What is in the PAS Database Part III: Temporal Uncertainty and Trends in PAS Dating

This post will examine the contents of the Portable Antiquities Scheme database (PASd) from the perspective of temporal information relating to the finds records. The key fields containing finds record dating information in the PASd are:

broadperiod            Indicating the board historical period of the find, as defined by the Controlled Vocabulary section of the PAS website.
fromdate The earliest probable date of find deposition.
todate  The latest probable date of find deposition.

Note that, unlike for other object types, for single finds of coins the fromdate and todate fields tend to give either the earliest and latest probable date the coin was minted based on numismatic information, or (less commonly) the regnal years of the ruler under whom the it was struck. Also, the vast majority of non-numismatic finds have both a fromdate and todate filled in, but for coins dated to their precise year of minting it is quite common that only the fromdate field has been filled to designate said year.

In addition periodFromName and periodToName fields may indicate if a find potentially falls into more than one historical period: e.g. a find dated fromdate 1400 and todate 1800 may have the broadperiod POST MEDIEVAL and periodFromName MEDIEVAL designations. These fields are worth paying attention to as even in the historical period most non-numismatic PAS finds are dated to a dated-range of a few centuries, and therefore a considerable number straddle period boundaries. The fields subperiodTo and subperiodFrom may give further information on sub-periodisation (early, middle, late) but these are not filled in consistently. Further dating information may be given in the general description field, or in other fields depending on the find type.

The below charts show breakdown of the dating accuracy of finds falling into the four major broad periods of the historical era: Roman (43-410), Early Medieval (410-1066), Medieval (1066-1540) and Post Medieval (1500-1900). The precise dates broad periods are assigned are may be of limited archaeological relevance and arguably arbitrary, but they are useful from the perspective of database management and querying. Coins, the single most numerous object type through the entire historical period, are typically much easier to date than other object types; as such coins and non-coin objects have been separated into their own charts.

Finds other than single finds of coins:

b3 noncoins all.png


A majority of finds (that are not coin finds) in all four periods have been dated to a probable range of at least 200 years and (expect for the Roman finds) some three-quarters are dated to within 300 years. This shows that while the bulk of the finds data in the PASd does not yield itself to close chronological analysis, there is considerable potential for investigating broad patterns and changes in the material culture over a timeframe of several generations or centuries. It also suggests differences in the finds recording process between different periods. A noticeably larger percentage of Early Medieval finds than Medieval finds have been dated to within just 100 years (34 vs. 20 per cent). This could be because recorders tend to identify early medieval objects more confidently to particular sub-periods by their morphological features.

The precision of dating for single finds of coins is considerable higher:

b3 coin all.png


Most coin finds (except for the Early Medieval period) can be dated to within 10 years of their date of being struck, and over three-quarters to within just 50. Such precision has the potential to yield invaluable information on money use and economic developments on local, regional and national scales alike. This data on mint dates should not, however, be confused with deposition: coins may have continued to circulate for decades or generations, or in some cases even for a few centuries, after being struck. Here it is useful to turn to statistical models for likely circulation periods and boundaries that have been established through analysis of the composition of coin hoards. The PAS website contains further information on coinages, and for the medieval period a very useful introductory guide has been written by Richard Kelleher (Fitzwilliam Museum) and Barrie Cook (British Museum).

We will now take a closer look at how the Medieval dataset is dated. The majority of PAS finds are recovered from the plough zone, and are therefore typically divorced from their close archaeological context. The objects are usually dated by their morphological features, which means that for most their chronology is necessarily fairly loose. The graph below gives a breakdown of the dates entered into the todate and fromdate fields in the PASd between 1000 and 1600 in the broad period Medieval category for objects other than single finds of coins.

b3 noncoin dating.png

As is clear, century and half-century dates are by far the most common, joined by the period boundary dates of 1066 and 1540. Other dates are comparatively uncommon, rarely numbering more than a few hundred records. This preference for round book-end dates reflects the temporal structure of the PAS database as a whole; a similar patterns has been discussed by Julian Richard, John Naylor and Caroline Holas-Clark in their detailed analysis of the Early Medieval material in the PAS and the Early Medieval Corpus of Coin Finds (see the Viking and Anglo-Saxon Landscape and Economy-project, 2009).

The fact that it is possible to date the numismatic material to a far finer degree with confidence is demonstrated by the below graph, which sets out the same information for single finds of coins.

b3 coin dating.png

Looking only at the fromdate and todate fields, a significant minority of up to 11 per cent of all broad period Medieval single finds of coins have been dated to within the year they were struck. Even where such precision is not possible, national recoinages, the introduction of new coin types and (to a lesser extent) dating by regnal years makes it possible to narrow down the dates of a coin with greater certainty than other object types. This is reflected in the most common coin fromdates and todates:

Screen Shot 2017-09-25 at 18.08.41.png

Ten most common fromdates and todates of medieval coins. Red: during major recoinages; Yellow: release of a new coin type; Blue: regnal date

Moreover, the PASd contains several other fields dedicated to numismatic information, and it is often possible to enhance the degree of precision given in the fromdate and todate fields by examining for example the description, typeTerm, and fields containing the obverse and reverse descriptions.

The next post will discuss aoristic analysis – a powerful technique for statistically investigating data where uncertainty in dating is an issue.


One thought on “What is in the PAS Database Part III: Temporal Uncertainty and Trends in PAS Dating

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s