Digital Error Correction

A digital issue has a finite list of valid answers (often just two); an analog issue spans a continuum with infinitely many. CF holds that error correction is only possible for digital issues. The argument borrows from computing: a CPU treats any voltage near 100 as “on” and near 0 as “off,” so small fluctuations get rounded back to a valid value and corrected automatically. Analog signals drift, because between any two values lie infinitely many others and no algorithm can pin down the exactly-correct point on the real number line. David Deutsch argued this for hardware; Temple’s claimed original move is applying it to epistemology.

The payoff is CF’s binary judgment of ideas: true/false and refuted/non-refuted are digital, so a criticism either shows an idea fails or it does not. CF rejects “stronger and weaker arguments” precisely because argument strength smuggles an analog scale into the one place where error correction is needed. The same critique targets credences, percentages, and gradations of certainty: rounding a continuum to whole-number percentages still leaves far too many valid values and distorts ratios.

This connects to evolutionary epistemology: knowledge grows by variation and the digital selection (keep/reject) of ideas, the same mechanism that makes life’s information-processing error-correcting. CF concedes the method is fallible and only catches errors amenable to simple categorization; deeper, systematic errors still need creative criticism. But cheap, reliable correction of many errors leaves fewer for costly methods, which is why CF insists epistemology must be digital, not analog.