Organizational Learning Curves, Angles, and Lines: Recognizing Diverse Types of Learning in the United Nations’ Humanitarian Operations

Nancy E. Wright
Political Science Program
Graduate Center
The City University of New York

E-mail:  nwright1014@yahoo.com

The 1990s marked the beginning of an ongoing era in which the UN received substantial criticism for making poor choices and circumventing its responsibilities within the normative framework in which it was created. [1]   This seemingly flawed performance has stood in glaring contrast to an international system that, in the absence of the Cold War and corresponding recurring Security Council gridlock, appeared more receptive to multilateral approaches to meeting humanitarian needs than at any point in history since the UN’s founding.

In the aftermath of disappointments, failures, and criticism from donor governments, humanitarian and human rights, and development non-governmental organizations (NGOs), the UN’s humanitarian organizations such as UNHCR  have commissioned numerous evaluations and undertaken documentation and analyses of “lessons learned” in Somalia, Bosnia, Rwanda, Kosovo, Afghanistan, and other countries embroiled in internal violent conflict and struggling with complex emergencies. [2]

What, if anything, are humanitarian IGOs learning from the experiences of the first post-Cold War decade?  Some analysis would argue that UN humanitarian operations are repeating their mistakes.  This paper argues that the UN’s humanitarian organizations are learning, and in fact have learned a great deal from the changing nature of conflict in the 1990s.  Existing theories of IGO learning, however, do not sufficiently account for this learning, because these theories were developed to examine IGO learning in the absence of violent internal conflict, during which an IGO’s work, namely humanitarian assistance, assumes a strategic value.  In contrast to many other areas of the UN”s work, the extremely high normative significance associated with humanitarian assistance frequently means that in its learning process an IGO confronts struggles and contradictions between what its knowledge acquisition indicates can be done and what it understands should be done.  Exacerbating this condition is the fact that IGOs, unlike NGOs, are accountable and intertwined with states, including state interests, both by virtue of comprising part of the UN system and depending on states’ voluntary contributions.

Overall, learning among international, intergovernmental organizations has not received extensive scholarly attention. [3]   Existing analyses tend to focus on environmental or other technical and scientific issue areas in which an IGO during relative peacetime receives and assimilates new information from groups of experts, such that the organization either adapts or transforms its policy direction and future activities as a result. [4] Similarly, scholarship on learning among IGOs—especially IGOs working in the midst of violent conflict—is relatively sparse but growing, in the aftermath of post-Cold War conflict, as noted previously.   Analyses of learning in the humanitarian sector have increased, especially under the leadership of such organizations as the Active Learning Network for Accountability and Performance in Humanitarian Action (ALNAP). [5] Literature on the role of norms and ideas in shaping policy is abundant, and specific examples are identified in this paper’s section on normative learning.  These works, however, concentrate on the influence of norms and ideas on states rather than on intergovernmental organizations such as the UN and its various agencies. 

Another major body of literature on organizational learning is oriented primarily toward the private sector, with a relatively small selection of works targeted to the learning behavior of non-profit and not-for-profit organizations.  These classic works on organizational learning focus on how individuals within organizations acquire and apply knowledge to behavior, thus shaping the organization itself. [6]   Within this body of literature works on humanitarian sector learning are extremely limited. [7]   These works tend to emphasize the relationship between individual and collective learning, obstacles to learning, and recommendations for creating and improving the organizational learning environment.

The framework presented here is distinct in two ways from the organizational learning literature noted above.  First, it specifically addresses IGOs, as opposed to the private sector or, in a few cases, not-for-profit or non-profit organizations.  Second, and more important for the present state of IGO learning in the humanitarian sector, this framework is designed to analyze what IGOs learn more than how they learn.

This analytical framework synthesizes normative and ideational theory and the concept of organizational learning.  This paper argues that norms can have a uniquely significant—and often confusing—effect on an organization’s understanding of what it should learn.   Inherent in evaluations of humanitarian operations is a particular tension between expectation and performance.  Unlike the usual tension in which expectations measure performance to the best that is deemed feasible, expectations related to norms measure performance to what is deemed morally and ethically acceptable.  As a result, feasibility, especially feasibility in the context of the situation at hand or vis-à-vis performance of other organizations within a given system, is only marginally taken into account.    An organization’s failure to perform in accordance with standards of moral and ethical appropriateness is highlighted as an abdication of responsibility, an inexcusable lapse of conscience.  The possibility that the organization simply might not be capable of fulfilling such responsibility in the context in which it is being required to work is often not adequately considered.  Instead, evaluators seek to avoid repeating past mistakes, thereby assuming that failure is the result of a mistake in the first place.  Furthermore, even if the mistake is correctly identified as such, evaluators also assume that it is and will be a mistake in the future as well, hence the concept of “lessons learned.”

The organizational learning framework presented here assumes that at some level learning takes place, both in preparation for and as a result of organizational activity, even if there is no evident application of that learning in future endeavors.  From this assumption immediately arises the question, what is the proof of learning in the absence of that future application?  The answer lies in understanding a more inclusive definition of learning.

What Constitutes learning?

Learning can be fundamentally defined as acquisition of knowledge.  Most analysts of learning, however, would add to that definition the application of knowledge acquired. [8] This applies especially to studies of organizational learning, where without evidence of knowledge application, gauging learning is extremely difficult.  This study acknowledges that acquisition of knowledge, even without application, constitutes a degree of learning.  As will be noted throughout this article, this framework acknowledges this and other application-oriented definitions of learning.  However, it also accounts for the possibility of acquisition of knowledge without application of that knowledge.  In other words, by expanding the definition of learning, non-application is not necessarily synonymous with failure or refusal to learn. [9]   Learning defined as acquisition of knowledge is evident in stated recognition and perceptions of issues, events, and situations. For example, if the leader of a humanitarian organization receives and assimilates information about a complex emergency, s/he has learned at a basic level.  If the leader then imparts information and gives a directive to other members of the organization, that membership –and therefore the organization itself—has also learned. 

Linking acquisition and application yields three additional levels of learning: (1) learning as knowledge acquisition with the intent of application, whether or not application is realized; (2) learning as knowledge acquisition and application; and (3) learning as knowledge acquisition, followed by rejection of application. [10]   The first and third levels pose a fundamental methodological problem, namely how can one determine intent, and how can one determine whether rejection of application is deliberate, unless organizational leaders explicitly state their intentions?  Yet if one limits identification of the first and third levels to cases where intentions are explicitly identified, how can one account for cases where they are not?  As subsequent paragraphs explain, this study focuses on three variables—mandates, resource allocations and evaluations—as both inputs and outputs of learning. The framework presented here is limited to intent to apply and rejection of application as reflected in changes in mandate, resource allocation, or evaluations of prior performance.

In addition to enlarging the scope of what constitutes learning, this framework rests on the premise that organizations engage in several types of learning, and that at times certain types may supersede or even obfuscate others.  This paper presents IGO learning not as a linear process, or even as a sequence with interruptions of reversal and circuitous paths.  Rather, it presents a framework that accounts for different types of learning and for the relationship among those types.  It is designed to identify not only the ways in which IGOs learn or do not learn, but also to illustrate how purposive learning of one type may actually obfuscate, pre-empt, or displace other forms.

This study views the state as an important but not exclusive actor.  Although constrained to a significant degree by states, IGOs do enjoy considerable autonomy in shaping international coordination. At the same time, this paper does not altogether agree with the social constructivist school, which views state interests as constructed from without and therefore also more subject to external influence by international organizations and even individuals. [11]   Rather, state interests greatly influence the capabilities and policy direction of IGOs, especially those that depend on voluntary contributions from states as donors.  At the same time, within those substantial constraints, leaders of IGOs nevertheless may exercise a range of options, and the choices they make both have consequences for their intended beneficiaries and influence choices of the very donor states on which they depend.

Diverse Forms of Learning in International Organizations

Political Learning

Political learning is defined here as the recognition of opportunities and constraints resulting from the capabilities and limits of individuals and of the organizations within which these individuals make and carry out decisions.   Key components are bounded rationality, incomplete and/or inaccurate information, political self-interest, and, for the United Nations system, dependence on member states. 

Works on bureaucratic and individual cognitive and other psychological constraints proliferate, especially in the study of public policy, both foreign and domestic. [12]   These analyses identify two levels of constraints, primarily on key individual decision-makers.  One is that posed by the inevitability of incomplete and/or inaccurate information, insufficient time and other resources to make and carry out informed decisions, and lack of cooperation on the part of other actors. The other level of constraint is human fallibility, including the failure to consider all options that are available, the refusal to acknowledge external limits, and ideologically driven insistence on certain preferences, regardless of their consequences.

Organizational leadership that engages in political learning recognizes these constraints and endeavors to offset them wherever possible.  Examples are applying sensitivity to timing when introducing controversial initiatives, garnering support for measures before presenting them to a larger forum, and employing the skills of a “gadfly” to counter what Irving Janis long ago labeled “groupthink.” [13]   Optimal political learning means mastery of what Stone has identified as “the art of political reasoning” and what George has called “political rationality.” [14]

For the United Nations, many key elements of political learning have been  encompassed in the concept of “political will.”  The organization has been viewed—and has viewed itself—as first and foremost a membership body of sovereign states, and therefore a body with little or no effective enforcement authority.  The superpower gridlock of the Cold War era consistently underscored the apparent lack of UN efficacy.  In addition, harsh criticisms of incompetence and ineffective management, especially by the United States during the 1980s, reinforced the idea of an organization severely constrained, if not at times paralyzed, by its own ills.    Efforts to respond on the UN’s behalf included a call for the organization to adopt a “comparative advantage” approach, that is, to be more selective in the types of problems to which it would respond. [15]   According to these critics, only those issues inherently international in scope and character, e.g., global warming, use of outer space, should be the UN’s primary concern.   All other matters should be left to states or regional bodies, which had the resources and the desire to address them, or to NGOs, which had the organizational and political flexibility to deliver services without bureaucratic encumbrance. [16]

The preoccupation with political will assumes the principle of state sovereignty as a given within the realm of political learning.  Absolute sovereignty may be compromised, and a measure of its corresponding power may be ceded, but only with the state’s acquiescence, thus recognizing the state as the preeminent source of power and authority in the international system.  (As subsequent paragraphs will explain, normative learning, in contrast, views sovereignty not as an absolute that may be altered only by state consent, but rather as a principle that may be challenged by other principles of international law.)

Yet, as noted previously, this analysis holds that humanitarian IGOs like UNHCR also embody dimensions of power and authority, despite their dependence on individual states.  In fact, their autonomy is twofold—the authority and power of the individual leader of the organization and the organization itself; in the case of UNHCR, “both an individual, represented in the High Commissioner, and a bureaucracy with its own distinct culture and value system.” [17]     These sources of authority function interdependently in ways that either enhance or undermine learning.  In the case of the latter,  lessons identified with the intention of application may be lost before they can be re-applied, precisely because of constraints imposed by those in control of resources.  For example, lessons about preparedness resulting from the tragedy of Rwanda were identified but not realized in Kosovo, because UNHCR experienced financial cutbacks between the two conflicts that prevented it from having much needed resources. [18]

This frequent scenario points to three limits to the application of  political learning by humanitarian and other IGOs dependent on donor states.  The first is the limited autonomy and authority of the organization itself.  Regardless of how well an organization like UNHCR might “learn its lessons,” its ability to apply those lessons to the future is  limited when such application exceeds existing levels of financial support .  Second is the reverse of the first, namely the limit of an IGO’s ability to succeed in garnering support from its donor states.  The third limitation is an IGO’s inability to at least partially extricate itself from donor states, either by seeking and obtaining alternative sources of funding or by identifying ways to apply lessons learned even in the absence of additional resources.

The notion of shared responsibility for performance failure or success points to the dual identity of agent and structure among the actors involved.  States are agents, especially in their roles as donors, but they are also structures that together comprise UN membership. [19]   UNHCR, embodied both in its leaders and as an institution unto itself, is an agent in its capacity as administrator of humanitarian relief, and it is also a structure to which agent states respond and in which they have input.  The UN Secretary-General, to whom UNHCR reports and from whom it takes direction, is also both agent and structure, as the one who issues directives and whose office is the authority of the UN structure.  Thus a major challenge in political learning for IGOs is not only recognizing bureaucratic and international constraints, but also distinguishing between which must be accepted and which can be overcome by exercising internal organizational autonomy.

Operational Learning

Operational learning is the acquisition of knowledge regarding how effectively to carry out policies and programs that are the products of the other learning types.  Operational learning consistently asks the following key questions:  what worked, what did not work and why, and how can performance be improved? 

Operational learning—or at least the attempt at operational learning—tends to increase dramatically in the aftermath of extreme success, and even more dramatically in the aftermath of extreme failure.  The desire to repeat successful operations is strong but more short-lived; however, the reflection on failure, though agonizing, tends to persist as the organization tries desperately not to repeat past mistakes.  Ironically, this preoccupation with the past often leads to new mistakes in the future, or the avoidance of repeating something that may have been a mistake in the past, but is a sound application for the future. 

Of the five types presented here, operational learning tends to be more isolated from the other types, due to the fact that it focuses on programs and tasks, rather than on more comprehensive policies.  It makes concerted efforts (however unsuccessfully) to dissociate from political elements that supposedly interfere with effectiveness, it focuses little on either context or relations with other entities (unless a jointly administered program is being evaluated), and its norms are deeply embedded and rarely questioned.  In other words, for the operational learner, the quest is to assess not the mission or the principles underlying the mission, but rather how well the organization carried out its responsibilities to fulfill that mission.

Perhaps the most puzzling and frustrating dilemma of operational learning is the failure of lessons identified and apparently learned to be applied in subsequent cases.  Astri Suhrke [20]   examines this trend in her comparison of UNHCR’s emergency responses in Northern Iraq and Kosovo, respectively.  Drawing upon Jack Levy’s definition of learning as “’development of new beliefs, skills or procedures as a result of the observation and interpretation of experience’” (noted earlier in this article) [21]   Suhrke notes the importance of distinguishing between the acquisition of new beliefs or skills and the application of those beliefs or skills to policy. [22] She asks “whether weaknesses in the agency’s response to the first crisis had been reduced in the second crisis, and, if so, whether this was due to a change in beliefs, skills, and procedure that at least partly reflected the earlier experiences.” [23]

The standard of success for operational learning is the extent to which the measure accomplishes the intended objectives; the standard of failure is the extent to which the measure fails to accomplish them.  Accordingly, operational learning is limited in several ways.  One is the limited capacity of the IGO to carry out stated objectives as a result of inherent constraints within the organization.  The second consists of limits imposed by ambiguities in identifying objectives.

Contextual Learning  

Contextual learning is defined here as the acquisition of knowledge about a particular event or situation and its surrounding elements, including recognition of similarities and differences between that event or situation and others.  Key components of contextual learning are immediate realities and larger surroundings.  Contextual learning, or assimilation and application of knowledge about the external environment in which an individual or organization is operating, consists of two levels.  One is the immediate context, usually characterized by a combination of events that prompt action by the organization and existing organizational realities, including available resources, and perception of the events at hand. The other is the larger environment, including past as well as present realities and perceptions.  For the UN, this means the larger international political order and changes within it, as well as its own past and present role within that order.

Several now well-documented characteristics are prevalent in the post-Cold War era, many of which are the source of continued debate.  One such characteristic is the extent to which the collapse of the Soviet Union has led to a more multipolar world with less superpower rivalry than that experienced during the Cold War.  The end of U.S.-Soviet gridlock in the UN Security Council at first seemed to signify an era that was more conducive to multilateral cooperation.

A number of leading realist scholars, however, have offered convincing challenges to this vision, especially in the wake of post-Cold War violence.   These include traditional realists whose foundation is the work of Morgenthau, scholars of offensive realism, such as Mearsheimer, and defensive realist theorists such as Waltz and his contemporaries. [24]]

The second characteristic is a particular dimension of what has been dubbed the “CNN effect,” of media coverage of conflicts such as those in Somalia, Bosnia and Rwanda, namely the entry of the media as a policy actor. [25]   The extent to which the media have inserted themselves into the policy process is debatable; however, whether real or perceived, it is a contextual element that has contributed to the difficulty in policy-makers’ ability to act without having to account for media interpretations of events and their implicit recommendations regarding the appropriate policy response. [26]   Faced with vaguely defined conflict and foreign policy direction, journalists have had the ability to exercise newfound latitude by presenting their own interpretations to the public. [27]  

A third complicating element is the confusion regarding the similarity and difference in post-Cold War conflicts.  The belief that every conflict is unique has been challenged by some experts in humanitarian assistance operations. [28]   Still, there is equal danger in proceeding on the premise that a template for humanitarian response to internal conflict exists.  The challenge for humanitarian IGOs has been to determine to what extent lessons identified from past conflicts apply to future conflicts that may resemble their predecessors on the surface, but in fact embody very different dimensions.

One difference is the transition away from traditional peacekeeping toward more comprehensive efforts, including protection of civilians   Another is the dramatic growth in numbers of internally displaced persons (IDPs) and the UN Secretary-General’s call to address their needs, including the appointment of a Special Representative on Internal Displacement.  Some changes are directly attributed to the end of the Cold War, most notably the elimination of the U.S.-Soviet veto gridlock in the Security Council.  Others occurred after the Soviet Union’s demise, but cannot with certainty be attributed to the end of the Cold War, though an association is suggested.  These include increased targeting of civilians, often with the ultimate goal not of conquest but of obliteration,  and corresponding ambiguity in identifying enemies and allies, especially in internal conflicts, increased military involvement in humanitarian relief efforts, and challenges to traditional interpretations of state sovereignty. [29]

A major limit to contextual learning is the inability of IGOs to modify their acquired knowledge to the particular situation at hand. Reliance on models and standard operating procedures can lead to organizational behavior that is irrelevant and often detrimental.  Similar obstacles can emerge from limits to relational learning, described in the next section.

Relational Learning

Relational learning is defined here as the acquisition of knowledge about the relationship of one organization to another, including impacts that one organization’s changes in mandate have on the work and perceptions of the other organization.   Relational learning is distinguished from contextual learning and political learning by the fact that, although it shares the characteristics of respectively comprehending external environments and bureaucratic structures and processes, relational learning focuses specifically on the relationship of one organization to another organization within the same system.  Singling out this relationship becomes important when examining the perceived and actual impacts of mandate change, especially mandate expansion, to areas previously addressed by another organization.

Relational learning becomes especially salient—and more distinguished from political learning—when one or more organizations within the same system undergo mandate expansion.  A key example during the conflicts of the 1990s is the decision by humanitarian organizations to include in their mandates post-conflict reconstruction.  The acceptance of responsibility toward societies emerging from conflict has evolved as part of efforts by humanitarian and development organizations alike to “bridge the relief-to-development gap.” [30]   Humanitarian organizations traditionally responsible for prompt delivery of emergency provisions have recently carried out more long-term development initiatives to support fragile societies emerging from conflict. [31]

A major limit to relational learning is present when the organizations are not learning concurrently, or are experiencing learning types that are antithetical to each other.  Under these conditions relational learning is not likely to be successful.

Normative Learning

Normative learning as defined here is the acquisition of knowledge about collective values and principles, i.e., the essential elements of norms themselves, that may influence policy direction, including acceptance, rejection, and selection of some norms over others, as well as confusion and conflict with respect to norm selection [32] .  Key components of normative learning are existing norms, emerging norms and norms in contention.  Normative learning differs from political, operational, contextual, and relational learning in three related ways.  First, it is driven less by acquisition of knowledge and more by beliefs and values about what is known.  Second, normative learning centers on decisions and behavior according to what should be done rather than what can be done.  Other types of learning are concerned with capabilities and effectiveness; normative learning is concerned with principles.  Third, because the sources of norms are not easily identified, it is accordingly difficult to examine the normative learning process.   While this analysis maintains that beliefs and values are present in all types of learning, normative learning is distinguished by the fact that beliefs and values actually drive the learning process.

Existing normative and ideational theories have yet to be able to explain specific choices or preferences of decision-makers.  One can point to normative influences in areas like international human rights protection and humanitarian assistance; however, the inconsistent nature of activity in these areas, along with the frequent joining of human rights and humanitarian initiatives with other goals, such as alliance-building, make it difficult to determine where one source of motivation stops and another begins. Yet norms have and continue to play an important role in international relations.  They are used to pressure governments to change their behavior, to affirm the virtues of some governments and to criticize others, and to build a transnational consensus and support based on principles that apply to all humans, regardless of nationality. [33]

The genesis and development of norms themselves are not clearly understood.  Martha Finnemore has examined the phenomenon of norm entrepreneurs, and Finnemore and Kathryn Sikkink together have developed an analytical framework that identifies norm “tipping points,” or thresholds where norms take on a new dimension of significance and influence in international behavior. [34]

Ann-Marie Clark has examined at length the influence of Amnesty International on the development, proliferation, and influence of human rights norms, even as challenges to state sovereignty.  She contends that

Existing works on norms do not explain why states adhere to some norms and not others.  Equally relevant to this study, these scholars do not explain why IGOs may emphasize one set of norms over another.  Herein is yet another question:  where is the locus of authority about norm decisions?  That is, if institutions are frameworks for norm development, does norm influence depend more on states or on international institutional, e.g., UN leaders? 

Clark asks, if human rights norms are understood as common societal rules, one is led to ask whether new norms might be traced to the composition of international institutions or to changes in their structure.   She explores the possibility that the structure of the UN itself might be responsible for facilitating and fostering norm generation.  A second potential genesis of norms is change in UN membership over time. [36]  

Each of these explanations has drawbacks, however.  The Charter allows only limited involvement by NGOs, including human rights advocates, and explicitly calls only for the “promotion” of human rights, with no enforcement mechanism.  Regarding evolving UN membership, Clark notes that while the addition of newly independent states as UN members following decolonization certainly augmented pressure for such measures as awarding self-determination and ending apartheid, a more widespread human rights accountability was not forthcoming from that same membership.

Eschewing state-centric explanations for norm emergence, Clark also argues that the possibility of the conditioning normative influence of “both formal and informal institutions” and non-state actors is overlooked, as scholars instead view institutions as existing only in service of state interests. [37]   What is needed,” she asserts, “is a deeper understanding that accounts for how norms gain authority and how normative authority interacts with the motives of state and nonstate actors.” [38] She argues that norms transcend shared understandings of socially shared interpretations of behavior; rather, continuity and change depend on a more deeply rooted interpretation. [39]   In this way she departs from scholars such as Audie Klotz, who attributes the eradication of apartheid to “shared (thus social) understandings of standards and behavior.” [40]   Moreover, the degree of state compliance with norms is not necessarily a reliable indicator of norm efficacy.  Rather, “whether a specific action supports a norm depends in large part on how it is interpreted, justified, and criticized by actors and observers.” [41]

Ideational theory includes the analysis not only of norms but also of ideas that are not directly traced to moral or ethical foundations.  Defining ideas as “beliefs held by individuals.” [42]  Judith Goldstein and Robert Keohane differentiate among three types of beliefs, namely “world views, principled beliefs, and causal beliefs.” [43]   World views, “conceptions of possibility,” are “embedded in the symbolism of a culture and deeply affect modes of thought and discourse.” [44]   They are not purely normative since they include views about cosmology and ontology, as well as about ethics.  Nevertheless, world views are entwined with people’s conceptions of their identities, evoking deep emotions and loyalties. [45] The second type of ideas, principled beliefs, consist of “normative ideas that specify criteria for distinguishing right from wrong and just from unjust.” [46]   The authors explain that principled beliefs take definite stances regarding issues and practices.  World views, in contrast, can encompass diverse and even dichotomous vantage points. [47]   Goldstein and Keohane’s third category is causal beliefs, or “beliefs about cause-effect relationships which derive authority from the shared consensus of recognized elites, whether they be village elders or scientists at elite institutions.” [48]  The cornerstone of causal beliefs is the concept of a strategy for goal achievement, usually grounded in scientific or another form of expertise. [49]   The way in which causal beliefs are conceived changes more frequently than conceptualization of world views or principled beliefs.  Unlike Finnemore, Clark, and other normative theorists, Goldstein and Keohane do address the question of how policy makers choose ideas.  To this end, they caution that innate qualities of ideas alone cannot explain policy makers’ preferences for one idea over another. [50]   Rather, interest still drives choice, but ideas, often deeply embedded, determine the pathway of interest fulfillment. [51]

The studies noted above have contributed much to IR theory by identifying and illustrating alternative sources of power to the state and by delineating the relationship between scientific knowledge and decision-makers within IGOs.  In this way, they pose a noteworthy challenge to classical realism and neo-realism’s state-centric assumptions. 

The scholarship on epistemic communities noted earlier, while not devoid of normative elements, nevertheless focuses primarily on the role of experts in clarifying complex issues so that leaders of IGOs can make more informed decisions.  The normative dimension is present in efforts by epistemic communities to intervene in negotiations and direct them to a desired outcome, e.g., environmental protection; however, normative influence is not examined as a separate phenomenon; rather it is incorporated into the justification of policy preference on the basis of scientific evidence.  For example, to the extent that epistemic communities influenced the successful conclusion of the 1987 Ozone Protocol in Montreal, they did so by presenting evidence that failure to act would likely have dire consequences for air quality.  The norm of environmental protection—that is, the belief that governments, acting through international institutions such as the UN, have responsibility to protect the environment—was assumed.  Furthermore, the assumed norm constituted a minimal source of contention, because political and economic security stakes were relatively low. [52]

The studies just presented have contributed to date little toward understanding organizational learning in situations of conflict and chaos. [53]   Moreover, the cases in which these scholars consider learning to be most complete feature those organizations of the UN system, e.g., the World Bank, that enjoy the greatest degree of organizational autonomy. [54]

The previous sections on political, operational, contextual, and relational learning respectively have closed with an identification of limits or challenges inherent in each learning type.  The limits on normative learning differ, however, because they encompass both ambiguity and a seemingly no-win situation.  With respect to the former, as noted in the summary of scholarship on norms and ideas, neither the origin nor the selection of norms is clear.  Therefore, an IGO cannot justify its norm selection on the premise that, based on past experience it is more feasible, or that it is the best that can be done.  With respect to the latter,  satisficing according to constraints on capabilities, whether real or perceived, suggests an abdication of moral responsibility; yet acting on normative grounds without the resources to fulfill the mission at hand often results in  performance failure, with dire consequences for the intended beneficiaries.

Unlike the limits associated with political, operational, contextual, and relational learning, limits associated with normative learning tend to be deemed unacceptable.  Partial fulfillment of “good” or constructive norms, such as assistance to refugees or protection of human rights, is considered a compromise of principle. Similarly, the complete eradication of destructive, or “bad” norms, such as those underlying slavery, torture, or genocide, is considered equally necessary. To be content with incremental success is to turn one’s back on the realization of humanitarian and human rights principles.  Thus the standard of successful normative learning is the fulfillment of ethical responsibilities; non-fulfillment signifies failure.

All five learning types influence each other, sometimes by occurring together, other times by occurring apart from each other but nevertheless affecting each other.  Normative learning, however, plays a unique role in contrast to the others, because normative learning usually involves a higher priority on principles than on feasibility or even appropriateness to the situation.  At the same time, three characteristics of normative learning can complicate how norms are actually assimilated.  These are norm periodization, norm ambiguity, and norm conflict.

Norm Periodization

Events often lead to norm emergence, norm debate, and norm re-appraisal.  An established norm that drives decision-making by a state or an organization can be reconsidered and even rejected in the aftermath of disaster.  Norm periodization occurs on two levels.  The first is strengthening and weakening of norms according to the evolution of decisions and events.  The second is the overall salience of norms once sequences of events have reached closure. 

Such is the legacy of humanitarian intervention in Somalia, driven initially by a strong norm to intervene on humanitarian grounds that subsequently weakened and was superceded by a normative learning that resulted in inaction. [54]   The impact of the Somalia case on the reticence of states and the UN alike vis-à-vis Rwanda cannot be overestimated.  What if the Somalia debacle had occurred after the genocide in Rwanda?  Would the United States have been so hesitant to act?  Would the United Nations Secretary-General and Security Council have exercised similar hesitancy, with or without the support of the United States?  Would humanitarian assistance have been administered in a manner that helped lead to such tragedies as Goma? [55]

Norm periodization can lead to decisions and actions that are not necessarily commensurate with the situation to which those decisions and actions are being applied.  In fact, perception of the influencing event itself is not necessarily congruent with reality.  Ironically, Somalia has been equated both with a reluctance or refusal to intervene in future cases on the one hand and with missed opportunity on the other.  Reflecting on Somalia, political leaders could have resolved to act even more quickly in Rwanda; instead, Somalia had the opposite effect.  The importance of phenomena like those in Somalia for this learning framework is the extent to which the associated decisions and actions, themselves the products of different types of learning, in turn led with respect to Rwanda to a re-appraisal of the norm to intervene on humanitarian grounds, and especially in response to either acts of genocide or genocide itself.

Norm Ambiguity

In June 1992 the Secretary-General issued “An Agenda for Peace:  Preventive Diplomacy, Peacemaking, and Peace-keeping,” at the request of the Security Council. [57]  The Agenda was developed on the premise that

The Agenda for Peace described in detail what the Secretary-General proclaimed as “the changing context” of the post-Cold War era.  While certain fundamental changes certainly did occur, as was addressed in the section on contextual learning, certain features of the Cold War era and even decades preceding World War II remained.  Therefore the proclamation by the Security Council, the Secretary-General, and others of a new age of efficacy for the UN stemmed as much from a normative as from a perceptive or analytical basis.

The Secretary-General further asserted

At the same time,

Thus the Secretary-General upheld the norm of state sovereignty, but only on the condition that states would be willing to cede a certain degree of sovereignty in the interest of international peace and security, with respect and promotion of human rights and social and economic well-being.

The application of authority that the Agenda for Peace represented was based on a belief of what should characterize the post-Cold War era, and not what actually has characterized the years following the Soviet Union’s collapse.  Authority typically encompasses principles, for authority calls for what should be done or how something should be done.  Authoritative statements founded on a blurred line between normative desires and contextual realities often lead to performance that is considered a partial or complete failure.   Furthermore, the ambiguity of expectations that accompanies such statements makes both learning and performance by an organization difficult to assess.  What exactly is expected?  What is the threshold of success, and how is that threshold determined?  In numbers of lives saved?  In numbers of people fed?  In the absence of recurring violence in a post-conflict society?  Or is it to be measured in the degree to which resources for humanitarian assistance operations are increased, including early warning or military protection of humanitarian workers?

This inconsistency dramatically amplified with the release of the Supplement to an Agenda for Peace. [61]   Three years after the Agenda for Peace was issued, the Secretary-General, in his Supplement, asserted

The purpose of the Supplement, therefore, was to highlight selective certain areas where unforeseen, or only partly foreseen, difficulties have arisen and where there is a need for the Member States to take the ‘hard decisions’ I referred to two and a half years ago. [63]

The Supplement went on to note several major changes since the Agenda had been issued three years before, including: (1)  the prevalence of internal over interstate conflicts, specifically a “new breed of intra-state conflicts” embodying “certain characteristics that present United Nations peacekeepers with challenges not encountered since the Congo operation of the early 1960s.” [64] (emphasis added); (2)  the use of UN forces to protect humanitarian operations, along with an increase in the tendency of combatants to divert relief supplies for their own purposes; [65] and a change in the nature of UN operations in the field, namely, the implementation of peace settlements, including civilian as well as military matters [66] (emphasis added).

The aforementioned passages from the Supplement to the Agenda for Peace at once call upon the UN to assume new duties that depart from traditional peacekeeping, and caution the organization that the obstacles to peacekeeping and peace-building are formidable and the results tenuous at best.  The Secretary-General underscores the anticipated need of the legitimate use of force under the Charter, and even calls for intervention to continue beyond achieving an immediate cessation of hostilities.  Yet he also reinforces the principles of neutrality and impartiality in carrying out any humanitarian intervention.  He acknowledges the prevalence of internal conflicts; yet he fails to address either the perceived or actual outcomes of the persistent and prolonged UN intervention that he advocates.

The Supplement contains directives fraught with norm ambiguities which reflect inherent contradictions and tensions.  This norm conflict is the result of particular characteristics of situations in which those norms have been applied, to both the benefit and the detriment of recipients.  Efforts to maintain neutrality and impartiality in the dissemination of emergency relief, the assumption by peacekeeping forces of humanitarian responsibilities, and the extension by humanitarian organizations into post-conflict development activities are all examples of UN humanitarian initiatives with mixed results that trigger and intensify norm conflict, described below.

Norm Conflict

Emerging norms and norms in contention collide with established norms on which decisions and principles have traditionally been based and justified.  A key example is the norm of state sovereignty, the cornerstone of the UN’s  creation.  Challenging this traditionally accepted right of a state to non-intervention has been the concept of sovereignty as responsibility. This principle holds that the right of non-intervention is valid only insofar as a state fulfils its responsibilities to those that reside within its borders.    Failure to do so warrants intervention on humanitarian and human rights grounds. [67]

The distinction between norm ambiguity and norm conflict is extremely difficult to ascertain.  Challenges to embedded norms may create ambiguity, followed by conflict, as the challenges intensify.  Conversely, salient conflict among norms can lead to norm ambiguity, as political leaders attempt to circumvent or reconcile conflicting elements in an effort to articulate policy direction or win support from state and non-state actors. 

Legitimizing intervention as warranted on humanitarian grounds leads to yet another source of norm contention, namely the conflict between human rights and humanitarian law.  Justification for military intervention on humanitarian grounds carries the risk of human rights violations when force is applied.  Human rights law is based on zero tolerance for the loss of human life. [68]   Conversely, humanitarian law assumes—and therefore implicitly tolerates—some degree of loss of life, including civilian deaths, in situations involving lawful collateral damage.

On the one hand, humanitarian intervention is justified in the name of responding to human rights violations and safeguarding the rights of oppressed populations.  On the other hand, it does little to help determine what type and to what extent intervention should be used.  Are the human rights of an oppressed population worth the human lives of both those intervening and those who may die in the crossfire of conflict?

Moreover, within humanitarian law there is yet another layer of conflict regarding humanitarian assistance, whether or not that assistance is rendered in conjunction with military intervention on humanitarian grounds.  Traditional interpretations view humanitarian assistance as both neutral and impartial.  The International Committee of the Red Cross (ICRC), the first institution to provide humanitarian relief to civilian victims of warfare, defines itself as “an impartial, neutral, and independent organization whose exclusively humanitarian mission is to protect the lives and dignity of victims of war and internal violence and to provide them with assistance.” [69]

Other perspectives, however, assert that humanitarian assistance is inherently partial and non-neutral, that it always has been motivated by the interests of strong states, who are inevitably the assistance providers, and that it often exacerbates and prolongs conflict.  Those who maintain this perspective point to three related characteristics of both humanitarianism as it is currently practiced:  authorization always comes from the most powerful states in the international system; those states inevitably link humanitarian assistance to state interests; and in the current international political order, states are ill-equipped to serve as moral agents, regardless of intent. [70]

Key normative dilemmas include questions of authority or authorization, selectivity, and the absence of a general doctrine.  These problems, of course, are frequently related.  Authorization to intervene and a general doctrine spelling out the guidelines, protocols, and procedures of humanitarian intervention necessarily proceed from a recognized authority.  The UN Security Council’s mandate to authorize intervention does not preclude these problems, since Security Council authorization reinforces the inevitable influence of the body’s permanent members. 

Yet an equally compelling counterargument to these criticisms is the position that not intervening for any of the reasons identified above can be equally morally unacceptable. [71]    Rather, agreement on such shared fundamental principles as the sanctity of human life and the abhorrence of life-threatening disasters should form the basis of intervention on strictly humanitarian grounds.

These debates are not limited to humanitarian assistance that is preceded by military intervention, but also are present with respect to humanitarian assistance that is permitted, even welcomed, by the recipient.  Still others, however, while accepting the premise that all humanitarian aid is inherently political, nevertheless question whether this characteristic is synonymous with causing harm.  As one scholar/practitioner has stated, “If there is moral ambiguity in providing aid, there is also moral ambiguity in providing no aid.” [72]

Another dimension to the debate is the possible absolution of responsibility it affords perpetrators of injustice.  Rendering assistance may alter undesirable conditions, but it also can remove from those who created those conditions the responsibility for changing them.  One question that follows is, which political action is preferable, especially if, as has been the case with many civil conflicts, those responsible for creating unjust and unacceptable conditions are not likely to be brought to justice until multitudes have died?  If any or all of these multitudes could have been saved, would their lives be deemed more important than the risk of non-accountability?

Having presented the five types of learning—political, operational, relational, contextual, and normative, this paper now turns to the inputs and outputs of the framework  by identifying examples of each learning type vis-à-vis a series of crises or issues, the nature of diverse learning occurring with the IGO during that time.  The next paragraphs explain the selection of inputs and outputs, namely changes in mandate, changes in resource allocations, and evaluations.

Inputs and Outputs

This organizational learning framework employs mandate change, resource allocation, and evaluations as both inputs and outputs.    Mandate change both drives and results from learning.  The decision to alter an organization’s mandate leads to the quest for how those changes can best be implemented.  At the same time, mandate change can also be an outcome of learning, when hindsight leads to re-appraisal not only of performance, but to the very foundations of an organization’s responsibilities.  Resource allocation influences learning, because it provides the template for what is deemed feasible.   It becomes an output of learning when lessons learned include recognition of the need to shift resources from one area to another.  Finally, evaluations are inputs, in that they guide future performance.  Even if recommendations are not explicitly followed, the presence of an evaluation casts its shadow; and evaluations as outputs of learning are probably the most familiar variable.  Evaluations, especially internal evaluations and solicited external evaluations, are the clearest indicators of attempts to learn from the past.  (Unsolicited evaluations provide tools for learning but do not necessarily signify the organization’s desire to learn.)  Therefore, organizational evaluation is an output of attempted learning.  At the same time, evaluations of past activities also shape decisions about future behavior, thus making evaluations inputs to learning as well.  Furthermore, designating evaluations as both an input and an output allows for analysis of the evaluations themselves.  This is critical to an understanding of the degree of each learning type that is present within the organizations being studied.  Evaluations that emphasize the need for better coordination without addressing the reasons for lack of coordination in the first place may reveal signs of increased operational learning but little or no contextual, political, or relational learning.  Evaluations that malign an organization’s reticence to act in response to a crisis may show high normative learning but a lack of operational or contextual learning.  Therefore, evaluations as an output may be indicators of incomplete learning, and evaluations as an input may be indicators of inadequate learning tools.

Evaluations are themselves the outcome of a perceived or recognized need to gauge performance in the light of expectation and reputation.  The recent past casts a long shadow, while the distant past is recalled only sporadically.  More specifically, both the recent and distant past, when used as points of reference for a comparison, are value-laden.  Why did this year’s operation not work as effectively as last year’s?  Do this year’s failures signal an overall organizational decline from the (assumed or apparent) exemplary performances of five or ten years ago?  Conversely, positive performances, on the other hand, act as a compass for the future to ensure the replication of success.

In summary, expectation and reputation are traditional key components of evaluations, yet ironically rarely is either clearly defined, either prior to performance or during evaluation.  Rather, two key aspects of expectation are assumed.  First, observers assume that the IGO will follow its mandate, that is, that its leaders will determine appropriate courses of action according to the organization’s stated responsibilities, authority, and limits.  A change in mandate would precede any decision to undertake activities external to the existing mandate.  Second, terms of performance effectiveness or success are implicit in expectation, but quantification of terms ironically may follow performance, e.g.  how many lives were saved is measured against how many perished.

Identifying Learning Types through Analysis of Inputs and Outputs

Identifying the various learning types through the inputs and outputs of changes in mandate, resource allocations, and evaluations is a difficult but essential task.  Not every input/output will reflect every learning type; rather, certain learning types are likely to surface in mandate change, while others will be reflected in resource allocations or evaluations.  Mandate change, especially mandate expansion, is likely to embody stronger elements of normative learning.  The decision to restrict an IGO’s mandate, however, could be indicative of political, operational, contextual, or relational learning, due to recognition of bureaucratic and leadership constraints, capacity limitations, the realities of a given crisis, or the responsibilities already entrusted to another organization.

Changes in resource allocation usually will reflect the other four types more than normative learning, because resource allocations are the means by which mandates are realized.  In other words, normative learning, found in mandate change, addresses organizational responsibilities; political, operational, contextual, and relational learning, reflected in resource allocations, address the means and the extent to which those responsibilities are carried out.

Possible Hypotheses Derived from the Diverse Learning Types Framework

The learning framework presented in the preceding pages yields several possible hypotheses that can be tested using case studies of humanitarian IGOs’ work in the field.

First, in the face of emerging norms, operational learning accelerates, as the desire to respond to those norms prompts organizations to seek ways to act. Second, as normative learning and operational learning drive each other forward, contextual learning becomes obscured.  

Third, organizations with narrow mandates are likely to be more attentive to contextual learning when the context is familiar.  This can be attributed to the fact that organizations with narrowly defined mandates are more likely to generalize their plans.  It is also somewhat paradoxical, in that one would think an unfamiliar context would draw more attention.  This paradox can be explained by cognitive dissonance.  In order to deal with an unfamiliar situation, an organization with a narrow mandate may cope by avoiding full attention to the reality.   Fourth, an organization with a broad mandate, such as UNDP, will be more attentive to context, because of the diverse components required to fulfill its mandate.  Fifth, relational learning occurs rarely, and when it does occur, it does so predictably when one organization alters its mandate and undertakes responsibilities traditionally entrusted to another organization.

Summary and Conclusions

Beginning with the assertion that, despite intense criticism and some performance failures, humanitarian IGOs have learned and continue to learn from their responses to the conflicts of the first post-Cold War decade.  Understanding this learning process requires an analytical framework that accounts for different types of learning, with a particular focus on what the IGO is learning and what it expects and is expected to learn.  Since these diverse types—political, operational, contextual, relational, and normative—do not always complement each other, an organization may appear not to learn from past experience when in fact learning is taking place.  Of the five types presented, normative learning, which has been so prevalent in much of the humanitarian work of the past decade, can complicate the learning process by challenging organizations beyond their capabilities and beyond the feasibility as defined by the context of the crisis.  Exacerbating this complexity is the fact that to date scholars of norm theory have yet to determine why either states or organizations choose some norms over others.  This uncertainty with respect to norm selection obscures understanding of organizational learning.  This learning framework presents norm periodization, norm ambiguity, and norm conflict as characteristics that shape norm selection in the normative learning process.

The identification of learning through examination of mandate change, changes in resource allocations, and evaluations as both inputs and outputs enables a greater comprehension of the types and degrees of learning taking place.  Hypotheses derived from this analysis can subsequently be tested with case studies of humanitarian IGOs’ work in conflict and post-conflict situations.