To me one major purpose of a reference is to not only record the material (as was the case with early works such as
Cohen), but also to attempt to arrange it
per the way it was produced so that we can tie it to the
history and make connections across mints. If the
mint locations are fixed, then this
comes down to assigning to
mint (not always so simple –
RIC gets it wrong in a number of places) and determining issue order and dates. Arranging the coins in this way then becomes the basis of assigning reference numbers. In attributing a coin we can then relate it to the arrangement of the material and use that a starting point for any further studies.
This may seem to suppose too much about the purpose of reference numbers, but otherwise we may as well just assign sequential numbers to all known coins, then one can look it up and say “I have a specimen of 42!!” … and so what? As Richard Feynman says, it’s the difference between “knowing the name of something and knowing something”. A reference number that is just a tag is a name, but a reference number that provides a doorway to the
history is knowledge.
So then, if we accept that one purpose of a reference number is to identify the issue, how else should we subdivide them? Minimally we should distinguish major
reverse types and different emperors, but what else? I guess this is the
crux of the matter. Too much further subdivision based on
bust types,
obverse legends and
reverse type minutiae is the direction
RIC VI-VII seems to have taken, creating a combinatorial explosion of
attribution factors, and this self-created narrow-type problem of “is it a
type or is it a variant”. On the other hand, too little subdivision and we fail to make numismatically (and financially!) significant
bust and
obverse legend varieties, etc, easily searchable.
If these were the only options, then I think that broader (more inclusive)
types are preferable as they make the composition of the material when searching more immediately apparent. In combination with a well defined
attribution scheme (see my “don’t”s below) the
type vs variant problem would mostly go away, as the broader
types would be inclusive of many things now considered as variants of some narrower
type.
An altogether different approach to reference numbers, taken by some recent references, is to make references structured rather than linear, therefore avoiding the combinatorial explosion, as well as accommodating unseen variants. We could then, for example, have a reference number of structure issue.emperor.reverse.bust.legend.sub-type, and
software could be designed to allow us to
search on all these components, even for unseen variants catalogued
per this scheme.
As a computer geek, an even more flexible scheme that
comes to mind, designed for the computer age, would be make reference designations a
collection of (tag, value) pairs such as {(issue, PTR S-A) (
reverse,
GENIO POP ROM) (emperor,
Constantine)} etc. The value here would be that it could be extended by adding additional tags beyond those originally conceived by the reference, so as a collector or researcher you could further organize your
collection, or reference material, by adding tags of your own devising. This would of course
work best with an online reference. The whole
crux issue of broad vs narrow
types is really one created by the need to define linear reference numbers for dead-tree reference works!
Finally, it doesn’t matter what reference number scheme you use if the discriminative criteria you’re mapping to reference numbers are ill-defined! This should go without saying, but
RIC violates it freely, so I’ll say it anyway: reference numbers should be based on non-subjective numismatic criteria such that
attribution is objective and unambiguous based on inspection of a coin.
For example:
- don’t attribute based on the date the coin was issued (e.g. before/after someone else’s death, e.g.
RIC VI vs
VIII) unless it’s possible to determine that by inspecting the coin !
- don’t attribute based on things that are a matter of design execution rather than intent (e.g. did
the celator depict the epaulets on an LDC
bust, or where did he have to break the
legend to
work around the
bust)
- don’t attribute based on stylistic criteria (e.g.
bust size) that exist along a continuum
- don’t attribute based on loosely controlled variables such as
flan size that exist along a continuum
All these sort of things belong in footnotes.
Ben