworks up to critical scrutiny. 5 We can—and if we are to minimize self-delusion—we must.
I fear, however, that a centralized early warning system, driven by cost considerations, will privilege a single framework which it will not hold up to critical scrutiny. This is no irrational fear. A main task of any global early warning system will be to assess the stream of political events for any signs of impending political explosions. As I have detailed elsewhere, 6 applications of the scientific attitude to political event analysis have up to now consistently privileged one worldview over all others. They rely on Western news accounts, particularly The New York Times, for their source material. They rely on codings with high inter-coder reliability correlations, but only among coders with similar cultural backgrounds, and they presuppose in their methodological choices that political actors everywhere perceive the world just as the analysts do. Because political event analysts impose meanings on the event stream, they generate substantive conclusions that unsurprisingly reinforce their own political worldview.
In this regard, I like to tell a story concerning an event analysis conducted in the 1960s and described in the pages of Science. 7 Researchers at Stanford University wanted to devise a scale of the severity of international interactions. To that end, they placed on index cards statements describing various interactions with the states’ names removed (e.g., “Country X recalls its ambassador from Country Y,” “Country X moves troops to its border with Country Y,” etc.). They then instructed four graduate student coders to sort these statements into a normal distribution from the least severe to the most severe. Analyzing the results, they found that three of the codings correlated at about 0.90. One student, however, produced codings that correlated with the others at 0.75. So they disincluded these codings from their severity scale. They failed to mention, however, that the codings they retained were produced by white American males, while the codings they discarded were produced by an Egyptian female. 8
Such scientific practices are the rule, not the exception, in studies that analyze the stream of interstate political events. And they are far too partisan to serve as a basis for any sustainable system of global early warning. 9 They arise, of course, from the training of social scientists, which in the twentieth-century has suffered under certain positivist delusions concerning (a) the value-neutrality of data and methods, and (b) the uncritical acceptance of natural science as a model for social science. I should add that these ills do not infect only academic analysts. They also contaminate state intelligence analysts, and may in large measure be responsible for some colossal policy blunders. Unless we conceptualize global early warning right, we cannot expect it to prove any more reliable than the intelligence that state leaders receive.
To grasp the special problems of social analysis—problems that are not well understood even among social scientists—it is useful to reflect on Aristotle’s aitia, or the reasons for or causes of things. 10 Aristotle understood any event, depicted as X in Figure 1, to be the joint product of four distinct categories of cause.
Material cause refers to what today we might consider compositional constraints on the (efficient) causal agency of X. For Aristotle, material cause refers to “that out of which a thing comes to be and which persists..., e.g., the bronze of the statue, the silver of the bowl...”
Formal cause pertains to the shape, form, or concept of X. It is, for Aristotle, “the form or the archetype, i.e. the definition of the essence, and...the parts in the definition.”
Final cause refers to the contribution of the ends of purposeful agents in producing X, or, for Aristotle. “that for the sake of which a thing is done.” 11
The diagonal in Figure 1 demarcates the boundary between those aspects of Aristotelian causation that pertain to nature and those that pertain to human artifice. Modern natural science has been successful, even stunningly successful, at explaining natural events, concentrating (in Aristotelian terms) on efficient and material cause. Natural scientists need not concern themselves with final cause, because they lack any reason to inquire into the ends of purposeful agents. Natural science can also ignore formal cause, as they can impose by convention definitions and other forms on the non-human objects of their inquiries. Those objects never object. Some slippage may be found in psychology (partly a social science in any event) and in biology to the extent that it attributes purposeful agency to particular life-forms (it has by now rooted out the tendency to attribute purposeful agency to nature itself). But, by and large, natural science confines its explanations to efficient and material causes or reasons.
The social sciences or what Herbert Simon termed the “sciences of the artificial,” 12 must straddle this diagonal divide. One neither understands nor explains social phenomena without reference to the ends of purposeful agents (final cause). And one runs the risk of misrepresenting social phenomena to the extent that one’s impositions of conventional meanings, worldviews, frameworks, concepts, etc.—one’s attributions of formal cause—diverge from those of the social actors under study. To paraphrase Jürgen Habermas, the social sciences cannot impose meaning on the objects of their science. They must instead negotiate meaning with them, because the human objects of their science are themselves potential subjects of science.
In light of these considerations, descriptions of social science as “immature” mislead us. They wrongly presuppose that social science will one day grow up and acquire the attributes of natural science. Likewise, ascriptions of the natural sciences as the “hard sciences” and social sciences as the “soft sciences” are mistaken. It’s the social sciences that are hard. Relatively speaking, the natural sciences are easy. Due perhaps to unfortunate latent effects of positivism in their training and in the intellectual environment, many social scientists today buy into this category mistake. They try to fashion social inquiry entirely on the model of natural inquiry. They have physics envy. People have the right to be wrong, of course, and often enough this tendency can be overlooked in the interest of comity. But when social scientists borrow methods from the natural sciences and apply them uncritically to social phenomena, they can cause real harm.
To see how, consider the distinction in set theory between a set’s extension, or its membership, and it’s intension or the necessary and sufficient conditions for assigning an individual to the set’s extension. Because we can treat concepts as sets, they too have extensions and intensions. For instance, the concept “insurgency” has as its extension the population of insurgencies in human history. It also has an intension, or the necessary and sufficient conditions for assigning a social movement to the extension of the insurgency concept. 13 When observers do not share intensions, their ideas of concept’s extension sometimes diverge. For example, was the US civil rights movement an insurgency? They might think so in Luanda, but think not in Washington. Who’s right? And who’s to say who’s right? The trouble is that judgments regarding intensions (formal causes) vary across interpreters.
Most of the analytic modes thus far proposed for global early warning would apply statistical techniques in one way or another. This presents a problem. Statistical methods, through their reliance on probability theory, find their grounding in extensional set theory. Where intensions differ and effect differences in extensions, a statistical analysis becomes less rational than it might appear on the surface. It does not enlighten; it reproduces a worldview. It is not the reasoned application of scientific inquiry. It is the unreasonable (and usually unconscious) imposition of tacit dogma. Operationalization becomes a form of colonization. By excluding the codings of the Egyptian woman, for instance, the Stanford researchers did not construct an index of severity. They constructed an index of severity as white American males perceive it. Such social-scientific practices provide a secure foundation neither for theory construction nor for policy prescription.
The difficulty can be cast in terms of construct validity. A measure that operationalizes a social concept lacks construct validity to the extent that it relies on a meaning for the concept that is particular to one human community to the exclusion of others. More directly put, a singular interpretation of any social construct that takes plural interpretations underrepresents that construct. Results that use such measures do not generalize across human communities. And policymakers who use them to inform their decisions do so at great peril—not just to themselves, but to us all.
These comments should not be taken to mean that statistical modeling has no place in social science or in the social-scientific analyses that will inform much of global early warning. They mean only that it should be done right. Doing it right in the context of global early warning means recognizing that no analysis speaks for itself. Every analysis is informed a priori by a host of simplifying assumptions, model specifications, and construct operationalizations, each of which may be deeply controversial politically. Likewise, each analysis a posteriori informs an argument. It is a premise that supports or warrants an argumentative conclusion. But to the extent that the assumptions, specifications, and operationalizations of an analysis supporting an argument presuppose the argument’s conclusion, the argument is circular and thus fallacious.
Unfortunately, such circularities are often quite surreptitious. Analysts themselves do not see them because their assumptions, specifications, and operationalizations make sense to them. We should understand this as a natural by-product of the frame-dependency of mortal humans. We necessarily interpret reality, whether in a statistical model or not, from within the frameworks of our own experiences. We have no other interpretive aids. Consequently, we cannot reasonably ask individual analysts to step outside their interpretive frameworks—to see political reality the way others see it. We can, however, demand of analysts that they expose their models to the criticism of others—particularly others who do not necessarily share their picture of social and political reality. We already do this in our professional practices by presenting our work at conferences and by publishing our results. We ought to demand this also of analysts who would generate alerts to the Secretary–General.
The inappropriateness of imposing conventional meanings for many constructs, together with the relative bareness of our toolchest of methods for modeling the “formal cause” and “final cause” aspects of the reasons behind social events, render the social sciences incapable of erecting sustained, consensual, or even temporarily dominant paradigms. They are locked in a state of permanent scientific revolution. Across social science disciplines several paradigms co-exist, sometimes quite uncomfortably. Only recently have any efforts emerged to construct non-extensional methods appropriate for social analysis—methods that do not impose meaning but instead problematize it. 14 But these techniques are nowhere near ready for widespread application.
Given the inadequacy of existing techniques for analyzing social affairs, then, there is real danger in centralizing any global early warning system. The danger lies in its control by a single authority. Whether that central authority is the UN Secretary–General, a hegemonic state power, or even a committee of bureaucrats from a range of states, it will analyze world political realities from the perspective with which it is most comfortable. Other perspectives will be slighted. Consequently, the analyses and warnings it produces will favor some interests over others. Any centralized system of early warning would become an instrument of domination.
4. The Web as Alternative
Although I have argued against the wisdom of any centralized system, the idea of global early warning promises such enormous human benefits that alternative proposals ought to receive serious consideration. Independently, John Mallery (of MIT, speaking at a York University workshop) and Gottfried Meyer–Kress (in a University of Illinois working paper) 15 have proposed using the World–Wide Web to carry out the functions envisioned for global early warning. I believe the idea has much merit, for several reasons.
As a function of its distributed control, costs would also be distributed. Consequently, no petulant power could hold the system hostage by refusing to allocate funds for its maintenance. Likewise, no combination of states could terminate the system by a vote of the General Assembly.
Because the web is not state-based, other actors could participate in a web-based system. The UN’s “Agenda for Peace” envisioned a system by and for states and inter-governmental organizations. But a web-based system could also include NGOs, think tanks and universities, private corporations, religious groups, interest groups, and even individuals.
The wide range of participants ensures that the interests of the poor and powerless will find voice on the system. NGOs and religious organizations in particular, but others as well, can post statements from the powerless. They can serve as a conduit for messages from locations with limited or absent web access. They can even provide a venue for dissidents in oppressive regimes to warn anonymously of conditions on the ground that might not otherwise survive official communication filters.
Since web pages are retrievable as soon as they are posted, data, analyses, and warnings would be disseminated with great immediacy. The UN bureaucracy, on the other hand, would likely produce great delays.
The immediacy of transmission and breadth of dissemination of web based early warnings and analyses enhance the likelihood that a wider range of actors would take steps earlier to ameliorate potentially explosive conditions. Consumers of the intelligence would not be limited to the Secretary–General. State officials and non-state actors would also receive the information.
Due to its decentralization and the wide range of participating actors, analyses on a web-based system can readily be replicated with alternative data sources and alternative model specifications. And, due to the web’s immediacy, replications would more quickly appear.
Because data can be posted as well as analyses and warnings, parties with alternative data sources—possibly proprietary data or even state-secret data—could seamlessly insert their data into an analysis and re-run it to test whether the data change implied different results. They could then decide whether to release those data by posting them.
Certain technical challenges and political obstacles must be overcome in order to realize a decentralized web-based global early warning system. The technical challenges pose no threat to the system’s creation. They more accurately represent implementational opportunities designed to enhance the facility. The political obstacles, however, do threaten the system’s creation, and removing them will likely require concerted actions by non-state actors interested in promoting world peace and justice.
4.1. Technical Challenges
In particular, techniques should be developed for describing data matrices and model specifications, so that:
codebooks, when provided, need not get into details of data formats.
any participant’s statistical package can parse a matrix and analyze the model without the need to recode instructions specific to that package;
any participant can easily replace a data vector with one from another dataset before reestimating the model;
any participant can add variables and data not present in the initial analysis; and
any participant can respecify the functional form of any model.
Whether these are universal standards for description or meta-data protocols, 16 these techniques would shorten the time from the initial appearance of an analysis to its first replication under an alternative theoretical and/or construct conceptualization.
Some attention ought to be paid to the development of facilities to demystify analytic techniques for the educated, but not necessarily technically sophisticated, practitioners who would most use such a facility. For example, a web-based early warning system might, in collaboration perhaps with publishers, maintain an on-line corpus of narratives that explain a wide variety of techniques. An analyst could then hyperlink her analysis to those narratives. For example, an analysis that used, say, a vector autogression analysis could point to a vector autoregression page that in turn hyperlinks to a half-dozen or so explanations of the technique at various levels of technical sophistication and with appropriate citations to the literature.
One could go further. Facilities could be provided for political analysts to annotate political speeches with verbal comments as well as analyses that use the tools of linguistic pragmatics and argument analysis to uncover the deeper, contextualized meanings of the utterances of political figures. 17 Each analysis could itself incorporate annotations, themselves annotatable, that criticize, refine, or deepen the original analysis. What emerges from such a facility would be a global virtual conversation concerning the political ramifications of political speech.
Less technically ambitious, but more immediately useful to practitioners less technically trained, the facility could maintain scores of conversations that concern emerging problems of contemporaneous global concern. Such a facility would maintain conversations that vary across several dimensions:
across the quality dimension, these conversations would range from instantaneous postings in chat rooms open to white papers posted by states and other institutions, to fully peer-reviewed essays.
across the substantive issue dimension, these conversations would include postings about issues pertaining to politics, the economy, the environment, energy, agriculture, refugees, etc.
Within this multi-dimensional conversation environment, informal groups would naturally emerge that concern themselves with particular substantive issues in particular regions.
If designed poorly, a decentralized system could generate a Babel of voices over which meaningful and substantively important information would not be heard. I expect, however, that serious users of information will quickly distinguish the voices of reason from the voices of unreason. Moreover, by varying conversations across the quality dimension, users who wish to avoid the Babel could simply restrict their attention to a delimited range of conversations. Other users, deeply concerned about a particular issue in a particular region, could monitor the chat rooms to discern signals from the noise. When they find such signals, they might investigate further and possibly alert conversants in other, higher-quality conversations.
Once participants in any conversation come to widespread consensus concerning the existence and nature of an actual or potential calamity, they will naturally turn their attention to prescribing interventions. These interventions will doubtless be carried out by political institutions, but there is no reason that discourse over these issues should be confined to those institutions. The model of the conversation should prevail. Participants should engage themselves in reasoned argumentation over how best and who best to intercede. For this purpose, it would be helpful if a web-based early warning system included facilities to support joint strategic planning. 18 These facilities would allow a multiple users on the web to specify scenarios regarding the levels of flows of material, relationships between those flows, expert opinions on likely impacts of interventions, etc. With these models, users could ask counterfactual, or “what if” questions (e.g., “what if the Red Cross were to ship 25% of its medical supplies from country X to country Y?) and interrogate the model to assess the impact of this action. Others, viewing this model as it is posted on the web, could criticize its specifications or its data assumptions, reporting perhaps that the poster had overestimated the quantity of Red Cross supplies in country X, or noting perhaps that those supplies were needed elsewhere. Or perhaps a Red Cross official would annotate the model with an account of the logistical support needed to move those supplies in a timely manner. The general point is that such a facility would support timely, global, collective deliberation among experts and officials for the development of rational plans for addressing actual or potential disasters.
I can only sketch provisional thoughts about web-based global early warning here, of course. And not all of them are original, as many have been discussed in various forms with many colleagues in a variety of venues. I hope only that these ideas will spark criticism, where warranted, to improve them, and that others will generate additional ideas for web-based facilities.
4.2. Political Obstacles
The “free-rider problem” constitutes the main political obstacle to web-based early warning. Early warning is a public good. That is, because it would provide services from which all would benefit and none would be excluded, everyone has an incentive to “free-ride” on early warning—to enjoy the benefits it promises without contributing to the costs of its construction and maintenance. If too many perceive non-contribution to the costs of the system as in their self-interest, the system will never materialize. 19 Given this problem, how does web-based early warning get off the ground?
Some prospective contributors—notably the humanitarian NGOs—are unlikely to free-ride, as their actions arguably spring more from altruistic motives than from self-interested ones. But these organizations are hardly in a position to underwrite such a project. They themselves distribute public goods, for which they typically seek charitable contributions.
Overcoming free-ridership is a primary purpose of public institutions with coercive taxation powers. In the context of global early warning, the most obvious public institution for this purpose would be the global one—the United Nations. In an earlier exchange on web-based early warning, however, Ted Gurr reported that UN officials have responded coldly to the idea. 20 This is unsurprising, of course, since any short-term self-interested calculations of UN officials would suggest that they control the facility centrally in the Secretariat. But central control of early warning is not in their enlightened, long-term self-interest, because, as I have argued above, the UN cannot sustain this arrangement politically. Perhaps UN officials can be convinced that its institutional position with respect to maintaining an early warning system is structurally unsound.
Even if UN officials insist on their central control of early warning, they cannot prevent development of a decentralized, web-based alternative. Because they need the information and analyses, they will freely participate in any decentralized system created without their approval. Initial support may be provided instead by private foundations oriented toward the creation and maintenance of international peace and security. Those individuals and organizations who wish to engage themselves in the serious conceptual and implementational work needed to realize such a system should thus create proposals and solicit funds for this purpose. With relatively modest foundation support to underwrite the initial implementational work, other entities—such as private corporations, religious organizations, states, and even the UN—would then be encouraged to contribute sums regularly to sustain the effort.
A decentralized, web-based early warning would affect state sovereignty no less than would a UN-centered system. On this basis some will doubtless oppose the idea. However, the observation that state sovereignty is not conceptually rigid effectively counters it. The scope and meaning of state sovereignty for interstate relations and for security has changed dramatically in the 350 years since the Treaty of Westphalia. Contemporary strains on the notion of sovereignty flow largely from the major 20th century advances in transportation and telecommunications technology, of which the World–Wide Web and any web-centered early warning system are part. But these advances do not in any sense “threaten” state sovereignty. They merely represent the motive forces behind reconceptualizations of the notion of state sovereignty—reconceptualizations that states themselves foster as a consequence of their own self-interested actions.
5. Conclusion
I have proposed web-based global early warning as an alternative to centralized conceptions. I have argued that an early warning system centralized in the UN is politically unsustainable, and that any centralized early warning system is politically unwise. Whereas the UN conception appears to be founded upon the early warning practices of state intelligence agencies, the decentralized alternative proposed here is founded upon the notion of critical argumentation in the service of reasoned deliberation. It conceives early warning not as a technical activity to be conducted solely by bureaucrats in the UN Secretariat or elsewhere, but as a conversation designed to engage people globally—from all locations and from all perspectives—in a forum in which we can share information about contemporary conditions, deliberate over the meaning of these conditions, and come to reasoned consensus over rational courses of collective action. Ironically, these are the very purposes for which the United Nations was established.
Endotes
Note 1: United Nations, General Assembly. 1992. “An Agenda for Peace: Preventive Diplomacy and Related Matters.” A/RES/47/120. 91st plenary meeting. December 18,1992. Back.
Note 2: Available at http://www.un.org/Docs/SG/SG-Rpt/. Back.
Note 3: Cf. Paul D. Johnston, “No Cloak and Dagger Required: Intelligence Support to UN Peacekeeping.” Intelligence and National Security. 12 (1997). Back.
Note 4: Cf. Hilary Putnam, “Beyond the Fact/Value Dichotomy,” in Realism with a Human Face. Cambridge: Harvard University Press, 1990, pp. 135–141. Back.
Note 5: Karl Popper, “Normal Science and Its Dangers.” In Criticism and the Growth of Knowledge. Imre Lakatos and Alan Musgrave, eds. Cambridge: Cambridge University Press, pp. 51–58. Back.
Note 6: Gavan Duffy, “Events and Versions: Reconstructing Event Data Analysis.” International Interactions. 20 (1994)147–167 Back.
Note 7: Moses, Lincoln, Richard A. Brody, Ole R. Holsti, Joseph B. Kadane, and Jeffrey S. Milstein. “Scaling Data on Inter-Nation Actions.” Science. 156 (1967)1054–1059. Back.
Note 8: Personal communication with Nazli Choucri, the coder whose sorts were discarded. Choucri is now a professor of political science at MIT. Back.
Note 9: A singular exception is the recent work of Peter Brecke, formerly of ORCI, once Choucri’s student, and now at Georgia Tech. Brecke’s analyses show sensitivity to multiple perspectives. See, for example, his use of multi-cultural event accounts in “Defining the Dependent Variable For Conflict Early Warning: A Taxonomy of Violent Conflicts.” Paper presented to the 1997 Meetings of the International Studies Association. Toronto. March 1997. Back.
Note 10: I here extend a point advanced in Hayward R. Alker, Rediscoveries and Reformulations: Humanistic Methodologies for International Studies. Cambridge: Cambridge University Press, 1996, p. 74. Back.
Note 11: Aristotle, Physics. 11.3. Quotations from The Complete Works of Aristotle. Jonathan Barnes, trans. Princeton: Princeton University Press, 1984. Back.
Note 12: Herbert A. Simon, The Sciences of the Artificial. Cambridge: MIT Press, 1969. Back.
Note 13: “Social movement” here is considered the genera of which “insurgency” is a species. The necessary and sufficient conditions might be supplemented by a list of ceteris paribus conditions to obviate a central problem in Wittgenstein’s Philosophical Investigations. Observers might differ over the conditions they take to be necessary and sufficient and those they believe to be ceteris paribus exceptions. This is equivalent to the contention that intensions differ across observers, which is just my point. Back.
Note 14: For examples, see the essays and citations in Part One of Alker, op. cit., and Gavan Duffy, Brian Frederking, and Seth A. Tucker, “Language Games: Dialogical Analysis of INF Negotiations.” International Studies Quarterly, 42 (1998): 271–294. Back.
Note 15: Gottfried Mayer–Kress and C. Barczys. “The Global Brain as an Emergent Structure from the Worldwide Computing Network, and its Implications for Modelling. The Information Society. 11(1995), to appear. Also available as Technical Report CCSR–94–22, Center for Complex Systems Research, Beckman Institute, University of Illinois, Urbana–Champaign. See also Gottfried Mayer–Kress, Gottfried, B. Bender, and J. Bazik. “Hyper-Media on the Internet as a Tool for Approaching Global Problems: A Tele-Conferencing Experiment.” Code No 94–0023. Unpublished Scholarly Papers, Information and Communication, CIESIN Human Dimensions Kiosk, http: / /www.ciesin.org/kiosk/home.html. Presented via the Internet and telephone at the Hunger Research Briefing and Exchange, Brown University, April, 1994. Also available as Technical Report CCSR–94–6, Center for Complex Systems Research, Beckman Institute, University of Illinois, Urbana–Champaign. Back.
Note 16: Needing to integrate data seamlessly from multiple national sources, the European Community has placed some emphasis on developing meta-data protocols. See http://idaresa.univie.ac.at/. Back.
Note 17: John Wilson, Politically Speaking. Oxford: Basil–Blackwell, 1990; Duffy et al. op. cit.; Alker, op. cit. Back.
Note 18: Several implemented systems that support joint decision-making and planning might be appropriate for exporting to the web. One is the system described in John G. Harhen, Knowledge Based Modelling to support the Determination of Manufacturing Strategy. Ph.D. dissertation. University of Massachusetts. 1990. The “Analytic Hierarchy Process,” described at http://www.ahp.net/, is another. Back.
Note 19: Cf. Mancur Olson, The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge: Harvard University Press, 1965. Back.
Note 20: Gavan Duffy, Ted Robert Gurr, Phillip A. Schrodt, Gottfried Meyer–Kress, and Peter Brecke, “Forum—A UN Early Warning System: Internet or Not?” Mershon Review of International Studies. 39 (1995): 315–326. Back.
Email this citation