In general, it appears that "rhythm" of speech in one form or another improves intelligibility(*). Syllable "stress"— making particular syllables locally prominent in some way compared to other syllables — is one component of rhythm. It may not be strictly a necessary feature of language, but it appears to be a perceptually useful feature. It is possible that all languages use stress in one way or another (see Hirst & Di Cristo, "Intonation Systems", for a survey of how a selection of different languages from different language families use stress and intonation generally).
Now, although stress per se is an apparently universal feature of language, lexically contrastive stress as in English — i.e. where the stress position can be determined by the identity of the word in question and the difference between two words can be determined by stress pattern — is by no means universal. In many languages, such as French, which syllables carry which type of stress is essentially determined at phrase level and not dependent on the identities of particular words (though in such a language, you can still get certain classes of words that cannot carry stress vs others which do).
So in summary: rhythm/intonation is perceptually important to languages and stress patterns form part of what we perceive as "rhythm". But the precise system of stress used varies from language to language and there is little necessity for English to have the particular stress system that it does.
(*) See e.g. Tajima et al (1997), "Effects of temporal correction on intelligibility of foreign-accented English", Journal of Phonetics 25:1–24. Even with speech in which, say, the actual quality of vowels is quite far from that of a native speaker, by correcting specifically the timing of those vowels (and other segments in the speech), these researchers found that speech becomes more intelligible despite the fact that the speaker is actually pronouncing 'the wrong sound' as it were.
A. They sell all kinds of toys in here. Pick something and I'll buy it for you.
B. I don't like anything round.
A. Well, what would you like then?
B. Something red.
A. Okay, something red, but what would you like?
B. I hate that bear.
A. I don't want to know what you hate. What would you like?
B. Jane likes drawing.
A. What would you like?
So the meanings break down as:
What would you like? (as opposed to some other property such as when or how you would like it)
What would you like? (as opposed to what you wouldn't or don't like)
What would you like? (as opposed to what another person would like)
What would you like? (as opposed to what you might hate, or remember, or do something other than liking with)
Best Answer
The stress you sense actually reflects your own assumption about what the writer of each sentence intends to emphasize. This is in no way fixed or consistent. There is no rule about where to put the stress in such a phrase. If any stress is placed, it is placed on the part the writer wants to emphasize. Here's an easy way to look at it: If I wanted to emphasize that the last sentence pertained to a YOUNG man, I could stress that word and not put ANY stress on any of the others.
Don't forget that the idea of stressing any of these words comes primarily from spoken inflections, not from any intrinsic structure of a "four word phrase," which isn't a known, recognized, or established grammatical entity. In your examples, in fact, you are linking sets of words that really bear no grammatical linkage. The first is a four-word proper name, the second is a three-word proper name preceded by an adjective, and the third is a single noun preceded by two adjectives, one of which is modified by an adverb. Therefore, "four-word phrase" in no way defines what we are looking at here, and we would have no expectation that the stresses WOULD fall in the same place.