Quirky British realizations of place and proper names are are source of transpondential humor. The classic is how the Marquess of Cholmondeley realizes "Cholmondeley" ('chumly'). The realization of 'Featherstonehaugh' as 'fanshaw' is another one.
There is a rather consistent rule in English that the way a person pronounces their name is always correct. The same rule applies to how local residents pronounce a place name. But foreign names, and foreign placenames moved into the US are subject to nativization (as with many languages). There are also a number of foreign place names which have received a 'native' English pronounciation/spelling (Cologne vs. Köln, etc).
Here in Iowa, we have the town of Nevada, /nəˈveɪdə/ instead of the 'correct' /nəˈvædə/, /nəˈvɑdə/, /nɛˈvædə/, etc.
The Duke of Marlborough may properly expect Americans to pronounce 'Marlborough' as he pronounces it, but as for the brand of the cigarette, this has the regularized American pronounciation.
With given names, American English regards different spellings of an etymologically related name as a different name. Anthony is not Antony, Margaret is not Margarita, Mark is not Marc is not Marcus is not Markus. Piotr is not Peter. Spanish newspapers perversely insist on callking Isabel II, but every native speaker of English says she's Elizabeth II, and that Isabel is a whole nother name altogether.
Getting back to my thesis,  I can say that the 'by the book-ness' of American English reflects to some degree my own English. If a word has a letter, that letter should be pronounced. Words like 'knot' are slightly disturbing, while words such as 'balm', 'calm' are quite disturbing to one's linguistic maps. Doing "Cholmondeley" as 'chumley' strikes us as a crime against the language.
As I have thought about this in the last 24 hours, I think I might add that anyone with a full college education is greatly influenced by the written form of his language.  For languages where the writing system and the spoken norm are at odds with each other, some sort of 'stratum develops (does 'graphostratrum' exist? If not, it's a concept that needs a word to describe it).
----- Original Message -----
From: Piotr Gasiorowski
To: phoNet@egroups.com
Sent: Sunday, April 23, 2000 4:47 AM
Subject: Re: [phoNet] English by the book.

----- Original Message -----
Sent: Sunday, April 23, 2000 6:28 AM
Subject: [phoNet] English by the book.
Spelling-pronunciations seem to be more common in the USA than elsewhere in the English-speaking world, and this reflects the Americans' respect for the written word as well as being typical of an "adoptive" language. When people feel linguistically insecure (as immigrants often do), they tend to rely on the spelling of difficult words whenever in doubt.
Here are a few characteristic examples:
Marlboro(ugh) US ['mαɹlbɝoʊ], UK ['mo:lbɹəʊ]
St James US [seɪnt'ʤeɪmz], UK [sn̩̩'ʤeɪmz]
clerk US ['klɝk], UK ['klα:k]
Berkeley, CA [bɝkli], Berkeley, Glos ['bα:klɪ]
Anthony US ['ænθəni], UK ['æntənɪ]
Thames, CT [θeɪmz], Thames, UK ['temz]

There is no real parallel for this phenomenon in any other language. We are speaking of how the *written* representation of a language fundamentally influences the speech patterns of its speakers.  The nearest parallel might be what is happening in North Africa, where Arabic is replacing the Berber languages, but this is a long term process, not the single generation language-switch found in the US. 
If you tilt your head just so, it's possible to say that American English can be described as 'native-speaker E2L'.