November Research Update

Amici!  It is now December!

I am sure most all of us are happy to see the last of 2020 slipping by.  It seems generally concluded, on no unsound basis, that this has been an annus horribilis.

Most of my time this month has been consumed with teaching: the compressed COVID-19 semesters meant that semesters ended before Thanksgiving rather than (as is typical) about two weeks after it, with all of the breaks and holidays removed.  That compressed schedule meant that November played host to a month and a half's worth of teaching and grading, along with the closing weeks of initial job market applications.

Meanwhile, ACOUP continues apace.  "AC:Valhalla and the Unfortunate Implications," a piece I generally thought of as a space-filler for a busy November before the real meat of the upcoming look at the Dothraki and textile production, turns out to be likely the most popular blog post I've written (it is, as I write this, just a hair behind Siege of Gondor, Part I on total views and doubtless will pass it soon).

The blog itself crossed 2,000,000 page views on Nov. 21st.  Of course I am not doing this project for raw view totals (or I wouldn't be writing lots of pedantic history essays about the structure of pre-modern farming).  But this is, of course, an exercise in public engagement and of course, to do public engagement you have to engage the public.  Outreach that doesn't reach out isn't very useful.  So I am absolutely thrilled to see the reach that this project is having, doubtless the result of a lot of word-of-mouth promotion (as I noted last month).

Now for something a bit more substantive:

Archaeological Dating

As I have been doing some revisions to my current article draft prior to peer review submission and a good portion of my argument has to do with the dates of various archaeologically recovered artifacts, I figured I could talk a bit about how those objects get their dates.  Naturally, this is going to be a summary of the matter; I am a historian, rather than an archaeologist and while I work with material culture (the fancy word historians use for most archaeological evidence, where the stuff we recover is physical evidence of a culture) I am not dating objects myself, but rather accepting and analyzing dated objects that archeologists provide for me with their work.

The system of dating objects is complex, with many different methods that are used together to try and get a sense of the 'when' of an object.  But firs to want to note that generally when we talk about dating the 'when' of an object, we do not mean the date of its manufacture or use - we specifically mean the date of its deposition, when it went into the ground for us to fine.  Almost every dating method is focused on that question (mostly because we lack evidence for the other questions) and that is important to keep in mind: an object may have been in use for years or even decades before it was thrown away, buried or otherwise deposited.

I should note, if you are hoping for some of the very fancy scientific radio-carbon stuff, you won't really find it here.  Most of those dating methods are fantastic for the deep past, but their error-bars are often measured in centuries and so for the Greek and Roman past, they're often not that useful.  Dendrochonology - tree ring dates - are a bit better, but a complete Mediterranean dendrochonological sequence is still a ways off (though increasingly coming into view, which is very exciting) and most sites don't have lots of wood to date either (it tends to decay).  The most important uses of dendrochronology and carbon dating have tended to be to provide date-ranges to areas and periods without solid historical records (Bronze Age Greece, Pre-Roman Gaul) to root the stylistic periods (so for instance, figuring out that Late Helladic IIB, a period of bronze age Greek archaeology defined by specific styles of objects and material culture assemblages, corresponds to 1550-1405BCE).  So while these dating methods are useful, they're not the most common tools.

Also, we want to note two kinds of dates: a terminus post quem ("limit after which," TPQ) is a date that the object can have been deposited no earlier than, while the terminus ante quem ("limit before which," TAQ) is a date that the object can have been deposited no later than.  

Let's start with the most direct sort of dating and more to the more complex stuff.

Coins.  Coins are lovely; we find a lot of them and they often date to the tenures of specific officials (who we can date) or the reigns of emperors (that we can date) which means that coins are some of the few objects that sometimes self-date.  For the Greek and Roman world, numismatists (coin specialists) have, over the years, constructed a fairly complete sense of when coins of every type were issues, meaning that we can often know, within a decade and sometimes to the year, when a given coin was minted.  Obviously, a coin cannot have been deposited before it was minted, so this gives us a firm TAQ.

But there's more that you can do here.  Say you have a whole bunch of coins (fairly common), all of which have a range of dates deposited together.  Generally it is safe to assume that the whole deposit probably dates quite close to the newest coin in the bunch, because the average coin horde (or person's purse) probably does have a few coins minted in the past couple of years, along with coins longer in circulation.

Now where this gets handy is that you can use those coins to date other objects in the deposit.  So if the coin evidence lets us narrow down the deposition date to, say, the 50s CE, then the other objects deposited with it can also be dated to the 50s CE.  Now, how do we know which deposits go together?  Well that leads to:

Stratigraphy:

A stratum is a layer of earth.  As time passes and dirt accumulates on the ground, it does so in identifiable layers (due to shifts in land use, climate, etc.).  Older deposits will be lower in the ground, newer deposits higher in the ground and deposits in the same stratigraphic layer are likely to be more or less contemporary.

Now of course human (and animal) activity throws all sorts of complications into this.  Humans might dig through layers and then make new deposits (burials, frequently), or construction might disrupt the stratigraphy.  Geology is also complex - stratigraphic dating generally only works for a single site, or even a single zone within a site, because conditions even a short walk away might be different, leading to different dating for the layers.

(And please note, the value of stratigraphy varies site to site.  In some areas - especially arid, upland areas - the soil layer with workable stratigraphy might be very shallow because there just isn't much erosion or soil deposition up there.  This is, for instance, a problem with the Renieblas (Numantia) site (a set of Roman camps from the second century).)

Nevertheless, careful study of the ground itself like this can let an archaeologist know that, for instance, all of the objects of a given burial were deposited together, or that a layer of trash represents a trash dump (a 'midden') active for a continuous, specific period.  That sort of thing is extremely important because it allows objects that can be firmly dated to inform the dates of objects which cannot be, if the two can be associated either to a single deposit, or a specific stratigraphic layer at a site.  Which is important when combined with:

Site Dating

Sometimes you know the year a town burned down.  Or the year a fort was set up, or abandoned (typically because a historical text tells you).  Destruction layers are typically quite visible in stratigraphy and sites that are abandoned also provide a workable TAQ for the objects there.  In cases where both the foundation date and abandonment date are known with some precision (common for Roman forts that correspond to known periods of occupation or specific campaigns) that can provide useful dating information which can structure the understanding of the stratigraphy (you can see this used to date objects at the Roman fort at Vindolanda, for instance).

These days, site dating is often used in conjunction with other forms of dating.  For instance, the various camps at Numantia (the Romans sieged the site multiple times) were dated initially to specific campaigns; Alicia Jimenez, who has resumed excavations there recently is pressing to change some of those dates, possibly shifting the site chronology about a decade later, on the basis mostly of coin and material culture evidence.  I am not yet fully convinced of this, but then I haven't seen the full assemblage.

Typological Dating

Here is where things get fun and complicated.  After decades of work, you - the plural you of 'archaeologists' - tend to accumulate quite a lot of objects of a given type.  Let's say amphora, a kind of ceramic container used for bulk transport, especially overseas.  Some of these objects have been dated, by the above methods.

You can then construct a typology of amphora (in this case, the first Roman amphora typology, which still gets used, was by Heinrich Dressel and are thus termed 'Dressel' types, thus you'll see 'Dressel IB type' and such), laying out the various types and their known dates.  Frequently, you will find clear evolution between types, much like you might see if you looked at historical fashion from, say, the 1800s - each decade clearly connected to the one before it and the one after it, but also different in its own right.

Now to be clear, to construct a clear typology like this, you need a lot of samples.  You also need them to be the sort of objects that move around, because your end goal is to be able to use the firm chronology at Site A to inform the unclear chronology of Site B (which means both sites need to have both objects).  Amphorae, the ancient equivalent of the Amazon Box, are practically perfect for the job.

Because once you can nail down each 'style' of object to a given set of decades (often made easier with large collections in sites with known destruction dates), you can use those dates to date other deposits.  So you find, say, a sword and a spear (which you cannot date typologically, since we find so few of them) deposited along with a bunch of Greco-Italic Amphorae (common in a burial, if these are grave goods) that you can date, you can date the entire deposit to the amphorae.

And, of course, the more swords and spear-heads you find the better able you may be to construct a typology of those, either to discuss the development of the weapons (obviously a keen interest of mine) or to date other things.  Weapon typologies, I will say, are finicky, but certain armor typologies are less so (particularly helmets, which tend to get used as symbols on coins and thus the design trends can be dated quite firmly).

Dates vs. Periods

Now one problem you will note here is that what we are actually doing is dating things relatively to other things.  This sword (in the ground) is about as old as this amphora (in the ground) which is about as old as this other amphora (in a museum) which was in the destruction layer of Pompeii which we can date to 79 AD, so all of these things are late first century AD.

What do you do if you are working in a period, or a society, or a place where there are no firm dates?  For instance, let's say you are working in Pre-Roman Gaul.  The Gauls wrote no history of their own, they didn't mint coins we can date with precision (some of them did mint coins, it's a funny story, remind me to talk about it some time) so unless the Romans come crashing into your site at some point, you have no firm dates at all.

What do you do?

Well, you date in periods.  In the case of Gaul, you start with a 'type site' (that is, an initial excavated site that serves as the basis of understanding for later discoveries).  For the last 400 years or so before the Romans show up (so 450-50BCE), in Gaul, that type-site is La Tene.  So you have a pile of artifacts from this site, of varying types.  With stratigraphy, you can know that some of them are older and some of them are newer, but you have no firm dates on any of them.

So we subdivide the period, numbering it (or lettering it): La Tene I, II, III or A, B, C, and D, etc, which just represent the stratigraphic periods, along with some basic inferences from how objects of a given type evolve over time.

Now, when you find another site and you find objects in it that match La Tene A and B, but not C or D, you can see, "Ah, this site is La Tene A/B; it was active in the early La Tene period, but not later."  And of course, should one of those sites have trade goods which can be firmly dated (say, an amphora of Italian wine), then you can incorporate that information into understanding when each of these stylistic sub-periods was.  These days, dendrochonology and some other fancy scientific tools can also give you date ranges, but generally such tools only work at certain sites.

That said, this sort of dating gets tricky fast, because the dating system is so far divorced from your firm dates and moreover is mostly dated on 'fashion' as it were.  And just like how today fashion can 'ripple' outward over time, the same is true of archaeological types.  La Tene B might not start at the exact time in every place it shows up!

Consequently, archaeologists working in Gaul tend to stick to dating sites based on periods and will get awfully squeamish if you ask them to put years next to those dates.  In part, it is a laudable desire to avoid providing false specificity to these objects, when the archaeologists themselves might not be sure when La Tene C1 really starts or ends at a given site, but just as often it seems to be the result of a desire to avoid advancing a theory because a theory could be wrong and proven wrong at some, unspecified later date.

I have to admit, I find that squeamishness a bit frustrating, because it forces me, as the historian, to be the one to put a chronological date range on an object, since using an object historically almost always requires a sense of its absolute date (even if this is expressed in a wide range, e.g. "object dates between 150-1 BC").  That is frustrating, because, of course, I have less information than the archaeologist on the ground to make those determinations, just because I lack a broader awareness of the site and its peculiarities.

I would also say, on this point, that there is often a real disconnect between what the archaeologists are doing in terms of compiling evidence and what historians are doing in terms of analyzing it.  By and large the two disciplines talk to each other far less than they should and archaeologists mostly write for other archaeologists, rather than for historians, which clearly I think is a bit of a mistake, if an understandable one.

All of which was occasioned by me chasing down footnote material in site reports over the last few days, often dealing with archaeologists that are either unwilling to offer a date range for their site at all (and instead stick entirely to periods) or essentially mumble their dates or precise periods by burying them in the site report rather than including them in any summary material.  Fortunately for me, my MA was in Classics, I have studied with and know quite a number of classical archaeologists and I am generally fairly literate in archaeology (although I do find foreign language site reports hard to follow - technical language in German, in particular, gets difficult fast with field-specific compound words).

Just one more tool in the historian's toolkit!