Plain-language definitions of the philosophical concepts behind the eleven argument styles.
These are the concepts named in the
philosophical toolkit table — many of them
tools from Baggini & Fosl's The Philosopher's Toolkit. The concept terms in that table link
straight to the matching entry below.
Monism vs. pluralism (Hedgehog/Fox)
A monist thinks one big idea can explain most of what matters; a pluralist thinks reality needs many separate explanations that don't reduce to one. The hedgehog is a monist by temperament; the fox, a pluralist. SEP ↗
Other things equal, prefer the simpler explanation — don't invent more moving parts than you need. The splitter's worry is that "simpler" can mean "ignoring real differences." SEP ↗
Deduction vs. induction (Chain/Cable)
A deductive argument, if its premises are true, guarantees its conclusion — but one false premise breaks it (the chain). An inductive argument makes its conclusion likely by piling up evidence; no single strand carries it (the cable). SEP ↗
Defeasible reasoning (Chain/Cable)
A defeasible conclusion is one you accept now but would give up if new evidence came in. Cable-style arguments are defeasible by design — a feature, not a weakness. SEP ↗
Serial vs. convergent support (Chain/Cable)
In serial support the reasons hang together as one load-bearing line, so if any link breaks the whole inference fails (the chain). In convergent support several independent reasons each push toward the conclusion on their own, so the case survives even if one strand is cut (the cable). SEP ↗
Generalism vs. particularism (Principle/Case)
A generalist thinks good reasoning starts from general rules and applies them to cases. A particularist thinks the right call can depend so much on the details that no rule captures it — you have to read the case itself. SEP ↗
Reflective equilibrium (Principle/Case)
Going back and forth between your general principles and your gut judgments about specific cases, adjusting each until they fit. It's how principle-first and case-first reasoners can meet in the middle. SEP ↗
A necessary condition must hold for something to count (no oxygen, no fire); a sufficient condition is enough on its own to guarantee it. Defining a concept by such conditions is the principle-first reasoner's instinct. SEP ↗
Carving nature at its joints / family resemblance (Lumper/Splitter)
Plato's image: a good category cuts reality where it really divides, like a butcher finding the joint. Wittgenstein's rival idea: some categories (like "game") have no single defining feature — members just share overlapping family resemblances. Lumpers trust the joints; splitters doubt they are joints. SEP ↗
The sorites ("heap") problem (Lumper/Splitter)
One grain isn't a heap; adding one grain never seems to make a heap — yet somehow you get a heap. It shows how hard it is to say exactly where a vague category begins. SEP ↗
Ambiguity & equivocation (Lumper/Splitter)
A word is ambiguous when it carries more than one meaning; equivocation is the error of sliding between those meanings inside a single argument. Splitters watch for it; lumpers risk it by treating two senses as one. SEP ↗
Category mistake (Lumper/Splitter)
Ryle's term for assigning something to the wrong logical type — like touring every college and then asking to see the university, as though it were one more building. The risk cuts both ways: splitters can multiply categories until a whole gets mistaken for just another part, so lumpers and splitters alike must track which logical type a thing belongs to. SEP ↗
Falsifiability (Bold/Cautious)
A claim is falsifiable if there's some possible observation that would prove it wrong. Popper's point: a bold, falsifiable claim is valuable because it sticks its neck out and can be tested. SEP ↗
Believe-truth vs. shun-error (Bold/Cautious)
William James named two goals that pull apart: get to the truth and avoid believing falsehoods. Chase truth too hard and you'll believe rashly; avoid error too hard and you'll commit to nothing. Bold vs. cautious is which you weight. SEP ↗
Certainty & probability (Bold/Cautious)
Claims come in degrees of confidence: the bold thinker asserts flatly, as if certain, while the cautious one attaches a probability and hedges. The cautious style treats belief as a dial you can turn up or down, not a switch that's simply on or off. SEP ↗
Three ways to test a claim by attacking it: the elenchus is Socrates' cross-examining until a belief contradicts itself; a counterexample is a single case that breaks a generalization; a reductio assumes a claim and derives an absurdity from it. SEP ↗
Abduction / inference to the best explanation (Refuter/Builder, Mechanism/Pattern)
Reasoning from a puzzling fact to the hypothesis that would best explain it ("the best account of these tracks is that a deer passed"). The builder's engine — and the mechanism-hunter's. SEP ↗
Verstehen vs. Erklären (Narrative/Analytic)
Two German words for two kinds of grasp: Erklären is to explain from the outside by causes and laws; Verstehen is to understand from the inside, by reconstructing what something means to the people involved. Analytic argument aims at the first; narrative at the second. SEP ↗
Knowledge by acquaintance vs. by description (Narrative/Analytic)
Russell's distinction: you can know of something by direct experience (acquaintance — what red looks like) or only through a description (knowing that 700nm light is "red"). Stories trade in acquaintance; analysis in description. SEP ↗
Thought experiments (Narrative/Analytic)
An imagined scenario — the trolley problem, Mary's room — built to pump an intuition or test a principle without leaving the armchair. It serves analytic argument while borrowing narrative's vividness, which is why it sits right on the seam between the two. SEP ↗
Intuition pumps (Narrative/Analytic)
Dennett's name for a vivid little story rigged so that one answer feels obvious. Useful for surfacing intuitions — and worth distrusting, since changing an incidental detail can pump the opposite verdict. SEP ↗
A priori vs. a posteriori (Empirical/Conceptual)
A priori knowledge can be had by thinking alone, before experience (2+2=4); a posteriori knowledge requires observing the world (water boils at 100°C). The empirical style leans a posteriori, the conceptual style a priori. SEP ↗
Hume's fork & the analytic/synthetic distinction (Empirical/Conceptual)
Hume split claims into relations of ideas (true by meaning — "all bachelors are unmarried") and matters of fact (true by how the world is). Kant's near-cousin: analytic truths unpack a concept; synthetic truths add information about the world. Quine later argued the line is blurrier than it looks. SEP ↗
Nomothetic vs. idiographic (Quantitative/Qualitative)
Nomothetic inquiry looks for general laws across many cases (the quantitative impulse); idiographic inquiry tries to understand one particular case in its full richness (the qualitative impulse). SEP ↗
Thick vs. thin concepts; thick description (Quantitative/Qualitative)
A thin concept is bare and general ("good," "large"); a thick concept fuses description and evaluation ("cruel," "courageous"). Thick description (Geertz) reports a scene with enough meaning-laden detail that you grasp what it meant to the people in it. SEP ↗
Objective vs. subjective (Quantitative/Qualitative)
An objective claim aims to hold independently of any particular viewpoint; a subjective one reports how things stand for a given perceiver. The trap is assuming this lines up with the quantitative/qualitative split — good qualitative work still aims at understanding others can check, not at mere subjectivity. SEP ↗
Criteria (Quantitative/Qualitative)
The standards that fix when a term genuinely applies — what would have to be true for something to count as, say, a fair trial. Settling criteria before measuring is what keeps a count from tracking the wrong thing. SEP ↗
Causation vs. constant conjunction (Mechanism/Pattern)
Hume noted we never actually see one thing make another happen — only that they regularly occur together (constant conjunction). The pattern-thinker is content with the regularity; the mechanism-thinker insists there's hidden machinery doing the making. SEP ↗
Covering-law vs. mechanistic explanation (Mechanism/Pattern)
One view says you've explained an event once you've shown it follows from a general law plus the conditions (the covering-law model); the rival says real explanation means displaying the mechanism — the parts and activities that produce it. SEP ↗
Underdetermination (Mechanism/Pattern)
When the same evidence is equally compatible with several explanations, the evidence underdetermines the theory. It's why a clean pattern can't, by itself, tell you which mechanism is behind it. SEP ↗
Alternative explanations (Hedgehog/Fox, Mechanism/Pattern)
Considering alternative explanations means listing the rival accounts that would fit the same facts, instead of just confirming the one you favor — it's what makes an explanation the best one rather than merely a possible one. The fox and the pattern-thinker keep several rivals in play; the hedgehog and mechanism-thinker risk settling on a favorite too soon. SEP ↗
Dialectic vs. eristic (Adversarial/Collaborative)
Aristotle's two kinds of argument: dialectic aims to reach the truth together through reasoned exchange; eristic aims only to win. Adversarial argument is healthy when dialectical, toxic when it slides into eristic. SEP ↗
Principle of charity / steelmanning (Adversarial/Collaborative)
Interpret the other person's argument in its strongest, most reasonable form before you respond. It keeps an adversarial style honest and a collaborative style from going mushy — you beat (or join) the best version of the view, not a caricature. SEP ↗
Fallacies (ad hominem, straw man) (Adversarial/Collaborative)
Recurring patterns of bad argument. Attacking the person rather than the claim (ad hominem), or refuting a weakened caricature of it (straw man), are the adversary's temptations — the very moves the principle of charity forbids. SEP ↗
Sources: Baggini & Fosl, The Philosopher's Toolkit (2003);
Peirce (1868); Dancy, Ethics Without Principles (2004); Rawls (1971); Hume, Enquiry (1748);
Kant (1781); Quine, "Two Dogmas" (1951); Russell (1912); Ryle (1949); Windelband (1894); B. Williams (1985);
Geertz (1973); Hempel & Oppenheim (1948); Salmon (1984); Mercier & Sperber (2017); Longino (1990).