Dec. 31

Upgrading to Philosophy 2.0

There was no theorizing about ghosts in the machine at an annual meeting of philosophers last Friday. Instead, they embraced technology’s implications for their field, both within the classroom and beyond.

Looking for a job?

See all 312 new postings

Browse all job listings:

Faculty: 2,352
Administrative: 2,117
Executive: 155
FEATURED EMPLOYERS

Two presentations outlined how computer software could change the way philosophy is both taught and disseminated. One professor discussed how artificial intelligence can help to improve individualized instruction, while another laid out a radical framework for online publication that would leave most of today’s academic press apparatus in the dust.

This being the annual meeting of the American Philosophical Association’s Eastern Division, the largest of the regional subgroups , none of the scholars’ suggestions went unchallenged. The Baltimore session, arranged by the APA Committee on Philosophy and Computers, led to a lively question-and-answer period among the 20 or so audience members who likewise saw technology as a vital component in the future of philosophy as an academic discipline.

Still, as the session’s title implied, the participants were also interested in exploring “the Ethics of Emerging Technologies,” not just how to implement them.

Marvin Croy, a professor at the University of North Carolina at Charlotte (and a former chair of the APA committee and current vice president of the International Association for Computing and Philosophy), described his search for a high-quality but cost-effective way of promoting individualized learning through artificial intelligence. He believes that making such programs worth the effort is more difficult than it would initially seem because people tend to play down maintenance costs and overestimate the extent to which automation can lead to greater accessibility of teaching content.

The reason for those misconceptions, Croy argued, is that adaptive learning techniques require AI, and good AI algorithms require long-term empirical research into how students learn and which methods predict classroom success. Moreover, he said, if a computer program that employs AI increases the range of students being taught, any economy of scale would be counterbalanced by the greater diversity of learning approaches reached — and that would require further development into more sophisticated processes to encompass them (and more money).

Bypassing that vicious cycle requires some brains, and not just the human kind. The problem becomes: How can a program learn how an individual student thinks, and use that insight to offer constructive suggestions as he or she works online?

One of Croy’s attempts to solve that problem involves a system designed to provide intelligent help to students constructing deductive proofs. As they graphically map out the steps from a given initial proposition to the provided end point, the software ideally provides helpful suggestions to students who can be working both forward and backward at the same time.

In looking for an algorithm that can offer hints “in a way that doesn’t cost us an arm and a leg,” Croy noted, the software employs a mathematical model called a Markov decision process that can map students’ steps toward the solution and “learn” the chosen path as they work. Such proofs can be solved in varying sequences, so the possibilities are numerous.

“They do stuff that I wouldn’t have expected them to do,” Croy said of the students. By anticipating the logical direction of the students’ reasoning, the program can ideally guide them along the way.

To see if such techniques are empirically useful, Croy also tested to see if he could predict students’ performance in his class early on, based on results from a computerized test of “justified thought” — for example, choosing from a multiple-choice list whether a given logical sequence was an example of modus ponens, modus tollens or neither. By dividing one class of 50 into two groups, one whose grades were below 65 percent and those with 65 or higher, Croy found that the test predicted their performance “fairly well.”

This being a meeting of philosophers, he touched on a few of the ethical implications of his work, such as the potential of conflicting roles as both a teacher and a researcher within the same classroom. “It does put you in a very strange position,” he admitted, since students could be both pupils and subjects at the same time. One clear solution, he said, was to seek informed consent. At the same time, Croy raised the question of whether technology should seek to replace or supplement student-teacher interactions.

In his own experience, he said, “the class is a lot better today than it used to be a year ago.”

A ‘Radical’ Rethinking of Scholarly Publishing

Harriet E. Baber of the University of San Diego thinks scholars should try to make their work as accessible as possible, forget about the financial rewards of publishing and find alternative ways to referee each other’s work. In short, they should ditch the current system of paper-based academic journals that persists, she said, by “creating scarcity,” “screening” valuable work and providing scholars with entries in their CVs.

“Now why would it be a bad thing if people didn’t pay for the information that we produce?” she asked, going over the traditional justifications for the current order — an incentive-based rationale she dubbed a “right wing, free marketeer, Republican argument.”

Instead, she argued, scholars (and in particular, philosophers) should accept that much of their work has little market value ("we’re lucky if we could give away this stuff for free") and embrace the intrinsic rewards of the work itself. After all, she said, they’re salaried, and “we don’t need incentives external [to] what we do.”

That doesn’t include only journal articles, she said; class notes fit into the paradigm just as easily. “I want any prospective student to see this and I want all the world to see” classroom materials, she added.

Responding to questions from the audience, she noted that journals’ current function of refereeing content wouldn’t get lost, since the “middlemen” merely provide a venue for peer review, which would still happen within her model.

“What’s going to happen pragmatically is the paper journals will morph into online journals,” she said.

Part of the purpose of holding the session, she implied, was to nudge the APA into playing a greater role in any such transition: “I’m hoping that the APA will organize things a little better.”

— Andy Guess

Want it on paper? Print this page.
Know someone who’d be interested? Forward this story.
Want to stay informed? Sign up for free daily news e-mail.

Advertisement

Comments