Allegedly, the greatest accomplishment of Wittgenstein’s On Certainty (OC) (1969) is to have developed an account of the relationship between knowledge and certainty that puts paid to sceptical challenges inspired by the limits of knowledge and justification. While this view is roughly shared by many philosophers, all consensus breaks on the specific way in which Wittgenstein’s epistemology deals with scepticism (Schönbaumsfeld 2016). The focus of this dispute are Wittgenstein’s basic certainties, sometimes called “hinges” or “hinge certainties”, a metaphoric expression used to describe a particular kind of propositions:

“[T]he questions that we raise and our doubts depend on the fact that some propositions are exempt from doubt, are as it were like hinges on which those turn”(OC 341).

Wittgenstein characterises the so-called “hinge propositions” in different ways. OC identifies them as the ground of our language-game, the foundation of our beliefs (OC 253), the propositions that stand fast for me (OC 152) and the background against which all distinctions between true and false are made (OC 94). Whether such propositions constitute genuine certainties or not, Wittgenstein’s language makes constant use of terms that suggest that they are such a thing.

If hinges are certainties, OC shows that they are objective, rather than subjective. Wittgenstein rules out the possibility of being mistaken about them (OC 194). The only way in which a mistake can be made is to judge “in conformity with mankind” (OC 156), that is, within a community of speakers who agrees on some basic rules. Any individual who takes part in the language-games of this community and is familiar with its rules knows what this means. Except for cases of memory loss, brain damage, or more unusual situations a person cannot fail to remember that he has always had five fingers (OC 157). And because familiarity with these rules is a basic requisite for engaging in certain practices, Wittgenstein believes that there are occasions in which “a man cannot make a mistake” (OC 155).

By furnishing knowledge with a solid, unassailable frame of reference, hinges rest at the very foundation of our cognitive scaffolding. On the face of it, any ordinary empirical proposition is supported by a prior and indisputable certainty that gives it meaning. Hinges are so ingrained in our belief system that their very existence can deflate any sceptical scenario. For instance, any attempt to introduce a Cartesian-like doubt of the kind of whether I am or not a brain-in-a-vat (BIV) will necessarily be underwritten by a limited set of basic and fundamental assumptions that provide this doubt with meaning and context. Cartesian doubts can only fly if these assumptions are taken for true. Thus, the context in which the BIV hypothesis is embedded assumes that there is a world, that this world has creatures in it, that some of these creatures have brains, that under very special circumstances, these brains can be manipulated, that if a mad scientist succeeds at manipulating them, these brains can conjure up highly deceptive scenarios that can potentially replace reality. Such assumptions inevitably precede any sceptical consideration, so that even radical scepticism cannot fly without them. As a result, sceptical doubts cannot be seriously entertained by any subject without her implicit commitment to some underlying hinge. As it turns out, then, hinges appear to be semantically constitutive:

“If you are not certain of any fact, you cannot be certain of the meaning of your words either” (OC 114).

Likewise, Wittgenstein wants to show that hinges can also handle universal doubt:

“A doubt that doubted everything would not be a doubt” (OC 450).

Wittgenstein understands that knowledge is ultimately parasitic on certainty, although not on any kind of certainty, but on the remarkable certainties embodied by one’s hinges. So by questioning those certainties one would end up questioning the very exercise of questioning in which one is engaged.

Because of this, Wittgenstein is keenly aware that hinges are not ordinary propositions, and that believing that they are so reveals, in his view, a grievous misunderstanding of the rules of the language-game in the context in which someone may invoke them.

This article is divided in seven sections. In the first section, I argue that unlike empirical propositions, hinges cannot be known. This involves that they are not backed up by evidence and are hence unjustifiable in the context of the justified true belief (JTB) framework. In this respect, R. Rhees suggested that they are dispositions that lie implicit in many of our thoughts and do so nonpropositionally. In the second section, I discuss Moyal-Sharrock’s dispositional theory, which portrays hinges as nonpropositional certainties that may be turned into propositions in some contexts. In the third section, I analyse two controversial aspects of Moyal-Sharrock’s theory, the most central of which is the groundlessness of certainties. These problems are also central to Pritchard’s account of hinges, that is part and parcel of his so-called “minimal interpretation” of OC. I discuss this interpretation in section four. Section five argues that the minimal interpretation, which rests on the groundlessness interpretation, finds little support in Wittgenstein’s texts and renders the main argument of OC problematic. Section six puts forward a viable solution to the minimal interpretation by stressing that hinges are world-pictures (Weltbilde) that both reflect reality and are consistent with knowledge and one’s overall evidence. Finally, section seven draws some conclusions about how hinges underpin cognition.

1 On Hinge Certainties

What kind of propositions are hinges? Are they empirical or logical? And if so, are they epistemic? There is a consensus that hinges are not a priori. Insofar as they describe empirical truths that do not appeal to any special or restricted set of experiences, we may assume that whatever their content is, they result from ordinary experience. Although Wittgenstein’s language may suggest that they can be logical or a priori principles, this impression is deceptive. Hinges result from our cognitive engagement with reality, whether directly—in the way of perceptual inputs—, indirectly, that is, by second-order analysis of perceptual inputs, or by testimony. So, if they arise from experience, could they be considered knowledge-providing veridical propositions under the JTB framework?

The answer to this question is negative. OC emphatically remarks that, despite Moore’s intention to disprove scepticism, Moore’s case against the sceptic fails to produce any knowledge of his purported certainties. Propositions like “I know that this is a hand”, which Moore holds up to be more evident than the sceptical argument, do not work in the way Moore expects them to do. The content of “I know that this is a hand” is not all the more warranted by the fact that one claims to be certain of it. As Wittgenstein puts it:

“How do I know that it is my hand? Do I even here know exactly what it means to say it is my hand? —When I say “how do I know?” I do not mean that I have the least doubt of it. What we have here is a foundation for all my action. But it seems to me that it is wrongly expressed by the words “I know” (OC 414).

On Wittgenstein’s view, the words “I know” can only be legitimately used in contexts “when one is ready to give compelling grounds. “I know” relates to a possibility of demonstrating the truth” (OC 243). We use knowledge-claims in contexts in which reasons for a particular claim can be accepted or rejected. In situations in which error can automatically be excluded because the context of the assertion is clear, using the words “I know” to claim certainty only increases the difficulty of understanding how knowledge can produce such certainty. In fact, in most cases a person does not claim to know what she takes to be most certain. If she were to do so, one should treat this ascription like any other. But as Wittgenstein contends, Moore’s propositions are not standard claims. He writes:

“If a blind man were to ask me “Have you got two hands?” I should not make sure by looking. If I were to have any doubt of it, then I don’t know why I should trust my eyes. For why shouldn’t I test my eyes by looking to find out whether I see my two hands? What should be tested by what?” (OC 125).

“Giving grounds, however, justifying the evidence, comes to an end; —but the end is not certain propositions’ striking us immediately as true” (OC 204).

If hinges appear to be unjustifiable, how are they evidentially grounded? How can they represent reality? At first glance, we may be tempted to treat hinges as self-justifying “basic beliefs” or “über propositions”: the non-inferential beliefs that provide a basis for other beliefs in a belief system. Yet this is not their role in our cognitive system. It is not even clear whether they may even be categorised as beliefs, that is, as “propositional attitudes” in the standard sense. More on this will follow soon. What concerns the possibility that hinges are self-justifying claims, the answer is negative. Although the images used in OC to describe hinges (“the grounds of our language-game”, “the foundation of all our assumptions”, “the background against which all distinctions between true and false are made”, “a world-picture” (OC 167)) may suggest that they are self-justifying, the book stresses that if their certainty “stem, or could stem, from justification, it is not a hinge certainty” (OC 94, Moyal-Sharrock 2016, p. 26). More surprisingly, the book contends that hinge-certainties are neither true nor false, neither justified nor unjustified. If so, these propositions should not be expected to give us the unequivocal evidence our belief system requires to be warranted. What is more, the fact that hinges stand outside the framework of JTB epistemology rules out the possibility that Wittgenstein may have developed a foundationalist theory (Williams 2005; Glock 2016).

Of course, the ambiguity of many of Wittgenstein’s intuitions has inspired different hypotheses about hinges. Take e.g., Rhees’ interpretation. Rhees argues that although hinges are empirical and can be re-evaluated on the face of contrary evidence, most hinges stay largely unformulated in one’s mind (Rhees 2003, p. 70). It is awkward to assert such basic claims. The resistance of hinges to take propositional form is a key insight that I will elaborate in the next sections. Rhees believes hinges to be dispositions to cognitively engage the world in a certain way. Unlike most of what is usually said, an individual may have never propositionally expressed or couched them using vocabulary. This ambiguity does not seem to affect their singular status. To test the hypothesis that hinges are dispositions rather than propositions, Wittgenstein asks himself: how do we know that “there is no stairway in this house going six floors deep into the earth, even though I have never thought about it?” (OC 398). The fact that this building does not have six underground floors never crossed my mind. Because of it, I have no evidence to back it up. So how do I know it? If I never contemplated this idea and may have never voiced it, how could I disprove a contrary claim? All I can say in this respect is that my presumption that this building is not so deep always stood fast for me, that is, that it is unquestionable: my observation does not look like the evidence of such a belief. Likewise, the idea that

“the people who gave themselves out as my parents were really my parents” (…) “may never have been expressed; even the thought that it was so, never thought” (OC 159).

On such basis, sceptical hypotheses that capitalise on the mere possibility of our having been born to different parents barely deserve credit, not because they are unacceptable, but because nothing indicates that they can be true. Accordingly, it would be unreasonable for me to give such taunts any credit.

Likewise, Wittgenstein observes that children are not taught that the objects they find around them are real. He writes:

“Children do not learn that books exist, that armchairs, exist, etc., etc., – they learn to fetch books, sit in armchairs, etc., etc.” (OC 476).

As a result, while the basic convictions that buttress our belief system in a general way do seem to have a special status, this status is not propositional.

2 The Dispositional Hinge Theory

The idea that hinge certainties are not propositions, but singular attitudes was first proposed by Moyal-Sharrock (2005). She considers the nonpropositionality of hinge certainties as the most ground-breaking elucidation of OC (2016, p. 32). She points to two basic features of hinges: one is that they are so basic, so “anchored in all my questions and answers” (OC 103), that they are part of the internal scaffolding of thought. The other is that inasmuch as they play a grammatical role, they are not bipolar. As Wittgenstein remarks, “the end is not certain propositions’ striking us immediately as true” (OC 204). So long as hinges are neither true nor false they fail to exhibit any propositional content. By implication, Moyal-Sharrock believes them to be “ungrounded, nonpropositional, and not the underpinning of knowledge, rather than its object” (Moyal-Sharrock 2016, p. 25).

What can be made of this interpretation? Of course, Wittgenstein did not draw any explicit distinction between propositional and nonpropositional certainties. And yet, the distinction may be said to remain implicit in OC. Although he considers propositions like “I have a hand”, “I have a brain”, “the world existed long before my birth”, and other similar assertions paradigmatic hinges, he does not think that they are standard propositional beliefs. Following Rhees, they can be seen as unarticulated, largely unspoken assumptions that sit in the back of our mind. In most contexts, it is either nonsensical or plainly embarrassing to voice them. Because of this, it seems more sensible to treat these certainties as implicit, rather than as explicit content. Wittgenstein appears to be aware of this particular issue as he seems to acknowledge the relativity of the concept of proposition in the “propositions of logic”—as he also describes hinges:

“Here one must, I believe, remember that the concept ‘proposition’ itself is not a sharp one” (OC 320).

By stressing that there is not a sharp boundary between the so-called “propositions of logic” and empirical propositions, Wittgenstein reminds us that the basic certainties underpinning thought are not standard propositions. A hinge can display the external features of a belief, but this does not mean that it is in any way propositional. Whether a subject goes on to express her certainty or not—as Moore arguably went on to do with mixed success, its primary status is dispositional, and OC marshals several arguments why we should consider them to be so. I regard this intuition as a critical insight to correctly identify the epistemic role of hinges in our cognitive system.

If this is true, that is, if Wittgenstein saw the dispositional character of hinges as central to the argument of OC, why didn’t he make this idea explicit? As per Moyal-Sharrock, Wittgenstein might have been unaware of it when he started the book, although of course, this is something that we will never know.

Moyal-Sharrock distinguishes two kinds of ‘certainties’ (she actually calls them “objective certainties”, in opposition to their subjective side, which her argument does not address):

  1. (a)

    a doxastic category, whose status in our belief system is foundational or basic, and.

  2. (b)

    a doxastic attitude, whose objects are foundational and non-propositional (2005, p. 79).

As a doxastic category, a hinge proposition reflects a foundational truth. Externally, the proposition exhibits the external features of an empirical proposition including bipolarity. It can be analysed, be found lacking in evidential support, challenged and reformulated to meet new epistemic demands. If at all feasible, one may try to justify it, even though doing so would inevitably draw the kind of criticism that Wittgenstein levelled at Moore on his attempt to justify his self-evident propositions (OC 151). In contrast, Moyal-Sharrock describes the doxastic attitude as phenomenological, that is, “as a certainty or sureness; an assurance; a conviction; a being sure; a trust” (2005, p. 81). She portrays it as a kind of “animal” or primitive trust that could perhaps be best described as a know-how.

As doxastic attitudes, Moyal-Sharrock sees hinges as dispositions to act in certain ways. In her view, they are best characterised as certainties that show themselves in what one says and does (2005, p. 81). Because they are so entrenched in the fabric of practical reason, I act in the certainty that I cannot be mistaken about them. So for grammatical reasons, that is, reasons that are internal to our cognitive scaffolding, I cannot be mistaken in identifying what a hinge lays out. In contrast, the so-called propositional Doppelgänger of hinges that is expressed by its doxastic counterpart can be doubted, become true or false, justified or unjustified as any ordinary proposition.

On my view, the distinction between doxastic categories and attitudes is critical to determining what Wittgenstein’s hinges are supposed to be and deserves credit. Given the shifting roles that hinges often seem to play, the distinction between (a) and (b) makes Wittgenstein’s position more consistent, highlighting the contrast between Moore’s propositions and Wittgenstein’s certainties. If our cognitive scaffolding cannot be said to rest on the safety of certain propositional beliefs, but on more basic dispositions, Wittgenstein will succeed in moving the conversation out the context in which antisceptical claims are made. With this, he will force his opponent to consider how hinges can be liable to Agrippa’s trilemma, for the trilemma concerns fully-articulated propositional beliefs.

3 Ungrounded Certainties

Despite Moyal-Sharrock’s view of hinges as doxastic attitudes is compelling, it is not unproblematic. The problems involved in her interpretation are at least two.

First, there appear to be tensions between the phenomenological character of hinges and the fact that they remain largely unarticulated in one’s mind. If so, how can be they felt? One could assume that hinges can only be felt when expressed in one’s behaviour. But even when they get expressed, what would their phenomenology be? How are i.e. certainties concerning general properties of nature like the ontological stability of objects—the fact that chairs do not suddenly disappear only to spring again into existence, and certainties about the regularities of nature felt? Certainties like these exhibit no particular phenomenology. I take this to be an essential commitment of the dispositional theory. Therefore, it is unlikely that there is something that “it is like being basically certain” (2016, p. 33), or that each hinge has its unique self-distinguishing phenomenology. Other Wittgenstein’s commentators have made arguments to the same effectFootnote 1.

Second, and in line with Wright’s (2004) and Pritchard’s (2005) view, Moyal-Sharrock’s (2016) analysis seems to be guided by the idea that hinges are the “ungrounded ground” (p. 38) of knowledge. This expression is used to indicate that hinges provide some foundation for knowledge without being themselves epistemically warranted. As argued, Wittgenstein’s insistence that hinge propositions are not bipolar and stand outside the domain of justification seems to force on us the belief that they are outside the categorial framework of JTB epistemology and that believing hinges to be so would misrepresent them.

The “groundlessness” interpretation is the dominant view in the hinge literature. This view describes hinges as “animal” (OC 359): as some form of practical commitment to the way things are. Some commentators also portray them as reflex-like, instinct-like, impulses that dictate our beliefs and shape our behaviour in the context of human action. Moyal-Sharrock emphasises this dimension by arguing that they “are really animal or unreflective ways of acting”. Being a by-product of our practical engagement with the world in a larger epistemic story, they arise because of causes rather than because of reasons (2016, p. 37, ft. 16).

Another way to describe them is to say that they embody “a (…) ‘primitive’ trust (OC 475)” (Moyal-Sharrock 2016, p. 35) that is sanguinely interpreted

“not [as] an epistemic lack, failing or limitation—not something that would allow the sceptic to triumph—but, on the contrary, descriptive of something so robust, so anchored, so unquestionable that it is not built on the back of justification or grounds” (Moyal-Sharrock 2016, pp. 36–37).

They are then the enacted certainties that inform an agent’s behaviour in ways that she may have never been entirely conscious of. Accordingly, rather than acting in awareness of the fact that I have a body, I am said to act in the full certainty of having a body (Moyal-Sharrock 2016, p. 32).

The above picture presents us with a stark choice: either to embrace maximally certain but ultimately unjustifiable certainties, on the one hand, or make them the non-inferential beliefs that provide a foundational basis for knowledge, on the other, with no middle way. For reasons already examined, the second option is incompatible with Wittgenstein’s view. Accordingly, we are forced to bite the bullet and stick with the idea that hinges are orthogonal to the categorial framework of JTB and at variance with traditional epistemology, representing, as it were, a sui generis epistemological project with little or no connection with today’s epistemology. Because this position assumes that hinges cannot be rationally warranted, Moyal-Sharrock concludes that they are “arational” (Moyal-Sharrock 2016, p. 40). In other words, their content may be parsed or analysed, but we should not expect this analysis to lead to better or more rational hinges because the mechanism by which they are acquired is an arational one. In this respect, they fall outside the space of reasons.

While the groundlessness interpretation may sensibly capture the idea that hinges are not standard knowledge and enjoys the benefit of being dominant, I do not think that it provides a plausible solution to the challenges raised by hinges. Nor does Moyal-Sharrock’s “logical pragmatismFootnote 2” solves them for reasons that I will spell out in Sect. 6.

4 The “Minimal Interpretation”

Pritchard argues that the groundlessness interpretation emerges as a result of Wittgenstein’s criticism of Moore, particularly, of Moore’s use of hinge propositions as if they were straightforwardly empirical. We examined these criticisms in Sect. 1. But claiming that optimal certainties like our having two hands is not evidence-based, and that asserting a first-person knowledge claim about a hinge proposition is conversationally inappropriate (Pritchard 2005, p. 198), Wittgenstein’s position has unsettling consequences for his theory.

Here is the main one. If any claim to know a hinge proposition is improper, Pritchard sees that our reasons do not go “all the way down” to rock bottom because “our practice of offering grounds for our beliefs, and for our doubts, presupposes a backdrop of hinge claims that are both groundless and yet immune to doubt” (Pritchard 2005, p. 198). This prevents a wholesale rational evaluation and the idea that such a rational evaluation is coherent—or even meaningful—because it can only be done by assuming some prior commitment to rational evaluation itself. Yet if this commitment is ungrounded, it is also sealed to enquiry. What results from this is a stand-off in which hinges are supposed to be safe from sceptical doubts, but at the same time not rationally evaluable.

Pritchard notes that in general, we can make two valid interpretations of the nature of hinges in OC. In the so-called “minimal interpretation”, Wittgenstein is said to have accomplished a radical revision of the structure of reasons and evidence by:

  1. (a)

    reassessing the way we look at the structure of reasons, such that first-person knowledge claims aimed at disproving a hinge proposition become entirely inappropriate—the wrong move in the particular language-game of our basic beliefs;

  2. (b)

    introducing the idea that hinge “foundations” are neither self-justifying nor justified by anything an agent believes;

and as a consequence of (a) and (b), by (c) assuming that “any doubt on a hinge proposition will be necessarily groundless” (Pritchard 2005, p. 198). The reason why it will be so is that any possible claim to know the object of a hinge proposition will find itself at odds with the internal features of these certainties. Indeed, if hinges are said to be the underpinning of knowledge, rather than its object, hinge propositions are effectively left in an awkward situationFootnote 3.

This constitutes the basis for the “minimal interpretation” of Wittgenstein’s certainties. The “stronger interpretation” doubles down on the impossibility to know our basic certainties. Since sceptical doubts about hinge propositions are meaningless, they do not deserve any role in the language-game of our basic beliefs. Recall that in the Tractatus, Wittgenstein argued that “scepticism is not irrefutable, but obviously nonsensical, where it tries to raise questions where no questions can be asked” (1961, 6.51). At that time, Wittgenstein reasoned that scepticism had no entitlement to be seriously taken. If the late Wittgenstein still held to this view, he could not be bothered by the inquisitive attitude of the sceptic who repeatedly questions the reasons that back up the agent’s beliefs. Her demands should be outrightly dismissed because “both Moore and the sceptic try to say what simply cannot be said, and this accounts for the incoherence of both sceptical and non-sceptical assertions” (Pritchard 2005, p. 199). Implicit in it is the assumption that both sceptical and anti-sceptical claims are motivated by “semantic, rather than epistemic grounds” (Pritchard 2005, p. 199). Because sceptical claims violate our semantic rules and these rules turn out to make our conversation possible, Prichard believes that the stronger interpretation is simply unviable.

Pritchard thinks that an epistemological response to scepticism is preferrable to a non-epistemological. On this premise, he rejects the stronger interpretation. But even when this interpretation is pushed aside, he believes that the trouble for hinge theory does not stop here. On the face of it, it is very difficult, if not utterly impossible, to extract an anti-sceptical moral from the minimal interpretation, and hence, to argue that in the end, OC can command a successful answer to scepticism. Here’s how Pritchard describes the situation:

“If one is looking for a primarily epistemic response to scepticism from Wittgenstein’s remarks in this text then one is likely to be disappointed. At best, it seems that all that Wittgenstein offers in this regard is the kind of pragmatic defence of our belief in hinges that Wright was trying to pass off as being a genuinely epistemic approach to the problem. It thus appears that if there is to be a plausible anti-sceptical thesis inspired by On Certainty then that thesis had better be understood along primarily, semantic, rather than epistemic lines” (Pritchard 2005, p. 217).

In the end, the groundlessness of hinge propositions effectively frustrates any desirable epistemic analysis. By forcing us to either accept the absolute primacy of hinges or be pushed out of the conversation altogether, Wittgenstein treats hinge certainties as a very special class of propositions. And this, despite this move seriously weakens his overall position before scepticism.

In Pritchard’s view, this weakness generates a kind of “epistemic vertigo”, a genuine form of felt ambivalence about one’s overall epistemic situation. The vertigo is primarily caused by the paradox lying at the very foundation of knowledge. If we are to assume the inevitable consequences of accepting the overarching architecture of Wittgenstein’s hinges, his account “looks very much like a version of radical skepticism” (Pritchard 2017, p. 566).

Pritchard’s reading of OC finds support among others, in Wright’s and Moyal-Sharrock’s interpretation of hinges as ungrounded, reflex-like, arational devices. Yet on my view, this interpretation fails to do justice to OC and is more menacing to Wittgenstein’s project that what we might be inclined to think. Not only because of the failure of hinge theory to deflect scepticism, but also because of the hidden threat of relativism. If hinges are genuinely groundless, that is, entirely outside the realm of evidence and justification, any proposition, no matter how much its content contradicts the facts can potentially become a hinge for somebody. The intuition that drives this unsettling prospect is that if any individual has a unique cognitive rapport with reality, and this rapport is the result of practices and rules that greatly vary across individuals and societies, hinges may turn out to be incommensurable with each other to the point of frustrating any unprejudiced conversation about them. Individuals may implicitly acknowledge this weakness but learn to put up with its epistemic consequences as long as conflict may be avoided. “You have your own hinges; I’ll have mine. It’s ok”.

5 A Grundlos Interpretation (OC 166)

I intend to argue that the “groundlessness” reading of OC—the dominant view, rests on weak ground for two reasons: first, it relies on thin textual evidence, and second, it leaves aside other pivotal ideas of OC. Specifically, it pays no heed to the dynamical relation between knowledge and reality that characterises knowledge acquisition. In this relation, the discovery of new evidence can cause us to review our belief system in ways that radically change our beliefs. This suggests that the links between hinges and knowledge are probably both more subtle and enduring than what a superficial reading of OC suggests.

Let us start by examining the textual evidence supporting the dominant view. The word “groundlessness” (Grundlosigkeit) only appears once in the whole book (OC 166). In this paragraph, Wittgenstein sums up an idea that he had developed in the preceding paragraphs, namely, that the general facts of history to be found in history textbooks and oral traditions that came down to us seem so well attested by direct and indirect evidence that they are usually accepted without scrutiny. The fact that most of what can be read in geography textbooks about Australia or in history textbooks about Napoleon is taken for granted shows how knowledge builds upon knowledge to inspire certainty. The decision to hedge our belief system by scrutinising these facts, while it may perhaps be genuinely spurred by well-meaning safety concerns, does not make our knowledge any safer. There is a practical reason for it; OC 164 says that enquiries must come as a matter of fact to an end.

At this point, Wittgenstein claims:

“The difficulty is to realize the groundlessness (Grundlosigkeit) of our believing” (OC 166)

I read the German term Grundlos as an implicit affirmation of the depth and ramification of our epistemic links as well as of our practical inability to trace these links and drive this search to any successful conclusion. Tracing the epistemic link of every claim to some prior indisputable evidence is a futile quest that leads nowhere. Knowledge that is well attested by history or practice must be therefore trusted. If this is what is meant, I do not see how OC 166 can give us unequivocal evidence for the idea that the term Grundlos is intended to advance the proposition that hinges are ultimately groundless in the sense envisaged by the dominant view. The fact that a systematic analysis of the epistemic validity of every claim leads nowhere does not imply that knowledge of the general facts of geography and history is impervious to evidence. If the context provides meaning to this assertion, the context of OC invites us to think that the term Grundlos simply describes our practical inability to trace the links of every belief.

6 The Rationality of Hinges

This gives us some reason to exercise caution. The other often-quoted passages lending support to the dominant view describe hinges as a “form of life” (OC 358) and as “something animal” (OC 359). Allegedly, these passages make it seem as if Wittgenstein is asking us to blindly trust our most reliable convictions be they justified or not. Their lack of evidential support is the price to be paid for the sureness and stability that our epistemic system gets in exchange.

And yet, this interpretation is largely implausible. When Wittgenstein observed that hinges are a “form of life”, he subsequently noted that this idea “is very badly expressed and probably badly thought as well” (OC 358). The suggestion that our basic certainties can be “something animal” comes in the next paragraph (OC 359). So, the expression “very badly expressed” reveals that Wittgenstein held reservations about the implications of some of the images used to convey the singularity of hinges. But if they are or were indeed poor images of what he had in mind, what did he exactly meant when he wrote these passages? While this may be impossible to decipher, it is conceivable that he was trying to express how hinges become entrenched in our basic attitudes and dispositions—both theoretical and practical—to become second nature. In this way, the certainty that such dispositions convey do not simply inform one’s beliefs, but also one’s behaviour. This may the reason why OC 357 says that the words “I know” express comfortable certainty.

If this interpretation is correct, Wittgenstein is not asking us to blindly trust our hinges. When he argued that hinges are part of our form of life and something animal, he was pointing to a radically new way to conceive the scope and extent of epistemology. At the most basic level, our cognitive scaffolding provides us with a wholesale rational outlook of reality, and not just with arguments that have been repeatedly confirmed by investigation, experience or practice. This outlook is there to set knowledge into perspective, not to justify it, or to provide us with a fallback system if empirical knowledge fails.

OC characterises hinges as “world-pictures” (Weltbilde) that belong to “our foundation” or “our frame of reference” (OC 83, 411) . Generally, one learns these Bilde “as a child”. Since then, they have laid unperturbed at the basis of our cognitive scaffolding for years, if not decades,preserving this status because nothing that was ever considered undermined their credibility. As a result, they encroached into our cognitive system so that the rise of possible defeaters became almost unthinkableFootnote 4. For instance, my call to New York is underpinned by the basic conviction that the earth exists (OC 210). And while the idea that I should pause to remind me that the earth exists before making a telephone call sounds ridiculous, this fact is not irrelevant to it either. By informing thoughts and actions with a larger frame of reference, foundational world-pictures give stable support to our belief system and an enduring sense of cognitive reliability.

Let us now look at one of the best-known passages of OC. After the Grundlosigkeit paragraph, Wittgenstein shifts his attention to Lavoisier and the world-picture he inherited:

“Think of chemical investigations. Lavoisier makes experiments with substances in his laboratory and now he concludes that this and that takes place when there is burning. He does not say that it might happen otherwise another time. He has got hold of a definite world-picture (Weltbild)—not of course one that he invented: he learned it as a child. I say world-picture and not hypothesis, because it is the matter-of-course foundation for his research and as such also does unmentioned” (OC 167).

The interpretation of hinges as semantic devices focuses on Wittgenstein’s characterisation of hinges as “norms of description” (OC 167) or rules (OC 494) to say that their role in our belief system is to provide the rules of our language-games. To the extent that empirical propositions cannot be understood without hinges, they enable or open up meaning in the most basic sense. But Lavoisier’s example illustrates that their role is not simply grammatical. Hinges provide a wider description that places Lavoisier’s scientific practices within the larger framework of understanding in which his experiments make sense.

Lavoisier understood that the world-picture on which his experiments rested was fundamentally safe because it was seamless with his evidence. For all he knew, they made sense. Further, the basic laws of chemistry, the stable set of experimental conditions, the reliability of his experimental results also confirmed it. This was part of Lavoisier’s Weltbild, the set of basic presuppositions that made up his world. If these Bilde were radically upset, much would necessarily have to follow suit. But to the extent that this possibility was remote, Lavoisier’s Weltbild was objective, that is, in conformity with facts. If this is so, hinges can be said to describe or illuminate reality, and if they do so I do not blindly accept them. If I did, I would fail to know how they stand to facts and so, how they are they supposed to provide any stable support to our belief system.

Are Bilde verifiable? Certainly, they do not exhibit the bipolarity of propositions. When Wittgenstein discusses how hinges solidified into our cognitive scaffolding, he notes that in my recognising a towel as a towel, there is no process of verification involved; in a sense, I just act fully aware that the towel is a towel—why should I verify what is plain in sight? Wittgenstein distinguishes then this realisation, which he compares to a “taking hold” of something, from “a knowing” (OC 511). But while I do not verify that the object I am using to dry my hands satisfies an intended description, this does not mean that I have a vague or indeterminate idea of the kind of object I am using, or that knowledge of this object is currently irrelevant.

To elaborate this idea, I will argue that hinges are objective, nonpropositional content. To show why, I will distinguish a double dimension of hinges: the first is practical. Their practical dimension is the main focus of OC, and the one that absorbs most of Wittgenstein’s attention. OC unpacks many of the practical commitments of hinges. As such, they tend to trigger instinct-like, reflex-like responses in their bearer. Consider the following paragraphs:

“It is just like directly taking hold of something, as I take hold of my towel without having doubts” (OC 510).

“(…) the game can be learned purely practically, without learning any explicit rules” (OC 95).

“(…) in the end I rely on these experiences, or on the reports of them, I feel no scruples about ordering my own activities in accordance with them” (OC 603).

Insofar as hinges shape the agent’s conduct to reflect her acting certainties, they are practical devices that inform her dispositions for action. Moyal-Sharrock describes them as “a disposition of a living creature that manifests itself in her acting in the certainty of having a body” (2016, p. 560). Since Wittgenstein repeatedly analysed these certainties and the groundlessness interpretation primarily rests on this dimension, I will leave this disposition aside.

The second and most critical dimension is the rational. In a broad sense, hinges could be described as a cognitive disposition to envisage reality rationally. This disposition ensures that time is not wasted in what is cognitively trivial for having been confirmed many times over by observation. At the same time, this does not involve that hinges are random or arbitrary points of departure (Baldwin 2012) for any possible conversation. We know that they are largely consistent with reality, having been so many times confirmed by it, as well as by human practices. While not being themselves knowledge, their content is constantly scrutinised by the others and ultimately safeguarded by the facts. For instance, my hinge certainty that the chairs and tables in this room do not disappear every five seconds only to spring again into existence is consistent with my perceptual evidence and safeguarded by the facts. And so is my certainty that fresh blood looks red, etc. Similarly, Wittgenstein lets us understand that certain attitudes toward third persons are indisputable (2009, II, iv, p. 178). In the case of third persons, their very existence or the fact e.g. that they are in pain or depressed must be fully consistent with the attitudes we take to them.

It follows from this that hinges can be supported by rational arguments. But the authoritative process works negatively. When the rational disposition is in place, the authority of the putative hinge holds unless proof against it is presented. In this way, if anyone presents a meaningful case against one of my hinges, I may be surprised or irked by this challenge and reluctant to review my hinge. But if I overcome this reticence, I could later say why I hold it. Wittgenstein never suggested that the possibility that hinges get challenged is outlandish. If evidence emerges that throws me “entirely off the rails” (OC 517) or we happen to meet someone who “does not accept our whole system of verification” (OC 279) we might be forced to re-examine our Weltbild. I could bite the bullet and argue why I consider my world-picture objective, that is, in conformity with facts and even more so than its rival hinge. Of course, my doing so may not be sufficient to ensure that a rival hinge fails to be in conformity with the same facts, but in the event, it is to be hoped that sharing a same language-game, that is, a common method to resolve rational disputes will settle the dispute. What matters is that if my world-picture is rationally embedded, I am always in a position to argue why its outlook stays in conformity with facts. So, if I can confront challenges intended to upset my hinges using standard epistemic procedures, our most fundamental certainties are not just arational, “animal”, or unreflected dispositions.

In the following, I will spell out three main reasons to buttress the view that hinges are rational dispositions to envisage reality and derive a few consequences from it:

(1) While hinges are at their core dispositional, following Moyal-Sharrock’s view, they can become doxastic categories and be propositionally expressed in sentences like “I have two hands”, “the world existed long before my birth” or “I am not a brain in a vat” in certain contexts. At the same time, in so doing they become bipolar and lose their status.

(2) Hinges reflect the dynamical relation between a Weltbild and empirical knowledge. There is a natural continuity between our world-pictures and our entire set of beliefs. Consider any scientific paradigm that has taken years or decades to establish itself. And consider likewise particular scientific theories such as gravitational waves or abiogenesis. Both paradigms and theories make up a seamless body of scientific knowledge. Some of its parts may be more likely to give if tensions between them arise. In this context, Wittgenstein likens the “propositions of logic” to the riverbed on which empirical propositions flow. The propositions of logic resemble scientific paradigms in their development and stability. While generally stable, nothing prevents that some riverbed propositions give and “change back into a state of flux” (OC 97), thereby losing their status. By contrast, empirical propositions that stabilise can eventually harden can become part of the riverbed. So it seems natural to think that the dynamical relation between both kinds of propositions can only be real if both share a common set of rules. If this were not the case, how could the “fluid” propositions of day-to-day theories, hypothesis and empirical observations become “hardened” (OC 96)? Thus, both scientific paradigms and scientific theories must be viewed as mutually interlockedFootnote 5 and supporting.

(3) If hinges are “the foundation of all operating with thought” (OC 401) or its bedrock, they must be consistent with our entire set of beliefs. Conformity with facts is not enough to create a unified belief systemFootnote 6. Insofar as hinges largely reflect how evidence is internalised, it is understood that they must be consistent with this system.

Consequently, if hinges are rational in the sense that (1) they can be turned into propositions, (2) reflect the dynamical relation between world-pictures and empirical knowledge, and (3) must be consistent with our belief system, the assumption that they are ungrounded, that is, entirely unsupported by facts, or that this support is irrelevant, should be reviewed. Hinges were born as pieces of empirical knowledge and must hence be indirectly supported by it.

This brings with it an importance consequence. Contrary to Pritchard’s assessment, the logic of OC underscores the possibility of warrant without justification and of immunity to error without bipolarity. If this warrant is not unsupported by facts, we need no longer think that Wittgenstein’s epistemology inspires any “epistemic vertigo” over the possibility that hinges are straightforwardly certain by epistemic fiat.

At the same time, the enactive theory of hinges that portrays them as arational, or as “ways of acting”, captures the practical, but not the rational dimension of hinges. The “animal” view fails to do justice to the role that hinges play in our rational evaluation of beliefs. On closer inspection, the picture that we get is that hinges are what knowledge would be like if all evidence were to become unequivocally certain for us, that is, optimally self-evident. Because many times reliable evidence is hard to come by, partially obscure, and deteriorating, the exercise of justification knows no end. Hinges bypass this difficulty by being—as internal dispositions—beyond justification to provide a solid basis for our beliefs.

Ultimately, by articulating hinges as dispositions to rationally envisage realityFootnote 7, we can expand the scope of analysis of OC to demonstrate that dominant view, which holds that Wittgenstein’s hinges are groundless, should be taken with scepticism, and that Wittgenstein’s project is still relevant.

7 Conclusion

The final picture of the relation between hinges and knowledge might still seem unconclusive. A possible complaint to the view I have articulated may be that my dispositional analysis revives the mythological character (OC 95) of hinges—an impression that OC does little to dispel. While certain mythology of hinges is inevitable given the ambiguity of Wittgenstein’s theory, I have shown that the critical role that hinges play in establishing our world-pictures is matched, although in a different way, by indirect empirical support. This support is not the reason why our hinges became entrenched in our cognitive scaffolding, because they are epistemically prior to empirical knowledge—recall that Wittgenstein rejects the idea that we “learned” (OC 315) our language-games. Yet knowing that they receive indirect evidential support helps us to see that they are rational and not powerless against scepticism.

Understanding that hinges have an important role to play in the space of reasons undermines the idea that they are ungrounded. This, in turn, ends up dissolving Pritchard’s epistemic vertigo. After removing it, it is not clear why we should accept Pritchard’s negative assessment of Wittgenstein’s antisceptical strategy. We rather should say that Wittgenstein’s epistemology casts the regress of justification on a different light. The sceptic’s demands for ultimate justification can only come to a halt if it is implicitly acknowledged that cognition—when it reaches “rock bottom” (OC 248)—is unassailable. Standard sceptical hypotheses like the supposition that I may be a BIV can only fly if hinges grant them context and meaning. Such a context is embedded in prior observations that solidified into our cognitive scaffolding for years, if not decades. This is how Wittgenstein’s nonepistemic solution to scepticism can deal with the problem that triggers the regress of justification. And while this interpretation assumes the reciprocal supporting relation between hinges and knowledge, I think that the reasons that motivate this acceptance are solid. Hinges are not standalone epistemic devices. While nonpropositional and nonepistemic, their commitments are ultimately rational—that is, amenable to reasons, and in keeping with reality in their idiosyncratic way.