[-empyre-] Towards a theory of digital poetics
Juan B. Gutierrez
jgutierrez at caviiar.org
Fri Mar 13 14:21:33 EST 2009
Here is where the definition proposed by Simon becomes very useful:
creative practice of association. All the distinction I tried to
establish between communicative act and communicative event was geared
toward the conditions necessary for communication. We can think of this
foundation as the epistemology of digital poetics. The creative practice
of association happens after the communicative event has been established.
Christopher Strachey's love-letter generator (1952) is a good example
for Davin's proposition. A computer program written by Strachey writes
love letters. The intentionality is not that of the computer's. The
intentionality belongs to Strachey, and is directed to the reader of the
output of the computer program, with the intention of demonstrating how
a computer could produce a human-like output (50s… these are the Turing
test years). For example, the following letter is generated:
YOU ARE MY BEAUTIFUL ARDOUR: MY AMOROUS LOVE. MY FONDNESS
DEVOTEDLY LUSTS FOR YOUR DEVOTION. YOU ARE MY ARDENT WISH. MY ANXIOUS
DESIRE AFFECTIONATELY PRIZES YOUR TENDERNESS.
M. U. C.
You can request a computer to generate many more letters at
http://www.alpha60.de/research/muc/ A new execution of the program
MY WISTFUL AFFECTION EAGERLY TREASURES YOUR EROTIC WISH. MY
RAPTURE AVIDLY HOLDS DEAR YOUR RAPTURE. MY IMPATIENT LOVE LIKES YOUR
BURNING APPETITE. MY LIKING EAGERLY PANTS FOR YOUR AFFECTION. YOU ARE MY
M. U. C.
The reader can elect to perform an exercise of interpretation around the
generated letter (i.e. his intentionality) directed to others who want
to learn how the love generator works. Or in more poetic terms, a poet
could read the generated letter and create an aesthetic interpretation
(that is his intentionality) directed to his audience. If the
directionality of the poet points to himself, then there is no
communication, and all the categories we have described remain safe. We
can even think of a reader assigning aesthetic value to the letter, with
similar processes as before.
We can use another analogy here: John enters Jane's building and runs up
four stories to get to her door. When she opens, he surprises her with a
passionate kiss. They spend the next twelve hours and four hundred pages
(Proustian?) together in her apartment drinking tea heated in a kettle.
This could fit the description of a romance novel (400 pages? Perhaps
existentialist too). But if John takes a tele-trasnporter to get to
Jane, stamps on her mouth the same passionate kiss, and heat their tea
in nanowave oven does this story belongs now to the sci-fi genre?
Probably not. We normally conceive sci-fi narratives as pieces that
shake the foundation of science and the related consequences for human
life (or existence in general) at the epistemological level. Just adding
randomly imagined gadgets to a narration with the purpose of entering
the sci-fi realm is the surest path to disband readers… and to receive
Now, let us connect the previous analogy to our reality. Is narrative
"digital" just because we read it in a computer? Does the I-Ching's
poetics acquire a digital dimension if we program the cards and see them
in a computer screen? Probably not. We could conceive as digital poetics
those pieces that could not exist without digital media. As I said
before, we have very powerful and numerous analytical devices to deal
with poetics that do not require the digital world (even if they could
live in it as a matter of convenience).
The solution to understand the I-Ching and other (potentially) literary
examples that challenge boundaries is considering three layers in the
literary construct: presentation, processing, and information, which is
a much more flexible categorization than the classic
writer-medium-reader… but the present communication is too long already,
and I do not want to punish the entire list. Let me know if you want to
keep the discussion.
Juan B. Gutierrez
davin heckman wrote:
> I really like this discussion. But I cannot help but wonder about
> which direction the poetic event works. Can the act be an event for
> the poet? Maybe it is always an event for the poet? When thinking of
> electronic literature, particularly generated texts, I think that
> there is the act, which directs some intention through the computer.
> But sometimes the event is what the poet discovers, rather than what
> we read, and then we just mess around with the traces. Like reading
> an epitaph.
> A more archaic way to think of this might by the I Ching, which is an
> act, directed through the randomizing process, and the event is
> experienced by the initiator.
> It's not such a sharp distinction, but here the interface certainly
> troubles the notion of writer-medium-reader.
> On Thu, Mar 12, 2009 at 11:07 AM, Juan B. Gutierrez
> <jgutierrez at caviiar.org> wrote:
>> Thanks to Simon and all others for such a wonderfully catalytic
>> discussion. Simon says:
>> «What I am seeking to do here is to separate poetics from human intent
>> and authorship and regard it instead as a phenomena of things.»
>> «Indeed, the bulk of this research is being carried out by teams of
>> computer scientists and linguists and involves artificial systems
>> interpreting and responding to body language, facial expression, vocal
>> tone, gesture and speech without building precise models of what is
>> occurring but functioning as sets of dynamic contingencies and
>> probabilities that may or may not require resolution prior to action. I
>> consider what they are seeking to do as digital poetics – which is
>> possibly why we find it easy to work together on artistic projects.»
>> Digital has become a wildcard that loosely means "through he use of a
>> computing device" in colloquial language. When we ask about digital
>> poetics, there are two dimensions of these questions: (i) the aesthetics
>> of a piece of art (probably of linguistic nature) that cannot be without
>> a computing device, and (ii) the aesthetics of a piece of art (probably
>> of linguistic nature) generated by a computing device.
>> I agree with both definitions of poetics suggested by Simon: (i)
>> "creative practice of association" and (ii) "The motile engagement with
>> the interplay of dynamic elements". My position is that it is impossible
>> to separate poetics from a fundamental attribute of human beings
>> (intentionality -- precise definition below), and that poetics generated
>> by machines is impossible to occur unless machines have this same
>> attribute; therefore, poetics is not an intrinsic attribute of things,
>> but an elaboration, a derivative, of intentionality.
>> Efforts to produce autonomy and some sort of intentionally have come a
>> long way since Turing. Complex probabilistic systems give raise to some
>> human-like behaviors, especially in regards to the ability of
>> classifying patterns. But we are very, very far away from producing an
>> entity with intentionality. In my work in biomedical research at CAVIIAR
>> (Advanced Research Center in Artificial Intelligence), I have found that
>> a coupled ensemble of what we call today "intelligent systems" can
>> exhibit emerging properties not present in the individual components.
>> Perhaps some day in the future (next year? next century? next
>> millennium?) we will be able to produce a system capable of
>> intentionality. But models, no matter how complicated we believe they
>> are, are only rudimentary and insufficient at this point. We have to
>> produce new mathematical tools (in a broad sense, including logic) to
>> solve this problem… if we ever can… Penrose has suggested that
>> consciousness is the result of sub-quantic phenomena, i.e. absolutely
>> out of reach for our present instrumentation and modeling ability.
>> For the reasons described above, I believe that the quest of defining
>> digital poetics needs to abandon the path of poetics generated by
>> machines, and passes through the influence of the media in human poetic
>> activities. Particularly, interactivity. Non-interactive digital pieces
>> with aesthetic intention can be studied with the analytical devices we
>> already have in, for example, literary criticism. Interactive digital
>> pieces require a new foundational framework for their study. The work of
>> Aarseth goes in this direction, but I find it insufficient.
>> Wardrip-Fruin has elaborated more, but something is still missing. I
>> have a proposal (BTW, elaborated with Laura Borras, Mark Marino and
>> Pablo Gervas), which *might* get published soon.
>> Now, let us ask, what does it mean to have interactivity in digital
>> media? I begin by defining what I call *communicative act* and
>> *communicative event*. The communication is a gregarious act; we can
>> only speak of *communicative event* when in it takes part two or more
>> interlocutors. For instance, if a shipwrecked puts a message in a
>> bottle, it would be a *communicative act*. But nothing guarantees that
>> the message would be read. In another example, if someone establishes a
>> monologue during a stroll by the mountains, that would be a
>> *communicative act*. But when someone reads the message, or listens
>> while the walker talks, then there is a *communicative event*.
>> The difference that we want to establish is that *communicative act* is
>> the one that shows intentional mental states (according to the meaning
>> of *intentionality* that we define below), whereas the *communicative
>> event* happens only in the interactivity between two subjects. Thus, the
>> shipwrecked performs a *communicative act*; there is a *communicative
>> event* when somebody else reads the message and occurs information flow.
>> The capacity of the *communicative acts* (including, but in a ampler
>> sense, the speech acts) to represent objects and states of things of the
>> world is an extension of abilities biologically more fundamental of the
>> brain, that is, *intentionality*. This attribute relate the organism to
>> the world by means of mental states such as belief or desire, and
>> specially through action and perception (John Searle has a wonderful
>> discussion about *intentionality* in "The Construction of Social
>> Reality", 1997).
>> Only some mental states and events, not all, have *intentionality*. The
>> beliefs, fears, hopes and desires are intentional; but there are forms
>> of nervousness, happiness, and anxiety not-directed that are
>> non-intentional. Thus, there are depression states that are not
>> depression of, or of joy that is not joy of, or sentences that do not go
>> directed to. Try to guess, for example, to whom is directed the
>> carpenter's exclamation when he hammers his finger by accident.
>> We will only speak of *intentionality* when the acts have
>> *directionality*, the property of the *communicative act* that makes it
>> act of communicating something. The emitter shows (*communicative act*)
>> a state or mental event directed to something, how late it is, for
>> example, to somebody by his side (*communicative event*), by means of a
>> raised hand and pointing a wristwatch while yawning. We use the term
>> *directionality* in two senses: the heat towards which it goes directed
>> the mental event, and the person to whom the gesture goes directed. It
>> is not of interest to us to limit the use of the term; rather we will be
>> able to determine the directionality of *directionality* according to
>> the context in which we use it. We will only speak of *communicative
>> act* when *intentionality* and *directionality* exist. In the
>> correspondence of the interlocutor occurs the *communicative event*. Let
>> us note that a *communicative event* can occur with a mistaken
>> *directionality*, e.g. the message in a bottle of a shipwrecked person
>> begging for help can arrive at the beach of an inhabited island by
>> another castaway.
>> When the reader interacts with a work of electronic literature, for
>> example, he must execute an action by means of a *communicative act*,
>> which has a *directionality* aiming at the computer. The computer
>> receives the reader's action and a *communicative event* is completed.
>> Then the computer processes the information and it feeds back to the
>> reader an information fragment that corresponds to that man-machine
>> interaction, that is to say, it executes a *communicative act*, which
>> has a *directionality* aiming at the reader. Thus a cycle of feedback is
>> completed. Since the information that is given to the reader depends on
>> the processes that happened in the machine, we can claim that *the main
>> characteristic of the digital media is that it acts on the message*.
>> That is the fundamental difference between the interactivity in digital
>> media with respect to interactivity in physical media. Until that
>> characteristic is used, the texts in digital format will be an extension
>> of the paper universe. I propose this as a cornerstone for the
>> definition of digital poetics.
>> This also raises the question about *intentionality* and
>> *directionality* of computer-generated narrative. Unless
>> *directionality* is present, automatic text is not generating a
>> *communicative act*, but a simple mechanical response in the reader that
>> triggers brain activity (interpretation, language, suggestions, etc.)
>> without the phenomena that usually accompanies it (*communicative act*).
>> Juan B. Gutierrez
>> Research Fellow
>> CAVIIAR, Inc.
>> Simon Biggs wrote:
>>> Before I respond to a number of good points raised by others I wish to
>>> say that my original intent with this thread/provocation was to start
>>> a discussion in the Empyre “bar”, So I am glad that there have been
>>> responses of the ilk we have seen. That said, I am sorry the bar
>>> snacks are tasteless and that, in particular, the drinks have no kick
>>> whatsoever. I know this makes it hard to work up the appropriate
>>> spirit for this sort of discussion, but that’s the digital for you.
>>> Jim wrote:
>>> In 'new media', there's a sense of the importance of theory such as
>>> Manovich's work. But not much sense of the importance of the
>>> theory of
>>> computation to an understanding of the phenomenology of computing.
>>> Similarly, my concern was to ensure that the dual topics of poetics
>>> and the digital were at the heart of any discussion on what digital
>>> poetics (or ePoetry) might be. Many debates on this seem to forget
>>> these aspects of the area of practice, taking a narrow view of
>>> poetics, assuming that it means poetry, and generally ignoring the
>>> implications and consequences of computational theory. In doing this
>>> the effect is to elucidate a view of ePoetry that is little different
>>> to traditional practices that embrace concrete poetry, visual poetry,
>>> multimedia and the concept of the open work of art. Each of these
>>> forms of practice are, indeed, part of the genealogy of digital
>>> poetry, in large part because many key practitioners in the field come
>>> from these backgrounds, and to understand digital poetry it helps to
>>> be knowledgeable about them. However, none of these practices are
>>> digital poetry.
>>> I agree that poetics are extremely difficult to define but I would
>>> stand by my general definition – although I might reword it. The
>>> creative practice of association could also be written as a motile
>>> engagement with the interplay of dynamic elements. What I am seeking
>>> to do here is to separate poetics from human intent and authorship and
>>> regard it instead as a phenomena of things. In that sense Juan is
>>> right that this definition cannot be used to define poetry. That would
>>> be a tautology anyway. However, it is a way of seeking to situate
>>> poetics in relation to other things and remove the legacy of
>>> Romanticism, and the centrality of the human, from the discussion.
>>> The point about formal and informal (human) languages, raised by Juan,
>>> is also good. However, research by colleagues here in Informatics
>>> illustrates that whilst they might agree early computing emerges from
>>> formal logic they see current developments far from those early
>>> beginnings. They are concerned with complexity and fuzzy logic, their
>>> objective being to create affective computational models. These are
>>> not strict formal systems, although they are fundamentally linguistic.
>>> Indeed, the bulk of this research is being carried out by teams of
>>> computer scientists and linguists and involves artificial systems
>>> interpreting and responding to body language, facial expression, vocal
>>> tone, gesture and speech without building precise models of what is
>>> occurring but functioning as sets of dynamic contingencies and
>>> probabilities that may or may not require resolution prior to action.
>>> I consider what they are seeking to do as digital poetics – which is
>>> possibly why we find it easy to work together on artistic projects.
>>> Whilst this work is distant from Turing’s original concepts it
>>> continues to embody them at its core, taking in the work of Winograd
>>> and others along the way. One very specific concept that is being
>>> pursued by this research team explicitly employs Dawkins work on
>>> Memes, seeking to create self-perpetuating affective linguistic
>>> systems that are able to interact with one another as well as with
>>> their environment. I don’t think they are looking at Kant to inform
>>> their work – although perhaps they should? However, I would argue that
>>> computing need not, and is not, constrained by strict formal systems
>>> that allow no choices outside a given system. Artificial reasoning has
>>> developed far beyond its early limitations and in the process
>>> illustrated that human cognition is far more constrained than we
>>> previously imagined (and certainly far more so than Kant proposed).
>>> So, I dispute that there is much difference between human and
>>> non-human languages and would include in that not only the language of
>>> machines but also the languages of nature.
>>> Sally makes a good point when she asks whether I am conflating things
>>> that can be digitally described with being digital. Perhaps I am but I
>>> am employing an expanded definition of what the digital can be. My
>>> fingers are digital. My DNA is digital. My computer is based on
>>> digital systems. The language I write with here is inscribed, through
>>> various protocols, as digital. Where do we draw the line between what
>>> is digital and what is not? The big question, of course, is whether
>>> there are any spaces in-between. At that smallest of scales what is
>>> going on? Is it particles or waves, or both? Last week I was at an
>>> event where the Professor of Physics was having his portrait unveiled.
>>> That Professor was Peter Higgs, responsible for formulating the
>>> concept of the Higgs Boson (the particle they built the CERN collider
>>> to find). I doubt he knows the answer to that question yet. Perhaps it
>>> is neither. However, down to at least the scale of the Higgs Boson
>>> everything is discrete, even if we do not perceive it to be so. That
>>> looks sort of digital to me.
>>> Simon Biggs
>>> Research Professor
>>> edinburgh college of art
>>> s.biggs at eca.ac.uk
>>> simon at littlepig.org.uk
>>> AIM/Skype: simonbiggsuk
>>> Edinburgh College of Art (eca) is a charity registered in Scotland, number SC009201
>>> empyre forum
>>> empyre at lists.cofa.unsw.edu.au
>> empyre forum
>> empyre at lists.cofa.unsw.edu.au
> empyre forum
> empyre at lists.cofa.unsw.edu.au
More information about the empyre