Temporal Reasoning in Geosciences

Claim, Evidence, Reasoning. One school of thought in science education places great emphasis on fostering students' ability to articulate a claimabout an aspect of the world, back up that claim with evidence, and construct a coherent line of reasoning to show that the evidence does indeed support the claim.

In geology, the evidence often has to do with the timing, or sequence, or rates, of events in the past. Dozens of geologist-lifetimes have been invested in figuring out to constrain what happened at what time in earth history. And then thousands of geologist-lifetimes have been invested in using these techniques to attach dates to bits of rock or mud. Geology students spend entire courses learning to think about dates, times, and ages, via fossils, via magnetic signature of rocks and mud, via stable isotope ratios, via unstable isotope ratios, via geometry of cross-cutting relationships.

So what is the big deal about dates and ages? Why spend so much time and effort on these factoids?

Because, I would argue, temporal evidence supports distinctive and powerful forms of reasoning, which give rise to strong and important claims about process and causality.

What exactly are these forms of reasoning, "temporal reasoning"? Here is a concept map for temporal reasoning in geosciences, part of a concept map about "Time in Geosciences," one of a series of concept maps we are constructing for the Synthesis of Research on Thinking & Learning in Geosciences.

The main cluster of nodes on the temporal reasoning concept map has to do with four kinds of evidence that can lead to at least tentative or partial scientific claims: Sequence, co-occurence, rate, and cyclicity.

• Sequence: Sequence constrains causality. If A occurs before B, then A can cause or influence B, but B cannot cause or influence B. Humans use this line of reasoning in daily life, and it is a staple of crime novels and detective shows. It seems so obvious as to not need stating, but the idea that this line of reasoning has a place in supporting scientific claims is not necessarily obvious to students.
• Co-occurrence: If A and B occur at the same time, then A could cause B, or B could cause A, or C could cause both. This sounds like a pretty feeble constraint, but if you previously had no idea what caused either A or B, then this co-occurence constraint moves the inquiry forward from "inexplicable" to "multiple working hypotheses," a major step forward.
• Rate: Rate constrains power. For example, if it took a geologically instantaneously interval of time to transport and deposit a large volume of sediment, let's say a cubic kilometer, the interpreter has a small handful of plausible causal processes, including landslides and turbidity currents. All of the plausible mechanisms involve large energy release per unit time, i.e. high levels of power. If the same volume of sediment was deposited over millions of years, then a variety of low power (i.e. low energy release per unit time) causal processes become plausible.
• Cyclicity: The observation that a natural phenomenon varies over time in a methodical way, in such a way that a predictable pattern repeats after a predictable length of time, can give rise to two sorts of claims. The first type would be claims about future behavior of the Earth System. Humanity's first reliable predictions about future behavior of the Earth System were about cyclic behavior, such as the annual flooding of the Nile. The second sort of claim would be, once again, about causality. The most straightforward, but not the only, explanation for an observed cyclic phenomenon is that its proximal cause is also a cyclic phenomenon. For example, Pleistocene glacial/interglacial cycles can be explained by the periodic small variations in the orbital parameters of the Earth's rotation.

Another node of the concept map has to do with the timing of observation versus the causative processes. One's position on this node constrains the types of reasoning that can be assembled to build a chain of reasoning that connects an observable Earth product or structure with the process that caused it. There seem to be three possibilities:

• Observable active processes:In such a situation, the process that caused the product is active, and functions on a fast enough time scale that formation or modification of the product can be observed. For example, a tidal current is causing sand ripples to form and migrate. The product, the ripples, can be mapped and measured. The causative process, the current, can be measured, and its changes over time can be monitored. Changes over time in product/ripples can be compared and contrasted with changes over time in the causative process or processes. If certain changes in process reliably co-occur with specific changes in product, one can make a claim for a causal connection.
  
• Active process, too slow to observe: In such a situation, the process and conditions that gave rise to the product are still in place and active, but the process is too slow to observe on the timescale of a typical research project. Soil formation or erosion of a mountain range would be examples. The casual processes can be observed and measured, but accounting for the observations in their full magnitude requires a bold extrapolation in time.
• Product of prior processes: In such a situation, the process and circumstances that gave rise to the product or structure are no long present or active. For example, metamorphic rocks at the Earth's surface have been divorced from the process and circumstances that formed them. The current temperature and pressure of such rocks has zero explanatory power for elucidating how the rocks were formed. Instead, claims about causality have to be grounded in more indirect lines of reasoning.

The final node of the concept map sketches a connection between temporal reasoning in geosciences and temporal reasoning in other historical sciences. These disciplines are mapped in order of decreasing timescale: cosmology, geology & paleontology, archeology, history, developmental psychology. One characteristic that all of the historical sciences have in common is that any specific observation or occurence is attributable to a combination of unchanging truths, plus the particular circumstances of the moment, plus historical contingency. For example, when an earthquake shakes a building, the amount of shaking experienced is partly explainable by the universal truths of the physics of wave propagation. And it is partly explainable by the specifics of that fault's motion and the distance from the epicenter to the observation point. But a full explanation also must take into account the tectonic history of the region, which controls how shattered the rocks are along the seismic wave path and thus the wave's attenuation, and the sedimentary history beneath the building, which controls the likelihood of liquifaction. Likewise, to flesh out a full explanation of the causes of the second world war, historians take into account not just universal truths about human nature, and the immediate circumstances of the Axis nations and their individual leaders, but also the historical development that set up the circumstances. To understand a troubled adolescent, a psychologist needs to consider not just the current circumstances, but also the lifelong history of events and influences.

Temporal reasoning does not play much of a role in the parts of chemistry, physics, and life sciences that are typically taught in K-12 education. As a consequence, students often come to us without the expectation that temporal reasoning is part of science. They may use temporal reasoning in their every day lives, reasoning about the behavior of the people around them. They may recognize that such lines of reasoning can be used by historians or archeologists. But they don't consider temporal reasoning to be part of their science tool kit. We are teaching a whole new way of creating and supporting scientific claims--not just a different body of content--from their previous science courses.

---

Thanks to all of the participants in the journal club discussions on geological time and temporal thinking of the Synthesis of Research on Thinking & Learning in the Geosciences project, especially discussion leaders Cinzia Cervato and Bob Frodeman. Lynn Liben contributed the insight that Developmental Psychology is an historical science, sharing the characteristic that whatever attributes and behaviors you observe at one moment in time are a consequence of the circumstances at the moment layered on top of the lingering impacts of earlier events.

Related post:

• Extrapolating beyond the last data point

References re "claims/evidence/reasoning":

• Toulmin. (1958). The uses of argument. Cambridge: Cambridge University Press. Seminal book about how people present an argument, using law and science as the main domains of consideration. Germ of the claims/evidence/reasoning idea, but with more complex terminology and several additional elements.
  
• 
  
  Duschl, R. A. (2000). Making the nature of science explicit. In R. Millar, J. Leach & J. Osborne (Eds.), Improving science education: The contribution of research (pp. 187-206). Buckingham, UK: Open University Press. Makes the case that students collect data and look at data, without really addressing the connection (i.e. the reasoning) that connects the data to the interpretation.
  

• 
  
  McNeill, K., & Krajcik, J. (2007). Middle school students' use of appropriate and inappropriate evidence in writing scientific explanations. In M. Lovett & P. Shah (Eds.), Thinking with Data: the Proceedings of the 33rd Carnegie Symposium on Cognition. Mahwah, NJ: Lawrence Erlbaum Associates. Simplifies Toulmin's idea down to three elements (claim/evidence/reasoning), and applies the idea to middle school science education.