If we are to provide a wider context for events visible in a
single record of the past, we need to have some method to
relate this event to other events in other records. For
example, we might wonder whether a period of population
growth evident in the genetic record is matched by specific
innovations apparent in the archaeological record that may
indicate why a larger population could be supported.
The natural way to achieve this cross-referencing between
multiple records is to relate them all to chronological time,
by dating events and processes visible in the different records.
Consequently, methods for dating are of prime importance
in the analysis of all the records of the past outlined earlier in
this chapter. Dating methods are typically dependent on
cumulative processes (often known as ‘clocks’ for obvious
reasons) that occur, on average, at known uniform rates and
can be accurately quantified. However, dating methods in
different disciplines are not equally reliable:
The physicochemical methodologies used for dating
in paleoclimatology, paleontology and archaeology
frequently depend upon the decay of radioisotopes.
This process is stochastic but on average highly
uniform and accurately quantifiable, and as a
consequence can produce very accurate dates, although
these sometimes need to be calibrated by adjusting for
past fluctuations in radioisotope levels (see Box 9.4).
Dating in linguistics is difficult and controversial, as
rates of language change are highly variable. As a
consequence, linguistic processes are generally not
reliably dated.
Dating the genetic record can utilize a number of
different ‘molecular clocks’ (see Section 6.6), but can
also be controversial. First, these molecular clocks are
difficult to calibrate accurately, and second, statistical
confidence in these dating estimates is determined at
least in part by the history of population size and
subdivision (demography), about which we have little
information.
Having used chronology to identify a correlation between
events and processes observed in different records of the past,
there is a temptation to ascribe causal relationships. A good
example in the scenario given above would be that
technological innovations revealed through archaeology
allowed the population growth that is apparent within the
genetic record. However, we must realize that it may not be
possible to prove these kinds of causal relationships with the
same degree of certainty available to other branches of
science. The study of evolution is in many ways a historical
discipline; being limited to a single past prevents us from
demonstrating causal relationships using the principle of
reproducibility.