Astronomers do not Date Sappho’s ‘Midnight’ Poem

Once again the internet is all excited by some scientists’ findings that solve a historical mystery. In this case, “UTA scientists use Planetarium’s advanced astronomical software to accurately date 2500 year-old lyric poem” (as the University of Texas at Arlington press announcement puts it). Unsurprisingly, UTA’s “press release” (by which I mean “propaganda”) misrepresents the article. Despite the link to the article in the “press release,” nobody at UTA—either in media relations or in the planetarium—apparently could be bothered to read the article. I shouldn’t, therefore, be surprised that most other people trafficking in this story have likewise ignored the article. While not surprised, I am disheartened to see that even purportedly reputable, pro-science sites that typically demand “evidence” and “data” expend no effort to read the original article, i.e., to base their posts on evidence. We read over and over again some variation on “astronomers date 2,500-year-old Sappho poem,” when, in fact, article does not determine nor does it claim to determine a date for Sappho’s poem (though the authors assume a particular year).[1] This episode raises three issues:

1. UTA’s propaganda about the article and the subsequent coverage of it expose the naïve assumptions people make about a universal applicability of scientific expertise.
2. The original article reveals the problems that plague scientists’ efforts to work in areas outside their own domains of expertise.
3. Pretending that such work is interdisciplinary—that “[t]his research helps to break down the traditional silos between science and the liberal arts, by using high-precision technology to accurately date ancient poetry,” as the dean of UTA’s College of Science put it—confuses dilettantism for expertise and rigor.

Finally, in this particular instance, the article suffers from serious problems that should have stood in the way of its publication, at least in its current form.

The Rogue Classicist has a nice post on UTA’s initial propaganda as well as the general contours of subsequent coverage, see: “Problems with the ‘Scientific’ Dating of Sappho’s Midnight Poem.” His critique revolves around the twin poles of critical thinking and source criticism. While neither is, in principle, unique to any discipline, different disciplines view different problems as worthy of critical thinking and different sources as open to criticism.

I want to emphasize how the UTA “press release” as well as the reposts and other summaries are possible because they assume that scientific expertise is somehow universal, or at least extends unproblematically into non-scientific fields and supersedes whatever expertise is unique to that field. Scientific expertise, it seems, gets at universal truths—in this case, the Pleiades are a constellation that obey certain, known equations that describe how the universe has always worked. If you assume the superiority of some ambiguous, ill defined but all pervasive scientific methodology that uncovers to timeless laws of nature, then there is little reason to check the original article or to ask questions about it. It’s science.[2]

If we turn to the original article, “‎Seasonal Dating of Sappho’s ‘Midnight Poem’ Revisited” by Manfred Cuntz, Levent Gurdemir, and Martin George, we see immediately how the questions that scientists tend to ask and the answers they identify are not valid for historical work.[3] Here we come back to the issue of source criticism. Scientists will often read certain aspects of historical sources (typically documents) as unproblematic reflections of reality, usually a reality they have expertise in studying today.[4] In this case, the astronomers, Cuntz et al., assume:

• that Sappho’s poem unproblematically reflects a reality that Sappho experienced;
• that the constellation we know as the Pleiades is the same thing as the πληΐαδες in the fragment and that we can reduce the πληΐαδες to Alcyone, the bright star “near the geometrical center of the most prominent part of the cluster;”
• that μέσαι δε νύκτες is the exact same thing as our midnight, i.e., 12:00 AM, a precise moment Sappho certainly knew and meant, they claim, because she checked a clepsydra;[5]
• that Sappho wrote the poem when Sappho saw the stars in Pleiades set before midnight, i.e., 12:00 AM.[6] They double down on this assumption in note 9, where they assert on no evidence whatsoever: “…it is more reasonable to assume that she [Sappho] made her astronomical observations and wrote the poem at about the same time.” What? Why is that a reasonable assumption? Reasonable on what grounds? Reasonable to whom? Reasonable to you because that’s what you would do if you wrote poetry?
• that Sappho died around 570 so it is therefore valid to use 570 as a date for their analysis and, then, for people to conclude that Sappho was writing in 570. So powerful is 570, that Cuntz et al. offer precise dates in that year: the poem had to be composed between January 25 and March 31.[7]

With each of these assumptions they have reduced historical possibilities and poetic language to quantifiable and quantified data. They have reduced Sappho to an astronomer and the poem to a research report.

These assumptions, which go largely unexamined, then support a project that uses purported celestial phenomena to establish when (during the year) Sappho wrote the poem. While these assumptions might be defensible, they remain assumptions that no classicist or historian could have made without flagging them.[8] A more interesting and defensible article would admit these assumptions and then conclude something like: if the Pleiades in this fragment refers to the constellation and if midnight is taken to be sometime halfway through the night, then this fragment seems to describe a late winter scene. But that conclusion is not new, and that article has already been written.

In 1990 two scholars from Delft University of Technology, Herschberg and Mebius, published a more careful reading of the same fragment (they seem to have completed their research a year earlier, as reported in the annual reports for 1989). Based, it seems, on astronomical calculations, they conclude

For the Pleiads to have been visible after dark and to have set before midnight, the time of year is necessarily between mid-January and late March in the modern calendar.

They point out that the poem contains “implicit astronomical information, which must have contributed to the poem’s expressiveness to contemporary audiences,” and highlight how the poem conveyed a particular setting. They don’t assume that Sappho made any observations, which she then reported in her poem. They also don’t pretend to determine when the poem was written.[9] It is difficult to see how the Cuntz et al. article advances our knowledge of Sappho.

Unfortunately, Cuntz et al. and their article reinforces a facile (and asymmetrical) notion of interdisciplinarity that confuses dilettantism for expertise. The tools and methods of science are brought to bear on a set of non-scientific questions, with no regard to the possible misfit between the purposes for which those tools were developed and the valid ways to investigate those non-scientific sources. Here Cuntz et al. are dilettantes in the domains of history and classics, but the trappings of science and quantification give them the patina of expertise and rigor. To be sure, there is often a misfit when a set of tools and methods developed for one domain of knowledge is uncritically applied to a different domain—Cuntz et al. are just examples of a broader problem that plagues so-called interdisciplinary work. We justifiably recognize domains of expertise, even in closely related fields (e.g., physicists generally don’t do chemistry, and I’d rather not have an OBGYN give me a root canal). When scholars venture into new fields they should draw on and work with experts in those fields. In this case, however, you have astronomers running roughshod over history and classics with no apparent awareness of their own ignorance. Far from breaking down traditional silos between the sciences and the liberal arts, this article and the cavalier approach of its authors reminds us that those disciplinary silos exist for reasons and that moving between them requires considerable effort. When done well, when drawing on experts in those silos, interdisciplinary scholarship is probably worth the effort. But it does require considerable work as well as humility. When done poorly, interdisciplinary work invites mockery and derision.

Finally, there are problems with this article’s scholarly integrity. On the one hand, a cursory review of the text reveals too many passages that are only lightly filtered lines from various Wikipedia entries. In many cases, Cuntz et al. cite the same sources for the same passages that the Wikipedia entry cites, suggesting further that they relied primarily on Wikipedia (in one instance they cite Wikipedia but not the page they borrow from). For example, nearly half the discussion of Sappho comes from the Wikipedia page on Sappho (with an additional sentence from the Poetry Foundation entry on Sappho)[10]

The historical discussion of the Pleiades is also compiled in lightly or unedited form from Wikipedia pages on the Pleiades and on the Pleiades in Folklore.

On the other hand, the authors cite sources they either don’t understand or haven’t read. So, for example, they cite James Evans’s The History and Practice of Ancient Astronomy when mentioning the Pleiades in Babylonian culture. Cuntz et al. say:

The Pleiades also have been epitomized by the Babylonians, as conveyed by the astrolabe and a fragment of a circular star list (Evans, 1998) [my emphasis].

But the Babylonians didn’t have astrolabes. Astrolabes weren’t invented for centuries, many centuries. Even our earliest texts describing astrolabes don’t appear for more than a millennium after the Babylonians. A quick look in Evans shows that he does use the term “astrolabe” but uses it to refer to circular fragments of star lists. He qualifies his use:

This [the circular fragment] is usually called a circular astrolabe. However this name is not especially apt, for the word astrolabe is also used for two kinds of astronomical instruments that were developed in late antiquity and the Middle Ages. Circular star list therefore might be more suitable.[11]

So, contrary to what Cuntz et al. say, the Babylonians did not have astrolabes and circular star lists. They had circular star lists, of which fragments still exist, lists that some people refer to as “astrolabes.” Cuntz et al. could insist on using the term astrolabe—if they wanted to confuse or mislead readers—but then they don’t get to say “astrolabe and a fragment of a circular star list” [my emphasis]. One or the other, but not both. And as Evans points out, circular star list would be the better choice.

Other, similar examples include: Cuntz et al.’s citation of Renée Raphael’s review of a recent translation of Galileo’s Siderius Nuncius suggests they didn’t read Raphael’s review. Although Raphael says nothing about Galileo’s sketch of the Pleiades, by citing her review as they do, Cuntz et al. imply that her review supports their claim. There is no reason to cite Raphael’s review, particularly since this paragraph comes, almost verbatim, from the Wikipedia page on the Pleiades. In another example, in their conclusion they attribute a claim to Joan Schmelz, although the particular blog post they refer to in the notes is clearly marked as a guest post written by Stuart Dean, a former attorney who now self identifies as an “independent researcher and writer.”

In the best case, Cuntz et al.’s “‎Seasonal Dating of Sappho’s ‘Midnight Poem’ Revisited” would simply confirm what Herschberg and Mebius concluded two decades ago (and did so better and more efficiently). There is nothing new here—their newer methods do not justify more than a paragraph. We do not, however, have a best case scenario. We have a poorly constructed article that makes strong claims about the past. Because the authors all lack expertise in the field, they don’t realize that their methods and understanding of the past are, as a colleague put it, “risible.” Their conclusions border on indefensible. The writing and style is, well, Wikipedian, especially in the historical sections. We also have an article that risks violating scholarly norms and practices with respect to citations and intellectual integrity. Yet, regrettably, countless sites and so-called news outlets repeat the article’s problematic conclusions without ever bothering to look at the original article, without holding up their end of the implicit contract, i.e., checking and confirming their sources and consulting with relevant experts. Such sites have an obligation to evaluate their own sources, especially when they are a simple click away.

Alas. We seem to be trapped in an echo chamber of dilettantism where the value of shoddy “scholarship” is validated by slapdash “reporting” which, in turn, reinforces both the “scholarship” and the “reporting” on it. In our dystopian future its going to be dilettantes all the way down.