The issue of separation

Latour talks throughout the book how we are not Modern because can’t keep categories separate. We cross politics with science with ethics and others all the time. And so, according to Latour, we have never been modern because of our use of “hybrids.”

But at one point in the book, he mentions how there are three categories we do strive to keep apart (or the postmodernist do, at least): nature, society, and discourse. Latour calls this a terrifying image:

“…a nature and a technology that are absolutely sleek; a society made up solely of false consciousness, simulacra and illusions; a discourse consisting only in meaning effects detached from everything; and this whole world of appearances keeps afloat other disconnected elements of networks that can be combined haphazardly by collage from all places and all times. Enough, indeed, to make one contemplate jumping off a cliff” (64-65).

It made me think of something else our society tries to keep separate and that also creates a terrifying image: race.

The US has always upheld the idea of the melting pot, that we are a mix of cultures and races, that anyone can come to the US from wherever and become “American.” And yet, we go completely against this image of a melting pot and label everyone separately. We don’t view each other as “American,” instead we see people as “African American,” or “Asian American,” or “Irish American.”

It leads to incidents such as what the girl in the video has experienced. We don’t view those of different race as the same as everyone else. You might live in the America, and you might be an American, but we will view you as “Asian” and treat you accordingly.

It seems to me that Latour is arguing that we can’t achieve the separation we hoped to achieve as modernists, though we try. I would argue he views complete separation of everything as impossible and destructive. And I think the racial example is a good illustration of his fears.

Agency and Corrupt Schools

Miller's discussion of automation and agency is a good argument that is sure to be debated by educators and administrators for years to come. The idea of an automated grading system has clear pros and cons that continue to be bantered about among the powers-that-be within primarily affluent school districts that can afford to play with the technology.

Feasibility is obviously one of the biggest issues facing the development and utilization of such technology. Those questions are for educators to figure out before measures are taken to potentially ease their after-class workload. My mother is a high school math teacher that works harder than anyone I've ever seen grading tests and assignments and going the extra mile for her students at all hours of the day and night. I doubt that the increased free time she would receive would outweigh the amount of ownership she would lose within her own classroom and over her own assignments by allowing a machine to spit out grades that she would then feel the need to verify.

To me, the most important portion of the debate is the ethical dilemma surrounding the decision to automatically grade papers, tests, and even speeches. Does putting the agency into the hands of a machine detract from the ability for teachers to teach and students to learn in the classroom? Does it create a standard that is so universal that teachers begin to teach students to perform in ways that enhance scores within the computer system? Does this potential change in philosophy stifle the creativity of bright minds, both the ones doing the teaching and the ones doing the learning?

With so many questions still to consider, there is one case in which having a computerized grading system is a far better option than what is currently happening: academic fraud. A district in Atlanta has recently come under fire after allegations surfaced that suggested that teachers would have "grade-changing parties" to doctor standardized test results and raise and lower the marks of students. The level of corruption extended to the very top of the district, leading to nationwide exposure and a call for a drastic overhaul of the entire structure of the school district.

It is clear that in such circumstances, the agency must not rest in the hands of corrupt teachers. Instead, something must be done to ensure fairness and accuracy...something like creating an automated grading machine. Often times teachers are unfairly lumped together with bad ones to create the image of a profession that simply cannot get out of its own way. Reforms are enacted that shackle the abilities of quality workers to teach so that bad ones who fall through the cracks can be held in check. Therefore, whatever system is devised needs to take into account the individual being given agency or being denied agency before blanketing the entire slew of the nation's teachers with a negative stigma.

This could be done in a cost-efficient way to benefit the students and residents of Atlanta; the problem is that in many places, the system is not broken and with so many questions, a solid system could be replaced with a creaky one. I firmly believe that teachers like my mom should be given the freedom to conduct their classrooms in whatever way they deem appropriate, but teachers who drag my mom's profession through the mud should be barred from inflicting harm on children by messing with a universal standard of judgment.

Object Lesson IV: The One About Robots

In her article “What Can Automation Tell Us About Agency” Carolyn Miller describes the problem facing contemporary notions of “agency,” calling attention to its decentralization from traditional human subjects. She goes into a detailed discussion on the “subject of computerized assessment automations”, analyzing the responses of both writing and speech instructors to a survey regarding the perceived viability of such grade-bots in the classroom (their response is generally incredulous), and determines that, if these automata can function as an audience, then agency is neither possessed by the speaker nor their audience but exists as a sort of “kinetic energy” which “is a property of the rhetorical event or performance itself.” Marilyn Cooper, however, opposes this assessment, arguing that “deeds are always done by someone, and replacing the doer of the action, the agent, with an amorphous force like kinetic energy leaves us with no basis for assigning responsibility for actions."

Cooper’s idea of “agency as a matter of action” emphasizes free will, or the embodied intention to enact change, and it is on this theme that I made a connection to Richard Powers’s novel Galatea 2.2. In it, a writer and a computer scientist work together in designing an artificial intelligence; their goal is to have it pass the Turing Test by answering a graduate-level question on the subject of literature better than a human student. They dub their A.I. “Helen” and attempt to teach her everything that will help her succeed in the test—this, they come to realize, means Helen needs to know almost everything about everything. In order for her to comprehend Shakespeare, for example, she requires not only the literary context of the story but the historical context in which it was written, and this context is in turn informed by a seemingly endless string of connections that encompasses human experience.

Spoiler alert: …………….Helen fails. Although she develops into an extremely sophisticated A.I., she realizes that everything she knows is being fed to her by the writer and she is incapable of real “experience.” She asks her creators to “see everything” for her before shutting herself down, an apparent act of free will. Drawing from Cooper’s terminology, Helen’s decision was the result of her inability to construct her own disposition, or ethos. She lacked the crucial element which enables, and indeed motivates, human agents to participate in rhetoric: the ability to affect change. As Freeman notes, humans, as “intentional beings”, have the capacity “to construct and pursue their individual goals within the contexts of their societies.” This, I believe, is the essence of “agency” and the reason why Helen was doomed to despair her role as a thinking machine, one possessing extensive databanks of information but lacking that essential capacity to see her actions as causal and therefore valuable—she must instead leave her creators with a desperate and impossible plea, “see everything,” a perfect linguistic distillation of thought without direction or sense of purpose.

These readings have forced me to question, reject, reaccept, reject, doubt and reconstruct any notions of agency I had before; what I find myself walking away with is not “kinetic energy,” however, but a sort of human clairvoyance that projects into the future while drawing from the past.

Little Did He Know. . . .

In "Rhetorical Agency as Emergent and Enacted," Marilyn M. Cooper argues that agency is "based in individuals' lived knowledge that their actions are their own" (421). The agents, who may or may not be aware of their own intentions, bring about changes through these actions. For Cooper, the crux of the agency debate seems to rest in the responsibility that agency demands of the agent. According to the article, all "concrete others" already have agency; instead of arguing for the empowerment of these agents, we must argue for the responsibility of the already empowered agent (443).

The discussion of agency in this week's readings reminds me of one of my favorite movies, Stranger Than Fiction. In the film, the main character, Harold Crick, begins hearing a voice narrating his mundane and unremarkable life as an IRS auditor. While trying to discover the origin of the voice, Harold meets a free-spirited baker named Ana who inspires him to live the life he always wanted. Eventually, Harold discovers that the mysterious voice belongs to reclusive author Karen Eiffel, whose works always result in the death of the main character.

Stranger Than Fiction: Harold and Karen Meet

Harold finds Eiffel and attempts to convince her not to kill him, but after reading the ending she has already written (but not typed), he decides that his death is necessary to save the life of a young boy. Eiffel ultimately decides not to kill Harold, however, because if a man knows he's going to die, but "dies willingly, knowing he could stop it, then isn't that the type of man you want to keep alive?"

The film's premise obviously opens it up to a whole host of implications that I don't have time to address in this blog post. The two I want to focus on are the idea of rhetorical responsibility and free will.

In a larger-than-life way, Karen Eiffel's words have an extreme form of rhetorical agency. At the beginning of the film, Karen is unconscious of the inescapability of her agency; after meeting Harold, however, she begins to realize the effects of her words. This realization instills her with a sense of responsibility and leads her to question how many other people she has killed as a result of her novels. After meeting Harold, Eiffel is torn between preserving the perfectly tragic ending she has already conceived for the novel and saving Harold's life.

Despite Eiffel's control over Harold's surroundings and knowledge of his thoughts, Harold does exude his own sense of agency. Harold's actions are not strictly dictated by the rhetorical situation or his past experiences, but rather, they "emerge from the ongoing process of his becoming the person he is, as he respond[s] to the world he encounter[s]" (426). Harold becomes more open to opportunities for living a rewarding life he makes choices based on the "larger narrative" of his life. Even after meeting Eiffel, Harold has the agency to decide his own fate and feels responsibility for the way his actions will affect others. He acts "intentionally and voluntarily," yet he is not the sole cause of what happens (439).

Harold's situation also relates to Cooper's discussion of free will. Harold does have free will in that he exercises the power to act. His actions, while "free," are never his own; they are influenced by the personal, historical, and societal (and in Harold's case, literary) context in which he performs them. Instead, Harold's freedom derives from his self-awareness of "having an experience while in the midst of it--of tacitly recognizing the sense of the unfolding experience without halting it or distancing [himself] from it" (440). In thinking about free will and agency from this perspective, Harold actually seems to have more agency after meeting Karen Eiffel and becoming aware of her involvement in his life. Harold recognizes her involvement in the situation and, rather than distancing himself from it, makes the decision pursue his own course of action within the experience. By doing so, he not only retains his own agency, but also makes himself responsible for the decisions that result from that agency.

Agency and a Trivia-Answering Computer

While reading Carolyn Miller’s essay, “What Can Automation Tell Us About Agency?”, I was reminded of  the Jeopardy tournament this past spring in which an IBM supercomputer made up 90 servers, named Watson, competed against two former human  Jeopardy champions. By the end of the tournament, Watson had completely dominated the humans, besting the second place contestant by over $50,000. As impressive as Watson performed, it was by no means omniscient; during Final Jeopardy of the first match, Watson did this: 

Miller notes that "if agency is a kinetic energy, it must be a property of the rhetorical event or performance itself" (147). Therefore, would Miller agree that Watson, the supercomputer, can only possess agency if it is situated in a game show environment, where it influences the responses of other contestants and the event as a whole? I would argue yes, because the computer was designed to interact in that specific environment. The agency of the computer only functions as "kinetic energy", not something possessed, but rather performed and acted out. If the computer was the only contestant on the show, it could be said that Watson holds one component of agency - the capacity to act - but ability to "instigate change (effect)" would be largely lost (146). The performance dimension of agency that Miller first focuses on can also relate to how Watson influenced the agency of the two human contestants. If Watson had answered every question correctly before the two other contestants - which would mean they were never allowed a chance to answer, or even talk - then Watson would be limiting their participation, and in effect, agency.

Response - What Can Automation Tell us About Agency?

Within the first five minutes of reading this essay, I began having flashbacks of standardized tests and the wave of anxiety they always brought with them.  I specifically remember sitting in the GMAT and seeing a girl break down into tears, throw her materials in the floor, and walk out the door screaming about how wrong it was to have a computer picking different questions for different people.  I felt bad for her, but I was also relieved to know at least someone was as stressed out as I was.

I think the SAT, GRE, and GMAT are all very accurate indicators of ability.  Evaluating a potential student should obviously be a combination of testing, previous work, and references, but the idea of a grading system for speaking is very interesting.  I think that Miller does a great job of explaining both the benefits and the limitations of using this hypothetical system.  

AutoSpeech-Easy has the potential to be very beneficial to education if used properly, but could also create a lot of problems if it is not.  I think it would be a great supplemental tool for students as they prepare for a public speaking class, or before they are giving a presentation.  I do think it would be an inaccurate measurement for the effectiveness of a speech just because the variables that it is eliminating.  As Miller said, "To grade via computer takes away one of the hardest parts of public speaking; the public part."  The entire reason a public speaking course exists is to help people alleviate their concerns of addressing a topic in front of people.  During the course I took in undergrad, I honestly can't even remember what the speeches were about.  I just remember that by the end of the semester, I was much more comfortable speaking in front of people.

It also clouds many of the ideas presented by Herrick in "An Overview of Rhetoric."  Herrick talks about the significance of the arrangement of a speech, the planned delivery, and adjusting the speech according to the feedback of the audience.  This cannot be done when speaking to a computer that is analyzing the content of the speech, but not necessarily the placement.  

I think that using some sort of automated analysis software for speech could potentially be a great tool.  However, oral communication is too complex to use it as the sole measurement for the effectiveness of a speech.  The message of a speech and how it is received will always be determined by the people receiving it.

Simulated Audiences and Rhetorical Agency: How Machine Stand-Ins Help Us Feel Alone Together

In her piece “What Can Automation Tell Us About Agency,” Miller opens up by drawing attention to the computer systems of automated assessment as a new product of interest to rhetoricians. She relates the dilemma of agency to technology studies by discussing the challenge of the distinction between action and motion in agency through human and non-human, or machines and natural forces (143). She explains “by positing a machine as audience, automated assessment systems for both writing and speaking denaturalize rhetorical action, challenging and uncovering our intuitions about its necessary conditions” (140). By this, Miller suggests that rhetorical agency is at stake and attempts to illustrate how our resistance to automated assessment can inform the current debate on the nature of rhetorical agency (140).

A lot of this discussion, and particularly the examples Miller illustrates, ties in directly with the book Alone Together by Sherry Turkle. In the first part of her book, Turkle discusses the “robotic moment” and sociable robots and how they are creating new unsettling relationships and instabilities in how we understand privacy, community, intimacy, and solitude. She poses a troubling observation that robots are filling gaps in the society and we hope to use them as a solution to our own imperfections, as an easy substitute for the difficulty of dealing with others. In her examination of the relationship between humans and robots, Turkle suggests we don’t mind giving human qualities to inanimate objects, much like Miller’s comments on our eagerness to attribute agency to technologies.

Miller makes the connection to Turkle in explaining that “automated writing assessment is an applied branch of natural language processing, which in turn is a branch of artificial intelligence” (138), like robots. She continues by discussing our resistance to automation as rooted in a commitment to agency, since they remove agency from the audience and deal with the confrontation between action and motion (141). She addresses three dimensions of rhetoric as missing in this: performance, audience, and interaction. In relation to performance, Miller comments that speaking requires a living presence (145) and thus prompts agency as a kinetic energy (147).  Instead of serving that purpose, the computer instead serves as a mediator and a surrogate. This is because it does not provide a relationship between two entities that will attribute agency to each other, what she says produces the kinetic energy (149). Therefore, interaction is fundamental for agency because it creates the kinetic energy of performance and puts it to rhetorical use (150). Interaction is also an issue that Turkle addresses. She points out that computers and robots no longer ask us to think with them, but ask us to feel for and with them as they become more sociable, affective, and relational (39). She argues that their interactivity prompts our minds to start projecting consciousness and we compensate by filling the gaps. A machine only needs to act clever and people will play along. Additionally, she points out that this simulation is enough to provoke empathetic urges. This ties in directly to Miller’s discussion of ethopoeia.

Miller mentions ethopoeia and society’s tendency to have a low threshold for it to illustrate how willing we are to attribute character to an interlocutor. She comments “we go out of our way to construct a human relationship” (151). She even makes a direct connection to the Eliza effect, or the attribution of intelligence and sympathy to a computer program, which is a reference Turkle uses throughout Alone Together to explain her ideas (151). Turkle argues that the new relationships we have with robots create a loop, drawing us into the complicities that make it possible; but we are playing along, willing to defer to what the robots are able to do (100).

Miller’s article illustrates indirectly how we are motivated by technologies to live a more rushed lifestyle and expect almost instantaneous results for everything. This even includes feedback in grading assessments. This ties in directly to Turkle’s point that as technologies advance, we expect more from them and less from each other. Turkle comments that as society gradually seems to be growing more “machine ready,” it seems less shocking to put robots in places where people used to be (146). So we must ask, what are we sacrificing when we look to robots as substitutions or surrogates? Turkle says “As we learn to get the ‘most’ out of robots, we may lower our expectations of all relationships, including those with people. In the process, we betray ourselves” (125). Both authors discussion of machines opens up daunting philosophical, psychological, and moral conversations.

Overall, Miller points out that “automated assessment technologies could take advantage of our eagerness to attribute agency” (152), prompting the dilemma of where to draw the line between the human and non-human, symbolic and material. Even further, the attribution of agency concerns matters of moral judgment, human decency, respect, and acknowledgement. Therefore she asks whether or not we owe such acknowledgements and agency-granting attribution to automated assessment systems (153). Similarly, in her book, Turkle not only describes how the boundaries between people and things are shifting, but makes the reader question which of these boundaries are worth maintaining.

Turkle, S. (2011). Alone together: Why we expect more from technology and less from each other. New York: Basic Books.