Ch2.0.8 — Of the Method of rendering the Irrational Formula Rational
Summary: Extends the rationalization techniques of ch2.0.4-surd-rationalization to cubic radicands; introduces two ansatz methods (root and root ) when the constant term is a square; develops a bootstrap from any known solution; warns that each operation yields only one new , and many formulas admit only a finite number of solutions.
Sources: chapter-2.0.8
Last updated: 2026-05-09
The Problem (§112–114)
Find rational for which
has rational . Three structural differences from the quadratic case in ch2.0.4-surd-rationalization:
- No general solution: the cubic radicand admits no parametric formula. Each application of the method gives only one new value of . (source: chapter-2.0.8, §112)
- Bootstrap is mandatory: a starting solution must already be known, or be discovered by trial. (§113)
- Some formulas admit only finitely many solutions — unlike quadratic radicands, where one solution typically generates infinitely many. (§113)
The natural starting point is when (a perfect square): then trivially works, providing the seed. (§114)
Method 1: Two-Term Root (§115, §117)
Given (first term square), suppose
Squaring: . The first terms cancel; choosing kills the second; dividing by gives , hence
Example (§115): . Take root : squares give , so and . The formula evaluates to .
Example (§115): . Take root , force , so . Then , giving and root .
Method 2: Three-Term Root (§116, §117)
Suppose . Squaring:
Killing the first three terms requires and . Dividing the remainder by :
Example (§116): . Take root . Forcing the third term gives . Then , so .
Failure Mode: (§118)
When the second and third terms are both absent, both methods collapse to :
- Method 1 forces , hence .
- Method 2 forces and , hence .
The same obstacle reappears with in ch2.0.9-quartic-surd-rationalization (§131). For these “lacunary” formulas a separate seed solution is required.
Bootstrap from a Known Solution (§119–120)
If becomes a square at (with value ), substitute . The transformed formula is
whose first term is the square . Apply Method 1 or Method 2 to , then recover .
Example (§119): with seed . Substitute to get .
- Method 1 with root : forces , yields , hence .
- Method 2 with root : , , yields , hence .
Finitely-Many-Solution Phenomena
Only for (§121)
With seed : substitute to get .
- Method 1: , , — already known.
- Method 2: , , , — already known. Trying gives , whose first term vanishes, so neither method applies.
Euler concludes (with grounds for further confirmation) that becomes a square only at . Compare with the cube case in Ch2.0.10 §155.
Prolix Iteration in (§122–123)
Seeds are immediate. From (also a seed), Method 1 yields (with ), and Method 2 yields . Starting instead from gives via Method 1 and again via Method 2. Iteration produces increasingly cumbersome fractions and surprisingly does not recover the small known solution from the seed — Euler frankly calls this “an imperfection of the present method.” (§123)
Factor First When Possible (§124–126)
Some formulas dissolve trivially once factored.
Example (§124): . Since the formula must be a square and is already one, the cofactor must be a square. Set : produces a square for every . So formula ; formula ; etc.
Rule (§125): Always check first whether has rational roots (necessarily divisors of by rational-root-theorem). Each rational root contributes a factor .
Special case (§126): When both and , the formula reduces to . The square factor leaves only to satisfy: — a complete parametric solution.
Alternative Form of Method 1 (§127)
If is not fixed to remove the second term, Method 1 instead leads to the quadratic
Now the question is to make the quartic a square — which falls under Chapter IX. This shows the cubic and quartic radicand problems are intertwined.