Alhassanain(p) Network for Heritage and Islamic Thought

II- The Naturalistic Turn in Logic

It is easy to see that the idea that a theory of premiss-conclusion reasoning’s  normative presumptions vary in some nontrivial way with its descriptive adequacy bears nointrinsic tie to theories contrived in the heavy-equipment way. So it would be a mistake to think that a theory’s approximation to empirical fidelity is owed to its mathematical complexity. It lies instead in the fact (if fact it be) that its still-considerable falsities remain eligible for the duties of real-life approximation. It is too early to declare the heavy-equipment industries a dead loss for inference-drawing on the ground. But it surely can’t hurt to start looking for alternatives.

I have an idea about where the search might pay off. It might pay off in a logic reconstructed in the way that traditional epistemology was adjusted by Quine and others in 1969 and following. Just as Quine proposed thenaturalization of epistemology, so I propose the naturalization oflogic .[25]The pivotal point of it all is this. What is the use of admitting to one’s logic agents, actions and the like, if we don’t admit them as they actually are on the ground - warts and all? Of course the last thing that Quine would ever agree to for logic is to do what he himself did for epistemology. For Quine, logic was first order classical quantification theory and nothing else. Quine wanted no truck with people in his logic, formally idealized or in the flesh. He clung to this conservatism until late in his career, when he grudgingly allowed that physics might require a nonclassical quantum logic, and constructivism in mathematics an intuitionist one.[26]The idea of naturalizing logic does not originate with me. In the modern era alone, it is actively proposed by Dewey and sympathetically entertained by Toulmin and Finocchiaro.[27]In this second section, I’ll give a sketch of how the naturalization of logic might go, with some indication of the good of it. To this end, I’ll call upon the reflections of Nat, an imaginary naturalizer.

Nat holds a naturalized logic to an adequacy condition of “empirical sensitivity”. To bring it off, the logician would familiarize himself with the data that the cognitive sciences seek to account for. He would develop an informed acquaintance with the findings of the empirically best-confirmed of those theories. On pain of abandoning his logic, the naturalizer would offer a satisfactory account of its empirical disconformities with the data and findings of the partner sciences.  Nat was also quick to appreciate a fact openly on view on the ground if we take the trouble to look. It is that the human animal is a knowledge-seeking organism and that premiss-conclusion reasoning is an important facilitator of its achievement. Accordingly, in addition to its empirical sensitivity, a naturalized logic of reasoning should display anepistemic sensitivity as well. This is already shaping up to be a coherent methodology:

Respect for data : Logic should develop a healthy respect for the data and should respond to them with empirical and epistemic sensitivity. And correlatively it should respect thedifficulty of paying them proper respect.

Consider in this regard Gerd Gigerenzer’s alarm about the “data-bending” he sees in the application to cognitive processes of methods developed by theories of statistical computation and theory testing in the 1940s and ’50s. Under the influence of such assumptions, theories of cognition in experimental psychology

were cleansed of terms such as restructuring and insight, and the new mind has come to be portrayed as drawing random samples from nervous fibers, computing probabilities, calculating analysis of variance, setting decision criteria, and performing utility analyses.[28]

Taken together, these assumptions conceptualize the human reasoner as an “intuitive statistician”. This “radically changed the kind of phenomena reported, the kind of explanation looked for,and even the kind of data that were generated .”[29]  What is more, researchers who adopted the methods of inferential statistics were unaware of this change, since these methods had become canonical in psychology.[30]

Data-bending is misconstrual of how the data actually are, owing to the wrong sort of data-loading assumptions. It is conceiving of facts on the ground in ways that facilitate pre-conceived theoretical outcomes. This is nothing to make light of. We have known at least since Bacon that data aren’t self-announcing, and yet that they can’t be grasped at all without some prior or concurrent conceptualization of them. This generates a nasty problem for the experimental theorist. He can’t proceed without conceptualizing his data, and yet the line between conceptualization and misconceptualizations is easily transgressed. There are no algorithms for the avoidance of data-bending. But there are some useful lessons to be learned. One is to be careful. Another is to respect these difficulties.

Nat attaches a singular importance to these lessons. He also views with suspicion the most prominent justification for sticking with an empirically false theory. According to that view, it is not the aim of such theories to be descriptively adequate; the goal is to establish rules that are normatively authoritative for human practice on the ground. Fundamental to Nat’s project is his rejection of this assumption, at least until such time as it might come to have a convincing independent defence. Like the rest of us, Nat knows full well that science frequently adopts false idealizations, such as the infinite cardinality of populations in population genetics. These are, he says “virtuous distortions”[31], and are so when they are compensated for by the confirmation of the theory’s observational predictions. Normatively idealized theories of empirically realized and normatively assessable human performance can’t meet that test - consider here the perfect information assumption for ideally rational decision systems. So the theory’s only payment alternative would appear to be its success at thenormative checkout. But Nat’s view is that there is little in the way of a settled consensus about what doing well there actually amounts to. In the absence of a sounder grasp of normative legitimacy, Nat thinks that the better option is to try to handle the problem of empirical infidelity in a different way.

Nat’s own response is twofold. He stresses the priority of attending with care to the actual details of human reasoning on the ground, without preconception and well in advance of assessments of goodness or badness. His worry about preconceptions is an ecumenical one. He dislikes the preconceptions of orthodox logic. But he dislikes no less the preconceptions of orthodox psychology.[32]

Nat has a working idea of how he thinks the naturalization process should go. To this end, he has availed himself of a fable, a kind of thought-experiment about the ways of naturalization. Nat imagines that for some years now a visiting team of cognitive anthropologists from a distant and unknown extraterrestrial place has been hard at work in our midst. Its earthly mission is the examination of human cognition. These cosmonauts from afar are well-equipped for their work. Themselves organic beings, they have been able to acclimate to the particularities of planet Earth. Themselves cognitive agents, they have some acquaintance with how cognition works under the ecological constraints of habitat. They are also accomplished field linguists and intelligent problem-solvers. Nat has adapted this fiction of the visiting scientist from van Fraassen’s notion of an epistemic marriage, which envisages an epistemic partnership between dolphins, extraterrestrials and us.[33]Just as Wittgenstein would wonder whether if a lion could talk, we would be able to understand him, the visitors were interested in whether the information-processing activities of us rose to the bar of cognition in a manner that would be discernible to their own methods of enquiry.[34]

The visiting anthropologists are here to do some descriptive epistemology, to take the earthly cognitive pulse as best they can with the exploratory resources available to them. They will run their investigations in compliance with their own understanding of best scientific practice. They will leave themselves free to consult the published record ofour own conception of best scientific practice, but without  prior commitment to defer to ours when it conflicts with theirs. Neither will they consult our philosophers, not even the ones who do logic; at least not until their own work is finished. Why would the visitors bother with philosophers? They are not themselves philosophers; they are scientists conducting a naturalistic examination of the naturally occurring phenomena of cognitive behaviour in their subject population. This exclusion is neither ideological nor hostile. It simply reflects the plain fact that scientists do their thing and philosophers do their largely different thing. Philosophers and scientists can share a common subject-matter, but they each tend to cut their respective cakes largely in independence of the other. A sensible course for each would be to arrive at a decision about what happens now. There might be some advantage for each in exposing one another’s views of the matters they have in common to a reciprocal scrutiny. But the visitors’ point, and Nat’s too, is that this scrutiny is better reserved until the scientists have done their business.

In this same spirit, the visitors decided that it was not part of their mission to discern how on-the-ground human inference-makersdefined inference - or knowledge or belief or whatever else. Nor would they circulate questionnaires asking their subjects to tell them whether they thought that this, that or the other thing isbona fide knowledge or good reasoning. The visitors have long been aware that, among their own kind, people do quite well at knowing and reasoning without being very good at saying what these things are. Until they learned otherwise, they would assume the same for us. From this came another foundational insight:

Being good at and knowing what : Being good at knowing things and reasoning well does not require knowers and reasoners to know how to define what knowledge and good reasoning are, or to specify the conditions that bring these things about.

Substitution here for “reasoning” of “remembering”, “imagining”, “seeing”, “high-fly ball catching” and a plethora of others generates a host of sentences most of us would consider too obviously true for words. Consider, for example, our knowledge by looking that it has snowed overnight. Everyone in town with eyes to see knows this to be so. But hardly any of them has much of an idea about how the mechanics of visual cognition actually play out. Why wouldn’t this also be true for “knowing” and “reasoning”?

Perhaps it will be protested that if we are so good at knowing things why are we so bad at knowing what knowing is? It is an excellent question but a feeble protest. Does the questioner really think that his question carries the force of rebuttal? If so, he would do well to explain why.When, in the late stages of their enquiry, the visitors relaxed their exclusion of earth-bound philosophy, they were astonished to learn - disapprovingly so - that some of the first of our great philosophers were in thrall to the idea that a person’s knowledge requires a prior or concurrent grasp of a real definition of it, that a concept can’t be instantiated in human behaviour in the absence of the behaver’s “analytic grasp” of it. This, the visitors thought, amounted to a scepticism so corrosive as to make the attainment of knowledge a generally impossible target. When some of our own earthbound  professors of epistemology protested in turn the excessiveness of the visitors’ alarm, they answered in unison: Socrates andles autres could not have thought this way had they paid scrupulous attention to the ups and downs of what really goes on in human life; especially in theagora.

Perhaps the visitors were a trifle hasty. They didn’t have time to immerse themselves in the history of earthly epistemology. But they’d had enough exposure to it to have tasked themselves with the four questions. One is whether having a real definition - or conceptual analysis - of it is a condition on knowing what knowledge is. A second is whether knowing what knowledge is is a condition of there being any. The third is whether the ancients were inclined to favour affirmative answers. The fourth is whether this favoritism has a discernible presence in modern-day analytic approaches to epistemology. Not having the time or inclination for extended consideration of these matters, they came together on three summary positions;first , that any notion that a scientific knowledge of knowledge is to be got by a conceptual analysis of “knows” is a misbegotten idea;second that the only room for big-box scepticism in the science of human knowledge is by way of the default rule that big-box scepticism in the science of human knowledge should not be so much as entertained, never mind rebutted, except for weighty cause; andthird that the earthly epistemological tradition betrays too little heed of these constraints; not without exception, but dominantly so. They thought that a thoughtful examination of the cognitive routines of human life make it clear that just about the last thing that could be true of it is radical scepticism of any broad kind. On the contrary, they thought, it was empirically evident that human beings are good at knowing things - not perfect but good; that they have lots and lots of it about lots and lots of different things; that the human cognitive harvest is both abundant and diverse.

A further shock was administered when, shortly before heading for home, the group began to look into what earthboundlogicians had been saying about these things. They were taken aback to discover the confidence and wide-spreadness of the dogma that premiss-conclusion reasoning is no good when it fails to be deductively valid or at least inductively strong in the technical sense familiar to the statistico-experimental sciences. What surprised them most was the utter lack of behavioural recognition of this would-be fact in the subject population. Most of the reasoning that passes muster there - and is evidently accepted as good - fails both these standards. This left the visitors with two choices. They could condemn their human subjects as across-the-board losers in the reasoning game. Or they could reject the validity-or-inductive strength condition as a general requirement for good premiss-conclusion drawing. Of course, they chose the latter. How could they have not? They were natural scientists who faithfully respected the necessity of respecting the data on the ground. I hardly need say that Nat was quick to sign on, and in short order would see that one of the essential tasks for a naturalized logic of inference is the discovery of the conditions under which this “third way reasoning” would be properly achieved; and in the process to see that the resulting third way logic would be the natural home for virtually all the nonmonotonic logics currently on offer, once such compensating adjustments as might be required were worked out.[35]

It was equally clear to the visiting team that there is another abundance for the naturalized logician to pay attention to in the cognitive ecologies of the subject population. Knowledge exists there in abundance. Error is another of its abundances. The human animal makes lots and lots of errors about lots and lots of different things. A logic of reasoning must bring these abundances into a benign harmony, in response to yet another empirically discernible feature of human cognition.

Enough-enough : Human beings know enough about enough of the right things enough of the time for survival and prosperity and, from time to time, for the erection of civilizations of dignity and lingering worth.

There is no question here of our prosperityentailing an abundance of enabling knowledge. Cognitive abundance is our visitors’ working hypothesis. It is not ruled out that there is none better. It is not ruled out that they might come to think that a more plausible working hypothesis would be that the beliefs that serve us well are mainly false, in which case,right belief would separate away fromtrue belief. But, if anything was clear to the anthropologists, it was that beings like us are awash in what we might call alethic beliefs, which are believings-to-be-true. This gives the new working hypothesis a bit of a twist, providing that the best way of achieving prosperity is believing-to-be true propositions that are false. This handed the visitors a chuckle and the idea was dropped like a hot potato. Where, they asked, is it empirically discernible in the belief-prosperity data that this story would hold in the general case?

Nat followed the visitors in thinking that an attentive naturalizer would lodge his curiosity about good and bad reasoning in a default principle for the logical theorist:

NN-convergence : Take it as given in the absence of particular reasons to the contrary that humans reason well when they reason in the ways that humans normally reason in the conditions of real life.

Nat was careful (and well-advised) to add an important caveat. The NN-convergence principle isnot a safe default for all aspects of human cognitivity; its application here is reserved for premiss-conclusion reasoning. It tells us to judge premiss-conclusion reasoning in roughly the same sort of way that we’d check the subject’s pulmonary behaviour. What this amounts to is that, for the most part, a human agent’s premiss-conclusion reasoning is the right way to reason when his conclusion-drawing mechanisms are in good working order, and at present working in the right way, engaging good information in the absence of hostile externalities.

Breathing is like that too. By and large, the goodness of good breathing depends on the ship-shapeness of the pulmonary equipment. At bottom, as we might say, good breathing is not down to us; it is down to our equipment. Sometimes, however, itis down to us. We can’t sing grand opera, or achieve a place on the Vancouver Canucks, if we don’t breathe in the required ways. We can’t bring this off until we learn to do this. Almost always we’ll have to be taught by an expert, and oftener than not it won’t work. Hardly anyone has the breath forTosca or for the seventh and deciding game of the Stanley Cup playoffs.[36]

These same passivities and activities are discernible in reasoning. Most of it happens passively and largely out of sight of the mind’s eye. Most of it is down to good machinery. Indeed, it isalways down to good machinery, but sometimes it is also down to us - to our disciplined, patient and skilled application of the expert routines that knowledge sometimes requires. Think here of the incompleteness of formal arithmetic.

We’ve already said that our visitors aren’t philosophers or logicians of the orthodox sort. They are cognitive anthropologists. This is not Nat’s situation. Nat is a philosopher whose dissent is launched from within the very orthodoxies he finds fault with. It might be appropriate for the visitors not to take a philosophical position on knowledge. But Nat can hardly proceed to completion without allowing his epistemic reach to carryepistemological implications as well. Nat knows that philosophical theories of knowledge broadly partition into two paradigms, one of them more historically dominant than the other. This first, he calls the Command and Control Model, and the other the Causal Response Model. We see in this division a nice correspondence with the already noted distinction between active-case knowledge, whose attainment is significantly down to us, and passive-case knowledge, whose attainment is largely down to our equipment. Having noticed the statistical dominance of the passive over the active, together with the passive underlay of even the active, Nat concluded that the right base epistemology for a naturalized logic is the CR model, provided that, where indicated, it accommodates the CC model as a proper subtheory for various kinds of knowledge acquisition “up above”, in which the knowing agent has an active and self-aware role to play. Think again of the second incompleteness theorem.

The idea that a naturalizedlogic needs a baseepistemology might strike some readers as implausible - anyhow puzzling. Why would I assume it? The answer is that in the human world reasoning is transacted by cognitive beings, and one of its principal functions is the facilitation of knowledge-seeking and the attainment of epistemic goals. Nat is a philosopher who wants a philosophically tenable account of what makes such reasoning in pursuits of such ends the right way to reason or, as the case may be, the wrong. A logic so designed cannot be judged in the absence of a philosophically convincing understanding of what it is that is facilitated or attained when these goals are in play and properly handled.[37]

Given Nat’s naturalistic leanings, this seems much the right choice. Right or not, it is a fateful one. It effects a considerable scrambling of the once-pacific distinction between reasons and causes, and it makes of knowledge a much more causal phenomenon than an intellectually wrought achievement. The same holds for reasoning. By a statistically large measure, our premiss-conclusion reasoning is right when the conclusions we draw are causally induced by belief-producing devices when working as they should. In shorter words still, in the general case you don’t have to be smart to reason well; you have to be healthy.

An interesting case in point, and one to which the visiting team paid a lot of attention, is the utter widespreadness of the phenomenon of being got to know things by being told them by others. Both models recognize the phenomenon, but they give it quite different theoretical treatments. The central point of contention is whetherjustification is a general control point for knowledge-transmission, with the CC-theorist voting aye and the CR-theorist nay; that is, nay as a general condition on transmission. The CR-reservation came down to this: If we pay close attention to what happens on the ground, the presence of justificatory involvement ismarkedly less discernible than the levels of recognizable cognitive satisfaction among transmitters and recipients alike. As the visitors came to appreciate, the CC-crowd has spared no effort to attribute the workings of justification even in the absence of its recognizable behavioural presence in the general case. Nat has noticed this too. The visitors didn’t quite know what to make of it. But Nat knew. What he made of it was that the CC-crowd weren’t paying sufficient heed to the respect-for-data rule.[38]

The on-ground data also disclose how much of this causal belief-inducing activity proceeds out of reach of the mind’s eye, without notice or attentiveness, without deliberation or overt case-making. As Nat puts it, most of human inference is inference “down below”. If this is so, it plays straight into the normativity question. If we wish to remain faithful to the NN-convergence thesis, we will need to derive a naturalistic account of errors of reasoning; for it is to precisely these that the badness of bad reasoning is owed, is it not? Such accounts don’t lie idly about. They are not numerously available or free on board. They have to be laboured after. First and foremost perhaps is the reconciliation of our two abundance theses - the abundance of knowledge and the abundance of error. Trailing along is a perfectly natural puzzlement about how, if we’re so good at reasoning, why are we so bad at avoiding errors.

Nat has an answer to this which (very sketchily) comes to this: A standing liability for the human knower, no less than the human high-jumper, is that there is only so much he can do. There is only so much that it makes sense for him to want to do - or have the slightest interest in doing. He must learn to live and to set his sights within his cognitive means. As with any design-constrained and resource-limited activity, cognitive success calls for economic alignment of capacity with resource-availability. Nat noted that human individuals are graced with very efficient feedback mechanisms. It is a welcome advantage, enabling a better record at error detection after the fact than before. He concluded from this that it is more economical for a resource-bound individual to correct mistakes after the fact than to avoid them before commission. He also observed that, in case upon case, the human animal is more adept at conclusion-drawing than he is at premiss-selection. Think of the former as errorsof reasoning, and of the latter as errorsin reasoning. The dominant fault of premiss-selection ismisinformation ; and it is markedly easier to be misinformed about something than to draw from it the wrong conclusion. On thinking it over, Nat came to the view that

Misinformation and misinference : There are significant global variations in peoples’ well-informedness - matching kindred variations in region-to-region levels of ignorance - and yet comparatively uniform performance-levels in human conclusion-drawing.

Nat’s naturalizing focus is on premiss-conclusion reasoning, chiefly on the drawing side. If, as it appears, this is something we’re uniformly good at, then errorsof reasoning, when committed at all, must arise from external hostilities or equipment failure. Generally speaking, when we make mistakes of conclusion-drawing, we are dog-tired or strung out, or leveled by a stroke; or the conclusion-drawing equipment isn’t - like the nineteen year old Ford - in quite good enough working order; or we are awash in information-overload. If this is right, something else is also bound to be right. It is that

The thickness of error : Error is a natural phenomenon, with a dominantly descriptive character, but also imbued with what Bernard Williams callsthickness . Roughly speaking, this means that to find error in a person’s performance is to see it as an utterly natural phenomenon but one that is out of joint with the how things are supposed to go.[39]

Since his first course in logic, Nat has known that logic from its very beginnings sought for the attainment of a decent theoretical command of what Aristotle callsfallacious reasoning. Since his second course in logic, Nat has also known that the fallacies programme is nowhere in sight in any of the going mainstream logics.[40]It is not hard to see why. We’ve already said that the modern orthodoxies were built for the relief they promised for metaphysical and epistemological anxieties in the foundations of mathematics. They weren’t built for human reasoning, even for when it is transacted fallaciously. This, of course, can’t be Nat’s own position. Nat wants a logic for real-life premiss-conclusion inference. He wants his logic to solve the normativity problem. He thinks that the correct account of errors of reasoning will be the key to its solution. Surely, one would think, no account of reasoning errors could be complete if it didn’t revive and make some headway with the fallacies project.

Nat knows that on the traditional approach fallacies are errors of reasoning having certain distinguishing features. One is that they are attractive, hence inapparent. Another is that they are universal, in the sense that virtually all of us are disposed to commit them with a frequency higher than our error-makings in general. Yet another is their incorrigibility; that is, even after detection and correction, rates of post-diagnostic recidivism are extremely high. Like the rest of us, Nat is also familiar with the traditional list of the fallacies - hasty generalization,argumentum ad verecundiam (argument from authority),argumentum, ad ignorantiam (arguments from ignorance), and so on. It wasn’t long before Nat made a discovery that genuinely surprised him. He saw that a proper regard for the respect for data principle discloses that virtually all the items on the traditionallist of fallacies have no discernible presence in the instantiation-class of the traditionalconcept of fallacy. Accordingly,

Concept-list misalignment : Virtually none of the fallacies in the traditional list lies in the extension of the predicate “is a fallacy” as traditionally interpreted. Either they aren’t errors, or they are not attractively inapparent, or not universal or incorrigible.

Here is an example that especially impressed Nat. Nat is on his first trip to Brazil, a country he knows little of, but enough to know that animals called ocelots are resident there. One day, Nat and his Brazilian host are tramping the countryside. “Look”, exclaims Luis, “an ocelot!” Nat is surprised. “Good heavens, Luis, I had always imagined ocelots as two-legged, not four.” Nat has come to see that, on the basis of a single encounter, ocelots are four-legged. On thinking it over a bit, he also realized that the true generalization “Ocelots are four-legged” is not falsified by the plain and subsequently discovered fact that this other ocelot, Ozzie, a beloved resident of the zoo of his friend Luis’ hometown, is three-legged. And in no time at all it became apparent to him that this kind of hasty generalization is, in the human species, as common as dirt (as the saying goes), and that to a quite marked degree the generalizations hastily drawn are actually right, not wrong. Whereupon we have it that

Hasty generalization : Hasty generalization is not a fallacy in the traditional sense. Indeed, comparatively speaking, it is hardly ever wrong when actually performed.

The obvious question now is whetheranything instantiates the traditional concept. It is, as I write, an open question in Nat’s logic.[41]

The naturalistic turn pulls logic and cognitive science in the opposite direction from the mathematical turn. For three decades and more, the mathematically shaped enquiry has searched out an affiliation with a closer attachment to agent-centred, goal-directed, resource-based, time and action systems. These enrichments are a considerable complication for, whose management more basic formal equipment must be upgraded with new machinery of correspondingly greater complexity, sometimes problematically so. Nat’s worry, like my own, about these heavy-equipment upgrades is that the more complex they are to handle, the likelier they are to invite the solace of a new batch of simplifying purpose-built performance norms. Nat isn’t opposed to the enlargement of capital assets as such. His reservation about heavy-equipment upgrades is that they leave the normativity problem undealt with. His present inclination is to enrich the logic of premiss-conclusion reasoning with naturalistic assets, especially those of them that improve our grasp of the on-the-ground management of error - its avoidance, its commission, its detection and repair. For it is here that he sees promise of a principled solution to the normativity problem. The empirical turn is an attempt to reshape this enlargement of capital assets, by lightening up on the notion that theory-building is intrinsically theorem-proving. Theorem-proving is not demanded for population biology. Why, asks Nat, should it be demanded here?

What Nat proposes for logic is what Quine proposed for epistemology. Not everyone thinks that Quine’s is a tenable project or, so far, a well executed one. But no one should think that these days naturalized epistemology is anything but a well-accepted part of mainstream epistemology. It is vanishingly unlikely that a friend of naturalized epistemology would dislike Nat’s proposal for logic because he distrusts the role of naturalism in philosophy. But he might well dislike it because it lies in the nature of logic not to take well to naturalistic intrusion. This, after all, was Quine’s own position. At the core of it all is modern logic’s deeply dug-in loathing of psychologism. Epistemology leaves lots of room for psychology, and logic leaves none. Naturalized epistemology may now have found a place in the big leagues, but this is the last thing that one could say of logic. This makes Nat’s proposal a radical one for logic if not any longer for epistemology. It also makes Nat’s proposal a contentious departure from the still well-favoured normative idealization approach to the social sciences.  In a nutshell, what Nat wants is logic’s reinstatement of psychologism.

Nat fully acknowledges the efforts of the newer developments in heavy-equipment logics to do better on the score of on-the-ground inference-friendliness. His chief reservations are two. The heavy-equipment technologies don’t solve the normativity problem; and bulking up the formal machinery hasn’t closed the gap between the logic’s theorems and settled practice on the ground. As far as psychology goes, there are weighty autoepistemic considerations to take respectful notice of. If the heavy-equipment crowd thought that there was room for psychology in their projects, they’d have put some in. But they haven’t. so they don’t.

If the case against the normative presumptions of heavy equipment logics could be made to stand, it carries like consequences for the normative presumptions of the ideal models approach to the social sciences generally. This is getting to be quite a bit of nay-saying. If acquiesced to, all of normatively presumptive science would be put on hold until the normative authority problem is properly sorted out. This separates Nat from virtually all the going traffic in agent-based logics, including logics of probabilistic reasoning, belief-change and decision, epistemic and justification logics, fallacy theory, discourse analysis and normative psychology. Of course, enquiries of every kind are needful of starting points and of assumptions that frame their conceptual spaces. These are assumptions that lend enquiry its procedural and organizational shape. This produces a pair of important consequences for the would-be dissident.

One is that by the very nature of received opinion, there is not much pent up enthusiasm for paradigm-overthrow. The other is that, for want of practice, the orthodoxies’ disciples aren’t much good at defending them. There is a story making the rounds in which the dialethic logician Richard Routley once challenged the logically more strait-laced David Lewis to prove the classically interpreted law of noncontradiction.[42]There was point to his challenge. Routley thought that, when interpreted the right way, the law of noncontradiction didnot preclude the truth of some select contradictions. So the challenge to Lewis was to show that this couldn’t be so. When it came down to it, Lewis didn’t bite; he refused to be drawn. He told Routley, in effect, to grow up and stop horsing around. Lewis’ was a telling response. It was an outright and unconsidered dismissal. The point of this littletableau is dialectical. Orthodox assumptions carry and are protected by high levels of dialecticalinertia . So Nat would have been foolish not to have anticipated that his own dissensions might receive scant attention in the high courts of received opinion.[43]On the other hand, Quine and others prevailed to good effect in the aftermath of 1969, against the grain of stiff resistance. So who knows? Perhaps Nat’s naturalistic prospects will have brightened forty or so years hence[44]

Acknowledgements : For comments on earlier drafts or discussion of closely related matters, I thank most warmly Johan van Benthem, Franz Berto, Dov Gabbay, John Greco, Maurice Finocchiaro, David Hitchcock, Jaakko Hintikka, Frank Hong, Ralph Johnson, Lorenzo Magnani, Christopher Mole, Adam Morton, Ahti-Veikko Pietarinen, Shahid Rahman, Harvey Siegel, Robert Thomas and Yi Zhao.

Notes