I do not count the time
Who knows where the time goes?
— Sandy Denny
When I was a graduate student, my fellow student Anne Raymond (now at Texas A&M), taught me an important lesson about discussing a graph during a presentation. She insisted that you must point at each axis and describe what it says, including describing the range of data on the axis, before discussing the content of the graphs. You should move your hand (or pointer) along the axes to make clear what information is conveyed. I have since taken her advice as gospel, use it in all my talks, and try to pass the style on to my students. And, of course, I use it while I teach.
I am currently teaching Earth System History, an upper-level-undergraduate/graduate course that I informally call “historical geology on steroids.” Instead of covering historical geology as an introductory course, I cover the subject on an advanced level, including readings of recent papers from the primary literature. Other current papers are introduced during lecture. Given the nature of the topic, almost all of these papers contain a graph in which some quantity, such as an isotope value or biotic diversity, is plotted against time. We cannot discuss rates of change without making an appropriate graph, where time is carefully calibrated. Or as the EarthTime project put it “No dates, no rates.” Unlike in the classic folk-rock song, geologists count the time.
It is when I highlight the axes that the trouble begins. There is simply no consistency in how time is plotted. On some slides (I think most), time runs from left to right, with the youngest date at the right edge. So, as your eye tracks from left to right, you are going forward in time. “Years ago” are decreasing (becoming less negative). It is almost as likely, however, for time to run right to left, with the youngest date at the left end. As your eye tracks, you are going backward in time. In terms of the axis, the “years ago” are becoming more negative. There are also numerous graphs where time runs from bottom to top, mimicking the sequence of ages in a geologic column. Gazing down the column sends you back in time. Graphs with a horizontal time axis starting with the oldest age at the origin are a 90-degree clockwise rotation of the vertical axis, whereas those with the youngest age at the origin correspond to a counterclockwise rotation. A related issue is how the values are labeled; sometimes there is a negative value, for years (or kilo-, mega-, or giga-; Ka, Ma, Ga) ago, or sometimes the negative is assumed, implied by the “years ago.”
Regardless of how the axes run and are labeled, I still need to make sure the class is clear about them. Sometimes the directions shift between consecutive slides on the same subject, not only making it necessary to re-explain the axes, but making it harder to compare the two graphs.
So, I am going to make two suggestions. First, let us agree to a consistent arrow of time for graphs with time on the horizontal access. My personal preference if for time to become younger towards the right. Axis labels should omit negative values, since they are implicit in Ga, etc. Second, I encourage my colleagues interested in cognition and learning to explore if the way the time flows on a graph has any impact on the perception and understanding of the underlying information (I would be interested in hearing if any such research already exists).