Keywords

1 Introduction

Epistemic contextualism (hereafter just “contextualism”) has become one of the most prominent positions in contemporary epistemology. According to this theory a proper analysis of our ordinary linguistic practices would establish a viewpoint on the semantic behaviour of knowledge ascriptions that can improve our understanding of some remarkable epistemological questions, such as the sceptical problem or the lottery paradox. In the last years the debate on contextualism especially focused on these aspects of the theory, neglecting some other relevant issues that arise from its internal structure. In this paper we will concentrate our attention precisely on one of those issues: we will try to single out which conversational and epistemic mechanisms are involved in the variation of the truth-conditions of knowledge ascriptions. After a brief introduction of contextualism and of some useful contextualist terminology, in Sect. 3 we will analyse different accounts of the mechanisms involved in a context’s change - Lewis’s view, the veto power view, the gap view, and the intentionalist view, - with the purpose of identifying the virtues and the lacks of each proposal. In the end we will conclude that both Lewis’ view and the veto power view are inadequate, and we will maintain that the gap view seems to be superior to the intentionalist view since it vindicates our intuition that, in that cases where the conversational partners do not agree on which epistemic standard should be in place in their conversational context, the speakers are, at least under a certain respect, contradicting one another.

2 Contextualism

Contextualism is the semantic thesis that the truth-conditions of knowledge-ascribing and knowledge-denying sentences of the form “S knows that p” and “S doesn’t know that p” depend upon certain features of the context of the knowledge attributor - i.e. the context of the person making the assertion. Therefore, accordingly, a sentence as:

  1. (1)

    Christian knows that he’s looking at a dandelion

can be true if uttered in a certain context C 1 at the time t and false if uttered in another context C 2 at the same time t without contradiction since the two utterances of the same sentence would express two different propositions (Cohen 1987: 1; Lewis 1996: 550; DeRose 2009: 2; Rysiew 2016). Contextualists usually account for this peculiar idea about the semantics of knowledge ascriptions resorting to the so-called “argument of the ordinary language”. According to this argument the analysis of our ordinary usage of knowledge ascriptions would prove that what we will deem as “knowledge” in certain contexts we will not recognise as such in others (DeRose 2009: 47). Consider, as an example, the sentence (1) and assume that Christian is in an ordinary epistemic position with respect to the proposition that he’s looking at a dandelion (the proposition q). As the contextualists predict, it seems that we will probably judge an utterance of (1) as true if we are in the context of an ordinary conversation, but as false if we are in the context of a botanical symposium. Indeed, the circumstances of an ordinary conversation would presumably establish a set of conditions at which a subject can count as knowing that q that are less strict than that defined in the context of a botanical symposium. In the context of an ordinary conversation, in fact, we can imagine that to Christian it will be simply requested to be able to tell the difference between a dandelion and oxeye daisy in order to make true a sentence as (1), while in the more demanding context of the botanical symposium he will probably have to be able to list all the relevant botanical characteristics of a dandelion in order to count as knowing that he’s looking precisely at that flower. Thus, for Christian is an ordinary epistemic position with respect to q - and therefore he’s able to tell the difference between a dandelion and oxeye daisy, but he can’t list all the relevant botanical characteristics of a dandelion, - an utterance of (1) will be evaluated as true in the first context but false in the latter.

It should be noted that, traditionally, knowledge-ascribing sentences are not held to be context-sensitive expressions. According to the classical “invariantist” view there is one and only one set of truth-conditions for knowledge ascriptions.Footnote 1 Some clarification about the contextualist main claim are thus in order. To illustrate the contextualist stance it’s useful to resort to a compelling analogy often employed by the advocates of the theory (Cohen 1986: 580, 1999: 60; DeRose 2009: 166). Consider gradable adjectives as “tall” or “flat”. The truth-conditions of this kind of adjectives almost certainly vary in different contexts since it is the context of utterance that establishes the parameter that defines how those terms apply - e.g. a context can determine the maximum height in order for a person to count as “short”. Therefore, since knowledge ascriptions seem to vary in degree of “strength” or “goodness”, each context can define the strength or the goodness of the epistemic position of a subject S such that S can count as a knower in that context. Another way to illustrate the context-sensitivity of “know” is to regard this term as an indexical like “I” or “here” (Cohen 1999: 61; DeRose 1992: 925). To this kind of expressions are connected the well-known notions introduced by Kaplan (1989) of “character” (the rule associated by convention to an expression that sets the contextual parameters to locate the reference of the occurrences of the expression in contexts) and “content” (the intension, or the semantic value of the expression). Making use of this distinction, we can thus associate to a knowledge ascription of the form “S knows that p” a character such as “S has a true belief that p and she is in a strong (or good) enough epistemic position with respect to p”. Clearly, this element of the meaning of “S knows that p” will be stable among contexts, while how much strong or good an epistemic position must be in order to make true a knowledge ascription will change from context to context (DeRose 2009: 3).Footnote 2

2.1 The Strength of the Epistemic Position and the Epistemic Standard

In the previous paragraph we have spoken of contexts resorting to a popular terminology among contextualists that involves expressions as “strength of epistemic position” and “epistemic standard”. From a very general point of view, being in a strong epistemic position with respect to a certain proposition one believes is for one’s belief in that proposition to have an adequate amount of the properties the having enough of which is what’s needed for a true believe to constitute a piece of knowledge. The epistemic standard set in the context is, instead, “the standard for how strong a position a subject must be in with respect to a proposition … for a sentence attributing knowledge to her in the context in question to be true” (DeRose 2009: 7).

Contextualists have proposed different hypotheses to clarify and specify how those notions should be understood. According to an influential interpretation proposed by Keith DeRose, the strength of the epistemic position of a subject S with respect to a proposition p should be understood in terms of truth-tracking or of “sensitivity”: S’s true belief that p tracks the truth or is sensitive to the truth value of p if S would not have believed that p if p had been false (DeRose 1995).Footnote 3 Consider Christian’s true belief that he’s looking at a dandelion: he certainly believes this both in the actual world and in some other nearby possible worlds where he is, in effect, looking at a dandelion. Consider, however, other nearby possible worlds where it is not the case that Christian is looking at this flower for he’s looking at a rose or at a storksbill: in these worlds where it is not the case that Christian is looking at a dandelion he is perfectly aware of this fact - using DeRose’s own words, Christian’s belief “matches the fact of the matter” (1995: 34), - hence, in these worlds Christian does not believe that he’s looking at a dandelion. There are, however, other possible worlds in which notwithstanding it is not the case that Christian is looking at a dandelion he will believe this anyway - think, for example, to a sceptical scenario in which Christian is deceived by an evil demon. Nevertheless, for these worlds are really far from the actual world, we could consider Christian’s belief sensitive and conclude that he’s in a pretty strong epistemic position with regard to the proposition that he’s looking at a dandelion (DeRose 1995: 35). Therefore, according to this model, the strength of the epistemic position of a subject S depends upon the possible worlds which are considered “fairly near” in the context: the epistemic standard defines the set I of possible worlds which in its turn determines the sensitivity of S’s belief that p, and if S’s belief can track the truth in all the possible worlds considered fairly near (i.e. all the possible worlds which belong to I) then S’s epistemic position with respect to p is strong enough - and hence a knowledge ascription as “S knows that p” can be judge as true in the context.

Yet, the notions of epistemic standard and of strength of epistemic position can be considered differently. Another popular way of defining standards consists in making use of the theory of relevant alternatives. According to this theory, knowledge is strictly tied to our ability of discriminating between epistemic alternatives; indeed, it seems that we would probably not say that Christian knows who Marian is if we would know that he cannot tell apart Marian from her monozygotic twin Vivian, even if he would say - precisely pointing at Marian, - “I know that she is Marian”. Therefore, accordingly, in order to count as knowing a certain proposition p a subject S should rule out all the relevant alternatives to p (Goldman 1976; Schaffer 2005). We can thus regard each context as a certain set - or as a “segmentation” using Antonia Barke’s words (2004: 354), - of alternatives that are considered to be relevant in that context.Footnote 4 If we consider again our example involving (1), we can outline the two contexts in the following way: in the first case - the one sets in an ordinary context, - to Christian it will be asked to rule out certain ordinary alternatives, while in the latter - the botanical symposium context, - he could have to rule out much extreme alternatives: for example, in the more demanding context also the remote possibility that the flower in front of Christian is a dandelion* (an extremely rare flower which is quite identical to the dandelion and is distinguishable from it only by a very skilled botanist) could be included in the set of alternatives that Christian must rule out to count as knowing that he’s looking at a dandelion.

Clearly, the sensitivity approach can be understood in a way that it’s rather similar to the relevant alternative approach: the fairly near possible worlds can be interpreted as the worlds in which are located the alternatives that are considered relevant. However, opting for the sensitivity approach could be a dangerous step. Indeed, Ron Wilburn maintains that this approach presupposes a metric to evaluate the modal distance (the amount of metaphysical differences between possible worlds) that is “implicitly calibrated” by a metric for judging epistemic distance. According to him:

this implicit calibration threatens circularity by effectively presupposing the truth of the very claim for which the contextualist must argue, i.e., that various skeptical scenario worlds (in which truth-tracking is sabotaged) stand at a greater distance from actuality than do more prosaic worlds (in which truth-tracking remains in order) (Wilburn 2010: 260).

Evaluating Wilburn’s objections goes beyond the aims of this paper, but for the relevant alternative approach is not certainly saddled with these kind of worries, here we will resort to this reading of the notions of epistemic standard and strength of epistemic position to characterise contextualism.

3 Context Changes

3.1 The Rule of Accommodation and the Rule of Attention

Now that we have a clearer understanding of some relevant contextualist notions, we can begin to examine the dynamics of contexts changes. Consider two speakers involved in a conversation: in which manner each conversational partner can manipulate the epistemic standard in place in the context? A first answer to this question has been developed by David Lewis. The American philosopher maintains that the variations of the truth-conditions of knowledge ascriptions are induced by certain conversational manoeuvres; in particular, according to Lewis the mere mention of an epistemic alternative would be sufficient to install a certain epistemic standard in a context (1996: 559). For a better understanding of this proposal we should consider it in the light of Lewis’ claim that at any stage of a well-run conversation is associated a “conversational score”:

Sentences depend for their truth value, or for their acceptability in other respects, on the components of conversational score at the stage of conversation when they are uttered. Not only aspects of acceptability of an uttered sentence may depend on score. So may other semantic properties that play a role in determining aspects of acceptability. For instance, the constituents of an uttered sentence - subsentences, names, predicates, etc. - may depend on the score for their intension or extension (Lewis 1979: 345).

According to Lewis, the conversational manoeuvres of the speakers can manipulate and change the values of the score during the conversation. Conversational contexts are in fact governed by a “rule of accommodation” which posits that:

If at time t something is said that requires component s n of conversational score to have a value in the range r if what is said is to be true, or otherwise acceptable; and if s n does not have a value in the range r just before t; and if such-and-such further conditions hold; then at t the score-component s n takes some value in the range r (1979: 347).

Therefore, if during a conversation a speaker utters the sentence “Smith’s Ford is in the parking lot” this utterance will immediately manipulate the conversational score posing the presupposition that Smith owns a Ford and, consequently, making rather odd a caveat as “… and he owns a Ford”. The rule of accommodation, thus, seems to be responsible for the set of the epistemic standards. Consider again the case where Christian is looking at a dandelion. If someone says that:

  1. (2)

    Christian doesn’t know that he’s looking at a dandelion

claiming, for instance, that the flower at which Christian is looking could be not a dandelion, but rather a very similar flower, the mention of (2) - according to the rule of accommodation, - will manipulate the conversational score changing the epistemic standard in place in the context so that (2) can be true.

If it is in virtue of the rule of accommodation that the standards can change, we could expect that in the same way in which the standards can be raised they can also be lowered. However, if according to Lewis the standards for many terms can be both raised and lowered in many occasions (1979: 339), the same manoeuvre cannot be employed when it concerns the epistemic standards. Lewis maintains that, in order to count as knowing a certain proposition p, a subject S must rule out any epistemic alternative that he is not properly ignoring. However, according to the “rule of attention” (1996: 559) an epistemic alternative is properly ignored only when it is, in fact, ignored simpliciter - that is, if the subject is attending to it then the alternative is not properly ignored. Therefore, since the statement of an epistemic alternative entails the fact that this very alternative has been attended by the subject, it follows that when the epistemic standard has been raised it cannot be lowered anymore.

From the conjunction of the rule of accommodation and the rule of attention it follows an extremely prudent characterisation of our epistemic rationality: indeed, according to this view any epistemic alternative - even the most farfetched as, for instance, the brain-in-a-vat hypothesis, - should be considered and ruled out by the epistemic subject. Lewis’ view can thus be labelled as a very “sceptical-friendly” view since it implies that when sceptical hypotheses are put in place sceptical standards are automatically installed.

It should be noted that Lewis’ proposal did not obtain a positive outcome among contextualists: Michael Ashfield, for instance, has defined it as depicting “a worst-case-scenario for the contextualist” (2013: 122), while in many passages of his works DeRose denies his support to the rule: “… actually I have a lot of sympathy for the thought that the mere mention of the alternative is not sufficient for making it relevant …” (DeRose 2000); “Nor do either of us [DeRose and Cohen] accept the view that is the next most commonly ascribed to the contextualist: that, in the situations under consideration, the sceptic’s extraordinarily high standards prevail” (DeRose 2004). An in-depth analysis of Lewis’ position is developed by Barke. According to her (2004: 214) a first problem of Lewis’ viewpoint lies in its characterization of the dynamics that rules the set of the epistemic standard as a conversational one, since it would not be able to account for the cases in which no conversational context is involved - as the one of a solitary reflection. As Barke notices, it could be replied that solitary considerations are cases of “conversation with oneself,” but nevertheless, according to her this “does not seem to be a very attractive solution”. Yet, Barke does not propose any argument for the rejection of a response that, at least to us, seems to be perfectly adequate; indeed, just as a conversation between two speakers, also a monologue - or an interior monologue in the case of a solitary consideration, - can be governed by a conversational score: if speaking with himself S thinks to a certain parameter for the predicate “being tall” and then, later, he changes this parameter for he’s thinking to basketball players, this change can be perfectly described in terms of variations in the values of a conversational score. Another remark posed by Barke concerns the ease that characterised the conversational move of rising the standard. Indeed, for many other context-sensitive terms we can manipulate the score in both ways: we can tighten the standard for tallness speaking about basketball players and then we can broad it driving the conversation to eight years old children; in the case of knowledge ascriptions, instead, a return to a loose standard cannot be achieved by mentioning a low-standard knowledge ascription. Clearly, this difference between the conversational mechanics of terms as “tall” or “flat” and knowledge ascriptions depends upon the assumption of the epistemic rule of attention. As Lewis claims, “speaking of knowledge despite uneliminated possibilities of error just sounds contradictory” (1996: 549) and the statement of a not ruled out epistemic alternative seems to prevent an effective ascription of knowledge; as an example, it seems that we would not say that Christian is looking at a barn if we would know that the countryside in which Christian is it’s full of fake barns that are indistinguishable from real barns. Yet, according to Barke this explanation of the asymmetry between the manoeuvres of rising and lowering the standards is not promising for she considers the rule of attention highly implausible. Indeed, according to Barke “discussing our presuppositions and agreeing on the question of which ones are reasonable … and which ones are not it’s an integral part of our epistemic practices” (2004: 358), and clearly Lewis’ position does not vindicate this part of our epistemic custom.

Since contextualists generally aim to construe a theory that it’s in consonance with our ordinary knowledge-attributing practices (Kompa 2014: 59; Rysiew 2016), the asymmetry between the manoeuvres of rising and lowering the standards that results from Lewis’ view seems to suggests that an alternative account of these mechanisms should be preferred.

3.2 The Veto Power

As Barke rightly points out, our ordinary epistemic practice seems to grant to us the possibility to stop that conversational manoeuvres that attempt to rise the epistemic standard. Appealing to this intuitions, many contextualists have contested the rule of attention maintaining that, just as in the case of other context-sensitive terms, the attempt of raising the standards can be resisted also when knowledge ascriptions are involved. According to this view, to the speakers is conferred a “veto power” over the changing of the conversational score (DeRose 2009: 140). Stewart Cohen describes this plausible conversational manoeuvre in these terms:

The pressure toward higher standards can sometimes be resisted. One device for doing this is adopting a certain tone of voice. So in response to the sceptic, one might say, “C’mon, you’ve got to be kidding - I know I am not a brain-in-a-vat!”. If this is the dominant response among the conversational participants, then everyday standards may remain in effect. In such a case, the speaker unmoved by skeptical doubt is not failing to adjust his ascriptions to contextually determined standards. Rather, such a speaker is managing to keep the standards from rising (Cohen 2001: 93).

The veto power view seems to perfectly conform to our ordinary epistemic custom. Moreover, it vindicates the contextualist’s claim according to which the standards are determined not by a rule as the rule of attention, but by the presuppositions, the purposes and the practical interests of the speakers (Cohen 2001; Kompa 2014). Michael Blome-Tillmann posits an interesting distinction among these lines. According to him the rule of attention should be restated for simply attending to a certain epistemic alternative it’s not enough for making it impossible to ignore in a “epistemologically relevant sense” (Blome-Tillmann 2009: 247). In order to prove that, Blome-Tillman proposes this case:

Imagine you saw your teenage son sneaking away through the window of his room late at night. When you confront him the next morning he replies somewhat desperately: ‘How do you know I left the house? I mean, for all you know you might have dreamt it. It was late at night, wasn’t it?’ (Blome-Tillmann 2009).

Now, according to the rule of attention the objection of the son would have the automatic consequence of raising the epistemic standard to a sceptical level; this consequence, however, seems to be extremely implausible to Blome-Tillmann since he maintains that in order to make a certain epistemic alternative relevant we should take it seriously. Yet, Blome-Tillmann example appears to be rather ambiguous since, as Daniele Sgaravatti rightly observes, the son’s response sounds so odd that it is even dubious if it could be considered as a proper contribution to the conversation. However, if the son’s objection would be sincere then the veto power view would grant to the father the possibility to overlook the son’s objection. Note that according to Sgaravatti a better contextualist response would be “You might be right that I do not know in an absolute sense what happened: we hardly know anything in that way. But this is just not relevant now!” (Sgaravatti 2013: 7). Yet, it seems that Sgaravatti’s remark is missing the point, since according to the advocates of the veto power this manoeuvre is effective also in that cases where the epistemic alternatives are sincerely affirmed by their proponents.

This veto power view seems to be really promising because it efficaciously prevents the standards to raise automatically. Unfortunately, this view has an highly undesirable consequence for it seems unable to account for cases of radical disagreement concerning the application of the epistemic standards. Imagine a dispute between a common-sense epistemologist and a sceptic: when the sceptic will try to raise the standard the common-sense epistemologist will immediately employ the veto power. Now, can the sceptic employs the veto power too preventing the lowering of the standard? Who wins? The one who first employed the conversational manoeuvre? But in such a situation it seems extremely implausible to tie the efficacy of the veto power to the temporal sequence of its employment in the conversational context. Thus, the result of this peculiar case of disagreement seems to be an opaque impasse since it’s not clear which value should assume the conversational score according to the veto power view (DeRose 2009: 141). Hence, also the adequacy of this view seems to be questionable.Footnote 5

3.3 The Gap View and the Intentionalist View: Single vs. Multiple Scoreboard Semantics

In order to give to the speakers the possibility to freely manipulate the truth-conditions of knowledge ascriptions we should thus opt for an account able to explain the outcomes of the cases of disagreement about the epistemic standard. DeRose proposes an account along these lines (the “gap view”) according to which in the cases where two speakers - as, for instance, a sceptic and a common-sense epistemologist, - do not agree on which standard should be applied in their conversational context an utterance of “S knows that p” assumes the following truth-conditions: the sentence is true if S meets the extremely high standard of the sceptic, false if S doesn’t meet the common-sense epistemologist ordinary standard and neither true or false if S does not meet the sceptic standard and meets the common-sense epistemologist standard.

As DeRose observes, this view has the remarkable virtue of respecting two strong intuitions that we have about the cases of disagreement in question: the first, according to which the two speakers are contradicting each other, and the latter that posits that the truth-conditions of each speaker’s assertion should match his personally indicated content (DeRose 2009: 144–148). On the other hand, however, the gap view seems to be saddled with some thorny problems. First of all, DeRose’s proposal could be considered a rather sceptical-friendly approach: in a dispute between a sceptic and a common-sense epistemologist, even if the sceptic does not manage to install his extremely high standard in the conversational context, at least he is able to prevent the knowledge ascription of the common-sense epistemologist. Furthermore, according to Montminy (2013) the gap view would entail another serious problem related to those subjects (that Montminy labels as “dogmatic Mooreans”) who impose low epistemic standards and refute any raising of them. “A dogmatic Moorean” Montminy writes “… insists that he ‘knows,’ even though he cannot rule out salient alternatives” (2013: 2348). In order to explain why those subjects would represent a problem for the contextualist who endorse the gap view, Montminy resorts to the well-known airport case developed by Cohen (1999: 58). Suppose then that John and Mary are at the Los Angeles airport and that they want to know whether their flight has a layover in Chicago. They ask to another passenger, Smith, who replies that he knows that the flight stops in Chicago for he has just checked his flight itinerary. Since Mary and John have a very important business contact who’s waiting for them at the Chicago Airport, Mary points out that Smith’s itinerary may be misprinted and concludes that Smith doesn’t know that the plane will stop in Chicago. In the original case John agrees with her. In Cohen’s intention, this fact should suggest to us, the readers, that the sentence “Smith knows that the flight will stop in Chicago” can assume different truth-conditions depending upon its context of utterance. However, according to Montminy, assuming that John is a dogmatic Moorean can cause serious troubles to the contextualist who endorses the gap view since this stance entails that if John does not agree with Mary, then she cannot succeed in establishing high epistemic standard in their context. Therefore, the case cannot be used anymore to illustrated and to support the contextualist position. Since examples as the airport case represent the main argument in favour of contextualism (Rysiew 2016), it seems that the gap view has the undesirable consequence of jeopardizing “the chief motivations” for the theory (Montminy 2013: 2348).

For the abovementioned reasons Montminy claims that the gap view should be abandoned. Montminy proposes an alternative position, the “intentionalist view” (2013: 2351–64), according to which the epistemic standard involved in the content of an utterance of “S knows that p” is defined only by the speaker’s intention. Clearly, contrary to DeRose’s position, Montminy’s proposal is a multi-scoreboard view for every speaker has his own conversational scoreboard.

An important virtue of the intentionalist view is that it isn’t a sceptical-friendly approach: indeed, even if the sceptic affirms truly that “S does not know that p” (for he is intending very high epistemic standard) also the common-sense epistemologist is speaking truly when he says that “S knows that p” (for he has in mind ordinary epistemic standard). Moreover, this view does not affect the contextualist cases since the objection of the dogmatic Moorean John cannot prevent Mary to adopt high epistemic standards.

Yet, also the intentionalist view faces some problems. Indeed, the distinction that Montminy draws between his intentionalist view and the classical contextualist stance it’s not perfectly clear. Cohen and DeRose - the two main proponents of contextualism, - generally describe their position as one according to which the truth-conditions of a knowledge ascription are determined by certain features of the ascriber’s context, as the ascriber’s purposes and practical interests. Think, for example, to DeRose’s bank cases (2009: 1–3): in the first scenario of his famous example DeRose ascribes knowledge to himself with regard to the proposition “The bank will be open on Saturday morning” despite his wife has pointed out that the bank could be closed on that day. DeRose considers this error possibility, yet he does not “take it seriously” or, better, he judges his own contrary evidence strong enough to rule out it. However, in the second scenario of the example DeRose’s practical interests raise the standard (if the bank will be closed on Saturday morning he won’t be able to deposit a very important check) and he prefers to admit that he doesn’t know whether the bank will be open on Saturday morning. Therefore, it seems that the “purposes” and the “practical interests” of DeRose’s contextualism are not so different from the “intentions” of Montminy’s view. Furthermore, Cohen explicitly says that the epistemic standards are defined by “the intentions, the expectations, and the presuppositions of the members of the conversational context” (Cohen 2001: 92). Nevertheless, one could object, as Montminy does (2013: 2352), that according to the classical contextualist view others factors than the practical interests, the intentions, and the purposes of the knowledge ascriber can define the truth-conditions of his knowledge attributions. For instance, one could say that, according to the classical contextualist view, the fact that a certain region R is full of fake barns should affect the truth-conditions of Henry’s knowledge ascription that “Tom (who is in R) knows that he’s looking at a barn” even if Henry does not know that R is full of fake barns. However, it seems that the classical contextualist view does not posit anything like this: according to DeRose and Cohen if a certain epistemic alternative is ignored by the ascriber this possibility cannot affect the truth-conditions of his knowledge ascriptions. Furthermore, even if the ascriber is aware of a certain epistemic alternative it’s up to him to decide if taking it seriously or not. Yet, if so, it seems that the truth-conditions of the knowledge ascriptions of the ascriber strongly depend by his conscious intentions.Footnote 6 After this important clarification, it appears that we can conclude that the true characterizing element of Montminy’s proposal is the fact of being a multi-scoreboard view. However, it’s by no means certain that a multi-scoreboard view is preferable to a single-scoreboard view.

Has we have seen Montminy claims that the intentionalist view should be preferred to the gap view since the latter would be too sceptical-friendly. Could we consider this as a real problem for DeRose’s multi-scoreboard position? The answer seems to be negative. The contextualist anti-sceptical argument simply claims that from the fact that we do not know ordinary propositions according to the sceptical standard it cannot be inferred that we do not know that very propositions according to ordinary standards.Footnote 7 Thus, what it is important for the contextualist is to preserve the legitimacy of our knowledge ascriptions made in ordinary contexts, a legitimacy that the gap view does not prevent.

On the contrary of the gap view Montminy’s single-scoreboard position doesn’t jeopardize the chief motivation for contextualism. However, it seems that contextualism could peacefully survive to this consequence. After all, as DeRose rightly points out (2009: 53–59), only certain properly construed cases are able to represent a compelling argument for contextualism. Hence, putting aside cases as the one where the speakers do not agree on the epistemic standard does not seem to restrict too much the contextualist’s motivations.

A more debatable aspect is tied instead with the intuition that when two speakers do not agree on which standard should apply to their context they are, in effect, contradicting one another. According to DeRose this intuition is confirmed by the fact that, in a case as the one that sees a sceptic opposing to a common-sense epistemologist, each speaker is explicitly indicating that he’s contradicting what his opponent is claiming. The gap view seems to do a nice work in accounting for the contradiction that we intuitively recognise in such cases, what about the intentionalist view? Montminy’s position involves an error theory for it maintains that the two conversational partners are mistaken in believing that they are contradicting one another (Montminy 2013: 2361). As Montminy rightly points out, an error theory is posed also by the classical contextualist view since even according to this position the knowledge ascriptions assessed in different contexts express different propositions, with the only difference that:

proponents of the [classical] view need to explain away only the intuition of contradiction across contexts, whereas intentionalists have to account for why we are mistaken about there being both inter- and intra-contextual contradiction (Montminy 2013, p. 2361).

Therefore, the intentionalist view has to deny our intuition that in that cases where the speakers do not agree on which epistemic standard should be in place in their context they are, in effect, involved in a genuine dispute. However, it’s difficult to imagine how such a denial could be efficaciously supported. As we have seen before our intuitions strongly suggest that the two speakers, the sceptic and the common-sense epistemologist, are contradicting one another. Furthermore, it is quite clear that the sceptic and the common-sense epistemologist are intending to do so: in fact, they do not simply argue, respectively, that their personal assertion of “S knows that p” is false or true according to their personal epistemic standard, but that in the conversational context in which they are - and that includes also their conversational partner, - it should be applied a certain epistemic standard instead of another. The purpose of each speaker is clearly to impose his epistemic standard to his conversational partner. In order to illustrate the strength of these intuition we will resort to an example. Consider the sentence “Louis is tall”. If it seems extremely plausible to assume that there isn’t a real contradiction between two utterances of this sentence made in two contexts in which are in place different standards for “tall”, it seems much less plausible to maintain that, when two speakers are debating on which standard for tallness should be adopted in their conversational context, they are not involved in a genuine dispute.

4 Concluding Remarks

In the previous paragraphs we have analysed the different dynamics that have been proposed to explain how the epistemic standards can be changed. Our aim was to scrutinize each position in order to find the better mechanisms of contexts changes, from both a conversational and an epistemic point of view. We have argued that the account proposed by Lewis reveals its undesirable limits for it sacrifices its consonance with some of our plausible epistemic customs: speakers can, after all, employ conversational manoeuvres in order to prevent the raise of the epistemic standards in the conversational contexts. We have seen that prima facie the veto power view efficaciously vindicates this intuition, but also that it faces a serious problem that consists in its inability to handle those cases in which the speakers do not agree on which standard should prevail in their contexts. Two proposals appear able to deal with this problem, the gap view and the intentionalist view. The two approaches conform to the contextualist anti-sceptical argument and, contrary to Montminy opinion, they both seems to not seriously undermine the main arguments for contextualism. The pivotal difference between these two view seems to concern their treatment of our intuition regarding that case where the speakers do not agree on which epistemic standard should be in place in their contexts. We have maintained that the main virtue of the gap view’s single scoreboard semantics strategy consists in its ability to account for the intuition that in the abovementioned cases the conversational partners are, in effect, contradicting one another. The intentionalist view clearly cannot obtain this results and this seems to be its main limit: indeed, the intuition that in the cases in question the speakers are contradicting one another seems very difficult to be explained away. Therefore, we conclude that, at least under this respect, the gap view appears to be superior to the intentionalist view. Yet, for this superiority depends upon a minute and not definitely decisive aspect, the problem surely deserves additional attention and further inquiry.