We oer this model as a substructure for discussing photocutionary acts with the rigor that
has long been available for word documents. Among the myriad of questions at play we have the
• How are we to envision the boundary between image-as-bits and image-as-photons?
• How do we speak of photon metadata—or do we?
• Can we/ought we to extract information beyond photographers’ intent?
• Are there meaningful dierences between measuring, reproducing, expressing, …?
2.2 STORY TWO: FIRST STEPS
e pertinent portion of Brian’s doctoral research was concerned with the making of abstracts
[being] primarily a matter of generating surrogates which will represent the dierences in structure
and extra-topical aspects of lm and video works by which a user might make a relevance judgment
either during browsing or during inspection of a group of documents supplied by a retrieval system.
A summary of that thinking, derived from my article (O’Connor, 1985) follows.
e lm or tape document is only an intermediate step, much like a computer program, for the
repeatable generation of a text. Light or electromagnetic radiation is modulated by the encoding
apparatus onto the substrate, producing a text which Augst (1980a) notes is: “un-attainable, in the
sense rst used by Bellour of introuvable, by being literally and guratively unquotable, everlastingly
slipping through in the instance of being identied, seized for closer scrutiny.”
Conceptualizing information as a complex relationship between societal and idiosyncratic
conventions of sender and receiver and medium, which changes an individual’s internal representa-
tions of the world argues, as Maron (1982) suggests: “e function of the document retrieval system
cannot be to retrieve all and only the relevant documents … because the system does not know
which are the relevant documents—that information simply does not exist.”
2.2.2 INADEQUACY OF CURRENT ACCESS TOOLS
Broadly speaking, words are general signs connected by convention to a concept, and made more
specic by syntax, whereas pictures are specic and made general by their context (in the moving
image document this would largely be juxtaposition with other images), which acts according to
Pryluck (1976) to “reduce unique attributes of an object to general attributes of a class of objects”.
e two sources of meaning in an image document can be seen as the object or event recorded
2.2 STORY TWO: FIRST STEPS
18 2. FIVE STORIES TO A MODEL OF VIDEO STRUCTURE
and the way in which it is presented. Pryluck (1976) also notes that the variety of recording and
post-production elements “emphasize or obscure some attributes of the object, thus aecting the
meaning consequent on the attributes of an object.”
Words and photographic images can be seen as two regions on a spectrum of modes of
symbolic description with very little in common. Pryluck (1976) assertss: “In an attempt to identify
those characteristics which are common to the sign systems and those which are unique ... beyond
the fact that both language and image communication mediate experience for recipients, few com-
mon characteristics were identied. e structures of language and image communication were
found to dier in almost every detail.”
Novitz (1977) and Pryluck (1976) point out that pictures are specically related to a par-
ticular object or event, whereas a word is general and refers, in Pryluck’s words “to a set of critical
attributes, but says nothing of other attributes which may be crucial to some meaning of the object.”
By making changes in the image coding variables one can reduce or enhance the “unique attributes
of an object” and thereby control the degree of generalization.
Metz (1974), in a similar vein, focuses on the primary dierences between words and moving
1. Shots are innite in number, contrary to words, but like statements, which can be
formulated in a verbal language.
2. Shots are the creations of the lm-maker, unlike words (which pre-exist in lexicons)
but similar to statements (which are, in principle, the invention of the speaker).
3. e shot presents the receiver with a quantity of undened information, contrary to
the word. From this viewpoint, the shot is not even equivalent to the sentence. Rather,
it is like the complex statement of undened length.
4. e shot is an actualized unit, a unit of discourse, an assertion, unlike the word (which
is purely a lexical unit), but like a statement, which always refers to reality. [
e meaning of two or more shots put together can be quite dierent from that of the indi-
vidual shots (the “third” eect); viewer perception of the same shot can be quite dierent when it is
put together with dierent shots (Kuleshov eect); timing and ordering of shots can aect perceived
meaning. Worth (1981) denes “sequence”as “a deliberately employed series used for the purpose of
giving meaning rather than order to more than one image-event and having the property of convey-
ing meaning through the sequence itself as well as through the elements in the sequence”.
Pryluck, Teddlie, and Sands (1982) used an “image-based” and non-verbal procedure to ex-
amine the “construction of meaning from sequenced images” and concluded: