WILLIAMSON:KNOWLEDGE & ITS LIMITS PAPER

by Timothy Williamson

Knowledge and its Limits presents a systematic new conception of knowledge as a kind of mental stage sensitive to the knower's environment. It makes a major contribution to the debate between externalist and internalist philosophies of mind, and breaks radically with the epistemological tradition of analyzing knowledge in terms of true belief. The theory casts new light on such philosophical problems as scepticism, evidence, probability and assertion, realism and anti-realism, and the limits of what can be known. The arguments are illustrated by rigorous models based on epistemic logic and probability theory. The result is a new way of doing epistemology and a notable contribution to the philosophy of mind.

Genres: Philosophy, Nonfiction

353 pages, Paperback

First published November 2, 2000

Reviews

[JDP · Rating 3 out of 5]

this book is only -- and i mean this -- only for scholars interested in contemporary analytic epistemology. williamson is clearly a brilliant thinker, but this book showed me everything that is wrong with externalisms in epistemology. it was challenging, provocative, and interesting, but i judge almost all of it to be incorrect. the low stars are on account of its falseness as well as its difficult readability.

[Bruce Crown · Rating 2 out of 5]

Difficult text to understand and simply a rebellion of the JTB debate viz. Plato all the way to Gettier. But still constrained by all the old dogmas imprisoned in history… Williamson thinks he's innovating and doing something new by "not adhering to dogmatic prejudices" but he creates a dogma that requires you to prejudice yourself.

[Leo Loos · Rating 3 out of 5]

A hard book to get a grip on. In my opinion even a bit disappointing. Williamson does many fancy maneuvers but left me at the same point as many sceptics would when it comes to hard/edge cases. Though knowledge is a mental state, we have no way of knowing that we are in a position to know anything and the tools given to us are utterly prone to the same problem. The idea that justified evidence for instance would make knowledge more probable seems not to solve the problem either. In the end, sometimes we do know and sometimes we merely believe (falsely) but there is no way for the subject to differentiate between these states.

[George · Rating 3 out of 5]

Williamson is brilliant, but this book is unnecessarily hard and opaque to read.

[Aung Sett Kyaw Min · Rating 2 out of 5]

One of the arguments put forward by Williamson in favor of taking knowledge to be conceptually more basic than belief in its various modifications (rational, true, etc.,) is that doing so dissolves the problem of regress of epistemic justification. As the problem goes, epistemic responsibility demands that the content of our beliefs and/or belief-states themselves, be justified by reasons for these beliefs which must be justified by yet further reasons, etc., Either justification traverses through an endlessly linear or circular chain of reasons or it comes to a stop. If it comes to a stop, then how is this last item justified? Look like it must be justified by appealing to another justifier and so on, ad infinitum

Fortunately, this regress becomes more tractable once we consider the option that it is not, as it is commonly held, belief, but rather knowledge that can function as the justifier. In Williamson's view, the epistemic regress of justification in its various guises arises due to the internalist cleavage between 'absolute' justification (a Bayesian justification relative to one's knowledge) and 'relative' justification (inferential, deductive, etc.,) which is bridged by the knowledge-first program's identification of absolute justification, evidence, and knowledge. Hence, Williamson's plea for a 'modest foundationalism'.

In other words, Williamson's strategy to dissolve the problem of the regress of justification-affording reasons is by substituting the factive mental state of knowledge (or one's evidence) in place of basic beliefs. Assuming that the causal pathway is non-deviant, and that the agent is able to grasp all the relevant propositions associated with a perceptual experience at hand (Williamson 197), then the resultant state of knowledge enjoys a kind of positive epistemic standing by default, according to Williamson. This strategy displays some affinity with how certain foundationalists understand basic beliefs as justifiers whose original justification is not derived from something else. Some sense datum theorists, for example, hold that one's beliefs about sensory images as pure appearances are by default endowed with automatic justification. However, and this is crucial--they also maintain that the latter could be stripped of this privilege and made sensitive to the demand for epistemic justification and/or warrant in appropriate dialectical contexts.

Likewise, Williamson construes one's evidence (or what is the same, knowledge) whose extensionality in his view far exceeds the bounds of the agent’s first-person phenomenal awareness in the present (KaiL 193), as enjoying a kind of justification by default (KaiL 9). This or that piece of evidence or a package of propositions in one's total body of evidence can be locally detached from the latter and tested against it. But we cannot subject all evidence to doubt all at once or else we will be left with an insufficient amount of knowledge to justify our beliefs.

The traditional epistemological program has not ruled out the possibility of utilizing knowledge as a justifier that can endow the status of justification on beliefs. If knowledge is what justifies one’s beliefs and if the factive mental state of knowing enjoys a degree of prima facie justification as a ‘limiting case’ (KaiL 9) in the probabilistic rather than first-person evidentialist sense of the term justification, then the regress of justification seems to stop at knowledge. However, this means that in order for the knowledge-first foundationalist solution to the problem of regress to work Williamson has to narrowly construe the notion of reasons for beliefs to mean evidence-providing reasons or evidence that increases the odds of its target belief or hypothesis being true.

Assuming one is a reasonably responsible epistemic agent, responsibly held beliefs would thereby constitute quite a decent chunk of the members in one’s total belief system at any given time. Their justifiers--responsibility-affording reasons, likewise need only to be self-evident and transparent to the epistemic agent who then actually endorses them on these grounds without these reasons being in any way attuned to the environment beyond the agent’s inner theater of phenomenal presentation and mental representation. Unsurprisingly, Williamson has no patience for such talk of reasons as purely internal states agents have transparent access to and which they could infallibly diagnose and endorse in an effort to follow this or that norm of rationality (i.e. epistemic responsibility). The positive argument to this effect is offered in Chapter 8 of KaIL. Try as we might, we as rational thinkers are not always in a position to infallibly identify what properties a piece of evidence has or know what the norms of rationality expect of us. Let us grant the assumption that Williamson’s argument for the non-transparency of evidence and rationality holds. This leaves us with Williamson’s own conception of evidence informed by the intersubjective and fallible way evidence is deployed in scientific inquiry (KaiL 193). Evidence is propositional through and through. It is propositional because propositions are the sort of things that a knower must grasp and which can then function as evidence in favor of an explanatory hypothesis (KaIL 195). For instance, sense-experiential states count as a form of evidence because from them one can derive propositions (KaiL 197-8). Putting these pieces together, Williamson’s solution to the problem of epistemic regress is as follows.

As the epistemic agent draws proportional support from evidence for one’s beliefs, there is a distinction to be made between ‘explicitly evidence based beliefs’ and ones that are merely implicit (KaIL 191). Presumably, belief x that is further away from the soil of experience is more likely to be explicitly evidenced based. Such a belief is heavily mediated by the agent’s prior beliefs concerning the evidence for x. For instance, it is apparent that a physicist’s belief that Loop Quantum Gravity is a promising solution that can reconcile quantum mechanics and gravity is explicitly evidence-based. On the other hand, my belief that a soccer ball launched in the air will fall to the ground in an arc is only implicitly evidence-based. This is because such a belief is casually sensitive to evidence in the form of propositions that planets tend to have strong gravitational pull, that there are currently no strong winds blowing that will carry the ball further up, etc.

Crucial to Williamson’s solution to the problem of epistemic regress is the distinction between explicitly and implicitly evidence-based beliefs and the claim that, while an explicitly evidence-based belief must be justified or preceded by an evidenced-based belief, the latter does not necessarily have to be an explicitly evidence-based belief. It is enough that the explicitly evidence-based belief y be preceded by an implicitly evidence-based x which is ‘causally sensitive’ to the evidence about x. Since this causal sensitivity or disposition involves no reasoning it is not sensitive to epistemic justification in a way a belief would. This causal sensitivity is what makes Williamson’s account of evidence a potential solution to the problem of epistemic regress of justification because the explicitly knowledge-based or evidenced-based belief that y is merely causally sensitive in a noncognitive and non-normative way to prior evidence-based belief x that would justify it. Thus, perceptual evidence, which has propositional content, can justify beliefs and hypotheses without the latter needing to be explicitly justified further, at least according to Williamson.

Does Williamson's solution succeed in shoring up skeptical concerns? In my opinion, what might problematic about Williamson’s solution to the problem of epistemic regress is that implicitly evidenced based beliefs which are justified by evidence/knowledge in the form of propositions that so and so can accrue more or less justification over time or when a challenge is posed to the agent to justify the very causal sensitivity of the evidence-based belief that x to other pieces of evidence. Propositional contents are simply true or false, not innately causally sensitive to one another. What might be causally sensitive to one another are mental-states rather than the contents of the mental states themselves. Williamson is of the view that one needs concepts in order to see that x in order to have evidence or otherwise one only sees a situation in which x (KaiL 31, 199-200). Concepts are acquired. This means causal sensitivity to prior evidence that Williamson refers to as property of the mental-state is rather a disposition that mental states of an agent initially acquire and which are continually reinforced over time. If one grants that they are acquired and reinforced through previous and ongoing normative attitudes adopted by agents, then the threat regress seems to loom once more. This way the skeptic can call into question the evidenced-based mental state's causal sensitivity to another evidenced-based mental state because it turns out that causal sensitivity itself is not an innate receptivity but one that is acquired and one for which the agent can take epistemic responsibility. Can Williamson’s foundationalist reply block the skeptic’s challenge?

One obvious way for an epistemic agent to respond to the skeptic’s move to challenge the justification enjoyed by a piece of evidence x in the total body of evidence available to a given community of which the said agent is a member is to suspend whatever prior propositional attitude they might have to x, detach x from the total body of evidence and evaluate whether x is justified relative to the remaining background knowledge. But Williamson stipulates that enough of the latter must survive intact or else inductive justification and Bayesian conditionalization on said evidence is bound to misfire (KaiL 10). When one is adrift at sea and rebuilding one’s raft, one must leave enough pieces of the raft intact. As a foundationalist solution, however, Williamson’s reply is found wanting because it is unable to respond to the skeptic’s global challenge in an immanent way and without appealing the specificities of one’s own theory of evidence which the skeptical would not grant. This is partly because Williamson’s ultimate justifiers--mental states of knowing in the form of propositions that buttress beliefs--can gain or lose justification.

[Lucas · Rating 2 out of 5]

Perhaps the culmination of post-Kripkean analytic philosophy

[Jonathan Badgley · Rating 5 out of 5]

I never took a formal epistemology course so reading this book was always going to be a stretch for me. I’m thankful that I didn’t give up on it after the first two chapters, though it took me three attempts to make progress. Having a strong familiarity from my own casual readings in epistemology with truth as justified, true belief (with modifications on justification all the rage), I was caught off guard/ unsure whether or how I could understand chapter one where Williamson presents knowledge as a Factive Mental State Operator. I wasn’t expecting so much odd philosophy of mind in a book on epistemology. Surely enough, the first chapter is just the ground work, and the rest of book opens up to addressing a number of really invigorating and exciting ideas in formal epistemology.

Chapter notes, very brief:

1. Knowledge is a Factive Mental State
2. Broadness: externalism
3. Primeness: If one is disposed to respond rationally to future evidence, then one’s future prospects are better if one now has knowledge than if one has mere justified true belief
4. Anti-luminosity: One could know something without being in a good position to know that one knows it
5. Margins and Iterations: Connect knowledge and safety. If one knows, one could not have been easily wrong in a similar case
6. Surprise Examination in light of all that’s been said
7. Sensitivity: we should not replace safety with sensitivity as a requirement for knowledge
8. Skepticism: We are not always in a position to know what our evidence is
9. Evidence: Knowledge is evidence, evidence is knowledge
10. Broadly bayesian account in which evidence can be gained as well as lost (pg 206)
11. Assertion: extremely simple claims about the practice of assertion, but illuminating. Shout out to Brandom
12. Structural Unknowability: reviving and reinforcing Fitch’s argument against weak verificationism: for any p, if p then it’s possible to know p
Conclusion: we don’t know nearly as much as we think, externalism, and other interesting details!

[Sreena · Rating 4 out of 5]

"Knowledge is not simply true belief, but the most general condition under which a belief could be true."

What is knowledge?
By critically examining various theories and perspectives, Williamson argues that knowledge is not merely justified true belief, as some philosophers have suggested. Instead, he proposes that knowledge is best understood as a state of mind that is both reliably formed and supported by the available evidence. This redefinition forces us as readers to confront their preconceived notions of knowledge and pushes us to a deeper understanding of its limits and potentialities. And this is what the central theme of the book is.

On of my favorite section
One of the interesting section of this book was when Timothy introduces about a doctor diagnosing a patient's illness. Through this scenario, Williamson explores the nuanced relationship between knowledge and belief, highlighting the potential consequences of misdiagnosis and the inherent uncertainties in medical practice. Such practical illustrations make the book engaging and accessible, even to those unfamiliar with formal philosophy.

Some parts of the book I felt bit repetitive or overly detailed, making it challenging to maintain a consistent reading flow. However, these minor shortcomings do not detract from the overall value and significance of the book.

[Paul Hampson · Rating 5 out of 5]

A new paradigm for epistemology at the time, Knowledge first epistemology, in which the concepts of belief and justification are parasitic on knowledge, was an exciting read. There needed to be some careful, tight argument, which Williamson presents in an accessible manner, since the main formulas are in the appendices.
The book has helped me clear up some thoughts I have on externalism, luminosity, transparency, and the norms of assertion, which are hinted at, and so valuable to, anti-realist accounts of meaning. His treatment of the Fitch cases make an interesting read and a good overview of the literature, and the treatment of evidence and probability illustrate to me what is right with the anti-realist intuition, whilst adopting a realist stance.

I cannot recommend this highly enough!

[Nur Duwek · Rating 3 out of 5]

Don’t read this book unless you really really like epistemology and symbolic logic. Very hard to read, but some really interesting and thought-provoking ideas in the philosophy of knowledge.

[Bailey Ingham · Rating 4 out of 5]

Knowledge first? What about vibes?

[Andre Parola · Rating 5 out of 5]

TW shows us how it’s done!

[Kaidi Pan · Rating 4 out of 5]

Classic.

[Dario Vaccaro · Rating 5 out of 5]

"Knowledge and its limits" is an incredible achievement in formal epistemology, developing a new approach to the analysis of knowledge and its related concepts and tackling dozens of questions and problems while providing original (and usually brilliant) new ideas for each of them. Easily one of the most groundbreaking works in philosophy in the 21st century. While I disagree with the project Williamson sets out for himself, it is a project all epistemologists ought to engage with.

[zimu]

Chapters read: Scepticism, Evidence

[Jon Norimann · Rating 3 out of 5]

This is heavy reading! Williamson works on finding out what does knowing something means and if one can know something. At the core lies the so called Gettier problem. Humes anti induction proof, that some may remember from school, is obviously related aswell.

Gradually the net is widened to discussing related areas like can you know you know something, can you know the truth is inside an interval, can you know the correctness of what you use to get to know something.

10-20% of the book is filled with logics heavy statements of the form a=b & b=c => a=c etc only more complex. Just skip those and you have an interesting discussion of the phenomenon knowledge. Well worth a read but only for those with special interest in the area.