Key Takeaways
Treat geometry as the art of choosing which differences matter.
Geometry reaches beyond shapes. The book’s central claim is that geometry is a way to organize information: decide what counts as a point, how distance works, and which differences can be ignored. That makes geometry useful far beyond diagrams, from tracking infections to comparing political maps and describing language. A mosquito’s wandering, for example, can help explain why a poll of a few hundred people often reflects a much larger population, and why hidden bias can still distort it.
Abstraction is a choice. Geometry does not simply reveal a ready-made structure. It builds a useful representation by setting aside features that do not matter to the question. The same discipline lets a person reason about a triangle, a stock price, or a legislature without pretending those things are literally alike.
What's striking is that geometry appears here as a practice of model design, not merely a branch of mathematics. Choosing a distance or a definition resembles choosing a measurement in science: the choice determines which patterns become visible. This is powerful because it makes hidden assumptions inspectable. It also creates a responsibility. A model can clarify a problem while quietly excluding important features, such as who is missing from a survey or whose interests a political map privileges. The value lies not in abstraction alone, but in making its choices explicit and revisable.
Use proof to test your reasoning, not to decorate obvious claims.
Proof trains intellectual honesty. Abraham Lincoln studied Euclid while rebuilding his legal and political career. He said geometry taught him what it meant to demonstrate a claim, and his habit of tracing an argument’s boundaries helped make his public reasoning unusually clear. He even spent two days trying to square a circle, a construction later proved impossible, then accepted that he had failed.
A proof should change confidence. A proof is not a ceremonial list of steps attached to something everyone already believes. The Pythagorean theorem becomes meaningful when a diagram shows how four copies of a right triangle can be rearranged to reveal the equality of areas. Good teaching lets learners feel the shift from plausible intuition to justified conclusion. Formal reasoning can expose hidden assumptions, but it works only within the axioms it begins with.
The strongest case for proof is not that every practical decision can be reduced to deduction. It is that proof makes the link between assumptions and conclusions visible, so another person can inspect it. That resembles reproducibility in experimental science: a result gains credibility when its path can be checked. Yet proof cannot certify that the starting assumptions fit the world. A flawless argument built on a bad premise remains misleading. Lincoln’s lesson is therefore best read as a discipline of accountability, not a promise that logic alone settles moral or political questions.
Choose what counts as the same before measuring anything.
Sameness depends on the question. In geometry, two objects can count as the same if one can be transformed into the other using moves that preserve the feature under study. Rigid motions preserve lengths and areas; resizing preserves angles but not area. Topology, which studies properties unchanged by continuous bending or stretching, ignores size and exact shape. A straw can be shortened and flattened into a ring without changing its topological character, while a pair of pants has a different arrangement of openings.
Symmetry reveals structure. An isosceles triangle can be understood as a figure that remains unchanged when flipped. This makes its equal angles feel less like an isolated theorem and more like a consequence of symmetry. In physics, symmetries also connect to conserved quantities, such as energy or momentum. The right definition of sameness depends on what the observer needs to preserve.
This is a compact version of a deep scientific habit: define equivalence relative to a purpose. Chemists group substances by composition, while clinicians may group them by symptoms or treatment response. Neither grouping is universally correct; each answers a different question. The danger is forgetting that the chosen equivalence relation is a choice. A classification that is useful for one task can erase distinctions vital to another. The book’s geometric perspective helps prevent that mistake by asking, before comparing objects, which transformations are allowed and which properties must remain fixed.
Model correlated errors before trusting a clean margin of error.
Randomness does not guarantee balance. Ronald Ross studied a mosquito making repeated random moves. After many steps, its typical distance from home grows roughly with the square root of the number of steps, not in proportion to the total. The same pattern helps explain sampling: in a hypothetical poll of 200 evenly divided Wisconsin voters, a 120-to-80 result would be unusual if respondents were chosen without bias.
But errors can travel together. If some voters are more likely to answer the phone, the sample may lean in one direction. Polling errors across states can also share a common cause, making a sweep of outcomes more plausible than calculations assuming independence suggest. This mattered in 2016: forecasts that treated swing-state errors as mostly independent gave Donald Trump much lower odds than models allowing for shared polling bias.
The distinction between random error and systematic bias is central to sound measurement. Increasing a sample can shrink random fluctuation, but it does not repair a skewed sampling process. Correlation poses a second trap: several measurements may look like independent confirmations while reflecting the same underlying error. This is why uncertainty estimates need assumptions about how data were collected and how errors relate. The argument is especially useful outside polling, including medical studies and business dashboards, where a precise-looking margin can create confidence that the design does not deserve.
Map every move, then accept when perfect search is impossible.
Games have a tree structure. A game tree records every possible position and move, branching from the start to the final outcomes. In Nim, players remove stones from piles, and a small tree can be labeled backward to show which positions are winning or losing. The method works in principle for any finite game with alternating moves and no randomness.
Perfect play can still be unreachable. Checkers has an enormous number of possible positions, but computer program Chinook eventually proved that perfect players must draw. Marion Tinsley, the dominant human champion, could compete at a level no person had matched, yet the machine could analyze far more possibilities. The distinction matters: a game can have a definite mathematical answer while its full solution remains too costly to compute. And even when a game is solved, human players may value beauty, tension, and judgment more than guaranteed victory.
The game-tree model separates logical possibility from computational feasibility. That distinction appears throughout computer science: an algorithm may be correct and complete in theory, yet unusable at the scale of a real problem. Approximation is not automatically a defect; it can be the rational response to limited time and computing power. Still, a search strategy can inherit blind spots from what it prunes or treats as valuable. A machine that wins reliably has solved one objective, not every human question about why the activity matters.
Use fast tests to find answers when exhaustive proof is infeasible.
Trial and error is a real method. To generate a very large prime for encryption, one can pick a candidate and test it rather than search systematically through every number. Fermat’s Little Theorem supplies a quick check: every prime passes a certain arithmetic test, though some non-primes pass as well. A failed candidate is discarded; a successful one is a probable prime, not an absolute guarantee.
Asymmetry enables secrecy. Public-key cryptography such as RSA relies on multiplying two huge primes being easy while recovering the original factors is, with known methods, extraordinarily difficult. This one-way difficulty lets a person publish a key that others can use to encrypt messages without revealing the secret factors needed to decode them. The lesson is practical: imperfect tests and hard problems can be useful when their limits are understood.
This example challenges the idea that mathematics is valuable only when it delivers certainty. In engineering, medicine, and security, decisions often rely on tests with known error rates rather than exhaustive verification. The crucial questions are whether the test is reliable enough for its purpose and what happens when it fails. Cryptographic confidence also depends on current computational limits, not on a theorem that factoring can never become easier. Treating security as a maintained assumption, rather than a permanent fact, is the more responsible interpretation.
Constrain the search, but test whether the model generalizes.
Learning is a search problem. Gradient descent is a procedure that repeatedly adjusts a strategy in the direction that reduces its errors on examples. To recognize cats, a program receives labeled pictures and searches among possible rules for a useful classifier. Neural networks make this search manageable by limiting the kinds of strategies considered and arranging their adjustable parameters into layers.
Training success can mislead. A model can memorize its examples and still fail on unfamiliar ones, a problem called overfitting. It can also be too simple to capture real patterns, which is underfitting. Gradient descent may settle at a local optimum, a point where nearby adjustments do not improve performance even though a better solution exists elsewhere. The practical challenge is to find a model complex enough to learn, but disciplined enough to work beyond its training data.
The account connects machine learning to a broader principle of scientific inference: fitting observations is not the same as explaining or predicting new cases. Separate test data provide one check, but they cannot guarantee performance under conditions unlike those represented in training. A system can also learn social biases embedded in its examples, then reproduce them at scale. The search space and evaluation criteria therefore matter as much as computing power. The metaphor of choosing a landscape is useful because it makes clear that algorithms explore only the possibilities their designers allow.
Use models to locate turning points, not promise exact futures.
Epidemics change as they spread. In a simple early-stage model, each infected person passes a disease to an average of R0 others. Above 1, cases tend to grow; below 1, they tend to shrink. As susceptible people become fewer, or people change their behavior, transmission changes too. The SIR model tracks people who are susceptible, infected, and recovered, helping describe how an outbreak rises and falls.
Extrapolation is fragile. William Farr correctly anticipated that Britain’s 1865 rinderpest outbreak would soon decline, but his method badly mistimed how quickly it would end. During COVID-19, curves that fit early data could still fail when behavior, policy, or the shape of the decline changed. A useful model can indicate whether a system is accelerating, stabilizing, or receding without claiming to know the exact count on a future date.
Models are most useful when they clarify conditional statements: if transmission and behavior continue in a certain way, these outcomes become more likely. That framing protects against treating forecasts as fate. It also matters because public predictions can change behavior, which then changes the process being predicted. A disease model is not observing a sealed physical system; people respond to warnings and rules. The challenge is to combine mechanistic understanding with frequent recalibration, while communicating uncertainty without making uncertainty sound like ignorance.
Redraw distance around actual connections, not just physical space.
Distance is not always geographic. A map can measure travel time, family relationships, or the likelihood that two words appear in similar contexts. The 1872 horse-flu outbreak in North America spread quickly along rail lines, reaching distant coastal cities before some nearer inland towns. The railway had changed the practical distances between places. Modern disease spread follows both ordinary proximity and fast transport links.
Networks create small worlds. A network is a set of points connected by links, such as people who know each other or web pages that link together. Facebook data cited in the book found that the average shortest chain between users was fewer than five steps. Word2vec similarly places words near one another when they occur in similar contexts. Its gendered associations expose patterns in language, but they record how people write, not the intrinsic meaning of words.
Changing the distance measure can uncover the mechanism that a conventional map hides. In transport planning, for instance, travel time may matter more than mileage; in public health, a person’s contacts may matter more than their neighborhood. Network models also reveal how a few long-distance links can accelerate spread. Yet short paths are not the same as equal access or social closeness. A network may connect people in a technical sense while leaving them separated by power, trust, or resources. The metric answers one question, never all of them.
Judge district maps against neutral possibilities, not proportionality alone.
Fairness is not a seat-share formula. In Wisconsin’s 2018 elections, Democrats won statewide offices while Republicans retained a commanding majority in the state assembly. Equal population and tidy district shapes do not prevent partisan advantage, and proportional representation is not a reliable benchmark when voters are geographically concentrated.
Compare with plausible alternatives. Researchers used computers to generate thousands of legally valid Wisconsin maps, favoring ordinary criteria such as compactness and respect for local boundaries, but not partisan outcomes. They then compared each map’s election results with the enacted map. The comparison suggested that the Republican map produced an unusually durable advantage, especially in Democratic-leaning election years. This ensemble approach asks whether a map is an outlier among neutral alternatives, rather than assuming one mathematically perfect division exists.
The ensemble method reframes the problem from finding a universal fairness number to testing a counterfactual: what outcomes would comparable maps have produced without partisan intent? That is a more realistic statistical question, but it is not free of judgment. Researchers must decide which maps are legally possible, what counts as compact, and how the random sampling process explores the options. Those decisions should be public and open to challenge. The method’s strength is not that it removes values from redistricting, but that it makes one form of manipulation measurable.
Demand understanding that expands the world beyond a verified answer.
A correct answer is not always enough. Edwin Abbott’s Flatland imagines a two-dimensional being who cannot picture a higher-dimensional world until a visitor lifts him beyond his plane. Geometry can challenge the limits of what seems self-evident, rather than merely reinforce established authority. Poet Rita Dove’s poem about proving a theorem turns that discovery into a world whose walls open outward.
Proof serves understanding. The mathematician Grigori Perelman proved the Poincaré Conjecture, a major result about the shape of three-dimensional spaces. The book’s point is not simply that a theorem has been settled, but that mathematical work matters when its ideas help people see and reason more clearly. Proof is a scaffold for understanding, not a substitute for it. The ultimate aim is to make previously invisible structure available to thought.
This ending gives the book’s broad tour an epistemological center: mathematics is not only a machine for producing true statements, but a way to expand what can be understood. That distinction is important as automated systems become better at generating correct outputs without offering humanly graspable reasons. Still, explanation is not the only measure of value; some results are useful before their full interpretation is clear. The durable standard is whether a method creates dependable insight, invites scrutiny, and enlarges the questions people can ask.
Analysis
Shape is a panoramic argument about mathematical representation. Its recurring move is to start from an ordinary puzzle, such as counting a straw’s holes or predicting a poll, then reveal that the answer depends on how the problem has been geometrically organized. A shape may be a set of points, a game tree, a network, or a space of possible strategies. The book’s unity comes less from one theorem than from this repeated act of choosing what counts as distance, sameness, connection, or evidence.
That method produces a valuable account of mathematics as both disciplined and creative. Formal proof checks the path from assumptions to conclusions, while intuition helps identify which path is worth following. Computation extends human reach, but the checkers tree, neural-network landscape, and redistricting ensemble all show that practical results depend on what can be searched and what is omitted. Mathematical certainty is therefore local: a proof may be absolute inside a chosen system, while the system’s relevance remains an empirical or ethical question.
The book is especially persuasive when it carries abstraction into public life. Polling, epidemics, language models, and electoral maps become legible through shared mathematical structures, but their differences remain consequential. A random walk can illuminate sampling without making voters into mosquitoes; an ensemble can detect a suspicious map without defining democracy for everyone. This is a subtle strength: analogy transfers tools, not identity.
The main limitation is also the book’s ambition. A large number of historical episodes and technical ideas arrive in rapid succession, and the memorable metaphor can sometimes make a difficult field seem more settled than it is. The author often signals uncertainty, especially around forecasting and gerrymandering, but readers should preserve that caution. The book is best understood as an invitation to mathematical judgment: identify the model, inspect its assumptions, and ask what the chosen geometry makes visible or leaves out.
Review Summary
Shape explores geometry's relevance in everyday life, from pandemics to politics. While praised for its engaging writing and accessibility, some readers found it meandering and overly detailed in parts. The gerrymandering chapter received particular acclaim. Opinions varied on the book's structure and focus, with some finding it enlightening and others struggling to follow. Overall, readers appreciated Ellenberg's humor and ability to connect mathematical concepts to real-world issues, though the level of difficulty may not suit all audiences.
People Also Read
Glossary
Topology
Studies shape through deformationA branch of geometry concerned with properties that remain unchanged when a shape is continuously bent or stretched without cutting or gluing. Topology ignores precise lengths and angles, focusing instead on features such as connected pieces and holes. The book uses a straw and a pair of pants to show how counting holes depends on a careful mathematical definition.
Symmetry
A permitted shape transformationIn the book’s geometric usage, a symmetry is a transformation allowed when deciding whether two objects count as the same. Rigid motions preserve lengths and areas; other choices, such as stretching or bending, preserve different properties. The symmetries selected determine what distinctions matter, what quantities remain invariant, and what kind of geometry is being used.
Markov chain
A process with state-based stepsA sequence of changing states in which the probability of the next state depends on the current state, rather than on the entire history. A mosquito that moves between two bogs according to fixed probabilities is a simple example. Markov chains can model text, disease spread, games, website links, and other processes that unfold step by step.
Law of Long Walks
Long-run state frequenciesThe book’s name for the tendency of a random walk on a suitable finite network to settle into stable long-run proportions. A walker may move unpredictably at each step, yet spend a predictable share of time at each location over a long period. Google’s PageRank uses this idea to rank web pages by how often a link-following walker would visit them.
Gradient descent
Iterative error reductionAn optimization method that repeatedly adjusts a strategy in the direction that reduces a measured error. In machine learning, a model is assessed on examples, its parameters are nudged to improve performance, and the process is repeated. The method can be computationally effective, but may settle at a local optimum or produce a strategy that fails on unfamiliar cases.
Overfitting
Memorizing examples too closelyA failure of a model that matches its training examples so precisely that it performs poorly on new cases. A cat classifier that remembers the exact pictures it was shown, but cannot recognize a new cat or a slightly altered image, is overfitting. The opposite problem is underfitting, when a model is too simple to capture meaningful patterns.
SIR model
Tracks infection compartmentsAn epidemic model that divides a population into susceptible, infected, and recovered groups. Rules describe how people move between those groups over time, allowing researchers to explore how an outbreak rises, peaks, and declines. The model is useful for understanding dynamics and thresholds, but its projections depend on assumptions about transmission, immunity, and behavior.
Eigenvalue
A system’s inherent growth factorA number that describes how a particular transformation scales a corresponding special vector or state. In interacting systems, such as populations divided by age or location, eigenvalues can reveal the rates that govern long-run growth or decline. The book connects them to geometric progression, Fibonacci-like patterns, epidemic spread, and the ranking of web pages.
ReCom geometry
A random walk among district mapsA way to explore possible electoral maps through repeated recombination. The method selects two neighboring districts, joins them, then splits the combined region into two valid districts, producing a new map. Repeated moves generate an ensemble of alternatives that can be compared with an enacted map to assess whether its partisan results are unusual.
efficiency gap
Difference in wasted votesA measure of partisan advantage based on the difference between the parties’ wasted votes, expressed as a share of total votes. Votes are counted as wasted when cast for a losing candidate or when they exceed the number needed to win. The measure is easy to calculate, but can change sharply with small shifts in election results and does not by itself settle questions of fairness.
ensemble
A set of simulated alternativesA collection of maps or models generated under specified constraints, used as a comparison group. In redistricting, an ensemble of maps can represent outcomes produced without an explicit partisan goal while respecting legal and traditional criteria. The comparison can reveal whether an enacted map is an outlier, though results depend on how alternatives are generated and sampled.
FAQ
What's Shape: The Hidden Geometry of Information, Biology, Strategy, Democracy, and Everything Else about?
- Exploration of Geometry's Role: The book delves into how geometry influences various fields, including information theory, biology, and democracy. It emphasizes that geometry is not just about shapes but also about the relationships and structures underlying complex systems.
- Interdisciplinary Connections: Author Jordan Ellenberg connects mathematical concepts to real-world applications, illustrating how understanding geometry can provide insights into everyday life and decision-making.
- Mathematics as a Tool: Ellenberg argues that mathematics, particularly geometry, serves as a powerful tool for understanding and navigating the complexities of the world around us.
Why should I read Shape by Jordan Ellenberg?
- Engaging and Accessible: Ellenberg presents complex mathematical ideas in an engaging and accessible manner, making them relatable to a broad audience. His writing style combines humor with deep insights, making the book enjoyable to read.
- Real-World Applications: The book provides numerous examples of how mathematical concepts apply to real-world situations, from politics to personal decision-making, demonstrating the relevance of math in everyday life.
- Broadening Perspectives: Reading Shape can help readers develop a new perspective on how they view information and relationships in various contexts, enhancing critical thinking skills.
What are the key takeaways of Shape by Jordan Ellenberg?
- Geometry's Ubiquity: One of the main takeaways is that geometry is present in many aspects of life, influencing everything from social networks to biological systems. Understanding these geometric principles can lead to better decision-making.
- Importance of Models: The book emphasizes the significance of mathematical models in predicting outcomes and understanding complex systems. Ellenberg illustrates how models can help us navigate uncertainty and make informed choices.
- Interplay of Math and Reality: Ellenberg highlights the interplay between mathematical theory and real-world phenomena, showing that math is not just abstract but deeply connected to the world we live in.
What are the best quotes from Shape by Jordan Ellenberg and what do they mean?
- “Geometry is the art of reasoning well from badly drawn figures.”: This quote by Henri Poincaré encapsulates the essence of geometry as a tool for logical reasoning, even when faced with imperfect representations. It suggests that the process of reasoning is more important than the precision of the figures.
- “A proof is crystallized thought.”: This quote emphasizes the significance of proofs in mathematics, as they encapsulate insights and make them shareable. It highlights the value of structured reasoning in understanding complex ideas.
- “All models are wrong, but some are useful.”: This quote encapsulates the essence of modeling in mathematics, acknowledging that while no model can perfectly capture reality, they can still provide valuable insights and guide decision-making.
How does Jordan Ellenberg connect geometry to democracy in Shape?
- Gerrymandering Analysis: Ellenberg uses geometry to explain the mathematical underpinnings of gerrymandering, illustrating how shapes of electoral districts can influence political outcomes. He shows that understanding geometric principles can help identify and combat unfair practices in democracy.
- Voting Systems: The book discusses various voting systems and their geometric implications, emphasizing how different methods can lead to different representations of public opinion. Ellenberg argues that geometry can provide insights into creating fairer electoral systems.
- Mathematical Integrity: Ellenberg advocates for the importance of mathematical integrity in democratic processes, suggesting that clear and fair geometric reasoning can enhance the legitimacy of electoral outcomes. This connection highlights the role of mathematics in promoting social justice.
What is the significance of the "random walk" concept in Shape by Jordan Ellenberg?
- Understanding Movement: The random walk concept is used to describe how entities, like mosquitoes or stock prices, move in unpredictable ways. Ellenberg illustrates that this randomness can be modeled mathematically, providing insights into various phenomena.
- Applications in Various Fields: Ellenberg connects the random walk to fields such as finance and biology, showing its relevance in understanding complex systems. This demonstrates how mathematical concepts can transcend disciplinary boundaries.
- Predictive Power: The random walk model helps in making predictions about future states based on past behavior, emphasizing the importance of probability in understanding randomness. This predictive aspect is crucial in fields like economics and ecology.
How does Shape by Jordan Ellenberg address the relationship between geometry and artificial intelligence?
- AI Algorithms: Ellenberg discusses how geometric principles are foundational to many algorithms used in artificial intelligence, particularly in machine learning. He explains that understanding geometry can enhance the development of more effective AI systems.
- Gradient Descent: The book introduces the concept of gradient descent, a method used in machine learning to optimize algorithms. Ellenberg explains how this mathematical technique is analogous to navigating a landscape to find the highest peak.
- Real-World Applications: Ellenberg illustrates the practical applications of AI in various fields, emphasizing that geometry plays a critical role in shaping these technologies. This connection highlights the interdisciplinary nature of modern mathematics.
What is the "Survivor Theorem" mentioned in Shape by Jordan Ellenberg?
- Game Theory Application: The Survivor Theorem is a mathematical principle derived from analyzing the game of Nim, which Ellenberg uses to illustrate strategic thinking. It states that if the number of flags is a multiple of four, the first player loses; otherwise, they can win.
- Proof Structure: Ellenberg demonstrates how to prove the theorem using logical reasoning and tree diagrams, showcasing the importance of structured thought in mathematics. This proof serves as an example of how game theory can inform decision-making.
- Broader Implications: The theorem has implications beyond the game itself, as it encourages readers to think critically about strategy and outcomes in competitive situations. This principle can be applied to various real-life scenarios.
How does Jordan Ellenberg use personal anecdotes in Shape?
- Engaging Narrative: Ellenberg incorporates personal stories to make complex mathematical concepts more relatable and engaging for readers. His anecdotes provide context and illustrate the human side of mathematics.
- Reflecting on Experiences: The author shares his own struggles and triumphs with geometry, allowing readers to connect with him on a personal level. This approach demystifies mathematics and encourages a more approachable view of the subject.
- Encouraging Curiosity: By sharing his journey, Ellenberg inspires readers to explore mathematics themselves, fostering a sense of curiosity and wonder about the subject. His narrative style makes the book accessible to a wide audience.
How does Shape by Jordan Ellenberg redefine the perception of geometry?
- Living Subject: Ellenberg presents geometry as a dynamic and evolving field, rather than a static set of rules. This redefinition encourages readers to see geometry as relevant to contemporary issues and innovations.
- Interdisciplinary Connections: The book highlights the connections between geometry and various disciplines, demonstrating its applicability in real-world scenarios. This approach broadens the understanding of geometry's role in society.
- Emphasizing Beauty: Ellenberg argues that geometry is not just about calculations but also about beauty and creativity. This perspective invites readers to appreciate the aesthetic aspects of mathematical thought.
What is gerrymandering, as explained in Shape by Jordan Ellenberg?
- Manipulation of District Boundaries: Gerrymandering refers to the practice of drawing electoral district boundaries to favor one political party over another. Ellenberg illustrates how this manipulation can distort representation and undermine democracy.
- Mathematical Analysis: The book uses mathematical principles to analyze gerrymandering, showing how geometry can reveal the unfairness of certain districting practices. Ellenberg provides examples to demonstrate the impact of gerrymandering on electoral outcomes.
- Call for Reform: Ellenberg advocates for reforms to combat gerrymandering, emphasizing the importance of fair representation. He argues that understanding the geometric implications of districting can lead to more equitable electoral systems.
How does Shape by Jordan Ellenberg encourage critical thinking?
- Challenging Assumptions: Ellenberg encourages readers to challenge their assumptions about mathematics and its applications. He demonstrates that math is not just a set of rules but a tool for understanding and engaging with the world.
- Interdisciplinary Connections: By connecting math to various fields, the book fosters interdisciplinary thinking. Ellenberg shows that complex problems often require insights from multiple disciplines to find effective solutions.
- Empowering Readers: The book empowers readers to think critically about societal issues, particularly in the context of democracy and representation. Ellenberg provides the tools to analyze and question the systems that govern our lives.
Download PDF
Download EPUB
.epub digital book format is ideal for reading ebooks on phones, tablets, and e-readers.