IndisputableMonolith.Numerics.Interval.Log
Numerics.Interval.Log supplies verified interval bounds for the natural logarithm, anchored on the inequality log(x) ≥ 1 - 1/x for x > 0. Researchers computing phi-ladder energies or superconducting critical temperatures cite these bounds when composing power and exponential expressions. The module lifts elementary inequalities to monotonic interval statements that integrate directly with rational-endpoint arithmetic.
claimThe module establishes $log x ≥ 1 - 1/x$ for $x > 0$, together with monotonicity of the log interval map and containment results that bound $log([a,b])$ for positive intervals.
background
The module extends Verified Interval Arithmetic, which computes rigorous bounds on transcendental functions using rational endpoints that Lean handles exactly. It also imports algebraic bounds on the golden ratio φ = (1 + √5)/2 obtained from 2.236² < 5 < 2.237². The local setting is the numerics layer that supplies controlled-error bounds for Recognition Science calculations involving log, exp, and powers.
proof idea
The module collects lemmas that apply the standard inequality log(x) ≥ 1 - 1/x, lift it via monotonicity to interval versions, and add phi-specific and Taylor-error results. Each lemma follows from basic analysis facts already available in the imported Basic and PhiBounds modules.
why it matters in Recognition Science
The module is imported by Numerics.Interval.Pow to bound x^y = exp(y log x), by Chemistry.SuperconductingTc for φ-gap ladder energy scales in superconductors, by Numerics.Interval.Tactic for bound tactics, and by Papers.GCIC.Thermodynamics for phase-structure calculations. It supplies the logarithm component required for eight-tick coherence and phi-ladder mechanisms.
scope and limits
- Does not extend to complex-valued logarithm.
- Does not provide bounds for arguments ≤ 0.
- Does not derive new inequalities beyond standard analysis results.
- Does not include floating-point implementations.
used by (4)
depends on (2)
declarations in this module (27)
-
def
logLowerSimple -
def
logUpperSimple -
def
logIntervalMono -
theorem
logIntervalMono_contains_log -
def
logPhiInterval -
lemma
phi_minus_one_abs -
lemma
phi_minus_one_abs_lt_one -
lemma
complex_norm_ofReal -
lemma
log_taylor_error_bound -
lemma
log_one_add_bounds -
def
exp_taylor_10_at_048 -
def
exp_error_10_at_048 -
lemma
exp_048_lt_phi -
theorem
log_phi_gt_048 -
def
exp_taylor_10_at_0481 -
def
exp_error_10_at_0481 -
lemma
exp_0481_lt_phi -
theorem
log_phi_gt_0481 -
def
exp_taylor_10_at_0483 -
def
exp_error_10_at_0483 -
lemma
phi_lt_exp_0483 -
theorem
log_phi_lt_0483 -
theorem
log_phi_in_interval -
def
log2Interval -
theorem
log_2_in_interval -
def
log10Interval -
theorem
log_10_in_interval