Calling things “an art, not a science” has always been a pet peeve of mine. And I’ve heard people say things like, “there’s no best way to do it’. In particular, I’m taking a Responsive CSS course on Udacity and the guy said these things (if you listen closely, you could hear me cringe).
And then there’s the idea that art is like inherently intuitive, whereas science isn’t. I want to focus on the “art is inherently intuitive and not about breaking things down into components like science” part. My thought is that these people who say this are confusing their map for the territory. They may not know how to deduce what the perfect painting would look like, but that doesn’t mean that it’s not possible.
I know that there are different versions of these beliefs, and that I may be misunderstanding them. If so, please correct me. Anyway, what do you guys think?
To me it’s shorthand for “It took years of practice to get good and I can’t explain the process I use. The problem appears too complex for a scientific approach to be worth the effort. I certainly can’t think of any polynomial-time algorithm and my experience has led me to doubt that one exists.”
It’s a caution for beginners. The expert is basically saying that getting good at it will take a lot of work and finding a good systematic approach is likely to be a dead end.
They may not know how to deduce what the perfect painting would look like, but that doesn’t mean that it’s not possible.
One can certainly brute-force the issue by creating every possible painting and comparing them. Whether this can be done efficiently is another matter altogether.
One charitable interpretation is “it’s something you learn by doing, not something you learn by reading”.
“Art” has a bit of a double meaning, there’s the “something that’s pretty/pleasing/aesthetic/original/creative”, but there’s also the “craft” meaning, as in “the art of XXX”.
I want to focus on the “art is inherently intuitive and not about breaking things down into components like science” part. My thought is that these people who say this are confusing their map for the territory. They may not know how to deduce what the perfect painting would look like, but that doesn’t mean that it’s not possible.
Two reactions to this:
1) If someone says something can’t be broken into component parts, a more charitable reading is that they think that trying to do so is a waste of time and less likely to bring progress than just a lot of practice. Even if it’s possible in theory, that doesn’t mean it’s actually a good idea, so warning people against it can be totally reasonable, and isn’t “confusing their map for the territory”.
2) HOWEVER, in the case of art, most forms of art I can think of—drawing, painting, storytelling, animations, etc. - most definitely CAN be broken into component pieces, and often those component pieces can be broken into component pieces too, etc. - just check out the right section in any library.
You can’t learn to draw by reading a book, but a good book on drawing can tell you what individual skills you should practice, and how to do so.
Art is about relying on intuitive pattern matching and not following strict rules.
Computers can follow strict rules that you program into them but are bad at creative pattern matching.
A correct breaking down may very well result in 10,000 rules with complex interactions between them.
The human brain has a lot more than 10,000 neurons that are active during a particular decisions process.
In LW jargon I’d phrase it as: “It’s a system one thing, not a system two thing.” I think this is what most people do mean when they use “it’s an art, not a science.” When something is considered an art and not a science it’s something that can’t be done well by “just” following a set of instructions. Keep in mind that the popular view on science (when it’s positive) is seen as strictly adhering to the scientific method (form hypothesis-->test hypothesis-->adjust hypothesis) and that this is something that anyone can do. The difficulty of each individual step is ignored or seen as something that can be learned without much trouble. The pop-culture view on art is “mysterious process.”
And you’re probably right that you can distill (some) parts of art into “rules” about good art. I recently heard a radio interview with a professional photographer and he could explain why each photo was good or bad by adhering to a set of simple principles and he could explain those principles. I do think, however, that part of what makes him a great photographer is that those simple principles have become part of his system one thinking.
I’m a bit skeptical that this is true. I sense that the majority of people don’t actually believe that art is reducible.
Yes, you might very well be right. What I meant to say is that I think “system one, not system two” is the general sort of idea that people want to convey, not that it was the exact same thing.
1) Things fall downwards, and anyone who disagrees is objectively wrong.
2) Mozart is better than Iron Maiden, and anyone who disagrees is objectively wrong.
I think there are problems with the second statement, and not just that people will stop inviting you to parties. You’re trying to impose your opinions on other people.
There are approaches you could take to determine which art is objectively better. For designing a website, the answer is which gets more hits, or sells more products. But as far as art goes… you could take a democratic vote over which artist is better. You could, in a thought experiment, extend the vote to a sum over the whole of mindspace. You could weight by IQ.
But I don’t want my tastes in art to be determined by a democratic vote of what everyone else thinks. At the end of the day, the people who say “there’s no best way to do it” are not able to work out what the most popular website ever would look like, and saying that the currently most popular site is objectively best would stifle innovation.
They’re not trying to say philosophically universally correct points. They’re not theologians trying to determine whether God could create the perfect painting. They’re trying to teach humans.
But as far as art goes… you could take a democratic vote over which artist is better.
I think the most direct way to do it would be to measure people’s brain activity. I’d be skeptical of self reporting. I think people would be swayed a lot by what they think they should like. I’m pretty sure I’ve read some research on this but can’t recall what it was :/
But I don’t want my tastes in art to be determined by a democratic vote of what everyone else thinks.
It wouldn’t be. In this case “best way” would be defined as something along the lines of maximizing total happiness.
Measuring brain activity is certainly possible but brings all sorts of challenges, rendering this kind of approach far less direct than it may seem on the surface.
However your thoughts about expectations are very much supported by a fMRI study (Kirk et al. Neuroimage 2009, direct link to PDF ) manipulating participants’ expectations about artistic images with the following instructions:
There are two types of paintings. 50% of the paintings have been borrowed from the Louisiana Museum of Modern Art (Copenhagen, Denmark). The other 50% have been generated by the experimenter by means of a computer program (Photoshop)
Paintings were randomly assigned to these two conditions for each participant, and ratings of aesthetic quality were obtained for each painting as well. Instructions modulated brain activity: the same artistic images evoke different responses in prefrontal/orbitofrontal cortex depending on participants’ expectations, and the same images received higher aesthetic ratings if the participant believed they came from a prestigious museum.
The authors also correlated brain activity with the aesthetic ratings given to each image (collapsed over context). No regions were reliably modulated by aesthetic ratings, which should have been the case if there are some regions that are responsive to “artistic quality” detached from context.
Of course, plenty of potential problems with this study, perhaps the most salient to me being the use of entirely abstract artistic images and artistically naive participants. But it does highlight some of the challenges related to revealing information about aesthetics from brain activity.
It wouldn’t be. In this case “best way” would be defined as something along the lines of maximizing total happiness.
Still, rather than having one universal best art, the best approach is to have different peices of art to apply to different people. There may be a universal best piece of art for one person, but I imagine they would still want variety.
I suppose you could try to define away all subjectivity by appealing to utilitarianism, and maybe a post-singularity optimiser that can simulate all brains to gauge their reactions could create a perfect peice of art by those standards. Its still useless for discussing matters in the present.
Calling things “an art, not a science” has always been a pet peeve of mine. And I’ve heard people say things like, “there’s no best way to do it’. In particular, I’m taking a Responsive CSS course on Udacity and the guy said these things (if you listen closely, you could hear me cringe).
And then there’s the idea that art is like inherently intuitive, whereas science isn’t. I want to focus on the “art is inherently intuitive and not about breaking things down into components like science” part. My thought is that these people who say this are confusing their map for the territory. They may not know how to deduce what the perfect painting would look like, but that doesn’t mean that it’s not possible.
I know that there are different versions of these beliefs, and that I may be misunderstanding them. If so, please correct me. Anyway, what do you guys think?
To me it’s shorthand for “It took years of practice to get good and I can’t explain the process I use. The problem appears too complex for a scientific approach to be worth the effort. I certainly can’t think of any polynomial-time algorithm and my experience has led me to doubt that one exists.”
It’s a caution for beginners. The expert is basically saying that getting good at it will take a lot of work and finding a good systematic approach is likely to be a dead end.
One can certainly brute-force the issue by creating every possible painting and comparing them. Whether this can be done efficiently is another matter altogether.
One charitable interpretation is “it’s something you learn by doing, not something you learn by reading”.
“Art” has a bit of a double meaning, there’s the “something that’s pretty/pleasing/aesthetic/original/creative”, but there’s also the “craft” meaning, as in “the art of XXX”.
Two reactions to this:
1) If someone says something can’t be broken into component parts, a more charitable reading is that they think that trying to do so is a waste of time and less likely to bring progress than just a lot of practice. Even if it’s possible in theory, that doesn’t mean it’s actually a good idea, so warning people against it can be totally reasonable, and isn’t “confusing their map for the territory”.
2) HOWEVER, in the case of art, most forms of art I can think of—drawing, painting, storytelling, animations, etc. - most definitely CAN be broken into component pieces, and often those component pieces can be broken into component pieces too, etc. - just check out the right section in any library.
You can’t learn to draw by reading a book, but a good book on drawing can tell you what individual skills you should practice, and how to do so.
Art is about relying on intuitive pattern matching and not following strict rules.
Computers can follow strict rules that you program into them but are bad at creative pattern matching.
A correct breaking down may very well result in 10,000 rules with complex interactions between them. The human brain has a lot more than 10,000 neurons that are active during a particular decisions process.
In LW jargon I’d phrase it as: “It’s a system one thing, not a system two thing.” I think this is what most people do mean when they use “it’s an art, not a science.” When something is considered an art and not a science it’s something that can’t be done well by “just” following a set of instructions. Keep in mind that the popular view on science (when it’s positive) is seen as strictly adhering to the scientific method (form hypothesis-->test hypothesis-->adjust hypothesis) and that this is something that anyone can do. The difficulty of each individual step is ignored or seen as something that can be learned without much trouble. The pop-culture view on art is “mysterious process.”
And you’re probably right that you can distill (some) parts of art into “rules” about good art. I recently heard a radio interview with a professional photographer and he could explain why each photo was good or bad by adhering to a set of simple principles and he could explain those principles. I do think, however, that part of what makes him a great photographer is that those simple principles have become part of his system one thinking.
I think that’s a great way to phrase it!
I’m a bit skeptical that this is true. I sense that the majority of people don’t actually believe that art is reducible.
Yes, you might very well be right. What I meant to say is that I think “system one, not system two” is the general sort of idea that people want to convey, not that it was the exact same thing.
“an art, not a science” =
we cannot explain it explicitly well, yet (charitable interpretation)
we do not bother to measure our outcomes (uncharitable interpretation)
Consider the following two statements:
1) Things fall downwards, and anyone who disagrees is objectively wrong.
2) Mozart is better than Iron Maiden, and anyone who disagrees is objectively wrong.
I think there are problems with the second statement, and not just that people will stop inviting you to parties. You’re trying to impose your opinions on other people.
There are approaches you could take to determine which art is objectively better. For designing a website, the answer is which gets more hits, or sells more products. But as far as art goes… you could take a democratic vote over which artist is better. You could, in a thought experiment, extend the vote to a sum over the whole of mindspace. You could weight by IQ.
But I don’t want my tastes in art to be determined by a democratic vote of what everyone else thinks. At the end of the day, the people who say “there’s no best way to do it” are not able to work out what the most popular website ever would look like, and saying that the currently most popular site is objectively best would stifle innovation.
They’re not trying to say philosophically universally correct points. They’re not theologians trying to determine whether God could create the perfect painting. They’re trying to teach humans.
I think the most direct way to do it would be to measure people’s brain activity. I’d be skeptical of self reporting. I think people would be swayed a lot by what they think they should like. I’m pretty sure I’ve read some research on this but can’t recall what it was :/
It wouldn’t be. In this case “best way” would be defined as something along the lines of maximizing total happiness.
Measuring brain activity is certainly possible but brings all sorts of challenges, rendering this kind of approach far less direct than it may seem on the surface.
However your thoughts about expectations are very much supported by a fMRI study (Kirk et al. Neuroimage 2009, direct link to PDF ) manipulating participants’ expectations about artistic images with the following instructions:
Paintings were randomly assigned to these two conditions for each participant, and ratings of aesthetic quality were obtained for each painting as well. Instructions modulated brain activity: the same artistic images evoke different responses in prefrontal/orbitofrontal cortex depending on participants’ expectations, and the same images received higher aesthetic ratings if the participant believed they came from a prestigious museum.
The authors also correlated brain activity with the aesthetic ratings given to each image (collapsed over context). No regions were reliably modulated by aesthetic ratings, which should have been the case if there are some regions that are responsive to “artistic quality” detached from context.
Of course, plenty of potential problems with this study, perhaps the most salient to me being the use of entirely abstract artistic images and artistically naive participants. But it does highlight some of the challenges related to revealing information about aesthetics from brain activity.
Still, rather than having one universal best art, the best approach is to have different peices of art to apply to different people. There may be a universal best piece of art for one person, but I imagine they would still want variety.
I suppose you could try to define away all subjectivity by appealing to utilitarianism, and maybe a post-singularity optimiser that can simulate all brains to gauge their reactions could create a perfect peice of art by those standards. Its still useless for discussing matters in the present.