When items from foreign languages are pronounced by native speakers in normal speech, there is almost always an adaptation such that the foreign word is rendered using native phonetics. For instance, the other day Y-H-B was speaking to the redoubtable P. Honan and recalling the players of the Boston Red Sox who were known to me from my childhood in Japan. This conjured up the Japanese pronunciation of the name of my matchless hero Ted Williams (“The Splendid Splinter”), whose forename in that language is [detto], with the initial voiceless [t] distorted as its voiced counterpart [d], the final voiced consonant doubled by its voiceless counterpart, a vowel [o] in final position (to conform to the open-syllable structure of Japanese), and the surname rendered with a medial [r] instead of the authentic [l].
Although Japanese is a language whose speakers rarely if ever make an effort to pronounce foreign words authentically, all languages make such distortions to one or another extent, including English. These adapted renderings become standard, and no one is expected to pronounce such items “authentically,” although some speakers with a knowledge of the foreign original choose to do so in some instances. Thus no speaker of English is going to say the name of the capital of France Paris a là française, with stress on the second syllable and a silent final consonant, except as a joke, etc. This is simply a cultural and historical fact that English speakers replicate when using their native language.