Ever since the Linguistics bug bit me many years ago the contemplation of the tension between "orthography" and "phonology" has always fascinated me. I think it does for every linguist (great and small). Of the many things I can no longer remember of my time at Memorial U of Newfoundland one thing that I do clearly remember is that this issue was posed by the Phonology prof and I think it was in the context of whether it'd make better sense to change the English spelling system into the International Phonetic Alphabet (what linguists use to document and talk about language).
Though, over time, I've come to appreciate the more subtle aspects of such a question (and have acquired a better grasp of the vocabulary and historical context to present my philosophical positions on the issue) my position has, surprisingly, changed little - well, 'surprising' to me in any case.
Taking cue from the brilliant lectures I enjoyed in the Historical Linguistics course I remember immediately jumping into the group discussion in the Phonology course with the confidence only fools are afforded. The Historical Linguistics course was the closest I ever got to philology and I was fortunate indeed to have such an able student of Old English for a professor who taught me to appreciate the historical significance of written sources as documentations of and witnesses to language change.
My aippakuluk came back from her Toronto trip with an old Penguin Books publication titled, Linguistics, by David Crystal (1971) whose writing on the subject of "orthography" vs "phonology" is typical of how linguists frame the issue:
...at the beginning of the present century [1900s], much of the attention was taken up with the devising of appropriate techniques for the transcription of speech. This emphasis was also to be found, independently motivated, in America, where...the focus of interest was to make sure a detailed description of the dying Amerindian tribes - particularly of their languages...A phonetic transcription is no more than an attempt to make a permanent and unambiguous record of what goes on in our speech. The point which has to be emphasized is that to get such a record, we have to devise a fresh technique: our usual alphabets, which we use for everyday writing, are insufficient to do this task precisely. After all, there are only 26 basic letters in our English alphabet, but there are over forty basic sounds...We all know how English tries to get round this problem: it uses the same letter or letters for different sounds, as in the many ways in which the ough combination can be pronounced; and it gives the same sound all sorts of different spellings - the same /i/ appears in sit, women, village, busy and enough, for example. This method is both uneconomical (two or more letters for one sound), and, more importantly, highly ambiguous: we cannot predict all the time from seeing a group of letters how a word will be pronounced. English is particularly difficult in this respect, as we can see from groups of words like bough, bow (of a ship, or of a head) and bow (the weapon or the knot), where the second sounds like the first, but looks like the third. Many other languages show a much better correspondence between sounds and letters, such as Finnish or Spanish, and some, of course, like Irish Gaelic, have a much worse relationship. (p. 168)
True enough in what he says about the spelling system (orthography) of English being 'uneconomical' and 'ambiguous' (sometimes, and in some contexts), and that different writing systems of different languages have their relative strengths and weaknesses. But this analysis ignores the fact that there is a system, a real convention that works because it has developed over time as a standard of spelling and pronunciation. Whether this be by accident or design is largely immaterial precisely because it is a system of standards or agreed-upon conventional notations that account for not only efficacy, but, more importantly, aesthetics.
Now, that the abstract notion of 'aesthetics' would come into play in a writing system is a subtle thing indeed - here, I'm not talking about the form of the script (ie, syllabics vs Chinese ideographs vs latin script), but the ough combination of letters that are pronounced differently for different words and the bough, bow and bow examples that the good doctor speaks of in the above quote - but this 'supra-segmental' aspect of a given script allows for sight-reading and proficiency in the 'phonemic' systems (such as English and Hebrew for that matter) in direct contrast to the syllabic systems (such as Inuktitut or Japanese) which are truly a bit harder to master than the phonemic systems because the combination of both consonant and vowel values are 'built into' a given symbol in a phonetic system.
Many users of the Inuktitut syllabic system master 'sight-reading' and are able readers but the contrast lies in that there truly is a one-to-one relationship between symbol and sound in a phonetic system whereas in a phonemic system the context and the more abstract notion of aesthetics (and etymological, or word origin) value plays into how the /f/ sound in English is spelt 'f'', 'ph' and 'gh'.
In the Hebrew script this notion of aesthetics is even more dramatic because (outside the nichodot notation) the vowel values are not usually specified and the different ways of indicating the same sound with different symbols is purely aesthetic (for ease of differentiation of meaningful words by contrast in the script) rather than an indication of the word origin as what has motivated the evolution of the English spelling system, say, the way it has.
-Interestingly enough, the Normands that subjugated the anglo-saxons introduced the latin script that is presently used in the English-speaking world, and this historical fact determines the spelling of the gh combinations in present day English.