Could you teach literary/film criticism to a computer?

Question in the thread title: could you teach literary or film criticism to a computer?

I'm not talking about any currently existing machine; consider instead whether a computer that wasn't subjected to our current technological limitations could perform the exercise. Another way of wording the question could be by asking if a literary or film criticism could be reduced to an algorithmic operation (albeit one of great complexity).

In assuming that most of you will be refuting the idea, I'll be taking the position that yes, a very complex and powerful computer could be capable of such an activity, as all such criticisms take the form of an evaluation of the text, x, based on a set of criteria, y.

So if I were to evaluate, for example, gothic imagery in "The Fall of the House of Usher," I would simply identify a set of criteria that constitutes "gothic imagery" and program the computer to identify such criteria and produce an evaluation of the text accordingly. Note two things here: (i) the evaluation doesn't necessarily have to be one of "quality" ie. how well the text does this, it could be how similar the text is to this, how (not how well) it fits into this set of literature etc. and (ii) the computer would have the task of evaluating a specific work, and would not necessarily have to be capable of examining every work with the same programming.

I'm not particularly knowledgeable when it comes to textual analysis or computer programming, so it would be cool if people familiar with either activity could weigh in with their thoughts.

 

I believe they had a computer figure out an algorithm for successful pop songs, so I suppose it's possible.

 

You can do that now, if you just put in terms related to being gothic and have the computer find them.

The more in-depth version, predicting the emotional effect of a work, is interesting. A great story has to be plausible, which means the computer needs to be able to model reality and human behavior and patterns of speech. In other words...it could probably also pass a Turing Test. Which means that if it's possible, it won't happen for centuries if at all.

 

lol I don't think it would even be that difficult. most film and literary criticism is simply based on drawing connections between new works and existing ones. artificial intelligence would be very adept at doing that.

hell there have been cases where programs have randomly generated ACADEMIC papers that were then submitted to real editors for publication and were accepted because the refs couldn't tell it was just gibberish generated by taking parts from random sample of papers.

 

Meaningful criticism would be hard, but Turing test criticism? Sure. Criticism is stylized enough that it would be easy to fake.

 

Possibly, if you could give it emotions. I think emotions are necessary to judge most art. Especially film. The rush, suspense, feeling of loss and gain, fear, comedy, irony, etc.

Art is about relating to people. It documents the human experience. It invokes something inside us. A connection is made. As Joseph Campbell said:

Then you get into the collective unconscious, archetypes, etc. which are only powerful to humans because of our evolution and cultural history. Not sure they would resonate with any other kind of creature, organic or artificial. So because our art is about the human experience, I'm not even sure an alien could relate to it or judge it.

 

I think there are two questions that are being asked and answered in this thread.

If the question is whether a program can produce a written or oral piece of criticism that can intelligibly understood by a reader or listener then I think the answer is definitely yes.

But what you are answering seem to be a much more philosophy of the mind question about whether a machine can ever understand literature which is somewhat different and more ambiguous.

 

4/4 drumbeat and pop riffs on a guitar.
It's really simple, actually.

 

Only an android. Not the software.

 

Did Cmdr. Data's poetry teach us nothing? Not even an android could do it.

 

Data is a so so android, a good one could do it.

 

Sort of, it came out to around a 70% success rate if I remember correctly.

When you think about it, Humans are fairly logical creatures. It is a known fact humans enjoy triangles because of a predisposition to the number 3. We also naturally appreciate symmetry and the human form (ie. look up the 'Uncanny Valley' in relation to robotics). Film is just another medium, I'd be almost certain we will eventually create algorithms to analyze the art behind it.

 

Computers will eventually become self-aware when we learn to replicate the circuitry of the human brain. What TS is asking is trivial compared to what computers of the future will be able to do.

 

Yes and no.

Yes, given unlimited technological resources, you could build a "computer" that could learn literary criticism.

The thing is criticism is also dependent upon emotions and experiences of people. So you'd have to be able to program emotions and have the damn thing also experience daily events. Your opinion of a movie may very well be influenced by how recently you've taken a dump.

So it's not so much building a computer, but really building a (synthetic?) person. Yeah that's possible given unlimited technology, but is that really the question?

 

You forgot the cliffs, bud. Here you go:

http://www.cliffsnotes.com/study_gu.../summary-analysis/fall-of-house-of-usher.html


I think themes are hard. Imagery is easy.... What are we after? An analysis of the plot, or just fish out every castle, omen, walking statue and howling ghoul?

 

The question is not merely the former, but not necessarily the latter either. In terms of producing a coherent piece of analysis, I think a computer could manage that fairly easily.

I"m talking graduate level writing here though, involving creativity (making original connections), extensive contextual evaluation etc. The computer would not necessarily have to understand this process as much as simulate the understanding, similar to how the person in Searle's Chinese Room was able to use the language without truly understanding it.

Again, a good way of thinking about it is to ask if literary criticism can be broken down into an algorithmic process. I'd like to see someone describe the part of the process that couldn't be.

 

Lol, it's such a short story I didn't realize those even existed.

We're after any form of analysis that can't be simulated. I'm putting forth the argument that every type of criticism can be broken down into an evaluation of the text, x, based on a set of criteria, y.

Furthermore I'd say that any evaluation that can't be described as above is incomprehensible. That's not to say that the process has to be simple on the surface, but that with enough research over time the criteria will eventually be established, and then (why not?) replicated/simulated.

 

Or some sort of model that successfully reflects human nature/emotion. It would never be completely accurate, but I suppose it would do okay.

 

Again, like the subject of the Chinese Room, the computer doesn't need to be able to feel emotion so much as identify what the requirements for a specific emotion would be. It's enough to say "the following pieces of text would produce the emotion of sadness in most normally constituted humans".

Of course that raises the objection that the computer could only exist through the compounding of all human knowledge/experience throughout history, which while interesting doesn't reject the OP conclusively.

 

Intuition too. Which is a product of evolution and biochemistry. Could one teach a computer intuition? How could you make that an algorithm? It's a feeling, not a logical thought. It's like asking if you could give a computer a subconscious.