I recently finished chapter seven of Expressive Processing, so I have yet to read anything about the “Sim City Effect” or the conclusion to Wardrin-Fruip’s argument. However, the discussions of Brutus really started me thinking about authoring systems. I was in a small group last class that specifically talked about whether or not computers or machines could potentially be authors, so I’ve been thinking about this topic the past couple of days. Last class, my response to the general question of computer programs being authors was that I thought “in terms of copyright, the program’s creator would be considered the author,” which continues to strike me as a strange response to the question. Maybe I’ve just been reading too much Foucault lately…
However, my other two group members’ ideas were much more fascinating to me. One was adamant that computers would never be able to truly embody creativity and have authorship of their outputs. The other group member was perfectly willing to accept that AI would one day be capable of creative authoring, but that the processes used to do that would no longer fall under the umbrella of computation.
In response to the first position, I’d like to think more about Brutus. First, I was astounded when I read the “Dave Striver loved the university” story because it seemed so literary (245). There was nuance in the syntax and drama in the structure. I tried to imagine a very complicated underlying structure that somehow was able to create a story from isolated or hierarchical fragments, potentially using a question-input system like Tale-Spin and Terminal Time. As I read on, I saw that the system was a more elaborate love letter generator. Then, I was struck that Wardrip-Fruin continued his discussion and analysis of Brutus as an authoring system. It was so clear to me, personally, that its creators were the authors of the stories it output since they authored both the sentence components and the system for selecting those components. It takes until the section “Author of Brutus‘s Stories” for Wardrip-Fruin to give a hint as to why he participated in this illusion/game for several pages, the reason being that he suspects that even the creators were drawn into the illusion (258). I think that is super fascinating, but I don’t have anything concrete to say about it. Tale-Spin and Minstrel seem more like proto-authoring systems, but their tendency to break down when exposed to larger sets of possibilities is so opposite from how human authors are supposed to write, that is they’re supposed to become better with each scrap of knowledge they get, and I don’t want to be too generous when thinking about their potential. Universe and Brutus for me are authored works with the potential for varied outcomes, and in this way I think of them as being similar to Terminal Time in actuality, even if that is not how they’re conceived or perceived by the public. I guess that there still is no computer system I’ve come across that I would consider as serving as an author. Perhaps we really will never see that happen. Universe, Brutus, and Terminal Time all seem like procedural translators of authored texts at most.
In response to the second position, that a creative, autonomous, authoring machine would cease to be computational, I’m stuck. Perhaps I am just resistant to that idea because I’ve spent all this time and energy into learning some code, expanding my knowledge of computation, and understanding computers as procedural. I think I prefer to think about this as a blending of computation and creativity. From my limited understanding of this discipline, a good example of the two coming together is mathematics. Math we are taught as children is purely computational stuff (2+2=4), and later we are taught that computational math has some intersection with usefulness (namely money calculations and statistical biology, from my own experience). Personally, this is where I am with code; I see it as a tool that can let me arrive at a solution to a pre-conceived problem.
http://www.math.ttu.edu/FacultyStaff/research.shtml
However, higher-level math and mathematical research is super creative.At a certain point, you stop using real numbers at all and work exclusively with concepts. Maybe this is why the idea that you need to be great at math to be really good at code, something we all came across in the Paul Ford piece at least, is so pervasive. Thinking about code as writing is easy, but actually writing code may have more in common with the way mathematicians creatively think of problem-solving than it does with how fiction writers creatively think of problem-solving. Perhaps, then, it would indeed be possible to have a computer that authored ideas, if not linguistically sound stories, and the solution to this problem lies in the path math walks between computation and creativity.
I possibly veered too far from the readings this week for what this blog post assignment intended. If so, forgive me. I hope my ideas were worthwhile enough to excuse my indiscretion.
Hi Smalltalk,
You make some great points, and I think that ultimately, the question about whether computational processes can be ‘authors’ depends on how one attributes authorship. Is self-awareness necessary? Intention? Free will? These questions are rapidly getting away from the text, but I think there’s a connection to Wardrip-Fruin’s discussion of the ways that different softwares attempt (or don’t) to hide their processes. As he explained in the chapter about the SimCity effect, some programs (e.g., chatbots and Eliza) are designed so that the surface belies the simplicity of the code. On the other hand, simulation games require the user to understand and manipulate the game’s processes in order to succeed. It seems that in the latter type of game, the user has a stronger claim to authorship over the program’s outputs, because s/he is actively manipulating the program’s processes, albeit via the surface rather than by directly editing the code. In software that resists this kind of manipulation (e.g., Eliza, or Mac OS, or the software that runs my coffee machine), the original programmer has the greater claim to authorship because there is less room for user choice / unexpected outcomes.
One aspect of this question that I’m still having wrapping my mind around is the level awareness a generative “being” (organic or mechanical) needs to have regarding its product in order to claim authorship of it. My group ran into a similar concern last week while discussing Maher’s “Artificial Rhetorical Agents and the Computing of Phronesis” article. Maher argues that an artificial rhetorical agent would have to be “capable not only of making moral judgments but also explaining through the persuasive use of natural language – that is, rhetoric – the reasoning behind those judgments” (4). I kept hitting a mental wall when it came to the ARA taking it a step further and being able to generate thoughts and opinions about this process. After all, wouldn’t being able to articulate the reasoning behind its judgments amount to no more than regurgitating the steps built into its program? For instance, if the machine in question was a Google car and it had to make the choice between hitting a baby and hitting a dog, it would only save the baby because it was programmed to do so. Thus, it would explain this reasoning in these terms. It would have no concept about what is a baby and what is a dog and why one’s life should be privileged over the other. Anyway, I’m a bit off topic, but this is the same trouble I run into when it comes to computers as authors. At what point can it claim that it made the conscious decision to order text (or music, or gameplay, etc.) in a certain way to evoke a certain reaction? If it doesn’t have to do this (after all, if computers are not humans, should they be held to the same standards?), at what point do we grant the computer a sense of authorship over the computer’s programmer? Is it when it surprises both the programmer and the audience in what it produces? Is this still, technically, a collaborative endeavor? Or would the programmer have to relinquish ownership in the same way that human authors’ parents don’t claim any ownership over what their children produce? Will I ever stop asking questions and just phrase one of these thoughts as a statement?