33.08.02 · history-of-science / contemporary-science

Big science and global collaboration: CERN, space race, Human Genome Project

stub3 tiersLean: nonepending prereqs

Anchor (Master): Weinberg, A. M. — Impact of Large-Scale Science on Science (1961)

Intuition Beginner

After World War II, science changed scale. The lone genius in a small lab gave way to billion-dollar projects staffed by thousands of researchers. Alvin Weinberg named this new mode "big science" in 1961. The shift reshaped not just how discoveries were made, but who paid for them, who received credit, and what counted as a scientific result.

CERN, the European Organization for Nuclear Research, was founded in 1954 to rebuild European physics through cooperation. Its Large Hadron Collider is a 27-kilometre ring beneath the Swiss-French border. About 10,000 scientists from 100 countries work on its experiments. In 2012 its ATLAS and CMS detectors found the Higgs boson, the final missing piece of the Standard Model of particle physics. Each discovery paper carried more than 3,000 author names.

The space race was Cold War competition fought with rockets instead of armies. The Soviet satellite Sputnik shocked America in 1957. President Kennedy pledged a Moon landing "before this decade is out," and NASA's Apollo program delivered. In 1969 Neil Armstrong and Buzz Aldrin walked on the lunar surface. The Saturn V rocket was built by Wernher von Braun, a German engineer brought to America after designing weapons for Nazi Germany — a moral complication the celebration usually ignored.

The Human Genome Project ran from 1990 to 2003. A coalition of 20 institutions across six countries spent about three billion dollars to read all three billion letters of human DNA. The project released its data publicly every day. That policy fueled genomic medicine and set a template for open science (see 33.08.01).

In 2015 the LIGO observatories detected gravitational waves — ripples in spacetime that Einstein had predicted a century earlier. The detection capped 40 years of effort and more than a billion dollars of investment. Big science raises hard questions. Who decides which projects get funded? Can a paper with 3,000 authors represent genuine discovery? Has the scale of modern instruments changed what science even is?

Visual Beginner

Project	Lead organization	Years	Scale	Key result
CERN / LHC	CERN (intergovernmental)	1954–; LHC 2008–	~10,000 scientists, 100 countries; 27 km ring	Higgs boson (2012)
Apollo	NASA (US national)	1961–1972	~400,000 workers; ~$25B in 1960s dollars	Humans on the Moon (1969)
LIGO	NSF (US, with partners)	1992–; first detection 2015	~$1.1B; 40+ year effort	Gravitational waves (GW150914)

Worked example Beginner

LIGO — the Laser Interferometer Gravitational-Wave Observatory — shows how big science actually unfolds. The idea of detecting gravitational waves dates to the 1960s. Joseph Weber claimed a detection in 1969 using resonant aluminum bars. Other laboratories could not reproduce his results, and the claim collapsed into controversy. The episode taught the field that a real detection would require far larger and more expensive instruments.

LIGO was proposed in 1989 and funded by the US National Science Foundation in 1992. Two interferometers were built — one in Hanford, Washington, one in Livingston, Louisiana — each with arms four kilometres long. Construction and the first observing runs, from 2002 to 2010, detected nothing at all. A major upgrade called Advanced LIGO then began.

On 14 September 2015, days after Advanced LIGO switched on, both detectors registered the same signal: GW150914. Its shape matched the predicted waveform of two black holes merging 1.3 billion years ago. The announcement paper carried around 1,000 authors. In 2017 the LIGO and Virgo detectors caught two neutron stars colliding, and telescopes across the world turned to watch the afterglow — the birth of multimessenger astronomy (see 28.04.04, 28.06.02).

From idea to detection took more than 50 years and over a billion dollars. No lone genius made the discovery. It required physicists, laser engineers, vacuum technicians, statisticians, and administrators coordinating across decades and continents. That is what big science means in practice.

Check your understanding Beginner

Exercise (easy, multiple choice).

Who coined the term "big science" and when?

A. Vannevar Bush, in his 1945 report Science: The Endless Frontier B. Alvin Weinberg, in a 1961 article in Science C. Derek de Solla Price, in his 1963 book Little Science, Big Science D. John F. Kennedy, in his 1961 Moon-landing speech

Hint

One of these men was an Oak Ridge physicist reflecting on whether national laboratories and giant accelerators were crowding out smaller, curiosity-driven research.

Answer

Option B.

Alvin Weinberg, director of the Oak Ridge National Laboratory, published "Impact of Large-Scale Science on Science" in Science in 1961. He named and analysed a pattern he saw in post-war research: accelerators, reactors, and space programs so expensive they demanded national commitment. Derek de Solla Price's Little Science, Big Science (1963) then gave the idea its first systematic quantitative treatment.

Exercise (easy, multiple choice).

What did the Human Genome Project's "Bermuda Principles" (1996) require?

A. That sequenced human DNA be kept confidential for five years before release B. That DNA sequence data be released publicly within 24 hours of generation C. That human genes be patented by the funding governments D. That only American laboratories could sequence the human genome

Hint

Think about the contrast between the public project and its private competitor, Celera, which withheld its data.

Answer

Option B.

The Bermuda Principles committed the public Human Genome Project to releasing its sequence data into the public domain within 24 hours. This daily, no-strings-attached release was a deliberate counter to private, proprietary sequencing. The policy set a precedent for open science and kept the human genome unpatentable as a resource for everyone (see 33.08.01).

Formal definition Intermediate+

The study of big science has no standard formal apparatus of its own, but several quantitative concepts let us define its structure precisely. None of this is formalised in Mathlib; the definitions below are stated in ordinary mathematical language and are reviewer-attested.

Scale parameters of a research project. For a project $P$ define four scalar observables: total lifecycle cost $C (P)$ , number of contributing personnel $N (P)$ , a characteristic instrument scale $L (P)$ (an accelerator's energy, a telescope's mirror diameter, a genome's read length), and project duration $T (P)$ . Following Weinberg (1961), $P$ is big science when $C (P)$ or $N (P)$ exceeds the corresponding budget or headcount of an ordinary university laboratory by several orders of magnitude, so that the project requires state-level commitment, capital infrastructure, bureaucratic management, and multidisciplinary teams. Little science, by contrast, is the regime in which an individual researcher or small group works in a single laboratory on curiosity-driven questions (see 33.05.* on the Cavendish-Lab quantum revolution, and 33.06.* on the Watson-Crick-Franklin double-helix work — both paradigm cases of small science).

Collaboration network. Model the collaboration behind a big-science result as an undirected graph $G = (V, E)$ on $N$ researchers, with an edge joining two researchers who co-authored at least one paper or shared a subsystem responsibility. The number of potential pairwise ties is $(2 N) = \frac{1}{2} N (N - 1)$ , and the number of potential working subgroups is $2^{N}$ . This combinatorial explosion is the structural reason that coordination, rather than individual insight, becomes the binding constraint as $N$ grows — the origin of the bureaucratic apparatus Peter Galison called a "trading zone."

Authorship scaling. Let $a (t)$ be the mean number of authors per paper in a field. Empirically $a (t)$ grows roughly linearly or faster in big-science fields; particle-physics discovery papers now carry more than 3,000 names. When $a (t)$ vastly exceeds the number of distinct intellectual contributions a reader can evaluate, the link between a published result and any individual's reasoning is dissolved, and credit must be allocated through institutional and collective mechanisms rather than by reading a single name.

Price's growth law. Derek de Solla Price (Little Science, Big Science, 1963) measured that both the number of scientists and the volume of publications have grown exponentially with doubling time $T_{2} \approx 15$ years, so that $S (t) = S_{0} \cdot 2^{t / T_{2}}$ . Exponential growth of this kind is the background condition against which the transition from little to big science took place: as the population of researchers and the cost of frontier instruments both rose, large-scale coordination became both possible and, eventually, unavoidable.

Key theorem with proof Intermediate+

Theorem (Price's saturation bound). Let $S (t) = S_{0} \cdot 2^{t / T_{2}}$ be the number of scientists and let $P (t) = P_{0} \cdot 2^{t / T_{P}}$ be the human population, with $T_{2} < T_{P}$ — scientists doubling faster than the population. Then there exists a finite time $t^{*}$ at which $S (t^{*}) \geq P (t^{*})$ . Since this would require every human being to be a scientist, it is impossible, and so the exponential law $S (t) = S_{0} \cdot 2^{t / T_{2}}$ must break down no later than $t^{*}$ .

Proof. We seek $t^{*}$ with $S (t^{*}) = P (t^{*})$ , i.e. $S_{0} \cdot 2^{t^{*} / T_{2}} = P_{0} \cdot 2^{t^{*} / T_{P}}$ . Taking base-2 logarithms,

lo g_{2} S_{0} + \frac{t ^{*}}{T _{2}} = lo g_{2} P_{0} + \frac{t ^{*}}{T _{P}},

t^{*} (\frac{1}{T _{2}} - \frac{1}{T _{P}}) = lo g_{2} (\frac{P _{0}}{S _{0}}), t^{*} = \frac{lo g _{2} ( P _{0} / S _{0} )}{1/ T _{2} - 1/ T _{P}} .

Because $T_{2} < T_{P}$ , the denominator $1/ T_{2} - 1/ T_{P}$ is strictly positive, and $t^{*}$ is finite. At any time $t > t^{*}$ the model predicts $S (t) > P (t)$ , an impossibility, so the exponential law cannot persist beyond $t^{*}$ . This completes the argument.

Numerical illustration. Taking $S_{0} \approx 1 0^{7}$ scientists and $P_{0} \approx 3 \times 1 0^{9}$ people near the mid-20th century, $T_{2} \approx 15$ years and $T_{P} \approx 50$ years, we get $lo g_{2} (P_{0} / S_{0}) \approx lo g_{2} 300 \approx 8.2$ and $1/ T_{2} - 1/ T_{P} \approx 0.047 yr^{- 1}$ , giving $t^{*} \approx 175$ years. Price's point is qualitative rather than predictive: an era of sustained exponential growth in science is historically bounded, and must give way to a slower, saturating curve.

Corollary (cost-driven saturation of frontier science). If the unit cost per frontier result also rises — as the Livingston chart shows for accelerators, whose energy frontier has grown exponentially at a cost that grows at least as fast — then the total cost of frontier discovery grows faster than $S (t)$ . The implication, recognised by Weinberg and sharpened by every subsequent generation of facility planning, is that big science encounters diminishing returns: each new regime of discovery demands an instrument proportionally harder to fund. This is the structural pressure behind the recurring question of whether a 100 km, 100 TeV Future Circular Collider can or should be built (see the master discussion below), and behind the broader worry that particle physics may be approaching an impasse.

Exercises Intermediate+

Exercise (medium, short answer).

Weinberg proposed three criteria for choosing between big-science projects: scientific merit, technological merit, and social merit. Explain each criterion and give one reason a project might score highly on one and poorly on another.

Answer

Scientific merit asks whether the project addresses an intellectually deep question (does it extend a fundamental theory, close a gap, open a new field?). Technological merit asks whether the project builds instruments, methods, or trained people that will be useful beyond the result itself. Social merit asks whether the knowledge or technology produced serves a legitimate public need.

The Apollo program is the canonical example of a mismatch: very high technological and social merit (it built an aerospace industry and demonstrated national capability) but contested scientific merit, since robotic missions could have answered many of the same lunar-science questions far more cheaply. A pure-dark-matter direct-detection experiment inverts the pattern: very high scientific merit, modest technological spillover, and diffuse social merit.

Exercise (hard, essay).

Philip Kitcher argues for "well-ordered science": research priorities should be set by a process that would be endorsed by fully informed citizens representing all affected groups. Apply Kitcher's ideal to one current big-science proposal (the Future Circular Collider, a crewed Mars mission, or a large-scale geoengineering experiment). Identify at least two ways the actual decision process falls short of "well-ordered" and one institutional reform that would move it closer.

Hint

Consider who sits on the funding bodies, what counts as a relevant "affected group," and whether non-scientists have any real voice in setting priorities.

Answer

Take the Future Circular Collider. The decision is dominated by the particle-physics community and by national funding agencies whose advisory panels are staffed largely by physicists. Two shortfalls against well-ordered science stand out. First, the affected community is far wider than the user community: taxpayers who would fund the facility, researchers in competing fields whose share of a fixed budget would shrink, and citizens of host countries affected by construction all have stakes, but only the first group is represented in any technical sense. Second, the framing of the decision is narrowed by disciplinary incentives — physicists have strong professional reasons to favour a next collider, so the process lacks the disinterested scrutiny Kitcher requires.

A reform moving closer to well-ordered science would be a participatory priority-setting body with deliberative seats for non-scientist citizens and for researchers from neighbouring fields, charged with ranking big-science investments against alternative uses of the same public funds on a common scale of public value.

Advanced results Master

Big science and the transformation of scientific practice

The deepest analysis of how big science changes the practice of research is Peter Galison's Image and Logic (1997). Galison argued that a modern experiment is not the work of a single mind confronting nature but a "trading zone" — a contact surface where physicists, engineers, technicians, programmers, and administrators speak partially overlapping jargons and coordinate through local pidgins, shared instruments, and negotiated procedures rather than through a common theoretical language. These trading zones connect to the broader sociology of organisations (see 30.01.*): big science is bureaucratic science, with written procedures, division of labour, and chains of accountability that resemble industrial production more than the Romantic image of the solitary discoverer.

This raises a problem about observation itself (see 20.08., 29.03.). When a Higgs-boson "event" is a statistical excess reconstructed by software from petabytes of detector data, what is the object of observation? The detector, the trigger, the reconstruction algorithm, and the statistical procedure are all part of the instrument. Scientific realism and structural realism (see 20.08.02) have to confront the fact that in big science we rarely "see" the phenomenon directly; we see a heavily mediated signal whose trustworthiness rests on trusting thousands of people and millions of lines of code (see 33.07.* on the computing infrastructure that makes this possible).

Science funding and the state

Big science is inconceivable without state funding, and the institutional architecture was built quickly after World War II. Vannevar Bush's 1945 report Science: The Endless Frontier argued that basic research is a public good the federal government must fund, and its recommendations led to the creation of the National Science Foundation in 1950 and the dramatic expansion of the National Institutes of Health. The result was a military-science complex rooted in the Cold War (see 32.22.): the Department of Defense's Advanced Research Projects Agency (ARPA, now DARPA) funded the ARPANET that became the internet (see 33.07.), and advisory bodies like the President's Science Advisory Committee and the JASON group embedded scientific judgement in national-security decision-making (see 30.05., 20.07.).

The space race was the most visible expression of this complex. It was ideological proxy war (see 32.17., 32.21.): Sputnik demonstrated that the same rockets that could orbit a satellite could deliver a warhead, and Apollo answered with a demonstration of industrial capacity. The same logic shaped the darker side of big science — nuclear weapons, born of the Manhattan Project (see 32.22.*) and shadowed ever after by the Bulletin of the Atomic Scientists' Doomsday Clock, which reframes the 20th century as the first era in which humanity acquired the means of its own extinction (see 28.05.03 on the Fermi paradox and the Great Filter).

International collaboration as a counter-model

CERN was founded in 1954 with an explicit dual purpose: to rebuild European physics after the war, and to bind former adversaries together through civilian scientific cooperation so that the continent would not again tear itself apart (see 32.20., 32.22.). Its governance is intergovernmental — member states fund it by treaty — and its outputs are open and non-military. CERN has become the reference model for international big science, and its innovations extend to the organisation of knowledge itself: it pioneered large-scale open-access publishing through SCOAP3 and, at the application layer, gave the world the World Wide Web (see 33.07.01).

Other collaborations extend the model. The International Space Station (1998–2024) joined the United States and Russia with Japan, Europe, and Canada — former Cold War enemies cooperating in orbit (see 28.06.). ITER, the fusion reactor under construction in France, unites 35 nations around a shared plasma-physics goal (see 11. on statistical mechanics and plasma, and the long-standing promise of fusion energy). The Intergovernmental Panel on Climate Change synthesises the work of thousands of scientists across every discipline that bears on the climate (see 27.07., 20.08.). The Square Kilometre Array will pool radio-astronomy capacity across continents (see 28.06.02, 28.04.*). And the Human Genome Project, discussed below, showed that even competitive, commercially pressured science could be organised as an open international commons.

Big science and its critics

Weinberg himself was an early critic. His 1961 essay asked not whether big science was possible but whether it was wise, proposing the scientific, technological, and social criteria discussed in the exercises. The worry is that large facilities become self-perpetuating: once a multibillion-dollar accelerator or telescope exists, the community that built it has a strong interest in securing its successor, which can crowd out small, curiosity-driven, high-risk research — the research that historically produces the largest conceptual surprises.

Daniel Greenberg developed this critique in The Politics of Pure Science, arguing that concentrated funding produces a politics of expertise in which the priorities of a few well-placed scientists dominate. Philip Kitcher's Science in a Democratic Society (2011) reframes the problem normatively: he argues for "well-ordered science," in which research priorities would be set by a process an ideal, fully informed deliberating public would endorse. Kitcher's standard exposes a genuine deficit in big-science governance — the people who pay for and are affected by these projects have, at best, indirect and dilute influence over what gets built (see 20.08., 20.07., 31.06.* on public anthropology and science activism).

Open science and the enclosure of knowledge

Big science and open science grew up together, sometimes in tension. The Human Genome Project's Bermuda Principles (1996) committed the public consortium to daily release of sequence data into the public domain, precisely to forestall the privatisation of the human genome. The rival effort by Craig Venter's company Celera sequenced the genome faster but withheld its data, attempting to monetise the result — a contest that dramatised the choice between science as a commons and science as property. The gene-patent question reached the US Supreme Court in Association for Molecular Pathology v. Myriad Genetics (2013), which held that naturally occurring genes are not patentable, a decision with far-reaching consequences for genomic medicine (see 35.08.*).

CERN's open-access initiative SCOAP3 redirected library subscription funds to pay for open publication in particle physics, and the FAIR principles — findable, accessible, interoperable, reusable — generalised the open-data ideal across disciplines (see 33.07.* on open-source software as a parallel movement, and 33.08.01 on open science more broadly). The persistent tension is that openness is a public good requiring collective funding, while enclosure captures private returns, so the equilibrium depends on institutional choices that are never settled once and for all.

The future of big science

The frontier is now running into the saturation bound derived above. Particle physics faces the question of what lies beyond the Standard Model (see 12., 20.03.02): the proposed Future Circular Collider, a 100 km ring operating near 100 TeV, would cost on the order of tens of billions of dollars for a result that is not guaranteed. Astronomers are building bigger — the James Webb Space Telescope (2021, see 28.06.02) and the planned Artemis return to the Moon and crewed Mars missions (see 28.06.) — but each step multiplies cost and complexity.

Two emerging domains are reshaping the category. Brain science — the Human Brain Project, the US BRAIN Initiative, and the Allen Institute's connectomics work — aims at the same scale of organisation as particle physics but applied to the nervous system (see 29.02., 20.06. on consciousness). And artificial intelligence has quietly become big science: training a frontier transformer model with hundreds of billions of parameters requires compute clusters costing hundreds of millions of dollars, staffed by large teams at OpenAI, DeepMind, and a handful of others. DeepMind's AlphaFold cracked the protein-folding problem (see 17.01., 33.06.), and the governance of such systems — including their existential risks — has become a question of global cooperation comparable in stakes to nuclear non-proliferation (see 20.02.06, 50.* on AI governance).

Science in the Anthropocene

Big science is both cause and potential remedy of the planetary condition that Earth-system scientists have named the Anthropocene. The same industrial-scaled research enterprise that produced nuclear weapons and the carbon-intensive economy also produced the climate science that diagnoses the problem and the models that might guide geoengineering responses (see 27.07., 19.10.). Climate cooperation through the IPCC is big science pressed into the service of planetary governance, and it exposes the equity question at the heart of global science: who pays, who decides, and who bears the consequences (see 30.07., 31.06. on decolonising and climate-justice scholarship). The unit on contemporary science challenges (33.08.01) follows directly from here.

Connections Master

Big science connects to particle physics and quantum field theory (chapter 12) through CERN's experimental program: the W and Z bosons (Rubbia, 1983) and the Higgs boson (2012) are direct confirmations of the Standard Model, and the FCC question is a question about where the Standard Model's empirical frontier can still be pushed. It connects to general relativity (chapter 13, and 20.03.02) through LIGO: gravitational waves are a prediction of Einstein's field equations, and their detection turned GR from a theory tested only in the weak-field Solar-System regime into a theory confirmed in the strong-field, dynamical regime of merging black holes.

The astronomy and cosmology connections are dense (chapter 28, especially 28.02.04 on stellar endpoints, 28.04.04 on dark matter and dark energy, 28.06.* on space exploration, and 28.06.02 on space telescopes). The Hubble and James Webb Space Telescopes, the Square Kilometre Array, and multimessenger astronomy are all big-science enterprises whose scientific output feeds back into cosmological questions. Fusion research through ITER connects to statistical mechanics and plasma physics (chapter 11).

The molecular-biology connection runs through the Human Genome Project (33.06.01, 33.06.02). The sequencing of the human genome was the founding act of genomic medicine (35.08.*) and set the data-sharing norms that now shape biomedicine. The gene-patent controversies and the open-data movement link to media literacy and data ethics (chapter 36).

The computing connection is two-directional (chapter 33.07). Big science could not exist without the digital revolution: the World Wide Web was invented at CERN to manage collaboration, and modern experiments depend on distributed computing, petabyte-scale data management, and machine-learning analysis. Conversely, artificial intelligence has itself become a form of big science, with billion-parameter models trained on industrial-scale compute (chapters 17, 50).

The philosophy of science connection is central (chapter 20, especially 20.08). Big science forces sharp versions of every classical question in the philosophy of science: what counts as an observation (20.01, 29.03), whether instruments reveal reality or mediate it (20.08.02), how to distribute credit and trust in collaborative work, how values enter the choice of research priorities (20.08), and how democratic societies should govern expert knowledge (20.07). The sociology of organisations (30.01) and of globalisation (30.07) supply the institutional vocabulary, and the history of the 20th-century world wars and the Cold War (32.20, 32.21, 32.22) supplies the political conditions under which big science was built.

Historical and philosophical context Master

Big science is a historical category before it is a sociological one. Before 1940 the dominant model of discovery was the individual researcher or the small school: Rutherford's Cavendish Laboratory discovered the nucleus with apparatus that fit on a table; the Watson-Crick-Franklin work on the double helix involved a handful of people in two Cambridge laboratories (see 33.05., 33.06.). The war changed the scale of organisation overnight. The Manhattan Project, the MIT Radiation Laboratory, and the code-breaking work at Bletchley Park showed that scientific results of decisive military importance could be produced by industrial-scale teams working under government direction. When the war ended, neither the funding relationships nor the organisational habits disappeared; they were redirected — through the NSF, the Atomic Energy Commission, NASA, and the defence agencies — into peacetime science.

Alvin Weinberg's 1961 essay named what everyone could already see. His concern was not that big science produced bad results but that it produced bad choices. A billion-dollar accelerator commits a generation of physicists and a large fraction of a field's budget to one instrument and one scientific question, and the researchers who built it have every incentive to secure its successor. Weinberg's three criteria — scientific, technological, and social merit — were an attempt to make these choices deliberate rather than merely cumulative. Derek de Solla Price's Little Science, Big Science (1963) added the quantitative frame: science had been growing exponentially for three centuries, and the growth was approaching the point where it would consume an implausible fraction of social resources. The saturation bound proved above is the formal kernel of Price's intuition.

The moral economy of big science

The Cold War origins of big science carry a moral weight the official histories often soften. Operation Paperclip brought Wernher von Braun and other German rocket engineers to the United States after they had designed the V-2 missiles that rained on London (see 32.20.). The Saturn V that carried Armstrong to the Moon was a direct descendant of that lineage. The same decade that produced Apollo also produced enough nuclear warheads to destroy civilisation several times over, and the Doomsday Clock of the Bulletin of the Atomic Scientists has tracked the proximity ever since. The historian's task is not to deny the engineering achievement of Apollo or the intellectual achievement of the Standard Model, but to hold both alongside the violence that paid for and organised them. Decolonial and public-anthropology scholarship (31.06., 30.07.*) has made the further point that the benefits of big science have flowed disproportionately to the wealthy nations that host its facilities, while the risks — nuclear proliferation, climate disruption — fall most heavily on those who had the least say in building them.

Trading zones and the question of observation

Galison's Image and Logic (1997) is the most philosophically ambitious study of what big science does to the concept of a scientific fact. Pre-war experimental physics had two cultures: the "image" tradition, which argued from pictures of tracks in cloud chambers and bubble chambers, and the "logic" tradition, which argued from statistical counts in electronic counters. Big science forced the two together. A modern collider detector is an engineered compromise between imaging and counting, built and operated by thousands of people who do not share a single theoretical language. The "discovery" of the Higgs boson was not an individual seeing a track; it was a collective judgement that a statistical excess, reconstructed from petabytes of data by a chain of algorithms each of which had been validated by a different subgroup, matched the predicted signal to a confidence of five standard deviations.

This forces a rethinking of the philosophy of observation (20.01, 20.08.02, 29.03.*). If a phenomenon is only ever accessible through a layered instrument that no single person fully understands, then trust in the result is trust in an institution, not trust in one's own eyes. Scientific realism has to be defended not by pointing to unmediated contact with nature but by arguing that the whole coordinated apparatus — the detector, the calibration, the analysis, the peer review — is collectively reliable. Structural realists argue that what big science preserves through theory change is the mathematical structure of the relations between observables, even when the underlying ontology is revised. Critics reply that this concedes too much: if all we ever hold onto is structure, then talk of "discovering" particles or waves is a rhetorical compression of a pattern-matching procedure rather than a metaphysical discovery. The debate is live and is sharpened, not settled, by the existence of big science.

Globalisation, governance, and the democratic deficit

The internationalisation of big science is both a moral achievement and a governance puzzle. CERN, ITER, the ISS, the IPCC, and the SKA demonstrate that states can cooperate on multibillion-dollar civilian projects across decades and across political fault lines — including, in the case of the ISS, cooperation between powers that point nuclear weapons at each other. Yet the decision-making in these bodies is intergovernmental and technocratic: priorities are set by funding agencies and scientific committees, with at best indirect input from the publics who pay. Kitcher's "well-ordered science" and Jasanoff's "technologies of humility" (see 33.08.01) are philosophical responses to this deficit; they ask how the priorities of big science could be made answerable to the people they affect.

The question is sharpening rather than easing. The cost of frontier facilities continues to rise, AI has become a big science of its own with governance stakes comparable to nuclear technology, and the Anthropocene forces big science into the role of planetary diagnostician and possible remedy. Whether the institutional forms invented for post-war reconstruction — the national lab, the intergovernmental treaty organisation, the open-data commons — can carry the additional load of governing AI, fusion, geoengineering, and climate adaptation is one of the open institutional questions of the century. The history of big science so far suggests cautious optimism: the same capacity for large-scale coordination that built the bomb also built CERN, sequenced the genome, and detected gravitational waves. The lesson is not that big science is good or bad, but that its direction is a political and moral choice — one that has so far been made mostly by experts, and that an older, more ambitious conception of democracy would place in more hands.

Bibliography Master

Primary sources

[1] Weinberg, A. M. "Impact of Large-Scale Science on Science." Science 134 (1961): 161–164.

[2] Bush, V. Science: The Endless Frontier. Washington, DC: US Government Printing Office, 1945.

[3] Price, D. J. de S. Little Science, Big Science. New York: Columbia University Press, 1963.

[4] International Human Genome Sequencing Consortium. "Initial Sequencing and Analysis of the Human Genome." Nature 409 (2001): 860–921.

[5] Abbott, B. P. et al. (LIGO Scientific Collaboration and Virgo Collaboration). "Observation of Gravitational Waves from a Binary Black Hole Merger." Physical Review Letters 116 (2016): 061102.

[6] ATLAS Collaboration. "Observation of a New Particle in the Search for the Standard Model Higgs Boson with the ATLAS Detector at the LHC." Physics Letters B 716 (2012): 1–29.

Secondary works

[7] Galison, P. Image and Logic: A Material Culture of Microphysics. Chicago: University of Chicago Press, 1997.

[8] Bowler, P. J. and Morus, I. R. Making Modern Science: A Survey of the History of Science. 2nd ed. Chicago: University of Chicago Press, 2005.

[9] Collins, H. Gravity's Shadow: The Search for Gravitational Waves. Chicago: University of Chicago Press, 2004.

[10] Hoddeson, L., Kolb, L., and Westfall, C. Fermilab: Physics, the Frontier, and Megascience. Chicago: University of Chicago Press, 2008.

[11] Krige, J. American Hegemony and the Postwar Reconstruction of Science in Europe. Cambridge, MA: MIT Press, 2006.

[12] Greenberg, D. S. The Politics of Pure Science. 2nd ed. Chicago: University of Chicago Press, 1999.

[13] Kitcher, P. Science in a Democratic Society. Amherst, NY: Prometheus Books, 2011.

[14] Dickson, P. Sputnik: The Shock of the Century. New York: Walker & Company, 2001.

[15] Mirowski, P. and Sent, E.-M., eds. Science Bought and Sold: Essays in the Economics of Science. Chicago: University of Chicago Press, 2002.