It's actually the other way around. You don't change your proposition when you change your prosody, instead you produce the correct output (including prosody) based on some system of rules that relates semantics to physical output. There can be a determinate relationship between prosody and semantic interpretation. This is easy to illustrate in a wide variety of languages.
Tone is one common example of a prosodic property that can be used to signal semantic distinctions. Chinese is well-known for having words with distinct meanings signaled only by tone, e.g. mā 媽 ("mom"), má 麻 ("hemp"), mǎ 馬 ("horse"), mà 罵 ("curse"). Many languages use tone as an inflectional property where one tone pattern signals the future tense and another signals the progressive; or some other difference like "definite" vs. "indefinite", "subject" vs "oblique".
Length is also commonly used distinctively to signal semantic differences, such as in Slovak which lengthens final vowels to mark genitive plural, or in Logoori where lengthening of the mood suffix vowel marks the progressive, and lengthening the vowel of the subject prefix indicates completive. Consonant lengthening is a marker of many grammatical distinctions such as case in North Saami.
Estonian foot structure is partially controlled by reference to case (for instance), which has a clear connection to semantic interpretation. This is not so common a manifestation of the connection between prosody and semantics, because foot structure tends to be based on simple syllable and mora counting principles, but there are languages like Russian and Tarahumara where stress is based in part on grammatical category. English exemplifies this in noun / verb alternations like (N) ˈpermit ~ (V) perˈmit or (N) ˈprogress ~ (V) proˈgress.
There is, finally, a murky conceptual domain of "word grouping", which if often treated as some kind of higher-level grouping of words, just as a foot is a higher-level grouping of syllables. This can give you the semantic differences between "light housekeeper" vs "lighthouse keeper" where we have the same three words grouped together differently (as reflected by the use of space) and different stress patterns. But again, the structure dictates the prosody, it's not that prosody changes at random and causes a replacement in semantic interpretation.
