There, the creation of Parliament as the joint representa-tive of landowners and bourgeoisie limited the monarch's power; inLatin America, local caudillos (political bosses) and Congress did.Structurally, however, the limitations of the analogy are marked by twobasic differences: in Latin America, the bourgeoisie was much weaker,and the availability of domestically generated capital to state makerswas much more restricted

In Tilly's classification, Latin America falls more easily into a thirdform of "capitalized coercionr" in which state makers used both coer-cion and capital to centralize power. Coercion and capital in LatinAmerica were used unevenly, however, for coercion was brutal butinefficient, and scarcity of capital remained most often the rule.

Of our threemain cases, Argentina-because of its wealthy landed upper classes andmore organized military-possessed the ingredients needed for such analliance, and indeed, it came the closest to this route

o satisfy its domestic creditors (and sometimes its inter-national creditors as well), the state tried to reduce its liability throughland grants and light taxes, pensions, or appointrnents to influentialposts in the central bureaucracy.

Even in Argentinar the strongest state among our cases,the central power found it extremely difficult-and at times even unde-sirable-to rely on revenue extraction through coercion. In most of theregiorl states depended heavily on customs duties because the stateswere ineffectual at taxation, making them extremely sensitive to disrup-tions in the prices of exports and imports. If tariffs were too high, thegovernment could succumb to plain smuggling; if they were too low,income plummeted.

Factors related to war and conflict resolution, on the otherhand, seem to create an easier and more common ground for com-parison.

Comparing Europe and Latin America, we come to the conclusionthat al*rough the pace of economic development and differences infactor endowments undoubtedly affect state formation and influencedemocratic or authoritarian outcomes, to stress these variables alonedoes not necessarily facilitate the formulation of a more encompassingtheory.

More-over, culturally, politicallg and institutionally one finds little resem-blance between the historical evolution of these Latin American casesand the European varieties of feudalism. In the new world, reasonsother than feudalism explained the hardy predominance of some "feu-dal" features, such as the preponderance of agricultural enterpriseswith low technology requirements, limited access to credit, poor com-munications, and frequent wars.

Latin America lacked an entrenchednobility, confronted very different international pressures, underwentno industrial revolution, formed a rather weak and late industrial bour-geoisie, held no colonies, and experienced colonial rule. To these, onemust add differences in demography, culture, and factor endowments.But it is precisely these contrasts that both mark ttre limits of theoriesemerging from the European experience and contribute to a sharperpicture of state making on both sides of the Atlantic

e challenge from a comparative perspective is to ex-plain differences in the timing of the centralization of power, to spell outthe conditions under which state makers succeed or fail, and to detectthe rationale behind more democratic or authoritarian outcomes.

f an appropriate degree of bureaucratic autonomy is an importantcharacteristic of high-quality government, then neither the Weberiannor the principal–agent models can stand intact as frameworks forunderstanding how bureaucracies ought to work.

to certain professional norms that seek to preclude certain types ofself-seeking behavior.

his is why Robert Klitgaard coined theformula Corruption = Discretion - Accountability (Klitgaard 1988).This is also why development agencies have been advising poor coun-tries to limit bureaucratic discretion in recent years

One of the worst forms is when bureaucracies losecontrol over internal recruitment and promotion to the politicalauthorities and are staffed entirely by political appointees.

A com-pletely autonomous bureaucracy gets no mandates at all but sets itsown goals independently of the political principal.

State-owned utilities, for example, often have mandates to simultaneouslydo cost recovery, universal service to the poor, and efficient pricing tobusiness clients, each promoted by a different part of the politicalsystem. These different mandates obviously cannot be simultaneouslyachieved, and generate bureaucratic dysfunctionality.

As state capacity varies substantially across functions, levels of govern-ment, and regions, one would ideally want capacity measures for allmajor government agencies. In Brazil, for example, it has been widelyrecognized that certain “islands of excellence” exist within the Brazil-ian state that would be missed by an aggregate measure.

The real mechanisms of perpetuation lay elsewhere. They will be found,most generally, with the varying power configurations that were originallyput in place during colonialism. It is these configurations – the basic con-stellations of power in state and society – that persist in Spanish Americancountries in spite of changing economic opportunities and actors, evolvingcultural practices, and fluctuating political conditions.

There are few relative improvements or declines; nocase has, for instance, moved from poor performance to good performance(or vice versa).

Given its relatively low level of economic development, Paraguay is moresurprising as an intermediate performer in the social arena. But its interme-diate position is quite clear from the data. Paraguay rarely, if ever, showsup in the bottom third of the region’s hierarchy, and it often makes veryrespectable showings (relative to regional and global standards). Paraguayand Colombia consistently place between fifth and eighth in the hierarchy.They have been roughly equal on literacy, life expectancy, and child mortal-ity. This intermediate ranking is expected for Colombia, because the countryis a regionally intermediate economic performer

hese four are also the places where, beginning in the late colonialand early independence periods, national myths were most effectively con-structed to portray the population as a single homogenous and mostly Euro-pean group. A necessary condition for such a mythology (as we have seen)was the near-complete disappearance of the indigenous population as alarge, spatially concentrated, and systematically subordinated stratum ofsociety

Countries that perform better on indicators of humanwell-being almost always do so year after year without interruption; the sameis true for countries that perform poorly relative to the rest of the region

The real long-term gains that were truncated by U.S. intervention mayhave been in the arena of social development. For progress in social devel-opment can occur in the midst of economic backwardness.57 And socialrevolutions may foster precisely improvements in human welfare. Let ustherefore turn to postcolonial trajectories of social development in SpanishAmerica, with an eye to noting any ruptures that occurred, perhaps withoutconcomitant changes in economic development

Such limited effects, when compared to those of the Mexican Revolution,raise the question of why social revolution inaugurated structural changesthat promoted long-run economic development in Mexico but not in Bolivia.We must remember that in Mexico, sustained high growth was driven bythe rapidly industrializing north, the part of the country that had beenmarginal under the Habsburgs but of increasing importance during theBourbon period. In a sense, the accomplishment of the Mexican Revolu-tion was to sweep away the colonial heritage in the Central Valley andthereby remove the fetters that had been preventing growth from fully tak-ing off in the country as a whole. The Bolivian Revolution also did awaywith many colonial legacies in the highland areas that had been the focus ofHabsburg colonialism. But Bolivia lacked a region that could play the rolethat northern Mexico did for Mexico

If we focusstrictly on the growth of the economy, it is hard to argue that the BolivianRevolution had much of a long-term impact one way or the other

The immediate impetus for change was theconclusion of the civil war of 1948, whose settlement most famously gavebirth to Latin America’s longest-lasting democratic regime.

Commentatorswho traveled in Central America during this period were already remarkingon Costa Rican exceptionalism. For example, Dana Munro considered CostaRica “a nation which is entirely different from any of the other CentralAmerican republics.” He argued that its relative political stability “enabledher to attain a prosperity which has entirely transformed the backward andpoverty-stricken community of colonial days

What really seems to have changed afterand as a consequence of the revolution was the performance of central andsouthern Mexico, those regions where colonialism had significantly hindereddevelopment. It was here, more than anywhere else, that the revolutiondestroyed the power of the large landowners and succeeded in its goal of eliminatingor at least significantly reducing “feudal” relations of production. In the process, iteliminated an important obstacle to increased agricultural production.

By contrast, in central Mexico, heartland of the Habsburg colonial project,the growth was much less impressive. Landlords here sought to cash inon the new market opportunities, but “‘feudal’ relations acted as a brakeon commercialization and a focus for subaltern resistance.

Growth was led by the northern region, precisely the area where the colonialheritage was most favorable for development. As Alan Knight stresses:In general, northern Mexico – economically dynamic, underpopulated, mestizo, andmore closely tied to the US economy – benefited from Porfirian growth.

Yet the basic trend of a rapid and dramatic rise forVenezuela after 1920, followed by abrupt fluctuations in annual growthrates, is not disputed. Nor is the claim that the world price for petroleumlargely dictated fluctuating growth rates

Hence, any Argentine reversal of development was mostly an artifactof poor performance throughout Spanish America (compared to westernEurope). The question of why Argentina declined in the mid-twentieth cen-tury is thus really a question of why Spanish America in general performedpoorly relative to richer countries during this time.

Their advantages relative toArgentina rested in part with the fact that they, unlike Argentina, werenever colonized by a mercantilist power at all.

Argentina’s failure to become an advanced industrial nation raises whatCarlos H. Waisman has called the “Argentine question.”16 By the mid-twentieth century, Argentina’s wealth was no longer on par with the Euro-pean average; and by the end of the century, its GDP per capita was onlyabout half that of Spain, Italy, and Austria – nations it had once outstripped.

Likewise, some shifts may result fromtemporary situations (e.g., a coup or a slump in international demand) thatare soon “corrected” when the country returns to the level of developmentthat existed prior to the “abnormal” circumstance

The few available estimates ofGDP per capita for the nineteenth century (or before) cover only a smalland unrepresentative set of the region’s countries.8 For recent decades, thethree most important datasets that cover all fifteen mainland Spanish Amer-ican countries are (1) Rosemary Thorp’s decadal index of national GDP fortwentieth-century Latin America; (2) the Penn World Table’s annual RGDP(real gross domestic product) per capita data for the 1950–2005 period;and (3) the World Bank’s annual GDP per capita data for the 1965–2005period.9 These are not the only indices available, but they are the ones withthe most comprehensive over-time data for the statistic (GDP per capita)

n Spanish America,as we shall see, ruptures that have improved a country’s relative economicstanding have occurred either because of the discovery of massive petroleumreserves (in Venezuela) or in the context of social revolutionary transforma-tions (in Mexico).

If stability in relative levels of development is the expected norm, thenperiods of discontinuity and change are what really need to be explained.Following Adelman, we can call these episodes of discontinuity in SpanishAmerican history “ruptures” – that is, “moments when the continuous timelines were irrevocably snapped.”5 Analysts of path dependence often urgeattention to precisely such breaking points and routes out of lock-in.

My argument is not that colonial institutions them-selves persisted into the present; rather, most of the colonial institutionsanalyzed here – and their affiliated actors – were gradually rearranged oreven abolished in the first decades that followed independence from Spain.It was the positions of the countries in the world hierarchy of developmentthat persisted over the long run. The basic socioeconomic legacy of Span-ish colonialism was precisely stable cross-national differences in levels ofdevelopment

rsistence is an old theme in work on Latin America. “More than inany other region,” writes Jeremy Adelman, “the formulation of the past aspersistence has shaped regional and national narratives of Latin Americansocieties.”

In liberal colonialism, the hypothesized pattern is just the opposite: higherlevels of colonialism are associated with higher levels of postcolonial devel-opment.

Level of mercantilistcolonialism is negatively related to postcolonial development, whereas levelof liberal colonialism is positively related to postcolonial development.

Conversely, liberal economic powers pursue lower levels of colonialismin more complex precolonial regions and higher levels of colonialism inless complex regions. More complex precolonial regions feature entrenchedprecapitalist institutions, and these institutions make it difficult for lib-eral colonizers to achieve market-based accumulation. A patrimonial state,entrenched precommercial agriculture, and multiple ethnicities within adensely populated polity – all of these are viewed as obstacles to efficientproduction and accumulation.

Hence, in colonial areas with a complex indigenous society, weexpect to see significant mercantilist activity, including the growth of a largesettler population and the introduction of institutionalized ways of life fromthe metropolis

The institutional complexity of the precolonial societypowerfully shapes the extent to which colonizers believe it is feasible anddesirable to impose their institutions and people

Level of colonialism refers to the extent to which a colonizing power installseconomic, political, and sociocultural institutions in a colonized territory

To provide the appropriate incentives, moreover, a liberal state works toensure law and order and to protect trade and production for profit; itavoids the most severe restrictions limiting and regulating commercial inter-ests.64 And the market is the main tool for shaping stratification; institu-tions that explicitly privilege status groups and impose hierarchical relationsof dependence are discouraged.

Under a mercantilist political economy, state authorities seek nationaleconomic self-sufficiency and organize productive activity to ensure favor-able trade balances and the accumulation of precious metal

Third, to identify the origins of differing levels of colonialism,I emphasize the institutional organization of the indigenous population atthe onset of colonialism.

in order to conceptualize how territories differentially experience insti-tution building during colonialism, we must analyze their contrasting levelsof colonialism.

First, in distinct contrast to other recent work oncolonialism and development, I hold that the institutional composition ofthe European colonizers – in particular, the institutions that constitute theirpolitical economies – needs to be a center of attention

Thus, attention must turn to colonial rules for securingindigenous labor, for assigning land rights, and for designating local politi-cal power holders. These institutions often connect ethnoracial categories topatterns of resource allocation; they can make ethnic identities into highlyenduring axes of contention for the people designated by the identities. Theconsequence for development may be especially grave when a large portionof the overall society is excluded, on the basis of their ethnoracial identity,from the fruits of economic prosperity

Building on North’s celebrated work on economic growth, Acemoglu,Johnson, and Robinson see institutions that guarantee property rights as thefundamental cause of long-run development.

When employing an institutional approach, therefore, it is essential to treatinstitutions as the objects of contestation among actors differentially impli-cated in their resource allocations. Only this perspective can capture theprominent role that power and conflict play in actual patterns of institu-tional formation and change.

In the second place, a distributional approach directs attention to issuesof conflict and power – issues that are essential for valid explanation butthat remain hidden so long as institutions are characterized as coordinat-ing devices

In the words of Jack Knight, “Institutions may or may not be sociallyefficient: it depends on whether or not the institutional form that distribu-tionally favors the actors capable of asserting their strategic advantage issocially efficient.”

Rather, I want to take sides and insist that institutions are first and foremostdistributional instruments and only secondarily coordinating mechanisms.

Consensus thus exists that, minimally,institutions (1) consist of formal or informal rules, (2) offer a guide to behav-ior, and (3) are consciously or unconsciously known by individuals in a givenpopulation

Any adequate explanation of long-run development needs to be attunedto the effects of geographic variables. Yet these effects are indirect, exertingthemselves only or primarily through mediating institutions. Consequently,in the analysis below, I will take geography quite seriously, but I will do sowith the goal of understanding how it shapes the establishment of colonialinstitutions, rather than by trying to make a direct link between geographyand development

shows that socioeconomic development in colo-nial regions did not depend primarily on agricultural potential. Countriesthat had similar agricultural endowments (e.g., Nicaragua and Costa Rica)sometimes turned out quite differently; likewise, countries with quite differ-ent agricultural settings (e.g., Colombia and Paraguay) sometimes ended upwith broadly similar levels of socioeconomic development.

In turn, I contend,differences in the organization of indigenous societies shaped the extent andform of European institution building during colonialism independently ofgeographical conditions.

Regions with climate andsoils appropriate for export plantation agriculture, they argue, are especiallylikely to be subject to a mode of institution building that promotes inequalityand inhibits socioeconomic progress.Undeniably, there is an association between temperate zones and greaterdevelopment and between tropical zones and lesser development in manyregions of the world, Latin America included. Likewise, there is no gainsay-ing that variables related to other aspects of climate are correlated with devel-opment in Latin America and elsewhere. But are these relationships causal innature? It is the position of this book that they are not: geographic conditionsare strongly associated with an antecedent factor, and this antecedent fac-tor – much more than geographically determined agrarian potential itself –gives rise to the institutions that drive levels of development.

The consequences of geography for institution building also vary depend-ing on the actor that pursues institution building. Actors who seek overseastrade, for example, may value strategic port locations in a way that overlandtraders do not. And changes in a given actor may lead it to react to the samegeography in distinct ways

Geographical effects simplydisappear once controls for institutional variables are introduced.

Their work has established beyond any reasonabledoubt that several features of geography are correlated with contemporarylevels of national development. Even the simple variable of distance fromthe equator performs reasonably well as a predictor of current levels ofGDP per capita: countries that are more distant from the equator tend tobe richer. Yet for the purposes of actually explaining levels of development,as opposed to identifying features that are correlated with development, wemust ask questions about how geography affects development, when geog-raphy affects development, and what specific features of geography affectdevelopment. For each of these questions, the existing literature providesinsights, but these insights need to be enriched by a more historically con-textualized approach if adequate explanation is to be achieved

With these insights at hand, we can begin to understand why so fewcountries have managed to achieve sustained high rates of growth and socialprogress. Most basically, a developmental state and a society lacking rigidhierarchical bonds of dependence are extremely difficult to construct. Inall developing countries, powerful actors derive resources from institutionalarrangements that promote inequalities. To dislodge these institutions andactors, a fundamental transformation of the basic distribution of powerwithin society is often required. Historically, such transformations haveoccurred largely in conjunction with foreign interventions, land reforms,revolutions, or wars that unseat dominant economic actors.

A country will experience a significant change in its relative level of develop-ment only when the variables that affect rates of growth (or rates of socialprogress) assume atypical values for prolonged periods of time.

This last point is of essential importance, because the vast majority of for-mer European colonies in Africa, Asia, and Latin America are now lower-or middle-income countries – the population that has exhibited the moststability in relative levels of development.

Specifically, at least in the post–World War II era, upper-middle-income nations have had the highest growth rates, and lower-income nationshave had the lowest. As a result, the global trend has been toward incomedivergence, even though there has been some convergence among the setof wealthy nations

A concern with relative levels of development differs in basic ways froma focus on rates of development. Whereas rates highlight variations in per-formance during specified intervals of time, relative levels cast the spotlighton differences that tend to endure across any given period. Precisely becauserelative levels of development are so persistent, one must explain them his-torically: their origins rest at some point before countries stabilize theirpositions in the hierarchy of development. By contrast, the causes drivingrates of development are typically far less historically rooted, and they mayinclude such short-range factors as natural disasters, business cycles, andpublic policies.

William Easterly and colleagues findthat the correlation for national growth rates across decades ranges from0.1 to 0.3, and hence that the performance of countries in one decade onlyweakly predicts performance in the next decade.

Stability marks relative levels of national development because differencesin rates of growth and social progress for countries are not stable. Episodesof sustained high (or low) growth rates are rare in the contemporary worldeconomy.

When treated as a theoretical construct, development may be defined,following Amartya Sen, as “the expansion of the ‘capabilities’ of people tolead the kind of lives they value – and have reason to value.”

Extensive colonialism featuringheavy settlement and institution building is not expected always or usuallyto leave behind rich and egalitarian new countries (as some analysts sug-gest). Nor is extensive colonialism predicted to produce impoverished andconflict-ridden states (as others suggest). Rather, the effect of a given levelof colonialism for long-run development depends on the political-economicinstitutions of the colonizing power

I examine the “fit” between the institutions of the colonizing nationand those of the colonized territory

Moreover, existing institutionalwork on colonialism in particular has suffered from vague conceptions ofinstitutions and overly generalized understandings of their effects that can-not make sense of basic historical facts about particular cases.

Concerning the geo-graphic perspectives, I argue that they too often assume that features of thenatural environment directly affect development; these approaches fail togive appropriate weight to mediating institutions. And they provide littleguidance for theorizing the time-variant effects of geography and the waysin which geographic effects depend on the arrangement of already-existinginstitutions.

One worthy approach is to explore why colonialismoccurred in certain places and at certain times;4 another is to generalizebroadly about the effects of colonialism for Europe, for the non-Europeanregions, or for the world system as a whole.5 Yet some scholars – espe-cially comparatively and historically oriented social scientists – will alwaysbe drawn to questions about the sources of alternative modes of colonialismand their legacies for nation-states. Why did similar or different forms ofcolonialism arise within the borders of what are now sovereign states?

While imperialism anddependency entail asymmetrical relationships between states, they do notinherently involve a loss of sovereignty or even the insertion of governancestructures under the control of a metropolis. Colonialism is set apart fromthese other kinds of interstate domination above all because it renders sub-ordinate (or makes obsolete) all prior political entities that could once layclaim to – and perhaps back up through coercive means, if necessary – finalauthority over territorial inhabitants.

In modern world history, colonialism is marked by a state’s successfulclaim to sovereignty over a foreign land. Under a colonial arrangement,major actors in the interstate system at least implicitly recognize the col-onizing state’s patrimony over the occupied territory; and this recognitionis founded in part upon the colonizing state’s proven ability to implantsettlers, maintain governance structures, and extract resources in the ter-ritory

The institutionsestablished during colonialism, furthermore, exhibited over-time effects,whether directly through their own persistence or indirectly through theactors and processes that they brought into being.

James K. Rilling and colleagues documented diff erences in the neural circuitry of bonobos and chimpanzees that relate to social cognition

Th e ability to understand and to engage in social interactions in the context of one’s culture and social institutions— social cognition— is essential to a person’s quality of life and even survival.

remack and Woodruff conducted tests with the chimpanzee Sarah to discern whether she could attribute mental states to actors on tapes; that is, did Sarah have a theory of mind?224 Th ey decided that she did, though other psychologists were unconvinced

Young chimpanzees also demonstrated intentional behavior by directing a well- regarded trainer to food rewards, misdirecting an untrust-worthy one, and suppressing the urge to look toward a baited container after fooling a bogus villain.2

Th ere are similar instances of other captive and free chimpanzees with-holding responses to preferred objects while dominant chimpanzees or hu-mans are watching them and misdirecting others before sneaking back to an incentive.228 Further, they act against conpecifi cs that have served them badly and cooperate with those that have not.229 Mitchell also documented instances of deception and hiding in a group of four captive western goril-las.230 Tactical deception and other manifestations of Machiavellian intelli-gence are attributes not only of humans and apes but also of nonhominoid primates.231

Initially, gorillas and gibbons failed to evidence self- recognition in mirrors via Gallup’s dye test.215 While the hylobatid failure seemed reasonable, con-sidering their other resemblances to monkeys, the results for gorillas were puzzling.216

Nonetheless, though chimpanzees are adept at mapping familiar tracts from direct views and tele vi sion screens, some fail to orient themselves to- ward incentives in a distant room on the basis of simple small- scale models (i.e., symbolic maps) like those comprehensible to 5- year- old humans.1

In sum, if apes and perhaps monkeys have latent mathematical abilities, they are either well masked or very rudimentary by comparison with human mea sures based on arbitrary symbols

Great apes performed signifi cantly better than gibbons, Old World mon- keys, and lemurs on 67 percent and 84 percent schedules. Gorillas and orang-utans were signifi cantly superior to chimpanzees on the 67 percent prereversal schedule. Th ere was no statistically signifi cant diff erence among the three great apes on the 84 percent schedule.148 Hence, Rumbaugh rightly challenged the myth of the mental supremacy of chimpanzees among apes.1

A twenty- fi rst- century meta- analysis of published studies has indicated that great apes are more intelligent than other nonhuman primate species

Köhler and other comparative psychologists noted that apes excel in prob-lems that require the use of tools. Further, they were especially interested in whether apes exhibited insight during problem solving.135 Köhler set the crite-rion for insight as “the appearance of a complete solution with reference to the whole lay- out of the fi eld.”136 For example, a subject might manipulate experi-mental objects and use them inappropriately for a span. Th en the subject may stop and remain quiet or may engage in unrelated activities. Abruptly, the subject returns to the situation and solves the problem straightaway. In this situation, the subject appears to have thought about the task before its actual resolution— the subject employed ideational pro cesses.1

Hominid brain mass has evidenced a threefold increase in size over the past 3 million years (Table 10.1).126 Brain size doubled between Australopithecus (3.0 Ma) and Homo erectus (1.5 Ma

Wernicke’s area appears to house the dictionary, which selects the words to be sent to Broca’s area where they are assembled syntactically. Broca’s

A negative consequence of increased brain size is that interhemispher

As in body size, the brain sizes of gorillas and orangutans exhibit consid-erably greater sexual dimorphism than the brains of chimpanzees.

White matter in the central ner vous system consists chiefl y of glial cells and myelinated (lipid- coated) axon tracts that conduct signals be-tween the ce re bral regions and between the cerebrum and the cerebellum and lower brain centers.

Grey matter is a major component of the central ner-vous system that is vital to muscle control, speech, sensory perception, mem-ory, and emotions.

Linguistic functions are lateralized to the right- posterior cerebellar hemi sphere, and spatial pro cessing is lateralized to the left posterior hemi sphere.20 Th e lateral asymmetry of chimpanzee posterior cere-bella is also related to complex activities— especially tool use, aimed throw-ing, and handedness— that mirror human conditions.15 Rilling summarized the distinctive features of hominoid brains as follow

Nevertheless, there are no-table indications of linkage between gross brain size and high- level cognition in extant Hominoidea because of “increased numbers of neural modules, increased size of the neocortex and of other higher neural pro cessing areas, and increased neural connectivity.

n gross size, the brains of elephants and whales surpass our brains, but relative to body mass human brains are the largest among placental mammals.1

All nine of the ape species in their sample evidenced food sharing with off spring, but adult gorillas, white- handed gibbons, and siamang evidenced no food sharing even though some partner choice occurs among all of them

Sharing with off spring is predicted by the relative diffi culty of pro-cessing dietary items, as mea sured by the degree of extractive foraging; not by overall dietary quality.• Interadult food sharing only evolved in species that already shared with off spring, regardless of diet.• Intersexual sharing coevolved with the opportunity for female mate choice.

In general agreement with White, Hohmann and Fruth summarized their observations on Lomako bonobos as follows:• Th ey share both plant and animal food.• Shared food items are large and heavy.• Food sharing generally occurred among individuals of diff erent ages and sexes, but females were more often in possession of food and shared more often than did males.• Females most frequently shared with infants, often with other females, and least often with males.• Infants received and took food more often from females other than their mothers.• Food transferred from males to females and among females at equivalent rates, but never from females to males or among males.11

Insofar as their habits are known, bonobos appear to be more pacifi c than chimpanzees. In some localities, they kill and eat other vertebrates. Th e roster of their mammalian prey species (n = 10) is a mere one- fourth that of chimpanzees (Table 8.2). Th eir primate prey include Demidoff ’s dwarf galagos and cercopithecine monkeys.96Lilungu bonobos caught and handled (sometimes fatally) two red- tailed guenons and an Angolan black- and- white colobus, but there is no evidence that they ate their captives

Despite the fact that chimpanzees sometimes brutally attack and kill other adults, there have been no reports of chimpanzee cannibalism in which the victim was over 3 years old

instance, an adult female attacked and killed the 3- week- old infant of a group member and shared the carcass with her adolescent daughter and infant son. About a year later, the daughter killed the recently born infant of the same female and shared it with her mother and brother. One month later, she killed another neonate in the group. Demographic rec ords for the Gombe chimpan-zee population indicated that over a three- year span only one infant survived longer than one month. Th e cannibalistic killings seemed to end, though Goodall intervened on one occasion to stop another attempted infanticide by the infanticidal daughter.

Among apes, chimpanzees are the only species reported to scavenge meat, albeit relatively rarely.69 Ingesting uncooked carrion might be unwise because it is likely to be contaminated with bacteria that can cause gastrointestinal distress or worse.7

Th e Ngogo data provide only limited support for the meat- for- sex hypo

Ngogo is the largest known chimpanzee community and comprises the most males. Th ey kill 6 percent to 12 percent of the red colobus population annually. Hunts are more likely to occur and are more successful where there are gaps in the canopy. Although the number of kills and the off take of meat per hunt increase with the num-ber of hunters, the per capita meat intake is in de pen dent of hunting party size.

Chimpanzees seem to have no vocal or gestural signal that is specifi c to sharing vertebrate versus vegetal foods. Moreover, laboratory experiments and fi eld observations indicate that generally chimpanzees are selfi sh, in part due to their limited understanding and sharing of intentions, and in part due to their indiff erence to the welfare of unrelated group members, though they might be altruistic and empathetic toward some group members, especially kin or reciprocating partners.

Chimpanzees seldom kill baboons outside Gombe National Park; after food provisioning was stopped, baboons were rarely victims there.41 Baboon killing was also very rare in Mahale Mountains National Park.42 Hence, it seems reasonable to attribute past victimization of Gombe baboons to the banana bonanza

Before Jane Goodall began her studies at Gombe, two documented adult male chimpanzee attacks on humans had occurred, presumably to snatch infants.35 When searchers found the body of one human infant, it had been partially eaten, presumably by the kidnapper, although it is also possible that dogs or other scavengers had eaten it

etween March 1968 and August 1970, 48 percent of forty- four Gombe chimpanzee predatory attempts— mostly via seizures and chases of baboons— were successfu

Chimpanzees employ a variety of methods to dispatch their prey:• Battering them against the ground or standing objects.• Biting them on the nape of the neck or spine.• Wringing their necks.• Eating them alive.• Pulling them apart while competing with others for the carcass

Th ere was no begging or active sharing; individuals simply waited un-til a possessor discarded or dropped a carcass to feed on it.2

Chimpanzees, bonobos, baboons, and capuchins constitute the rare nonhu-man primates that systematically hunt sizeable vertebrates, and bonobos and chimpanzees are the most inclined to share prey.2

Scenarios of man- the- hunter were refueled by pioneer reports of meat eating and hunting among Gombe chimpanzees, followed by recognition that it is characteristic of many chimpanzee communities across Africa.1 Concomi-tantly, as the United States fl aunted its fi repower in Vietnam, the aggressive male regained ascendenc

Th us, it appears that the silverback may not be placing his nest strategically in order to protect group members from nocturnal predatory attacks.2

h ey were especially attracted to the secondary forests when fi gs

Among apes, only some gorillas commonly lodge at night on the groun

Like bonobos, which sometimes nest in groups of ≥20 individuals (mean: 9), chimpanzee nesting is generally much more communal than that of orang- utans.213 Chimpanzee foraging parties nest together, split into smaller nesting groups, or join neighboring foraging parties for the night.217 Th ere was a cluster of twenty- three chimpanzee nests at Ugalla, Tanzania.218 A Gombe chimpan-zee nesting party produced a cluster of seventeen nests, and another resulted in ten nests in one tree. In de pen dent Gombe males tended to lodge ≥90 meters from females and juveniles that nested within a few meters of one another.2

Like orangutans, bonobos and chimpanzees prefer trees for night nesting (Plate 7). In some areas, they seem to select a par tic u lar species of trees for nesting. For example, at Yalosidi, though bonobos nested in 103 species of trees, nineteen species accommodated 81 percent of them. 107 At Lomako, bonobos nested in only twenty- six of fi fty- two (50 percent) potentially avail-able tree species in seventeen (of 48 total censused) forest plots that contained nests.213 Like Sumatran orangutans, Wamba bonobos sometimes used leafy branches as rain covers.2

Great apes usually build a new nest or reline an old nest each night, but there is little evidence that they possess full- fl edged nesting instincts.204 Th ey need safe new or refurbished nests each day because their foraging rounds take them to diff erent localities. Moreover, their transience probably assists in confounding competitors and potentially also predators.

Like cercopithecoid monkeys, gibbons roost arboreally with the aid of ischial callosities. By contrast, great apes support their bulk by building nests, in which they recline.

Although all great ape species build tree platforms— nests—arboreally at night, some African apes rest diurnally on the ground, and in some areas the heavier gorillas nest nocturnally on the ground.

In contrast, the Bwindi gorilla diet comprised 24.6 percent fruit and 75.4 percent leaves and THV. More than one fruit species (mean: 2.2) were found in 47.2 percent of gorilla fecal samples. Particles of decayed wood were found in 19.4 percent of gorilla fecal samples, and they also ingested small stones.161 Decaying wood provided 95 percent of their dietary sodium, though it rep-resented only 3.9 percent wet weight of their food intake.1

Adult gorillas have much longer gut retention times than chimpanzees (mean: gorilla, 50 hours; chimpanzee, 31.5 hours).147 Although fundamentally frugivorous, like some sympatric monkeys and chimpanzees, Lopé gorillas are outstanding for “their ability to eat large fi brous fruit, mature leaves and stems, and to overcome high levels of . . . phenols and condensed tannins.”

Th ere are rare isolated incidents of western gorillas and chimpanzees feed-ing peacefully together in fruit trees. For instance, in the Nouabalé- Ndoki Forest in the Republic of the Congo, Suzuki and Nishihara observed them feeding together in fi g trees four times.

Active competition can require a notable expenditure of energy and occa-sionally may lead to injury. Direct confl icts between diff erent species of apes in areas where they are sympatric are rare or at least are seldom observed by fi eld researchers. Bonobos and many species of gibbons are allopatric with re-gard to other ape species. But chimpanzee and gorilla populations are sympatric

Th e Asian and African forests that accommodate apes have markedly dif-ferent characteristics. Th ere are more plant families, genera, and species in Asian rain forests than in African rain forests.117 Southeast Asian rain forests are especially rich in species of the Dipterocarpaceae, which, along with other large- seeded species, experience irregular supra- annual peaks in syn-chronous fl owering and fruit production (masting) often followed by much lower yields during most years per de cade. Possible results of this phenome-non are lower mammalian biomass and specifi c richness, including fewer species of frugivorous cercopithecine monkeys but more colobine species in Southeast Asian rain forests.118 Th e densities of orangutan populations might also be lower in forests with relatively high densities of Dipterocarpaceae. However, when compared with the limestone karst forests, the lowland dip-terocarp forest accommodated the highest density of orangutans in the Sangkulirang Peninsula of East Kalimantan, perhaps due to the low diver-sity of tree species in the limestone karst forests.119

Like chimpanzees, bonobos feed at all levels of the forest.

they also feed heavily on leaves, bark, and pith like the moun-tain gorillas that also range in montane forests.

(>90 percent) of their feeding time ingesting nonreproductive parts of plants, often to the exclusion of fruits (Plates 8 and 9).92 Later fi eld researchers showed that western, Cross River, Grauer’s, and even some mountain gorillas are more frugivorous (including seeds) than the Virunga mountain gorillas (Table 7.6).93 Th eir data sets are chiefl y based on analyses of fecal samples and leavings at fi eld sites because of the diffi culties of habituating the subjects; the relative brevity or primary foci of their stud-ies did not include systematic dietary observations.

Stumpf generalized the di-etary composition of Pan troglodytes to be 64 percent fruit, 16 percent leaves, 7 percent terrestrial herbaceous vegetation, 4 percent bark and miscellaneous items, 4 percent animal prey, 3 percent seeds, and 2 percent fl owers (Plates 5, 6 and 7).80

Like many human mall and fast- food customers who forage for sugar, salt, and lipids, orangutans have the potential to fatten dramatically, which serves them well in the forest but can predispose them to diabetes in captivity.6

Th e capacity to toler-ate toxins increases with body size, especially in primates with simple versus sacculated stomachs (Figure 2.5). Further, the energy demand per unit body mass is less the larger the individual. Accordingly, whereas the smaller gibbons best rely on sugary fruits and tender young leaves, the great apes can more readily fall back on tougher leaves, bark, unripe fruit, and stems

Primates are probably endowed with somewhat diff erent abilities to taste, smell, and palpate potential dietary items, though one cannot assume that the senses are controlled by the same alleles.44 Likewise, although one might assume that primates hear with equal facility, this might not be the case.45 Although the hands of apes might appear clumsy because of the proportions of their fi ngers, they are extremely sensitive and adept at manually collecting and preparing well- protected food items.

Primates are unique among eutherian mammals in having color vision.39 Inso-far as we can be informed by current data, like humans, the apes and Old World monkeys are more or less equally endowed with trichromatic color vi-sion that enables them to discern relatively tender red and yellow proteinaceous young leaves and fl owers and colorful sugary and fatty fruit amid tougher mature green foliage (Plates 3, 4, 5, 6, and 7).

Chimpanzees can also subsist chiefl y on quantities of rather hard seeds and desiccated fruits— reminiscent of trail mix (gorp)— over notable periods of time. With the exception of Virunga mountain gorillas, the apes subsist on a remarkable variety of plant species and parts. Because of seasonal fl uc-tuations in the availability of preferred plant species, their subsistence staples diff er dramatically from month to month and season to season.23 Chimpan-zees and gorillas living near fi elds and plantations forage cultigens, thereby incurring the wrath of farmers.

Another major challenge for primate socioecologists is to identify and to determine the eff ects of preferred versus staple and seasonal fallback foods on foraging behavior and group structure and dynamics.16 Apes and other verte-brates adjust their daily and seasonal ranges, their feeding, and the traveling substrate and composition of their foraging parties according to the avail-ability of their favorite and fallback foods.1

Amylase in the mouth and small intestine releases glucose units of poly-saccharide starch in seeds and other structural plant parts for absorption in the small intestine. In the ceca and colons of anthropoids with simple stom-achs, and in the fi rst of the four chambers in the stomachs of colobine mon-keys, microbial fermentation of the polysaccharide cellulose, hemicellulose, lignin, and pectin that form the walls of plant cells produces short- chain fatty acids that are absorbed into the bloodstream (Figure 2.5).

In general, fruit pulp and edible seeds are sources of caloric energy via car-bohydrates or lipids. Leaves, stems, fl ower buds, and seeds provide protein for growth. Insects contain protein, amino acids, and lipids.8 Depending on the plant part, a notable amount of protein might not be available to an animal because it is bound to fi ber.9 All plant foods contain water, and some parts may be rich in vitamins and minerals. For example, relative to other forest fruits, on average, fi gs contain three times more calcium and have a calcium- to- phosphorus ratio suffi cient to promote bone growth.10 Figs provide fl uid and quick energy in the form of carbohydrates, and they contain amino acids, protein, and lipids when infested by fi g wasps and their parasites.11

Observations of foraging subjects.• Collection of leavings at foraging sites.• Analysis of stomach contents from dead subjects.• Analyses of urine and feces, including gross examination and DNA se-quencing of undigested animal and plant fragments.2• Stable isotopic analyses of bone, teeth, hair, and other tissues to indicate the kinds of plants and animals eaten.

As heterotrophs, hominoids must forage for foods that will provide them with sustaining energy for basic metabolic pro cesses, cell replacement, body growth, and, in females, gestating fetuses, lactation, and transporting dependent youngsters.1 On balance, daily foraging should not often require more energy than is required to fuel the daily quest, and one also must avoid predators, lengthy and potentially injurious altercations

Th e remarkable humanness of the Laetoli G footprints contrasts with the apelike features of the Hadar foot bones (Figure 6.15). Indeed, the longish, curved toes of the 3- Ma Hadar feet and features indicating absence of a me-dial longitudinal arch make them unlikely candidates for the species that made the Laetoli trails.99 Th erefore, I advise against ascribing the footprints to Australopithecus afarensis.Th e lack of consensus about what kind of feet made the Laetoli G homi-nid footprint trails is fueled by the following factors:• Conditions of the prints when they were originally excavated and weath-ering immediately thereafter were followed by reburial and further dam-age due to subsidence and growth of vegetation atop them.• After re- excavation and before the second reburial, relatively few experts and commentators had access to them.• Bias persists about what the feet of 3.66- Ma Laetoli hominids should look like, given that foot bones of broadly contemporaneous Australopithecu

Several optional conclusions are possible:• Th ere was a species of Australopithecus with virtually human feet.• Homo was present at Laetoli around 3.66 Ma.• Th e trackways were made by an anonymous genus for which we have no identifi ed skeletal remains; accordingly, they should be designated Ho-minidae genus and species indeterminata or Ichnanthropus bipes.

Th e large human heel serves as a powerful lever for the triceps surae mus-cle, which plantarfl exes the foot at the ankle joint during bipedal walking and running.

Humans have no ischiofemoralis muscle, and the gluteus maximus muscles are large and relatively compact blobs that attach near the hip joint (Figure 5.13). Accordingly, they can serve as powerful extensors of the thigh at the hip joint. Th ey are more active when we climb steep gradients, rise from squatting pos-tures, and run than they are during routine walking.65

Th e chimpanzee alternately rotates its entire body about the hip of each supporting foot, which enhances the impression of unsteadiness vis-à- vis smooth human striding.5

Although a human’s center of mass is much higher than that of a monkey, their quadrilaterals of support are similar in area. As long as the center of mass stays within the perimeter of the supporting cheiridia, the creature should not topple over

Because our pelvis is oriented like those of quadrupedal primates, human truncal erectness is chiefl y a suprasacral accomplishment, focused in the lower back.

Our stem ancestors were used to standing and running bipedally on branches and had lower centers of mass. Hence, they stood and moved bipedally on the ground, and, except during infancy, they rarely resorted to terrestrial quadru-pedal locomotion.27 During diurnal respites, they probably squatted or sat on the ground or on tree platforms

Th e protopongids advanced more in the direction of general arboreal climb-ing and engaged in more suspensory behavior than our protohominid ances-tors. Th eir forelimbs and chests enlarged, and their centers of mass were high in the torso. Consequently, when they came to the ground, they tipped for-ward and walked quadrupedally. Because their arboreally adapted fi ngers were long, they fl exed them, thereby paving the way toward knuckle- walking adap-tations in their forelimbs.2

Like orangutans and gibbons, human hands sport no trace of the knuckle- walking manual features. Still, looking toward the savanna and betting on the close ge ne tic similarities between Homo and Pan, Washburn claimed that our ancestors were knuckle- walkers.22 However, I looked to the trees and pro-posed instead that because all apes are climbers, climbing and other arboreal positional behaviors must have predisposed our ancestors for bipedalism.

Bipedalism is not unique to humans, but our par tic u lar form of it is distinc-tive: while most other mammalian bipeds hop or waddle, we stride. Nonethe-less, humans share remarkably similar locomotor neural circuits with other mammals, birds, and at least some reptiles

Further, inspired by his observations of wild gibbons in Th ailand, Keith proposed that brachiation is responsible for the development of many postcranial features that humans share with apes. Keith’s model has three successive stages: hylobatian (i.e., gibbon like), troglodytian (i.e., African ape like), and plantigrade (i.e., bipedal on the ground, like us).

Nonetheless, other explanations on the emergence of bipedalism might be entertained. Th ere is a marvelous variety of positional behavioral models to consider, including Keith’s troglodytian brachiator, Morton’s hylobatian bra-chiator, Washburn’s knuckle- walker, and my hylobatian vertical climber, bi-pedal forager, branch runner, and leaper.

though recent paleoanthropological evidence has uncoupled presumed links among bipedalism, stone tool use, and ce re-bral expansion

Like Proconsul, Sivapithecus lacks the distinctive postcranial traits of knuckle-walkers and acrobatic suspensory apes. However, some features indicate that, like Dryopithecus, they were well equipped for arboreal locomotion and forag-ing.148 Moreover, postcranially Sivapithecus is more derived toward extant hominoid conditions than Proconsul, Dendropithecus, and other Eastern Afri-can Early and Middle Miocene Proconsuloidea and Dendropithecoidea.1

Th e sternum of Pliopithecus is broad, which, like breadth of the lumbar vertebral bodies, implies that the thorax might have been wide transversely as in extant Hominoidea and Atelinae. Nevertheless, unlike modern homi-noids and ateline monkeys, its lumbar region might not be reduced, and its relatively narrow iliac blade argues against a hominoid breadth of the tho-rax.128 Although the lumbar vertebrae of Pliopithecus are more hominoid than those of Proconsul, the caudal reduction had not progressed to the con-dition in Proconsul heseloni and Nacholapithecus kerioi

In Eu rope, the early Middle Miocene postcranial bones of Pliopithecus vindobonensis indicate that it was basically an arboreal quadruped that could climb adeptly and perhaps sometimes foraged in suspended postures. But it had not developed the ricochetal arm- swinging complex, which character-izes modern Hylobatidae.15 Flea gle estimated its body mass to be 7 kg, which

African apes regularly plant their long, wide heels fully on the ground while moving and standing quadrupedally and bipedally. As in human feet, their heels are robust and often neatly rounded. But unlike humans, African apes have a fl exible midtarsal joint that allows the heel to be raised while the distal segment of the sole is still on the ground. Accordingly, they leave fl at- footed impressions.

Th e plantigrade feet of African apes are quite distinct from those of Asian apes and monkeys (Figure 5.20). Indeed, revolutionary evolutionists have cited the resemblance of gorilla and chimpanzee feet to human feet as par-ticularly compelling evidence for our close phylogenic affi nity with the

Manipulation of the wrists of gorillas, chimpanzees, orangutans, and gib-bons shows that the African apes are particularly limited in dorsifl exion (ex-tension), adduction (ulnar deviation), and abduction (radial deviation).

Th e term knuckle- walking derives from the fact that its employers fl ex the fi ngers so that the backs (dorsa) instead of the palmar surfaces contact the substrate.43 Actually, the intermediate phalanges, not the knuckles (i.e., the fi nger joints), are the primary manual contacts during knuckle- walking (Fig-ure 5.15). Th is unique form of dorsal digitigrady contrasts sharply with the palmigrade or ventral digitigrade postures of baboons and other terrestrially adapted monkeys.44Knuckled postures allow long- fi ngered pongid hands, which are useful for arboreal climbing and foraging, to function eff ectively as terrestrial supports and propulsive organs.

Orangutan brachiation diff ers markedly from that of gibbons in that it is nonricochetal.27 Moreover, between hand-holds, orangutans swing the free forelimb overhand, whereas brachiating lesser apes swing it underhand (Figure 5.9). In the lower canopy, where trees are young and supple, hefty males will sway them until they can reach the next support.28Like gibbons, orangutans employ a wide variety of feeding postures, rang-ing from seated to suspensory. Feeding orangutans commonly sit on fi rm branches and hold onto an overhead branch with one hand while gathering food with the free hand. Th eir long reach and great strength enable them to grab, draw in, and break off terminal branches that are inaccessible to many other nonfl ying mammals.29More distally in a tree, an orangutan hangs out from or below a branch, holding it with the ipsilateral hand and foot, while foraging with the oppo-site hand and foot (Figure 5.10). Th ey even hang headlong by their feet alone while manually collecting and opening fruits. However, three- point- suspension feeding postures are more common than two- point- pedal or unimanual postures.30Powerful prehension and overall fl exibility are at a premium for bulky beasts that climb and hang precariously in terminal branches, tens of meters above the forest fl oor.

Gibbons have the second longest hind limbs compared with their truncal height among the anthropoid primates. In this index, only humans surpass them (Figure 2.15). In gibbons, like humans, the quadriceps femoris muscle, which extends the knee, is massive. However, their gluteal and calf muscle

In gibbons, orangutans, African apes, and certain New World monkeys— the spider and woolly monkeys, howlers, and muriquis— the glenoid cavity of the scapula, which articulates with the head of the humerus, is oriented cranially (Figure 5.5). Th is feature facilitates positional behaviors in which the forelimbs are employed overhead.11 Further, in extant apes, the shoulder muscles, which raise the arm at the humeral joint, are well developed, and the clavicle, which struts the arm away from the chest wall for a wide range of movement, is long (Figure 2.6).12Gibbon wrists contain special arrangements of the carpal bones that en-hance rotation.13

Ricochetal arm- swinging is the most dramatic and probably also the most mechanically challenging component of hylobatid positional behavior.2 It is unique to gibbons among the tetrapods. Consequently, it seems reasonable to interpret the anatomical peculiarities of hylobatid forelimbs as features of a derived (apomorphic) adaptive complex for it.

Th e positional behaviors— posture and locomotion— of apes and the mor-phologies that underpin them are of special interest to anthropologists be-cause they might provide clues about the precursors and development of human adaptive complexes.

ubnational variation approachexamines the potentially uneven reach of state institutions through territory.

The weight of the state approach,in contrast, focuses on the effects of the state on societal actors to observe its powerin operation

The national capabilities approach focuses onthe resources at the disposal of the central state, which can be leveraged to exercisecontrol over society and regulate social relation

As the articles in this issue show, the concept of state infrastructural power is apowerful analytical tool when applied with care and precision, since the ability of thestate to penetrate society and implement its chosen policies is relevant to many areasof social research. The conceptual discussion in this article has provided a set ofguidelines for maximizing the payoffs of research into this important aspect of thestate.

Their causal mechanism is that “the reach of the state intorural areas”is “most important for the prospects of a nascent insurgency”(p. 80). Inother words, they argue that the territorial reach of the state is the crucial aspect ofinfrastructural power in determining the emergence of insurgencies. This theoreticalclaim has been echoed by other scholars, but Fearon and Laitin offer the first cross-national statistical investigation of the role of this aspect of state power ininsurgencies. In their statistical analysis, GDP per capita is used as a measure of thereach of the state, and it is shown to be strongly related to the rise of insurgency

scholar takingthe national capabilities approach—the first approach described above—wouldlikely focus on the quantity of tax which the state extracted, or on the size andcapacity of the tax bureaucracy: the resources available to the state for tax collection(Gallo 1991). A focus on the weight of the state as the measure of infrastructuralpower would examine how taxation changed local practices such as patterns oflandholding (Scott 1998).

The territorial reach of the state defines the geographic area within which itspolicies can be enforced. Herbst (2000), who asks why some African states are moreable than others to exercise what he calls “authority over distance,”also explores thisconception of infrastructural power. Many states, particularly in the developingworld, have only limited territorial reach—one example of this is the fact that at least5% of the Peruvian population, mostly in the Amazonian region of the country, wereunable to participate in that country’s 2006 elections because they had not beenissued official identity cards

The pervasive reach of the state, Straus argues, was facilitated by features ofRwanda’s physical and human geography.18 The hilly topography and the heavilysettled land “increase the capacity for surveillance, and they limit the opportunitiesfor exit and escape”(p. 215). As a result, whether sought as victims, or to compelparticipation, “Rwanda’s citizenry is eminently findable.”In Straus’account,Rwanda in 1994 combines a state with high despotic power (the ability to order awide range of policies implemented) with conditions that allowed the enforcement ofeven its genocidal policies. I

tate infrastructural power relates to the set ofrelationships that link these institutions of control to the local communities theypenetrate, and to central state elites.

He argues that Weber’s discussion of the stateincludes three layers: political power, the state, and the modern state. He describespolitical power as the base layer. Fundamentally territorial, political power predatedthe origin of the institution of the state, and is protected by the application of force.The state as the monopoly of legitimate force emerged over time and became layeredover political power, in a process studied by Charles Tilly (1975, 1992) and otherscholars of the emergence of the territorial state. The addition of legitimate authorityand the monopoly of force to the territorial basis of political power institutionalizedthe control of the state

I identify below three analytical approaches to infrastructural power. The firstapproach captures the capabilities of the central state, the second conceptualizesinfrastructural power in terms of its territorial reach,3 and the third focuses on theeffects of the state on society.

In prominent recent scholarship, state strength has been shown to be crucial tooutcomes such as the emergence and persistence of insurgency, economicdevelopment, and the quality of democracy. Two problems characterize the currentapplication of infrastructural power in these literatures. The first problem, addressedin the introduction to this issue, is that many scholars rely on infrastructural powerwithout referring to it as such and without drawing on Mann’s conceptual apparatus.

Three distinct analytical lenses ofinfrastructural power can be derived from his definitions: infrastructural power as thecapabilities of the central state, as the territorial reach of the state, and as the effectsof the state on society.

This then suggests why devising single aggregated measures forquality of governance can be inadequate and misleading

Trying to locate India on this matrix demonstrates some of the com-plexities of this analysis. India is famous both for high levels of cor-ruption and clientelism, and for simultaneously having excessive rulesand bureaucratic red tape. India clearly needs much greater statecapacity across the board. But does it need more or less autonomy? Theanswer to the latter question is probably both, dependent on specificcontext. Given the recent scandal, the agency handling spectrum auc-tions needs to be subjected to much stricter rules; on the other hand,the Hyderabad Municipal Water Authority needs to be relieved of itsmultiple and conflicting political mandates if it is to function properly.

Figure 2 illustrates how the optimal autonomy curves would differ forfour hypothetical countries of differing levels of capacity. For each, thecurve slopes downward at the extremes, because every bureaucracycan have too much or too little autonomy. But the lower-capacitycountries have their inflection points shifted to the left, while they areshifted right for higher-capacity countries

This isbasically the idea contained in Peter Evans’ concept of “embeddedautonomy”: Bureaucrats need to be shielded from certain influences ofsocial actors, but also subordinate to the society with regard to largergoals (Evans 1995

A final measure of the quality of government is the degree of bureau-cratic autonomy possessed by the different components of the state.Samuel Huntington makes autonomy one of his four criteria ofinstitutionalization; highly institutionalized political systems havebureaucracies with high autonomy. The opposite of autonomy in Hun-tington’s terminology is subordination (Huntington 2006).

n fact, it might be better toleave output as a dependent variable to be explained by state quality,rather than being a measure of capacity in itself.

his suggests an alterna-tive measure of government quality, a measure of final output. Onecould look at literacy, primary and secondary education test scores, orvarious measures of health to get some idea as to how governments areperforming.Attractive as output measures sound, there are several big and, in myview, decisive drawbacks to their use. First and most important,outputs like health or education are not simply the consequences ofpublic action; the public sector interacts with the environment aroundit and the society it is dealing with to produce results.

The most commonly used measure of capacity is extractive capacity,measured in terms of tax extraction. Tax extraction measures capacityin two ways: First, it takes capacity, however generated, in order toextract taxes; second, successful tax extraction provides resources thatenable the government to operate in other domains. Tax extractionrates can be measured both by the percentage of taxes to gross domes-tic product, as well as by the nature of taxation—that is, whether it isbased on income or wealth, or indirect taxation (as income and wealthtaxes are much more difficult to extract than indirect taxes)

If we accept this definition of the object we are trying to study, thenthere are at least four broad approaches to evaluating the quality ofgovernance: procedural measures, input measures, output measures,and measures of bureaucratic autonomy.

The current orthodoxy in the development community is that democ-racy and good governance are mutually supportive. I would argue thatthis is more of a theory than an empirically demonstrated fact, and thatwe cannot empirically demonstrate the connection if we define one toinclude the other

As a starting point, I am going to define governance as a government’sability to make and enforce rules, and to deliver services, regardless ofwhether that government is democratic or not. I am more interested inwhat Michael Mann labels “infrastructural” rather than “despotic”power (Mann 1984). T

The bias against thinking about state capacity is particularly strongamong rational choice institutionalists. Most in this school begin withMancur Olson’s assumption that states are predatory and that the chiefaim of political development is the creation of institutions like rule oflaw and accountability that limit the state’s discretion.

We havefewer measures of Weberian bureaucracy—that is, the degree to whichbureaucratic recruitment and promotion is merit based, functionallyorganized, based on technical qualifications, etc. One of the only studiesto attempt to do this was by Peter Evans and James Rauch back in 2000,but their sample was limited to 30 odd countries and produced notime-series data.

In other words, everyone is interested instudying political institutions that limit or check power—democraticaccountability and rule of law—but very few people pay attention tothe institution that accumulates and uses power, the state

It rejects output measures and sug-gests a two-dimensional framework of using capacity and autonomy as ameasure of executive branch quality. This framework explains the conun-drum of why low-income countries are advised to reduce bureaucraticautonomy while high-income ones seek to increase it.

Patrimonialism and Bureaucratization Re-diagnosing problems conventionallyviewed as deficiencies in democratization as issues of patrimonial rule and failuresof bureaucratization is meant to mobilize theories, insights, and data that havelargely remained outside the focus of the research on Latin American politics.