Analog Truth
Dec. 23rd, 2004 12:02 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
dr_tectonicSo how come I'm still running around half-frantic even though I got my Christmas shopping done early? Ah well. At least we're past the solstice, so now I get more sunlight in my day.
I borrowed a philosophy book from Jen & Nick the other day. Usually philosophy makes me want to tear my hair out, but this book is actually very interesting. I think it might be because the book isn't trying to convince me of anything; it just says, "here's what Kant believed, and here's the reasoning behind it," which means that I can then say, "Aha! And here and here are the spots where he's smoking crack," and figure out what to keep and what to toss.
I might post some summaries of various philosophical concepts and my responses to them, because I find that I'm thinking about a lot of these kinds of things lately, but if I do, I'll be sure to hide them carefully behind cuts, so as not to inflict my ramblings upon the innocent.
For example, I think that many problems in philosophy arise because "truth" means one thing when we're talking about abstract concepts and another when we're talking about the real world. In the world of pattern, which is sort of like Plato's "forms", statements are either true, false, or poorly-formed. You don't have things that are "mostly true" the way you do in the physical world.
Consider the statements "all triangles have angles that sum to 180 degrees" and "all mothers are women". The first statement is always true (as long as we're talking about Euclidean space); it's not like you can find one particular triangle that it doesn't hold for if you just look long enough. But for the second, well, everything is much messier. What about animals? Isn't a mare with a foal a mother? Okay, we say, what we really meant was "all mothers are female". But what about species that divide things up differently? Bees have queens, workers, and drones; which one is "female", and why? Is it ovaries that matter? Or the womb? What about seahorses? What about fish that change sexes under environmental stress?
Okay, maybe fish gender is irrelevant for philosophical discussions. Maybe we only care about humans. But even though the vast majority of humans fall into one of two sexual categories, there's a non-zero number that don't. Consider a woman who bears a child, then sexually transitions to male, including sex-change surgery. Is s/he a "mother"? Is s/he a "woman"? What about a similar person a few decades from now who gets gene therapy to replace X chromosomes with Ys? Science fiction, yes, but near-future, not far-future.
Anyway, the point is, when we talk about things in the real world (as opposed to purely abstract things), I think it makes sense to talk about truth being an analog and continuous value, not a discrete and boolean one. Things in the real world are more or less true, but never wholly so. We can be very sure something is or isn't true -- but not certain. Ivory soap is as pure as our truth can get.
Okay, I just rambled a whole bunch more than I intended to. Summary: I have this idea that abstract truth is boolean and real-world truth is floating-point. I think that might be philosophically interesting, and I might explore it more. Probably not a new idea, but as best I can tell from the book, none of the big philosophers has ever dealt with it.
For now, though, the truth is that I need to get to bed.
...As soon as I fold the laundry.
I borrowed a philosophy book from Jen & Nick the other day. Usually philosophy makes me want to tear my hair out, but this book is actually very interesting. I think it might be because the book isn't trying to convince me of anything; it just says, "here's what Kant believed, and here's the reasoning behind it," which means that I can then say, "Aha! And here and here are the spots where he's smoking crack," and figure out what to keep and what to toss.
I might post some summaries of various philosophical concepts and my responses to them, because I find that I'm thinking about a lot of these kinds of things lately, but if I do, I'll be sure to hide them carefully behind cuts, so as not to inflict my ramblings upon the innocent.
For example, I think that many problems in philosophy arise because "truth" means one thing when we're talking about abstract concepts and another when we're talking about the real world. In the world of pattern, which is sort of like Plato's "forms", statements are either true, false, or poorly-formed. You don't have things that are "mostly true" the way you do in the physical world.
Consider the statements "all triangles have angles that sum to 180 degrees" and "all mothers are women". The first statement is always true (as long as we're talking about Euclidean space); it's not like you can find one particular triangle that it doesn't hold for if you just look long enough. But for the second, well, everything is much messier. What about animals? Isn't a mare with a foal a mother? Okay, we say, what we really meant was "all mothers are female". But what about species that divide things up differently? Bees have queens, workers, and drones; which one is "female", and why? Is it ovaries that matter? Or the womb? What about seahorses? What about fish that change sexes under environmental stress?
Okay, maybe fish gender is irrelevant for philosophical discussions. Maybe we only care about humans. But even though the vast majority of humans fall into one of two sexual categories, there's a non-zero number that don't. Consider a woman who bears a child, then sexually transitions to male, including sex-change surgery. Is s/he a "mother"? Is s/he a "woman"? What about a similar person a few decades from now who gets gene therapy to replace X chromosomes with Ys? Science fiction, yes, but near-future, not far-future.
Anyway, the point is, when we talk about things in the real world (as opposed to purely abstract things), I think it makes sense to talk about truth being an analog and continuous value, not a discrete and boolean one. Things in the real world are more or less true, but never wholly so. We can be very sure something is or isn't true -- but not certain. Ivory soap is as pure as our truth can get.
Okay, I just rambled a whole bunch more than I intended to. Summary: I have this idea that abstract truth is boolean and real-world truth is floating-point. I think that might be philosophically interesting, and I might explore it more. Probably not a new idea, but as best I can tell from the book, none of the big philosophers has ever dealt with it.
For now, though, the truth is that I need to get to bed.
...As soon as I fold the laundry.
no subject
Date: 2004-12-23 12:03 am (UTC)Seem to me that you can always refine (define?) your language to take into account the wackiest exceptions you care to think of, but that we're better off by and large trying simply to make statements that are at least 80% true. Wiggle room is good, and it's why we have a judicial system, in large part. This works in regular-people life, too: it's why we have discussions.
no subject
Date: 2004-12-23 12:05 am (UTC)I find crisp definitions a useful property of toy worlds, and toy worlds are a great way to practice modelling and manipulation techniques, but to apply crisp definitions to things in the real world typically strikes me as silly. I feel a little odd even saying this, since it seems rather like saying that I find air a more reasonable substance to breathe than methane, except that as above I realize that there are people who believe otherwise, and even more people who want to even if they can't work out how.
At its best, philosophy sets up toy worlds in which various ideas can be tried out, then tries to cautiously apply the insights thus gained to the real world. The physical sciences have done pretty well with this strategy, other branches of philosophy have a less impressive resume but still have some accomplishments to their names.
At its worst, philosophy sets up toy worlds and then tries to pretend that you live in them. But, to be fair, in many cases it's not so much that the philosopher makes this error, but that his/her students (readers, etc.) do.
no subject
Date: 2004-12-23 01:09 am (UTC)Anyway, yes, truth as a fuzzy logic concept fascinates me -- thing is, fuzzy logic is way newer than Kant, I think. Or at least, it's not taken really seriously by most folks to this very day. Still, I like fuzzy logic. Fuzzy can be very good. ;-)
no subject
Date: 2004-12-23 05:25 am (UTC)Recommended book: "Reasoning about Uncertainty", by some set of authors that includes Moshe Vardi and Joe Halpern. MIT Press, sometime in the late 90s.
There are much better formalisms. Worth having some fun with.
no subject
Date: 2004-12-23 05:32 am (UTC)I didn't really start to get that all rules have exceptions, and that sometimes the important part was the rule, and sometimes it's the exception, until I started working with biologists.
In part, that's because I had spent all of my studies to that point working on areas where it really still feels like there are absolutes, as in your example of the triangles. That's the trouble with being a mathematician...
In the "real world", well, I'm not entirely sure that there aren't parts of the real world that aren't concrete, but I guess I'd be surprised. And I continue to believe that the ability to hold mutually contradictory beliefs is both a source of cognitive dissonance and evidence of maturity.
As true as Ivory soap
Date: 2004-12-23 10:36 am (UTC)Wow. That's almost Bumper-Sticker worthy. It requires just a touch too much context, but man oh man is it a great summary of an idea that some people have a really hard time grappling with.
I can understand where they are coming from. I have been a math geek for a long time, but for some lucky reason, I didn't fall into the trap as deeply as many other people. Maybe because I was offended when they pulled imaginary numbers out of the air. Or when they admitted that they defined zero factorial differently at different times, sometimes it's 1 and sometimes it's zero, and we mostly hope it's obvious in context. Maybe it's because I got blamed for fights that my little brother started with me. Who knows?
Chains of logic
Date: 2004-12-23 10:55 am (UTC)So, I when faced with a 'should' choice, I try to find the closest truth or most significant aspect to base the decision on. For most choices, you can construct a chain of logic from one general rule or another (like "treat people as you would want to be treated" or "don't let people take advantage of you"). It becomes a question of balancing the strength and directness of the various rules.
Re: Chains of logic
Date: 2004-12-23 11:13 am (UTC)I've been told that if you answer a question of the form "Is X right or wrong" with "well, that's a complicated question and I really need more details; are we talking in general, or did you have something specific in mind?", that's officially called "post-modern ethics". (I think.)
no subject
Date: 2004-12-23 11:11 am (UTC)Formulae and equations are mathematically true. (.T.)
Stories and explainations are accurate (with an implied "as far as we know or can know.)
Religious and personal beliefs are subjectively true (or "true for you").
I always mean what I say. While the vast majority of the time it will be true or accurate, even if it's inaccurate, not true, or an outright lie ("no, those jeans don't make you look fat") it's what I meant to say, and I'll act as if it were true from then on.
no subject
Date: 2004-12-23 01:15 pm (UTC)I admire that.
Personally, I often say things that leave me scratching my head going "Huh. I wonder if I mean that" or even "Huh... I wonder what I meant by that"... I open my mouth and things come out. They are sincerely reflecting things in my head, which is fine, but when I try to synthesize from that a "me" to which I can attribute a coherent stream of meaning, I often fail (and when I succeed, I suspect I'm often confabulating).
no subject
Date: 2004-12-23 02:13 pm (UTC)Beliefs are statements about the structure of someone's internal model of the universe, so while they're certainly accurate with regard to the person's thinking, the model itself may or may not be an accurate representation of the world. Which I think is exactly what you said.
I don't really follow you on your last truth-concept, though. You'll have to elaborate.
With regard to questions like "how do I look?", I follow Miss Manners, who says that it's important to distinguish between what is meant and what is actually said. Sometimes, people are asking for your candid opinion of their appearance -- but more often, what they are actually requesting is that you give them reassurance, which of course one does.
I think that's in line with what you said for the fourth thing, yesno?
Elaboration
Date: 2004-12-23 02:58 pm (UTC)On the second, well, that was the facetious example, but it goes deeper.
I would really like to say that I always speak the truth, but that is not 100% accurate; maybe 95%. I'm known for my information abilities: my knowledge of facts, figures, trivia and tidbits of information; and my ability to search out info either in reference materials or on-line.
There are times, however, when the truth, or an accurate representation of reality (especially for opinion questions) is not the best answer — extreme ones, as when a madman with a gun asks you if you're "feeling lucky" (or a martial artist asks if you're making passes at his partner); serious ones, as when someone needs reassurance; and trivial ones, like answers to telemarketers. Also, there're those cases when you're going to outright lie, without being asked, either to avoid a consequence or to gain a benefit. ("I'm so sick today." "I find you so sexy.")
So, while I cannot say that I'm truthful all the time, I take some satisfaction in the phrase "I always mean what I say." At the time I say it, I mean it. After that, I try to act and think as if I actually thought that way before I made the statement. While it may not have been "true" for me before I said it, it becomes "true" for me after I say it.
Also, I sometimes don't say anything when I could venture an opinion. Discretion is the better part of valour, still.
In the past, the very perceptive have picked up on the differences between "I always mean what I say." and "I always say what I mean." and "I always tell the truth."
I also recognize the irony in the fact that I use a slight evasion to apparently assert my honesty.
no subject
Date: 2004-12-23 03:53 pm (UTC)I own Passion of the Western Mind, by Richard Tarnas, which is a survey of Western culture and thought, and the couple chapters I read in that were very good, but I never got around to reading the rest.
As far as the actual speculation, I've been moving towards the idea that "truth" only has meaning when dealing with the physical world, where there is something to test against that will consistently return the same results. And that may even only apply in physics and chemistry, because biology is way too messy (explaining away bad data by saying "Oh, the cells weren't happy that day" appears to be a perfectly legitimate thing to say in biology)
As soon as you are dealing with anything human or social related, it all becomes a matter of perspective. I like the concept of E-prime, which is a version of English that abolishes the word "to be" and its declensions. Nothing "is" anything. It only appears to be such from a given perspective at a given time. And if people were constantly reminded by their language that theirs is only one view on things, it seems like we could avoid a lot of messes.
More commentary on such stuff at my review of Quantum Psychology, by Robert Anton Wilson, where I read about E-prime, and in my review of Language in Thought and Action, by Hayakawa. Or heck, any of my recent posts blathering on about social context.
no subject
Date: 2004-12-29 02:04 pm (UTC)I just posted a long reply to Chris's comment, down at the bottom, about real-world truth vs. abstract truth. The problem with the physical world test is that it doesn't give you the same results consistently. Even physics and chemistry are really messy with lots of outliers and errors and anomalies that you have to correct for and average out and curve-fit to get the data to display the nice pattern described by the math...
The E-Prime idea is interesting, but totally unworkable. For one, it's based on the Sapir-Whorf hypothesis, and strong Sapir-Whorf (vocabulary constrains thought) is pretty well discredited; the best you can get is "vocabulary can influence thought". For another, lots of languages omit the verb 'to be' in various circumstances, and I haven't noticed that Russian-speakers, for example, are particularly better at recognizing other viewpoints than anyone else. And finally, the idea of being or identity is a pretty basic one; if you remove "to be" and replace it with "seems" in most instances, I think that all it will accomplish is to incline speakers of the language to equate seeming and being. "Seems" will mean "be" for speakers of E-Prime! D'oh!
You could probably get something closer to the intended effect if you did something like what Lithuanian and many Native American langauges do, where you have an obligatory case marker for "evidentiality". In other words, to form a grammatically-correct statement of fact, you have to indicate the source or justification for the claim, whether it's direct or indirect observation, personal, second-hand, or something "everyone knows", etc. I've observed that MIT students tend to do that at lot in conversation, marking all their statements of fact with a source. Look, I just did it myself!
Let be be finale of seem
Date: 2004-12-29 03:33 pm (UTC)--Chris
Re: Let be be finale of seem
Date: 2004-12-29 03:50 pm (UTC)Re: Let be be finale of seem
Date: 2004-12-30 08:57 am (UTC)--Chris
no subject
Date: 2004-12-23 04:25 pm (UTC)no subject
Date: 2004-12-29 11:53 am (UTC)Isn't it possible that the problem you're discussing here is one of language and not truth? That truth doesn't fundamentally change in the "real-world," but just that flaws in our language are more apparent in discussing practical and specific cases than in discussing mathematical notions with built-in language sets (which are therefore more precise)?
(To illustrate this latter idea, I could bicker with your first proposition on triangles if I insist that any set of three angles is a "triangle." It's not so dissimilar from where you take the word "women" in your real-world proposition. It doesn't take long for us to narrow down what "triangle" really should mean, because the language for such is heavily formalized. It may take longer and we might have disagreements in narrowing down the terms "mother" and "woman," but it doesn't become impossible and it doesn't mean that there isn't a way to make the statement a verifiably true one.)
The result is the same: Statements made about the real world are always going to be More-Or-Less True with exceptions and contextual distinctions. But does that really mean that Hard-and-Fast Truth is nonexistent in the real world? Or just that it is difficult to ascertain and express?
--Chris
no subject
Date: 2004-12-29 01:10 pm (UTC)It's not a problem with imprecision in language. For one thing, it's meaningless to talk about truth outside the context of language.
The problem is that language is an inherently abstract system of representation, and "truth" is a description of a relationship between patterns. So if we're talking about other abstract systems, we can set up a perfect correspondence between representation and object, and "truth" only encompasses the consistency of the relationship.
Whereas if we're talking about physical reality, we're making abstract statements about observed patterns, but the observations are flawed and the patterns themselves are imperfectly instantiated. So there's always room for mismatch between the representation and the underlying reality. The notion of "truth" then has to deal not only with the abstract pattern relationships, which are boolean-true, but also with the relationships between representation and reality, which are analog-true.
To use
So in the triangle-versus-woman example, the problem isn't that we're using a flawed definition of the word. We all know what triangles are and what women are; those are decent representations of the abstract patterns in question. The problem is that women are real entities, and there are real-world entities that instantiate the abstract pattern "woman" imperfectly or incompletely.
And you can't solve that problem with better definitions. Pick a term, and no matter how narrowly you define it, I can construct an example that will straddle its definitional boundary. "Alive", "human", "male", whatever, as long as these terms are referring to real objects, there's a little bit of fuzziness, a small grey area at the edges where you can find boundary cases that are difficult or impossible to categorize.
The only statements that aren't fuzzy are basically identity statements ("the set of Floyds includes Chris, Rose, Andrew, and Liese"), and the process of refining definitions to exclude boundary cases eventually results in one big identity: "the universe is the way it is", which, while boolean-True, is not particularly useful.
So, to sum up, it's not that Hard-and-Fast Truth is absent in the real world, it's that the idea isn't even meaningful in that context beyond the non-linguistic atomic identity of "this-here-now". Any statement about patterns in the real world is More-Or-Less True because of the inherent mismatch or gap between abstract pattern and concrete instantiation of pattern. And it's not because language is imperfect, it's because reality is.
Fascinating!
Date: 2004-12-30 09:24 am (UTC)That's interesting, and I'm going to argue with it here to test it, not because I disagree with it. I might agree with it. I'm not sure. First, tell me: is this a notion that's plainly apparent to mathematicians or physicists? I think you made mention in another comment about chemicals or some such. If there's a scientific underpinning to this which is important, I'd like to hear it.
The first challenge to this notion that comes to mind is, If reality isn't perfect, what is? You seem to be arguing from an essentially Platonic stand-point, which is that these "patterns" are hard-and-fast True and perfect and reality is fuzzy and imperfect. But doesn't our knowledge of these patterns derive from observations in reality (and extrapolation beyond that)? And isn't it conceivable that the problem is not that reality is flawed, but that we either don't recognize all the patterns at work (perfectly at work) in reality or that we can't account for all the operative ones at work at a given observational moment?
I guess it seems to me that if we're going to start labeling things as imperfect, the first recipient should be ourselves, not reality.
Again, I still think the conclusions about how we should talk and think about abstract vs. real things is the same under either model. True/Accurate, for instance, is a useful--probably vital--distinction.
--Chris