SOCIAL SCIENCE AND ‘ NORMATIVE FACTS ’ What ’ s the big deal ? *

Stephen Turner’s Explaining the Normative is probably a puzzling work for the average empirical social scientist. Turner devotes more than two hundred pages and considerable argumentative efforts to build a compelling case against what he calls ‘normativism’. His main thesis is that what ‘normativist’ philosophers call ‘normative facts’ are not special or transcendental in any meaningful sense, but can be fully grasped by social science standard explanations and even by naturalistic ones. I think most social scientists would agree with this thesis, and would in fact consider it as a trivial truth. However, Turner goes further to suggest that ‘normative talk’ and notions as ‘normative correctness’ are just bogus talk or simply the folk way we use to describe some of our cultural practices, but have no explanatory relevance or rational meaning. I think that far fewer social scientists would buy this second claim. In all fairness, it should be acknowledged that Turner’s book is written having in mind a philosophical audience (or perhaps a ‘social theory’ audience) rather than a socialscientific or sociological one. But since the invitation for commenting on his book comes from a sociology journal, I will try to present what in my opinion would be a very usual social-scientific look on the issues Turner is concerned about. I advance that I fully share the general spirit of Turner’s criticism of ‘normativism’, as well as his sympathy for the present naturalistic and anti-metaphysical trends in social science and philosophy. However, I would like to point to some apparent disagreements with the kind of social science I am engaged with.

Stephen Turner's Explaining the Normative is probably a puzzling work for the average empirical social scientist.Turner devotes more than two hundred pages and considerable argumentative efforts to build a compelling case against what he calls 'normativism'.His main thesis is that what 'normativist' philosophers call 'normative facts' are not special or transcendental in any meaningful sense, but can be fully grasped by social science standard explanations and even by naturalistic ones.I think most social scientists would agree with this thesis, and would in fact consider it as a trivial truth.However, Turner goes further to suggest that 'normative talk' and notions as 'normative correctness' are just bogus talk or simply the folk way we use to describe some of our cultural practices, but have no explanatory relevance or rational meaning.I think that far fewer social scientists would buy this second claim.
In all fairness, it should be acknowledged that Turner's book is written having in mind a philosophical audience (or perhaps a 'social theory' audience) rather than a socialscientific or sociological one.But since the invitation for commenting on his book comes from a sociology journal, I will try to present what in my opinion would be a very usual social-scientific look on the issues Turner is concerned about.I advance that I fully share the general spirit of Turner's criticism of 'normativism', as well as his sympathy for the present naturalistic and anti-metaphysical trends in social science and philosophy.However, I would like to point to some apparent disagreements with the kind of social science I am engaged with.

Four views on 'normative facts'
The book's playing field is inhabited by two players: on one side we have the 'normativists', who claim that 'normative facts' are inescapable; that they involve such concepts as 'correctness' or 'validity', which are 'conditions of possibility' of our social life; that these facts involve a 'surplus value' in front of standard social-scientific descriptions, and that therefore they are outside of the ordinary stream of causal explanation.On the other side, we have the 'antinormativists', who think that those facts, if they really exist as such, are part of the causal structure of the empirical world, are to be sociologically described as local beliefs, and do not have any special transcendental quality or 'validity'.Turner clearly takes side with the latter.
However, two more players also contend to win the game.According to the third one, Turner is right that the normativist claim to immunize normative facts from causal explanations is untenable.But at the same time, he thinks that Turner goes too far when he infers from this that all normative talk is mysticism and metaphysical circular thinking, and when he suggests that we could perfectly eliminate normative facts from our explanatory accounts of social reality.The third player (who I think incarnates the present advanced social science's view) thinks, to say it in Turner's words, that the 'surplus value' in normative facts may be real (though ontologically subjective, as most social realities) and that normative facts should and can be integrated into our scientific explanations of social actions and social reality.
Figure 1 may clarify this, by showing that there are two different questions merged in Turner's characterization of the debate, and that to disentangle them may throw more light on it.The debate between Turner (lower left-cell) and the 'normativists' (upper-right Can 'normative facts' be 'naturalized' or integrated in the ordinary stream of causal explanation?

Four different views in the debate on 'normative facts'
cell) is the topic of the book.However, as my classification suggests, I believe most present and historical social science trends are simply ignored in that debate, particularly those situated in the upper-left cell (I will not refer here to those in the lower-right cell for reasons of space).

Turner's 'social science'
A first puzzling fact is that in Explaining the Normative, social science is often simply identified with some selected 'classics' plus some particular trends in relativistic anthropology and sociology so that the 'conflict between normativism and social science' suspiciously looks very much like the classical 'rationality vs. relativism' debate which took place from the sixties to the eighties of the past century.This is certainly a very limited and outdated conception of 'social science'.It is also surprising that Turner tends to identify descriptivist ethnography with 'causal explanation', when the first has traditionally denied the very possibility of the second.It is curious, too, that in a book entitled Explaining the Normative it is hard to find one single explanation of any normative empirical social fact, while contemporary social science and evolutionary theory are full of examples (see, for example, Bicchieri 2006;Binmore 2005;Boudon 2001Boudon , 2003;;Camerer and Fehr 2004;Elster 2007Elster , 2009;;Esser 2009;Fehr and Gintis 2007;Gigerenzer and Sturm 2012;Hedström & Bearman 2009;Hechter and Opp 2001;Kuran 1995;Petersen et al. 2012).
If one takes a closer look at contemporary analytical sociology, behavioural and experimental social science, heuristics and biases theory, behavioural rational choice and collective action theories, computational sociology and economics, evolutionary game theory and psychology, social norms theory, research on prosocial motivations and social justice perceptions, etc. (not to mention contemporary political science, social stratification studies, institutional design and institutional economics, biological anthropology, or many other sub-disciplines and fields), I suspect that some of the issues disputed in the book would hardly survive as more than historical, scholastic debates.So my criticism, in a nutshell, would be: too much hermeneutic social theory and ethnography from the 1970s, too much 'armchair' philosophy, and too little contemporary social science and applied normative theory.

The social-scientific explanation of 'normative facts'
The empirical social scientist who is committed to usual scientific methods and procedures would get over the discussions contained in Turner's book in a rather pragmatic way.Many fertile and informative works in contemporary social science simply take for granted some solutions (or dissolutions) of those discussions, so in Marxian words they exemplify that 'the proof of the cake is to eat it'.The usual (at least for me) social-scientific view would go like this: social norms, conventions, normative beliefs, ethical or political beliefs and attitudes, and the like, are considered as propositional mental states individuals may hold ('beliefs' and 'desires'), whose causes and effects may in principle be investigated RIS, VOL.71.Nº 1, ENERO-ABRIL, 191-225, 2013. ISSN: 0034-9712. DOI: 10.3989 empirically.As such, normative beliefs do not pose more epistemological or ontological problems than factual beliefs or other cognitive mental states. 1  Understood in this way, 'normative' may be a proper label for some type of propositional contents of mental states (or some or their properties).It is not clear to me how problematic it is that they are 'empirically inaccessible' (Turner 2010: 1): in a sense, all mental states are (just as some distant planets), but, in another, humans have abilities to infer and attribute to others mental states and their propositional contents with some degree of confidence.Moreover, psychology and the social sciences have notably refined different methods and sophisticated techniques to improve the epistemic quality of these kind of inferences and attributions (as astronomy has done in order to detect distant planets).Therefore, it is not true that 'normative facts' are 'not part of the ordinary stream of explanation' (íbid.), for their efficient causes and subsequent effects on behaviour and social institutions may be a matter of empirical scientific research.
Take, for example, Bicchieri's recent and widely discussed book The Grammar of Society (2006), where the most refined tools of contemporary sociology, behavioural economics, and social psychology are combined in order to produce a compelling definition and explanation of social norms as equilibria reached from mutual expectations and sanctions in interactions where some salient feature psychologically triggers a 'norm' frame.This is an empirically founded and theoretically well-constructed account of (some kinds of) normativity in social interaction, where, ontologically speaking, nothing more than propositional mental states and individual actions is needed.

The 'disconnection' between the 'causal' and the 'normative' worlds
The paradox is that, while Turner wants to criticize 'normativists', he nonetheless seems to accept one of their basic assumptions, the 'disconnection' between the 'causal' and the 'normative' worlds, that is to say, that normative facts are to be either eliminated or located out of the causal chains of the empirical world.If you think like that and you are scientifically minded you will probably be led to the conclusion (as Turner is) that there is nothing like 'the normative'.But why should we not take normative facts as a type of facts (or properties of some facts) in the causal world, just as any other?Then we would really be rejecting the normativists' assumption of the 'disconnection' and integrating normative facts into the causal world without eliminating them.
1 Unless we start from the assumption that 'mental states' are themselves problematic entities, but in that case all the 'normativity' debate is simply eliminated as senseless, along with any reference to intentions, beliefs and desires, be they 'normative' or not.Note that if we discard propositional mental states as legitimate entities in social science explanations, then we are inescapably led to some sort of behaviourism; and behaviourism was abandoned in social science and psychology for very good reasons.In previous work, Turner himself claims that the 'folk psychology' used in mainstream sociology is something more than a 'Good Bad Theory' (Turner 2002: 7-8).
If we accept (as I think we should) that the adjective 'normative' refers just to an intensional property of the contents of some mental states, and we also accept the standard desire-belief explanations of human action as scientifically legitimate, then there is no disconnection at all: some normative mental states may be causing some actions (and therefore, through complex composition processes, some macrosocial phenomena), and may themselves be caused by other previous events, whether they are other normative mental states, non-normative ones, or just non-mental events.This is the standard way of explanation in the social sciences, and in general terms I hardly see what is wrong with it (nor what alternative way of explanation is available).
Turner points out that normativists might reply: this is just a descriptivist account in which you use the adjective 'normative' in a purely taxonomic way, but the fact is that the 'normative validity' of some actions and beliefs might be an essential part of the explanation or the understanding of those phenomena, so this purely 'taxonomical' use of 'normative' won't do.I think that normativists are right when they point out that 'correctness' or 'validity' is an important issue (and I will comment later on how accepted sociological explanations like Weber's or rational choice theory's support this claim).But I agree with Turner that they are wrong when they imply that 'normative validity' is 'constitutive' in a somehow mysterious way, and not simply explanatory in a conventional causal sense.Let me explain.
From a scientific, explanatory point of view, it is not a problem that besides first-order normative mental states we also have second-order mental states on the 'validity' of the former (or of some actions caused by them).Whether we judge the concordance of actions or beliefs with a given rule, or we judge the 'validity' (logical, technical, moral, semantic, etc.) of the rule itself, 'normative' is still a predicate of mental states, which are in the stream of causal explanation.Boudon (2001) has shown how we can plausibly reconstruct causal chains that take into account different orders of normative beliefs.In fact, whether a normative prescription or belief is 'valid' or not is not always to be determined by the community or subjectively, but often can be judged in an objective way through contrasting with external brute facts or, to use Searle's term (1995), with institutional facts.For example, the validity of the normative claim that 'it is truth that Everest is the highest mountain on earth' is to be empirically determined, and if it is found to be a valid claim, this property may of course have a causal influence on some individuals' beliefs and actions.Similarly, the validity of the claim 'it is truth that I am married' is also to be empirically determined even if the relevant facts are social or institutional.Nowhere in these examples is there a break in the causal chain from objective facts to beliefs in the validity of claims and to relevant effects in individuals' beliefs and actions.So where is the 'disconnection' to be placed?Turner is right that there is not a 'normative account' of the world which runs parallel to the 'causal-empirical' description; but he does not fully acknowledge that some facts within that 'causal-empirical' description have normative properties which are causally relevant (through mental causation of actions).
The basic problem of normativism according to Turner is that, at some point, 'something normative has to come out from something nonnormative ' (2010: 192).But the trick here, as he argues, is to define 'normative' in such a way that fabricates the pro-RIS, VOL.71.Nº 1, ENERO-ABRIL, 191-225, 2013. ISSN: 0034-9712. DOI: 10.3989 blem itself in a tautological and circular way.If one defines 'normative' just as a specific property of the propositional content of some mental states, the problem is not bigger than the classic one of how mental states come out of something which is not mental (or, to go all the way through, of how something solid comes out of something non-solid, or how something organic comes out of inorganic matter, and so forth).In fact, why focus just on normativity?We could pose the same 'problem' for any property we could attribute to mental states: how does something 'cognitive' come out of something noncognitive?, how does something 'expressive' come out of something non-expressive?, how does the 'sense of beauty' come out of something senseless?, and so on.Defined as a property of the propositional content of some mental states, the problem disappears as a transcendental one and becomes a tractable, empirical question (which does not mean an easy one): how prescriptive propositional contents of mental states come out of non-prescriptive ones and non-mental facts.

The role of 'normative validity' in social science explanations
Turner is right that Weber already viewed values and norms as sociological facts.But he was not an ontological or epistemic relativist like other more recent sociologists or anthropologists.He gave a central role (and a normative role) to rationality in the explanation of actions, beliefs, and social institutions.This poses a problem for Turner's too simple narrative of a confrontation between Weber's sociological point of view and 'normativism': in his famous Zwischenbetrachtung, Weber explicitly wrote that rationality (and not only sociological factors such as interests, traditions, or social emotions) is inescapable in order to explain some social phenomena and actions: "The rational, in the sense of the logical or teleological 'consistency' of an intellectual-theoretical or practical-ethical stand, exerts and has always exerted power over men"; he even called this "the effect of ratio" (1915-19: 528).Weber's famous metaphor of ideational systems and worldviews as 'rail switchmen' (íbid.: 247) who impose their own 'logic' or 'legality' over sociological and economic facts is equivalent to saying that rational internal (normative!) relationships between ideas and beliefs may have their own causal efficacy in producing social phenomena.So was Weber, after all, a 'normativist' according to Turner's characterization?
The answer, in my opinion, is that the confrontation is partially made out of straw men.I think Weber correctly meant that it is perfectly possible to insert 'the normative' in the ordinary stream of causal explanation, while, at the same time, acknowledging that the 'validity' of some normative reasons is causing (in exactly the same sense of causality) some of the social actions and phenomena we want to explain.That is what Weber in fact applied in his writings when, for instance, he talked extensively about how the 'psychological and pragmatic implications of religions' (íbid.: 234) were often more important to explain behaviour than material interests or economic structures; or when he noted the consequences on behaviour of the conflict between a religion's 'rational aspirations ' and its 'irrational' components (íbid.: 506).This is a sound and common methodological strategy which has often allowed for explanatory success in social science.
For example, Elster's work on the 'civilizing force of hypocrisy' (Elster 1998) shows how impartial and egalitarian ideas are difficult to oppose publicly, and Harsanyi (1969) already showed years before how 'impartiality' is a 'normative' force in achieving social and strategic equilibria.Other well-known social phenomena such as 'pluralistic ignorance' (Bicchieri 2006;Kuran 1995), which explains intriguing facts about the prevalence of unpopular or inefficient social norms, could not even be formulated or detected if we forbade ourselves to describe individuals' beliefs about their peers' commitment to the norm as false but rational (because they rationally infer such a commitment from behaviour that is apparently consistent with the norm).The fact that the beliefs are false is an essential part of the explanation itself, but the social scientists who study pluralistic ignorance are not engaging in mysticism or metaphysical arguments about 'normativity' in any recognisable way.Under another description which ignores the issue of the 'validity' of the beliefs, the substantive causal explanation of the phenomenon (not any metaphysical and tautological 'surplus value') would be gone. 2f course we could ask what makes 'valid' beliefs produce some causal effects instead of others, as well as under which conditions 'valid' beliefs are more likely to be held (some branches of applied epistemology have given interesting answers to these issues).We can even try to give a naturalistic explanation of all that.For example, the classical 'money-pump' argument (Davidson et al. 1955) offers a plausible evolutionary explanation of transitivity in preference ordering: non-transitive individuals would be economically exploited by transitive ones and thus would go extinct.It is true that recent research on heuristics and biases shows that strict rationality is systematically violated by humans, probably for adaptive reasons (Kahneman and Tversky 2000;Gilovich et al. 2002).But the point is that even in these cases there is something non-arbitrary in terms of 'validity' that explains the adoption of some rules, beliefs, or decisions (for a good recent review of the issue, see Gigerenzer and Sturm 2012).So I would say that Weber's intuitions, but also Boudon's and Elster's, are confirmed by contemporary cognitive and evolutionary psychology: some types of ideas are psychologically more 'appealing' than others on evolutionary and adaptive grounds.All this evidence does not fit well with Turner's suggestion that explanations in terms of contextual and culturally acquired "dispositions and habits" do "all the explanatory work" (2010:110).

Decision theory and the rational choice approach as 'normative'
Take rational choice theory and decision theory in general: it is widely accepted that they may be used in a normative as well as in an explanatory way.One can use both the axioms of the theory to assess the rationality of a choice or action, and to explain it if one assumes some degree of rationality in the agent (which more often than is thought seems like a realistic assumption).In Elster's words, a rational action is defined as "behavior that is optimal from the point of view of the agent and performed, moreover, because it is perceived to be optimal" (2000:28, emphasis added).The same idea is expressed by Boudon and many others by saying that in a rational choice explanation the ends-means rationality of the action is its own explanation.But the way Turner (and perhaps the normativists) presents the discussion seems to involve that if the explanatory use is possible, then the normative use is futile or illusory, and the reverse.But why should this be like that at all? Users of rational choice theory do not have deep philosophical problems when they use the theory in both senses, but rather see them as fully compatible and often complementary.Once again, the proof of the cake is to eat it.
In fact, the postulates of rational choice theory are purely logical and mathematical inferences from a very simple set of assumptions.It is a formal theory, not a substantive one, as game theory is a mathematical theory which may find applications in very different fields.So if the 'normativity' of the logical rules or mathematical truths that form the axioms of the theory is an ontological or metaphysical problem, then the same applies to natural sciences, which have been using those tools extensively.The question for all the contenders in the normativist debate is then: why should normativity be a problem in the social sciences if it is not in natural or 'hard' ones?Turner tends to write about normativity in a social-moral sense or in a 'conventional' sense (semantic rules, social conventions, and the like), but not in an epistemic, logical, or mathematical sense.But why not?The problems with normativity, as he defines them, should be similar in both cases.

Searle to the rescue: Institutional facts vs. brute facts
In other words, the normativity of decision theory poses no less problem than the normativity of law.Why does Turner not focus on decision theory as a 'paradigm-case' of his discussion?My impression is that, if he had done so, he would have more rapidly and easily come to this conclusion: 'correctness' in decision theory has the meaning of whether an action or decision is causally adequate and efficient (or even optimal) to produce some result.So here is a kind of normativity which is explicitly and obviously linked to the causal structure of the world and as such to the 'ordinary stream of causal explanation', to adaptive and evolutionary selection explanations of their standards, etc., but which is not eliminable, since it is doing 'part of the explanatory work' in the study of human behaviour.The fact that 'legal norms' are not so obviously explainable in that way shows, first, that we are placing completely different things under the label of 'normativity', and second, that the ones that are not directly linked with brute facts (because they are 'institutional facts' or 'assignment of status functions' in Searle's terms) may ask for an historical or sociological explanation (however, as Searle convincingly shows, even in this sense normativity is not ontologically disconnected from the material world, since at the end of the status function's chain there is always a brute fact).
To disentangle both kinds of 'normativity' may throw some light on the debate: the first kind (the 'decision-theory-like' normativity) is very directly linked with the causal structure of the world (following the model of Searle's agentive functions), but the second (the 'legal-norms-like' normativity) is not, since legal norms are status functions.Some kinds of 'normativity' may then be 'universal' (because they are based on agentive functions), some others are not (because they are ontologically subjective status functions which may vary culturally in time and space).Applying logical inference rules, efficacy and efficiency technical rules, and evaluating congruence with given rules, are all normative operations that fall under the first kind, while most conventions, semantic rules, ethical, moral, or social norms, may fall under the second.I suggest, then, that Turner could make some use of Searle's sharp ontological distinctions to clarify the philosophical debates on 'normativity'.

Ontological vs. epistemic objectivity in normative facts
Searle's distinction between 'ontological objectivity' and 'epistemic objectivity' may also be useful for the discussion (but surprisingly is never mentioned by Turner).The distinction, for instance, is what makes the difference (denied by Turner) between a Maori's claim that 'hau' exists (Turner 2010: 61) and a normativist's claim that there are rational and valid claims about normative issues.'Hau' talk claims ontological objectivity, and in that sense is wrong.But most mainstream political philosophers make claims about the epistemic objectivity of normative facts.If they are institutional facts we are then justified in having a space of (valid or invalid) reasons about them, which is independent of the issue of how they are brought about causally ('independent' in that it is different business or 'game', not in an ontological-causal sense).This claim does not involve any kind of 'mysticism' or 'bogus' talk: the claim that a goal by Messi in a football match was to be declared null or not does not involve any metaphysical mystery; it is possible to determine the goal's validity objectively (for example by using cameras).And the question about its validity is independent from the causal analysis of how the rules of football historically emerged, of why Messi was playing better football today than in previous matches, or of why the referee declared the goal null.
The same distinction applies to other paradigmatic cases analyzed by Turner.For example, with regard to Kelsen's idea that the normativity of law is necessary to describe some legal-social facts (2010: 74s); Turner suggests this is equivalent to saying that the existence of God is necessary to properly describe a church service.But the analogy does not hold: again it is quite easy to see that in the latter case the God-believers are claiming ontological objectivity (God's existence as ontologically independent of the believers' beliefs), while in the former Kelsen (or whoever) is only claiming epistemic objectivity (whether a decision or rule is 'legally valid' is to be determined in an epistemically objective manner, though 'legal validity' and 'legal norms' are not ontologically independent of human beliefs).

Naturalization vs. elimination of normativity
In recent years, evolutionary theories, game theory, or the heuristic and biases approach have offered quite robust ways of advancing in the naturalization of human rationality and normative beliefs.If 'the conflict between normativism and social science', as Turner puts it, was such, then we should be seeing a strong quarrel between social scientists who rely on the idea of 'explanation by reasons' and the like (such as Elster, Boudon, analytical sociologists, etc.) and those who try to find evolutionary and naturalistic explanations for the origin of rationality and normativity.But what we see (even in Weber's historical openness to what we call today 'naturalistic explanations') is exactly the opposite: both groups work very closely, they are establishing growing links between them, they are using each other's work to strengthen their own, and they are often an active part in shared projects in both types of research.
How is this possible?My answer is again: because there is no necessary opposition between some type of 'normativism' and 'naturalistic' social science.In short, because the naturalization of 'normativity' does not mean its elimination.My impression is that most social scientists who actively try to build causal explanations of normative facts would at the same time agree with Hilary Putnam's claim that "The elimination of the normative is attempted mental suicide" (Putnam, 1982: 20).Causal reduction (and 'naturalization' is just a form of it) has to do with tracking causal links back to more and more elementary sources of the phenomena under study.Elimination has to do with encoding the same information in simpler ways, without informational loss.Clearly, normative beliefs may and should be causally reduced (it would be miraculous that they are not causally linked with the rest of facts in the world).But they will hardly be eliminated, since the information that normative explanations and language provides is not provided by statements about neuronal activity or interaction of elementary particles.It is not surprising, for the same applies to a Shakespeare play, to a digital photograph, to a jury's decision, or to a Coke: they are all examples of how the combinations of certain elementary entities in certain ways cause a more complex phenomenon which may in turn be taken, under a certain relevant description, as having explanatory power for the generation of other kinds of phenomena.
So why should we act as if there is a problem with all this?Biology was virtually reduced to chemistry, and chemistry to physics, but chemistry and biology have not been eliminated as scientific disciplines which build their own explanations (with the reasonable condition that they are consistent with what we know of their reduction base).To empirically explain normative beliefs in a scientifically rigorous way (whether the explanations are based on neurology, biology, sociology, cognitive sciences, or whatever) is perfectly compatible with keeping normative facts as meaningful, and even as part of some social scientific explanations.

Two legitimate tasks for normative talk
The proposed view does not throw social science in the arms of the elimination of the normative: first, because normative correctness may have an explanatory role when stu-dying some types of social behaviour; and second, because we can legitimately engage in rational and meaningful discussions on the validity and correctness of normative beliefs/theories according to several evaluative criteria.
I gave examples of the first, explanatory task in the preceding sections.An example of the second are contemporary political philosophy discussions on distributive justice since Rawls' A Theory of Justice (1971), which have provided lots of interesting results; for example, on how to apply distributive justice rules to particular cases, which are the acceptable distribuenda, what is equality and how it is to be measured, which trade-offs we face between different normative ideals, and so on.Most interestingly, they have done so in constant dialogue and mutual enrichment with sophisticated social science (for example, with normative economics, political science, institutional design theories, social stratification studies, or cognitive psychology).It was Rawls himself who suggested that ethical and political philosophy should draw on some regularities and facts about human behaviour discovered by the social sciences.
In short: even when empirical contrast is not directly possible, as is the case with moral or ethical beliefs, the task of determining their 'validity' may still be a rational and legitimate one, with its own methodology (like Rawlsian reflective equilibrium, for example), and without that task implying any disconnection with the empirical causal world at all.

'Normativism': weak and strong
Let me conclude: the best social-scientific work and the best normative theories we have available at present both agree that 'normativity' can and should be integrated in our empirical world without eliminating issues about 'normative validity', and even allowing for some explanatory role for them when studying human behaviour.Under this lens, it seems like the 'normativist challenge' (or, if one takes the other side, the 'naturalistic challenge') merits little concern by social scientists and applied normative theorists.Normativists' 'strong' thesis that 'normative facts' are some kind of Platonic 'essence' (Turner sometimes call this 'fundamentalism') is trivially false.However, a 'weak' normativist thesis (as I have described it here) is perfectly acceptable by a scientifically-minded worldview.
The real gains in our understanding of what we have called 'normative facts' will come, I have little doubt, from empirical research in a variety of interrelated disciplines.A Spanish philosopher of science, Manuel Sacristán (1983), used to say that before having a good philosophy of social science, we probably need to have lots of good social science.I completely agree, and I think we are about to see real progress on these issues in the years to come; progress that perhaps will make some of the philosophical discussions I have commented on here pointless.
Figure 1.Four different views in the debate on 'normative facts'