from Harmer, 1952, p. 282), the record creator is an English king, who writes
formally to his subordinates:
King Edward sends greetings I inform you that I have given to Bishop Giso
the land at Wedmore
An even earlier (and less formal) example is a business letter written by a cattle
overseer in ancient Egypt, about 3000 years ago (translation from Wente Meltzer,
1990, p. 31):
This is a missive to inform my lord that I am carrying outevery assignment
that has been charged to me. I am writing to inform my lord that a message
should be sent
Modern equivalents using the word 'inform' can easily be imagined:
Dear Ms Bloggs,
I write to inform you that the Board has decided
Even when creators of 'assertive' records do not use a word such as 'inform', the
propositions they state ('We have sold the property'; 'I have completed the task';
'The Board has resolved to issue new shares') could perhaps be seen as information
that they wish to convey to readers of the records concerned. Advocates of the view
that records 'contain information' rarely discuss propositions, but what they mean
by 'information' would seem to be the propositions that record creators state.
Suppose, for example, that a Board Secretary writes:
Dear Ms Bloggs,
The Board has agreed the terms of the new share issue
Here, a proposition is the only explicit content of the record. The Secretary is stating
a proposition about what the Board has agreed to do, but the record appears to be
conveying autonomous information, because the act of making a statement
remains implicit and is concealed from the reader.
However, any concept of information must take account of the possibility of
misinformation. It is legitimate to ask how far records can be said to contain or
provide information, if some of the propositions set out in them seem inaccurate,
mistaken, biased, or distorted to tell of an ideal past. It could be argued that the
cattle overseer supplied his manager with the information that he was carrying out
every task assigned to him, even if several of his tasks had actually remained
untouched. But it is also possible to contend that, when propositions are fallacious,
they cannot appropriately be labelled as information. For Luciano Floridi (2004,
pp. 42-46), a 'general definition of information' does not require information to be
truthful; truthfulness is a condition only of a 'special definition of information'.
This distinction remains characteristic of Floridi's work. Other philosophers have
considered the issue without resolving it. Some speculate that information could be
either true or false; others insist that falsehoods cannot be information (Hennig,
2014, pp.251-252).
Truth, of course, is itself a contested notion; not every commentator would accept
that propositions can be characterised as definitively true or false. A less
foundational stance suggests that propositions can only express perceptions of the
world, and that stating a proposition entails consciously or unconsciously selecting
one way of representing the matter to which the proposition refers, while excluding
others that could be equally plausible. Stating a proposition is a social practice that
necessarily reduces complex realities to manageable verbal forms.
Whenever record creators assert propositions, a cautious approach may lead us to
conclude that those propositions are open to dispute or at least to variable
interpretation. The studies cited earlier in this chapter, which seek to demonstrate
that medical records or social work files are tendentious, all focus on records that
depend on the assertion of propositions about past events. Because these records are
almost always constructed at an interval of time after - and often also at a place
distant from - the actions and events they describe, they are liable to distortion or
bias in favour of the interests of their creators or the organisations for which their
creators work.
Concerns about the reliability of propositions asserted by records creators need not
be limited to records that report retrospectively on past events. Making a statement
of any kind entails producing a particular representation of the way things were, are,
or are thought likely to be, and it cannot be a neutral practice. Of course, a simple
assertion of, for example, a cost estimate offers less scope for improvisation than
more discursive or creative forms of record, but all assertions involve record creators
in choosing to present their message in a particular way. Even when institutions or
legal systems attempt to impose regulated vocabularies, there is almost always space
left for authorial choice, which in turn may lead us to question the objectivity of the
propositions asserted in a record.
Almost certainly, however, when we encounter such propositions, we will be less
inclined to deny that the creator of the record has asserted them. Although we may
choose not to believe the Egyptian overseer's claim that he was working on each of
his assignments, we will be unlikely to doubt that he asserted a proposition to that
effect. Likewise, when we read a file of job application forms, we may question the
veracity of statements made by individual applicants, but such questioning does not
diminish our understanding of the file as a record of the statements that were made
during the application process.15
Seeing records of this kind as 'information' is less fruitful than seeing them as
representations of propositions asserted and of the acts of asserting them. In
contrast to the popular view that information wants to be 'free' and enjoy 'a life of
archives in liquid times
108
geoffrey yeo information, records, and the philosophy of speech acts
15 We can draw very similar conclusions about the metadata, or descriptions, that records'
custodians or other agents create. Records professionals seek to secure the contextualisation of
records by surrounding them with appropriate metadata, but metadata are not exempt from error,
distortion, social constraints, or human judgement. Metadata are created at specific moments of
time; they always have contexts of their own, and they can be understood as records of
propositions that were asserted in the course of a descriptive process. Although metadata are
commonly characterised as 'data about data', it is equally possible - and, from a speech-act
viewpoint, more productive - to see them as records of assertions made about other records,
entities, or relationships.
109