From the question to a thesis, from a thesis to a topic

Do not ask "are humans good or bad", there is no answer to it. Ask "is moral intuition innate": both sides have an empirical claim and a counterexample. The same infant and primate studies get read differently by the two sides. Wherever the same evidence has two readings, there is an arguable question.

Instead of defending empathy, tie it to a measurable claim: does empathy mislead us systematically when we decide? Focus on a single victim, blindness to the group. Bloom defends a thesis that runs against intuition, de Waal shows the evolutionary basis of empathy. The two books answer each other directly.

If you are going to defend the situationist account, do not ignore that the Stanford prison experiment has taken heavy methodological criticism; your evidence gets knocked out right there. Browning's data on a police battalion is firmer ground. A historical case, a lab experiment and a philosophical frame all test each other here. This is the topic where practicing the comparison of evidence types pays off most.

Question the concept itself: is calling something "evil" an explanation, or a label that ends the explanation? A thesis can be built on whether the concept does any work. Neiman puts the concept at the center of modern thought, Midgley explains it as an accumulation of ordinary failings, Nietzsche historicizes the pair of concepts itself. Three different jobs loaded onto the same word; a definitional argument is rarely this concrete.

Argue about the interpretation, not the result: does the participation rate show the power of obedience, or the persuasive power of the experimental design? Blass writes the story of the experiment and the experimenter, Zimbardo carries the same frame into another setup, and Cialdini ties the mechanism to everyday persuasion techniques. A case where the evidence itself is in dispute, and one still taught in textbooks as a certainty.

Turn the question "do prisons work" into "what are they trying to achieve". Reform, deterrence and revenge are different criteria; you cannot evaluate without saying which one it fails on. Foucault redefines the purpose of the institution, Garland works with measurable policy data. Practice at using a philosophical frame and empirical data in the same essay.

The thesis that the perpetrator is not a "monster" is the easy part; the hard part is separating how much of it is the situation and how much is ideology. Browning and Goldhagen are diametrically opposed on that split. Two historians use the same archival material and reach opposite conclusions. Showing two sources drawing different conclusions from the same evidence is the strongest move there is in an essay.

Treat honesty not as character but as behavior sensitive to conditions: under which condition do how many people cheat? Ariely shows experimentally that cheating is condition sensitive, Bazerman frames the same behavior as violations committed without noticing, and Kahneman gives you the cognitive basis of the mechanism. A two layered argument earns a higher grade.

Instead of saying "psychopaths are evil", ask where the threshold of responsibility is: if one personality structure reduces moral responsibility, should the same logic apply to other personality differences? Oakley builds a biological account, Wilson looks for crime at the intersection of personality and environment, Fromm explains destructiveness through social conditions rather than character. Three different levels of explanation aimed at the same behavior.

Pinker's thesis rests on a population adjusted measure, and the objection lands exactly there. Build the argument not on "did it decline" but on "which measure answers which question". Showing how a statistical claim is constructed through the choice of measure lands right in the middle of the evidence analysis item in a course essay.

The sentence "algorithms are biased" is not a thesis. Ask: who picks the criterion that counts a system's output as 'fair', and is that pick a technical decision or a political one? O'Neil and Benjamin give you concrete cases; Russell and Bostrom frame the same problem as one of goal specification. For an engineering student, a debate that is both familiar and still open.

Do not retell the trolley problem. Ask: does translating a moral principle into a specification change it? Is a principle that has become codable still the same principle? Wallach and Maurer discuss the translation problem directly; Bostrom and Russell carry the same problem up to the level of goal specification. The tension between abstract ethics and engineering constraints is exactly the kind of arguable question ENG 101 wants.

Drop the question "will robots take our jobs" and move to "who collects the gains while technology raises productivity". That is a question with evidence behind it. Acemoglu and Brynjolfsson draw different policy conclusions from the same data. Two economists framing the same fact differently is a ready made disagreement for an essay.

Pick a frame that defines privacy not as 'secrecy' but as who uses the data and for what purpose; the "I have nothing to hide" objection collapses inside that frame. Zuboff gives you a property frame, Lyon the institutional history of surveillance, and Foucault the claim that being seen is itself a means of control. Comparing three different levels makes a definitional argument concrete.

Accept that the same platform feeds both organizing and polarization, then ask about the mechanism: does the difference come from the platform, from how it is used, or from the existing media order? Tufekci shows the same tools both building movements and making them fragile; Pariser and Vaidhyanathan give a design based account; Lippmann set up the same problem for newspapers a century ago. A mechanism thought to be new turning out to be an old version of itself is good thesis material.

The "tragedy of the commons" is not inevitable; Ostrom's field data shows cases where it was avoided. Build the thesis on the question "when does it collapse". Ostrom's empirical finding refutes the classical theory directly, and that is what the Nobel was for. Open source software and knowledge commons can be argued with the same frame.

Blaming one faulty part is easy. Ask: how did the organization normalize the risk? Vaughan's concept of the "normalization of deviance" makes that measurable. Perrow treats the accident as the inevitable result of system complexity, Petroski gives the history of design decisions, and Norman puts the error on the interface rather than the user. For an engineering student, a frame that stays useful after the assignment.

Do not squeeze the argument into the "security or privacy" dilemma; ask the technical question: can a backdoor be kept open only for the authorized party? Schneier writes both the technical and the policy side, and Singh gives the century long history of the same conflict. Where a claim of technical impossibility collides with a policy demand, building a thesis is easy.

Do not ask "is the smart city good or bad". Ask: which problem gets picked because it is measurable, and which one stays invisible because it is not? Scott's concept of legibility shows what making something measurable renders invisible; Jacobs and Lynch ask the same question from the point of view of the city's user. Directly applicable for an urban design or engineering student.

Argue about the use of models, not the models: does a model that turns uncertainty into a number reduce the uncertainty or hide it? Taleb and Tetlock propose two different solutions to the same problem: reducing fragility, and calibrating forecasts. Comparing the two prescriptions carries a thesis on its own.

Instead of racing the geography account against the institutional one, ask: if institutions are a consequence of geography, does the explanation still stay institutional? Acemoglu-Robinson and Diamond answer each other openly. For an economics or international relations student, a debate that is both familiar and still open.

Do not try to prove that inequality rose, argue about the measure: within a country or globally, income or wealth, before tax or after? The answer changes with the measure. At a global scale Milanovic shows a different trend from Piketty's national finding. The same fact looking inverted at two scales is strong thesis material.

Turn the question "does aid work" into a question about scale: does an intervention proven in a randomized experiment also work at national scale? Banerjee and Duflo produce micro evidence from field experiments; Moyo and Sachs answer the same question at macro scale in diametrically opposed ways. The external validity argument stays valid for an economics student after the course too.

Instead of questioning whether merit can be measured, ask what it does: is it used to justify outcomes, or to correct the distribution? Young invented the concept as satire; Rawls, Nozick and Walzer answer in three different ways the question of which criterion makes a distribution just; Bourdieu asks where the measured 'talent' came from. A rare case where the history of a concept turns into the argument.

Not "is the nudge good": when you decide on someone's behalf, whose preference is it? The paternalism objection is not technical here, it is a matter of principle. Thaler and Sunstein defend the intervention, Kahneman and Ariely give the experimental basis of the mechanism, and Mill's classic liberty argument builds the objection itself. A three position debate makes it easier to balance the thesis.

Do not race free trade against protectionism ideologically; ask about the historical order: how did the countries that defend free trade today industrialize? Chang's historical claim shows a direct contradiction: the countries that defend free trade today industrialized behind protection. Wade tests the same claim against East Asian data. Being transferable to the Turkish context is an extra advantage.

The question "is migration good or bad" cannot be answered. Ask: for which group and over which time scale do the costs and gains appear? Betts treats the current refugee regime itself as broken, Castles shows why reducing migration to a single cause does not work, and Gellner shows that the question 'who counts as a citizen' comes out of the construction of the nation. The topic closest to being arguable with Turkey's own data.

Compare the supply constraint account with the distribution account: are housing prices rising because of zoning constraints, or because housing became an investment vehicle? Glaeser defends the supply side, Harvey and Desmond the distribution side. Two causal accounts of the same fact, a ready skeleton for an essay.

Treat the gecekondu not as a problem but as a solution: whose problem did it solve, and whose shoulders did the cost land on? Karpat's field work runs against the official discourse of the period. Testing a phenomenon you can observe in your own city against the academic literature is an advantage on the evidence item.

The thesis that the nation is "constructed" is no longer in dispute; what is in dispute is how much of it is construction. Anderson and Smith part company on that point. Hobsbawm's concept of the invented tradition can be tested against concrete examples. Turkish examples are plentiful in the literature, so evidence is easy to find.

Bring the rupture and continuity dilemma down to a single institution: law, education or the bureaucracy. The general "revolution or evolution" question does not fit in one essay. Zürcher stresses continuity, Bernard Lewis rupture, and the two are describing the same events. Two different framings of the same event is exactly the material you want.

Instead of judging the language reform on a success and failure axis, ask about the criterion: was the goal intelligibility, a break, or a new identity? Geoffrey Lewis calls the reform a "catastrophic success"; that phrase is itself a thesis argument. It can be combined with Anderson's frame.

Telling the story of one refuted finding is not enough. Ask: if the system produces the wrong finding, is the error individual or is it the incentive structure? Popper makes falsifiability the criterion, Kuhn argues that science does not actually work that way, and Kahneman admits what working with small samples produced in his own field. The gap between the criterion and the practice is a ready made tension for a thesis.

Do not try to prove the scientific consensus, that argument is over. Ask: with a consensus in place, why is policy not moving, is it uncertainty, interest, or framing? Oreskes shows with archival documents that the uncertainty was manufactured deliberately, Klein ties the problem to the economic order itself, Wallace-Wells gives you the consequence, and Beck explains how risk gets distributed. Four different levels of explanation of the same fact, ideal for comparison.

Comparing risk by total death counts is a weak argument on its own. Ask: can the type of nuclear risk (low probability, high consequence) be compared on the same scale as other risks? Perrow argues that in complex systems the accident is structural, Beck sets out how risk is distributed in modern society, and Rhodes gives the history of the technology itself. The criterion problem in risk comparison is productive for an engineering student.

Aiming at extending life and aiming at improving it are different criteria. Build the thesis on the choice of criterion. Gawande's clinical observations conflict with institutional incentives. For a medicine or psychology student, a frame that is both close to home and not sentimental.

Finding an image "beautiful" or "powerful" is not analysis. Ask: which way of looking does the image make natural, and what does it leave outside the frame? Berger and Sontag hold the same phenomenon from different places: one property and the gaze, the other witnessing and desensitization. Core texts for a fine arts or communication student.

Argue directly with the documentary's claim that it "shows the real": do the choices of editing, sound and sequence weaken that claim to reality, or construct it? Nichols divides the genre into modes, Bazin defends an ontological claim. For a communication or film student, a frame that stays in use after the assignment.

Instead of labeling sampling as theft or as art, ask: does the definition of originality change along with technology, or is it the law that lags behind? Demers collects the legal cases, Lessig treats the same practice as the norm of cultural production, and Frith questions the criterion of 'authentic performance' itself. For a music student, a directly professional question.

Question the criterion of fidelity: does judging a film by its faithfulness to the book block you from judging the film on its own terms? Hutcheon rejects the fidelity criterion openly, Leitch criticizes the debate itself. A ready made disagreement for comparative literature and film.

Do not assume a relation between price and aesthetic value; ask about the institutional mechanism: who sets the price, and with what information? Bourdieu shows how value is produced institutionally, Berger the relation of the gaze to property, and Gombrich how the canon was formed. The three together are enough to test the assumption that 'price measures aesthetic value'.

The sentence "something is lost in translation" is not a thesis. Ask: is there a criterion that measures the loss, or should we speak of transformation instead of loss? Venuti frames the translator's invisibility as a political matter, Steiner treats the loss as inevitable, and Baker and Munday break the decision down at a technical level. The core debate for translation and interpreting.

Try measuring publicness as use rather than as a design feature: is a well designed square that nobody uses a public space? Whyte collects direct observation data, Sennett builds a conceptual frame, Gehl proposes a measurement method. Testable in the field for architecture and urban design.

Tie the question "how should I study" to evidence: why do the methods that look effective intuitively (rereading, highlighting) come out badly when measured? Make It Stick sets up exactly this contradiction: the method that feels easy teaches least. Carey collects the same finding at a popular level, Hattie compares the effect of hundreds of studies, and Dewey shows the limit of the assumption that 'experience teaches'.

Treating distraction as a personal weakness is a weak thesis. Ask: if attention is a product being sold, where does responsibility sit? Wu gives the business model, Carr the cognitive effect, Gazzaley and Hari the measured outcome; Postman set up the same mechanism for television. Explanations at four different levels, ideal for comparison.

Instead of debating the purpose of the university in general, pick one tension: research or teaching, or public good or employment? Newman treats the institution as an end in itself, Nussbaum objects to the market frame, and Freire asks whose interest education serves. Writing about your own institution makes evidence easier to find.

The cognitive advantage claim has been seriously questioned over the last decade. Build the thesis not on the finding itself but on which type of evidence the claim rests on. Grosjean refuses to see bilingualism as deficient monolingualism, Pinker explains language ability at the level of the species, and Ellis and Wardhaugh give you the learning and the social context side. The same phenomenon measured by three separate disciplines.

Do not assume data analysis is "more accurate". Ask: is it that the measurable things are the important part of the game, or that they are counted as important because they are measurable? Lewis tells the story of the shift, Anderson-Sally show the limits of the same method in football. For a student interested in sport, real life evidence is ready to hand.

Question the panic assumption: does a crowd dissolve the individual, or does it produce a new rule of behavior? The answer changes how you design an evacuation. Canetti reads the crowd as a form with its own logic, Zimbardo and Aronson build the same phenomenon out of experimental social psychology. Two readings of the same phenomenon, one theoretical and one experimental.

Treating car dependency as a consumer choice may be wrong. Ask: when, and by whom, was it redefined who the street belongs to? Jacobs and Gehl read the street off the behavior of its users, and Lynch measures how the city is represented in the mind. Directly observable for a student living in Ankara.

Put persuasion techniques and argument face to face: does a valid argument have to be persuasive, does a persuasive text have to be valid? Cialdini shows the mechanisms experimentally, Lakoff the framing power of language. Both tie directly into this course's own subject.

Drop the cliché that "statistics lie"; pick a specific mechanism: the base rate, a chosen axis, survivorship bias. That Huff's 1954 book still holds is a finding in itself; Wheelan and Rosling give current examples, and Silver and Kahneman show that the error is in the method, not in the reader.

Repeating the correlation and causation distinction is not enough. Ask: which extra assumption, once added, gets you a causal conclusion from observational data? Pearl offers a formal frame, Angrist a quasi-experimental method; two different answers to the same problem. Technical but arguable, so it suits this course.

Frame invisible labor not as a moral issue but as a measurement problem: how does labor that never enters the national accounts distort policy? Criado Perez shows the data gap with concrete examples, Hochschild turns emotional labor into a measurable concept. Framed as a measurement problem, the thesis becomes provable.

Argue the limit through harm rather than "offense": how is the harm of speech measured, and if it cannot be measured, on what basis is the line drawn? Waldron defends a limit, Mill gives the classic counterframe, and Matsuda's collection shows to whom and how the harm is distributed. A clean three position debate.

Take the "I have nothing to hide" objection seriously and refute it: if privacy is not a right to keep secrets, what is it, and why does that definition make the objection invalid? Zuboff builds the issue through property, Schneier through infrastructure, and Foucault through being seen changing behavior; Orwell gives the literary extreme. Ideal for an essay structure that starts with the counterargument.

Counting animal rights as an extension of human rights is an assumption. Ask: if the criterion is the capacity to suffer, how far do the consequences have to be carried? Singer builds a utilitarian frame, Regan a rights based one, Nussbaum a capabilities based one, and Donaldson one based on political membership. Sources reaching the same conclusion through four different justifications, strong for an assignment.