More precise distinctions and terminology can help us
analyze the issues raised in this article. New terms generally get accepted only
gradually as people see a use for them, so I will concentrate here on the
conceptual distinctions rather than on a creating a consistent nomenclature.
The term hypertext is often used in ways that conflate or
confuse three analytically distinguishable notions. First, there is the
hypertext medium, which is analogous to the functional properties of a book, as
opposed to either its textual contents or its physical make-up. Functionally,
the hypertext medium is one in which information units on various scales (see
below) may be linked to one another (at the same scale or across different
scales) in such a way that an almost effortless action of the user activates the
link, causing the presented information (anchor or source unit) to shift
immediately to a different unit (target or successor unit). Depending on the
implementation, the anchor item may be replaced visually by the target item, or
both may be displayed simultaneously in separate or overlapping visual spaces.
Formally this is the hypertext “reading” medium, and it may be combined
technologically with a hypertext “authoring” medium, in which a similar
action creates a link. In many cases, the choice of which items to link requires
a means of viewing or navigating among many items at once. For readers, there
may be an “overview” function and/or a “choice” function; the former
allows us to see many items, or their icons, and possibly some web of
relationships among them, and the latter enables us to choose to which of
several targets a link from a single anchor will take us. For authors, links may
be created by marking a unit as an anchor and then navigating to the target and
confirming the link, or by creating a link between the icons representing the
two units in some overview.
The medium is a functionally defined concept. It is a
medium in which we can do hypertext activity (traversing and authoring). It is
implemented by a technology in much the way that the functionality of the
“book” medium is implemented in a sheaf of bound paper sheets. There is a
difference (as noted by Arseth 1997) in that the underlying technology of
hypertext (a special case of his cybertext) may be dynamic in a sense
that is not true for the bound-paper book. The book takes no action on its own,
it is inert. The dynamics of its use arises entirely from the actions of the
reader (or writer). This is not necessarily true of hypertext technology, in
which, for example, the target of a link may change as a result of prior actions
by the user (or someone else), which modify the “state” of the hypertext.
The technology implements a set of rules by which the response of the medium to
an external action depends not merely on the physical action, but on an
“interpretation” (rule application) of that action in the context of other
factors (e.g. prior actions). It is always true that the dynamics of a semiotic
medium depends on the material system of user-plus-technology-of-the-medium, but
for hypertext (and generally for the class Arseth calls cybertext media) their
interactions are non-trivial and factorizable only if you know the underlying
rules of the technology. In a trivial or non-trivial sense, hypertext is an
“intelligent” medium, based in an algorithmic, and so a semiotically
mediated technology. The technology in which the hypertext medium is implemented
consists of all the material systems (with their inherent physical dynamics) and
the operations (both human and machine, typical practices and algorithmic
rules-for-operation) necessary to produce the functionality of the medium (cf
Harris 19.. on the technologies of writing).
Finally there is hypertext as a count-noun, a particular hypertext web, which consists of the set of information items, their relations to one another, and whatever else defines the state of the web in relation to any possible action of the user. We can define the web across its possible states by saying that this is the set of rules for the possible succession of states (which is not of course the same as the succession of presented units).
In the case of print media, we are accustomed to
distinguishing between at least two and sometimes three notions of “the
text”. There is the object-text in the sense of the material object with which
we make meaning, the material signifiers. In this sense, two otherwise identical
copies of the same book are two distinct texts. Then there is the linguistic
text, in the sense of the sequence of linguistic signifiers at the lexical and
grammatical level, and what is conceptually akin, the multimodal text in the
sense that includes the “markup” of those features of the visual
presentation of the linguistic signs which are significant for the genre (e.g.
contrasting typefaces, page designs, etc.). It is in this second sense that we
say that two different printed editions of the “same” text are or are not
the same for different purposes. Finally there is the meaning-text: the
interpreted sense made of the linguistic text, through the material medium of
the object-text. The meaning-text is not quite so simple as a sequence of
signifieds or interpretants, because it is not linear or even two-dimensional.
The topology of meanings-made is still something of a mystery, though we can at
least say that it includes patterns of meaningful relationships among items on
many scales. The fact that it may be made in time experientially also does not
imply linearity, because experiencing itself occurs on multiple, mutually
embedding and overlapping timescales.
In the case of hypertext meaning, the object text consists of the sequence of presented visual images, which may include navigational images as well as images of primary content items. The semantic or semiotic text at the level of the signifiers consists of the content and navigational items presented through these images (see below). And the meaning-text consists of the interpreted traversal through the web. We may use the term “trajectory” for the sequence of presented signs, and “traversal” for the interpreted meaning experience created by making meaning along the trajectory. Note that we make meaning with the presented signs, but not just about them, since we always bring to the meanings we make with texts, images, or sounds far more than the information value they have in themselves, or even in relation to one another. We also bring intertextual information about their typical and specific uses in other texts, and their associations with events, feelings, other signs, etc.
When we speak of any of the three senses of text in
hypertext (or hypermedia): the material signifiers, the semiotic sequences of
signs, or the interpreted meaning-text, we find that there is both an
extensional organization and a dynamic scaling. That is there are units (syntagms)
at various extensional scales (word, clause, paragraph, lexia; element, figure,
grouping, composition, whole image; note, phrase, section, whole composition),
and these come in various types and with various functional roles for their
constituents. And there are the timescales of reading/authoring, from almost
instantaneous actions and acts of recognition, to extended cumulations of sense
and retrospective and prospective interpretations and expectations, to the scale
of very long traversals through a web.
Textual scales are well known, at least up to units such as
those at the paragraph scale (which is really a scale of semiotic-textual
organization), which tend semantically to fill roles in a rhetorical or genre
structure in conventional printed texts. In traversals through and across
hypertext webs, the expectations of rhetorical structures (e.g.
problem-solution, event-consequence, claim-evidence) and genre structures remain
relevant, but there is far more emphasis, at larger extensional scales (i.e.
across more lexias) on pattern-construal of more complex kinds (semantic webs),
which may be embedded in or linked to one another (not in the hypertext as such,
but in the meaning-patterns of the interpreted traversal). I do not consider
here the special case of narrative, which I believe is a highly simplified
special case of general textuality; i.e. the meaning patterns gleaned by readers
from narrative are special cases of general semantic webs (or what I have called
elsewhere thematic formations, Lemke 1983), and narrative uses only a limited
subset of the possible meaning-making resources of language, and does so in
highly conventionalized ways that are recognizable and more easily interpretable
from their kinship to experiential recounts and simple stories heard and told in
childhood.
For semiotic modalities other than written language,
matters are somewhat less clear. In figurative painting and related realistic
depictions of familiar scenes, there are conventions of object aggregation (e.g.
the parts of the human body, whole bodies, groups of persons; furnishings,
architectural elements, persons-with-artifacts/natural objects; inset scenes,
whole compositions, triptychs, multi-panel series, etc.). For film, we have
scales which accord with timescales, e.g. event, shot, sequence, etc. And for
music, but not necessarily for other kinds of sound signs, we also have a unit
hierarchy, at least in classical western music, and also tied to timescales.
There are similar hierarchies for actions and some kinds of event-genres and
ritualized activities.
The dynamics of the use of a hypertext web implies
timescales for various operations, actions, and extended activities. There is
the almost automatic momentary recognition of a word or phrase, or of a familiar
image or sound. Then the more durational cumulation of sequences of these,
usually within a lexia (now in the sense of all that is presented for viewing
and listening without activating a link), and making integrated sense of them.
Then the longer times of looking back over a whole lexia, and of integrating its
meaning, provisionally, into the cumulative sequence of lexias in memory, and
imagining what items may be presented from various links, and deciding what
information or presentations we want and in what order, deciding which links to
click, etc. There is the timescale of actively navigating, across two lexias,
back and forth along a trajectory, around loops in a web, and eventually over a
timescale comparable to that of a whole “sitting” with the hypertext, and
finally over many such sittings to the cumulative meaning-experience of first
one traversal that spans some significant fraction of the available lexias, and
then of several traversals through and across the same web.
These dynamics are modified of course with the addition of
multiple media in several ways. Complex images may take longer to interpret and
recall and may be returned to more often for study than texts which occupy the
same amount of visual space. If the implementation of the hypertext medium
permits it, more than one lexia might be displayed at the same time to permit
connections to be made between images, or between images and text. Sounds are
conventionally presented simultaneously with some other (visually presented)
item, but being non-persistent by their nature, may need to be repeated several
times before they are as easily cumulated into the interpretation of the
traversal-to-now or into its interpreted meaning-patterns.
A hypertext web potentially includes a pattern of links among units at each of the extensional scales, for each of the media/modalities (text, images and graphics, videos and animations, sounds).