I’ve named this writing project The web of belief because I’m drawn very much to Quine’s metaphor. Beliefs form a network, with some beliefs more central and fundamental, while others are peripheral and subject to revision. My web of belief can be different from yours, but they also have much in common. It’s a helpful way to think about how beliefs hang together.
In the introduction to Volume I of his philosophical papers, David Lewis says:
Our “intuitions” are simply opinions; our philosophical theories are the same. Some are commonsensical, some are sophisticated; some are particular, some general; some are more firmly held, some less. But they are all opinions, and a reasonable goal for a philosopher is to bring them into equilibrium. Our common task is to find out what equilibria there are that can withstand examination, but it remains for each of us to come to rest at one or another of them. If we lose our moorings in everyday common sense, our fault is not that we ignore part of our evidence. Rather, the trouble is that we settle for a very inadequate equilibrium.
I agree, theory is opinion. Science is continuous with philosophy, philosophy is continuous with literature, it’s all opinion. Now, to be clear, some theories are more firmly held. In ordinary conversation, in its day-to-day sense, evolution by natural selection and general relativity are not just mere ‘opinions’. They are backed by science as a collective human activity. In some sense, these theories go beyond mere opinion. Still, in a more abstract sense, it’s opinion.
Let’s put it this way: an opinion is “a view or judgment formed about something, not necessarily based on fact or knowledge”; notice it can be based on fact or knowledge. My opinion, my view, my judgment, is that Quine’s metaphor, the web of belief, is the best way to make sense of the world in the most general sense. Sellars famously said: “The aim of philosophy, abstractly formulated, is to understand how things in the broadest possible sense of the term hang together in the broadest possible sense of the term.” Different realms of understanding—scientific, moral, and everyday experience—connect within a coherent framework. Quine simply points towards one way to think about how all this hangs together, one perspective, which, in my view, is the most helpful.
Quine’s view is summarized in a book called The Web of Belief he wrote with J.S. Ullian and I want to say a few words about this book here.
Chapter 1: Introduction
From the introduction:
nearly any body of knowledge that is sufficiently organized to exhibit appropriate evidential relationships among its constituent claims has at least some call to be seen as scientific. What makes for science is system, whatever the subject. And what makes for system is the judicious application of logic. Science is thus a fruit of rational investigation.
They go on:
what is needed for scientific inquiry is just receptivity to data, skill in reasoning, and yearning for truth. Admittedly, ingenuity can help too.
The main objective of the book can be summarized as follow:
to broach many of the criteria by which reasonable belief may be discriminated from unreasonable belief. But not only are the criteria not foolproof; they do not always even point in a unique direction. When we meet the Virtues for assessing hypotheses we will find that they require us to look at candidates for belief in multiple ways, to weigh together a variety of considerations. Decisions in science, as in life, can be difficult. There is no simple touchstone for responsible belief.
There’s one last quote from the introduction I want to include here:
Of course science is not the only discipline that brings enlightenment; literature and the arts teach us too. One pities the person who derives nothing from poetry or music. It has been supposed by many that appreciation of the arts bears little relation to knowledge as such, as if appreciation springs from the heart while knowledge resides in the brain. But some recent writers, like Nelson Goodman, have argued convincingly that such appreciation has a much more substantial cognitive component than has been widely realized. “Cognitive”, we note, comes from a Latin word for acquiring knowledge or coming to know.
I’ve always felt that literature and the arts bring us much more than we normally assume or accept. In some of my moods, I feel it brings more than science, although I recognize this view is problematic, pretty much all of our living standards today are due to science. In any case, if our goal is to understand how things in the broadest possible sense of the term hang together in the broadest possible sense, the priority for me has always been on the side of literature and the arts, so I’m granting a special status to the disciplines in the web of belief. Thinking about how science, literature and the arts fit together is the goal.
Let’s put it this way: The web of belief encompasses all beliefs, and within mine, I hold that science, liberalism, and something like will to power were pivotal in driving the Industrial Revolution, the rise of democracy, and our high standard of living. At the same time, I deeply value art and philosophy pursued for their own sake. They probably are the highest good. I acknowledge that our wealth is built on systems of oppression, yet the ultimate goal is freedom from oppression. These ideas are in deep tension, and it frustrates me—but this is the reality I’m grappling with, and I want to find a way to reconcile them.
I’ve often felt that values are at the core of what truly matters. Science isn’t merely about uncovering what is; it’s fundamentally a value-driven endeavor, a discussion about what we choose to study and prioritize.
Chapter 2: Belief and change of belief.
Nature of Belief:
Disposition, not activity: Belief is a mental state, a readiness to respond in characteristic ways when situations arise, rather than an active process or feeling.
Retention and abandonment: Beliefs can be enduring (e.g., historical facts) or fleeting (e.g., trivial observations). They are abandoned either because they become irrelevant or because they conflict with other beliefs.
Belief and Evidence:
Beliefs are shaped by evidence but also influenced by other factors, such as subconscious stimuli, misunderstanding, or wishful thinking.
Rational belief is tied to evidence, and beliefs unsupported by evidence are eventually discarded when conflicts arise.
Truth and Knowledge:
Truth: Defined as the property of sentences that would be rightly affirmed.
Knowledge: A subset of belief grounded in sufficient evidence and truth. Belief does not qualify as knowledge unless it is both true and well-supported.
Consistency and Conflict Resolution:
Beliefs often exist in harmony until contradictions surface. Resolving conflicts involves assessing evidence to retain the most supported beliefs and discarding weaker ones.
Example of belief revision: The authors illustrate this process with a murder case involving three suspects and their alibis. Contradictions among beliefs are resolved by weighing evidence for each belief and eliminating the least supported ones.
Method of Belief Revision:
Divide and conquer: When faced with inconsistent beliefs, focus on a minimal set of conflicting beliefs. Resolve the conflict by examining evidence and revising or discarding some beliefs. Then trace the consequences of this revision through related beliefs.
Probing evidence: Inquiry stops when sufficient consistency is restored or when no further evidence is obtainable.
Belief Spectrum:
- Beliefs can range from full acceptance to disbelief to nonbelief (suspension of judgment). The process of belief revision might lead to one of these outcomes.
Philosophical Observations:
- The authors avoid resolving metaphysical questions about the “objects of belief” (e.g., whether beliefs correspond to sentences or abstract propositions). Instead, they focus on practical criteria, such as whether someone assents to a sentence as true.
Chapter 3: Observation
Observation as the Basis for Revising Beliefs
Observations act as the ultimate test for our system of beliefs.
When an observation contradicts expectations or predictions, it necessitates revising one or more beliefs.
Similar to the murder mystery analogy from Chapter 2, conflicting observations require identifying and adjusting interlocking beliefs to restore consistency.
Direct vs. Indirect Observations
Direct observations: Sensory experiences we witness firsthand (e.g., seeing an event occur).
Indirect observations: Reports, memories, or written records that rely on inference.
The ultimate evidence for our belief systems rests on direct sensory observations, but indirect sources (e.g., testimony, written records) play a significant role.
Observation Sentences
Definition**: Sentences that describe events or situations observable by multiple people at the same time and that all competent speakers of the language would agree upon.
Example: “The cat is on the mat” qualifies, while “My cat is on the mat” does not, as the latter requires additional context about ownership.
Observation sentences are learned ostensively (by direct association with sensory experiences) and form the foundation of language and physical theory.
Role of Language in Observation
Language shapes how we describe and understand observations, but observation sentences remain grounded in publicly observable phenomena.
Terms in observation sentences are often learned by pointing to objects or events (e.g., “yellow” or “ball”), making them accessible to shared understanding.
Revising Scientific Theories
Observations that challenge existing theories necessitate their revision.
Scientists prioritize maintaining consistency between observations and theories but may delay abandoning a theory until a viable alternative is available.
Example: Historical scientific revolutions (e.g., Galileo’s motion studies or quantum mechanics) emerged from prolonged crises caused by unresolved anomalies in earlier theories.
Observation and Subjectivity
While observations are intersubjective (agreed upon by multiple observers), discrepancies can arise due to differences in expertise or perspective (e.g., a trained eye sees a “condenser” where others see a “metal box”).
Philosophical concerns over subjectivity are addressed by focusing on public, verifiable observations rather than private sensory experiences (e.g., “I feel pain”).
Limits of Observations
Observation sentences are nearly infallible when describing immediate sensory experiences, but their reports (e.g., memory or testimony) are subject to error.
Persistent anomalies or ignored observations (e.g., reports of UFOs or occult phenomena) highlight the tension between theoretical commitments and conflicting data.
The Role of Observation in Science
Observations remain the boundary conditions for all belief systems, particularly scientific theories.
Scientists often dismiss observations that cannot be integrated into existing theories unless significant evidence justifies their consideration.
A good scientific theory minimizes the need to waive observations and prioritizes coherence with observed data.
Philosophical Perspective on Observation Sentences
Observation sentences, as defined, are distinct from introspective reports like “I feel pain,” which lack intersubjective verifiability.
These sentences connect directly to shared, observable phenomena, reinforcing their role as the foundation for evidence and theory in both individual and collective inquiry.
Chapter 4: Self-Evidence
Definition of Self-Evident Beliefs
Self-evident beliefs do not rely on observation or other beliefs for support. To understand them is to accept them.
Examples: “Water is wet,” “No bachelor is married,” “Every brother has a sibling.”
Some beliefs (e.g., “There have been dogs”) may seem self-evident but rely on evidence from observation.
Self-Evidence and Meaning
Self-evident truths often seem tied to the meanings of words (e.g., “bachelor” means “unmarried man”).
However, this connection does not fully explain self-evidence; it merely points to its intuitive and linguistic basis.
Logical Truths
Logical truths are a subset of self-evident truths, such as “Every horse that is white is a horse.”
Logical truths can be identified as instances of valid logical forms (e.g., “Every A that is B is an A”).
The importance of logical truth lies in its ability to imply other truths through logical implication.
Demonstrability
Absolute demonstrability: Truths deducible from self-evident premises via self-evident steps.
Logical truths can always be demonstrated in this way, but not all self-evident truths are logically true.
Proofs often involve chains of trivial steps that collectively yield non-obvious or even counterintuitive results.
Implication and Coherence
Logical implication connects beliefs and theories, ensuring coherence within the “web of belief.”
Joint implication occurs when several premises together imply a conclusion, requiring revision if contradictions arise.
Self-Evidence in Mathematics
Mathematical truths, once thought to be entirely self-evident, face challenges:
Set theory paradoxes (e.g., Russell’s paradox) show that some intuitive assumptions lead to contradictions.
Gödel’s incompleteness theorem reveals that not all mathematical truths are provable from self-evident axioms.
Mathematics now often relies on hypothetical axioms, judged by their practical utility, much like scientific theories.
Limiting Principles
Broad philosophical principles (e.g., “Nothing can come from nothing” or “Every event has a cause”) are often treated as self-evident but are prone to conflict:
The steady-state cosmological theory, which posits continuous creation of matter, challenged the principle ex nihilo nihil fit.
Quantum mechanics challenges the principle of universal causation.
These principles guide theoretical choices but may be abandoned when they impede scientific progress.
Self-Evidence in Moral Precepts
Moral claims (e.g., “One should not inflict needless pain”) are sometimes called self-evident, but this often reflects a decision to treat them as basic and immune to debate.
Even moral maxims can conflict (e.g., “Do not lie” vs. “Do not cause unnecessary harm”), raising questions of consistency.
Chapter 5: Testimony
The Function of Testimony
Language facilitates:
Action: Getting others to act as desired.
Knowledge: Learning from others’ experiences.
Testimony extends our observational reach, acting as a “vicarious sense” that amplifies our understanding of the world.
Early testimony relies on observation sentences, learned ostensively (via direct sensory association), which are relatively reliable.
Challenges of Testimony
Errors and Lies: Testimony becomes less reliable as we move beyond simple observation sentences.
A high degree of trustworthiness in testimony is essential for language and communication to function.
Even lies require effort, as language learning conditions speakers toward veracity (truthfulness) and listeners toward credulity (belief).
Balancing Credulity and Skepticism
Humans are naturally predisposed to believe others due to language learning, but unchecked credulity can lead to errors.
Healthy skepticism requires:
Assessing the plausibility of the testimony (e.g., the speaker’s motives and evidence).
Evaluating the source’s credibility (e.g., expertise, independence, and possible biases).
Examples:
In courts of law, testimony is scrutinized based on motives and potential biases.
Historical errors, such as the claim that Monaco is eight square miles, highlight the dangers of over-relying on non-independent sources.
The Role of Evidence
Testimony should be evaluated based on the speaker’s access to evidence:
Reference books are often more reliable than hearsay because of systematic data collection.
Science relies on empirical methods, making its testimony more robust.
Common knowledge (“everyone knows it”) can still be misleading, as seen in historical examples like geocentrism or the ether theory.
Testimony in Science
Science relies on testimony backed by empirical evidence and reproducible results.
Testimonies of experts (e.g., Einstein’s relativity or the Big Bang theory) often seem counterintuitive but gain credibility through cumulative evidence and successful predictions.
Absurdity in testimony (e.g., quantum mechanics or space-time relativity) becomes acceptable when supported by sufficient evidence.
Faith and Absurdity
Faith in testimony, as in Kierkegaard’s philosophy or Tertullian’s “I believe because it is absurd,” highlights the tension between belief and evidence.
Belief in the “absurd” often stems from deference to authority or sources deemed credible, rather than genuine acceptance of contradictions.
Scientific examples show how testimony can initially appear absurd but become credible with stronger evidence.
Practical Lessons
Independence of Sources: Multiple corroborating sources are valuable only if they are independent.
Humility in Knowledge: Many widely accepted beliefs may later be disproven, underscoring the need for caution and openness to revision.
Weighing Evidence: Acceptability of testimony depends on balancing:
Evidence supporting the source.
Plausibility of the claim.
Contradictory evidence.
Chapter 6: Hypothesis
Limits of Absolute Demonstrability
Rationalists once believed that all truths could be derived from self-evident beginnings through self-evident steps. This has been proven incorrect:
Gödel’s theorem shows that even elementary number theory cannot be fully derived from self-evident truths.
Natural truths (e.g., “giraffes are mute”) rely on generalizations from observations, which themselves are not self-evident and can fail, as in the case of assuming all swans are white.
The Role of Hypotheses
Definition: A hypothesis is a form of educated guesswork that helps explain observations and predict future events.
Hypotheses connect the observed to the unobservable, filling gaps where direct deduction from self-evident truths and observations is insufficient.
Example: In the murder mystery (Chapter II), hypotheses were created and revised to explain the evidence and account for new discoveries.
Virtues of Hypotheses
Quine and Ullian propose five “virtues” that make hypotheses plausible, practical, and scientifically valuable:
Virtue I: Conservatism
A hypothesis is more plausible if it minimizes conflict with existing beliefs.
Example: In choosing explanations for a magician’s card trick, sleight-of-hand is preferred over telepathy, as it better aligns with prior beliefs about human capabilities.
Virtue II: Modesty
A modest hypothesis assumes less, sticking to humbler and more common explanations.
Example: Explaining a wrong phone call as a dialing mistake is more modest than attributing it to a burglar checking if anyone is home.
Virtue III: Simplicity
Simplicity favors hypotheses that are less complex and easier to generalize.
Example: Newton’s universal gravitation, though more complex than “heavy objects fall downward,” unified celestial and terrestrial mechanics into a simpler overall framework.
Simplicity also influences how we evaluate curves on a graph or equations in physics, preferring those with fewer terms or smoother shapes.
Virtue IV: Generality
A hypothesis gains plausibility and importance if it applies broadly across cases.
Example: Discovering that copper wire conducts electricity leads to the broader hypothesis that all copper conducts electricity.
Virtue V: Refutability
A hypothesis must be falsifiable; it must allow for conditions under which it could be proven wrong.
Example: Astrology often lacks refutability because its predictions are vague and unfalsifiable, unlike scientific hypotheses that invite tests and potential disproof.
Balancing the Virtues
Hypotheses must balance conservatism, modesty, simplicity, generality, and refutability to succeed.
Example: Einstein’s theory of relativity balanced conservatism (retaining Newton’s mechanics in limited cases) with simplicity and generality (explaining both terrestrial and celestial phenomena).
Ad Hoc Hypotheses
Ad hoc hypotheses explain specific cases without general applicability or predictive power, making them undesirable.
Example: Speculating that Uranus alone defies physical laws to explain its orbit would have been ad hoc. Instead, Neptune’s discovery preserved existing theory while expanding its scope.
Evolution of Hypotheses
Hypotheses are shaped by culture and experience:
Inherited Hypotheses: Many hypotheses are passed down culturally (e.g., scientific principles, societal norms).
Testing and Revision: Hypotheses are continually revised based on new observations or failures of prediction, as seen in the transition from Newtonian physics to Einsteinian relativity.
Scientific revolutions often arise when simplicity and generality demand breaking with conservatism.
Conclusion: Observation and Hypothesis
The heart of hypothesis-making lies in observation:
Observations provide the evidence for and against hypotheses.
Hypotheses, in turn, predict further observations, which confirm or refute them.
Scientific laws, mathematical principles, and even everyday beliefs function as hypotheses within a larger web of belief, constantly tested and refined as new evidence emerges.
Chapter 7: Induction, Analogy, and Intuition
Induction: Learning from Experience
Definition: Induction involves generalizing from observed cases to future cases, expecting similar things to behave similarly.
Example: We expect toothpaste to emerge from a tube when squeezed because it has done so consistently in the past.
Induction is driven by Virtue IV (generality) and Virtue III (simplicity). Traits we notice and project—like “green” in emeralds—are those that feel simple and familiar.
Challenges of Induction
Goodman’s Riddle of “Grue”:
If we define “grue” as objects that are green before midnight but blue afterward, past observations of green emeralds also support the expectation that future emeralds will be “grue.”
This paradox highlights that not all traits are equally projectible.
Problem of Arbitrary Traits:
Traits like “being prior to 1978” or “followed by more life” are not projectible despite being shared by past observations.
The key question is: What makes some traits projectible while others are not?
Biological and Evolutionary Basis:
Traits like “green” are likely projectible due to evolutionary survival value.
The process is not infallible, as our instincts for projecting traits are shaped by neural limitations and historical successes.
Analogy: Drawing Parallels
Definition: Analogy involves inferring similarities between cases without forming an explicit generalization.
Examples:
If past boiled lobsters turned red, we expect the next boiled lobster to do the same. This leap from case to case is based on analogy.
Recognizing a voice and predicting personality traits based on its similarity to someone else’s voice.
Analogies are often implicit and based on multiple shared features between cases.
- Weak analogies (e.g., basing predictions on superficial resemblances) improve with experience and better discrimination of relevant features.
Analogy and Scientific Generalization
Analogies can extend general laws:
- Example: If one serum immunizes against a disease, a similar serum might work for a related disease caused by similar bacteria.
Analogies sometimes point to broader generalizations (e.g., laws about diseases and serums) that can later be made explicit.
Intuition: Subconscious Reasoning
Definition: Intuition refers to beliefs formed without conscious reasoning or clear evidence.
Example: Judging someone as insincere based on subtle, unnoticed cues like facial expressions or tone.
Intuition is not mystical; it often relies on implicit analogies, forgotten experiences, or unnoticed observations.
Criticism of Intuition:
Explicit appeals to intuition often lack explanatory power.
Pseudoscientific claims (e.g., “auras” or “vibes”) exploit intuition without offering meaningful mechanisms.
Machine Analogies and Artificial Intelligence
Human intuition, like recognition of faces or voices, parallels how machines process and compare data.
The study of artificial intelligence can illuminate how humans use stored knowledge to recognize patterns and make decisions.
Induction vs. Analogy
Induction:
Generalizes from specific cases to form broad rules or laws.
Example: Observing that all boiled lobsters turn red and concluding a general law that “boiled lobsters are red.”
Analogy:
Draws direct inferences between specific cases without articulating a general rule.
Example: Expecting the next boiled lobster to turn red based on past experience without stating a general law.
Analogy may bypass explicit generalization but still relies on implicit patterns.
Improving Our Reasoning
Refining Projectibility: With experience, we refine our sense of which traits are projectible.
Example: Whales were once classified as fish, but scientific understanding adjusted the analogy based on new traits.
Scientific Advances: Science and induction reinforce each other; better generalizations improve science, and scientific insights refine inductive reasoning.
Limits of Induction and Analogy
Induction and analogy are fallible, as they depend on subjective notions of simplicity and projectibility.
They reflect evolutionary and cultural constraints, which can bias or limit their accuracy.
Nonetheless, they are indispensable tools for navigating the complexities of the world.
Chapter 8: Confirmation and refutation
Testing Hypotheses
Process: A hypothesis is tested by observing whether its consequences align with reality.
- Example: A hypothesis that eating sweet pickles causes headaches is tested by deliberately consuming pickles and observing the result.
Confirmation: When predictions align with observations, the hypothesis is confirmed but not proven.
- Limitations: Confirmation supports the hypothesis but does not exclude other explanations (many curves can fit the same data points).
Virtues in Hypothesis Testing:
Conservatism (Virtue I): The hypothesis should align with existing knowledge where possible.
Modesty (Virtue II): It should not be overly complex or assume too much.
Generality (Virtue IV): Broader hypotheses are often preferred.
Precision (Virtue VI)
Importance of Precision: Precise hypotheses generate more testable and meaningful predictions.
- Example: “Sweet pickles cause headaches within 12–13 minutes” is more precise than “Sweet pickles cause headaches eventually.”
Measurement and Concomitant Variation: Precision often involves quantifying variables and identifying relationships (e.g., boiling point of water varies with pressure).
Trade-offs with Generality: Increasing precision can sometimes limit generality (e.g., boiling point changes for impure water).
Confirmation in Experience
Confirmation by Instances: Observing instances of a hypothesis (e.g., green emeralds) adds to its plausibility.
Lawlike Generalizations:
Only lawlike hypotheses are confirmed by their instances.
Example: “All emeralds are green” is lawlike, while “All coins in my pocket on Monday were dimes” is not.
Goodman’s Riddle Revisited:
- Traits like “green” are projectible and support induction; “grue” is not.
Refuting Hypotheses
Asymmetry of Refutation:
- A single contradictory instance refutes a hypothesis (e.g., water boiling at 92°C under standard conditions refutes the boiling point hypothesis).
Complication of Supporting Beliefs:
Hypotheses rarely stand alone; they rely on a “supporting chorus” of background beliefs.
Example: Observing water boiling at 92°C could mean the hypothesis is wrong—or that the water is impure or the pressure mismeasured.
Imprecision and Isolation:
Imprecise hypotheses are harder to test because their predictions are vague.
Precise hypotheses carry assumptions that can complicate testing.
Probabilities and Confirmation
Probability of a Hypothesis:
- Assigning probabilities to hypotheses is straightforward in controlled cases (e.g., gambling) but challenging in broader contexts (e.g., “The universe began with a bang”).
Statistical Hypotheses:
Example: “93% of inoculations are effective” is a statistical hypothesis, tested through repeated observations.
Statistical hypotheses are distinct from assigning a probability to a hypothesis itself.
Challenges in Real-World Probabilities:
- Conflicting background information (e.g., voter demographics and affiliations) complicates probability estimation.
Refutability (Virtue V)
Definition: A hypothesis should be testable and potentially refutable by specific observations.
Examples of Poor Refutability:
Ad hoc adjustments: Modifying a hypothesis to fit conflicting observations without changing its core assumptions (e.g., excuses for a medium’s failed predictions).
Imprecise claims: Vague predictions (e.g., “pickles eventually cause headaches”) resist disconfirmation.
Chapter 9: Explanation
This chapter of The Web of Belief delves into the nature and purpose of explanations, their role in science and everyday reasoning, and the challenges of crafting meaningful explanatory hypotheses.
The Purpose of Explanation
Utility of Hypotheses: Hypotheses help predict the future and explain the past. This dual role enhances both practical problem-solving and intellectual curiosity.
Curiosity and Survival Value: The drive to explain satisfies curiosity, which has survival value by fostering deeper understanding and aiding future predictions.
Infinite Chain of “Why”: Explanations often lead to further questions, demonstrating the iterative nature of inquiry.
From Observation to Laws
Primitive Explanations: Early explanations often arise from observation and pattern recognition.
- Example: A tribe observing tides might notice their correlation with the moon’s position.
Improved Explanations: Over time, more precise and general laws emerge (e.g., Newton’s law of gravitation explains tides).
Limitless Depth: Even well-founded explanations, like Newton’s laws, inspire quests for deeper theories (e.g., unified field theories).
Characteristics of Explanations
Causal Connections: Explanations often aim to uncover causal chains leading to the event in question.
- Example: The Chicago fire traced to Mrs. O’Leary’s cow involves a network of contributory causes.
Statistical Explanations: These suggest causal connections without implying certainty (e.g., exposure to disease explaining illness, though not everyone exposed gets sick).
Mechanisms of Explanation
Physical Causes: Physics reduces phenomena to elementary forces (e.g., impacts, gravitational pulls), forming the basis of causal explanation.
Psychological and Teleological Causes:
Explanations of behavior often invoke motives and purposes, which can coexist with physical explanations.
Example: A man hurrying to the library by taxi may be explained teleologically (to meet a deadline) or causally (nervous system stimuli).
Teleological Explanations
Biology and Purpose: Explanations like “eyes are for seeing” or “willows lean to disperse seeds” imply purpose.
Darwinian Resolution: Darwin’s theory of natural selection reframes teleological explanations as shorthand for causal chains driven by survival and reproduction.
Explanatory Standards
Good Explanations:
Must advance understanding by revealing causal mechanisms.
Should avoid circular reasoning or restating the phenomenon (e.g., Molière’s “virtus dormitiva”).
Implication and Coherence:
The best explanations often imply what they explain, though statistical explanations may not.
Explanations gain credibility by their coherence and ability to account for known facts.
Explanations and Hypotheses
Mutual Reinforcement:
A hypothesis explaining observations gains credibility.
Conversely, observations that fit a hypothesis enhance its plausibility.
Argument from Exclusion:
When no alternative explanation exists, a hypothesis may gain credibility (e.g., Houdini’s escape mechanisms).
This reasoning, though powerful, is prone to abuse by charlatans and uncritical thinkers.
Dangers of Poor Explanations
Unfalsifiable Claims:
- Hypotheses that cannot be tested (e.g., “God’s will” or “unconscious desires”) lack explanatory power.
Motivated Reasoning:
- Explanations shaped by biases (e.g., attributing altruism to neurosis) can reinforce prejudices and resist refutation.
Charlatanry:
- Spurious hypotheses often exploit gaps in understanding to promote pseudoscience or cultish beliefs (e.g., extraterrestrial origins of medieval rockets).
Key Safeguards Against Flawed Explanations
Skepticism: Be cautious of unlikely claims and uncritical acceptance of reports.
Virtues of Hypotheses: Favor hypotheses that exhibit simplicity, generality, conservatism, and refutability.
Clarity: Demand clear articulation of claims. Persistently vague explanations often mask meaningless ideas.
Explanations Beyond Science
Mathematics: Explanations involve deductive reasoning, tracing truths to their logical foundations.
Human Behavior:
Psychological explanations must remain testable and open to refinement.
Caution is needed to avoid reinforcing stereotypes or untestable assumptions.
Chapter 10: Persuasion and Evaluation
Two Core Purposes of Language
Influencing Action: Language can be used to persuade others to act in desired ways.
- This includes commands, requests, and arguments aimed at implanting beliefs that motivate action.
Learning from Others: Language also serves the purpose of acquiring knowledge or understanding from others.
Persuasion and Sincerity
Utility of Lying: Persuasion doesn’t always require truth, as beliefs can be implanted deceitfully.
Ethics of Persuasion: In societies valuing truth, the ideal purpose of persuasion is to share sincerely held beliefs.
Propagation of Beliefs: Beyond practical motives, many seek to share beliefs for the sake of shared understanding.
Structure of Argument
Roots of Belief: Beliefs rest on supporting beliefs, which may include observations or shared assumptions.
Building Agreement: Persuasion involves appealing to beliefs the other party already holds and logically connecting these to the desired conclusion.
- Example: Mathematical proofs rely on shared axioms or previously accepted theorems.
Maxim of Shallow Analysis: In persuasion, arguments should stay as simple and general as possible to maintain engagement and avoid confusion.
Challenges in Persuasion
Observational Testimony: Convincing others of observations is weaker than personal observation due to the need for trust.
- Repetition of observations by skeptics can bridge this gap, though practical limitations often exist.
Credibility: A persuader’s credibility—earned through consistency, honesty, and sound judgment—enhances their influence.
Overcoming Resistance
Two Strategies:
Overwhelming: Provide abundant evidence to make opposing beliefs untenable.
Undermining: Directly challenge and weaken opposing beliefs by questioning their foundations.
Repercussions of Argument: In some cases, the persuader might be swayed by the strength of the counterargument, fulfilling the purpose of learning rather than persuasion.
Desire to Be Right
Two Desires:
Desire to Be Right: Reflects the pursuit of truth and understanding.
Desire to Have Been Right: Reflects pride and resistance to admitting error, which hinders learning.
Being Reasonable: Good reasoning doesn’t always guarantee correctness, but it maximizes truth over the long term.
Evaluation and Values
Values and Beliefs:
Evaluations (e.g., “this act is good”) are often tied to beliefs about means and ends (e.g., “this act leads to a desired outcome”).
Evaluations become indistinguishable from beliefs when focused on causal relationships.
Aesthetic and Moral Values:
Aesthetic Training: Appreciation of art can be cultivated through exposure, analysis, or emulation.
Moral Training: Altruism and ethical behavior can be nurtured through rewards, punishments, or vivid depictions of suffering or joy.
Natural Selection and Shared Values
Evolutionary Basis:
- Traits like altruism and fellow feeling have evolved to favor societal survival, fostering widespread agreement on basic moral values.
Moral Disagreements:
Complex issues (e.g., abortion, euthanasia) often reflect deeper value conflicts rather than disagreements about facts.
These can sometimes be addressed by reframing the issue as one of means rather than ends.
The Limits of Moral Theories
Grounding Morality:
Many moral theories, from religious doctrines to Kant’s categorical imperative, aim to provide ultimate foundations for values.
None has achieved universal acceptance, reflecting the inherent challenge of adjudicating ultimate moral ends.
Hope for Consensus:
- Despite disagreements, shared moral instincts provide a foundation for basic ethical consensus across societies.
My take
One may find The Web of Belief overly committed to a rationalist, utilitarian perspective on belief and human behavior, which I think neglects the irrational, excessive, and contradictory aspects of human experience. Its focus on coherence, utility, and evolutionary explanations feels reductive, failing to address the deeper, transgressive forces—like desire, taboo, and the sacred—that often shape our thoughts and actions beyond reason. To me, its insistence on clarity and plausibility simplifies the complex, chaotic interplay between knowledge, belief, and the ineffable dimensions of existence.
For example, Oscar Wilde would find The Web of Belief irredeemably tedious, a celebration of dull precision at the expense of beauty. Truth should not be a mechanical web to be woven through endless observations and hypotheses but a radiant fiction, shaped by beauty, wit, and the imagination’s refusal to bow to mere fact. Wilde would see the book’s virtuous dedication to reason as a symptom of its ultimate failing: the neglect of art, irony, and the marvelous in favor of a dreary worship of systems. If life is a canvas to Wilde, this book insists on painting it in grayscale, missing the colors that make existence worth living.
If I were channeling Marcuse, I would say that The Web of Belief represents a technocratic reduction of human thought, one that mistakes the refinement of method for liberation. It reifies the scientific and logical apparatus of inquiry, presenting it as neutral and universal while ignoring its entanglement with systems of domination. The book’s virtues—conservatism, simplicity, and modesty—reflect the intellectual passivity of a society content to reproduce existing structures rather than confront the contradictions of its historical moment. It is a manifesto for administering knowledge within the boundaries of an unfree world, blind to the emancipatory potential of critique and imagination. It seeks order, but in doing so, it betrays the dialectical tension necessary for meaningful change.
Yet, it’s obviously a correct perspective. What is correct is boring. What is exciting is either false, love, or art.