Despite the dictionary definition, "not sharing information" is the more fundamental of the two aspects of privacy. If surveillance seems to occur, yet the information collected is never shared with a human, in a sense privacy has not truly been violated, even if people may feel it has been.
The real privacy concerns people have today fundamentally revolve around who has access to what information, not what information is collected. One common straw man argument attributed to privacy advocates goes like this: "Privacy advocates want to never given anybody their information, such as their address. But if you wish to order a book to be delivered to your house, you must divulge your address to somebody in order to accomplish this. Since this information must be known by somebody in order to do business, it is not possible to maintain privacy and do business. Since business is important, privacy must fall by the wayside." The "straw man" nature of this argument lies in the definition of privacy as "not collecting information". Nobody disagrees that information is necessary to do business; it is what is done with the information afterwards that constitutes a privacy breach, not the collection itself.
(It is also worth observing that less information is strictly necessary then might be obvious at first, which is another subtle weakness in this straw man argument. One common example is a Post Office box, which does not correspond to a physical location as an address does. The black market uses other techniques to obscure sources and destinations, while still doing business. One method frequently seen in the movies in the context of paying off a kidnapper is the "dead drop". There are ways of making it difficult to collect useful information if the parties are motivated enough.)
I can't prove it, but I would suspect that the vast majority of information collected has some legitimate use, and is not just random surveillance for no apparent reason. Therefore, for the purposes of this essay, I will mostly drop "Freedom From Surveillance" from my consideration. The part of surveillance I want to talk about is captured entirely by considering how surveillance information is communicated.
To continue to refine the generic meanings of the term "privacy" down to what I think is truly fundamental, it is worth noting another distinction based on this separation. What we would typically consider "surveillance" is only (legally) available to the government; it is illegal for a private party to engage in many surveillance activities, for a variety of reasons: Powers reserved to the government (go ahead... try to get a search warrant as a private citizen...), the inability to place cameras in locations owned by the government, practical resource considerations. The only privacy issues concerning individuals and corporations are those concerning the sharing of information. Governments also have the ability to force information sharing to occur, especially in law investigations, because of the natural position of power it occupies. So while in theory I'd draw the separation as in the previous section, when following practical news stories, I tend to group them into two categories: Corporate privacy invasion, and government privacy invasion. Similar to the difference I drew between "censorship" and "free speech", while the effects of corporate privacy invasion and government privacy invasion may be somewhat similar, the methods used to accomplish them, and the corresponding countermeasures necessary, are quite different, and can not be lumped together as one issue without loss of clarity.