Sosa, general assumptions, and the skeptical trojan horse

Abstract:

For many hinge epistemologists, general, background assumptions are principles that help providing default or presumptive justification to our empirical beliefs. However, the ‘blanketing’ nature of a priori arguments to the end of supporting the rationality of general assumptions might be seen as the Trojan horse through which radical scepticism threatens the common sense picture of the world. Sosa’s recent distinction between background presuppositions and domain-defining conditions, as well as his claim that agents are not negligent for dismissing global scenarios as irrelevant to epistemic normativity, are instrumental to avoid an epistemic construal of über hinges, and thus, to a reassessment of the function they really perform in regards to ordinary practices of judgment.

Keywords: Agency; Epistemic Negligence; Hinge Epistemology; Humean Scepticism; Virtue Epistemology

Introduction

In his seminal article on Wittgensteinian certainties, Crispin Wright (2004, p. 42) came to grouping those ‘hinge commitments’ that are brought to light in On Certainty in three sub-classes: (i) judgments such as simple arithmetic operations and empirical certainties about which one could not be mistaken (“I have two hands”) - they are insulated from disconfirming evidence; (ii) particular beliefs such as “My name is M. G.” which are grounded in an overwhelming body of evidence together with interlocking coherence with many other particular beliefs, so that one’s vast array of specific beliefs make up a system of beliefs with mutual support; and (iii) general, background assumptions such as “There is an external world” and “We are not radically in error” which do not (and cannot) derive from evidence. It is plausible to claim that with these three sub-classes Wright was providing an ordered description of the core components of the common sense picture of the world.

Among those components, special attention has been given to general, background assumptions. Firstly, because as they are not based on evidence, their epistemic standing is (if any) far from clear. Are those commitments innate? Are they nothing more than natural inclinations bestowed by gentle nature? Can they be properly claimed? And if so, is there an a priori argument available so as to enhance natural trust in general hinges as nonarbitrary, rational trust? What is at least clear is that, contra Moore, given evidence plus inductive reasoning are not able to explaining and/or justifying basic, general commitments.²

There is, besides, the question as to whether Wright’s type-I and type-II Moorean certainties really are of the same overall category as type-III hinge commitments, and so, as to whether the former types are as epistemically fundamental as the latter is.³ This question might be expressed in two different, though internally related ways.

First, as the problem of whether empirical propositions for which one has found particular circumstances in which all (empirical) doubts are removed do have the same logical status of general assumptions - a logical status that, as hinge epistemologists are eager to underline, is bound up for the latter with the basic architecture of epistemic rationality.

It is not only that empirical certainties such as “I have two hands” are invulnerable to doubt only insofar as one implicitly incorporates a particular set of circumstances into examples involving those claims, so that circumstances in which an error concerning such ‘firm’ beliefs would make perfect sense are conceivable after all.⁴ More importantly, the trouble comes from the fact that at least as far as how human rationality is de facto constituted, general assumptions have a permanent, inbuilt logical status - one that in spite of the empirical form of such commitments, might be recognized independently of circumstances as well as of the truth-value of empirical beliefs, however ‘certain’ the latter may be. There is thus a striking disanalogy between the a priori status of general assumptions and the fluctuating role played within our system of reference by Moorean empirical certainties. Besides, it is not necessary a shift of logical status of empirical beliefs as a result of their shift in epistemic standing. An empirical belief can increase its epistemic standing up to the highest rank of being certain while its logical status remains unchanged.

A further question (and disanalogy) arises from the fact that empirical knowledge requires principles for the unification as well as for the objective import of one’s system of particular beliefs provided only by general, background assumptions.

The empirical rules that help to guide our ordinary epistemic navigation cannot acquire anything more through quotidian procedures of perception, psychologically explained habituation, and induction than comparative universality and widespread usefulness. Unaided by general beliefs, perceptual, mnemonic, selective and judgmental competences would only be able to ground a contingent system of merely contingent beliefs - a subjective system where there would neither be universal and necessary connections among sensible appearances (only constant conjunctions) nor, more importantly, there would be a principled way of both referring experience to objective externality and to refer it to an objective, no cultural-dependent rationality, even if the latter is nothing but, as with Coliva, a rationality relative to how human beings are de facto constituted (COLIVA, 2015, p. 128), to wit, even if it is nothing else than a system of rationality that is comprised of a set of principles that are just finitely universal and finitely necessary principles.⁵ For our commonsense system of beliefs to express, not just events in the mental life of the subject or rules relative to contingent forms of life, but events in the career of objects, general, background assumptions are needed. This is why Sosa underwrites the fact that “general beliefs and commitments are needed from the start” (SOSA, 2021, p. 201) to explain epistemic practices and judgments.

But what kind of explanatory account is that which general, background assumptions provide? Hinge epistemologists take usually for granted that type-III commitments are the first principles in the architecture of perceptual justification, principles that help providing default or presumptive justification (justification that might be defeated by particular, unfavoured circumstances) to our empirical claims. On this view, general assumptions would be applicable to specific situations and particular perceptual judgments so as to establish a reason-based relation between the various empirical beliefs and those general, background commitments which while acting as their reasons, cannot themselves be grounded in still more fundamental reasons. The series of reasons would thus end with fundamental assumptions that are neither self-explanatory nor metaphysically self-validating.

However, the above-mentioned contrast between general hinges and local epistemic commitments should alert us. It suggests that each of the two classes plays a different role within our cognitive practices and has a distinctive nature. Maybe there is, after all, an unbridgeable gap between them.

Hinge epistemologists have been mainly interested in the question of whether the series of reasons terminates arbitrarily, namely, in the problem of whether an argument to the end of vindicating the groundless yet rational nature of general assumptions is to be found. There is no doubt about the relevance of such a question. However, they have been mainly blind to the further question as to whether the presupposed principles, which as Reality tends to correspond to appearance (SOSA, 2021, p. 129) or I am not the victim of a massive perceptual and cognitive deception (COLIVA, 2015, p. 6) are general and indeterminate, can really be rendered determinately applicable as conferring epistemic default justification to particular perceptual judgments. Notice that this further question is (at least at face value) independent of whether general assumptions are able to gain a rational standing. An argument to this end might be available, and yet those principles might be epistemically idle regarding the epistemic standing of empirical beliefs.

Curiously enough, this is a lesson from Sosa’s early criticism of Davidson’s a priori argument against radical scepticism (SOSA, 2009, p. 116 - 189) - an argument that in spite of important differences, is of the same ‘blanketing’ nature as the recent views on the structure of perceptual justification proposed by hinge epistemologists.⁶ I will argue that the same sceptical Trojan horse that on Sosa’s reading Davidson’s account admits (SOSA, 2009, p. 117), is also admitted in epistemic construals of the function of general, background assumptions - where an epistemic construal is one that conceives of general hinges as assumptions that guide the reflective, epistemic agent to aptness. It is my view that there is (and there can be) no kind of logical reasoning from general, background assumptions to empirical, substantial beliefs which provides presumptive justification to the latter.⁷

In section 2, I will focus on Sosa’s recent proposal of a hierarchy of grades of knowledge which adds to the familiar categories of animal knowledge and full reflective knowledge the further category of securely (safely) knowing full well (SOSA, 2021, p. 161). If, on the one hand, this addition threatens to undermine Sosa’s telic account of knowledge as fully apt belief, thus raising an internal challenge for robust virtue epistemologists - one that must be properly answered; it also is instrumental to shedding light on the contrast between presuppositions that affect the epistemic standing of the performer and domain-defining conditions of performance (SOSA, 2021, p. 160). Drawing attention to the latter category, I will argue that besides being epistemically nonoperative within ordinary contexts of weighting reasons, general, background assumptions do not help improving the epistemic quality of beliefs.

In section 3, I will draw heavily on the second-order nature of the sceptical challenge, and on how global scenarios take the selective competence of epistemic performers as their target. This adds further pressure to the view that general hinges are assumptions through which one aptly attains apt beliefs, thus paving the road to a reassessment of the function they really perform. On our favoured view, general, background assumptions open up the logical space for epistemic evaluations to be possible. As such, they do not contribute as components of telic normativity to the epistemic status of empirical beliefs. Hinge epistemologists have not shown that, in order to inject objectivity and externality into our system of beliefs, we actually need to apply - or that we need to be able to apply - general assumptions to given experience. They have shown only that, for that purpose, we need to assume something indeterminate such as the fact that most experiences are veridical - an assumption internally related to the very possibility of epistemic agency.

It is to the advantage of Sosa’s recent contribution to the sceptical debate that by arguing that there is a necessary relation between the framework of epistemic evaluation and epistemic agency he manages both to offer an answer to the sceptical question as to whether general assumptions are rational and nonarbitrary, as well as to deflect the kind of empirical, Humean scepticism (the sceptical Trojan horse) which is of main interest for this article. As it is to the advantage of the present proposal, or so I think, that it is the only view on general assumptions and their function that fully accords with Sosa’s conception of epistemology as a ludic performance domain (SOSA, 2021, p. 151) sealed off from conditions for performing which are not (and cannot be) conditions for performing well. Epistemic explanations do not include our commitment to conditions without which performances, however their epistemic quality might be, would be cancelled.

1 Background presuppositions and domain-defining conditions.

In Epistemic Explanations, Sosa has come to improve his previous analysis of the case of Simone⁸ by introducing the notion of background presuppositions - presuppositions that are ways through which the knower has a second-order grasp that her animal (first-order) affirmation would be apt. On Sosa’s view most of our epistemic performances take place against the backdrop of implicit, second-order assumptions, as with the following case.

Simone is an experienced fighter pilot that after a long career is subjected to a simulation test. She is unaware of the test, and thinks that, as usual, she is piloting a real plane and shooting at real targets. Let us stipulate that in the present occasion, and despite the fact that she might easily have been led to the simulation screen, Simone is piloting a real jet-fighter, and has just hit a real target. How does the simulation scenario affect the quality of her current performance? Since the scenario is not actual, does it affect Simone’s performance at all?

It is clear that the actuality of the scenario would affect one’s performances by making them false. However, it seems also clear that the performer is within her rights to assume by default (insofar as there is no sign to the contrary) that the conditions for the performance are normal and propitious. According to our quotidian, normative framework, Simone is not guilty of epistemic negligence for taking for granted that she is shooting at real targets. Simone’s performance is normatively flawless, and achieves its goal. She both enjoys animal knowledge and full reflective knowledge; the latter because she is neither reckless nor negligent in presupposing that the situation is adequate. In this sense, Simone’s achievement is fully attributable to her executive as well as to her selective competences.⁹ However, her performance falls short of a higher epistemic standing in that the performance’s success is too fragile and insecure, as it can easily be appreciated by comparing Simone’s standing with her counterpart’s standing (Simone +) in a situation completely safe from simulation (and other defeaters). This is why Sosa places the category of secure knowledge at the highest rank of a hierarchical system of grades of knowledge (SOSA, 2021, p. 157).

The first question that arises from the previous account is that as to why Sosa adds a further, higher rank to the familiar categories of animal and full reflective knowledge - one that, besides, seems to do justice to the arguments of those epistemologists who by incorporating as conditions for knowledge the ability condition underlined by robust virtue epistemologists as well as the anti-luck condition proper of safety-based epistemologies, claim contra Sosa that knowledge is not (or, not only) a manifestation of competence.¹⁰

The answer lies in the fact that Sosa is dealing with a clash between two conflicting intuitions the strength of each, far from denying or minimizing, Sosa is realist enough to acknowledge. The task is how to accommodate the two intuitions by providing a proper place to each. There is, on the one hand, the recalcitrant intuition that the threat of simulation blocks Simone from really knowing - an intuition that aligns well with the intuition that Barney (of fake-barn cases) does not known in fake-barn territory, and that seems to support the view that knowledge is (at least in part) a matter of what happens in modally close possible worlds.¹¹ There is, on the other, the intuition that Simone is fully creditable for her success, to wit, that her success, even if insecure, is really an achievement due to a presupposition that is for her legitimate and proper to hold. The category of secure knowledge can naturally explain why Simone’s performance falls short of the higher epistemic standing intuitively enjoyed by her counterpart in modally safe surroundings, while retaining the insight that the quality of her shot (and her being creditable for gaining knowledge) is not affected in the slightest by how easily she might then have been under simulation, indiscernibly so. How is so?

The crucial point for Sosa is that though the epistemic standing of Simone might be better, her performance’s quality is not for that downgraded. It is not as if the achievement of Simone + would be better as an achievement than that of Simone, but instead that while the achievement of the latter is unsafe (fragile), that of the former is secure. Whether the achievement is safe or not does not contribute at all to the fact that it is an achievement. Secure and insecure knowledge are both equally knowledge. Safety might well be a modal property that involves counterfactuals, but that is compatible with Sosa’s robust actualism in epistemology - actualism being the position which claims that knowledge only is a matter of the actual manifestation of competences, independently of how easily those competences might have remained unactualized. Safety, which for many a year has been a nuisance for robust virtue epistemology, is thus elegantly accommodated within such a view, as a higher epistemic standing for agents who are not for that better performers. Beliefs which are secure in their aptness are not beliefs which are more apt. Simone + does not know better that she hits a real target than Simone. It is only that her knowledge is less fragile than that of the latter.

Things become, however, more complicated. Besides introducing background presuppositions within the framework of traditional virtue epistemology, and as a second original feature in his most recent account, Sosa also comes to distinguish a further class of background presuppositions that he calls domain-defining conditions (SOSA, 2021, p. 160).

To illustrate this notion, Sosa invites us to compare Simone with Fielder, a baseball player in a night game who is performing when, unbeknownst to him, the lights might too easily go out, a situation which makes of his achievement a fragile one. Do the fragile lights downgrade the quality of Fielder’s performance?

The crucial point for Sosa is that Fielder and Simone differ in that while if the lights go out, Fielder loses his ability to perform, Simone would retain her ability to perform (and to issue judgments) even when she is in the simulation cockpit, and her judgments are massively false. Contrary to what happens with Simone, the fragility of the lights does not directly affect Fielder’s achievement by making it insecure: it is his performance as such, independently of whether it is apt or inept, that is unsafe. If the lights were out, such situation would prevent him to perform at all. A domain-defining condition is thus one that is “constitutive of performing in a given domain, in such a way that performing in that domain requires satisfying that condition.” (SOSA, 2021, p. 160).

Fielder is neither negligent nor reckless in assuming as he does that the lights will stay on - a condition that must be met for him to perform. It is this assumption that underwrites his current, particular belief at the moment just previous to start making his catches that he will continue to see well enough to play (a belief which is, on the other hand, irrelevant both to Fielder’s status as a baseball player and to the quality of his catches). Domain-defining conditions are, therefore, of the same kind of background presuppositions (as a sub-class of the latter) in that they are assumptions that must be added to appearances (whether to one single appearance or to a series of congruent, successive seemings) in order to obtain evidential justification for empirical, quotidian beliefs. One might thus plausibly claim that by introducing those two notions, Sosa is advancing his own version of Hinge Epistemology - one that shares common ground with all the varieties of this recent trend in epistemology, mainly in that it also opposes the neo-Moorean proposal (as it is expounded by self-called Dogmatists and Liberals such as Huemer and Pryor) that perception or seemings suffice to grant one justification for the corresponding perceptual or inferential belief.

Let us take that Simone seems to see a real target. Given that this seeming is compatible with the fact that Simone is not really seeing a real target (maybe because she is only hallucinating one or because in the simulation cockpit what she is seeing is a fake-target), the experience by itself is not enough to warrant her belief that she is seeing a real target. It is also necessary, for that purpose, the relevant default assumption that if one seems to see a real target, and absent reasons to doubt, one does see a real target. Presuppositions of the same sort (Fielder’s assumption that the lights will stay on; Barney’s taking for granted that what appears to be a barn is really one, and so on) are thus required to confer epistemic justification to particular beliefs. Let me be clear from now that this is something I do not dispute.

The question that naturally arises at this point is as to why it is within his rights for the agent to assume that the conditions, internal as well as external, are propitious. Why is it so intuitive to claim that Fielder is within his rights to presuppose that the lights will stay on, that Simone is permitted to take for granted that she is shooting at real targets, or that Barney is properly presupposing that he is not in fake barn-country - and that, even when those presuppositions are either false or insecure?

It is my view that in the cases at hand - cases in which only local conditions are considered, so that general, background assumptions have not entered into the picture yet- there are two conditions for S to properly assume that P: (i) that the assumption’s content is such that both personal as well as collective experience teach us to expect it to be the case, namely, that the assumption is both grounded in overwhelming evidence and interrelated to many other beliefs so as to be what one might generally expect to happen if things are normal; and (ii) that the agent is relevantly sensitive to possible signs to the contrary (defeaters), at least implicitly.

Given all we know about the nature and use of barns, or about how infrequent blackouts are, it is far from arbitrary to assume that what appears to be a barn is really one, or that lights in public and well-tended venues are not prone to fail. As it is rational for Simone to assume that she is piloting a real jet fighter, given all she knows through previous experience and testimony about training routines, combat conditions, and usual scenarios. There is thus insight in how those tacitly presupposed expectations - which are crucially derived from experience- can be applied to particular beliefs as (defeasible) rules that confer on them presumptive justification. The trouble for über hinge commitments is that with regards to them such an insight is lacking, since general assumptions being not derived themselves from experience it is unintelligible how they might apply to it.

The main issue comes, however, from the fact that though epistemic agents are in their rights to assume by default that background conditions are met, and so, that though they are not negligent for ignoring to explicitly check local conditions which they are permitted to presuppose, they would be guilty of negligence if aware of (actual as well as probable) defeaters, they would opt for ignoring them, and perform instead.

Consider Simone has been privately informed of the likelihood of the simulation test. She is thus fully aware of how easily she might be placed in the simulation cockpit. However, she ignores the danger, and performs. Even though it happens that Simone is now piloting a real jet fighter and that she hits a real target, it is clear that in this scenario she has been negligent, even reckless. She is not fully creditable for attaining such success - one that is more the result of luck than due to full competence.¹²Abstention is the right attitude she should have taken, given all she knows. Norms internal to the epistemic domain, which are rules relevant to the assessment of domain-internal choices and decisions, leave no doubt as to the fact that, on pain of negligence, Simone should suspend judgment if aware of how unsafe and risky her performance is. This verdict is, of course, inherently related to the sensitivity to tale-telling signs to the contrary required for proper assumption, and thus, to the role played in epistemology by second-order monitoring on the part of the agent.

Is Fielder’s presupposition relevantly similar in this regard to Simone’s assumption? Would Fielder’s decision to continue performing when aware of how easily the lights might go out be negligent, or even reckless? Notice that the question is not whether Fielder is within his rights in bracketing as irrelevant safety conditions in view of backdrop assumptions that are presumed by default to be satisfied. It is instead as to whether Fielder should be blamed as negligent Simone is blamed - blamed for performing when he knows that he can’t continue to presume by default that circumstances are as expected. Presumably, one would be tempted to answer that as in the case of Simone, if Fielder were aware of the lights’ fragility, he should abstain for action, on pain of downgrading its quality.

The problem for the previous answer is that as in paradigm cases of benign luck (luck that does not affect the epistemic evaluation of beliefs)¹³, the danger that Fielder is facing is such as to cancel his performance if actualized. Which means that it is an external threat - one that is not internal to the framework of epistemic evaluation, and thus, that is of the nature of consequentialist factors. In view of how easily the lights might go out, Fielder has to decide whether to continue performing. His choice will be usually based, if at all, on normal cost-benefit considerations - but neither consequentialist considerations are epistemic considerations nor the external factors that Fielder may take into account to perform are factors constitutive of knowledge. Fielder’s decision may well be unwise, as it might be unwise the choice of the enquirer to attain the truth at the cost of bad health or of certain death.¹⁴ But he is within his epistemic rights to dismiss it and to bracket it away into irrelevance, even when fully aware of how easily might be actualized. In performing the agent would take a risk, but not a cognitive risk that he should not have taken for the sake of performing well. Again, factors pertaining to bare doing are not factors pertaining to how well or badly one is performing within a normative domain of action.

The interesting thing is that Fielder’s scenario provides us with a clear case of lack of entailment between the question as to the rationality of the agent’s presupposition and the question as to whether such assumption impinges into the quality of actions in general and of epistemic performances (affirmations) in particular. As it has been argued, while Fielder is rationally entitled to assume by default that the lights will stay on, nothing follows from that (adequate) assumption regarding the quality of his catches. The same gap might be projected to general, background assumptions. Could it be possible that we were entitled to presuppose the denials of global skeptical scenarios, while those assumptions did not contribute at all to the aptness of our empirical beliefs? Could it be that general assumptions were external factors that although must be satisfied for performance to take place, the agent can properly ignore (and their denial properly dismiss) without epistemic negligence?

Sosa’s distinction between background presuppositions and domain-defining conditions open up the logical space for raising the question as to whether general assumptions are of the class of background presuppositions included in the normativity of knowledge, or, on the contrary, they are of the class of domain-defining conditions which are not domain-internal, even though, if the latter option is correct, they would not be, as in the case of Fielder, local conditions. To make it possible to raise this question is by itself a portentous achievement.

We have thus to face two issues. On the one hand, the question as to whether general, background assumptions are rational and nonarbitrary - a problem that seems especially intractable since being underived from experience, those assumptions seem to hang in the air, as visceral commitments that are irrational. On the other, the problem as to whether they have a role to play within our quotidian, epistemic practices - a question to which our lack of insight into how they might be rules realized in examples adds pressure.

It has been usual among hinge epistemologists to ground the rationality of über hinge commitments in the supposed fact that they are applied to empirical beliefs as a matter of course - applied as conferring default justification on the latter. However, those assumptions might only be mere forms of thought without determinate realizations. As it will be argued, there is something of a paradox in the thesis that the rationality of hinges is supported by the fact that they lie inside the scope of the basic epistemic practice of producing, assessing and withdrawing from empirical beliefs - in that such thesis plays directly in the hands of the radical form of skepticism which it opposes. By contrast, and apart from being intrinsically convincing, Sosa’s way of gaining rationality for general assumptions has the further advantage of accommodating Humean skepticism without having for that to yield to its more disturbing consequences. On Sosa’s view, for hinges to be rational they cannot be epistemically relevant as factors on which nonnegligent performance depends. As in the case of domain-defining conditions, their function is deeper and much more fundamental.

2 The skeptical challenge and how to deal with it: a lesson from Sosa

There are two aspects of radical skepticism that are as generally ignored as they are crucial for the correct understanding of the skeptical challenge. Firstly, that those global scenarios familiar from skeptical literature - from the dreaming argument to BIVs- differ from local unpropitious conditions of performance that are checkable in that their actuality would be unnoticeable not only to the victim but to every possible evaluator within the empirical domain. It is in this sense that global scenarios are metaphysical scenarios - in the sense that there is no standpoint from which to discern whether they are actualized or not; in that they are compatible with the whole of our experience. Second, that far from being concerned with whether one knows that one knows that p, the sceptic is concerned with whether one is in his rights to claim that p. The sceptic’s point is thus, and borrowing from Sosa’s metaphor, that although the archer may in fact hit the target of truth - and hit it aptly- , he is forbidden in view of the possibility of global scenarios to claim to have hit it. Skepticism is thus a variety of bi-level epistemology - one that opposes externalist accounts of what it would be for a reflective agent to know.¹⁵

It thus follows from the global nature of radical skepticism that there is no sign that the agent might exploit to tell whether he is or whether he is not within a global scenario. As a consequence, agents would be constitutively blind to their overarching condition. From the skeptical perspective, one might thus claim that it would suffice with our becoming aware of the unnoticeable character of global scenarios to raise doubts on whether the reflective agent would retain her rational right to take hinges for granted in the face of those possibilities, which are, as it were, reflectively present to the conscientious performer without requiring any empirical sign to the contrary. Hinge epistemologists such as Wright (2004, p. 53) have been prone to argue that one is entitled to presume by default (absent signs to the contrary) that one’s basic commitments are satisfied. However, the point of the sceptic is that such a default justification only is right in the view of local defeaters, to wit, for those cases where it would make sense to think of the possibility of alerting signs. Absent those possible signs of defeaters, one’s trust in hinges is visceral and arbitrary.

However, the radical sceptic is neither arguing that we really are inside a global scenario nor even suggesting that global scenarios are too modally close for comfort. For all he knows, the sceptic can be happy enough to agree with us in that it is much more probably (or almost certain) that we are not massively deceived than that we are. For the sceptic, the agent would be deprived of knowledge even if the latter were not (and even if it were not easy for him to be) within a global scenario. Since on the sceptic’s view agents are deprived of a selective, second-order competence to take global defeaters into account, they only would be able to enjoy animal knowledge, and thus, to possess knowledge without having for that a second-order grasp that her animal affirmation would be apt. On this view, the aptness of our beliefs is not achieved though the guidance of the agent’s second-order general presuppositions.

For hinge epistemologists, general assumptions are factors within the scope of the normativity pertaining to quotidian, epistemic practices. If so, they are second-order presuppositions that guide our beliefs to aptness. The problem is, on the one hand, that rooted in no distinctive competence, über hinges are not able to perform their supposed function; and on the other, that if the sceptics are right (as they appear to be) in that victims of global scenarios (alike local victims such as Simone-inside-the-simulation cockpit) would retain their capacity to make genuine, though massively false judgments, then epistemic practices would stand as they are independently of whether hinges are true as well as of whether one (as a good sceptic) refuses to endorse those assumptions at the second order.¹⁶ This is why in my view through the epistemic role that hinge epistemologists provide to über hinges they really pave a high road for Humean sceptics to target their rationality.

The question raised by the sceptic as to whether general assumptions are arbitrary revolves around the question as to whether agents are epistemically negligent by default. One possible way out of the net of normative scepticism is, however, by taking advantage of the lessons learnt from radical, descriptive scepticism. The epistemologist could thus adopt an escape route of sorts as follows. He might concede to the sceptic that it is impossible to rule out global scenarios, while pointing out that actions and projects that agents cannot perform do not impose duties on us. As a consequence, agents would not be guilty of negligence for bracketing away as irrelevant (to their duties as epistemic performers) sceptical scenarios. On this view, the relation between descriptive and normative scepticism would be undermined, and quotidian agents would be within their rights to continue their practices unperturbed by metaphysical considerations, irrespective of whether global scenarios are modally close or even true.

This is, in its results (epistemic practices take, as it were, care of themselves), the kind of approach that I would favour. It needs, however, to be improved. It is not so much that there is a bitter aftertaste of defeat in approaches where radical scepticism is not neatly refuted but that general assumptions are still left hanging in the air, as arbitrarily presupposed as before. This is why it is in my view necessary to proceed on this line of thought towards the model provided by Fielder. On the latter model, agents are not negligent for dismissing easy possibilities that would cancel their performance if actualized while being rationally permitted to presume by default that conditions are propitious to perform. It is clear, however, that for Fielder’s model to be projected to general assumptions, their rationality must be regained in an a priori way. Besides, such procedure should be able to support the claim that hinges are excluded from quotidian, epistemic normativity as well as to explain why it is so. Is it such a rational re-appropriation of hinges feasible?

As Sosa has recently come to argue (SOSA, 2021, p. 166-171), a positive answer depends on whether, as it is taken for granted by radical sceptics, the victim of global scenarios retains her capacity to act and to think, and can thus be seriously be conceived as a cognizer. By raising doubts on this prima facie intuitive presupposition, Sosa has come to conclude that general assumptions are in the relevant sense of the family of domain-defining conditions, if only nonlocal and constitutive of the architecture of basic rationality.

Take, for instance, the dreaming argument. It is commonly assumed that if one were dreaming one would still be forming beliefs and making judgments - those very beliefs and judgments which one is dreaming of. However, all of the actions, decisions, and willful acts that happen in the dream, and because they are dreamt of, are cancelled out as real, that is, as things that one does, or that are really happening to one, while one is dreaming. There is no doubt that thoughts, understood as mental imagery that crops up in the mind of the dreamer for any number of reasons unrelated to the dreamer’s cognitive achievements, are attributable to the dreamer while he dreams of them. However, in dreaming there is no real thinking, no genuine act of deliberation, affirmation or even doubt on the part of an agent. The dreamer dreams of himself as thinking and acting, but he is only representing himself as thinking and acting without really being a cognizer and a performer. It is, therefore, as if the sceptic were adopting a passive and purely representational view of the human life and condition. The main objection to the sceptic is that he cannot really conceive of the dreaming world whose possibility he supposes, because he cannot conceive of himself as a thinker and a doubter in such a world. The sceptic is thus guilty of a performative contradiction of sorts.

Since the previous considerations can be applied to all global scenarios, Sosa’s master argument in favor of the rationality of hinges comes as a result to disclose the hidden nature of radical skepticism - in that the skeptical attack to the rationality of hinges is a self-undermining attack to subjectivity and agency, so that what the sceptic really does is suggesting that the feeling of ourselves as living agents is nothing more than an illusion. This is the core of Sosa’s new, extended Cogito, which might either be expressed like in early versions (SOSA, 2007, p. 20) as “I think, therefore I am awake”, or to assume a wider scope as in “I think, therefore hinges are true”. Notice, however, that the argument does neither increase the likelihood of the truth of hinges nor it rules out as logically or metaphysically impossible global scenarios. What it concludes is that since the sceptic cannot challenge as arbitrary our trust in hinges without challenging for that reason as arbitrary the very conditions that make that challenge possible, we are as reflective performers within our rights to assume that hinges are as certain as agency is. As a matter of fact, Sosa’s argument helps making it visible how deeply, logically interrelated are the subjective and the objective dimensions of human experience, thus contributing to the issue of providing closure for rationality without excluding the openness of empirical experience.

For the purpose of the present discussion, the crucial point in Sosa’s argument is that it brings out that global scenarios would cancel doings if actualized - and that, regardless of whether those doings are apt or inept. Trust in general assumptions is trust in that we are really acting - trust which neither says anything about the epistemic quality of our performance nor guide the agent to proper performance. If, on the one hand, global scenarios are not credit-reducing factors that the agent must consider to avoid falling epistemically short, it has been argued, on the other, that it does not follow how well one is performing from the assurance of one’s being performing. It is thus clear that once the epistemic construal of hinges is shown to be of the family of category mistakes, the ‘blanketing’ nature of a priori arguments to which so accurately Sosa called our attention in his discussion of Davidson’s epistemology ceases to be a problem - it ceases to be the Trojan horse through which the inroads of organized skepticism may ravage the citadel of common sense. General assumptions are marks of action - not (right or wrong) second-order ways of grasping the triple-S profile of a given performance.

Hinge epistemologists have been concerned to show how the metaphysical principles of ordinary practices of judgment are transcendentally grounded. It is my view that Sosa’s main contribution to Hinge Epistemology lies in the fact that he has convincingly argued that far from governing how we (must) act, those metaphysical principles are bound up with action itself, with the facts that we act and that acting is attributable to us. As such, it is as if our ordinary practices of judgment were permeated from their roots by the firmest and more permanent commitment to objectivity and externality - one that neither ordinary norms of epistemic assessment nor congruent experience are on their own able to provide. The question mark has been put, at last, deep enough down.¹⁷

References

• BONJOUR, L., SOSA, E. Epistemic Justification: Internalism vs. Externalism, Foundations vs Virtues. Oxford: Blackwell, 2003.
• COLIVA, A. Extended Rationality: A Hinge Epistemology. Basingstoke: Palgrave Macmillan. Doi: 10.1057/9781137501899, 2015.
  » https://doi.org/10.1057/9781137501899
• HADDOCK, A., MILLAR, A.; PRITCHARD, D. The Nature and Value of Knowledge: Three Investigations. Oxford: Oxford University Press, 2010.
• PRITCHARD, D. Faith and Reason, Royal Institute of Philosophy Supplement, 81, p. 101-118. Doi: 10.1017/S135824611700025X, 2017.
  » https://doi.org/10.1017/S135824611700025X
• RADFORD, C. Knowledge- By Examples, Analysis, 27 (1), p. 1-11, 1966.
• SOSA, E. A Virtue Epistemology: Apt Belief and Reflective Knowledge, Volume I. Oxford: Clarendon Press, 2007.
• SOSA, E. Reflective Knowledge: Apt Belief and Reflective Knowledge, Volume II. Oxford: Clarendon Press , 2009.
• SOSA, E. Judgment and Agency. Oxford: Oxford University Press . Doi: 10.1080/00048402.2015., 2015.
  » https://doi.org/10.1080/00048402.2015
• SOSA, E. Epistemic Explanations: A Theory of Telic Normativity, and What It Explains. Oxford: Oxford University Press , 2021.
• WITTGENSTEIN, L. On Certainty. Oxford: Blackwell. Doi: 10.2307/2217581, 1969.
  » https://doi.org/10.2307/2217581
• WRIGHT, C. Wittgensteinian Certainties. In D. McManus (ed.), Wittgenstein and Scepticism. London and New York: Routledge, p. 87-116, 2004.

• 2
  This point has been explicitly stressed by Sosa in his recent comparison between Moore’s defence of common sense and Wittgenstein’s late epistemological remarks. One of the lessons that Sosa claims we might learn from Wittgenstein’s critical comments on Moore’s epistemological task is that while “Moore is right to defend his common sense, including its core component beliefs”, he is “wrong to force our knowledge into the foundationalist framework of the tradition, with its intimate evidence epistemically supporting our body of beliefs.” (SOSA, 2021, p. 199). [Any reference to page numbers of Epistemic Explanations (2021) corresponds to the forthcoming, final manuscript of the book, which will be released in July 2021] Moore’s approach would thus essentially depend on self-presenting states, as one more instantiation of the myth of the given.
• 3
  Type-III hinge propositions are predecessors both of those very general and fundamental assumptions that for Annalisa Coliva are constitutive of epistemic rationality (COLIVA, 2015, p. 128), and of the fundamental certainty that Duncan Pritchard (2017, p. 111) calls the über hinge commitment that we are not radically in error - one that is the underlying core common to all our epistemic commitments.
• 4
  Wittgenstein himself remarks that for all the empirical beliefs that Moore takes to constitute certain knowledge it is always possible to conceive particular contexts in which quotidian doubt would be legitimate, as in Wittgenstein, 1969/2004, § 622: “For each one of these sentences I can imagine circumstances that turn it into a move in one of our language-games, and by that it loses everything that is philosophically astonishing.”
• 5
  Note that I am far from endorsing such finitist conception of rationality. It is only that defactoist rationality - rationality that depends on how human beings are contingently constituted- is not reducible to cultural norms and transient forms of life.
• 6
  This ‘blanketing’ nature of über hinge commitments might be made salient by noticing that from this standpoint the clairvoyant beliefs that BonJour’s Amanda arrives at thanks to her extraordinary ability (BONJOUR & SOSA, 2003, p. 28-30), the set of beliefs formed by the agent through imbibing cultural prejudices, and the perceptual beliefs formed by normal perceivers, would all of them enjoy the same presumptive and default justification. I think that this should give us pause.
• 7
  Where a factor (f) would contribute to justify p by helping to increase the likelihood of p being true, as a reason in favour of p.
• 8
  As it was developed in Sosa (2015, p. 146-153). In this first approach, Sosa argued that while Simone enjoys animal knowledge of her surroundings, she seems to fall short of full knowledge, since she is not in a position to know that if she affirmed perceptually, she would affirm aptly.
• 9
  In Sosa’s terminology, executive competences include first-order powers such as perception, memory, and the ability to subitize; powers which are able to operate - and to operate reliably- by their own, independently of rational evaluation (whether deliberate and thoughtful or merely implicit). Selective competences involve, by contrast, the performer’s being rationally and agentially sensitive to the current triple-S profile (seat, shape, situation) of a given epistemic performance.
• 10
  This claim is mainly associated with proponents of Anti-luck Virtue Epistemology such as Duncan Pritchard and Jesper Kallestrup. On this account, competences and their manifestations as apt believing are not sufficient for the truth of the corresponding attribution of knowledge. To this end, a safety clause has to be added to success due to competence. This results in a dual-condition account, as Pritchard has been prone to stress (HADDOCK, MILLAR & PRITCHARD, 2010, 54).
• 11
  Notice that the cases of Simone and Barney are of the same kind - both the two victims of unpropitious environments are full competent performers who through second-order presupposition aptly gain apt beliefs, though the aptness of their respective beliefs is not secured through the aptness of second-order selective competence.
• 12
  Given that reckless Simone is not being guided to aptness through second-order sensitivity to her surroundings, it is clear that even though she hits the target of truth, she falls short of full reflective knowledge. One would, however, be tempted to attribute animal knowledge to Simone. After all, her executive competences are manifested in her belief’s success. The problem for the latter attribution comes from the fact that it is difficult to make sense of negligent Simone as holding the belief that she is hitting a real target, even if she is hitting it. How can Simone be confident enough to hold that particular belief when she is aware of the likelihood of the simulation test? Maybe it would be relevant here to mention the non-doxastic account of knowledge, which argues that knowledge does not require belief. For instance, Radford (1966) argues that a student that writes down the correct answer unconfidently in an exam can still be recognized as knowing the correct answer - although the student lacks the relevant belief. On this view, while negligent Simone would fall short even of credal animal knowledge, she would still enjoy of subcredal animal knowledge. It would seem that Sosa makes room for the latter category when considering guessing (SOSA, 2015, p. 74-6). If so, it appears to be within the logical space of epistemic categories some position for alethic affirmations which are not beliefs.
• 13
  For example, the fact that S miraculously escaped a fatal accident that would have deprived her of her visual competence is irrelevant to assess how competent she is in forming her current perceptual beliefs by means of her sight. As it is irrelevant to the epistemic quality of her beliefs whether or not she was prone to suffer a heart attack. Maybe it only was by chance that the agent was able to perform, but that does not inject epistemic value to her performance.
• 14
  Sosa discusses in detail this issue in his response to Character Theory in general, and to Baehr’s proposal in particular. See Sosa (2015, p. 34-61).
• 15
  As Sosa has argued in different contexts. It is relevant that although his proposal is not sceptical in any plausible meaning of the word, Sosa has come to underline the Pyrrhonian sources of his thought - mainly in the second-order nature of Pyrrhonian scepticism, and in the project of accommodating the phenomenon of suspension within a complete epistemology. See among many others Sosa (2015, p. 215-232).
• 16
  Let me add that views such as the one proposed by Coliva (2015) are explanatory accounts of our epistemic practices which, in effect, and by means of a de facto transcendental deduction of hinges, rule out any explanation of epistemic practices that entails scepticism about those practices. However, one would be careful, or so I believe, to assume a transcendental principle for the explanation of a phenomenon, when the latter may be explained from other principles, such as congruent experience, which are more familiar. In this regard, it appears as if hinge epistemologists would be likely to confuse the concepts of objectivity and externality with the use of those concepts in quotidian practices. Empirical beliefs do not gain more epistemic weight by appealing to the existence of the external world. There is a point in practice from which a further reason would always be one reason too many. One cannot thus infer the epistemic nature of a concept from its daily uses in ordinary language. On the view I favour, there is a gap between the linguistic uses of concepts such as externality and their metaphysical nature.
• 17
  I am deeply grateful to Ernest Sosa for helpful comments on previous versions of this paper. Thanks also to Guido Tana for many insightful observations, as well as fruitful discussion and philosophical wisdom.

This article has been funded by the following projects:

• -«Epistemology of Religious Belief: Wittgenstein, Grammar and the Contemporary World» (PTDC/FER-FIL/32203/2017. Portuguese Foundation for Science and Technology-FCT).
• -«Vulnerabilidad cognitiva, verosimilitud y verdad» (FFI2017-84826-P. Ministerio de Economía y Competitividad. Programa Estatal de Fomento de la Investigación Científica y Técnica de Excelencia. Gobierno de España).
• -«Perspectivas personales. Conceptos y aplicaciones» (RTI2018-098254-B-100. Ministerio de Ciencia, Innovación y Universidades. Programa Estatal de I+D+i Orientada a los Retos de la Sociedad. Gobierno de España).

Publication Dates

• Publication in this collection
  28 June 2021
• Date of issue
  06 2021

History

• Received
  04 Sept 2020
• Accepted
  05 Mar 2021