Skepticism, Flat, and Knows

INTRODUCTION

In this paper I will discuss contextualism about the word “knows” to determine if it offers a plausible resolve to skepticism. To do this I will, first, present Brain in a Vat (BIV) skepticism. Next, I will explain the theory of contextualism as presented by Stewart Cohen and explain his application of the theory against particular skeptical arguments. Then, I will present a dilemma posed by Ernest Sosa for contextualism and, finally, will argue that contextualism cannot overcome this dilemma nor defeat BIV skepticism.

SKEPTICISM

The strength of the skeptic is the apparent appeal of her argument which is ipso facto the very essence of the paradox. She presents premises whose plausibility we intuitively accept, yet which lead to skeptical conclusions we intuitively reject; as is the case following:

1. If I know that I have hands, then I know I am not a disembodied brain in a vat (BIV) being deceived.
2. I do not know I am not a BIV.
3. Therefore, I do not know that I have hands.

We (the non skeptics) may repulse the thought of (II), but remember it is the plausibility of the premises our intuitions accept. Further, we may press the falsity of the conclusion as being more intuitively plausible and attempt a “Moorean maneuver” conjoining this falsity with (I) to refute the skeptic. This may save our intuition that we have knowledge by escaping BIV skepticism, although, as Cohen noted, we will have failed to explain the attractiveness this argument presents. “What we are confronting here is a paradox — a set of inconsistent propositions, (I), (II), and (III), each of which has considerable independent plausibility”(Cohen, 100). Therefore, a resolve of skepticism will both save our intuition that we know many things and explain the appeal of the paradox.

The alternative to a resolve would be biting the bullet, siding with the skeptic that we cannot and do not have knowledge. We grant our assumptions to be mere illusions and concede attainment of truth. This skeptical view is surely worrisome to hear and further troublesome to live by as it contradicts the intuitive and pragmatic claim that people know many things. For why must it be that knowledge is beyond our grasp, that we cannot even know simplicities as there being hands? The only way to change the mind of the skeptic is, perhaps, with an absolute guarantee, one possibly searched for by way of the entailment principle which (roughly) states that subject S knows proposition p based on evidence E only if E guarantees p (Cohen, p.96). Yet strong anti-skepticism– the defeating of skepticism so to also change the view of the skeptic– isn’t achieved through entailment as this principle is compatible with skeptical hypotheses. For any knowledge of hands could likewise be illusions of hands in which both cases obtain by the same evidence, namely the perceptual experience of hands. Albeit, it is not stipulated that the skeptic’s view must change under modest anti-skepticism. This approach calls for our (the non skeptics) satisfaction that we can justifiably believe and know propositions. Justification is the evidence or reasons supporting one’s beliefs, and just how good the evidence or reasons have to be to constitute knowledge is a major component of Cohen’s theory of contextualism.

CONTEXTUALISM

Contextualism is the view that ascriptions of knowledge are context sensitive. Cohen defends the contextualist view that sentence utterances containing predicates, such as “tall”, “flat”, and (in particular to Epistemic Contextualism) “knows”, etc., are assigned truth values determinable by the standards operating within the context of a particular utterance. Therefore, utterances of the form “S knows that p” are judged by assessing the correctness of the speaker’s use, if in regarding the operative standards of the context the utterance was applied correctly or incorrectly and thereby spoken truly or falsely respectively. Consequently, in fixed scenarios different speakers may say contradicting utterances of similar form while yet having both spoken correctly, truly. Again by this point, the same “S knows that p” uttered by distinct persons may result in only one speaker having done so correctly.

Out at the cliffs I say truly “this rock is flat” when suggesting a possible seat for a friend. A while later, after failing to balance the tripod supporting her camcorder she utters “this rock isn’t flat” correctly. As it’s the case that both utterances are true we can appropriately derive that the rock (the boulder that was the subject of both ascriptions) has been attributed the properties of being both flat and not flat. Contradiction! However, this contradiction will not manifest of the rock itself nor is it an objective truth about the external world. (It’s a contradiction! Though admittedly it fails to suffice either for a different reason.) Objectiveness and external world manifestations are not granted in epistemic contextualism. What the contextualist theory provides, rather, is a method for evaluating the truth value of ascription instances, sentence utterances. So a successful evaluation of the first utterance “this rock is flat” will look at the speaker’s use as it applies to the operative standards which in this case is the relation to the comparison class of seat-worthy objects. In other words, I spoke truly because my use of the term “flat” meant relative to ‘flat enough to for someone to sit on’. Subtle change in context to the time of my friend’s utterance “this rock isn’t flat” changes the standards operating. She, being the speaker, then used the term “flat” to mean relative to ‘flat enough to balance a tripod’ thereby making her use correct. By recognizing this shift in meaning with the uses of the term “flat” we should realize that there was in fact no contradiction, but instead a difference in the implicit comparison class referents.

I want to (very) briefly discuss David Lewis’s account of contextualism as it figures into the above example in order to draw a distinction between the two views and, in so doing, further illustrate the mechanics of Cohen’s theory. Lewis is another proponent of the idea that ascriptions of knowledge are context sensitive. However, he claims proper ignorance within the context containing utterances of the form “S knows that p” is the factor that determines their truth value. What may or may not be proper to ignore is any possible world W that one’s evidence does not rule out, remains uneliminated. Lewis claims evidence is perceptual and memory appearances and, to restate, just the subjective experiences of each (remember the contextualism divergence from objectiveness). A W is eliminated just in case that W fails to obtain identical evidence. Now, for any particular evidence for any proposition there will remain a fixed number of W’s which may not be eliminated and it is of these uneliminated W’s that shifts in context change which W’s may be properly ignored. Applied to the example above I spoke truly because I was ignoring those W’s that made salient the bumpiness of the rock beyond that which I was considering when I said it was “flat.” As my friend made salient those bumps shes could no longer correctly say the rock is “flat” and so too spoke truly. So on Lewis’s account the contextual term is fixed in that it always means simply what the term means and context determines which W’s the speaker is allowed to ignore. Uses of contextual terms on Cohen’s view, however, shift meaning when context changes so that the term means what it does relative to the comparison class picked out by the operating standards.

The above example outlined possible operating standards and their functions but was rather simple as it illustrated the mechanics applied to simpler predicates, in this case “tall.” Complexity follows in considerations of ascriptions containing the term “knows” and its cognates as this predicate will not refer to any particular comparison class. If someone uttered “I know I have hands” it just would beg the question to implicitly relate any comparison class such as ‘knows by being a human’ or ‘knows by being a glove wearer.’ For this reason the context sensitivity of “knows” shifts degrees of knowledge, knowing by high standards or knowing by low standards. So the context will now determine the degree of knowledge the speaker is relating to with each use of the term “knows.” As Cohen agrees with the claim that knowledge is justified, true belief, this shiftiness of knowledge is correlated with the shiftiness of justification– the evidence or reasons one has for believing a proposition which are then evaluated against the operating standards within the context.

What then are the factors that compose the operating standards within a context? “Knows” has two distinct categories of factors, namely evidence and risk. How reliable one’s evidence is will set the degree of justification attainable to a certain degree. So I might look in the sky today and utter “I know it will be sunny next Wednesday.” Surely we would not count this as knowledge, for how the sky appears today is not reliable evidence for the weather of next week. “Knowing” next week’s weather demands high standards, perhaps a .85 to .9 degree threshold of justification whereas my evidence only grants me .3 degrees of justification. If instead I were to check predictions of a highly rated weather station whose forecast predicted sun next Wednesday, this evidence might grant .85 degrees of justification which would then make my utterance correct. Notice, further, how each day nearer that of next Wednesday that I check the weather report, which still predicts sun, my degrees of justification may increase to .855 to .86 to .865, etc. This is because the fallibility of the meteorologist decreases as each day brings Wednesday closer to the present making weather easier to predict and therefore strengthening my evidence. Now, let’s say next Wednesday is my family reunion picnic where all my out of town relatives will come down for just that one day. This adds risk to the picture, for if it rains the picnic will not happen. As there is now a consequence if I am wrong in “knowing it will be sunny” this shifts the degree my weather report evidence provides and so I might no longer be correct in my utterance.

Salience relations are essential in context of ascriptions of “knows.” To use the example, it was the importance that the weather be good on Wednesday that made the chance of error salient. Only at the thought of cancelling the picnic did I recognize the possibility of rain, but once I did acknowledge this possibility it was then operative within the context of my ascription. As a chance of error was now salient it’s relation to the context had me, the ascriber, presuppose, imply, stricter standards. It will not always be the case, however, that I demand those stricter standards when uttering similar ascriptions of knowledge even with the chance of incompatible possibilities. This is because Cohen’s contextualism endorses fallibilism, which states regarding those incompatible possibilities that “S can know p on the basis of (reasons, evidence) R even if there is some alternative to p, compatible with R” (Cohen, 95). This allows us to know propositions even if there is a chance of error, though if context makes it so that the chance of error is salient our evidence will no longer suffice the standards demanded.

MAIN THESIS

Contextualism argues that uses of “knows” will mean different things depending on context by implying either ‘knows by low standards’ or ‘knows by high standards.’ In our everday use of “knows” we imply those lower standards and so speak correctly in utterances of “I know I have hands.” The lower standard implied sets a threshold of justification that our evidence suffices to meet and so express true statements. Yet, if an error is made salient, then our evidence will need to be strengthened before we may again express “knowing” correctly. Failure to recognize this contextual shift is attributed to the attractiveness of BIV skepticism.

When observing the BIV argument, it is common to mistake the meaning of the term “know.” Again (1) reads as: “If I know that I have hands, then I know I am not a BIV.” The utterance “I know I have hands” is true in contexts in which operative standards call for a threshold of justification the speaker’s evidence meets. But as the possibility of being deceived is introduced the context is shifted to demand a threshold of justification that is much higher, and not just for the second application of “know” in the premise but of both uses. And so by highlighting the failure to acknowledge contextual shifts the contextualist has met one condition of the resolve to skepticism, he has explained the arguments appeal.

But what of the other condition of the resolve? How are we to defeat this argument so we may reject its conclusion? It is here that Sosa posits his dilemma for contextualism for when it attempts to refute skepticism. First, Sosa worries that if contextualism were to agree with skepticism and in so doing deny the claim that anyone “knows” they have hands, then what good will it do for epistemology to determine the truth value of people’s sentence expressions. To settle this contextualism must claim that people express their beliefs and that these beliefs may have a great deal of justification. The context determines the threshold of justification and so by evaluating utterances by their ability to suffice the threshold it will be granted that true expressions carry a great deal of epistemic goodness. As it is, correct ascriptions of knowledge are expressions of justified, true belief. Here Sosa posits the second worry, namely that the BIV skeptical argument is presented in such a context that it prevents any progression from the nadir of justification. In other words, BIV skepticism falls under the category of global skeptical alternatives- those alternatives that “are immune to rejection on the basis of any evidence.” (Cohen, 103). Under the context of BIV skepticism there is no such evidence that could grant any justification as the salience of error is both commonplace in any evidence and demands stricter standards. So contextualism needs epistemic goodness to explain it’s worthiness to epistemology, yet cannot attain any epistemic goodness under the context of BIV skepticism.

The idea for contextualism was right, evaluate the sentence expression itself, the actual real world instance to determine knowledge. The theory even managed to answer the half of the resolve in which most other attempts fail. Though by its own construction, Contextualism has failed to defeat skepticism.