The nature of beauty: subjective or objective?

[clarsct]

I have to object to the 'average' arguement about human beauty, once again. (Re Mercutio and Beerina)

All the machine does when it 'averages' the face is remove any imperfections in the skin, and makes the face more and more symmetrical...which is NOT the average.

We do tend to find more symmetrical people more beautiful. But that symmetry is not the norm, by any means.

In other words, we do not see extremely beautiful faces all the time, they are rare. They are NOT average.


Sorry, had to get that off my chest.


Oh, and of COURSE beauty is subjective, else we would all compete for the same mate.


ETA: By what parameters is that machine measuring beauty, BTW? Could it be that the people who set up those parameters are biased?

 

[Z]

ETA: By what parameters is that machine measuring beauty, BTW? Could it be that the people who set up those parameters are biased?

A machine that measures 'beauty'?

By whose standard?

If there is such a machine, then whatever it seeks for its beauty parameters was undoubtably placed there by a person - whose criteria for 'beauty' was based on his/her subjective preferences.

I find it highly unlikely the machine is going to spit out a face that is pleasing to an Aryan-racist and a Pai Dong and a Seminole at the same time.

 

[Dark Jaguar]

I'm surprised this is even a question any more. Isn't it pretty well established that what's beautiful is basically just a function of the brain, as developed from genetic preprogramming and what they learn through their life time? That said, a program can be written keeping in mind whatever basic "rules" seem to apply to most humans to find an average idea of what most humans consider beautiful.

This will of course be different from a program designed to find the standard image of peacock beauty, but there will also be some similarities.

It's not built into the universe, it's built into the brain. It's subjective in some senses, and objective in others. Get an individual's programming down, and you can determine exactly what would be beautiful to that person, but it's not an absolute of nature like the speed of light or something. It's just a description in terms of mental programming.

Yeesh, I'm surprised philosophers are still mucking around in these outdated questions. Next I'll be seeing a thread about the definition of "art".

 

[Ichneumonwasp]

Bubblefish,

I think you need to be a little clearer on what you mean by objective and subjective. We use these words in ways that confuse the topic more than help. For instance, when someone says that beauty is subjective, they generally do not mean that it is completely and utterly subjective; rather there is a subjective component to it. One of the general ways of using "objective" means "true independent of viewpoint" or it can mean "true for everyone within a group". Your current examples seem to imply different uses of the term -- the computer example using the first, the appeal to symmetry using the second.

For beauty to be objective, it must be true for every observer. We clearly do not see complete agreement in individual examples or in standards of beauty across cultures. We do see basic algorithms that we all seem to use (I doubt it is true of every human, though) to decide upon the beautiful. This is reflected in the computer example, for instance. But that does not cover the field of beauty. It is merely one aspect of it, modified by several others, many of which are individual or group based.

You have essentially already lost the argument by stating that beauty is both objective and subjective. That means that beauty is subjective. Of course we share means of deciding what is beautiful. If we didn't we wouldn't be able to discuss the matter. Objective is absolute. Subjective is not. Any combination of objectivity and subjectivity amounts to subjectivity. It is subjectivity within a framework, but that is what virtually everyone has argued all along in this thread -- and that includes Zaay.

 

[WWHP]

Bubblefish,

I think you need to be a little clearer on what you mean by objective and subjective. We use these words in ways that confuse the topic more than help. For instance, when someone says that beauty is subjective, they generally do not mean that it is completely and utterly subjective; rather there is a subjective component to it. One of the general ways of using "objective" means "true independent of viewpoint" or it can mean "true for everyone within a group". Your current examples seem to imply different uses of the term -- the computer example using the first, the appeal to symmetry using the second.

For beauty to be objective, it must be true for every observer. We clearly do not see complete agreement in individual examples or in standards of beauty across cultures. We do see basic algorithms that we all seem to use (I doubt it is true of every human, though) to decide upon the beautiful. This is reflected in the computer example, for instance. But that does not cover the field of beauty. It is merely one aspect of it, modified by several others, many of which are individual or group based.

You have essentially already lost the argument by stating that beauty is both objective and subjective. That means that beauty is subjective. Of course we share means of deciding what is beautiful. If we didn't we wouldn't be able to discuss the matter. Objective is absolute. Subjective is not. Any combination of objectivity and subjectivity amounts to subjectivity. It is subjectivity within a framework, but that is what virtually everyone has argued all along in this thread -- and that includes Zaay.

I will address this on my other thread, it's the first reasonable rebuttal I have recieved, and my response is better served on the skepticism and the normal thread, where this discussion orginally began.

http://72.32.2.238/forumlive/showthread.php?t=73672&page=52

 

[clarsct]

For those that wish to maintain their philosophy and opinion with science and math.

Here is a PDF which attemtps to get computers to predict a beautiful human face.

To quote from the SCIENTIFIC and MATHEMATICAL paper.



http://www.mitpressjournals.org/doi/pdf/10.1162/089976606774841602

If a machine can learn beauty, then clearly beauty must be objective. machines have NO subjectivity, and thus, this clears the whole argument out of the water objectively.

Beauty is not soley objective.
Beauty is not soley subjective.

Beauty is BOTH objective, and subjective.

Toodles.

This is what I was speaking of when I said 'machine'. I am sorry, Z, i understand you have him on Ignore.

Call me old fashioned, but I tend to ignore people the 'wetware' way.

 

[Earthborn]

machines have NO subjectivity

Learning machines do. The fact that they needed a learning algorithm to get the computer to recognise beauty proves that it is subjective, because if it was objective it should have been possible to directly measure it without learning.

If a machine can learn something, it is not proof that it is 'universal' or 'objective'. You can teach a learning machine all sorts of things that are not universal or objective.

 

[UserGoogol]

We always seem to devolve back to "man is the measure of all things". Even attempts to "objectify" morals, as in Kant's categorical imperative ultimately lead back to some form of utilitarianism, which is ultimately based on a particular value system, which is ultimately based on emotion and feeling.

I have my doubts that ethics can be made objective in a meaningful way, but utilitarianism is based on emotion in an objective way. It is an objective fact that "Ted is happy." You can reach into his brain and see those neurons firing away. Technology has not yet developed the ability to precisely measure a person's happiness, but in principle it seems possible, so happiness exists in a very objective sense.

Although I agree with you that aesthetics can be made intersubjective but not objective in the stronger sense.

 

[Earthborn]

We do tend to find more symmetrical people more beautiful. But that symmetry is not the norm, by any means.

In other words, we do not see extremely beautiful faces all the time, they are rare. They are NOT average.

Average does not necessarily mean 'common'. If half a class scores a 0% on a test and the other half 100%, the average score of the class is 50% even if there is no one who has that score. Same thing is true here: very few people have average faces. Most of them have faces that have at least some characteristics that deviate from the average.

The "average is beautiful" people argue that while people rarely see perfectly average faces, their brains average out the faces that they see to become a 'template' for face recognition. The closer a face matches the template the stronger the sense of recognition is and this is what we call beauty.

There are problems with this idea, though. Many people consider some deviations from the average of some characteristics to be especially attractive. You might argue that this is because a person has seen different faces and therefore has built up a different average template. But then it becomes unfalsifiable; if people like average faces it is because their brains has averaged the faces it has seen, but if they like faces that are different than average it is also because their brains has averaged the faces it has seen.

The most we can say at this point is that what people on average think is beautiful is what faces on average look like. But individual people are likely to consider faces that are somewhat different from the average to more attractive than the actual average, because it will have more 'personality'.

 

[WWHP]

Learning machines do.

utterly impossible. AI does not have consciousness or experience.

The fact that they needed a learning algorithm to get the computer to recognise beauty proves that it is subjective, because if it was objective it should have been possible to directly measure it without learning.

Utterly impossible. Human intelligence also has to discover what beauty is.

If a machine can learn something, it is not proof that it is 'universal' or 'objective'. You can teach a learning machine all sorts of things that are not universal or objective.

Why do you call them learning machines? do we call airplanes flying machines? I'm sorry, it just sounds, er, old fashioned when you call them that!

If AI can learn to detect patterns, and those patterns are consistant with facial beauty, then that could only suggest that the pattern is digitized, and we can only digitize something that is quantifiable. If it has number and measure, then it's objective.

 

[Lonewulf]

utterly impossible. AI does not have consciousness or experience.

Yet.

We don't have true AI yet.

 

[Earthborn]

AI does not have consciousness or experience.

True (for some definitions of conciousness and experience), but irrelevant. It does not need to have it, to be 'subjective'. All that is needed for it to be subjective is that it considers something 'beautiful' depending on what it has learned.

Human intelligence also has to discover what beauty is.

So what is your point? Humans are also subjective.

Why do you call them learning machines?

I call them "learning machines" because computers are not necessarily learning machines, and learning machines are not necessarily computers. Calling the things discussed here "computers" gives people the incorrect impression that this is about machines that follow strict deterministic patterns and strict preprogrammed calculations on input. But a "learning machine" is something different (although it can be simulated on a computer) because it is specifically designed to change its behaviour in ways that are not preprogrammed, but rather to "program itself".

There are many different learning machines, such as neural networks and genetic algorithms, but it is still useful to have a word to describe them all.

do we call airplanes flying machines?

Yes, we do.

If AI can learn to detect patterns, and those patterns are consistant with facial beauty

That is to say: those patterns are consistant with what the machine's teacher wanted the machine to learn. The machine will therefore closely reflect the teacher's subjective preferences. It does not discover anything "universal" or "objective" about beauty, but rather it will learn what the teacher likes.

then that could only suggest that the pattern is digitized,

A learning machine is not necessarily digital. Learning is in fact something that is ideally suited for analog machines. Simulations of learning machines running on digital computers often try to simulate analog input by using lots of intermediary values instead of just ones and zeroes.

and we can only digitize something that is quantifiable.

True, but irrelevant.

If it has number and measure, then it's objective.

No, it's not. You can also quantify people's subjective preferences, and this is in fact something for which learning machines are often used; to learn what patterns of preferences people have and to adapt to users.

 

[clarsct]

Average does not necessarily mean 'common'. If half a class scores a 0% on a test and the other half 100%, the average score of the class is 50% even if there is no one who has that score. Same thing is true here: very few people have average faces. Most of them have faces that have at least some characteristics that deviate from the average.

I do hope you're not basing your arguement on sample size of two.
If you graded 50 students, not many would have 100% and not many would have 0%. But you'd see quite a few in the 80-90 range. You'd have a bell curve with many of these 'average' students in the middle of it. We do not see this curve when it comes to people we find attractive. We do not see lots and lots of beautiful people and a few average and a few ugly ones. We see lots and lots of average and few very symmetrical, and few that are gawdawful fugly. I'm sorry, but this example simply does not reflect reality.

The "average is beautiful" people argue that while people rarely see perfectly average faces, their brains average out the faces that they see to become a 'template' for face recognition. The closer a face matches the template the stronger the sense of recognition is and this is what we call beauty.

Perfectly symmetrical is rare, I'll grant you, but if symmetry were the average, there would a LOT of '9's out there. You'd have a curve that apexed at about 8-9, then dropped at 10. Once again, we simply don't see this happen. What we see is that fours marry fours and nines marry nines, most of the time.(Unless money or pregnancy got involved, or some other mitigating circumstance.) There simply isn't a lot of nines around. Plenty of sixes, but few nines.

There are problems with this idea, though. Many people consider some deviations from the average of some characteristics to be especially attractive. You might argue that this is because a person has seen different faces and therefore has built up a different average template. But then it becomes unfalsifiable; if people like average faces it is because their brains has averaged the faces it has seen, but if they like faces that are different than average it is also because their brains has averaged the faces it has seen.

The most we can say at this point is that what people on average think is beautiful is what faces on average look like. But individual people are likely to consider faces that are somewhat different from the average to more attractive than the actual average, because it will have more 'personality'.

This last part is particularly good. What happens when we see a bunch of nines and tens in advertising?
(Do we reject less good looking mates?)

 

[Earthborn]

We do not see this curve when it comes to people we find attractive.

That's because attractiveness depends on many characteristics. If you measure the noses of a thousand people then you'll find that most people have an average nose. If you measure the distance between the eyes, you'll find that most people have an average distance between the eyes. Same thing for the width of the mouth, the height of the cheekbones, the width of the jaw... etc. But you'll also find that very few people have the average of all these characteristics. Most people have a few of their many characteristics deviate noticebly from the average. Someone who looks "ordinary" is not someone who looks like a perfect average of all possible human characteristics. There are clear deviations that make it easy to recognise someone.

Perfectly symmetrical is rare, I'll grant you, but if symmetry were the average, there would a LOT of '9's out there.

No, you just need about as many people who deviate in a particular way on the right as people who deviate in that way on the left, for symmetry to be the average. Most people are fairly symmetrical.

What happens when we see a bunch of nines and tens in advertising?

Difficult to say, though I think it is something that may have a profound impact on our perceptions of other people. Many of the photoshopped beauties in cosmetics ads do look rather like the averaged out faces and we are expected to consider this to be what 'beauty' looks like. A person who would look like that would certainly draw people's attention in if she existed in real life. At the same time people are so used to such images that they start to look for imperfections, deviations from the average that make it possible to identify people, to give them 'personality'. Such differences will probably become more important and be considered more attractive the more we are bombarded with such averaged out looks.

 

[Ichneumonwasp]

I have my doubts that ethics can be made objective in a meaningful way, but utilitarianism is based on emotion in an objective way. It is an objective fact that "Ted is happy." You can reach into his brain and see those neurons firing away. Technology has not yet developed the ability to precisely measure a person's happiness, but in principle it seems possible, so happiness exists in a very objective sense.

Although I agree with you that aesthetics can be made intersubjective but not objective in the stronger sense.

OK, but this depends on using the word "objective" in a particular way, meaning "observable". It is an objective fact that Ted has neurons firing that we can see correlate with happiness. But that is not the use of "objective" being discussed in this case. "Objective", when it comes to ethics or beauty is either "true independent of observation" or "true for all individuals" (meaning the valuation, not the brute fact of an experience). That Ted is happy may be "objective", observable. But the reason for Ted's happiness is dependent on Ted's individual valuation system. Now, some of his values are shared by 90-99% of us, so we tend to view those values as "objectively true" -- like murder is wrong and symmetry is beautiful. But, the reasons behind his valuations are not an inherent feature of the universe (true independent of observer) or shared by all observers (unless proven otherwise).

That happiness occurs I think we can mostly all agree upon. That Ted is happy, I think we can also agree (since we have defined it thus). But the crux of the matter with ethics and aesthetics is the reason for the happiness. The neuron firing is simply a brute fact of the universe, something observable. The actual valuation -- where Ted's happiness is directed and why -- determines the content of utilitarianism. That is not objective in any sense that I have been using the word in this thread.

Or another way of putting it.......we can theoretically see Ted's happiness in his neuron firings, but we cannot see Ted's happiness about the double fudge chocolate ice cream cone in his hands (unless we are privy to every predisposition in him). And even there, we do not experience what Ted experiences, but only view from the outside the correlate of those experiences. "Subjective", in the way I have been using it, refers to those predispositions in Ted that are unique to him or to his culture and determine how and why he experiences pleasure and happiness. This is independent of our being able to witness the neuron firings that constitute his experience. However we want to cut it, Ted is still the only real observer here -- the only one able actually to experience the happiness.

 

[Z]

This is what I was speaking of when I said 'machine'. I am sorry, Z, i understand you have him on Ignore.

Call me old fashioned, but I tend to ignore people the 'wetware' way.

Thanks.

As I thought - subjective standards were put into the machine.

Of course, since he doesn't comprehend the distinctions between 'subjective' and 'objective', and the correlations between subjective evaluations and objective factors, he can't possibly comprehend this paper.

We can establish all the 'objective' factors for beauty in the universe, and the final evaluation of beauty will still be a subjective one, within the mind of the person observing. A woman can meet EVERY objective criteria for beauty, and someone is likely to NOT find her beautiful. Furthermore, someone is likely to declare beautiful someone who does NOT meet the objective criteria, on the basis of criteria that is not included in the list.

Hence, beauty is subjective.

 

[Ichneumonwasp]

my response is better served on the skepticism and the normal thread, where this discussion orginally began.

Um, OK, but I have a limited amount of time and don't generally traipse about to look at other threads in depth. I have barely sampled anything on this site outside the Philosophy and Religion subheading. If you would like to discuss it here I would be happy to do so. Perhaps cutting and pasting to this thread? If not, that is OK too, since I realize I am asking a bit much.

 

[Ichneumonwasp]

Usergoogol,

Or, to put what I wrote above into Mercutio's language -- subjective means one observer, objective means many observers (suing "objective" to refer to "observable").

This issue arises frequently in discussions of consciousness, so I don't want anyone to get the wrong impression about what I wrote above. The fact that only one observer has access to a particular experience does not make it somehow unexplainable. The information is simply priveleged. The same thing actually happens everywhere, but we don't think about it the same (probably because we are stuck on ourselves).

For instance, using your example above, we can see Ted is happy by witnessing his neurons firing in a particular pattern (theoretically). But we cannot directly experience Ted's happiness. This is because a description of an event is not the event itself. We can describe a computer's function. But that is not the computer's function -- it is a description of it. We can describe someone running and paint a vivid picture of the act. But that is not the person running him or herself.

We consider running and computer function to be mechanical processes, so we don't privelege them. All that mechanical stuff is somehow below us. But brain function is just a variant on the same process. With neurons, the action consists in ions crossing semi-permeable membranes, something we don't see or "visualize" very readily. But it's just the same thing as running, essentially. That is why I think we would have fewer arguments over consciousness if "consciousness" and "mind" were verbs and not nouns.

I just didn't want anyone to read into what I wrote above and decide that I think of "subjective" as somehow mystical or woo. Subjective experiences are brain function. They can be objectively observed (theoretically), but this can only provide a description of them. The objective observation can never be the experience itself.

 

[Malachi151]

The problem is the concept of "beauty" itself. "Beauty" is just an acceptance criteria.

If you create two different robots that are designed to select fruit on a conveyor belt as it passes by them, and you equip both with optical sensors, and you program one robot to select objects that are shiny and of a uniform color, and you program the other robot to select objects that are dull and patchy in color, then you have effectively created two different standards for "beauty" among the two robots.

Assuming then that these were interactive robots, if you asked robot A (shiny fruit), what kind of fruit they like, it would say "shiny and smooth", and if you ask the other one what kind of fruit it liked, it would say "dull an spotted" (or whatever).

That's all that's going on with living beings, human or otherwise.

We have variation in our programming so there is variation in our selection mechanisms, which we express in language by saying that "beauty is subjective", but all that really means is that we each have slightly different programs, or course produced via undirected mutation, sexual recombination, and natural selection.

It's just a matter of which specific features cause rewarding chemicals, such as dopamine, to be released in your brain.

 

[Ichneumonwasp]

The problem is the concept of "beauty" itself. "Beauty" is just an acceptance criteria.

If you create two different robots that are designed to select fruit on a conveyor belt as it passes by them, and you equip both with optical sensors, and you program one robot to select objects that are shiny and of a uniform color, and you program the other robot to select objects that are dull and patchy in color, then you have effectively created two different standards for "beauty" among the two robots.

Assuming then that these were interactive robots, if you asked robot A (shiny fruit), what kind of fruit they like, it would say "shiny and smooth", and if you ask the other one what kind of fruit it liked, it would say "dull an spotted" (or whatever).

That's all that's going on with living beings, human or otherwise.

We have variation in our programming so there is variation in our selection mechanisms, which we express in language by saying that "beauty is subjective", but all that really means is that we each have slightly different programs, or course produced via undirected mutation, sexual recombination, and natural selection.

It's just a matter of which specific features cause rewarding chemicals, such as dopamine, to be released in your brain.

Yeah, I think that nicely covers it. This is like most every "philosophical" discussion in that it falls by Wittgenstein's sword. The whole "controversy" depends on the multiple meanings and nuances of the words "objective" and "subjective" and creating false dichotomies on their backs. In saying that a computer has no "subjectivity" because it cannot feel, we simply move the bar on the argument to a new arena and use a different definition of the word "subjective" than was originally offered. The original distinction in the thread was between "objective" referring to "a universal feature of the universe indpenendent of observer" and/or "complete inter-subjective agreement" and subjective meaning "depending on an individual subject's perspective". Now "objective" has been defined as "externally observable" and subjective as "only internally observable and dependent on emotion".

Any computer programmed along certain lines will evaluation objects based on its individual programming. It has no choice. That involves the first use of "subjective" -- depending on an individual subject's perspective. That was really what was being discussed in the initial statement "Anlina Jolie is objectively beutiful". There is no issue about us being able to observe the process or understand the process of valuation. The issue in that statement deals with complete inter-subjective agreement. We don't all agee with the statement, so there is no complete inter-subjective agreement.

It's all just word games, moving the definitions about in order to cloud reality; which is why I started with my problems over the use of the word "objective". I have seen these ploys used over and over again. It is never done intentionally, I know, but it is frustrating.

Saizai's discussion about "does a building imply a builder?" involves the same issues, despite his valiant attempts to stop anyone from defining the words. Building is used either as a word to denote something specifically designed or to denote a particular function -- shelter, as in your great pictures of caves. There is ambiguity in the words used, so "philosophical" confusion arises.