Hardly any written norm can be found that would implement an "ideal" correspondence of phonemes and graphemes, Vachek says. Although Finnish and Serbo-Croatian come close to such an ideal, they do not wholly conform to it. He says the nonexistence of such pure cases is in full conformity with the fact that the function of the written norm of language differs principally from that of the phonetic (and phonological) transcription, which is incapable of speaking to the eyes as quickly and distinctly as the written norm demands.

Most written norms do respect the correspondences between phonemes and graphemes to a degree, but there are also some specimens of correspondences on some level higher than that of phonemes (p. 21-22).

Vachek also considers the correspondence on the level of words -- instances of which exemplify the operation of the logographic principle (p. 23). Example: right, rite, write, wright. And then there are correspondences on the level of grammatical morphemes (p. 25); examples: s-endings to indicate plural in English despite phonemically different allomorphs /-s/, /-z/, /-iz/. Another example in English is the past tense: played (/-d/), jumped (/-t/), and waited (/-id/).

Reforming the writing system to capture the phonemic distinction instead of the morphological one would be a retrograde step because it would render the morphological information less clear than in the present, traditional way of writing. p. 25.

Also: The modern English preterite whose allomorphs /-d/, /-t/, /-id/ are, as a rule, uniformly reflected by the written suffixal morpheme -(e)d ( p. 25). "It seems certain ... that all written norms constitute various kinds of compromises between the correspondences established on various levels" p. 25.