Meaning and Form in Mathematics

The most important thing to understand about meaning in a formal system is that a suggestive label is not enough: meaning is earned when relations among formal strings match relations in another domain. Day 01’s formal-search receipt separated legal symbol moves from claims about those moves. Today you will preserve that separation while deriving a pq-theorem, mapping its form to arithmetic, and rejecting an interpretation that fails on a counterexample.

The smallest complete model keeps Day 01’s formal spine intact and adds a tested bridge to a domain. The pq-system produces strings without consulting arithmetic, while additions are true or false without consulting pq derivations. An interpretation assigns counterparts to symbols; an isomorphism is a structure-preserving correspondence in which related parts and operations on one side match related parts and operations on the other. The new counterexample path tests that bridge without changing what formal evidence means.

Boundary. This lesson tests addition and one alternative interpretation over the pq-system. It does not claim that every formal system has one unique meaning, or that a finite sample proves an interpretation correct for infinitely many strings.

Meaning cannot be evaluated until the underlying mechanism is precise, or an attractive reading may smuggle in illegal strings. A hyphen-string is one or more hyphens; a well-formed string here has three hyphen-groups separated by one p and then one q, as in --p---q-----.

The pq-system begins with an axiom schema, a pattern that supplies infinitely many axioms:

xp-qx-

Here x is a metavariable for the same hyphen-string in both positions. The single rule says: if xpyqz is a theorem, then xpy-qz- is a theorem. It appends one hyphen to the middle group and one to the final group while leaving the first group fixed.

Take x to be --. The axiom is --p-q---. Apply the rule twice:

A pattern can be renamed arbitrarily, and arbitrary renaming creates poetry rather than understanding, so the bridge must preserve relationships. Under the intended interpretation, each hyphen denotes one, p denotes “plus,” and q denotes “equals”; the theorem --p---q----- then corresponds to 2 + 3 = 5.

The mapping works at two levels. At the symbol level, marks correspond to numbers and relations. At the structural level, the axiom schema expresses n + 1 = n + 1, and the rule carries a + b = c to a + (b + 1) = c + 1. The rule is therefore truth-preserving under this interpretation.

Source trace. Chapter II introduces the pq/addition isomorphism and the distinction between symbol-level interpretation and theorem-to-truth correspondence on printed pages 49–53. This lab is a conceptual reconstruction of that two-tier mapping; it uses switchable interpretations to expose where relation preservation succeeds or fails.

Protocol. Choose First number 2 and Second number 3, then set Interpretation to addition/intended and compare the pq theorem with its equation and truth readout. Switch Interpretation to multiplication without changing either number; the string stays a theorem while 2 × 3 = 5 becomes the visible counterexample. Finally choose swapped relation and record which symbol-role mismatch breaks the correspondence.

Limits. One counterexample can refute a universal interpretation, but any finite collection of successes cannot by itself prove a perfect infinite isomorphism. The multiplication and swapped-relation choices are diagnostic contrasts, not claims that those readings are historically proposed meanings. The lab also tests positive integers only, matching the pq-system’s limited expressive range.

The software-engineering analogue is not “code is math”; it is a serializer whose decoded records must preserve field roles and relationships across a round trip. For LLM agents, a tool schema maps model tokens to real operations only if argument positions and state effects remain aligned. In business, calling gross bookings “revenue” fails the same role-preservation test; in daily life, reading Celsius values as Fahrenheit may preserve order while destroying the intended quantities.

A handful of agreeable examples can make a bad story feel inevitable, with the consequence that labels substitute for a tested mapping. A counterexample is one valid case that contradicts a universal claim, and it is the fastest way to reject a proposed meaning.

Multiplication illustrates the method cleanly. --p--q---- happens to survive the reading 2 × 2 = 4, so one example rewards the wrong hypothesis. The next theorem in the worked family, --p---q-----, becomes 2 × 3 = 5, which is false. The interpretation fails even though the formal theorem remains perfectly valid.

Source trace. Chapter II contrasts meaningful and meaningless interpretations, multiple passive meanings, and the independence of a formal system from its interpretation on printed pages 53–56. This lab is a conceptual reconstruction of that testing problem; meaning menus, example budgets, and strictness settings are course additions.

Protocol. Select a Meaning proposal, start Examples tested low, and note whether the displayed sample supports or refutes it. Increase Examples tested while keeping the proposal fixed, and watch both the tested-case count and surviving/failed readout. Raise Test strictness so a single false mapped statement rejects the proposal, then compare the first counterexample with the formal derivation that remains valid beneath it.

Limits. Sampling can falsify a proposal but cannot certify an infinite correspondence unless the test reaches the axiom schema and production rule themselves. Test strictness is an evidential policy, not part of the pq-system. The available meanings are deliberately few, so absence from the menu says nothing about whether another meaningful isomorphism exists.

In testing code, property-based examples are useful because one failing input refutes a universal contract, while passing samples motivate but do not replace a proof or invariant. An LLM evaluation has the same asymmetry: one schema-violating action disproves “always safe,” but fifty compliant traces do not establish it. A startup should stress a pricing story across customer sizes rather than celebrate one convenient cohort; a daily-life belief such as “this route is always faster” deserves a rush-hour counterexample.

Interpretation makes three categories look like one, and the consequence is that grammatical nonsense, formal nonmembership, and factual falsehood get confused. A pq string is well formed when it has the shape hyphens p hyphens q hyphens; it is a theorem when a derivation exists; its interpreted equation is true or false in arithmetic.

Once a theorem reads as a true equation, it is tempting to use arithmetic intuition to manufacture strings, which would silently enlarge the formal system. A meaning is passive here because it lets an observer read existing strings but grants no new production rule; only the axiom schema and formal rule determine theoremhood.

This is the Requirement of Formality inherited from Day 01. Inside the pq-system, p and q are marks whose positions are checked mechanically. Outside, we may discover that their relations mirror addition or subtraction. The bridge explains the theorem set; it does not reach back and edit it.

Use this decision rule: first verify well-formedness, then verify derivability, then evaluate the chosen interpretation. Never use truth under an interpretation as an unlisted rule of production.

Day 03 needs a tested bridge rather than a favorite reading, so record an interpretation receipt with the three-line derivation of --p---q-----, the intended symbol mapping, the rule-level reason it preserves addition, and the multiplication counterexample 2 × 3 ≠ 5. The receipt should also mark which conclusions are formal, which are arithmetical, and which connect the two.

Checklist:

• [ ] Can you distinguish an axiom schema from one axiom it generates?
• [ ] Can you derive --p---q----- without appealing to arithmetic?
• [ ] Can you define interpretation and isomorphism as different parts of the bridge?
• [ ] Can you explain why one multiplication success does not save the interpretation after a counterexample?
• [ ] Can you keep well-formedness, theoremhood, and truth in separate columns?
• [ ] Can you state why meaning does not create new pq rules?

Day 03 consumes this receipt. It treats the theorem set as a visible figure, asks what can be said positively about the strings left outside it, and turns that visual question into a precise distinction between enumerable and decidable sets.

A reconstructed interpretation can look more authoritative than its source warrant, so the provenance and additions must remain explicit. The primary source is Hofstadter’s Chapter II, “Meaning and Form in Mathematics,” printed pages 46–66, read from the supplied local EPUB. The pq notation and central distinctions follow the chapter; all diagrams, interfaces, protocols, software/agent/business/daily-life applications, counterexample workflow, and interpretation receipt are original conceptual reconstructions.