1 The Analysis of Knowledge

Chapter Learning Outcomes

Upon completion of this chapter, readers will be able to:

1. Identify the main types of knowledge, the relationships among them, and their distinguishing characteristics.
2. Evaluate analyses of concepts, in particular the traditional analysis of knowledge.
3. Assess the value of conceptual analysis, including its relevance to other topics in epistemology.
4. Explain the role of analysis in shaping the history of the field.

Introduction: Conceptual Analysis

Knowledge is the central concept of traditional epistemology. But what is knowledge? This is the most basic question about the central concept, and so the appropriate starting place. Answers traditionally come in the form of conceptual analysis: a set of more basic concepts out of which the analyzed concept is built, arranged to form a definition. The concept “square,” for example, is analyzable into components such as “four-sided figure,” “right-angled,” and “equilateral.”^[1] Our focus here is the analysis of knowledge. But we’ll also consider critiques of this focus, which yield useful insights and prompt new directions of inquiry. The chapter closes with a reflection on the value of epistemological conceptual analysis.

Kinds of Knowledge

Before undertaking analysis, our target concept needs refinement. “Knowledge” is an umbrella term, capturing a family of related meanings:

1. Ability (procedural) knowledge: knowledge-how (e.g., I know how to ride a bike.)
2. Acquaintance knowledge: knowing a person, place, or thing (e.g., Plato knew Socrates. He also knew Athens well.)
3. Phenomenal knowledge: knowing “what it’s like” to have a given experience (e.g., Stella knows what strawberries taste like.)
4. Propositional knowledge: knowledge-that (e.g., Everybody reading this chapter knows that it is about knowledge.)

What the first three have in common is that they require direct experience with their objects. I know how to ride a bike because I’ve had practice; I don’t know how to fly a plane, since I lack training—despite having memorized the manual. Plato knew Socrates and Athens because he studied under the man and lived in the city; Plato knew neither Homer nor London because he neither met the poet nor visited the place. Plato knew of Homer, and propositions about him, but nothing concerning London. Stella knows what strawberries taste like (having eaten them before), but not what it’s like to be a bat given her lack of batty experiences (see Box 1).^[2]

Box 1 – Nagel’s Bat

In his influential 1974 paper “What Is It Like to Be a Bat?” philosopher Thomas Nagel explains that for something to be conscious, “there is something it is like to be” that thing—“something it is like for” that thing to be (436). Thus, consciousness essentially has a “subjective character” in that it requires a first-person “point of view.” As such, no conscious state can be fully grasped or explained from the purely objective third-person perspective (nor from a God’s eye “view from nowhere”). From this, Nagel draws a metaphysical conclusion: that the mental cannot be reduced to the physical. More pertinent to this chapter is an important epistemological implication: that we cannot know “what it’s like” to have experiences that are radically unlike those we’ve actually had. He uses his now-famous bat example to illustrate:

Bats, although more closely related to us than those other species, nevertheless present a range of activity and a sensory apparatus so different from ours that the problem I want to pose is exceptionally vivid (though it certainly could be raised with other species). Even without the benefit of philosophical reflection, anyone who has spent some time in an enclosed space with an excited bat knows what it is to encounter a fundamentally alien form of life.

I have said that the essence of the belief that bats have experience is that there is something that it is like to be a bat. Now we know that most bats (the microchiroptera, to be precise) perceive the external world primarily by sonar, or echolocation, detecting the reflections, from objects within range, of their own rapid, subtly modulated, high-frequency shrieks. Their brains are designed to correlate the outgoing impulses with the subsequent echoes, and the information thus acquired enables bats to make precise discriminations of distance, size, shape, motion, and texture comparable to those we make by vision. But bat sonar, though clearly a form of perception, is not similar in its operation to any sense that we possess, and there is no reason to suppose that it is subjectively like anything we can experience or imagine. This appears to create difficulties for the notion of what it is like to be a bat. (438)

Whereas Eastern and some recent Western epistemology emphasize experiential knowledge (see Monica C. Poole on feminist epistemologies in Chapter 8 of this volume), traditional Western epistemology emphasizes propositional knowledge.^[3] Such knowledge can be expressed with a that-clause, which expresses a proposition: a statement or claim with a truth value (that is, something that is either true or false).^[4] The proposition that this chapter is about knowledge is true; the proposition that it’s about waterfall photography is false.^[5]

Propositional knowledge can be interpersonally communicated or acquired by evidence or argument. By contrast, experiential knowledge can be neither argued for nor linguistically transferred. Try as I might to describe the taste of strawberries, it’s not the same as knowing what they’re like. Someone who has never had the pleasure will still learn something new upon their first bite.

Despite the importance of experiential knowledge, we’ll explore the traditional approach in this chapter. For brevity’s sake, then, “knowledge” here refers to the propositional variety.

The Traditional Analysis

The most influential analysis of propositional knowledge derives from Plato (ca. 428–347 BCE). In his Meno dialogue, Plato’s character Socrates (modeled after his real-life teacher) argues that “knowledge differs from correct opinion in being tied down” by “an account of the reason why” ([ca. 380 BCE] 2009, 98a).^[6] Table 1 shows how this translates into modern parlance.

Table 1 – Platonic to Modern Translation

Platonic Term	Modern Term	Abbreviation
Opinion	Belief	B
Correct	True	T
Account of the reason why (logos)	Justification	J
Knowledge	Knowledge	K

This translation yields the traditional analysis of knowledge, or “JTB” analysis: knowledge is “justified true belief.” On this account, there are three concepts that pairwise overlap, and knowledge is the convergence of all three (see Figure 1). Let’s consider each in turn.

A. Belief

Belief (specifically belief-that^[7]) means accepting the proposition as true (equivalently: assenting to the proposition, thinking that it’s true, agreeing with it, or holding it as an opinion/view). Belief can range from a slight leaning to moderate assurance to absolute certainty—the entire positive half of the confidence spectrum (see Jonathan Lopez on degrees of belief in Chapter 6 of this volume).^[8] Belief excludes both the negative half of the spectrum (disbelief, or belief that the proposition is false) and the neutral, halfway point (suspending/withholding judgment).^[9] Belief, disbelief, and suspension are the main doxastic attitudes (stances on the truth value of a proposition).

On the traditional analysis, knowing a proposition requires believing it. If a truth you’ve never thought of is “out there” awaiting discovery, you don’t know that truth. If you are now thinking about it but form no opinion (suspension), you still do not know. This is why, when asked about the truth value in cases of suspension, the natural answer is “I don’t know.” And if you have settled your opinion against the proposition (disbelief), you again do not know it. Suppose I ask, “Do you know that Marie Curie led the underground railroad?” You won’t say, “Yes, I do know that.” Instead, you’ll deny it, perhaps offer a correction. This reaction is not best explained by what is actually correct but by what you believe is correct, since you would respond in the same manner if the question were instead about a matter on which you were convincingly misled (say, by reading a misprint in a seemingly trustworthy textbook).

Bringing these points together gives us a process-of-elimination argument. So far, we have determined that you lack knowledge of (a) propositions you have not considered, (b) propositions on which you suspend judgment, and (c) propositions you disbelieve. The only remaining candidates for knowledge are propositions you do believe, such as that Marie Curie did not lead the underground railroad but Harriet Tubman did.

A word of caution: people often speak loosely. Loose talk is language that is inaccurate by strict literal standards—such as metaphor, hyperbole, approximation, and ellipsis (word omission). This phenomenon sometimes causes mistaken evaluations of conceptual analyses, since the aim of analysis is the strict literal truth. Consider the expression “I don’t believe it; I know it.” A natural interpretation is that one doesn’t merely believe it, where “merely” is omitted to achieve brevity (and for rhetorical effect). We use such elliptical speech routinely. Consider: “She’s not good at math; she’s great!” But if she’s not even good, she’s not great, since greatness is a degree of goodness. Let’s rephrase: “She’s not just good at math; she’s great.” This illuminates what was previously disguised—that the “not” negates a lesser degree rather than goodness altogether.^[10]

B. Truth

Belief is one thing; truth is another. There are unbelieved truths (the earth was an oblate spheroid long before it occurred to anyone) and believed falsehoods (such as Ptolemy’s geocentric model of the universe). The problematic phrase “true for me” confuses this issue. Ptolemy’s view may have been “true for him,” but this merely means he accepted it, not that it was actually true.

Acceptance and truth can come apart because human opinion is not a perfect measure of reality. We are capable of mistakes. Acknowledging this is not a weakness but an expression of intellectual virtues (such as intellectual honesty and humility) that motivate inquiry, open-mindedness, and collaboration. Just as we sometimes recognize our own mistakes, we sometimes recognize that others are mistaken. The situation may require speaking up about this (in an appropriate fashion); in other cases, we should keep it to ourselves. Either way, prospective falsehood is why it’s a bad idea to believe just anything anyone says. We often need to reflect for ourselves and formulate beliefs independently. Between intellectual deference and autonomy lies virtuous inquiry. (For more on social dimensions, see William D. Rowley on social epistemology in Chapter 7 of this volume.)

But what is truth? In Aristotle’s famous words, “To say of what is that it is not, or of what is not that it is, is false, while to say of what is that it is, and of what is not that it is not, is true” ([ca. 350 BCE] 2009, 1011b). This is an ancient precursor to a popular modern starting point—the correspondence theory: a proposition is true if it corresponds to reality, and false otherwise. While there are alternative theories, it is possible to interpret them as different takes on “correspondence.” Details won’t matter here.^[11]

Only true beliefs can qualify as knowledge on the traditional analysis. Suppose you claim to know the answer to a trivia question. The answer is revealed and you got it wrong. Your friend exclaims, “See, you didn’t know it!” This reaction is perfectly natural because false belief isn’t knowledge. This explains why teachers often grade factual questions based on whether students give correct answers: their purpose in such cases is to test knowledge, and whether students answer correctly is such a test—precisely because of the truth condition on knowledge.

Again, loose talk skews intuition. Several books and a Weird Al Yankovic album are titled Everything You Know Is Wrong. Even Mark Twain purportedly quipped, “It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.” However, Shakespeare’s King of Denmark had it right when he proclaimed, “what we know must be” ([ca. 1600] 1998, I.ii.98). For, if it ain’t so, you don’t really know it. At best, you merely think you know it. Knowledge is factive (entails truth), whereas belief is non-factive (possibly wrong).^[12]

C. Justification

We’ve seen that knowledge requires true belief. But even true beliefs can be unjustified. A justification is a good reason for belief (see Todd R. Long in Chapter 2 of this volume for theoretical accounts). On the traditional analysis, justification is necessary for knowledge. To understand why, suppose you are playing trivia again (apparently, you’re hooked):

“What is the name of those tiny bumps on blackberries?”

Your guess: Choice D – Druplets.

Desperate to win, you rationalize: “Yeah, this has to be right.”

The answer is revealed, prompting your proud reaction: “See, I knew it!”

Your friend remarks, “No, you didn’t. You were just guessing!”

Your friend’s response is natural. Absent good reason, one does not know.

Plato offered an analogy. Consider the statues of the mythical inventor and sculptor Daedalus, which were said to be so realistic they could run away. Unless they were tethered down, you never knew where to find them. Mere true beliefs are akin to the untethered statues: they are sometimes found by sheer luck. Justification is similar to the tethering: it anchors true beliefs in good reasons, turning them into knowledge. Another oft-used analogy is that justification functions as a good road map to the desired destination (truth). Knowledge, like the successfully navigated journey, like the tethered statues, enjoys a kind of stability. This makes evident why justification plays a crucial role in the value of knowledge (see Guy Axtell in Chapter 5 of this volume on epistemic value).

Here, too, loose talk misleads: “The thermometer ‘knows’ the temperature”—but surely lacks justification. The justification condition is also dubious if inflated, as in Plato’s description. Knowledge doesn’t require “an account of the reason why” a belief is true so much as a reason that it’s true.^[13] One can know that a computer works but be clueless why. A reason-that need not be sophisticated. No argument or scientific demonstration is necessary. Just turn on the computer and see it working, recall this from memory, or be told by the technician testing it. Nor do good reasons have to be perfect. The concept good is weaker than perfect (maximally good). If perfect reasons were required, justification would be impossible (mere mortals are always subject to error).^[14] Tolerating imperfect reasons fits everyday judgments. In grade school, I had reason to believe Newtonian physics: I had testimony from trustworthy teachers and textbooks and no reason to suspect oversimplification. My belief was justified—a belief I now recognize as false given quantum mechanics and Einsteinian relativity. Justified beliefs can be false—a view called fallibilism about justification (to be distinguished from fallibilism about knowledge).^[15] For this reason, a separate truth condition on knowledge is not redundant.

Another challenge to the justification condition is the common attribution of knowledge to infants and (non-human) animals. Are such attributions mere loose talk? It’s unclear. Do infants and animals have a kind of weak justification? Difficult to say. Perhaps they know without justification. If so, we can distinguish two kinds of knowledge. Infants and animals have lightweight knowledge (true belief) but lack heavyweight knowledge—the kind we seek beyond mere correct opinion, where guessing and poor reasoning are precluded (Hawthorne 2002). The traditional analysis is meant to capture this heavyweight variety.

Counterexamples to the Traditional Analysis

Since justification seems to distinguish mere true belief from (heavyweight) knowledge, its addition completes the analysis—or so it seemed to many for 2400 years! The JTB analysis became Western philosophical heritage until Edmund Gettier (1927–2021) with his three-page article in 1963.^[16]

Gettier argued against the traditional analysis by counterexamples (examples that refute). His counterexamples are cases of JTB that aren’t knowledge. Since the original examples are intricate, we will consider more straightforward examples with the same gist. Such examples are called Gettier cases.

You’re driving through sheep country. Passing a field, you seemingly see a sheep and think, “There’s a sheep in the field.” Normally, this suffices for knowledge: you have a belief, a visual perception supports it, and there’s a sheep in the field. The kicker: you’re looking at a sheep-shaped rock, or a wolf in sheep’s clothing!^[17] There’s no way to tell from your angle. You have no reason to suspect. How is it true, then, that there’s a sheep in the field? Unbeknownst to you, there happens to be one out of sight, in some far-off corner of the field. Intuitively, you don’t know there’s a sheep.

You may not initially share this intuition (I didn’t at first). Sometimes intuitions need to be massaged or pumped before they surface. Here’s an intuition pump. Consider a revised scenario: the real sheep has been removed. Since it was out of sight, you won’t be able to detect any change. So, for all you know, nothing has changed. This means your state of knowledge should be the same as before. But in the revised scenario, it’s clear you don’t know: a sheep you don’t know about can’t help you know there’s a sheep. Since your state of knowledge is untouched by the revision, you didn’t know in the first place.

Thinkers had discovered this Gettier problem long before Gettier rediscovered it and made it famous, including the fourteenth-century Italian logician Peter of Mantua (Boh 1985). As early as the eighth century, the Buddhist philosopher Dharmottara devised a case: a desert traveler seeing a water mirage where there is real water underneath a rock has a justified true belief without knowledge (Dreyfus 1997). Spanning time and culture, such intuitions are widely and independently attested.

Revised Analyses

Gettier never published a solution to his own problem, but he did prompt others to search for a fourth condition on knowledge. The idea is that knowledge is JTB plus some extra condition that rules out the problematic cases—JTB+ accounts. There’s insufficient space to review these proposals here. Suffice it to say that the extra condition remains elusive. Perhaps the problem is that JTB+ carves up knowledge such that the plus fails to match any natural concept. Cut out all the best-decorated pieces from a birthday cake; those portions may be nice. But the remainder has no identifiable shape.

Returning to Plato’s footsteps, it may be more promising to seek what distinguishes true belief from knowledge—a TB+ account. As Alvin Plantinga defined the term, warrant is that “elusive quality or quantity enough of which, together with truth and belief, is sufficient for knowledge” (1993, v). It follows that knowledge is sufficiently warranted true belief, yielding an sWTB account. Now our question shifts: What is warrant?

This shift has potential advantages. First, while the sWTB approach is compatible with JTB+ accounts, it is also compatible with abandoning the justification condition, as some prefer.^[18] So, sWTB may bypass this debate. Second, there’s a kind of unity to warrant that justification lacks. To see this, we need to explore the concept of epistemic luck: the kind of luck that affects one’s epistemic status.

Let’s take stock of the various forms of epistemic luck. Gettier cases are ones in which good luck cancels bad (Zagzebski 1994). In the sheep case, you’re unluckily misled by a sheep shape over here, but luckily this mistake is corrected by a real sheep over there. By contrast, lottery cases seem better construed as involving a single element of chance. Luck in Gettier and lottery cases doesn’t threaten justification. So, plausibly, the luck involved in acquiring truth via unjustified belief (e.g., pure guesswork) is yet another kind.

Matters aren’t so simple. Some epistemic luck contributes positively to knowledge. Suppose you read a newspaper and tell me all about it. I attribute knowledge to you. When I find out that you only read it because you luckily won a free subscription, I am not inclined to retract my knowledge attribution. This knowledge is founded on good epistemic luck, the kind which enables one to be lucky to know. Veritic luck is the knowledge-precluding kind, which includes all of the various forms identified in the previous paragraph: Gettier-luck, lottery-luck, and lucky guessing (Engel 1992). One fascinating aspect of warrant, unlike justification, is that warrant rules out all and only veritic luck.

But what connection between belief and truth accomplishes this? What exactly are the conditions that secure warrant and exclude veritic luck, resulting in knowledge? We don’t have space to explore all candidates. I’ll mention one promising direction as an example, which draws the parallel between belief and action. Imagine an expert archer, Artemis (the Greek goddess of wild animals and the hunt, also known as the Roman goddess Diana). Her aim is perfect. Her release is perfect. The arrow is going to hit the bullseye—until Poseidon (the Greek god of the sea) mischievously slams his trident into the seabed, causing an earthquake, which shifts the target. A simultaneous gust of wind from the breath of Aeolus (the keeper of the winds) alters the arrow’s path, serendipitously correcting course. In this scenario, skilled Artemis sees success, yet her skill is not the reason for success. Whenever her success is instead attributable to skill, it is to her credit rather than luck. Similarly, perhaps knowledge is “credit for true belief” (Greco 2003). Knowledge is achieved when intellectual skill/excellence/virtue manifests in success (truth). So, knowledge is virtuously achieved true belief (Sosa 1980). From this observation originates virtue epistemology, the study of intellectual virtue and its relationship to knowledge.

Conclusion: Post-Gettier Epistemology

Fast-forward several decades. Thousands of pages of ink have been spilled on the fourth condition, warrant, veritic luck, the knowledge-yielding virtues, and so forth. Some believe they have the solution. Others continue to pursue new solutions. Perhaps you will be the one to find it! For now, there’s no agreed-upon answer. We live in a post-Gettier age: the problem no longer occupies center stage. Still, it inspired what came next.

In the aftermath, some epistemologists came to suspect that knowledge is not subject to analysis—that no component can be added to (J)TB to get knowledge (Zagzebski 1994). If true, this doesn’t render knowledge mysterious. Some concepts are basic, and perhaps knowledge is one of them. Yes, knowledge may entail JTB, but this does not mean it can be divvied into neat chunks that seamlessly reassemble without remainder. This gave birth to knowledge-first epistemology, advocated most prominently by Timothy Williamson (2000).

Others abandoned concern with knowledge altogether. What Gettier (and lottery) cases reveal, they say, is that knowledge is a concept with quirks. Who cares whether one is Gettiered (or “lotteried”)? What matters is acquiring the truth, having good reasons, or achieving intellectual virtue more generally (e.g., understanding, open-mindedness, curiosity, humility).^[19] Thus, virtue epistemologists began investigating the intellectual virtues in their own right (Zagzebski 1996).

Whatever tack one takes, there is one remarkable thing on which we can agree: Gettier’s little paper permanently transformed the world of epistemology. It planted seeds in an ever-growing garden of fruitful new directions, producing some of the most fascinating work the field has seen: work on epistemic luck, epistemic value, intellectual virtue, and more. Thus, conceptual analysis, even when unsuccessful, reveals insight. Much of what follows in this book we owe in large part to that.

Further Reading

Engel, Mylan Jr. n.d. “Epistemic Luck.” In The Internet Encyclopedia of Philosophy. https://iep.utm.edu/epi-luck/

Gettier, Edmund L. 1963. “Is Justified True Belief Knowledge?” Analysis 23 (6): 121–3. https://fitelson.org/proseminar/gettier.pdf.

Graham, David A. 2014. “Rumsfeld’s Knowns and Unknowns: The Intellectual History of a Quip.” The Atlantic, March 27, 2014. https://www.theatlantic.com/politics/archive/2014/03/rumsfelds-knowns-and-unknowns-the-intellectual-history-of-a-quip/359719/.

Hetherington, Stephen. n.d. “Gettier Problems.” In The Internet Encyclopedia of Philosophy. https://www.iep.utm.edu/gettier/.

Plato. (ca. 380 BCE) 2009. Meno. Translated by Benjamin Jowett. The Internet Classics Archive. http://classics.mit.edu/Plato/meno.html.

———. (ca. 369 BCE) 2013. Theaetetus. Translated by Benjamin Jowett. Urbana, IL: Project Gutenberg. https://www.gutenberg.org/files/1726/1726-h/1726-h.htm.

References

Aristotle. (ca. 350 BCE) 2009. Metaphysics. Translated by W. D. Ross. The Internet Classics Archive. http://classics.mit.edu/Aristotle/metaphysics.html.

Boh, Ivan. 1985. “Belief, Justification and Knowledge: Some Late Medieval Epistemic Concerns.” Journal of the Rocky Mountain Medieval and Renaissance Association 6: 87–103.

Chisholm, R. M. 1966. Theory of Knowledge. Englewood Cliffs, NJ: Prentice Hall.

Dreyfus, George B. J. 1997. Recognizing Reality: Dharmakirti’s Philosophy and its Tibetan Interpretations. Albany, NY: SUNY Press.

Engel, Mylan Jr. 1992. “Is Epistemic Luck Compatible with Knowledge?” Southern Journal of Philosophy 30 (2): 59–75.

Gettier, Edmund L. 1963. “Is Justified True Belief Knowledge?” Analysis 23 (6): 121–3.

Glanzberg, Michael. 2018. “Truth.” In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta. https://plato.stanford.edu/archives/fall2018/entries/truth/.

Greco, John. 2003. “Knowledge as Credit for True Belief.” In Intellectual Virtue: Perspectives from Ethics and Epistemology, edited by Michael DePaul and Linda Zagzebski, 111–34. New York: Oxford University Press.

Harman, Gilbert. 1968. “Knowledge, Inference, and Explanation.” American Philosophical Quarterly 5 (3): 164–73.

Hawthorne, John. 2002. “Deeply Contingent A Priori Knowledge.” Philosophy and Phenomenological Research 65 (2): 247–69.

Hazlett, Allan. 2010. “The Myth of Factive Verbs.” Philosophy and Phenomenological Research 80 (3): 497–522.

Moon, Andrew. 2017. “Beliefs Do Not Come in Degrees.” Canadian Journal of Philosophy 47 (6): 760–78.

Nagel, Thomas. 1974. “What Is It Like to Be a Bat?” Philosophical Review 83 (4): 435–50.

Plantinga, Alvin. 1992. Warrant: The Current Debate. New York: Oxford University Press.

———. 1993. Warrant and Proper Function. New York: Oxford University Press.

Plato. (ca. 380 BCE) 2009. Meno. Translated by Benjamin Jowett. The Internet Classics Archive. http://classics.mit.edu/Plato/meno.html.

———. (ca. 369 BCE) 2013. Theaetetus. Translated by Benjamin Jowett. Urbana, IL: Project Gutenberg. https://www.gutenberg.org/files/1726/1726-h/1726-h.htm.

Radford, Colin. 1966. “Knowledge: By Examples.” Analysis 27 (1): 1–11.

Ryle, Gilbert. 1949. The Concept of Mind. Chicago: University of Chicago Press.

Shakespeare, William. (ca. 1600) 1998. Hamlet. Edited by Sylvan Barnet. New York: Signet Classics.

Sosa, Ernest. 1980. “The Raft and the Pyramid: Coherence versus Foundations in the Theory of Knowledge.” Midwest Studies in Philosophy 5: 3–25.

Weatherson, Brian. 2003. “What Good Are Counterexamples?” Philosophical Studies 115 (1): 1–31.

Williamson, Timothy. 2000. Knowledge and its Limits. New York: Oxford University Press.

Zagzebski, Linda. 1994. “The Inescapability of Gettier Problems.” The Philosophical Quarterly 44 (174): 65–73.

———. 1996. Virtues of the Mind: An Inquiry into the Nature of Virtue and the Ethical Foundations of Knowledge. Cambridge: Cambridge University Press.