Carbon dating, or radiocarbon dating, like any other laboratory testing technique, can be extremely reliable, so long as all of the variables involved are controlled and understood. Several factors affect radiocarbon test results, not all of which are easy to control objectively. For this reason, it’s preferable to date objects using multiple methods, rather than relying on one single test. Carbon dating is reliable within certain parameters but certainly not infallible.
When testing an object using radiocarbon dating, several factors have to be considered:
First, carbon dating only works on matter that was once alive, and it only determines the approximate date of death for that sample. For example, a steel spearhead cannot be carbon dated, so archaeologists might perform testing on the wooden shaft it was attached to. This provides good information, but it only indicates how long ago that piece of wood was cut from a living tree. Radiocarbon dating can’t tell the difference between wood that was cut and immediately used for the spear, and wood that was cut years before being re-used for that purpose. Nor can it tell if a much older spearhead was attached to a brand-new shaft.
Most archaeological items can’t be directly carbon dated, so their dating is based on testing done on nearby objects or materials. This makes the results subject to the researchers’ assumptions about those objects. If the spear head is dated using animal bones nearby, the accuracy of the results is entirely dependent on the assumed link between the spear head and the animal. This is perhaps the greatest point of potential error, as assumptions about dating can lead to circular reasoning, or choosing confirming results, rather than accepting a “wrong” date.
Second, radiocarbon dating becomes more difficult, and less accurate, as the sample gets older. The bodies of living things generally have concentrations of the isotope carbon-14, also known as radiocarbon, identical to concentrations in the atmosphere. When an organism dies, it stops taking in new carbon-14, and whatever is inside gradually decays into other elements. Carbon-14 normally makes up about 1 trillionth (1/1,000,000,000,000) of the earth’s atmosphere. So even brand-new samples contain incredibly tiny quantities of radiocarbon.
Eventually, the amount of carbon-14 remaining is so small that it’s all but undetectable. Tiny variations within a particular sample become significant enough to skew results to the point of absurdity. Carbon dating therefore relies on enrichment and enhancement techniques to make smaller quantities easier to detect, but such enhancement can also skew the test results. Normal errors in the test become magnified. As a result, carbon dating is only plausible for objects less than about 40,000 years old.
The other major factor affecting the results of carbon dating is gauging the original proportion of carbon-14 itself. Carbon dating is based on the loss of carbon-14, so, even if the present amount in a specimen can be detected accurately, we must still know how much carbon-14 the organism started with. Scientists must assume how much carbon-14 was in the organism when it died. Complicating matters is the fact that Earth’s carbon-14 concentrations change drastically based on various factors. As samples get older, errors are magnified, and assumptions can render carbon dating all but useless.
For example, variations in greenhouse effects and solar radiation change how much carbon-14 a living organism is exposed to, which drastically changes the “starting point” from which a radiocarbon dating test is based. Likewise, different living things absorb or reject carbon-14 at different rates. Two plants that died at the same moment, but which naturally contained different levels of radiocarbon, could be dated to drastically different times. Modern effects such as fossil fuel burning and nuclear testing have also changed atmospheric carbon-14 levels and in turn change the “starting point” for a radiocarbon test. All in all, setting the parameters of the carbon-14 test is more of an art than a science.
Contamination and repeatability are also factors that have to be considered with carbon dating. A tiny amount of carbon contamination will greatly skew test results, so sample preparation is critical. Even then, a large proportion of radiocarbon dating tests return inconsistent, or even incoherent, results, even for tests done on the same sample. The explanation given for these outliers is usually “contamination.” Inconsistent results are another reason why multiple samples, multiples tests, and various parallel methods are used to date objects.
Due to all these factors, it’s common for carbon dating results of a particular sample, or even a group of samples, to be rejected for the sole reason that they don’t align with the “expected” results. That’s not unusual in science, so far as it goes, but the relationship between assumptions and interpretations must be kept in mind. At best, it needs to be acknowledged. At worst, it can make carbon dating circular and self-confirming, though there are other means of dating that can reduce this risk.
In short, carbon dating is as useful as any other technique, so long as it’s done properly and the results are objectively interpreted. It is not, however, an inherently error-free or black-and-white method for dating objects.