Social inequalities almost always generate contested causal explanations. Is poverty caused by too many or too few government programs, by an irresponsible ‘culture of poverty’ or by historical injustice?Footnote 1 Is the gender wage gap due to gender discrimination or different personal choices?Footnote 2 Is the underrepresentation of historically disadvantaged groups in certain fields (e.g., mathematics, physical sciences, engineering, and philosophy) explained by inferior abilities or mechanisms of bias?Footnote 3 And so on. Because people disagree about the causes of such inequalities, they also disagree about whether interventions ought to be pursued and, if so, what kinds.
While philosophers are usually highly concerned with problems of social inequality,Footnote 4 it is often thought that the real work of solving them—of determining causal explanations and interventions—is not a job for philosophers. ‘Well, that’s an empirical question,’ goes the oft-heard conversational disclaimer (or conversational derailer) in philosophical Q&As. After all, we are not social scientists. It has thus been asserted that philosophers should avoid making claims about the causes of poverty and restrict themselves to discussing obligations for relieving it,Footnote 5 or that we should stick to normative arguments for racial reparations that do not rely on specific causal explanations of racial inequality.Footnote 6 But this, I think, is too quick. An overly sharp division of labour between empirical and normative inquiry belies the fact that causal explanations, especially of social phenomena, necessarily depend on certain kinds of value judgements, e.g., judgements of moral responsibility. Thus, it is incumbent on moral philosophers to appreciate how the structure of causal explanation itself admits them a role that they are uniquely positioned to take up: identifying and modifying the normative commitments that shape empirical inquiry.
The essence of my argument is this: if (as others have argued) giving causal explanations requires selecting some factors rather than others, and if this selection depends on broadly normative commitments, it follows that (moral) philosophers already have a role in determining acceptable causal explanations. In what follows, I demonstrate the antecedent by drawing together several disparate lines of thought from feminist epistemology, philosophy of science, and political philosophy, and bringing them to bear on a case study of a particularly well-discussed form of social inequality: poverty. I organize these into a unified account of the value-ladenness of causal investigation, and deliver the upshot: that, by working to alter prevailing moral expectations, moral philosophers are simultaneously working to reshape the space of acceptable causal explanations for social inequality. I conclude with suggestions for how philosophers can better carry out this work, once we recognize that empirical and normative inquiry are a two-way street.
I. Individualist vs. Situationist Explanations of Poverty
In this section, I interpret some background data on different causal explanations of poverty by applying the core insights of feminist epistemology and philosophy of science. The starting point here is the recognition that all inquiry requires choices (e.g., selecting which problems to investigate, which hypotheses to test, which methods of data collection to use, and so on). What feminist work demonstrates is that different social circumstances—historical, political, economic, cultural, and personal—condition inquirers to make those choices differently. These points are vividly illustrated in people’s preferred causal explanations of poverty.
A decades-long program of survey research, spearheaded by sociologist Joe Feagin, has examined the causal explanations people give for poverty.Footnote 7 It is an exceptionally robust research paradigm: studies have been conducted in all of North America and all 28 European Union countries, southeast Asia, the Middle East, and parts of Africa, not only on university students, but also children, social workers, recipients of social services, and policy makers. These studies typically follow Feagin’s original tripartite framework, divided into 1) explanations locating the causes of poverty in the dispositions and behaviours of poor people themselves, e.g., lack of effort or willpower, substance addiction, promiscuity; 2) explanations locating them in broader social structures and processes, e.g., insufficient social benefits, low wages, discrimination; and 3) explanations appealing to fate or bad luck, e.g., illness, disability,Footnote 8 accident, loss of family.
In this paper, I focus on the first two types because they map onto an especially significant distinction within social psychology: the distinction between what I will call ‘individualist’ and ‘situationist’ explanations of behaviour. Individualist explanations attribute a person’s behaviour to stable dispositional traits of the person herself, while situationist explanations attribute behaviour to contingent features of the situation in which she is acting. Arguably, the core discovery of post-WWII social psychology was that people are prone to a ‘correspondence bias,’ that is, a tendency to prefer individualist over situationist explanations, even in the face of obvious and straightforward situational constraints.Footnote 9 Indeed, many of social psychology’s most notorious, surprising, and significant results (e.g., the Milgram obedience experiment, the Stanford prison experiment, and situational priming) are surprising only against a background of underestimating the power of situational constraints—a fact that has garnered the correspondence bias the reputation of being the ‘fundamental attribution error.’Footnote 10
For my purposes, the most important finding about correspondence bias is that, even though present more or less universally, its strength varies according to systematic social patterns. These patterns are demonstrated clearly in people’s preferences for individualist vs. situationist causal explanations of poverty.
First, these preferences are moderated by the type of socialization prevalent within a specific social milieu, e.g., a culture or a discipline. Across countries, for example, the popularity of individualist or situationist explanations varies according to whether dominant conceptions of the self are ‘independent’ or ‘interdependent.’ Societies with independent conceptions (e.g., in North America and Europe) generally emphasize how the self is uniquely different from others; by contrast, those with interdependent conceptions (e.g., in East Asia) emphasize the importance of harmonizing the self with others.Footnote 11 Accordingly, researchers have found that individualist explanations, including those for poverty, are more common in independent than in interdependent societies.Footnote 12 Similarly, within a student body, commerce and engineering students are more likely to adopt individualist explanations of poverty than those in social science, and social science students grow increasingly likely to adopt situationist explanations as they progress through their programs.Footnote 13
Second, differences in causal explanation are moderated by positions of relative advantage and disadvantage within a social hierarchy. Roughly speaking, people in wealthier countries favour individualist explanations, while those in poorer countries favour situationist ones.Footnote 14 Moreover, within a given society (including but not only in the U.S.), people who are male, white, and middle class are more likely to give individualist explanations of poverty, while people who are female, black, and working class are more likely to give situationist explanations.Footnote 15
Social psychologists posit that these variations in the correspondence bias are mediated by different ‘lay theories of attribution’ grounded in cultural and religious values alongside lived experience. In independent societies that value standing out from others, people are socialized to act in accordance with personal preference, which supports a lay theory of individualist causes. Conversely, in interdependent societies that value fitting in, people are socialized to give greater weight to norms, roles, and situational scripts, which supports a more situationist lay theory.Footnote 16 Educated middle-class people more often have the liberty and resources to act on their personal preferences, while working-class people more often encounter situational constraints that dominate their preferences when they try to act. The former is consistent with individualist explanations while the latter is more indicative of situationist explanations. These findings thus provide striking support for feminist epistemologists’ claims that processes of knowledge-making are shaped by inquirers’ social circumstances.
The takeaway from lay explanations of poverty, then, is that choices of causal explanation are socially conditioned: they vary according to different social circumstances. In the next sections, I develop a more detailed account of what makes this possible.
II. Causal Selection: Where Causality Meets Morality
Why and how exactly do the social circumstances of different inquirers lead them to reach different causal explanations? In this section, I propose that this is due to an inherent feature of causal explanation itself: the need for causal selection. I synthesize work by philosophers of science, on the one hand, and moral and political philosophers, on the other, to pinpoint the place where different normative expectations make a difference to causal inquiry.
Let me begin by emphasizing that the task of identifying causal selection is different from the task of identifying causal structure, the latter of which consists in distinguishing between causal and non-causal factors—that is, between which factors are causally relevant or not.Footnote 17 Causal selection, instead, refers to the singling out of some factor as ‘the cause’ of some event or outcome, while designating other factors to be merely enabling conditions, even when both causes and enabling conditions are equally metaphysically necessary (and jointly sufficient) for the event or outcome. Such selection is crucial for causal explanation, because explanations necessarily foreground certain relevant factors at the expense of irrelevant others in order to answer the ‘why’ question at hand.Footnote 18 Outside very special contexts, for instance, it is simply not giving an explanation of some event or outcome to list every single factor that was individually necessary and jointly sufficient for it, or to cite, as Bertrand Russell would, the “whole state of the universe” at some prior time.Footnote 19 This means that I am not making metaphysical claims about the actual causal structure of poverty; my claims are rather psychological and sociological claims about how we make causal selection judgements. In other words, the question is not whether individualist or situationist factors are metaphysically necessary for poverty (since it is clearly the case that both individualist and situationist factors are present metaphysically), but why we should focus on some of the metaphysically necessary factors rather than others.
Consider a classic example of causal selectionFootnote 20: in a case where children playing with matches accidentally start a fire, we would ordinarily explain it by saying that ‘Striking the match caused the fire’ rather than ‘Oxygen in the air caused the fire,’ even though both were equally necessary for there to be a fire. Why do we choose the match-striking rather than the oxygen? Broadly speaking (and greatly oversimplifying), there are three major strands of thought: they appeal to abnormality, manipulability, and morality. On the abnormality view, we select as cause the condition that is abnormal or unusual, in the statistical sense, for those circumstances.Footnote 21 Because oxygen is typically in the air and hence statistically more normal, we choose the match-lighting instead. A second proposal is that we select causes based on what we can control.Footnote 22 On this view, we choose match-lighting as the cause of the fire because, unlike the presence of oxygen in the air, it is something we can manipulate in order to achieve the effect of producing a fire. Finally, a recent line of research in experimental moral psychology has suggested that morality, i.e., judgements of the wrongness or blameworthiness of an action, affect which factors people select as the cause of some event or outcome.Footnote 23 On this third view, we select those who are blameworthy as the cause of the fire, e.g., the older child who should have known not to play with matches, rather than the younger sibling who innocently performed the very same actions.
I do not have space here to defend any specific account of causal selection. Moreover, the three approaches are interwoven with one another in complex ways throughout the literature.Footnote 24 For my purposes, what matters is that (1) there is universal agreement on the importance of distinguishing between the task of identifying causal structure and the task of causal selection, and (2) while some deny the existence of any principled basisFootnote 25 for causal selection, those who do identify principles of selection inevitably resort to some sort of sensitivity to context, interests, or norms. This is typically achieved within a two-stage frameworkFootnote 26 in which it is claimed that there is no objective, observer-independent basis for selecting among causal factors at the level of causal structure, or for selecting the relevant contexts (interests, etc.) of inquiry, but that once a context has been fixed, there is an objective method of selecting certain factors as ‘causes.’ This latter type of objectivity is fixed through a process of social consensus and coordinated practice, which we might refer to as ‘conventional morality,’ because it is contingently settled upon for practical reasons.Footnote 27 Conventional morality is to be distinguished from ‘critical morality,’Footnote 28 which is objective in the former, stronger sense of designating what really ought to be the case for moral and political rather than purely pragmatic reasons. In general, then, it is widely accepted that we must rely on norms of conventional morality to select causes in a principled manner. This reliance is especially pronounced when the objects of inquiry are themselves the products of conventional morality, i.e., when they are social objects—groups, institutions, practices—of the sort well-studied by social ontologists.Footnote 29
In the interest of most clearly presenting my case, I will simply adopt a recent manipulationist theory of causal selection proposed by Maria Kronfeldner which smoothly unifies abnormality, manipulability, and morality.Footnote 30 According to Kronfeldner, we select as causes the factors we are willing and prepared to change. She distinguishes between what she calls ‘conservative’ vs. ‘forward’ control: conservative control is the ability to maintain some state of affairs and prevent changes, while forward control is the ability to manipulate or intervene on something in order to produce desired effects. Her proposal is that factors over which we exert only conservative control—because we want them to remain as they are, or because we do not care either way—are backgrounded as enabling conditions, while those over which we are willing and prepared to exert forward control are selected as causes. This distinction allows Kronfeldner to fold in the abnormality approach, since the willingness to consider some state of affairs ‘normal’ represents one kind of control, namely conservative control.Footnote 31 But Kronfeldner’s account also accommodates the morality approach, because it is often moral considerations in the critical sense that determine whether we are willing and prepared to control something.Footnote 32 It is the fact that we are conventionally (and critically) more willing and prepared to change children’s behaviour than the presence of oxygen in the air that makes the former a cause and the latter a mere enabling condition.
Kronfeldner’s framework thus allows us to integrate the insights of philosophers who have provided moral and political answers to the question of which factors we are willing and prepared to change, and which ones we are not. For once we admit the role of conventional morality in causal selection, we must admit that those conventional norms are in turn open to the influence of critical morality.Footnote 33 Joel Feinberg, for instance, argues that legal responsibility is “something to be decided, not discovered” because the question of whether some causal contribution was sufficiently important to some harmful outcome—as compared to all the other causal factors that also contributed—depends on the purposes of the law and the policies it intends to promote.Footnote 34 Arthur Ripstein, moreover, argues that the cause of some harmful outcome can only be selected by reference to some prior standard of care that determines what a person was responsible for doing; even if a person could have prevented some harm, she is not the cause of that harm unless she had a responsibility to do so.Footnote 35
Marion Smiley provides a particularly rich account of this relationship between causal and moral responsibility: since causal explanations require identifying who had the responsibility for preventing some harmful outcome, they therefore depend on social expectations about (1) who possesses that power of prevention, along with (2) how much we could expect them, in both the predictive and normative senses, to give up of their own interests to do so.Footnote 36 Causal explanations of social phenomena depend, as we have already seen, on norms of conventional morality that structure social groups, institutions, and practices. In Kronfeldner’s terminology, these are questions about who has (1) the ability and (2) the willingness to exercise control over some outcome—precisely those questions that determine our judgements of causal selection. But Smiley’s account emphasizes the additional role of critical morality here, for we can ask not only who actually possesses this ability and willingness, but also (1) who ought to possess this ability and (2) whether they ought to be willing to change. Our answers to these questions—which are not fixed, but subject to ongoing contestation—depend irreducibly on our views about moral responsibility, that is to say, on our substantive conceptions of social roles, relationships of power, and membership in a moral community. That women were once (and still are by some) expected per traditional gender roles to be responsible for taking care of the home, for instance, leads them to be viewed by some as the cause of men’s unemployment, rather than worldwide trends toward globalization and deindustrialization.Footnote 37 Since wealthy people are powerful, they are able to shape both legislative policies and the cultural imagination in ways that generate very little expectation that they will or should give up their wealth to fund, say, drug rehabilitation centres; we neither predict nor morally demand that they do so.Footnote 38 And because many people in the U.S. during the 1980s did not consider black South Africans to be part of their moral community, they did not feel any responsibility to participate in anti-apartheid boycotts.Footnote 39 It is for such reasons of conventional morality, Smiley contends, and not in virtue of some deeper metaphysical fact, that it sounds odd to say ‘Wealthy people cause drug abuse,’ or ‘American investors caused apartheid.’ But once we scrutinize the contingent norms of conventional morality by the light of critical morality, that can change—as I shall demonstrate in the next section.
To sum up here, theories of causal selection show us that causal explanation is itself morally conditioned: our selection of certain factors rather than others as the causes of some phenomena depends on what we are morally willing and prepared to control.
III. The Politics of Explaining Poverty
I have argued thus far that choices of causal explanations are socially conditioned, because causal explanation is itself morally conditioned. To illustrate, I return now to the case of poverty. I argue that disputes over situationist vs. individualist explanations of poverty persist so long as there are normative disagreements about what we are morally willing and prepared to control, and I defuse some potential objections to this account.
Rebecca Blank has reviewed in detail six theoretical frameworks implicit in causal explanations of poverty given by economists, five of which I consider here.Footnote 40 Many of these are situationist in orientation: economists theorize that poverty is caused by 1) non-market social forces such as racial discrimination,Footnote 41 2) lack of economic growth and functioning markets,Footnote 42 or alternatively 3) that it is markets (and capitalism more generally) themselves that cause poverty.Footnote 43 Other theories rely on individualist explanations, taking poverty to be caused by 4) individual dispositions and behaviours. These theories posit the existence of distinctive traits or a ‘culture of poverty’ among the poor: deviant values and behaviours such as a weak work ethic, pregnancy out of wedlock, and substance addiction.Footnote 44 Finally, some theories posit that poverty is itself caused by 5) efforts to alleviate poverty. This last framework comprises both a situationist version, according to which foreign aid disrupts local markets,Footnote 45 and the more common individualist version,Footnote 46 according to which poverty results from poor individual choices (e.g., to have more children, to work less hard) incentivized by social benefits.
Blank argues that the “continuous and divisive” nature of policy debates over poverty result from differing causal explanations that entail very different—often incompatible—policy interventions, e.g., expanding vs. limiting markets, providing vs. withholding assistance to the poor.Footnote 47 The problem, she points out, is that all the frameworks described are supported by some empirical evidence. What is important for my purposes is Blank’s assessment of the state of the debate, which cannot advance simply by procuring more data. Blank writes:
Academics often suggest that such debates can be solved by reference to the facts and the research results. While accumulated research knowledge can indeed shape the policy discussion, I am not very optimistic that research alone will solve some of these debates … [E]ven when facts are widely agreed upon, interpretation can vary … Individuals in the policy debate typically bring value assumptions and judgments to their reading of the research.Footnote 48
She goes on to identify the two “intellectual and value based conflicts [that] have been particularly divisive.”Footnote 49 The first is the efficacy and desirability of markets, and the second is efficacy and desirability of social benefits.Footnote 50 My point is that this is precisely what we would expect from Kronfeldner’s and Smiley’s analyses of causal selection: seemingly empirical disagreements about the causes of poverty actually derive from normative disagreement over the right moral values and political commitments to adopt.
Whether social benefits actually cause further poverty, for instance—that is, whether such benefits generate incentives for (bad) individuals to make bad choices—depends in part on moral beliefs about the extent and appropriateness of the government’s role in maintaining the well-being of its people. If we do not expect that the government has a significant role to play in people’s lives—if we are not prepared to use it to exert forward control on these outcomes—then we will not be inclined to say that punitive or insufficiently funded welfare services are a cause of poverty, even if those same policies form part of the conditions allowing people to fall into and remain in poverty. (One reason for rejecting this role of government would be if we subscribe to an ethos of ‘personal responsibility,’ which Iris Marion Young describes as the idea that “each must self-sufficiently bear the costs of its choices and has no moral right to expect help from others.”Footnote 51 I will return to this idea in the next section.) Mutatis mutandis for the disagreement about markets: whether one takes the absence or presence of markets to be the cause of poverty depends on one’s political beliefs as to how large a role the market (as opposed to the state) ought to play in people’s lives.
The upshot, then, is that, insofar as causal explanation is morally and politically conditioned, philosophers have a larger part to play in causal explanations of social inequality than is typically acknowledged. Fleshing out this claim will be my task in the next section. Before I leave this one, however, let me clarify these claims by considering a few potential objections.
It might be objected, for instance, that questions of causal selection—like those of causal structure—are also questions of metaphysics, and hence not interest-relative or pragmatically bound in the way that explanation is. Kenneth Waters, for instance, has argued that judgements of causal selection depend on certain metaphysical properties that are not interest-relative. For example, he claims, DNA is properly selected as the cause of biological development because it is an actual and not merely a potential difference maker—meaning that in addition to exhibiting the right counterfactual dependence relation with respect to some phenotypical outcome, DNA differs in reality across a population while other potential difference makers (other aspects of RNA synthesis, environmental factors, etc.) do not. In the cases of social inequality with which I am concerned, however, both individualist and situationist factors do differ across actual populations. There is real variation in character and personality traits, as well as in levels of marketization, labour protections, discrimination, and other social structural processes across affluent and poor groups, so these may all serve as actual difference makers for poverty.
Waters further argues that DNA is selected because of its causal specificity: slight differences in DNA will alter biological outcomes in varied and specific ways (while environmental factors, say, may not). But as James Woodward (Reference Woodward2010) argues in reply, the reason that a property such as causal specificity is used for causal selection is precisely because it is a property that conduces greater control and manipulability in ways that further biologists’ interests. In other words, insofar as causes can be selected on the basis of certain metaphysical properties, those metaphysical properties are themselves selected for reasons of practical interest (which in the case of social inequality is subject to moral and political debate). Finally, it should be noted again that a key dissimilarity between DNA and social phenomena is that in the latter case the very metaphysics itself is already normatively laden with conventional morality. The fact that such interest-relative, pragmatically bound factors are already built into the social ontology of phenomena such as poverty, then, opens the door for critical morality to play the role I have indicated.Footnote 52
It might also be objected that individualist and situationist factors are not competing explanations, but merely emphasize different micro and macro levels of the same phenomenon. This is certainly true as far as it goes;Footnote 53 as I stated earlier, both individualist and situationist factors enter into the actual causal structure of poverty. But, again, my concern here—and what is at play in the rancorous public debate over the causes of poverty—is causal selection. Because judgements of causal selection entail normative commitments to certain interventions and not others, people who advocate incompatible interventions (i.e., expanding vs. limiting government services) will be committed to competing causal explanations. Indeed, the mere suggestion of certain causal explanations can be read by others as deeply offensive, precisely because of the normative commitments they imply. To wonder whether female survivors of rape caused the violence inflicted upon them, for instance, suggests that survivors rather than perpetrators are the ones who ought to change their behaviours, i.e., that forward control ought to be exercised over survivors rather than perpetrators. But this implication that women should simply accept limitations on their freedom of movement, dress, etc. is inconsistent with powerful and widely accepted ideals of gender equality. That is why such a suggestion—especially coming from those who claim also to endorse such egalitarian ideals—is experienced as so repugnant (‘How dare you suggest that she brought it on herself?’).Footnote 54 Thus, while different causal explanations need not conflict if they support mutually compatible interventions, there are other cases in which causal explanations really do compete, i.e., when they entail incompatible interventions. Such is usually the case, I claim, with individualist vs. situationist explanations of social inequality.
Finally, it might be objected that, so long as we get the metaphysics right—i.e., that we have the correct causal structure—there really is nothing further to be said for causal selection and hence no basis for criticizing individualist or situationist explanations of social inequality. While this is true enough from the perspective of metaphysics, I think there is more to say from the ethical perspective. In particular, the social psychological evidence that I cited in Section I should give us pause. The mere fact that different researchers bring diverse background assumptions, priorities, and questions to bear on inquiry, i.e., that they are socialized differently, is an unavoidable fact—and not undesirable, since it enhances our ability to weed out biases and arrive at the truth. But we have reason for concern if, as the data suggest, there are systematic patterns in the way that the moral expectations behind different causal explanations map onto systems of advantage and disadvantage within a society. I take up this thought in the following section concerning the role of moral philosophy.
IV. How (Moral) Philosophers Can Help
If disputes over causal explanation ultimately depend on normative disagreements, it follows that philosophers working on those normative issues can fruitfully contribute to the debate. I should say up front that this is a modal or conceptual ‘can,’ not necessarily an attribution of present ability. My point is that the very logic of causal inquiry allots such a role to philosophers—not that it is easy or straightforward to take it up. But let me offer a proof of concept, as it were, by suggesting two ways in which philosophers might take up this role.
First, philosophers can identify the background assumptions operating behind empirical inquiry, especially when these are moral or political in nature. Lisa Gannett argues, for instance, that scientific attention to genetic factors as the causes of disease—as opposed to wider environmental factors contributing equally to disease—is driven in part by the political priorities of governments motivated to reduce spending: “Genetically engineered solutions make private investors money; serious attempts to counter poverty, environmental degradation, and tobacco, alcohol, and drug addiction just cost taxpayers money.”Footnote 55 Similarly, Philip Kitcher suggests that individuals who are themselves well-off may be more concerned to intervene on causal factors that threaten them, such as genes, than causal factors that do not, such as poverty.Footnote 56 And Kronfeldner details a historical shift in causal explanations of cancer in 1980s and 1990s: where once cancer was widely thought to be caused by carcinogen-producing industries, it has become increasingly common to view cancer as caused by individuals’ lifestyle choices, e.g., failure to eat organic vegetables and consume vitamins. In short, researchers are more willing and prepared to exert forward control over factors like genes and individual behaviour because they are correspondingly unprepared to advocate the far-reaching changes needed to address poverty and environmental degradation—the difference is political, not metaphysical. Specifically, we can trace the difference to an assumption that responsibility for well-being lies with the individual and not with society.Footnote 57 Indeed, this is the very same assumption of ‘personal responsibility’ that, according to Young, motivates individualistic explanations of poverty:
[T]he rhetoric of personal responsibility encourages an isolated, atomistic way of thinking about individuals. What it means to be responsible is for a person to maintain control over his or her actions and their consequences, and to make sure that they and only they bear their costs. A capable, responsible person does not depend on others … There is, then, a kind of tautology in the account of poverty that this rhetoric produces. These people are poor because they are not personally responsible, and the evidence of their lack of personal responsibility is that they depend on public assistance.Footnote 58
Philosophers are thus well equipped to excavate the different metaphysical and moral presuppositions—in other words, different norms of conventional morality operant in different sub-communities of researchers—that underlie competing causal explanations.
Moreover, by doing and disseminating philosophical theory, philosophers can aid social scientists in better understanding their own non-empirical commitments. A real-life example of this kind of philosophical facilitation is the Toolbox Project, a National Science Foundation funded initiative run by academic philosophers.Footnote 59 Because scientific researchers are socialized into disciplinary paradigms carrying philosophical commitments that become so ingrained as to be unnoticeable to researchers themselves, cross-disciplinary teams are often frustrated by disagreements arising from conflicting presuppositions. The Toolbox Project surfaces these presuppositions by having research teams fill out a questionnaire gauging their epistemological, metaphysical, and normative commitments. Through structured dialogue facilitated by philosophers, participants are able to utilize a shared vocabulary of philosophical concepts with which they can better recognize and negotiate—even if they continue to disagree—one another’s philosophical assumptions.
To take another example, social psychologists have long drawn on various philosophical traditions of thought to better understand different conceptions of the self that are central to psychological research. Pioneering feminist psychologists such as Carol Gilligan and Jean Baker Miller used feminist theory to challenge the centrality of the (masculine) autonomous self in prevailing research frameworks, arguing that women’s experiences in a patriarchal society led them down a different—but not inferior—path of psychological development.Footnote 60 Edward Sampson has drawn extensively on political philosophers’ critiques of liberal individualism to issue a call for the discipline of psychology to adopt a new post-modern conception of the self appropriate for understanding psychological processes in a globalized world.Footnote 61 And cultural psychologists have linked different traditions of ancient metaphysical and epistemological thought to conceptions of the independent vs. interdependent self—which, recall, is one factor in varying levels of correspondence bias in North American/European and East Asian countries).Footnote 62 Importantly, such normatively inflected psychological research has been taken up by law and policy researchers seeking to address the ideological stalemate between liberals and conservatives in the U.S. across a wide range of issues, such as poverty, racial inequality, and criminal justice.Footnote 63 Philosophers, then, are not total outsiders to science, but enablers of the background presuppositions that make it possible.
Second, however, moral and political philosophers—just by doing their work of generating normative arguments—can help not just to diagnose but to break these stalemates. In other words, they can advocate specific norms of critical morality in order to challenge existing norms of conventional morality. I thus propose that we view the work of moral philosophers as aimed at least in part at changing the normative commitments that underlie our causal understandings of social problems. Against a certain background of moral expectations—for example, that wealthy people or wealthy countries are fully entitled to as much as they can accumulate from the market—we would not be willing and prepared to intervene on them, and hence we would not cite them as the causes of poverty. But against a different background of moral expectations, according to which markets as currently realized are unfairly biased towards some at the expense of others, then we would be willing and prepared to intervene on them. It would no longer be implausible to say that ‘Markets cause poverty’ or ‘Wealthy countries cause poverty’ when they flout these moral expectations, e.g., by refusing to reform lopsided international trade agreements that benefit wealthier nations while disadvantaging poorer nations. The assumptions structuring these different sets of moral expectations lie squarely within the province of moral and political philosophers. It is not, however, limited to professional philosophers. The moral philosophical work of changing normative expectations is also performed by activists, community organizers, and politicians; in schools, churches, unions; and so on.
I think there is reason to think that something like this has indeed occurred in the debate on poverty. Across the spectrum of empirical disciplines that investigate causes of poverty—economics, psychology, sociology, political science, and anthropology—there has been a historical evolution from individualist to more situationist explanations.Footnote 64 In the 1960s and 1970s, for instance, psychological theories posited that individuals’ low levels of intellectual ability, motivation to achieve, or ability to delay gratification caused them to be poor; the field has subsequently undergone a shift toward theories emphasizing environmental influences, such as the detrimental impact of low socioeconomic status on cognitive development.Footnote 65 During the same period, anthropological and sociological theories citing cultures of poverty, i.e., the intergenerational transmission of deviant values such as promiscuity and dependency, largely gave way to theories portraying such values as contingent, adaptive strategies for coping with severely adverse conditions.Footnote 66 I find it plausible to think that it was partly a matter of moral progress, e.g., the successes of the civil rights and feminist movements in transforming acceptable moral views (as well as the demographic makeup of the scientific community), that drove the shift from individualist to situationist explanations in these fields. In other words, as more people adopted egalitarian commitments to reduce racial and ethnic prejudice, causal explanations of poverty shifted away from stigmatizing theories that painted these social groups as essentially different.
By bringing to light the moral stakes involved in competing causal explanations of poverty, then, we are in a better position to make arguments that penetrate to the real problem: in other words, we might give moral and political arguments for opting for one type of causal explanation over the other. Taking seriously the social psychological analysis of individualist vs. situationist causal explanations that I have given—in particular, preferences for the former amongst the relatively advantaged and for the latter amongst the relatively disadvantaged—forces us to see that systematic variations in the correspondence bias serve competing interests.Footnote 67 It is, in effect, to politicize the debate: to expose the political stakes beneath the veneer of ‘neutrality’ associated with a naïve picture of empirically minded social scientists analyzing pure facts. In a fascinating study, researchers took advantage of a natural experiment in Buenos Aires, wherein the state granted property rights to a subset of a population of urban squatters.Footnote 68 There was no way to determine ex ante which squatter households would receive land titles (since this depended on whether the original owners would be willing to accept state compensation), and those who did and did not receive titles were otherwise essentially identical in their social features. Ex post, however, those who received land titles were significantly more likely to express individualist beliefs (i.e., that it is possible to be financially successful without the support of a large group, and that those who work harder end up better off than those who don’t). Because the newly minted property owners now had different interests compared to their squatter neighbours, they adopted beliefs that would rationalize certain moral views (e.g., the justifiability of markets and private property schemes) along with individualist causal explanations of poverty. Forced into the open once these political aims and commitments have been surfaced, debates over causal explanations of poverty look very different.
To be sure, moving to the moral and political battlefield does not automatically settle the question of which causal explanations to adopt, because there is no pre-existing agreement amongst philosophers over the correct moral and political principles. In the current philosophical debate over effective altruism, for instance, disagreements over which interventions are most effective—amassing wealth for charitable redistribution within the existing economic order vs. wholesale overhaul of the existing economic order—serve as the exact mirror image of the causal debate, with the same mixture of empirical and normative assumptions. To take another example: luck egalitarians who accept inequalities grounded in individual choice are likely to be sympathetic to individualist causal explanations of poverty, while social egalitarians who view inequality in terms of relations of domination are likely to prefer situationist explanations. In general, there is no uniform agreement over the correct moral and political principles governing what people ought to be able and willing to control (which in turn govern our causal explanations). Furthermore, the discipline of academic philosophy is itself undoubtedly subject to the same competing interests and hence biases that I have described; the sheer lack of demographic diversity amongst practitioners ensures as much. In the absence of having already secured the correct moral and political theory, then, are we doomed to a ‘post-factual’ world in which there is no role for truth, but only brute contests of ideology? And can philosophers do any better?Footnote 69
These are serious challenges, beyond the scope of what I can address here. However, they are also the facts of life which no theory can do away with.Footnote 70 In response, let me make two points. First, even though philosophers are undoubtedly vulnerable to biases (and in this sense no different from social scientists and everyone else), it remains the case that we are trained to study and evaluate normative arguments, while others are not. It is only through doing some kind of philosophy that we can hope to establish standards for adjudicating between competing normative claims and hence causal explanations. In spite of our limitations, then, we can expect to contribute a degree of rigor and familiarity with normative thought that empirical researchers normally do not possess. Second, my own view is that we would do better to face up to the normatively inflected nature of fact finding and fact presentation than to pretend that things can be otherwise. In other words, it is better for moral values and political commitments to be worn openly on the sleeve, such that different interpretations of evidence (as presented, say, in the media) may be assessed critically against each other,Footnote 71 rather than smuggled in under the guise of an unattainable objectivity where they produce intractable causal debates. Once the ubiquity of these values is recognized, it is clear that philosophical analysis enabling us to better navigate the interrelationship between empirical and normative inquiry, such as I have put forth in this paper, places us in a better position to identify what is truly at stake.Footnote 72
To show my cards (at last), here is how ‘breaking the stalemate’ might go for causal explanations of poverty. We might argue on grounds of justice, not empirics, that, where there is evidence for both sides, we should throw our weight behind the poor and disadvantaged, lest we run the risk of committing what Miranda Fricker calls ‘testimonial injustice,’ in which the claims of a marginalized group are prejudicially denied credibility.Footnote 73 The World Bank’s 2015 World Development Report, for instance, describes the ways in which poverty experts—who have very different backgrounds from those whom they work to aid—exhibit cognitive biases that prevent them from understanding how conditions of poverty shape the mindsets of the poor; the poor thus “face additional hurdles to getting their voices heard, their concerns heeded, and their aspirations realized.”Footnote 74 Since the poor are likely to be subject to testimonial injustice, e.g., the prejudiced notion that they are ignorant or otherwise incapable of advancing knowledge claims about poverty, we should counterbalance this by deliberately paying more attention to their perspectives.Footnote 75
Or, justice might demand that we back the poor on the grounds that they deserve to be, as Monique Deveaux argues, ‘agents of justice.’Footnote 76 That is, the poor ought to take the lead in poverty alleviation initiatives—irrespective of expert opinion on their efficacy—on the grounds that such policies are addressed at their plight, and they deserve morally to exercise their agency and self-determination. This might mean, for instance, taking more seriously the claims of groups within the World Social Forum (WSF), a longstanding mass annual meeting led by organizers from the global South that defines itself as an ‘alter-globalization’ movement or “globalization from below” in opposition to the first-world, market-friendly business and political leaders that meet in the World Economic Forum (WEF).Footnote 77 One of the core claims of the WSF, against the aggressive market solutions advocated by the WEF, is that market-driven globalization is itself the cause of poverty.Footnote 78 Again, the point is that we can make moral and political arguments for siding with one type of causal explanation over the other, when debates persist despite ample accumulation of empirical evidence.Footnote 79
Finally, alongside these moral arguments, philosophers can provide further epistemic arguments that go beyond the reach of empirical data. Feminist, anti-racist, and Marxist theorists have long contended that the epistemological standpoints available from certain social locations may be epistemically superior to others because of their potential for better (e.g., first-hand) access to the mechanisms by which social inequality is maintained.Footnote 80 This would again suggest that we should prioritize the causal explanations and interventions preferred from the standpoints of (at least some of)Footnote 81 the poor.
If what I have argued heretofore is correct, then such moral-epistemic considerations are what we really need to adjudicate between conflicting causal explanations, not just more data. I suspect that if moral philosophers paid better attention to empirical work and understood its normative underpinnings, they might well shift away from advocating duties of individual charity by the privileged (premised on individualistic explanations) rather than duties to collectively organize for structural change (premised on situationist explanations). I think that philosophers’ routine use of ‘donating to NGOs’ as the prime example of ‘morally required action for addressing poverty’ falls wide of the mark precisely because it simply assumes that charity is effective (and that it is for social scientists to determine whether it really is). Doing so obscures the moral and political assumptions that also underlie such an individualist approach. By highlighting the role of normative commitments in causal explanation, then, I hope to thereby encourage philosophers to appreciate that issues of causality set aside as ‘just an empirical question’ do deserve to be scrutinized through the lens of moral and political philosophy.Footnote 82 We cannot so neatly divide the labour between philosophers and social scientists. Moreover, we should also critically self-reflect on what empirical findings (and disputes over themFootnote 83) suggest about how inquiry proceeds in our own discipline. The more we overcome social inequalities within our own discipline, the more we will be able to contribute to overcoming social inequalities more generally. These forms of empirical engagement are continuous with and required for responsibly performing our wider moral work.
V. Conclusion
I have argued that choices of causal explanation are socially conditioned, and that this is because causal explanation is itself morally conditioned by the necessity of causal selection. From this it follows that moral and political philosophers have a role to play in changing the expectations that underlie what people are willing and prepared to change, and hence which causal explanations and interventions they adopt. We perform that role precisely by doing the work of theorizing and pushing for moral progress.
To conclude, let me offer a few words to further elucidate the conception of philosophy I am advocating here. The value of philosophy, I suggest, does not lie primarily in the ability to deliver up results, that is, permanent moral truths or verdicts; rather, it consists in the process of unearthing, analyzing, and evaluating the deep assumptions that make inquiry possible.Footnote 84 As such, this means that philosophy should not make its contribution—shifting normative expectations and hence causal explanations—as a mere body of texts, but also as a living community of scholars in dialogue with different publics. A growing body of literature in social and political psychology suggests that pure argumentation and facts do not change minds.Footnote 85 This is not to say that we should abandon reasoned arguments, but that we must consider the wider social and psychological conditions that act as barriers or facilitators to the uptake of arguments. Building up dialogue-enabling social relationships takes time and is not easy, but is worth striving for. This dialogical model might also allow us to see ourselves in a new light, more akin to those in the empirical sciences: as offering up tentative and timely moral hypotheses, open to further testing in light of the way they play out in the world. I hasten to add that this way of conceiving of ourselves and our role is not intended to supplant more traditional forms of philosophical work, but to empower it. By doing philosophy—that is, by rigorously examining the assumptions that structure our investigations of the world—we occupy a unique and enduring role in improving it.
Acknowledgements:
I am grateful for feedback from audiences at the Social Impact of Philosophy Conference at la Universidad Iberoamericana, 2013 Princeton-Michigan Graduate Meta-Ethics Conference, 2014 FEMMSS and CSWIP Joint Meeting, the Social Equality Conference at the University of Cape Town, 2014 PSA Biennial Meeting, 2014 APA Eastern Division Meeting, the Ethics and Explanation Conference at the University of Nottingham, the Workshop on Philosophy and Poverty at the University of Salzburg, and the 2016 BSPS Annual Conference. I would also like to thank Elizabeth Anderson, Dmitri Gallow, Keith Horton, Simeon Newman, Laura Ruetsche, and my colleagues at Yale-NUS College for their comments on earlier versions of this paper.