Globalization

Why fewer people died

Three separate things reduced death rates, and they arrived in that order. Sanitation and clean water came first, in the nineteenth century in industrial cities and later elsewhere, and they cut the waterborne diseases the Topic 5.9 chapter describes. Vaccines came next, and the campaigns of the twentieth century were global in a way earlier ones were not: smallpox, a disease that had killed and disfigured people for millennia and devastated the Americas in Topic 4.3, was declared eradicated by the World Health Organization in 1980, the last natural case having occurred in Somalia in 1977. That remains the only human disease eradicated by deliberate campaign.

Antibiotics came third. Alexander Fleming observed penicillin's effect on bacteria in 1928; the work that turned it into a usable drug and then into mass production ran through the early 1940s, driven partly by wartime need. An antibiotic converts a large category of previously fatal events, an infected wound, pneumonia, a difficult childbirth, into treatable ones. Together with vaccination this pushed infant and child mortality down sharply almost everywhere, and life expectancy at birth rose across the century in every world region, though it did not rise equally and the gap between the highest and lowest national figures remained large.

Then births fell

The demographic transition is the pattern in which death rates fall first and birth rates fall afterward, with rapid population growth in the gap between them. It is a description of what has happened in most places rather than a law, and the length of the gap varies enormously: it took over a century in parts of Europe and a few decades in South Korea, Iran and Thailand. Where the gap is short, the population growth is smaller. Where it is long, growth compounds.

Why births fell is genuinely multi-causal, and the honest answer names several mechanisms rather than one. When child mortality falls, parents need fewer births to end up with the same number of surviving children, though the adjustment takes a generation. When children move from farm work to school they become a cost rather than a contribution. When women gain education and paid employment, the opportunity cost of each additional birth rises, and female education is among the strongest and most consistent statistical predictors of lower fertility across countries. And reliable contraception makes the desired number achievable.

The oral contraceptive pill was approved for contraceptive use in the United States in 1960, and more effective methods spread widely, though unevenly and against significant legal and religious opposition in many countries. The College Board's framing is that more effective birth control gave women greater control over fertility and contributed to declining fertility rates, and that word "contributed" is doing correct work. Contraception is a means; whether and how much fertility falls also depends on schooling, employment, child mortality and law, which is why countries with similar contraceptive availability show very different fertility.

The numbers, and what kind of numbers they are

World population was roughly 1.6 billion around 1900 and roughly 2.5 billion in 1950, and United Nations estimates put it past six billion around 1999 and past eight billion in 2022. Global fertility, measured as births per woman, is estimated to have fallen from around five in the early 1950s to under two and a half by the 2010s.

Now the discipline this volume asks for. None of those crossings was observed. There is no world birth registry, and the day the United Nations announced eight billion was a date chosen from a projection, publicized to make a point. The estimates are good, they are the best available, and they are model output with uncertainty attached, revised at every revision. A student who writes "United Nations estimates put world population past eight billion in 2022" is making a claim that survives scrutiny; a student who writes "the world reached exactly eight billion people on that day" is making one that does not.

The distributional point matters more than the total. The growth was not evenly spread, and the projections for the rest of this century concentrate it in sub-Saharan Africa while several countries in Europe and East Asia have populations that are aging and, in some cases, shrinking. Those are two different policy problems arriving at the same time in different places, and they are a substantial part of why migration became the politically charged subject the Topic 9.7 chapter describes.

Disruption is political as often as it is medical

An epidemic does predictable things to a society, and they are worth knowing as a set because they recur. It suspends ordinary economic life, because the measures that reduce transmission are measures that stop people gathering. It concentrates its damage on the people who cannot stay home, which in most epidemics means the poor, and on those in crowded institutional settings. It produces blame, usually directed at outsiders or at a stigmatized group, which is visible in the naming of the 1918 pandemic, in the treatment of gay men and Haitians in the early years of AIDS, and in harassment reported during later outbreaks.

And it strains the relationship between citizens and the state, because public health measures are coercive by nature. Quarantine, contact tracing, compulsory vaccination and closure orders are all exercises of power over individuals justified by a collective benefit, and societies have disagreed about that trade for as long as the measures have existed. Presenting that disagreement as ignorance is a mistake; it is a real argument about the limits of state authority, and it has been made by serious people in every epidemic in this chapter.

The institutions epidemics built

States began meeting about disease when cholera reached Europe: the first International Sanitary Conference was held in 1851, and its logic was frankly commercial, since quarantine of ships was expensive and every state wanted the others to bear the cost. That is the mechanism to keep. International health cooperation grows where an epidemic makes unilateral action useless, and it stalls where it does not.

The World Health Organization was founded in 1948 as a specialized agency of the new United Nations, which is the Topic 9.8 chapter's subject. It ran the smallpox eradication campaign that succeeded and the malaria campaign that did not, and the difference between them is instructive: smallpox had no animal reservoir, an obvious rash that made case-finding possible, and a heat-stable vaccine that gave lasting protection, while malaria had none of those. The severe acute respiratory syndrome outbreak of 2002 and 2003 exposed how little authority the organization had to compel reporting, and the International Health Regulations were revised in 2005 to create the formal emergency declaration used in the Ebola and coronavirus emergencies that followed.

Outbreaks also drive the science. The AIDS epidemic accelerated antiviral drug development and reshaped how clinical trials are run, partly because activist organizations demanded faster access to experimental drugs and won changes to the regulatory process. The Ebola epidemic of 2013 to 2016 pushed a vaccine that had sat undeveloped for years through trials, and it was approved in 2019. Messenger RNA vaccine platforms, under development for years without a commercial application, were authorized against COVID-19 in December 2020, less than a year after the virus was sequenced.

How to quote a death toll

Unit 9 is full of numbers that look precise and are not, and pandemic tolls are the clearest case. There are three different quantities, and the difference between them is not a technicality.

A reported count is deaths recorded and attributed by a health system, and it is limited by testing, by certification practice and sometimes by political pressure. A modeled estimate reconstructs the total from partial data. Excess mortality compares all deaths from all causes against the number that would have been expected, which captures both uncounted cases and deaths caused indirectly by an overwhelmed health system, and misses nothing because it does not require anyone to identify a cause.

So the rule for this whole volume: say what kind of number it is, give a range where the honest sources give a range, and attach a date. "Reported deaths were about five million; the World Health Organization's excess-mortality estimate for the same period was about fifteen million" is a stronger sentence than either number alone, and it is the kind of sentence that earns credit for sourcing.

Three questions, kept apart

Is the climate warming and are human greenhouse gas emissions the main cause of the recent warming? This is a question about physical evidence, and the assessed scientific literature converges on yes. The mechanism was measured in 1859, the concentrations have been recorded continuously since 1958, the isotopic signature identifies the added carbon as fossil, the warming trend appears in surface stations, ocean heat content, satellite records, sea level and ice extent, and the pattern of warming (the stratosphere cooling while the surface warms) matches what greenhouse forcing predicts and does not match what a brighter sun would produce. Every major national academy of sciences and the assessment reports of the intergovernmental panel state this position. Presenting it as an open scientific controversy misrepresents the state of the evidence.

How much will it warm, how fast, and what exactly will happen where? Genuinely uncertain, and scientists say so in print. The panel's sixth assessment gives a likely range rather than a number for how much warming a doubling of carbon dioxide produces, and the projections for particular regions, for the behavior of the great ice sheets, for tipping points and for extreme weather at local scale carry substantial and honestly stated uncertainty. Uncertainty here cuts both ways, which is a point worth making: it includes outcomes worse than the central estimate as well as milder ones.

What should be done, by whom, how fast, and at whose expense? Not a scientific question at all. It involves how much a future harm is worth against a present cost, how much risk a society should accept, which technologies to bet on, how the burden falls across countries and across income groups within them, and what governments should be permitted to require of people. Serious people disagree about all of it, and the disagreement is not evidence of scientific ignorance. It is the ordinary politics of distribution, and it is the part this chapter treats as a debate.

The equity argument, which is the real obstacle

Here is the argument that has stalled climate negotiations since they began, stated from both sides as its holders state it.

Lower-income and industrializing countries point out that the accumulated carbon dioxide in the atmosphere was put there overwhelmingly by the countries that industrialized first, that those countries became rich in the process, and that emissions per person remain far higher in wealthy countries than in poor ones. On that view, asking a country to forgo cheap energy before its population has electricity is asking it to pay for a problem it did not create. The common but differentiated responsibilities principle, written into the framework convention agreed at Rio in 1992, is that argument in treaty language.

Wealthier countries respond that current and future emissions are what determine future concentrations, that emissions growth after the 1990s came predominantly from industrializing economies, that China has been the largest annual emitter since the mid-2000s, and that an agreement binding only some countries cannot solve a problem defined by the global total, while also placing their own industries at a competitive disadvantage. Both arguments are coherent. Neither is refuted by the other, because they are appealing to different principles, historical responsibility against present arithmetic.

Kyoto and Paris: two designs for the same problem

The framework convention of 1992 committed states to the goal and left the mechanism open. The Kyoto Protocol, adopted in 1997 and in force from 2005, chose binding targets, and it applied them to industrialized countries only, in line with the differentiated-responsibility principle. The United States signed and did not ratify, Canada withdrew in 2011, and the countries with the fastest-growing emissions had no targets. Binding commitments turned out to be binding mainly on the states willing to be bound.

The Paris Agreement of 2015 inverted the design. Every party, not only the industrialized ones, submits its own nationally determined contribution, and there is no enforcement beyond publication, review and the reputational cost of missing your own target. The stated aim is to hold warming well below two degrees Celsius above pre-industrial levels and to pursue efforts to limit it to one and a half. The trade is explicit: universal participation was bought by giving up binding targets. Whether that trade was worth making is a live argument, and the honest position as of the mid-2020s is that pledged contributions in aggregate have consistently fallen short of the stated temperature goal, which is exactly what a voluntary regime would be expected to produce and is not, on its own, evidence that a binding regime would have done better, since Kyoto's binding version produced non-participation instead.

Extreme poverty fell, and here is exactly what that means

The World Bank measures extreme poverty as living below an international poverty line, set at 1.90 dollars a day in 2011 purchasing power terms and revised to 2.15 dollars a day in 2017 terms. On that measure the share of the world's population in extreme poverty fell from roughly a third or more around 1990 to under a tenth by the late 2010s, and the absolute number fell as well despite population growth. That is a very large change and it should not be minimized.

Now the qualifications a careful answer includes, none of which reverses the finding. The line is extremely low and describes destitution rather than adequacy; the number of people above it and still poor is much larger. The figures come from household surveys of varying quality and coverage, adjusted by purchasing power estimates that are themselves contested and revised. And a large share of the global decline is China, whose extreme poverty fell by hundreds of millions after 1978, so a claim that "globalization reduced poverty" is doing much of its work through a single country whose policy mix, as section 02 showed, was not the standard prescription.

Inequality moved in two directions at once

This is the fact that makes the whole argument confusing, and stating it clearly is worth a great deal. Between countries, inequality narrowed over the three decades after 1990, because China and India grew much faster than the wealthy economies. Within countries, inequality widened in a large number of them over the same period, including most of the wealthy economies and China itself. So a person asking "did inequality rise or fall" is asking a question that has two correct and opposite answers depending on the unit of analysis.

The most cited attempt to put both in one picture is the graph of real income growth by global income percentile between the late 1980s and the late 2000s produced by Christoph Lakner and Branko Milanovic, often called the elephant curve. It shows strong income growth for the middle of the global distribution, largely people in Asia, strong growth at the very top, and weak growth around the eightieth to ninetieth percentiles, which is roughly where the working and lower-middle classes of wealthy countries sit. The interpretation, that globalization lifted the global middle and the global rich while stagnating the rich world's workers, is powerful and it is contested: critics including Adam Corlett in 2016 showed that changes in which countries appear in the sample, especially Japan and the former Soviet bloc, account for a substantial part of the dip. Cite the curve, and cite the critique, and you are doing history rather than repeating a graph.

The unresolved argument about the lost jobs

Manufacturing employment fell substantially in the United States and in several European countries from around 2000, and the political consequences of that decline run through the Topic 9.7 chapter and beyond. Why it fell is a genuine and unresolved dispute among economists, and a student who presents it as one is doing exactly what the complexity criterion asks.

The trade case is associated with David Autor, David Dorn and Gordon Hanson, whose work from 2013 onward compared American local labor markets by their exposure to Chinese import competition and found large, persistent and geographically concentrated losses in employment and wages, with estimates in the range of two million or more manufacturing jobs lost to that competition in the 2000s, and slow adjustment in the affected places. The automation case points out that American manufacturing output kept rising while employment fell, which is the signature of rising productivity rather than of lost production, and studies such as one from Ball State University in 2015 attributed the large majority of job losses in the 2000s to productivity growth rather than trade.

The two are not fully separable, and that is the honest conclusion rather than a dodge. Import competition raises the pressure to automate; automation raises the productivity that makes competing possible. What the evidence supports is that both mattered, that their relative weight is disputed in the peer-reviewed literature, and that the local concentration of the losses, whatever their cause, is what made them politically explosive. Aggregate gains distributed thinly across consumers and concentrated losses landing on particular towns is a distribution problem, and it is the same one this chapter opened with.

Three ways to campaign without a government to petition

Attach the global claim to a local one. The Chipko movement, from 1973 in the Himalayan foothills of what is now Uttarakhand, opposed commercial logging by villagers physically placing themselves between the trees and the contractors. Its demands were about local rights to forest resources, and its effect included a moratorium on felling in the region. Wangari Maathai's Green Belt Movement, founded in Kenya in 1977, organized rural women to plant trees, which addressed fuelwood, erosion and income simultaneously and put those women into an organized relationship with the state; Maathai received the Nobel Peace Prize in 2004. Both are environmental campaigns that worked because they were also livelihood campaigns, and neither began by addressing an international body.

Make the global harm visible. Greenpeace, founded in 1971 when activists sailed toward a United States nuclear test site off Alaska, built its method on producing images that carried an argument. Sailing a small boat between a whale and a harpoon, or between a test site and a warship, generates a photograph that needs no explanation, and the target of the campaign is public opinion in the countries whose governments and companies are responsible. The method's strength is reach and its weakness is that a photograph makes an argument about a moment and not about a system.

Use the market as the lever. Fair trade certification, developed from the late 1980s with the first labeling initiative launched in the Netherlands in 1988 and international bodies including the World Fair Trade Organization following, tries to route a higher and more stable price to producers by giving consumers a label to choose. Independent evaluations have generally found modest and real benefits for participating producers, concentrated in cooperatives, and limited effects on the wider market, which is roughly what the mechanism would predict: a scheme that depends on consumers paying more can only reach as far as consumers willing to pay more.

What the record supports about when reform movements succeed

Pulling the whole chapter together, the campaigns in it did better under four conditions, and this is a genuinely useful analytical tool rather than a list.

They did better when there was a specific rule to change rather than a general condition to improve, because a rule has a decision-maker attached. They did better where the target cared about reputation, whether a government seeking loans and legitimacy or a brand selling to consumers who read labels. They did better where the movement could impose a concrete cost, a boycott, a strike, a divestment campaign, litigation, rather than only moral pressure. And they did better where they could point to a standard the target had already accepted, which returns to section 01.

The apartheid case has all four, which is why it appears in every textbook, and it is worth noticing that it also took over four decades. The campaigns that changed the distribution of economic costs have generally had the fewest of the four, which is a large part of why the Topic 9.7 chapter is about people who went from petitioning to protesting in the street.

What actually crossed: brands, formats and platforms

Coca-Cola was first sold in 1886 and by the second half of the twentieth century was on sale in most countries of the world, common enough that critics in postwar France coined Coca-colonization for the fear that American products were displacing local ways of living. Toyota, founded in 1937, sold cars on every continent through production and marketing networks that were not American at all, which is the counterflow again. McDonald's opened its first restaurant outside North America in the 1960s and opened in Moscow in January 1990, a date that carried more political meaning than any menu. eBay launched in the United States in 1995 and Alibaba was founded in Hangzhou in 1999, and between them they turned cross-border consumer exchange from something a person watched into something a person could do.

Now separate three things that get merged. The product crossed: the same bottle, the same car, the same handset. The format crossed: the ninety-minute feature, the three-minute pop single, the twenty-two-minute sitcom, the standardized fast-food service model, the platform interface. And the ownership concentrated: a shrinking number of firms controlled a growing share of what was distributed, a claim the American critic Ben Bagdikian tracked across editions of his book from 1983, counting the companies dominating that country's media falling from dozens to a handful over two decades. Those three are all real convergence and they are convergence in the machinery, which is the strongest version of the homogenization case.

The dispute, with names on it

The cultural imperialism position was set out most influentially by Herbert Schiller in 1969, arguing that the international flow of media ran overwhelmingly outward from the United States and carried commercial values with it, so that other societies' media systems were being restructured around advertising and American formats. Its evidence is trade data on media exports, ownership concentration, and the spread of the commercial broadcasting model. George Ritzer's argument from 1993 about McDonaldization is a related but distinct claim: what spreads is not American taste but a form of organization, efficiency, standardization, predictability and control, applied to more and more of life.

The opposing position rests on what audiences do with what they receive. Elihu Katz and Tamar Liebes published a study in 1990 that showed the same episode of an American television drama to viewers of different backgrounds in Israel and elsewhere and recorded how they discussed it; the groups produced systematically different readings, moral, political and narrative, of an identical text. James Watson's collection of ethnographies of McDonald's in East Asia in 1997 found the restaurants being used in locally specific ways, as places for students to sit for hours or for children's birthday parties, functions the company had not designed. Roland Robertson proposed glocalization in 1995 for the deliberate adaptation of a global product to a local market, and Arjun Appadurai argued in 1996 that global cultural flows are disjointed rather than coordinated, so that media, money, people, technology and ideology move in different directions at once and no single center controls the result.

What can be documented

Institutional procedure. The World Bank withdrew from Sardar Sarovar in 1993 after the Morse review, and created the Inspection Panel the same year. Environmental and resettlement safeguards were formalized across development lending in the following decade.

Debt. The Jubilee 2000 campaign, built out of churches, unions and development organizations and reported to have gathered over twenty million signatures worldwide, pressed for cancellation of the debts of the poorest countries. The Heavily Indebted Poor Countries initiative launched in 1996 and was enhanced in 1999; the Multilateral Debt Relief Initiative followed in 2005. Debt was canceled for more than thirty countries, and the relief was tied to conditions of its own, which critics noted at the time.

Medicines. In 1998 a large group of pharmaceutical companies sued the South African government over a law intended to allow cheaper access to HIV medicines. A sustained campaign by South African treatment activists and international organizations made the case a public confrontation, and the companies dropped it in April 2001. In November 2001 the WTO ministerial at Doha adopted a declaration affirming that the intellectual property agreement should be interpreted to support public health, including the right to issue compulsory licenses. That is a change to how a binding international rule is read, produced by campaigning, and it is the strongest single item on this ledger.

Supply chains. Campaigning against conditions in contract factories from the mid-1990s produced codes of conduct, monitoring bodies and auditing regimes across the apparel and footwear industries. After the collapse of the Rana Plaza building in Bangladesh in April 2013 killed over a thousand garment workers, a legally binding agreement on building and fire safety was signed the following month by a large group of mostly European brands together with global unions, while a separate and non-binding body was formed by mostly North American brands. That contrast, binding versus voluntary, is the most useful thing in this paragraph, because it lets a student compare two responses to one disaster.

What did not change, and the honest reckoning

The structural features the movements attacked mostly persisted. Capital remained far more mobile than labor, which the Topic 9.4 chapter identifies as the fact that decides most of the distribution. Conditionality continued in a modified form. Investor protections abandoned with the Multilateral Agreement on Investment in 1998 reappeared, chapter by chapter, inside bilateral and regional trade and investment treaties, where they attracted much less attention than a single global agreement had. And the movement's own great success at Seattle and Cancun, blocking multilateral rounds, contributed to negotiation shifting into those bilateral and regional deals, where a poorer country bargains alone against a larger one instead of within a bloc. That is a genuinely uncomfortable outcome and it belongs in the answer.

Set that against the four conditions Topic 9.5 derived for when a campaign succeeds: a specific rule to change, a target that cares about reputation, an ability to impose a concrete cost, and a standard the target has already accepted. Every item on the documented list above has at least three of the four. The medicines case has all four, which is why it worked. The general campaign against the structure of the global economy has, at best, one, which is why it did not.

Writing about the present without leaving history

This topic runs into live politics, and the rule that keeps an answer defensible is to make dated structural claims and to leave contested judgments as contested. What the historical record supports is this: through the 2000s and 2010s the political salience of trade, migration and the authority of international bodies rose in a number of wealthy democracies, and questions about who sets rules and who controls movement featured in major votes, including the United Kingdom's referendum on European Union membership in June 2016. Why voters decided as they did is studied by political scientists and genuinely disputed, with economic dislocation, cultural and demographic change, party competition and the campaigns themselves all argued for and none of them established as decisive.

So write the structure and date it, and attribute the interpretation. The concentrated losses that Topic 9.4 documents were real and geographically specific, and a constituency shaped by them existed; whether that is the main explanation for any particular election is a question this course cannot settle and should not pretend to.

The axis, and three tests of it

A coordination problem is one where every party gains from a common rule and no party gains by defecting once the rule exists. Which side of the road to drive on, which frequency a distress call goes out on, which format an airport uses for a flight plan. There is no incentive to cheat, so an institution that merely publishes the standard is sufficient, and no enforcement is needed. That is why the ITU, ICAO and the postal union work through every diplomatic crisis of the century.

A collective action problem is different: every party gains if all comply and each individual party gains more by defecting while others comply. Emissions, fisheries, arms control. Here publishing a standard is not enough, and the institution needs monitoring, and ideally a penalty. The Montreal Protocol of 1987 on ozone-depleting substances, which the Topic 9.3 chapter treats in detail, is the case where this was achieved: a small number of producers, cheap substitutes available, clear measurement, and trade restrictions against non-parties, and it has been ratified by every United Nations member state.

Climate is the same category and harder on every dimension, and the treaty design shows it. The Kyoto Protocol of 1997, in force from 2005, set binding targets for a subset of countries, and its central weakness was that the states with the fastest-growing emissions had no targets and the largest emitter at the time never ratified. The Paris Agreement of December 2015 made the opposite trade: it is nearly universal, and its national commitments are self-determined and carry no penalty for missing them, relying instead on transparency, periodic review and pressure. Whether that trade was right is genuinely argued. What is not in doubt is that it is a design response to the collective action problem, choosing participation over enforcement, which is exactly the trade the 1945 founders made with the veto.

And a compulsion problem is when the institution must make a state do something against its judgment of its own vital interests. Nothing built by sovereign states does this reliably to a powerful state, and the record is unambiguous. The International Criminal Court, established by the Rome Statute of 1998 and operating from 2002, has jurisdiction over the nationals and territory of states that joined, or over situations referred by the Security Council; the United States, China, Russia and India are not parties. The International Court of Justice hears cases between states that have consented to its jurisdiction. Both are real institutions doing real work within a boundary drawn by consent.

The one place sovereignty was actually pooled

The European experiment is the exception that shows what the rest of the system did not do. Six states pooled authority over coal and steel in 1951, formed a common market by the Treaty of Rome in 1957, and through decisions of the European Court of Justice in the early 1960s accepted that European law takes precedence over conflicting national law within its fields and can be invoked by individuals in national courts. The Maastricht Treaty signed in February 1992 and in force in November 1993 created the European Union; the Schengen arrangements removed internal border checks among participating states from 1995; euro notes and coins circulated from 2002.

That is a genuine transfer of sovereignty rather than a coordination agreement, and the conditions under which it happened are worth listing because they are so restrictive: a small initial group, comparable levels of wealth, a shared security threat, democratic governments, and a specific memory of what the alternative had cost. Other regional bodies have taken the coordination route instead. The Association of Southeast Asian Nations from 1967 is built explicitly on non-interference; the African Union, which replaced the Organization of African Unity in 2002, went further than its predecessor in providing for intervention in extreme cases; Mercosur from 1991 and the North American agreement in force from 1994, replaced by its successor in 2020, are trade arrangements rather than political unions. The variation is the evidence: states pool sovereignty rarely, and only in conditions that mostly do not obtain.

The dispute, and the actors that are not states

Whether international institutions matter is a live argument in the study of international politics, and both sides accept the same facts. The realist position, argued most sharply by John Mearsheimer in 1994, is that institutions reflect the distribution of power rather than altering it, that states comply when compliance suits them, and that expecting more is a false promise. The institutionalist position, associated with Robert Keohane's work from 1984, accepts that states are self-interested and argues that institutions still change outcomes by reducing the cost of making agreements, supplying information about who is complying, and making relationships repeat, so that a state that cheats today pays for it in the next negotiation. Notice that Keohane's mechanism is precisely the coordination axis above, and that Mearsheimer's evidence is drawn from the compulsion end. They are largely right about different parts of the same system, and saying so is a better answer than picking a side.

The last piece is the actors states did not create. The International Committee of the Red Cross dates from 1863, Amnesty International from 1961, and Doctors Without Borders from 1971, the latter founded partly on the argument that neutrality can amount to silence. These organizations have no territory, no vote and no army, and they matter through the mechanism Topic 9.5 identifies: documentation and publicity against a standard the target has accepted. Multinational corporations became global actors of a different kind, several with revenues larger than the output of many states, and the Topic 9.4 chapter explains how mobile capital shifts bargaining power away from governments. The map of world affairs, drawn in 1900 almost entirely between governments, contains by 2000 a great many actors that no government fully controls.

The vocabulary, and the two words that decide the grade

Topic 7.9 supplies the weighing language for ranking causes and it transfers directly. Add the two constructions this topic specifically needs. "Measured by X, the change was decisive; measured by Y, the continuity was." That sentence names two criteria and commits to both, which is what an extent question is asking for. And "the change was in the machinery and not in the structure", or its reverse, which names the layer.

The two words that decide the grade are extent and compared. If your answer does not contain a comparison and a standard, it has not answered an extent prompt however much it knows. And the phrases to avoid are the ones that concede the question: "there were many changes", "it was a mix of both", "in some ways yes and in some ways no". Each of those is true of every question ever asked and therefore says nothing.

A worked thesis

Take the prompt: Evaluate the extent to which science and technology brought change in the period from 1900 to the present.

Weak. "Science and technology changed the world enormously after 1900 through inventions like the internet, antibiotics, and the Green Revolution, although there were still problems like inequality." That is a list with a concession attached. It names no baseline, no criterion and no population, and it could have been written before the unit.

Strong. "Measured by depth of change in an ordinary life, the transformation after 1900 was among the most complete in human history: United Nations estimates put global life expectancy at roughly forty-six years in 1950 and roughly seventy-two by the late 2010s, driven by sanitation, vaccination and antibiotics rather than by any change in the social order. Measured instead by reach into the structures of power, the century changed remarkably little: the sovereign state multiplied and strengthened rather than eroding, the share of the world's population living outside its country of birth rose only from under three to about three and a half percent between 1990 and 2020 while goods and capital moved freely, and inequality between countries narrowed only as inequality within them widened. The most defensible position is therefore that technology transformed the conditions of life and left the distribution of power substantially intact, and the clearest evidence for that claim is that every technology in Unit 9 required a political decision to determine who received it, from the Green Revolution package reaching irrigated Punjab and not rain-fed regions without credit, to antiretroviral drugs reaching wealthy countries years before the regions worst hit by the epidemic."

Look at what that does. It names two criteria and gives the answer under each. It gives a baseline and dates every figure. It says what kind of numbers they are. It commits to a position rather than balancing. It supplies a mechanism, political decision determining distribution, rather than a list. And it is arguable: a reader could reasonably respond that a change in who survives childhood is a change in the structure of power, and that response would be a good essay too. Arguable is the property you are being graded on.

What to walk in holding

This is the last page of the course, so here is the compression. Four sentences, and they will carry you through most of what the exam can ask.

Connection is old; its price is what changed. Goods, faiths, technologies and diseases crossed Afro-Eurasia in the thirteenth century, and the circuit closed around the world from 1571. What the last two centuries changed is the cost of moving a thing, an idea or a person, and almost every question in Units 5 through 9 is downstream of that.

A name is not a mechanism. Whatever you are asked about, the marks are in how it worked, for whom, and at whose expense. The container saved money on the dock. The declaration worked because a state can be held to what it published. Conditionality moved a decision out of reach of a vote. Say the mechanism and the evidence stops being a list.

Every gain has a bill, and it usually lands somewhere else. Synthetic nitrogen and the dead zone, petroleum and the atmosphere, cheap goods and a closed plant, a cure that exists and a child who cannot reach it. Naming the gain and the bill in one sentence is the qualification criterion, and it works on nearly every prompt in the course.

Say what you are comparing to. Baseline, criterion, population; level of claim; criterion of significance. Every method chapter in this course has been teaching one version of the same habit, and it is the habit that turns knowing things into arguing something. That is the whole of the skill, and you have it.