It, of course, depends of the context, but radiocarbon dating dates the time of death of the living material used to make something or the bones of someone buried. If you want to date a statue made from wood, the 14C date you will get is when the wood ring grew, not when the statue was carved. If the statue is made of a fast growing, wide ring wood, you are considering a few years, if the statue is made of hard, thin ringed wood, the wood used might have grown over a period of a few hundred years. I onced dated a branch, 10 cm in diameter, that had 267 rings/years. You also have to consider how long the wood was dried before it was used. I onced read that wood used to make violins is dried for 40-50 years before it can be used (I am no expert in this thought). This means that if one dated the wood of a Stradivasius violin, one could revolutionarise the history of violin making, making it older and wrong. If you want to date something made from bones, one needs to assess if the animal (human or not) had an aquatic diet, as aquatic food often shows older dates, this is called a "reservoir effect". Regarding reservoir effects, look in the research of Ricardo Fernandes from Kiel (on Researchgate). I do not have cications with me as I am on vacation.

In addition to the "old wood" problem of dating wood or charcoal derived from long-lived trees, shellfish can have unique relationships with the carbon cycle (based on diet, metabolism and environment) that can lead to 14C ages anywhere up to 1000 years from their "true" age. Gastropods are particularly problematic owing to their feeding habits, but any marine (or freshwater) shellfish living in limestone environments can have major in-build radiocarbon ages. Have a look at the work by Fiona Petchey (University of Waikato Radiocarbon Dating Laboratory) on these issues for the Pacific Ocean.

What said by Marie Nadeau is tremendoulsy true and it seems, at least to my experience, that some "final users" of 14C dating data are not fully aware of these issues. In our lab we once found 14C ages on terrestrial gastropods shells which were more than 1000 years older than the expected ones. We then investigated more carefully what the reason for this could be by analyzing shells taken from living animals. We confirmed the same effect. In that study (you can find it on my profile) we attributed this to a contribution from 14C-depleted limestone in the diet. This of course should be a warning for those using these samples for dating purposes in Archaeology or Geochronology, for instance.

All these effects are "amplified" when dating recent samples by using the so called "bomb spike", for instance for forensics investigations. In this case the high resolution achievable ( of the order of few years) allows one to see age differences between different tissues sampled from the same human. In a recent study (also this in my profile) we could see age differences between different teeth, or betweem the trabecular and the cortical fraction of the same bone. This of course as the result of the different time of formation and/or the different carbon turnover rates.

It thus seems that we are now quite good in measuring the 14C content with very good precision and accuracy and the problem is now moving towards the understanding of WHAT we are really measuring in such a good way.

In a paper by Hebda et al. (2008) we radiocarbon date vole bones on a debris flow / sheet flood fan between much younger soil horizons.

"The bones occur between two prominent Ah horizons - the upper one

dating to 5520 ± 150 BP, and the lower one dating to 7132 ± 80 BP . The bones, however, yielded uncalibrated radiocarbon ages of 11 507 ± 52 years BP and 12 567 ± 49 years BP , 5000–6000 years older than the upper and lower bracketing soil horizons. "

The vole unit must have been remobilized from its original position and incorporated in a debris flow some 5000-6000 years later.

So the lesson here - understand the geomorphic context, and date more rather than less. In this case, only relying on the bracketing ages would have yielded results that were substantially too young.

Article AMS-dated Late Pleistocene taiga vole (Rodentia: Microtus xa...