Category Archives: philosophy

in defense of (some) implicit bias

Leave a reply

I hope that if there were an implicit bias test for Nazism, I would demonstrate a strong negative bias. Shown rapid-fire images of swastikas and Nazi leaders, I would be unable to associate them with positive words without strenuous effort. The reason is that I learned a deep aversion to National Socialism, based originally on reasons and evidence. It is now no longer efficient for me to use conscious effort to assess Nazis, their pros and cons. I have rightly translated a very well-founded judgment into a habit, which works like a constructed instinct. That way, I can reserve my limited attention and cognitive capacity for other issues.

In 1970, Charles Fried proposed as a philosophical thought-experiment a situation in which two people are drowning, one of whom happens to be your spouse. It was “absurd,” said Fried, that you should be impartial about which one to save. Fried was developing an argument against pure impartiality. But Bernard Williams famously replied that you shouldn’t even have to think about which person to save. That would be “one thought too many.” If you must reason about whether to save your spouse as opposed to someone else, you do not love your spouse. The problem with having to think in this case is not mere inefficiency (it might slow you down and increase her chance of drowning). It’s more basic than that. You do not have a “deep attachment” to another person unless—here I extend Williams’ argument—you have turned your preference for that person into an acquired instinct. Your ability to act on that instinct instead of reasoning is proof of a process that we call love.

In a really interesting new paper, “Rationalization is Rational,” Fiery Cushman argues that human beings, since we have limited cognitive resources, have evolved several different modes of representing things in our environment: reason and planning, habit, instinct, and norms. These modes require varying amounts of cognitive attention. Cushman also proposes that we have evolved mechanisms for shifting representations from one mode to another for efficiency’s sake. For instance, we intentionally learn the way home and then form a habit of walking home so that we no longer have to think about it. But we can also make a habit conscious and practice until we change it.

Many people are currently worried about two specific “representational exchanges,” in Cushman’s terms. One is rationalization. We think that we are making a conscious and reasoned choice, but we have actually formed an instinctive reaction that we then merely rationalize with explicit words. This phenomenon is widely taken to be evidence of human unreason and inability to deliberate. But Cushman sees it as an efficient process. We can’t go through life assessing everything explicitly, so we develop habits of reacting to categories of things and then justify our reactions when reasons are needed. So long as the learned habit was based on good thinking in the first place, it is an efficiency measure rather than a limitation. In turn, rationalization (giving reasons for something we have already decided) serves a useful purpose: it puts a habit into verbal form so that it can be debated.

The other problem that worries many of us implicit bias, particularly in the form of racial stereotyping. Tests of implicit bias show that various forms are common in the population as a whole.

Implicit bias research sometimes seems to flatten crucial moral differences. A subject might have a 3% bias against African Americans and a 24% bias against Millennials. This does not mean that generational bias is eight times more important, even in this individual’s case. Racism is structural, historical, connected to laws and institutions, and literally deadly. Generational bias is just one of those things we should probably think about. To assess the empirical data about bias, we need judgments about what is just and unjust.

Applying Cushman’s insight, I would go further. An implicit bias is not necessarily bad at all. It is actually a virtue (in the Aristotelian sense) if it reflects a process of reasoning and learning that we have stored as a habit. Being biased against Nazis and in favor of your spouse are virtues. Being biased against people of color is a vice. The difference lies in the content of the judgment, not the form.

It’s true that any bias can mislead. For instance, your appropriate abhorrence of Nazism might distort your views of justice in the current Middle East. Your appropriate bias in favor of your dearly beloved family members might cause you to treat strangers in unjust ways. It is characteristic of virtues that each is insufficient; we need a whole suite of them. And one important task is to bring even our best biases into conversation with other ideas and principles. But it wouldn’t be progress to temper your bias against Nazis or in favor of your spouse. That would just weaken your virtues. Progress means combating bad biases, developing good biases, and combining your good biases with more abstract principles of judgment.

See also: the era of cognitive bias; marginalizing views in a time of polarization; Empathy and Justice; Jonathan Haidt’s six foundations of morality; and don’t confuse bias and judgment (which is incompatible with this post).

decoding institutions

Leave a reply

Today I presented at Tufts’ Science, Technology & Society lunch seminar series on how knowledge and power interrelate. My basic thesis was that knowledge is produced by institutions, which are fields of power. Assessing knowledge therefore requires analyzing institutions (not claims about facts by themselves).

The general model I am assuming works like this.

Actors can be individual people or (at larger scales) such entities as firms, bureaus, or even nations. They have goals; mental constructs such as philosophies, identities, or ideologies; and relations with each other.

They interact in an Action Space, such as a market, a democratic election, or a scholarly publication. Their interactions vary, but actors always make choices shaped by rules, norms, and goods.

A “norm” is a shared expectation that has a positive moral valence. For instance, Robert K. Merton’s CUDOS Norms for science are values that are widely expected. An actual “rule,” on the other hand, structures outcomes but may not have a positive moral valence. Merton also coined the phrase “Matthew Principle” for the general rule that, in science, the person who is already most famous gets the most credit. That rule conflicts with the CUDOS norm of Universalism.

Action Spaces affect, and are influenced by, biophysical conditions, general social circumstances (e.g., poverty), and other institutions.

The institution as a whole has Inputs and Outputs. Insofar as the institution involves knowledge, Inputs may include ideas, opinions, and knowledge-claims and it may produce new ideas, opinions, and knowledge-claims.

We can assess the whole process in terms of value criteria, such as justice. Such assessments not only influence institutions; they are also shaped by institutions. In fact, we don’t have information or values that we can use for assessment except for those that have emerged from institutions. The interaction is reciprocal.

Each element of the whole system is a target for power. To use Stephen Lukes’ Faces of Power framework: one “face” involves actors influencing other actors within an Action Space; a second “face” involves changing the rules of the Action Space; and a “third face” involves changing either norms or the actors’ mentalities, or both. But we could add many more “faces” as we consider each element in the diagram.

We rarely assess knowledge directly, because we are rarely in a position to have justified true beliefs all on our own. Instead, we must assess knowledge as the product of institutions. But that is not a relativist claim, because some institutions are better than others. Assessing the value of an institution requires taking it apart and assessing its components.

See also: adding democracy to Robert Merton’s CUDOS norms for science; is all truth scientific truth?; tools for the #resistance; and a template for analyzing an institution

we should be debating the big social and political paradigms

Leave a reply

[The following post is inspired by Rogers Smith’s recent APSA presidential address, and I believe is consistent with it, but I incorporate some additional elements.]

Consider a sample of articles about politics, or society more generally. One might be an interpretation of some remarks by Thomas Jefferson. Another reports an experiment in which people are more likely to vote when contacted one way rather than another. A third is a critique of a prevalent word (say, “security”) from the perspective of Michel Foucault. A fourth shows that rational agents in a certain situation are better off when rules are externally imposed on them.

Each of these articles presumes, and contributes to, a larger paradigm that combines values, causal theories, and methods. Perhaps the interpreter of Jefferson believes that the United States is still founded on its written Constitution and that the intentions of the founders matter today. The experimentalist presumes that voting in a regime like the modern USA is a consequential act, and it would be good to increase voting by several percentage points if possible. The critic of “security” sees a world of pervasive injustice sustained by ideology, in the sense of distorting views that preserve the status quo. The rational-choice modeler thinks that all (or many) institutions can be analyzed as the interactions of utility-maximizing individuals.

(For the sake of space, I have omitted a classical Marxist, a deep feminist, a post-colonial theorist, a Prospect Theorist, a radical environmentalist, and many others who deserve places on the list.)

In these articles, you will not see much about their larger paradigms, worldviews or schemata. Their paradigms may not even be mentioned. Instead, you will see evidence in support of the specific claims of each article (whether in the form of statistics, quotations, or equations).

Of course, journals offer limited space, and there may not be room to present a whole paradigm. In addition, citing your scheme may only hurt you in the review process. Reviewers who happen to oppose your overall paradigm may be alienated, when they would have been persuaded by your detailed findings if presented alone.

For instance, imagine that you discover that texting people increases their turnout more than emailing them. That is more persuasive as a bare finding than as part of an argument for the significance and value of voting in a mass democracy within a capitalist market economy.

A certain form of positivism (or verificationism, or empiricism) is still widely influential. It holds that facts can be verified directly. Larger mental constructs are fashioned by us and are only valid to the extent that they match all the facts. If big, general ideas influence our beliefs, they are sources of bias. Therefore, we may need to disclose our paradigms as caveats, but we don’t want to focus on them.

Some people would hate to be described as positivists but end up in a similar place for a different reason. They presume that they have a right to their fundamental views as a matter of identity. “As a migrant from the global south, I explore the colonization of indigenous spaces …” This is not a disclosure of bias, nor a defense of a thesis; it is a claim to be recognized as a member of the scholarly community. To disagree is to deny the author’s place in the community.

I believe that the really important task is to select our worldviews, our big normative/conceptual schemes. It matters which of the available choices we adopt, and maybe we can create better ones. Therefore–as I think Rogers Smith argued–it is a real weakness in any intellectual community if the paradigms are implicit or merely stated, rather than explained, justified, and put into competition.

Within political science, many contrasting worldviews are available. You can walk down the hall of the APSA Annual Meeting past rooms in which everyone accepts the basic normative principles of contemporary electoral democracy, and other rooms in which people are quoting Slavoj Žižek about ideology.

It is hard to compare such worldviews or paradigms, because they have different normative, epistemic, and sometimes metaphysical premises. A finding within one paradigm does not disprove a different paradigm. A certain kind of relativism interferes with comparative assessment.

However, we can consider a whole body of specific findings along with their shared overall claims. We can ask whether this whole literature is coherent, whether it generates persuasive specific findings by its own lights, whether it informs practice in any useful way, whether it makes sense of other literatures, how it handles criticisms and rival views, whether it is responsive to new evidence and events, and whether its normative implications are defensible when fully articulated. Each paradigm has points of relative strength and weaknesses.

It would be helpful if people had that discussion in print, instead of always only writing within paradigms (at least for journal articles.)

See also trying to keep myself honest, how philosophy is supposed to work

Moral Foundations theory and political processes

Leave a reply

Moral Foundations theory is an important body of research that has generated significant findings. It belongs to a somewhat larger category of research about morality that has these features:

  • Many individuals are asked for their moral judgments of hypothetical or historical cases. (See this test, for example.)
  • The data from these individuals is analyzed to identify latent factors that statistically predict many people’s specific answers. Respondents may not be aware of the latent factors or even endorse them when they hear them, yet these factors are said to explain (statistically, and maybe causally) their actual judgments.
  • Some factors are found to be common across cultures and to have adaptive value for human beings as a species that evolved through natural selection. These factors are named “foundations.”
  • Variation among human beings depends on which of the foundations are more important for different groups. This is a structuralist theory, akin to Claude Levi-Strauss or Noam Chomsky. For structuralists, surface-level diversity results from variation in a finite set of underpinnings.
  • In ambitious versions of this method, morality is the list of foundations. Morality is what we think it is because of how we evolved. To debate the value of the various foundations or to argue for a different principle is pointless. There is no way to know the truth of moral claims, but science can reveal the foundations of our moral psychology, so those alone are real.

Despite my interest in learning from the specific findings of Moral Foundations research, I harbor many objections, summarized here. In this post, I would like to elaborate on one concern.

We human beings make individual judgments about cases and decisions. That’s what the data-collection of Moral Foundations Theory models. But we do many other things that also shape values. For instance, we buy things, which causes more of those things to be made and sold. We participate in bureaucratic institutions that manage themselves through chains of command, policies, and files. We join and quit groups. And we govern by making laws and policies at all levels.

As Owen Flanagan writes in The Geography of Morals (2016), “Moral stage theory conceives moral decision-making individualistically. The dilemmas are to be solved by singleton agents. This is ecologically unrealistic. Committees at hospitals decide policies on organ allocation; government agencies deliberate about monetary policy; officers deliberate about costs and benefits of military operations; and friends talk to friends about tough decisions.”

I’ll focus on political decisions, although we should also consider markets, bureaucracies, friendships, scientific disciplines, and other social forms.

People make laws and other rules to prohibit, regulate, tax, subsidize, and mandate various behaviors. These rules directly influence how we act and may also affect our values, if only because many of us display the Moral Foundation of “Authority.”

One might think that rules and laws are made by people who implement their moral judgments, which are influenced by their individual Moral Foundations. Laws would then simply be aggregated Moral Foundations.

Not so, for these reasons:

  1. An important determinant of the actual law is the choice of who gets to make it. That choice is often made by people other than those who do the governing. Millions of voters choose a president. A president plus 51 US Senators decides who sits on the Supreme Court (one opening at a time, spread over decades). In making these choices, people are not focused on the specific cases or controversies that come before the Court. Top of mind for a president may be finding a justice who can be confirmed, who impresses the electorate, and who is young enough to serve for a long time. These are different considerations from moral judgments about cases, but they shape the actual law. (Here I use a democratic example, but a dictator is also chosen, often by a military junta or by party cadres.)
  2. When multiple people make laws, they don’t do so by judging in isolation and then aggregating their votes. A secret vote may mark an important moment in a political process, but it is almost never the only moment. It usually follows argument, persuasion, and mobilization; and sometimes groups decide without voting at all. Communication plays an important role. But when we communicate, we are not merely expressing our moral judgments of concrete cases. We may be doing many other things: trying to go along with the group, trying to stand out, trying to look (or be) impartial or moderate, trying to waste time on a point of disagreement to prevent attention to a different topic that we fear (filibustering), trying to set a precedent for other topics, supporting someone else so she’ll help us later (logrolling), saying something to irk someone in particular, enjoying the sound of our own voices, and so on. Our moral judgment of concrete issues may be an input, but only one among many.
  3. The procedures for making collective decisions influence the outcomes. For instance, almost all procedures favor the status quo because it takes energy and agreement to shift it. Therefore, a group can live with a law or policy that every individual would prefer to change. This is common and it implies that law often fails to reflect the private opinions of the majority–even when everyone has an equal say, which is vanishingly rare.
  4. One way that laws and policies shift is that subgroups successfully advocate for changes, based not on abstract judgments but self-interest. I don’t think that attitudes toward sexual orientation changed over the past half century because the broad public decided to reconsider their views. I think sexual minorities felt compelled to take their struggles into the streets, the ballot booth, and the courts and won some significant victories. The resulting policies then began to change attitudes. (A different kind of example: girls’/women’s sports got a huge boost from Title IX, which originated when Congress rejected a mischievous amendment to exclude sports from civil rights legislation–a backdoor way to enact a policy that has changed everyday attitudes toward gender.)
  5. Often the subject of political or legal debate is not whether individuals should do specific things: marry, steal, take drugs. It is about our collective stance toward social constructs: the United States (or Russia), Christianity, the family, a forest. These things have histories, they are complex and internally inconsistent, and they reflect laws or norms that people have formed over time. Often we are not asked to assess concrete actions but big abstractions that embody, among other things, many previous concrete actions taken by many people for many reasons. A major question is what story we should tell about a large construct.

For these reasons, empirical moral psychology cannot stand on its own without institutional/political analysis. Moral Foundations Theory is strongest when it aims to predict how people will individually react to a situation that raises issues so general that it resembles the problems that confronted our prehistoric ancestors on the savanna. The theory is least helpful for explaining why the same group of people might change its stance toward a specific topic, as we see with sexual orientation since 1969.

See also Jonathan Haidt’s six foundations of morality; an alternative to Moral Foundations Theory; moral thinking is a network, not a foundation with a superstructure; and against methodological individualism.

trying to keep myself honest

Leave a reply

(Madrid) This summer–which is not over yet–has already been full of rich and challenging discussions for which I am grateful.

In June, I spent several days discussing some lesser-known works of Friedrich Hayek with a group of mostly classical liberals/libertarians.

In late June and early July, more than 160 experienced scholars, practitioners, and activists from many countries visited Tisch College for a series of linked events: the American Political Science Association’s Institute for Civically Engaged Research, a convening of city staff from 15 Cities of Service, a gathering of Bridge Alliance members, the eleventh annual Summer Institute of Civic Studies, Lead for America’s summer institute, and the Frontiers of Democracy Conference. These people certainly held diverse ideological views, but a strong voice came from participants whom I would associate with intersectional movement politics–people who favor bottom-up, extra-institutional movements to confront white supremacy, patriarchy, and related “-isms.”

And now I am in Madrid for the Ibero-American Meeting on Civic Studies. I am very much enjoying my academic colleagues from Argentina, Colombia, Ecuador, Guatemala, Mexico, Spain and Venezuela who hold diverse views. While here, I have also visited a traditionally “red” working class Madrid neighborhood and met with radical Spanish architects and have heard a senator from the PSOE (socialist) party lecture. They have given me a dose of European social democratic politics. In contrast to intersectional movement politics, this is largely about building mass institutions (unions and parties) for “the people,” understood as singular.

I remain basically an American center-leftist. Barack Obama is my favorite president and have sent a little money to Kamala Harris. But since I fear intellectual complacency and clichés, I am always grateful to have my presumptions challenged. Libertarianism, intersectional movement politics, and social democracy feel like a triangle of ideas that keep me (somewhat) honest from three directions.

I think I hear the classical liberals saying, “Society is too complex to be modeled, let alone regulated or planned, because it is a function of countless individual choices, and the millions of agents can react to any effort to constrain or guide them by changing their behavior. Opportunity costs are ubiquitous and especially difficult to measure. Talk of ‘social justice’ arrogantly replaces what individuals want in their own circumstances with a specific theory of what they should want and implies that someone has the right to enforce that. Instead, policy should be maximally general, durable, and predictable so that individuals can form and implement their own plans in their contexts.”

I think I hear intersectional activists saying, “People are dying as a result of racism and transphobia and sexual violence. That is because other people hold deeply seated world-views that categorize their fellow human beings into hierarchies and create boundaries. These world-views are fundamental causes of injustice and must be challenged. There is no substitute for the people at the top of the hierarchies [people like me] acknowledging their advantages and changing their own lives accordingly.”

And I think I hear the social democrats say, “When large numbers of ordinary people have organized themselves into unions, parties, and social movements, they have countered corporate capital and negotiated mixed economies that have generated equity and security along with prosperity. But such organizations require substantial discipline (constraining individual choice) and broad identities, such as ‘worker’ or ‘citizen.'”

See also: on hedgehogs and foxes; The truth in Hayek; identities, interests, and opinions