Yet geologists commonly take for granted the philosophical distinction between experimental and historical approaches in their research, and consequently receive criticism from a range of skeptical observers. “Nobody was there to observe it. You are simply making assumptions about the past and extrapolating the data over long time periods. This is not science because it is not falsifiable!” If you are a geologist (or a historian, for that matter), you are probably familiar with such claims, but I am willing to speculate that few of you have found necessary occasion to defend against them. So what do you do when a majority of the public discredits historical science, even mocking it as an oxymoron? Your best bet may be to continue in your research, realizing that when applied properly (confined by a common scientific method) a combination of historical and experimental approaches is capable of producing accurate and, most importantly, falsifiable results. But in the hope that I have spiked your interest, I want to consider a recent criticism of dating methods commonly used in geology.
And if you are not a geologist, then I hope you are still curious as to how the geologic timescale is constructed, and how we know whether those methods are reliable. So click here to download a PDF of the timescale, and let’s get into it!
In a 2008 Creation Research Science Quarterly article, Dr. John K. Reed examined what he termed “the starting rotation” of dating methods – that is, four geological methods used to assign ages to rocks. The rotation includes radiometric dating, biostratigraphy, astronomical tuning, and isotope chronostratigraphy (or chemostratigraphy). [If you have no idea what any of these words mean, then you are in for a treat, because all of them are fascinating and I’m here to explain!] After a brief discussion of each method, Dr. Reed concluded:
‘The current stable of “scientific” methods is riddled by uncertainty, and a very large
element of faith is needed to believe that they constitute a valid and verifiable chronometer of Earth’s supposed 4.5 billion-year past. In reality, there is no “silver bullet,” no single absolute clock that has measured uniformitarian history.’
So we are left with the impression that: 1) we have yet to find an absolute time-piece of Earth history; 2) there is much reason to doubt the validity of published dates; 3) the scientific nature of each method is sufficiently questionable to earn “quotation marks”; and 4) there is such a thing as uniformitarian history.
I want to draft my consideration of Dr. Reed’s claims over two articles. Below, I will consider the scientific background of geologic dating methods. In the next article, I will look more specifically at the individual methods and Dr. Reed’s assessment thereof.
Using a multiplicity of geological dating methods is like taking pages from a diary
For this analogy, I only require that you have an imagination. Imagine, for example, that you discovered a box of personal diaries from the burnt ruins of an old, countryside village. Your hope is to piece together the historical details of that village — maybe to better understand its reaction to political turmoil in the major cities? — and the diaries are your only hope. But there is one problem. The diaries are old and worn down, which renders them all incomplete. Furthermore, exposure to fire/smoke, and perhaps some water damage has erased the entry date for a majority of the pages. Is it still possible to apply a scientific method to reconstructing the history?
Let’s take a look at a single diary. It appears that in the 200 pages of entries, the entry date is still clear for 15 of those pages. This provides an absolute chronometer, meaning that it allows us to assign a real age to when those 15 pages were recorded. As for the rest of the entries, we can apply some relative methods of dating. For example, we can calculate the average number of pages between pages of a known age to get an idea of how often the person made a diary entry. We may also want to investigate the continuity of each record. In other words, phrases like “it’s been a long time since my last journal entry” can tell us where time gaps may exist in the record. Lastly, we can look at specific events (festivals, dates of birth/death of villagers, mention of a meteor shower or forest fire, etc.), but for this we require the other journals. If one journal contains a specific date for the marriage of villagers A and B, then we can assign that same date to journal entries from other diaries that mention the same marriage.
Of course, our historical reconstruction does not come without significant assumptions. Foremost, we assume the diaries were constructed by methods observed today: a living person drafted each page by their own hand, and that entries were made on sequential pages and reflected their thoughts at the time. We are assuming that the calendar age of the journal corresponds to our own calendar age, and that the author was not mistaken when he/she recorded the date. Our relative dating methods also rely on assumptions about the consistency of journal entries. Using specific events as markers assumes that each journal is referring to the same event, and that the date of the entry in which it was mentioned corresponds to date it actually took place (maybe the person was recalling an event from the year before?). And that is where the scientific method comes into play. In our reconstruction, we must apply a specific criteria to how we obtain dates, and how to decide whether assumptions in our method were falsified. If using our initial method tells us that according to Journal A, villager C was born in 1824, but according to Journal B, the same villager was born in 1794, then we have falsified the method and need to refine it. On the other hand, if our refined methods consistently predict the correct age of journal entries for multiple journals, then we have good evidence that the model is reliable. In other words, imagine now that a new laboratory method allows you to obtain the journal entry date from damaged pages. If the laboratory results are consistent the age you predicted for the entry, then your model is predictive and has great scientific value. If the combined methods produce the same history from each journal, then your method is also internally consistent. When it comes to historical science, the goal is to construct a model that is both predictive and internally consistent.
How does this apply to geology? Early on, geologists dealt primarily with relative dating methods in constructing the geologic timescale. One such method was biostratigraphy, which correlates rocks based on the types of fossils they contain (like using people mentioned in diary entries). [On a side note, it was not until the 17th century that scientists widely accepted that fossils came from once-living organisms. Sound crazy? Put yourself in the shoes of a Medieval/Classical scholar, and try to describe a process by which living matter can be turned into stone without sounding like an alchemist!] The work of Nicolas Steno was seminal to modern paleontology and stratigraphy, as he provided good evidence for the biological origin of fossils and suggested that the relative ages of rock layers could be estimated by stratigraphic relationships — namely: 1) sedimentary rock layers are younger than the rocks below them; 2) sedimentary rock layers were originally deposited horizontally and were laterally continuous; 3) rocks that cut through another type of rock are younger than the rock through which they cut. Within 200 years, geologists applied the methods of Steno to rock layers around the world and constructed a rough geologic timescale. There was still one problem, however. Although the timescale predicted which rock layers and organisms were older or younger than others (the order of events), it could attach a real date to neither. Geologists had no way to obtain specific dates for any of the pages, and thus lacked an absolute chronometer.
Calendar under construction
Early geologists attempted to estimate the age of rocks using known rates of sedimentation and extrapolating backward, but the method was limited and made too many assumptions about the continuity of the rock record (much like assuming a constant frequency and length of journal entries). By the mid-twentieth century, however, the discovery of radioactivity and isotopes allowed scientists to formulate a method (radiometric dating) that could potentially assign the absolute ages for which they had so hoped.
And so they went to work. Thousands of radiometric dates were acquired using elements like potassium and argon, rubidium and strontium, uranium and lead, etc., for which radioactive isotopes decayed at a known rate. Intrinsic to the method were several assumptions: a constant decay rate, known initial concentrations, a closed system, etc. In other words, they created a scientific model and applied it to the modeled geologic timescale that had been constructed. But the real test was whether the combined model was both predictive and internally consistent. Thus rocks from strata identified as Cambrian should yield radiometric dates older than rocks from Devonian strata, which should yield radiometric dates older than Triassic strata, and so forth. Furthermore, historic volcanic rocks (from eruptions that occurred in human history) should give approximately no age at all.
It is perhaps of no surprise to you that results from the first decades of geochronology were very promising. In general, rocks predicted to be old yielded very old dates (e.g. Fairbairn et al., 1967; Welin et al., 1980), while rocks predicted or known to be young yielded rather young dates (e.g. Dalrymple, 1969). Furthermore, radiometric ages of meteorites clustered around 4.55 billion years (Patterson, 1956) – the age assigned to the Earth itself. By this point, a history of geologic events (such as major extinctions and appearances of certain organisms, ancient lava flows, etc.) had been constructed using relative dating methods. Thus geologists worked hard to assign accurate ages to events that could be used as time-markers in the geologic record. If, for example, scientists could measure the age of lava flows coincident with the Permo-Triassic extinction (the largest known extinction in Earth history) in one part of the world, they could assign the same age to rocks that recorded the fossil transition in other parts of the world. In the decades to follow, a bulk of radiometric dating results showed the modeled geologic timescale to be both predictive and internally consistent to a reasonable extent, but the model was by no means perfect. Some rocks yielded very different dates, depending on the method used. Others yielded dates that were obviously too old (or too young) to be accurate (e.g. Brewer, 1969; Dalrymple, 1969). Early on, Pasteels (1968) summarized radiometric dating methods in use, and concluded with a rather prophetic exhortation:
“All methodological approaches to geological problems are interconnected. Geochronology as such does not exist; the interpretation of the results must take into account field, petrographic, geochemical, and geophysical evidences...It is hoped that the progress of interpretative geochronology will not be retarded, but that a clearer picture of many points presently debated will shortly emerge. However, when all difficulties of interpretation have been resolved, many fundamental questions...will also be resolved. The progress of geochronology depends on the progress of geology in
general, but it may also contribute towards this general progress.” (emphasis added)
general, but it may also contribute towards this general progress.” (emphasis added)
Making an “ASS” out of “U” and “ME”
Every scientific pursuit involves assumptions – this should come as no surprise. But the conclusions reached are only valid as long as the assumptions hold. When Lord Kelvin estimated the age of the Earth to be no more than ~24 million years, he assumed the Earth started as a sphere at a given temperature, cooling only by radiative heat loss and with no heat being added thereto. The discovery of radioactivity showed that significant heat was being added to the Earth, however, thereby invalidating his conclusion. Making assumptions in science is not a bad thing, rather it is a necessity, and assumptions must be tried and verified just as the interpretations that follow from those assumptions.
A scientific model is only valid to the extent that it corresponds to reality. Gravitational theory predicts a constant downward acceleration for all objects near the Earth’s surface (~9.81 m/s^2). But what if I tried to prove the model wrong by measuring the acceleration of a feather when I dropped it? Obviously my calculation will be much lower than gravitational theory predicts, but I have done nothing to invalidate the model (by the way, I’m referring to the model of how objects are predicted to respond to the force of gravity according to gravitational theory). The reason is that the model assumes no other force acting on the object (in this case, drag from air resistance) and therefore does not correspond to physical conditions in my experiment. When a geologist analyzes a rock to obtain a radiometric age, he/she does not consider the number to be an absolute age. Rather it is a model age for when the rock/mineral was last at a given temperature. Thus inconsistent (discordant) ages do not necessarily invalidate the model (radiometric dating), which makes assumptions about the physical history of the rock/mineral being analyzed. When a geologist obtains an age that contradicts the broader model of geologic history, he/she must also verify the assumptions intrinsic to the model. Note that by this line of reasoning, radiometric dating methods do not prove the age of rocks, or the Earth for that matter, any more than dropping rocks in a vacuum proves gravitational theory. Both attempt to construct an internally consistent model that explains the relevant data while making a set of assumptions about the universe.
Before you all run off as skeptics, I’ll let you in on a little secret: science doesn’t prove anything. The goal of scientific methods is to falsify hypotheses. Science is self-correcting in that hypotheses/models not corresponding to reality are frequently disproven, while models that explain reality very well are widely accepted. Yes, widely accepted models can be overturned and paradigm shifts commonly occur. Nonetheless, this happens through mounting scientific evidence against the prevailing model and in favor of a new one that better explains the data.
“All models are wrong, but some are useful”
By definition, scientific models are a simplified representation of reality used to understand how things work. As such, they are not meant to be infallible in their predictions. Geological dating methods are scientific models used to interpret Earth history. Radiometric dating is the only method capable of yielding an “absolute age” (i.e. our calendar date) for a vast majority of Earth history, but geologists recognize it as a model that is ever being refined. The reason I have spent so much time discussing models and falsifying hypotheses is that Dr. Reed seems to misunderstand this basic concept in his article, particularly when he claims that the assumption of deep time precludes dating methods from proving deep time (i.e. that certain rocks are many millions of years old). Furthermore, he criticizes the methods apart from their intrinsic assumptions, replacing them instead with his own assumptions about Earth history, and then pronounces the case closed. Finally, he misunderstands the use of multiple, overlapping dating methods in geology, and believes that the need for multiple methods compounds the uncertainty and unreliability of individual methods, rather than strengthening the model as a whole.
Take a step back to the ‘diary reconstruction’ analogy. Each approach to interpreting history from a single diary was riddled with uncertainty and relied on falsifiable assumptions. Yet when combined, and proven to be internally consistent and predictive, the uncertainties in our reconstructed history were reduced and we could make a solid case for its accuracy. In the next article, I want to discuss uncertainties in individual dating methods and show that in a majority of cases, individual methods are consistent and predictive of one another. Pasteels (1968) was correct in his assessment that the development of geochronology would depend on advances in geology as a whole. New technologies, which allow geologists to analyze minerals on the micron scale, have greatly improved our understanding of the physics behind radioactive decay and the retention of daughter elements, thereby explaining many of the discrepancies early researchers had suspected. Better documentation and correlation of fossil species has increased the resolution at which we can investigate periods of Earth history. Advances in magnetostratigraphy (a technique that analyzes the alignment of magnetic minerals in rocks) and continued research in the Deep Sea Drilling Project have provided an additional link between sedimentary and igneous rock records. Finally, studies in the field of chemostratigraphy (my own field) continue to provide some of the most important tests of all: 1) they verify key assumptions about the nature of the sedimentary and fossil records, by providing evidence that these layers/fossils represent isochronous intervals of Earth history; 2) they test whether other methods can accurately predict the proper age of rocks around the world; 3) they allow us to identify and interpret paleoclimatic and paleoceanographic events in Earth history, such as changes in geochemical cycles and the composition of the ocean/atmosphere. If Dr. Reed and other YECs want to dismiss these models or overturn them, it will require them to provide a new, internally consistent model that better explains the range of data. So far, this model does not exist.
Brewer, M.S., 1969, Excess radiogenic argon in metamorphic micas from the eastern Alps, Austria: Earth and Planetary Science Letters, v. 6, p. 321-331.
Briden, J.C., Henthorn, D.I., Rex, D.C., 1971, Paleomagnetic and radiometric evidence for the age of the Freetown Igneous Complex, Sierra Leone: Earth and Planetary Science Letters, v. 12, p. 385-391.
Dalrymple, G.B., 1969, 40Ar/36Ar analyses of historic lava flows: Earth and Planetary Science Letters, v. 6., p. 47-55.
Fairbairn, H.W., Moorbath, S., Ramo, A.O., Pinson, W.H., Hurley, P.M., 1967, Rb-Sr age of granitic rocks of southeastern Massachusetts and the age of the lower Cambrian at Hoppin Hill: Earth and Planetary Science Letters, v. 2, p. 321-328.
Pasteels, P., 1968, A comparison of methods in geochronology: Earth Science Reviews, v. 4, p. 5-38.
Patterson, C., 1956, Age of meteorites and the Earth: Geochimica et Cosmochimica Acta, v. 10, p. 230-237.
Welin, E., Lundegårdh, P.H., Kähr, A.M., 1980, The radiometric age of a Proterozoic hyperite diabase in Vrmland, western Sweden: Journal of the Geological Society of Sweden, v. 102, p. 49-52.