Posted by: Johan Normark | April 17, 2012

2012: The correlation issue again – GMT+2

Readers of the blog may recall that I covered Gerardo Aldana’s article on the problems of the GMT correlation constant(s) almost 1.5 years ago. He argues that the correlation issue has not been resolved and that it may be off by at least 60 days. One of his arguments relates to Stela 3 at the site of Poco Winik (or Poco Uinic) which contains a known Classic period inscription that probably mentions an eclipse. Aldana writes that, “the one record from the Classic period, then, with the potential for identifying a unique date astronomically does not conform to any version of the GMT. Rather than call into question the GMT, this has led to speculations about the authenticity of the record itself; more often than not, however, the Poco Winik data is left outside of the discussion.”

In a recent discussion on facebook that has partially spilled over on one of my old posts on Aldana’s article, Barbara MacLeod points out several interesting observations by her and Michael Grofe. The eclipse date in question on the monument is 5 Kib 14 Ch’en G6 (15 July 790 according to GMT+2). The reason why Aldana is skeptical is that the actual eclipse occurred on 16 July 790 and hence the eclipse would only suit a GMT+3 correlation, and this is one day off from the correlation favored by Lounsbury and three days off from the correlation favored by 2012ers. MacLeod believes it still fits the GMT+2 correlation because the Long Count was synchronized with the tzolkin and both changed at sunset and the haab date began at sunrise. Based on Malmstrom, MacLeod believes that “Maya Day 1425516 (5 Kib) began at sunset on July 15, 790 (J) and 14 Ch’en arrived at sunrise. It was still 5 Kib 14 Ch’en when the eclipse occurred at midday, but the JDN, which changes at noon, had shifted forward from 2009801 to 2009802” (quote from facebook, April 13). See my earlier post on when the day began.

One of the issues with the GMT+2 correlation is that this is not corresponding to the tzolkin being used in the highlands of Guatemala today (see my post on Barbara Tedlock’s work at Momostenango). There is a two day discrepancy between the Classic period tzolkin and the current one and this is one of the reasons why Thompson finally settled for the GMT correlation (the one favored by 2012ers). This is also the 260-day cycle being used in Central Mexico at the time of the Spanish conquest. On facebook MacLeod mentions Lounsbury’s explanation of the two-day discrepancy that I find convincing. He “refers to the strong Postclassic Central Mexican presence in the highlands of Guatemala prior to Spanish contact, and assumes that the political/religious structure of those communities was both (1) internally synchronized (down to the smallest villages) and (2) subject to authoritarian decrees which included reconciliation of the tzolk’in (and presumably the haab) with the 584283 [GMT] correlation constant found in Central Mexico” (quote from MacLeod on facebook, April 13).  I speculate that the Aztecs may have wanted to synchronize their New Fire ceremonies throughout their hegemonic empire.

In short, the evidence seems to point to December 23, 2012, being the date of 13 baktun.



  1. Here’s the link to Malmstrom’s discussion of this:

    He gets the credit for having actually solved the Poco Uinic problem, although I was not able to determine who he was from his website; Michael recognized his address.

  2. Is there anyone in the international Mayanist community supporting the radically different correlation proposed by Wells & Fuls which would be off from GMT by a whopping four calendar rounds? See Fuls’ table comparing several correlations. At a major Maya conference in Germany earlier this year Fuls presented his arguments again (based mainly on Venus in the Dresden Codex as far as I could follow his highly technical talk) – and there were no objections raised from the other speakers or the audience.

    • I can’t speak to the lack of objections other than to say that very few Maya scholars would have the expertise to respond off-the-cuff to a technical presentation on the Dresden Venus Table which defends an alternative correlation. Michael Grofe could likely have done it (he is currently in the midst of a serious family matter); I could not. But I can take the time to consider others’ objections–notably those of Harvey and Victoria Bricker (2011:93-94)–to the Wells-Fuls correlation. The Brickers support 584283. I refer anyone with a serious interest to that volume, which should be available in a good library; I am fortunate to have it. I will recopy the summary at the end of the chapter:

      “In summary, then, there is no longer any serious doubt about the correct correlation of the Maya and Western calendars. It lies most assuredly within the 11.16 or Goodman-Martinez-Thompson family, and the variant that is best supported, causing the least strain to *both* historical and astronomical constraints, is the Modified Thompson 2 correlation constant of 584283. It is not possible to derive a correlation constant from astronomy alone, and no candidate that ignores the historical data can be taken seriously. On the other hand, astronomical considerations, which have in the past been thought by some to prevent acceptance of the GMT, are now known to be part of the ever-growing body of evidence that supports it. As Lounsbury (1983a:24) said of the GMT a quarter-century ago: “The Venus chronology now is corroborative, as has been lunar chronology from the beginning. The problem of the Dresden Codex Venus Table can no longer be held against the correlation. Rather, it may be seen now as a type case illustrating that reputed failures of this correlation may derive not from wrongness of the correlation, but from faulty analyses of some of the problems to which it has been applied.” More in a sec.

  3. From p. 93 of the same volume, Astronomy in the Maya Codices, by Harvey M. and Victoria R. Bricker, American Philosophical Society, Philadelphia, 2011:

    “However, in 1983, Floyd Lounsbury solved the problem of the Venus table for the correlation by showing that the *base date* was not intended to mark a heliacal rise; rather, it was present as a springboard for calculating an *entry date* for the table by adding to it a distance number given in the introduction to the table.” and

    “Perceived deficiencies with the use of the GMT correlation, particularly as it concerns the Venus table, have led some scholars to eschew entirely the approach to correlation that is based on an investigation of calendrical continuity between Pre- and Postconquest times and to propose putative correlations derived from astronomical considerations only–for example, that of Brian Wells and Andreas Fuls (2000), whose correlation constant of 660,208 would place any given Maya date about 208 years later in time than the GMT.” more in a sec.

  4. Archaeologists conversant with the dating methods cited and who know the reliability of the sources used to date hieroglyphic texts may have more to say on the table of Wells and Fuls suggesting that their correlation better meets the dating expectations. One expectation that is significantly absent from their table is the eclipse of Poco Uinic Stela 3—also addressed by the Brickers amid their discussions of solar eclipses. I’ve done a few (quick, hopefully accurate) calculations based on the Wells-Fuls correlation constant of 660,208, which is four Calendar Rounds plus three days greater than 584,285. Four CRs forward from the 584285 GMT ( or 15/16 July 790—JDN 2009801—recall that the JDN changes at noon UTC) put a 5 Kib 14 Ch’en (the eclipse CR recorded on Poco Uinic 3) on 26 May, 998 Julian (JDN 2085721, correlation constant 660205). A cross-check shows these JDNs to be 75920 (4 x 18980) days apart.
    So let’s have a look at the NASA Five-Millennium Catalog of Solar Eclipses:
    We have a solar eclipse on May 28, 998 (Julian), which I must admit came as a surprise.

    But it was partial and not at all visible to the Maya (Lat 64N, Long 94W).
    All else equal, given a choice between a dramatic near-total eclipse occurring at Poco Uinic and one that was only partial and only visible in Greenland, my money is on the home-based drama. But all else is *not* equal; as the Brickers point out (and as their table shows), the Wells-Fuls correlation tosses the historical data and ratifies astronomical assumptions regarding the Venus table that Lounsbury argued to be incorrect.

  5. Thanks for the replies Barb. I will read them more thoroughly later but right now my son wants to play “Scooby Doo” on this computer…

  6. […] Followers of this blog may recall that I went to Vietnam last year. What I did not mention in my posts was that I met Pố Ngô, a Vietnamese historian at HCMC University of Culture. The university library has in its possession a previously unrecorded codex (the “Saigon Codex”). According to Ngô, the codex was originally stored at Bibliothèque royale in Paris, together with the Paris Codex. Both were acquired in 1832. In 1886, the year before León de Rosny printed a chromolithographic version of Paris Codex, the Saigon Codex was reported stolen. It is believed to have been brought to HCMC/Saigon by a former French-Croatian (Dalmatian) employee at Bibliothèque royale whose family settled in Vietnam in the early 1890s. The HCMC library was given the codex after the American War. Ngô has been studying the codex for several years and contacted me in 2011 to discuss its significance for the correlation issue between the Gregorian Calendar and the Maya Long Count (which has been discussed many times before on this blog). […]


%d bloggers like this: