Week 6: how do we know, and what don't we know

The rod and frame test · what it was built for from the reading

1 Witkin and Asch built the rod and frame apparatus in 1948. What were they originally using it to tell apart?

1. AMen from women.

   That is what it was turned into from the 1950s, and it is not what it was built for, which is the thing the case turns on.

2. BPeople who are comfortable in the dark from people who are not.

   Close to what Fausto-Sterling later argued it might be measuring. The question asks what it was built for.

3. CPeople who judge how things are oriented by using their surroundings, from people who can judge it without them.

   Field-dependent against field-independent. The spatial-ability reading came later, and the apparatus was never designed to test that.

4. DPeople with good eyesight from people with poor eyesight.

   No vision test is involved. Both groups see the rod and the frame perfectly well.

The rod and frame test · what you actually have to do from the reading

2 To complete the test you sit in a dark room in a chair that a male experimenter can tilt, and you have to keep telling him to move the rod a little more, and again, and again, until it is right. What did Anne Fausto-Sterling draw from that in 1985?

1. AThe experimenters were deliberately sexist in designing it.

   She makes no such charge, and the chapter is careful that these were respected studies. Her objection survives everybody acting in good faith.

2. BIt is unclear what the test measures, since finishing it requires a willingness to keep making demands of a strange man, which is not a fact about the brain.

   Women in the 1950s and 1960s were raised not to be demanding and not to make work for a man, and some would plainly have been uncomfortable alone in the dark with a male stranger. So the score may record who was willing to keep asking. Later studies that controlled for this found the gap shrank sharply.

3. CThe lighting conditions made the task harder for women than for men.

   Everybody sits in the same dark. The asymmetry she names is a social one.

4. DThe sample sizes were too small to support the conclusion.

   A statistical complaint, and hers is about what the apparatus is measuring in the first place.

The nine phases · why cut it up at all from the reading

3 School teaches that the scientific method is: form a hypothesis, test it, draw a conclusion. The chapter replaces that with nine phases. What is wrong with the school version?

1. AIt is not so much false as useless, because every interesting decision is hidden inside the word test.

   Unpacking test is the whole exercise. Which observations count, how to collect them, what they mean, when you have enough, what else might explain them: a person decides each one, and nothing in nature decides it for them.

2. BIt describes physics and not biology.

   The chapter makes no distinction between fields here.

3. CIt is false. Science does not use hypotheses.

   Hypotheses are real and phase 2 is where they come from. The complaint is different.

4. DIt is too long and the phases compress it.

   Nine steps is longer than three, so the move is in the other direction.

The nine phases · the two warnings from the reading

4 The authors attach two warnings to the list of nine phases. Which two?

1. AThe phases are value-free; and only the researcher can introduce values.

   The chapter exists to deny the first half.

2. BThe phases are not a sequence, since research loops back; and they interact, since how you framed the object decides what counts as relevant later.

   Both matter for an exam. A decision at phase 7 can send you back to revise phase 3, so the list is a tool for looking at inquiry, and no part of it is a timetable. And because phase 1 feeds phase 3 which feeds phase 5, you cannot quarantine values in the early phases.

3. CThe phases apply only to experimental sciences; and they were devised in the nineteenth century.

   Neither claim appears, and the examples run across fields.

4. DThe phases must be done in order; and skipping one invalidates the research.

   The first warning says the opposite of this.

The nine phases · which ones this chapter is about from the reading

5 Chapter 3 dealt with phases 1 to 3. This chapter takes phases 4 to 9. Why does the split fall there?

1. ABecause the first three are done by institutions and the rest by individuals.

   No such division of labour is described.

2. BBecause phases 1 to 3 are where the values are, and 4 to 9 are value-free.

   The reverse of the chapter's argument, which is that values reach into all of them.

3. CBecause 4 to 9 are the controversial ones. Hardly anybody denies scientists pick their topics for human reasons; the fight is over whether values reach into collecting and interpreting the evidence.

   Choosing to study breast cancer because it matters to people is uncontroversial. Claiming that a value shaped how the data got read is the claim people resist, so that is where the chapter spends itself.

4. DBecause phases 1 to 3 happen before funding is awarded.

   Funding is not the dividing line used.

Inductive risk from the reading

6 Deciding you have collected enough evidence is phase 6. The chapter says that decision is never purely factual. Why?

1. ABecause statistical significance thresholds are set by convention.

   Conventions appear in the chapter's discussion of effect size, and the inductive-risk point is about the cost of error.

2. BBecause funders impose deadlines.

   A real pressure on research and not the argument here.

3. CBecause more data is always better, so any stopping point is arbitrary.

   The chapter does not treat the decision as arbitrary. It treats it as a judgment with a shape.

4. DBecause how much is enough depends on the probability the hypothesis is true and on what it would cost to be wrong, and the second half is a value judgment.

   That is inductive risk. The cost of a mistake is not something the data can tell you, so somebody weighs it, every time, and what they weigh depends on what they think matters.

Inductive risk · both directions from the reading

7 The authors run the rod and frame case both ways, which is the move that turns it into an argument. What does wrongly accepting that spatial ability is fixed by biology cost?

1. AIt wastes research funding that could have gone to other projects.

   True of any wrong result and not the cost the chapter traces.

2. BIt makes the field look unscientific to outsiders.

   Reputation is not the currency here.

3. CNothing, since a false belief about brains has no practical effect.

   The chapter's whole point is that it does, and it names where.

4. DIt reinforces a stereotype, which bears on whether girls are encouraged into mathematics, engineering and architecture, and those are well-paid fields, so the cost appears in women's earnings.

   And the error spreads easily, because the stereotype is already everywhere and people weight evidence that agrees with what they already believe. The other direction has a cost too: wrongly rejecting a real difference loses the chance to understand it and may back policies that cannot work.

Phase 8 · the one that does the damage from the reading

8 Social and environmental explanations for the rod and frame results existed, and so did the possibility that the test measures assertiveness. Neither was ruled out. Why does the chapter treat that as the most serious failure?

1. ABecause concluding means inferring to the best explanation, and you can only compare explanations you have thought of, so who is in the room changes how well supported the conclusion actually is.

   That is the step from a point about individuals to a point about communities. An explanation that never occurs to you is not an explanation you have ruled out, and the strength of the conclusion depends on a list that somebody's imagination produced.

2. BBecause the alternatives were published and ignored.

   They were not in the literature waiting to be read, and the case is useful for exactly that reason.

3. CBecause the experiment was never replicated.

   Versions of it are still in use, and papers from 2019 and 2022 still draw the same conclusion.

4. DBecause the researchers knew about them and suppressed them.

   The chapter says the opposite, and is careful about it. These were respected studies and the alternatives simply did not come to mind.

Why the authors refuse to call it bad science from the reading

9 In hindsight the rod and frame research looks like sloppy design and hasty conclusions. The authors will not describe it that way. What is their reason?

1. ABecause these were peer-reviewed studies that still shape a live research field, and the ways values entered, through inductive risk and through which alternatives come to mind, look inherent to knowledge production, and unlike faults anybody could correct.

   That is why the chapter cannot end with a demand to try harder. If the entry points are built into how knowing works, the fix has to be structural, which is what the three accounts of objectivity are for.

2. BBecause no result has yet been shown to be false.

   Later controlled studies found the gap shrank sharply, so some of it has.

3. CThat the researchers were working to the standards of their time and cannot be judged by ours.

   A reasonable principle and not the one they use. Theirs is about the present, since the conclusion is still being drawn.

4. DThat calling it bad science would be impolite to a respected field.

   Nothing in the chapter turns on politeness.

Two things called objectivity from the reading

10 The chapter uses two older notions of objectivity constantly. What is the difference between them?

1. AMethodological: used in the sciences. Agent: used in the humanities.

   No such division appears.

2. BMethodological: objectivity as a matter of degree. Agent: objectivity as all or nothing.

   Degrees come up later, with Longino, and they do not separate these two.

3. CMethodological: the procedure is what makes it trustworthy, because anyone could run it and get the same answer.

   Agent: the person is what makes it trustworthy, because they are detached and have no stake. Both come from the value-free ideal, the view that good science carries no social, political or moral commitment anywhere inside it. Every feminist position in the chapter rejects that ideal, and almost none of them gives up objectivity, which is the combination to be able to state.

4. DMethodological: about experiments. Agent: about theory.

   Neither term is tied to a kind of work.

Spontaneous feminist empiricism from the reading

11 One response to sexist results is that the standard methods are fine and were simply applied badly, so following the procedures properly will filter the values out. Why does the chapter reject it?

1. ABecause practices thought to be good science can carry values inside them where nobody sees them, which is exactly the rod and frame test, and the obvious line between legitimate and illegitimate uses does not hold, since the phases feed each other.

   Nobody broke a rule in that case. The value was in the apparatus, in what it was taken to measure, and in which alternatives came to mind. So the job is no longer removing values but telling legitimate uses from illegitimate ones, and the tidy answer, values are fine in phases 1 and 2 and not after, fails because phase 1 decides what counts as relevant in phase 3.

2. BBecause it requires more funding than is available.

   A practical objection and not the one raised.

3. CBecause it was proposed by critics of feminism.

   Harding names the position, and the chapter engages with its content.

4. DBecause the standard methods have been shown to be unreliable in general.

   No such general claim is made, and the chapter keeps empirical standards.

Longino · what evidence is from the reading

12 Longino's account of objectivity follows from her account of evidence. What is that account?

1. AEvidence is whatever can be measured.

   Measurability is not the condition. The condition is about what connects a fact to a hypothesis.

2. BNothing is evidence on its own. A fact becomes evidence for a hypothesis only through background assumptions that connect the two, so changing the assumptions can stop it being evidence, or make it evidence for something else.

   She calls this contextual empiricism: empiricist because only observable things count, contextual because what they are evidence of depends on the surrounding assumptions. The objectivity argument then follows, since assumptions are shared and invisible to those who share them.

3. CEvidence is a fact that no one disputes.

   Disputed facts are still evidence on her account, which is why criticism has something to work on.

4. DEvidence is whatever a majority of researchers accept.

   Consensus appears in her account of communities and it is not her definition of evidence.

Longino · the housework case from the reading

13 The standard economic account says women earn less, so they have less bargaining power, so they do more housework. It predicts that as a woman's earnings rise her share falls. Julie Brines found in 1994 that when women out-earn their male partners, the men do even less housework than before. What does the case show?

1. AThat on the bargaining account the number is not evidence for or against anything, and only once you assume a marriage is also somewhere people perform gender does it become the central thing to be explained.

   Same number, two contexts, evidence in only one of them. A man who is out-earned has failed at the breadwinner part, so he refuses the part coded female. The feminist commitment that gender is relevant is what made the fact visible, and the displaced assumption had been shared across the field, which is why nobody had been troubled by the anomaly.

2. BThat economic explanations should be abandoned in favour of cultural ones.

   Her point is that the second assumption is added to the first, and the bargaining mechanism is not denied.

3. CThat the bargaining account is confirmed by a surprising result.

   The finding runs against what that account predicts.

4. DThat the data were collected badly.

   The chapter treats the finding as sound and asks what to make of it.

Longino · the condition with teeth from the reading

14 Longino names four conditions for transformative criticism: public venues for criticism, uptake, public standards, and tempered equality of intellectual authority. Which one does the real work, and why?

1. ATempered equality, because it guarantees every opinion counts equally.

   Tempered means the opposite: expertise still counts, and the rule is that nobody with relevant expertise is dismissed for gender, race or class.

2. BPublic venues, because without somewhere to publish, criticism cannot happen.

   Necessary and not sufficient. A community can host criticism and ignore all of it.

3. CPublic standards, because criticism needs something to appeal to.

   Also necessary, and it says nothing about whether anybody listens.

4. DUptake, because the community must not merely tolerate criticism but actually change its beliefs over time in response.

   Tolerating criticism costs nothing. The requirement that beliefs move is the one a real community can fail, and it is what makes objectivity come in degrees, since no community meets the four fully.

Longino · what diversity is for from the reading

15 On Longino's account diversity acquires a job it did not have before. What is it?

1. AA community that all shares the same background assumptions cannot see them, so difference is what makes criticism possible at all.

   That is an epistemic argument for diversity, on top of the moral one. If no individual can get out from under their own assumptions, the only thing that can is other people, and only people who do not share them.

2. BDiverse teams work faster, so more research gets done.

   Productivity is not the claim.

3. CDiversity makes the results more acceptable to the public.

   Acceptability is not the argument.

4. DDiversity is required by funders, so it has become a practical necessity.

   A real pressure and not her reasoning.

Harding · the inversion thesis from the reading

16 Standpoint theory runs on a claim with a name. What does the inversion thesis say?

1. AThat the social position of a researcher has no bearing on their findings.

   The value-free ideal, which standpoint theory rejects.

2. BThat dominant groups have systematically false beliefs about everything.

   Too strong, and the thesis is about a position being better for noticing certain things.

3. CThat people who are socially, politically and economically subordinated are, because of that position, better placed to notice the assumptions researchers in dominant positions are making.

   Two mechanisms, both concrete. A person inside the dominant framework cannot see its assumptions, because they are what seeing is done with. And marginalised knowers can state their own interests, so they can propose different accounts of what the object of study even is.

4. DThat knowledge should be inverted, meaning the conclusions of dominant science should be reversed.

   A pun on the word. The thesis is about who is placed to notice what.

Harding · the guard rail from the reading

17 The obvious objection to standpoint theory is that it hands anybody from a marginalised group an automatic advantage in knowing. What does the chapter say to that?

1. AIt accepts the consequence, since the advantage is the point of the theory.

   The chapter is explicit that this is not the claim.

2. BEpistemic advantage is not automatic. It does not come from identity or social position by itself, but from conscious awareness of the power relations that make up that position.

   Being a woman does not give you the standpoint. Understanding how gender structures the field does. This is the sentence that saves the theory from being silly, and it is the likeliest short-answer prompt in the chapter.

3. CIt abandons the inversion thesis in response.

   The thesis stands, with this condition attached to it.

4. DIt restricts the theory to the social sciences.

   No such restriction is made.

Harding · weak and strong objectivity from the reading

18 Harding defines two grades of objectivity. What separates them?

1. AWeak: objectivity about facts. Strong: objectivity about values.

   Both grades concern the inquiry as a whole.

2. BWeak: meeting the traditional empiricist standards of evidence, which covers phases 3 to 7.

   Strong: scrutinising all the phases, including whether the subject and the questions suit the goals of the research. Her diagnosis is that good science lacked the methodological resources to detect widely held sexist and androcentric assumptions, so the standards themselves are what need fixing. Androcentric means centred on men, treating male behaviour as the standard case.

3. CWeak: applied by individuals. Strong: applied by committees.

   The grades are about how much gets examined. Who does it is a separate matter.

4. DWeak: qualitative research. Strong: quantitative research.

   No method is privileged either way.

Wylie · the elevator word from the reading

19 Wylie starts from a complaint about the word objective itself, borrowed from Ian Hacking. What is the complaint?

1. AThat it is too technical for ordinary readers.

   Her problem is that it is too empty.

2. BThat it is an elevator word: a word that does no work of its own and only raises the standing of whatever it is attached to.

   Calling a result objective says it does not suffer from some failing, without saying what it has. So she gives it content: objectivity belongs to knowledge claims and not to researchers, and a claim is objective to the degree it shows the epistemic virtues, meaning empirical adequacy, inferential robustness, explanatory power, internal coherence, and consistency with other established knowledge.

3. CThat it was coined in the nineteenth century and carries that period's assumptions.

   No etymology is at issue.

4. DThat it means different things in different languages.

   Translation is not the problem raised.

Wylie · where standpoint gets in from the reading

20 Wylie makes objectivity a property of claims, which sounds like it has no room for social position. How does standpoint get into her account anyway?

1. AShe argues that marginalised researchers produce more empirically adequate claims.

   Her mechanism is about which virtue gets priority. Who scores higher on one is a different claim.

2. BShe abandons standpoint theory and keeps only the virtues.

   Her account is built so that standpoint improves the judgments the virtues require.

3. CThrough funding, since who pays decides which claims get made.

   Funding is not her mechanism.

4. DYou can almost never maximise all the epistemic virtues at once, so at every phase researchers must decide which to prioritise, and that depends on their purposes and goals, which depend on where they are standing.

   A model that fits the data best is often not the one that explains the most. Somebody chooses, and standpoint improves that choice. This is why her version is the most conservative of the three and the easiest to defend, since it needs no prior commitment about social position.

Agnotology · what the field studies from the reading

24 Chapter 5 turns the question round. Epistemology asks how knowledge gets made. What does agnotology ask?

1. AHow to teach difficult material.

   Nothing to do with pedagogy.

2. BHow scientific errors are corrected over time.

   Error correction is a different topic, and ignorance here is often not an error at all.

3. CHow ignorance gets produced and maintained.

   It got going in the 1990s with Robert Proctor. Feminists, critical race theorists and decolonial theorists developed a related line called epistemologies of ignorance. Sullivan and Tuana's definition: identifying the forms of ignorance, examining how they are produced and sustained, and asking what role they play in knowledge practices.

4. DHow to measure what a population does not know.

   Surveys of public knowledge are not the subject.

Ignorance · the thing most people get wrong from the reading

25 Some accounts treat ignorance as simply the absence of knowledge, or the lack of true beliefs. The chapter rejects that. What is its position?

1. AThat ignorance is always deliberate.

   The five kinds include ones nobody chose, and the fifth is not a failing at all.

2. BThat ignorance is always harmful.

   The chapter spends real time denying this: blinding a clinical trial, restricting how to build a biological weapon, and things said about you that you would rather not know.

3. CThat social and cultural forces leave gaps and also make certain things hard or impossible to know, so ignorance can be a substantive epistemic practice in itself.

   Linda Alcoff's phrase is the one to keep. Treating it as a simple absence misses every interesting case, because an absence does not have to be produced and maintained, and these are.

4. DThat ignorance can always be removed by better education.

   Willful ignorance in particular survives the evidence being put in front of somebody.

Tuana · the first two kinds from the reading

26 Tuana's first kind is what we do not know but do not care to. Her second is what we do not know that we do not know. Male hormonal contraception and the anatomy of the clitoris illustrate one each. Which goes where, and what separates them?

1. ABoth are the first kind, and the second kind has no medical examples.

   Tuana's own example of the second kind is the clitoris.

2. BBoth are the third kind, since industry concealed the facts in each case.

   Concealment by an industry is the third kind, and its case is oestrogen risks in contraceptives.

3. CMale contraception is the second and the clitoris the first, since contraception is the better-known case.

   Fame is not the criterion. The criterion is whether anybody ever considered it.

4. DMale contraception is the first, since it was weighed up and dropped.

   The clitoris is the second, since it was never treated as an object of study at all. Contraception was considered: men would be less willing to accept side effects such as reduced libido, while nobody worried about reduced libido in women, because of the standing assumption that women did not have much desire anyway. The clitoris was different. Study of the genitals was organised around reproduction, the clitoris was taken to play no part in it, so it was not a thing to be ignorant about. It was never rejected as unworthy, because it was never an object of enquiry.

Willful ignorance against self-deception from the reading

27 The chapter draws a careful line between self-deception and willful ignorance. What is it?

1. ASelf-deception is a moral failing and willful ignorance is a cognitive one.

   The chapter does not sort them that way.

2. BSelf-deception is unconscious and willful ignorance is deliberate.

   Both involve something the person is doing, and the difference is in what they do with the question.

3. CSelf-deception concerns oneself and willful ignorance concerns others.

   Both can be about either.

4. DSelf-deception is convincing yourself something is true because you want it to be.

   Willful ignorance is fearing it might be true, having an interest in not believing it, so avoiding the question altogether and building ways of minimising whatever evidence turns up. Tuana's phrase is an active ignoring of the oppression of others and one's own role in it. The self-deceiver reaches a conclusion. The willfully ignorant person takes care never to reach one, which is why it is the only kind where the ignorant person is doing work.

Mills · white ignorance from the reading

28 Charles Mills gave white ignorance its name in The Racial Contract, 1997. What is his claim?

1. AThat a society built on inequality gives the group on top an interest in seeing the world a particular way, and the not-knowing is maintained because it is useful.

   People with advantages have an interest in believing the world is fair and that what they have is deserved, and no interest at all in learning otherwise, because of what that would imply. So the refusal to understand the conditions others live under is doing a job.

2. BThat racial inequality is caused by ignorance, with institutions playing no part.

   His argument runs the other way: the institutions produce the ignorance.

3. CThat ignorance about race is impossible to correct.

   Spelman's analysis describes how corrections get deflected, and nobody claims it cannot happen.

4. DThat white people happen not to have been taught certain history.

   A gap in a curriculum would be the first kind of ignorance. His claim is stronger and about interest.

The two quotations put side by side from the reading

29 In 2017 Kate Winslet called the women speaking out about Weinstein incredibly brave and said it was deeply shocking to hear. In the same year, starring in a Woody Allen film, she said she did not know him or anything about that family, that you step away and say you do not know whether any of it is true or false, and that he is an incredible director. What does the chapter draw from the pair?

1. AThat she is not prepared to say the allegations are false, because that would mean weighing Dylan Farrow's testimony, which in the other context she said should be believed, and she has no interest in its being true, so the comfortable position is not knowing.

   Actively ignoring the evidence and reaching no conclusion in either direction. That is willful ignorance with a career attached, and the chapter uses two people saying both things in one year so that the pattern cannot be read as one person's inconsistency.

2. BThat celebrities should not comment on legal matters.

   The chapter is analysing a kind of ignorance. It issues no advice.

3. CThat she changed her mind between the two interviews.

   The two were given in the same year, which is why they are printed together.

4. DThat the Allen allegations are weaker than the Weinstein ones.

   The chapter sets out why it is not a word against a word: eyewitness corroboration, experts finding no sign of suggestion, a judge denying custody over the risk, and a prosecutor finding probable cause.

The fifth kind · loving ignorance from the reading

30 Tuana's fifth kind is what we cannot know: limits on perception, on time and money, and on understanding experiences you have not had. Lugones and Ortega turn it into something other than an excuse. What do they ask for?

1. AThat the limits be removed through better research methods.

   The fifth kind is a real limit, and the only one that is no failing.

2. BThat researchers study only their own communities.

   The opposite of sustained engagement across difference.

3. CThat the limit be acknowledged and the subject then dropped.

   Dropping it is what they are arguing against.

4. DAccepting that some experience cannot be fully shared, while refusing to ignore, dismiss or stereotype the people whose it is.

   Ortega adds that reading about the experiences of women of colour is not enough and produces its own arrogance, so what is wanted is sustained engagement and real relationships. Loving ignorance. The limit is real and it licenses nothing. Lugones calls for a loving perception of those who have been marginalised, and Ortega's sharpening is the part that stings, since it is aimed at people who think they have done the work by reading.

The five ways ignorance gets made · discrediting from the reading

31 The chapter names five production strategies: neglect, suppression, discrediting, and two more. Discrediting attacks either the knowledge or the knower. How does attacking the knowledge usually work?

1. ABy showing the researchers were paid by an interested party.

   That is attacking the knower, and the question asks about attacking the knowledge.

2. BBy proving the claim false.

   Disproof would be ordinary science. The strategy works without it, and that is where its cheapness comes from.

3. CThrough uncertainty, which is the tobacco strategy: executives knew they did not have to show smoking was safe, only that the science was unsettled.

   They did it by funding conflicting research, exaggerating the uncertainties, and putting non-scientists on television to attack the science. The institutional vehicle is the think tank: it employs researchers, publishes reports, is funded by whoever chose to fund it, has no students and no requirement to be peer reviewed, and its output looks like scholarship, so it is useful to somebody who wants a conclusion to appear independent.

4. DBy having the research retracted from journals.

   Retraction is not the mechanism described.

The difficulty the chapter raises against itself from the reading

32 Having set out discrediting as a way ignorance gets manufactured, the chapter immediately raises a problem with its own category. What is it?

1. AThat nobody has ever been successfully discredited.

   The chapter gives several cases where it worked.

2. BThat discrediting mimics reasonable practice, since it is often right to ask whether a witness is credible or an expert qualified.

   Their example of a fair challenge is Scott Atlas, a radiologist with no training in epidemiology, advising the White House on masks. Telling fair criticism from manufactured doubt is genuinely hard, and the chapter says so, where it could have pretended the category was clean.

3. CThat discrediting was invented by the tobacco industry and so is historically recent.

   No such claim about its origins is made.

4. DThat discrediting only works on scientific claims.

   Their examples include Indigenous knowledge and the testimony of individuals.