I can't imagine why anyone would ever want to share an entire speech bubble, without any context (as in a screenshot).
On the other hand, it's obvious that given a message like "My address is 123 Xyzzy Rd, come anytime," the user would only want to select a subset, not the entire message.
Right, it saves typing when using some of the words elsewhere, and or when there's something in the bubble like a URL or awkward Unicode character. Copying the text is much quicker than regenerating it from scratch—or it ought to be!
Also, with an address and or phone number it saves making copying errors (I regularly copy text this way).
