Reference CartwrightTowards the end of their reign, the logical positivists found themselves in bitter disagreement as to what extent the methods and axioms of the natural sciences can be justified by our abilities to grunt and point. What began as a project to epistemically ground the natural sciences ended as an argument about cavemen. Although such a story might be a good one, the consensus, among the positivists' rivals and the positivists themselves, seemed to be that ahead lay a difficult road indeed if grunts clarify physics. Understandably, that road would be one not traveled, and so our weary fact-warriors walked off elsewhere, leaving the detritus in their path and the rest of the world to deal with the consequences. And consequences there were. Take the current, popular idea that the natural sciences need not be governed by moral concerns; or if they do, such moral governance only need be justified by appeals to governmental, technological, or, happily in this most learned of eras, divine authority. Although we may not be able to connect such popular ideas directly to the positivist program, it is nevertheless no great leap to say that their program was motivated in strong ways by a blind acceptance of the abilities of technological advancement and the triumphant march of the natural sciences.

Not only is there a growing body of literature that details the development of the positivist movements (Cartwright Reference Cartwright1996; Friedman Reference Friedman1999), there is also a growing cadre of philosophers (and intellectuals more generally) who question the basic assumptions behind what is now known as the project of the “right-wing” of the Vienna circle. Putnam's newest installment thus stands as a welcome addition in the effort to clean up the mess left behind by our positivist predecessors. His general intention is to argue that accepting the notion that facts have nothing to do with values not only entails insuperable conceptual difficulties, but more importantly, has unwanted practical consequences (1). And so Putnam's particular goal is to show that when economists embrace such a view, policy decisions become disastrous (52). Such then, is Putnam's attempt to explain not only the negative argument that blindly accepting this notion has grave practical consequences, but the positive argument that examining the practical consequences of our theoretical commitments must be considered part and parcel of our philosophical responsibilities.

Every negative argument, however strong, needs a whipping boy. Here, Putnam chooses Carnap. The choice is an obvious one, since he, along with Ayer, are the leading representatives of the positivist movement. But there is at least one difficulty with Putnam's choice. Given that Putnam chooses to spend much of his time arguing with the ‘leading representatives’ of the positivist movement, the choice ends up confining his gaze merely to the right-wing of the movement. Hence, because of Putnam's desire to argue in broadstrokes against the basic commitments of the positivist program, he ends up neglecting an important and burgeoning literature detailing the historical emergence of positivism (see, e.g., Stadler Reference Stadler2000, Reference Stadler2003; Cartwright Reference Cartwright1996). Disregarding any concrete engagement with the structure of and developments within the positivist program, Putnam's new collection engages only the basic commitment pushing the program along: the distinction between analytic and synthetic statements. Analytic statements, claimed some, are tautologies and express no facts about the world. They are merely syntactically specified conventions, the acceptance of which leads necessarily to the acceptance of the tautologies, both logical and mathematical, that flow from them. Thus included in this class of statements are only logical laws and mathematical axioms. On the other hand, synthetic statements express facts about the world. Such facts are expressions of an observation and such observation is the business of the natural sciences. So, where the truths of logic and mathematics are merely conventional, an expression of fact is, in ideal cases, true by virtue of the objective world. Now the only business left is in attending to the technical problems. But the technical problems quickly led the positivists towards their death-knell. One famous debate was over the criteria that make a statement meaningful.

Obviously, if the only kinds of statements capable of meaning are synthetic statements, then the answer is going to be along the lines that a statement is meaningful when it is either directly confirmed by experience, or reducible to such direct confirmation. Such a criterion is itself not directly confirmable, and so the criterion of meaning renders itself meaningless. Even though the positivists tried to overcome these difficulties (see Friedman Reference Friedman1999), Putnam's strategy is not to argue with this view in particular, for he has done so extensively elsewhere. Rather, through the work of John Dewey, he shows how such a view rests on inflating what is, at times, a useful and important distinction into a metaphysical dichotomy. The difference between the two is crucial, claims Putnam. In fact, positivism only gets its import from taking what seems like an obvious and ordinary way to distinguish, in practice, between descriptions and logical or mathematical identities and hypostasizing that distinction into two mutually exclusive theoretical categories. A dichotomy between analytic and synthetic statements forces us to group one, seemingly similar, class of statements entirely apart from another, seemingly unrelated, class of statements, allowing no overlap in between. A distinction, on the other hand, has a range of application, committing us to no particular classificatory scheme within our paradigm examples of analytic and synthetic truths (11). With this strategy in hand, Putnam goes for the real gold: the dichotomy between facts and values.

Correctly tracing the distinction between facts and values to Hume's dictum that no ‘ought’ may be derived from an ‘is,’ Putnam shows how from Hume's time on, the distinction was inflated into a dichotomy. He claims that some of the Scot's remarks were taken to imply a strict division between two separate classes of judgments. Given Hume's view that meaning depends on representation, of course it follows that he holds no such ‘matters of fact’ to represent virtue or the concept of justice. But it does not follow that statements including terms like ‘good,’ ‘right,’ ‘ought’ and so on are devoid of meaning; nor does it imply that there is a class of statements only expressing ‘matters of fact,’ and that that class has rigid boundaries. By Hume's empiricist lights, claims Putnam, we are no more committed to the idea that each class contains statements having mutually exclusive properties, nor need we solve any philosophical problems with our distinction between matters of fact and matters of value (16). In fact, in deflating the distinction, we have released both classes from the strain brought upon them by claiming, as many moral theorists following the emotivists have, that there is an ‘essence’ to each category which makes the paradigmatic statements falling in one or the other class what they are. The upshot, to Putnam, is that it remains possible and useful, in some contexts, to distinguish between two (or more, he stresses) kinds of judgments. Often, it is useful to do so in order to see more clearly what it is that makes an ethical word ethical, and, on the other hand, what makes a descriptive term descriptive. This seems to be the explanatory power of ‘deflating’ the metaphysical dichotomy between facts and values into a distinction. “But,” he says, “nothing metaphysical follows from the existence of a fact/value distinction in this (modest) sense” (19). Indeed, it remains to be seen if anything stronger actually does follow from such a seemingly weak thesis, a question to which we will return in a moment.

Having put himself in this position, Putnam argues that while it is useful to distinguish between facts and values in this weak sense, it is actually the case that facts and values are richly intertwined. Unlike his negative arguments, Putnam here attempts to show that not only are epistemic values presupposed in the natural sciences, but that certain ‘thick’ normative concepts cannot be disentangled at all. That is, while the natural objects of the natural sciences are determined extensionally and explained causally, the way in which a scientific theoretical apparatus supports such explanation cannot be causal. Each theory—and the selection of a particular theory over and against its competitors—presupposes normative values such as coherence, simplicity, explanatory power, and so on. Thus a scientific theory is itself supported by the selection of values implicit in its practice (31). Such an insight is a good one. But, it is by no means new. Even as early as the 1930s, Neurath had distinguished between what he labeled the “domain of determination,” the level in which empirical evidence contributes to theory selection, and the “domain of underdetermination,” the level in which social and political factors contribute to theory selection. Putnam's oversight here seems to be in large part due to the fact that most of the considerations he discusses in the collection are linguistic matters. Hammering home the point about the dichotomy between synthetic and analytic statements leaves him in a bit of a lurch when it comes to issues in theory selection. And debates over theory selection have long been raging. Moreover, from those debates has emerged a large literature questioning the fact/value distinction along the lines Putnam does in this collection. Helen Longino (Reference Longino1990), Steve Fuller (Reference Fuller2002), and many others have been hard at work in recent years attempting to show that not only do the sciences presuppose epistemic values, but more generally, epistemology is bankrupt without that presupposition. Putnam utterly fails to mention or even cite any of this literature, significantly weakening his attempted deflation of the metaphysical dichotomy between facts and values.

Nevertheless, this aspect of his discussion we shall call his ‘entanglement thesis,’ to be distinguished from those arguments holding that a fact/value dichotomy is necessary, which we shall call the ‘disentanglement theses.’ For Putnam intends the force of this new collection of essays to be an investigation into a philosophy of language that can support the entanglement thesis. By so doing, he intends to show not only that the practices of the sciences presupposes normative practices and judgments, but that objectivity cannot be separated from such practices. Thus, he says, “not only is there no reason to think that the sorts of judgments I have been talking about—judgments of reasonableness—can be reduced to non-normative judgments; there is not even a serious sketch of such a reduction” (145). Talk of objectivity apart from these kinds of judgments is, on his view, mere fantasy.

Of course, strong claims need strong justifications. Entanglement works insofar as Putnam can show how values are presupposed not only in the practice of science but also at the level of language. For if he is to cash in the entanglement thesis by developing a philosophy of language that supports it, then it is necessary to show how not only scientific practices presuppose epistemic values, but also that all kinds of values are presupposed even at the level of individual predicates. However, lest we see this as a claim that there is no difference between epistemic values and ethical values, Putnam needs to show how certain paradigm predicates cannot be simply described as either factual or ethical. The object is to demonstrate that value-talk is as much a part of the “right description of the world,” as any kind of descriptive speech. Description and evaluation depend on each other. But this does not imply that they can be reduced to each other. Such an implication is unwanted to the extent that it lands us back in a dichotomy. Hence, predicates such as ‘cruel’ can be used in either a descriptive context or an evaluative context; but neither may be assimilated to the other. Such predicates, along with ‘rude,’ ‘suffering,’ ‘brave,’ and so on represent the kinds of predicates that can be reasonably attributed to an agent only if the language user has the capacity to “identify imaginatively with an evaluative point of view” (39). In turn, that capacity depends on a reflective ability which is irreducible to any other viewpoint (i.e., logical, ontological, or epistemic). Yet neither need we posit strange natural entities called ‘values,’ nor must we commit to a third realm to make our values objective. From his pragmatist point of view, were we to commit to value realism or platonism, would we thereby be able to better explain values? It is at this point that Putnam engages with the discussion between Bernard Williams, R.M. Hare, J. Mackie, Philippa Foot, and John McDowell and cohorts.

Contrary to those ethicists like Williams, Mackie, and Hare who accept some positivist version of the ‘total conception of the world,’ Putnam argues that it is precisely these ‘thick’ concepts that not only show the irreducibility of values to descriptions, but moreover show the interdependence of the two. Thus, an agent may use ‘cruel’ as a description in the sentence ‘the cruel acts of the government provoked insurrections,’ given that it is delivered in a context where ‘cruel’ functions descriptively. Likewise however, an agent may use ‘cruel’ as an evaluation in the sentence ‘your husband is very cruel,’ given that it is delivered in a context where it functions as an evaluation. A ‘thick’ ethical concept, Putnam claims, “simply ignores the supposed fact/value dichotomy and cheerfully allows itself to be used sometimes for a normative purpose and sometimes as a descriptive term” (35). Moreover, such concepts have values that may be ‘objectively’ outweighed by other factors. The way to gauge when one action or object should be valued more than another is not to point to an agent's subjective desires (internal reasons) or to point to the contingent factors for a decision (external reasons). For Putnam, talk of internal, as well as external, reasons are modes of discourse that require rational justification. And if no hard distinction between descriptive and evaluative reasons can be had, then it follows that, from the level of scientific theory selection to the level of the reasons behind an individual's values, all justification is, in a sense, ‘objective.’ Thus a theory that claims evaluative concepts get their meaning from a speaker's subjective motivation and preferences is untenable (R.M. Hare), as is a theory that claims such ethical concepts are reducible to a descriptive component and an attitudinal component (J. Mackie). Rather, it is internal to an action or a choice that their evaluation “should include the possession of reasons that should be seen from within that very life as having force, and that constitute the way of life as the way of life it is” (88 ff.). It is only by failing to recognize that the total body of language—normative, descriptive, evaluative, and otherwise—interlaces at crucial points that one can submit to a disentanglement thesis.

Investigating some of the consequences of this failure is perhaps the most crucial import of this collection. For it is in his analysis of the work of Nobel-laureate economist Amartya Sen that Putnam's discussion gains powerful insights into the real consequences of adopting the disentanglement theses. As is well known, for many decades now, economics and political science have been dominated by so-called rational choice theory. Just as mathematical logic and the attempt to formalize language led the positivists to declare a division between factually meaningful and metaphysically nonsensical statements, so were economists led to divide between the factual, predictable behavior of economic agents and the agent's subjective desires. Putnam marks out two different phases of economic theory. The first is the classical phase. The second, emerging phase, is led by Sen and Vivian Walsh. In the first phase, the mainline approaches to defining rationality of economic behavior were either internal consistency of choice or the maximization of self-interest. Neither approach in classical economics supports an evaluative point of view. For, as Putnam says, from the scientific point of view “the idea that the economist could and should be concerned with the welfare of the society in an evaluative sense was rejected” (54). It was rejected precisely because evaluation was thought not to provide any foothold in logistic and technical, economic issues. Disentangling values from facts allowed economists to develop important theoretical tools for the explanation of economic fluctuation, market behavior, and so on. In other words, dissociating facts and values increases efficiency. But when combined with a dichotomy that rejects any use for ‘value’ in the explanation of social welfare, Putnam thinks the results can be devastating.

The devastation is particularly striking in welfare economics. Guided by disentanglement theses, the economist no longer needs to consider questions of the particular welfare of a society. Rather, the question only regards maximizing an agent's rational choices. But such choices are always framed in terms of the most obvious decision an agent makes—on the condition that ‘being rational’ is equated with ‘being self-interested.’ Using mathematical models, the economist can predict the behavior of social organizations by presupposing that every actor is rational and accounting for deviation by the complementary concept of irrationality. Economic predictions like these are claimed to be value-neutral, since a person's values are thought to be subjective, and so not in the province of science proper. On the other hand, in a capitalist society, each actor always chooses in his or her own interests. These utilitarian criteria predict the optimal economic function of a society, and with them the disentangler describes the welfare of a society in terms outside of the intersubjective welfare of the particular agents constituting that society. Along these lines, each member is presumably accounted for as a rational agent. If at least one agent does not maximize on his or her welfare in that society, then such a society does not function in an optimal way (so-called Pareto optimality). In a vivid illustration of this point, Putnam says this generates the implication that “[d]efeating Nazi Germany in 1945 could not be called Pareto optimal… because at least one agent—Adolf Hitler—was moved to a lower utility surface” (56). However, Putnam's discussion of the Pareto criterion is somewhat stifled by confusing Pareto-optimality with Pareto-improvement. That is, a situation is Pareto-optimal if and only if there are no alternative situations in which any agent would be better off. On the other hand, a situation is a Pareto-improvement if and only if some individuals are better off and no individual is worse. Hence, while defeating Nazi Germany is not a Pareto-improvement, it does not follow that the defeat is not Pareto-optimal. But the point Putnam makes with the illustration is not altogether lost. Following Sen, Putnam challenges these assumptions on the same basis that he challenged the notion that the practice of science is value-free. For by choosing Pareto optimality as our criterion, presumably we endorse the view that every agent has an equal right and ability to maximize on his or her own welfare. Hence, even our technical means for predicting economic behavior presuppose substantial evaluative concepts. The implication for economists, claims Putnam, is that “to the extent that people's motivations are significantly influenced by their ethical reasoning, we will need to take account of—and to make ‘descriptive’ uses of—a variety of thick ethical concepts in the description of economically relevant behavior” (64).

Here, second phase economics stakes its claim in its ability to describe economically relevant behavior while nevertheless anchoring itself to the insight that economics need be able to explain interpersonal comparisons of value. The question of whether it is better to tax the rich for the welfare of the poor should indeed be a question dealt with by mainstream economics. But the macro-level, as mentioned above, does not seem to provide Putnam with enough traction for his argument. In fact, he claims, economists need be concerned even at the micro-level of descriptions of personal values. For example, in classical theory, a person may choose a commodity A over a commodity B if and only if A yields more utility. Given that a person chooses A, he or she is rational; given that a person chooses B, he or she is irrational. Now, say that our agent chooses B. If we ask for the reasons for this choice, our agent will respond that she values B more than A. Such a response constitutes not a reason given for B over A, but, on the classical model, a subjective value. But, Putnam claims, what if our agent has chosen B because it better fits with her personal values, her hopes, and her dreams? It then becomes obvious that an agent's choice of A over B or vice versa is simply not reflexive. That is, B is valued more than A. If the decision procedure is irreflexive, then unless we simply want to call the agent irrational, the choice of commodity B must be explained using different criteria. Those criteria need introduce the agent's personal, individual goals and long-term needs into the description of the reasons for which the agent chose B over A, when A clearly would have been more ‘useful.’ But, the criteria that explain such behavior cannot, according to Putnam, appeal to a person's utility functions. Rather, it must be able to explain such micro-level behaviors in terms that capture an individual's particular preferences in terms of that person's long-term ethical behavior. It is these kinds of descriptive criteria that Putnam sees second-phase economics championing, and with it, the revival and reintroduction of ethical concepts into the science of economic behavior at the personal level. Indeed, Putnam's collection is largely oriented towards showing that in order to give a coherent account of micro-economic choices, conceptions of the good life—specifically ethical conceptions—must be introduced (94). Hence, Putnam argues (with Sen), normative concepts cannot but be presupposed by any practically effective and theoretically relevant economic description.

It is here that the entanglement thesis does its best work. Putnam's strategy shows that not only is a value irreducible to a fact, but that facts presuppose normative and ethical concepts in order to cash themselves in. Putnam then applies this specifically to economics, showing its intuitive plausibility and its practical effectiveness. Yet, we may ask, does the strength of the attack on disentanglement theses support the strong entanglement thesis? Indeed, from the claim that we cannot disentangle values from facts it certainly does not follow that they are therefore interdependent. Putnam obviously recognizes that we need to be able to see how the very use of basic linguistic items is already a social phenomenon oriented towards the good life. But for Putnam, considerations of the good life enter at the level of an actor's agency and autonomy. Both are concepts that assume a person has the ability to correctly apply a predicate in different contexts—and so it seems that ‘autonomy’ cannot be the preferred concept used to explain interpersonal welfare. Moreover, Putnam neglects a large body of recent critical literature on the subject in favor of advancing general theses as to the projected collapse of the fact/value distinction. Such an oversight seems to mirror the same kind of neglect in details that Putnam attributes to the positivists. In fact, one of the difficulties that Putnam's collection faces is that he needs to show how the philosophy of language he develops may be extended to wider domains and related issues. In this collection at least, that promise is not delivered, for much of the collection focuses exclusively on questions of language, to the detriment of recent developments in the philosophy of science. But such difficulties do not weaken Putnam's ability to show—on a grand scale—that an investigation into the extent to which facts and values interlace is a necessary one. This broad thesis is where Putnam begins the collection. In the tradition of the pragmatists, he shines in being able to finish by highlighting and engaging with real-world, practical consequences of accepting versions of disentanglement theses and the real-world, practical consequences of committing to an entanglement thesis without having to commit to realism, antirealism, materialism, or any of their offspring doctrines.