Can anything really be true by definition, and could that be how we know that 2 + 2 = 4?
My links: https://linktr.ee/frictionphilosophy.
1. Guest
Check out his book, “True by Definition”!
https://global.oup.com/academic/product/true-by-definition-9780197846872
2. Book Summary
Thomas Donaldson’s True by Definition argues for a qualified “analyticity bullism”, the view that the analytic/synthetic distinction is a useful philosophical tool, though a less powerful one than its champions have claimed. The book also works as an accessible guide to the literature, and it uses elementary integer arithmetic as its running case study. Donaldson starts from a Fregean idea: a statement is “narrowly” analytic if it is a logical consequence of “proper” definitions. These are definitions that are strictly true and foundational, like the stipulative definitions in a mathematics textbook, which are used as premises but never proved. He defends their existence with his “Very Important Deferral to Experts Principle”. If stipulation couldn’t fix meaning, as Cappelen holds, the standard practice of mathematics would be systematically mistaken. And because “mathematicians are not aliens”, proper definitions exist outside mathematics too, for words like “kitten” and “oncologist”. He then gives the bears a lot. Following Wittgenstein on “game” (against Suits) and Kripke on names and natural kind terms, he accepts that most philosophically interesting words have no proper definition. That pushes him to a “broad” notion of analyticity. It covers statements derivable from “basic analytic statements” such as “every game is an activity for one or more players” or “every red thing is coloured”, which are foundational, necessary and explanatorily brute even with no definition behind them. To call a statement basically analytic is to make an empirical claim about a community’s conventions, which rules out the “lazy evasion” of simply declaring a thesis analytic. Conventions can also be unjust or contradict each other, which opens the door to conceptual engineering; his example is the definitional objection to same-sex marriage. His treatment of implicit definition brings in presuppositions: such a definition succeeds only if its presupposition holds. So a broadly analytic statement can be untrue (“d is the largest prime”) or a posteriori (the 1983 definition of the metre presupposes that the speed of light is constant). Incompletely defined terms like “psmall” produce truth-value gaps, and he gives up bivalence for those cases.
Part 2 sets “definitionalism” against a mathematical sceptic. When the regress of proofs runs out at the axioms, Donaldson turns down BonJour-style rational insight. Instead he develops the empiricist line running through Hilbert, Poincaré, Reichenbach and Ayer, on which the axioms of integer arithmetic implicitly define “zero”, “integer”, “+” and the rest. On this reading, for example, the least number principle just records the decision to exclude ½ from the integers. The obstacle is the presupposition. On a “reference-based” theory of implicit definition, the presupposition is that infinitely many qualified referents exist, and nothing an empiricist can reach secures that. Set theory starts a regress, an infinity of thoughts runs into Russell’s dilemma, and infinite or infinitely divisible space is speculative physics. Donaldson replaces this with a “truth-based” theory: a definition succeeds if there is a suitable assignment of truth conditions. For a domain-expanding definition like the one his character Ashni introduces, that comes to the axioms being “everywhere coherent”, meaning consistent with the facts of her old language at every possible world. A finite physical universe is then no objection. But the sceptic’s challenge simply moves to consistency. Gödel’s second incompleteness theorem blocks proving it from a privileged subset of the axioms, and inductive arguments from no contradiction having turned up are weak, because the theorems proved so far are an unrepresentative sample lying close to the axioms. Donaldson’s answer is a refined indispensability, or “no miracles”, argument built on “ordinary geometry”, the kind woodworkers and engineers use. Its success is explained by a “Mirroring Principle”: at everyday scales, physical lengths, areas, volumes and angles almost exactly match their counterparts in ℝ³. That explanation requires the underlying axioms to be coherent. Because the argument needs only coherence, not truth, it escapes Field’s nominalism and Maddy’s point about idealisation. The conclusion is that elementary arithmetic is analytic yet a posteriori. He explicitly limits this to “obvious” mathematics and leaves out higher set theory.
Part 3 turns to metaphysics. Using Lange’s “explanation by constraint”, Donaldson defends a version of Ayer’s thesis that analytic truths are true in virtue of meaning: the truth of a narrowly analytic statement can be proved from linguistic premises alone. He reconciles this with Boghossian’s “truism” by reading the truism as a claim about grounding explanations only. Whether analytic truths are fully grounded in linguistic facts stays open, and he criticises Gillian Russell’s reference-determiner account as hard to apply empirically. Chapter 10 explores the “Parallelism Hypothesis”, the claim that analytic entailment between sentences mirrors grounding between facts. It would give the physicalist something valuable: the fact that Mars has a moon, another one, and no more would ground the fact that the number of Mars’s moons is prime. The price is three things. First, zero-grounding for axioms like “zero is an integer”. Second, rewriting definitions into an unnatural, somewhat ad hoc form. Third, a formal theory nobody has yet supplied. Donaldson finds the hypothesis attractive but faces up to these obstacles. Chapter 11 argues that if the only constraint on interpreting arithmetic is that the axioms come out true, then any one standard interpretation brings infinitely many with it. Numerals are therefore radically indeterminate in reference. Even so, “3 + 2 = 5” is true because it holds on every standard interpretation, and truisms like ‘”zero” refers to zero’ still come out true. Neo-Fregean abstractionism could secure determinate reference, but only if Parallelism holds. The afterword explains why he was a bear for years: he kept seeing analyticity used as a deus ex machina, and he shows it with David Lewis’s identity-theory paper (”I have an analytic truth / You have an intuition / They have a metaphysical prejudice”). He gives three reasons why declaring a claim analytic never settles anything on its own: analyticity is an empirical claim, presuppositions can fail, and a convention may need reform. On top of that, the bridge principles from language to metaphysics are hard to formulate. He still maintains that, used carefully, the distinction does real philosophical work.
3. Interview Chapters
00:00 - Introduction
01:15 - Overview of book
03:19 - Bullish vs. bearish
05:41 - Definition
09:06 - Herman Cappelen
13:53 - Semantic theories
15:13 - Proper definitions
22:31 - Deferral to experts
24:58 - Analytic truths without definitions
32:01 - Implicit definitions
39:58 - Analytic but false?
45:05 - Mathematics
59:06 - Worry
1:11:53 - Realism and nominalism
1:22:36 - Amie Thomasson
1:25:24 - Illegitimate cases
1:29:30 - God
1:31:30 - Determining analyticity
1:40:32 - Red and green all over
1:44:54 - Reflecting
1:46:55 - Coherence
1:48:26 - Grounding
1:58:22 - Physicalism
2:03:38 - Main takeaway
2:06:18 - Value of philosophy
2:09:56 - Conclusion
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fric.substack.com/subscribe
More episodes from "Friction"



Don't miss an episode of “Friction” and subscribe to it in the GetPodcast app.








