We might call the second type 'value-added metadata', that is, description and
arrangement schemes imposed by an archivist, other records expert, or user, or
automatically captured and repurposed trace metadata. Such metadata describes the
object in hand, augments what is not self-evident of the circumstances of its
creation and handling with the intent of making apparent not only its evidential
aspects but also the information it contains "about particular persons, situations,
events, conditions, problems, materials, and properties in relation to which the
question of action comes up" (Schellenberg, 1956). It makes the object more usable
for subsequent users, who may bring different backgrounds from those of the
creators of the object, and who may wish to use the objects for completely different
purposes than those for which they were designed. For example, a handwritten
ledger kept by a clerk of courts listing applications of Intent to Naturalize would
likely have an individual entry corresponding to each application and listing the
applicant's name, city or county of current residence, date of birth, date and port of
arrival in the country and a code linking the entry to the Intent to Naturalize form.
The primary access point to the ledger would be the date when the application was
made and the ledger would list the applications chronologically in the order
received. A genealogist trying to find an entry for an ancestor in these ledgers (and
thereby, to retrieve the relevant application form) would need to know the date
upon which the application would have been filed an unlikely data point for the
genealogist to know. Value-added description, therefore, might consist of abstracting
the ledger to create a database index to the ledger that would permit random
searching on any of the data elements contained in the ledger entries as well as the
ability to compile various statistical reports on application rates and applicant
demographics.
With a continuum approach, which "maps the creation of records as traces of
actions, events and participants, their capture into systems [broadly defined]", their
organization into an archive and their pluralization beyond the confines of their
original creator and the archive (McKemmish, 2016, p. 139), both of these types of
metadata are created at all points in the life of the object. With the life cycle
approach, value-added archival metadata would usually be created post-archival
accession. Beyond the obvious temporal differences in when the metadata are
created in these two approaches, there is an important qualitative difference
between value-added data created close to the time when the action with which it is
associated is performed (e.g., creation, active bureaucratic use or migration) and
that created later, sometimes many years later, without firsthand knowledge of those
actions but with the benefit of hindsight. Although the latter value-added metadata
are still created with reference to the object of description itself, they are also based
upon the archivist's assessments of the past and with knowledge of intervening
events that might cast the resources in a somewhat different light from that in
which they were perceived originally or at an earlier date. Again, it would be
interesting to contemplate this phenomenon with reference to that of second-hand
knowledge as described by Wilson (1986) with reference to bibliographic
description. It should also be noted that in legal theories of evidence, the closer in
time to an action that a record is created, the more likely it is to be reliable. By
implication, this also would be the case for those metadata that are essential to
supporting the reliability of that record. Ultimately, even in a life cycle approach,
however, these types are non-exclusive because, depending upon one's standpoint,
all metadata could be read as 'metadata as trace' in the sense that it is always
evidence of what has happened to the object over time and of how it has been
interpreted and valued. This discussion also neglects a third, less tangible category of
use and interpretation metadata-that which is brought to bear by a user (e.g.,
scholar, genealogist, lawyer, artist) in locating and using or exploiting an object in
the form of pre-existing knowledge, purpose, and intellectual, ethical, political or
artistic stance.
These kinds of categorizations are somewhat crude ways of conceptualizing
metadata that better support practical than philosophical understanding. From a
more conceptual perspective, can we identify any other characteristics that might
help us to apprehend the nature and role of metadata? The following discussion
suggests a few areas for further contemplation:
Metadata is what makes something a record: Fundamentally, because they are
co-constructions, any discussion of the nature of metadata requires an
understanding of the nature of a record as a conceptual and juridical as well as an
information object, and vice versa. The circular definition of metadata as 'data
about data' that was often cited in other information fields has proven to be an
overly reductivist way to talk about the kinds of highly contingent objects with
which the body of archival and recordkeeping theory and practice grapples. A record
can convey information, but its primary distinguishing characteristic is its
evidential capacity to bear witness, formally or informally. How well it does that
depends on how much metadata is associated with it, as well as how trustworthy
that metadata is. Without metadata associated with its various contexts, there may
be a poor record, or even no record at all, merely discrete pieces and points of data. In
other words, a record must have metadata, and the sufficiency and reliability of that
metadata speak to the quality of the record in terms of its comprehensiveness and
trustworthiness.
Metadata can be a record: Conceptually, metadata is a paradox, for besides serving as
a running commentary on, and validator and representation of the record(s), it can
also be construed as a record in its own right (at various levels of completeness and
granularity) and will have its own associated metadata. By extension, metadata as
records need their own metadata not only to be reliable, but also to be
understandable and referential. For example, embedded in the wall of the entrance
to one of the Oxford colleges is a plain concrete slab which reads as follows "This
Foundation Stone was Laid by Her Majesty the Queen 2 May 1968). In this case,
explanatory metadata disambiguating to which Queen it refers, and linking the
placing of the foundation stone to the broader history of the college, has been
presumed to be unnecessary since it will instead be contributed by the contextual
knowledge brought by the reader of the stone. As already discussed, explications of
metadata in archives and recordkeeping often fail to take into account the metadata
that users bring or need to bring to objects in their sensemaking process, and such
metadata may not always be relied upon to exist, or to travel well across time and
knowledge bases. For example, a visitor from another country and in a future time
may not easily be able to figure out that it was British Queen Elizabeth II who laid
the foundation stone. Metadata for films provide us with a rather different example
of meaningfulness and understandings-viewers are expected to understand
how the ordering of actor credits can indicate such things as the standing and pay
of actors. Uncredited cameos in cast lists may suggest a lack of confidence in the
quality of the film or their own performance, on the part of the unlisted actor.
Metadata can and must be used to demonstrate the fixity of records even while the
archives in liquid times
220
anne j. gilliland 'the wink that's worth a thousand words': a contemplation on
the nature of metadata and metadata practices in the archival world
221