In general, structures like megaliths are dated by their archaeological context.
In most cases, a socket will have been excavated when the megalith was erected. When archaeologists excavate that socket, we're looking to retrieve evidence which might be used to provide a date for the socket, and so for the date the megalith was erected.
For C14 dating, that evidence might be provided by fragments of charcoal buried when the socket was back-filled around the megalith, or tools used to excavate the socket ( things like antler picks etc.).
Any organic material - not just artefacts that were deliberately buried at the time - can be used to obtain a C14 date.
When we excavate a site we're looking for edges that indicate features. We can trace the edge of the socket - and any later features that have been dug into it. The socket is a 'context', the megalith is a 'context', and the fill of the socket is also a 'context'.
We use these contexts, and the observed relationships between them (e.g. context x cuts into context y, therefore context y must have been there before context x was dug, and so context y is older than context x) to construct a Harris matrix. This allows us to understand the relative chronology of the site. The Wikipedia article has some helpful diagrams that illustrate the process.
Techniques like C14 allow us to get an absolute chronology for particular features/contexts within that site.
If the socket dug for a megalith (or any other structure for that matter) is left open for any length of time, the bottom of the hole tends to have a build up of silt, which we can observe during excavation. We thus have a fairly good idea of how closely the date that the megalith was erected relates to the back-fill in the hole. If there is no appreciable accumulation of silt in the hole, then the fill must be contemporaneous with the raising of the megalith, as must any organic inclusions in that material.
In the absence of excavation and dating, megalithic structures cannot be dated with confidence, as in the recent case of the stone circle in Aberdeenshire that was initially though to be a Neolithic recumbent stone circle, but which was actually built in the 1990s. (These monuments often didn't have an excavated socket, and so can be particularly hard to date).
Colin Renfrew's 'Before Civilization : the radiocarbon revolution and prehistoric Europe' is probably available from your local library, but it is worth noting that it is also available to read online at archive.org