Some Comments on Turner on Old World Iron
Written by
Categories: Ancient

Some Comments on Turner on Old World Iron

an outline map of Eurasia with coloured dates marked on it, all multiples of +100 or -100 except for the year 1
Map 1 from Turner 2020. “First acceleration in the use of iron across Afro-Eurasia … When iron becomes a material used for multiple object types … iron is used on a much greater scale 100 years after the proposed date and on a much smaller scale 100 years before the proposed date.”

Someone associated with the SESHAT project has taken Andre Costopoulos’ suggestion to focus on things which leave good archaeological evidence like metallurgy. They wrote a study of the spread and improvement of iron technology across the Old World. That is a topic that I am an expert on, so how does the paper hold up?

  • Turner, Edward A. L. (2020) “Anvil Age Economy: A Map of the Spread of Iron Metallurgy across Afro-Eurasia.” Cliodynamics 11.1

Because archaeologists in different countries define the start and end of the Iron Age in different ways, Turner defined four things to look for in books and articles on archaeology. He looked (1) for a date midway between iron being rare and used for a few purposes and iron being common and used for many purposes, (2) for the date at which iron versions of a common tool become more common than bronze versions, (3) for the first common use of iron helmets, and (4) the first common use of high-quality steel swords. He walks through the usual history of iron technology spreading out of northern India and Syria after 1200 BCE, quickly spreading east and west from Burma to Iberia, then booming in North China at the end of the Warring States period around 300 BCE and slowly spreading into the Baltic region and the northern forests (with a digression on 9th century BCE finds from the lower Amur River Valley in Manchuria). Creating this survey was a lot of work and it asks questions people are interested in that archaeology can answer.

Turner’s paper is structured like a natural-science paper, so to decide whether you agree with it you have to read the list of sources in appendix A and the writeup of the data in appendix B. These let you see what lies behind the dates on his maps. After I first looked at his maps, I had specific questions to focus on such as why Turner thinks Thrace was later to adopt iron technology than Greece and where his data on iron helmets in the Sasanid empire comes from. I also had questions about definitions, especially what kinds of swords count as “high quality” for him and how he deals with the fact that studies of the metal of early iron swords focus on swords from a few regions. If you look at his chart of data on helmets, you can see that these handful of points are not enough to fill out the kind of maps he draws, and that in eastern Europe, the areas which he presents as slow to adopt iron are quick to adopt iron helmets (click on the images to enlarge them!)

an outline map of Eurasia with coloured dates marked on it, all multiples of +100 or -100 except for the year 1
Map 3 from Turner 2020 “First militarily significant use of iron helmets … The spread of iron helmet data is intended to reflect the moment iron helmets entered into use for elite soldiers.”
Two side-by-side outline maps of Southwest Asia and the Balkans with coloured dates marked on them
Details from map 1 (first date of widespread iron use) and map 3 (first date that iron helmets become common) from Turner 2020. Parts of south-eastern Europe which we do not call “Greece” have iron helmets and iron use simultaneously in 300 BCE, while parts we call “Greece” have iron in 700 BCE but iron helmets only in the year 1

When I read Appendix A and Appendix B, I see a pretty good bibliography of research in English but only a handful of references in other languages (about 1 in 100 or 1 in 200 of the works in the bibliographies).[1] Archaeology is published in the language of whatever government is willing to fund it, not in a single language like English. So the less archaeology from a region is published in English, the later this article’s estimates are, especially for specific types of find like swords and helmets. I think this article overlooks sites in Thrace like Golyamata Mogila because Bulgarian archaeology is mostly published in Bulgarian and Russian and because scholars in eastern Europe can’t promote their work in the USA and UK as well as scholars at Stanford or Oxford can.[2] This article has more accurate dates from Greece, Syria, and India because there are whole departments at universities in English-speaking countries which study the Iron Age in those regions.

To study world archaeology in detail, you need a team which can read half a dozen world languages. Otherwise, you will learn about things which get powerful people who write in your native language excited, and miss things without advocates writing in your language. Because this article is based on finds published in English, it misses the chance to make some exciting points, such as that some prestigious ancient cultures like Athens were backwards in iron technology compared to less famous ones like Lydia and Thrace. We see hard-to-make objects like one-piece iron helmets and breastplates in the North Aegean and Anatolia long before we see them in the Greek peninsula (Appendix B p. 39 notes the iron helmet from Sardis, but rather than use this Anatolian helmet to understand Anatolian iron technology, it speculates whether Greeks could also have had iron helmets by the same date).

I can make some suggestions about specific details:

  • on Iron Age Eastern Europe, see E.V. Černenko’s book on Scythian armour, Daniella Agre’s bilingual book on Golyamata Mogila, and Marek Verčík’s book on large knives. My PhD thesis and first book have the citations.
  • for swords (data type: steel in Apx. A) the dates on the map don’t always agree with the definition of “high quality steel” on appendix B pages 30-32. The author seems to really like wootz / crucible steel which is well defined (despite the controversy about Alan Williams and crucible steel ULFBERHT swords); but the map shows the Roman empire having swords of “high quality steel” in the year 1. I don’t see the basis for that in appendix B, and its 200 years too late: Philon of Byzantium mentions Iberian swords which could be placed on the head and bent until they touched both shoulders then return to their form, and Peter Connolly saw some La Tenè swords from around that date survive being bent and released. As far as I know, other than Pleiner’s study of ‘Celtic swords’ and a few studies of weapons in Danish bogs there are no studies of many swords from a culture in Iron Age Europe and Southwest Asia which would let us trace changes in the quality of swords, just studies of a few objects at a time. When archaeometallurgusts study a large set of sharp tools, they usually find a wide range of quality, so studies of a few objects are usually misleading.
  • helmets (Apx. A lines 909-960) are missing two of the three iron helmets (including two early Spangenhelme) from Anatolia and Cyprus in my PhD thesis / forthcoming book. Its also missing Petros Dintisis’ Hellenistitische Helme, Encyclopedia Iranica s.v. Helmet i. In Pre-Islamic Iran (and Reallexikon der Assyriologie s.v. Helm!), and Bishop and Coulston’s overview of Roman military equipment. I think an arms and armour scholar could have helped the author find books and articles.
  • where the dates for iron helmets in Iran in the map come from is not clear to me, the idea seems to be that the iron helmets in the siege mine at Dura-Europos on the Euphrates are evidence that iron helmets were worn all over the Sasanid empire. That is a plausible guess, but there is a serious lack of documents and published archaeological finds from Iron Age Iran after most Iranians stopped burying their dead with grave goods, and in northern and eastern Iran the best archaeology is published in Russian and French. I can see an argument for only counting helmets with a one-piece iron bowl, since its easy to assemble many small pieces of low-grade iron into a helmet.
  • Appendix A line 955: the “hood of the (iron) armour” in the Gadal-Jâma contract is around 400 BCE not 500, and in Babylonia not Persia. Its probably fair to use this as evidence that arms and armour in the Achaemenid empire were at least as sophisticated as in the Neo-Assyrian empire, but see previous point.
  • some of the entries in Apx. A are texts written 50 to 500 years after the events they describe but the methodology section does not walk through the problems of combining archaeological data and literary sources
  • some of the sources cited in Apx. A are very general or just touch on the history of iron in passing. A study like this should be based on site reports or summaries of those reports.
  • Apx. A line 910: CE should be BCE
  • the bibliography in appendix B has formatting issues eg. O’Reilly 2000 has been broken into two entries on pages 42 and 55, some entries are just a URL with no other information, Kim 2014 and Kirkland 2006 are on the same line, Yılmaz 2012 is just a title in ALL CAPS with no other bibliographical information. Its not completely clear to me how the bibliography in spreadsheet appendix A relates to the one at the end of PDF appendix B and to the maps in the main article.

How to think about gaps in the evidence (“known unknowns” and “unknown unknowns”) is a hard problem philosophically speaking. I can see a case for being hard-headed and insisting on clear archaeological or documentary evidence for a “no later than” date, but that would make areas where soil conditions are good for preserving iron and where lots of archaeology is published in English look ‘advanced’ and areas with acidic soil, different customs, or archaeology rarely published in English look ‘primitive.’ And if the maps include data points based on quick-and-dirty research, these should be marked differently from data points based on specific archaeological finds or summaries by specialists. Most readers will not be able to find and assess the basis for each individual data point, and when they see a map or chart they expect that all the data comes from one kind of source, but in world history such global data sets do not exist. We have to piece together our models using the different kinds of evidence which survive from different places and times, and because the details are so important we can’t see the whole data set at once.

Because of the big problem of only using archaeology published in English, and the small problems about swords, helmets, and specific sites, I don’t think this data set is quite ready to be used. Right now, areas whose archaeology is often written about in English tend to look ‘precocious’ and areas whose archaeology is not often written about in English tend to look ‘backwards.’ I have to assume that the dates from northern Europe or Russia or East Asia or Africa have similar problems to the data in areas I can check. But I think that the basic idea of this project is reasonable, and that the hard work and hard thinking in this article could be the starting point for a team project by scholars who know archaeology published in Chinese, Russian, French, German, and English. I hope that the SESHAT team continue to talk to archaeologists and historians about what kinds of questions our data for ancient history can and cannot answer, and experiment to find projects that are big enough to be exciting but small enough to complete with the available resources.

Keep me hammering away with a monthly donation on Patreon or or even liberapay

Edit 2020-11-22: Added closeups of the parts of the maps which my knowledge allows me to assess to make the differences between map 1 and map 3 clearer

[1] About ten citations out of the 960+ citations in spreadsheet Appendix A [works by Claude Rapin, Alexander Sedov, Sergey Nefedov], and two bibliography entries out of the 380+ bibliography entries in pdf Appendix B, neither of which seems to be cited [Jarrige 1979 and Rapin and Isamiddinov 2013]. For comparison, about a quarter of the works cited in my first book are in German or French, and I had to avoid talking about places where too much archaeology is published in other languages. ↑ back to top ↑

[2] Turner’s data on Thrace comes from one chapter and the preface to one book:

  • Nikola Theodossiev. “Ancient Thrace Between the East and the West.” In Catalin Nicolae Popa. Simon Stoddart. eds. 2014. Fingerprinting The Iron Age. Oxbow Books. Oxford.
  • Peter A. Dimitrov. 2009. Thracian Language and Greek and Thracian Epigraphy. Cambridge Scholars Publishing. Newcastle upon Tyne.

It looks to me like Turner’s map treats the ‘Celtic invasions’ as the beginning of common iron use in Thrace, and I don’t think that idea would survive a week wandering through Bulgarian museums. But its not my field! ↑ back to top ↑

paypal logo
patreon logo

2 thoughts on “Some Comments on Turner on Old World Iron

  1. Edward Turner says:

    Thank you Sean for taking a serious look at my map/s.

    The data collection for this project began in about 2016.

    As I state in the paper the ‘first acceleration in the use of iron’ is the primary map of the study and was the only one mentioned in the discussion. The only one to be used for analysis.

    I consider the ‘spread of iron helmets’ and ‘high-quality steel sword’ maps to be extras, ideas that were comparatively short on data that were intended to advance or introduced discussion of the idea of them (better than no map at all). I hope the idea of improving them catches on. The high-quality steel sword map also has a methodological question mark that needs more expertise than I have to further refine. I believe the idea of restricting the spread of high-quality steel to swords to be correct (this prevents early dates on the basis of trivial finds, and makes the map possibly meaningful), the question is what counts a high-quality steel sword: it has to be much more than a minimal definition of a steel sword i.e. an iron sword with some carbon content.

    It is interesting you highlight south east Europe as an area where there can be much improvement as I had great trouble finding good sources for this region in English. The impression I got was the pre-Celtic tribes were primarily bronze-using and the switch to iron military technology came when the Celts swept through. I could only rely on the sources I had to indicate whether the level of iron-using that existed before the Celts.

    The ‘first acceleration in the use of iron’ map can and will be improved. As I said: “The dates in this paper are based on recent scholarly positions, which are changing all the time and often subject to controversy and revision.” I believe the more sophisticated methodologies I put forward for assessing the historical spread of iron (and steel) – e.g. the first acceleration in use rather than first presence of iron smelting – and the data I have collected (useful baseline) will help progress the work.

    As you say the next iteration of the map ought to involve scholars who understand multiple languages who work in universities in parts of the world outside the sphere of English-speaking academia.

    Edward Turner

    1. Sean Manning says:

      Hi Edward, I agree, it would have been fair for me to mention that the article asks readers to focus on the “first common iron use” map and treat the “helmet” and “sword” maps as tentative. They are just closer to my areas of expertise than the chronology of the Bronze Age / Iron Age transition (I focus on Southwest Asia in the middle of the first millennium BCE), and so I could give more specific comments.

      Marek Verčík (Barbarische Einflüsse p. 64) calls out Illyria as “a region with concentrations of early iron swords and cleavers” and suspects that the Greek single-edged swords were borrowed from Illyria, Macedonia, or Thrace. Surveying the spread of iron use in eastern Europe is something that archaeologists might be able to get a EU research grant for.

Write a comment

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.