Reasoning by Analogy is like Poisoning Kittens

Vikram Bath

Vikram Bath is the pseudonym of a former business school professor living in the United States with his wife, daughter, and dog. (Dog pictured.) His current interests include amateur philosophy of science, business, and economics. Tweet at him at @vikrambath1.

Related Post Roulette

96 Responses

  1. Patrick says:

    Analogies are meant to clarify, draw parallels, spark enlightenment.

    In our modern version of discourse, though, they rarely do that. First, it’s at least as common that I see an analogy used to muddy a debate than clarify one. Second, it’s hardly common for anyone to admit they are possibly wrong, so the analogy must be rejected as forcefully as the original argument.

    It is impossible to make useful, fully general statements about things as abstract as “things” when all things differ in meaningful ways. Each thing differs from each other thing. That’s why we each thing gets its own name and why some of them are controversial and other are not.

    Humans are good at dealing with classifications, and classifications are useful *because* they allow us to treat unique things as general things. This is what enables us to have a civil society in the first place… we agree that while some things are different from other things we will still treat them in a common way for a common purpose, which is the last bit that everybody seems to forget nowadays.

    Going back to birth control, buying insurance for an employee that covers some particular thing is different than paying an employee fungible dollars she uses to buy that same thing.

    Why? This is the thing in the fungible-money v. insurance conversation that I haven’t been able to get anybody to articulate. Other than “it just is”.

    And if you think it isn’t different, you still haven’t added to the discussion unless everyone else also already agreed it is the same.

    Well, no, not exactly.

    If I think it isn’t different for reasons, and somebody else things it *is* different for reasons, but those reasons are ones that we agree aren’t suitable for our common purpose, then their different reasons don’t mean jack. Similarly, if my reasons are ones that aren’t suitable for our common purpose then my reasons don’t mean jack.

    The free speech debate is often similar. We’ve agreed (and by “we”, I mean “the general legal structure of the U.S.”) that there is no exception for speech we find odious (we’re slowly coming around to obscene, but we’ve got a ways to go there).

    Somebody can say “I don’t think the American Nazi Party ought to parade because I find it objectionable” and that’s fine, but when they take it the additional step of “… so I don’t think we should allow them to parade, legally”, you lose, because our common purpose doesn’t allow you to treat that thing unlike a Fourth of July parade because you like one and you don’t like the other.

    This is why the “these two things are different” rejoinder is often pretty weak sauce argumentation, as it’s just a cover for not engaging in the discussion of why the two things are the same or not.Report

    • Chris in reply to Patrick says:

      There is a ton of research on how much analogy helps us to think about and communicate difficult concepts. This is not surprising, given that the parts of our brains that deal with higher level cognition (and higher level perception, and even some lower level stuff) basically comprise an analogy machine. However, because the similarities and differences (and there are different kinds of differences, which is important in actually working out analogies) we’re likely to see are the ones we’re focused on, much of the time we’re going to produce really bad analogies in discourse.Report

      • Patrick in reply to Chris says:

        The funny thing about bad analogies is that they usually articulate one of two things about the person who is making the bad analogy:

        (a) what they think the classification is supposed to be
        (b) how badly they are willing to munge the classification to get their tribal inclination validated.

        So they’re useful in that regard.Report

    • Vikram Bath in reply to Patrick says:

      Thanks for the great comment!

      Analogies are meant to clarify, draw parallels, spark enlightenment.
      In our modern version of discourse, though, they rarely do that.

      Similar to my “moral rights” post, this post is meant to take a ground-floor view of how analogies work in practice rather than what they are meant to do. I don’t mean to say that there has never been a good analogy or that it would be impossible to make one. I do mean to say that if everyone were to stop using analogies, the productivity level of the average discourse would improve greatly.

      You are right about classification and the usefulness of certain generalizations. I was probably being a bit overly snarky in my paragraph about “things”. If you prove Mike has certain rights, you shouldn’t have to start all your work over again to prove from scratch that Phil has rights. But I’d note that these aren’t the kinds of things we usually call analogies. We call it an analogy when we prove that Mike has rights and want to say that that means a toaster also has rights.

      Going back to birth control, buying insurance for an employee that covers some particular thing is different than paying an employee fungible dollars she uses to buy that same thing.

      Why?

      I assume you mean to ask why is it *meaningfully* different. It actually is different in practice for at least the reason that the money is sent to a different party by the employer. Now whether that is an important difference is another question, but it’s one I would view as a distraction because it isn’t the original problem that people were concerned about.

      If I think it isn’t different for reasons, and somebody else things it *is* different for reasons, but those reasons are ones that we agree aren’t suitable for our common purpose, then their different reasons don’t mean jack.

      I think if there is a disagreement, then you will find that there also is no consensus on the reasons that are and are not suitable for our common purpose. And you’re then even further trapped in the meta discussion about what makes for a good analogy when all anyone is really interested in is birth control.

      This is why the “these two things are different” rejoinder is often pretty weak sauce argumentation, as it’s just a cover for not engaging in the discussion of why the two things are the same or not.

      Well, your own example relied on there being a difference between saying the Nazi Party shouldn’t parade and legally preventing them from parading. I think we’d both agree that the “those two things are the same” rejoinder would also constitute weak sauce.

      To clarify, it was never my intention to say that “these two things are [importantly] different” is always a good retort to a well-crafted analogy. It is instead that by offering an analogy in the first place, you open yourself up to receiving the “these two things are [importantly] different” retort and spiriting off into the distance about how different they really are, what constitute meaningful differences, and whether alternative analogies might be better. All that time is better spent talking about the thing itself.Report

      • Glyph in reply to Vikram Bath says:

        I do mean to say that if everyone were to stop using analogies, the productivity level of the average discourse would improve greatly.

        I don’t know about this.

        I mean, sure, in an ideal world, the average discourse would then be:

        Person A: I believe [X] because [Y].

        Person B: I believe [not X] because [P].

        Person A: Ah, yes, you are correct. It’s so obvious. How could I have not seen that?

        Back here in the real world of storytelling human beings, we arrive at provisional conclusions all the time, derived from all sorts of random inputs (philosophy, emotional reaction/gut instinct, life experience, peer pressure, anecdote, observation, statistical analysis, rough heuristic, superstition), and we hopefully revise these conclusions as new data and experiences come in.

        Some (maybe most) of that data comes in the messy form of stories and analogies, and not neat mathematical columns, because for most of us that’s not how our brains work (Schilling possibly excepted). It’s often not enough to say “mutatis mutandis” and change one small aspect, the entire frame or context needs to be switched out for a minute. You need a new “lens” for the thing to look any different.

        Believing that reasoning would be better on average minus analogies because some people use analogies badly, is like believing that surgery would be better on average minus scalpels because some people wield them clumsily and get messy blood everywhere. 😉

        (Full disclosure: I LOVE analogies, I find them extremely helpful in “changing the mental viewing angle” on a problem or question, and probably couldn’t stop using them even if I wanted to. So long as all interlocutors commit to using them to elucidate and not obscure, and fairly acknowledge their limitations and discontiguous aspects, I think they are aces.)Report

      • Some (maybe most) of that data comes in the messy form of stories and analogies, and not neat mathematical columns, because for most of us that’s not how our brains work (Schilling possibly excepted).

        I’m not sure why you say that. Schilling seems to me to be the person *most* likely to come up with seemingly-unrelated-but-actually-valid connections to whatever the subject is.

        I LOVE analogies

        Most people do. That’s precisely because they replace something unfamiliar and uncomfortable with something familiar and comfortable. But that comfort may come at the price of obscuring important attributes of the thing that was making you uncomfortable in the first place.Report

      • Glyph in reply to Vikram Bath says:

        That’s precisely because they replace something unfamiliar and uncomfortable with something familiar and comfortable. But that comfort may come at the price of obscuring important attributes of the thing that was making you uncomfortable in the first place.

        It’s not a matter of “comfort” (well, it can be, but people who are forever convinced that the complicated can always be made simple can’t really be helped).

        Just as often, an analogy may be employed to go the other direction – to show that what seems familiar and comfortable, can be viewed as monstrous from another angle.

        Sure, the tool of analogy may obscure what you need to see, instead of throwing light on it.

        And a flashlight may blind you if you point it the wrong way, instead of at what you are trying to look at.

        You seem to be moving beyond simply acknowledging the (very obvious) limitations of analogies and their uses, to stating that discourse would (on average) be better off without it.

        Personally, I doubt we’d ever have gotten ANYWHERE without it.

        In a way, analogy is kind of like (wait for it) a rhetorical version of the scientific method. Analogy “cites” something that is assumed to be commonly-agreed-upon and understood, then uses that “cite” to “hypothesize” the “expected” outcome of the newer question before us.

        (“Well, we all know that pure water flows downhill; so what do we think syrup, which is similar in many respects, will do when we dump it out?”)

        Can that tool be misused? Sure, happens all the time. But without it we’d rarely figure out how to move forward, past

        Person A: I believe [X] because [Y].

        Person B: I believe [not X] because [P].

        Person A: No.Report

      • Brandon Berg in reply to Vikram Bath says:

        So long as all interlocutors commit to using them to elucidate and not obscure, and fairly acknowledge their limitations and discontiguous aspects…

        Doesn’t this basically mean “So long as everyone already understands the thing being analogized?” Analogies are great, as long as they’re not needed?Report

      • Patrick in reply to Vikram Bath says:

        I think if there is a disagreement, then you will find that there also is no consensus on the reasons that are and are not suitable for our common purpose.

        Well, that’s what we have the rulebook for.

        I’m okay with saying, “If your only actual objection is because of your feelings, then you really ought to examine whether or not it’s okay – in general – for people to demand legislation on account o’ the feels.”

        “If that’s okay with you, well… okay. But then in two weeks when someone starts arguing about their feels, and you start talking about principles, I’m not going to feel to sympathetic to your cause.”

        Many if not most of our culture wars boil down to “it’s a-ok for me to demand this because of feels and that because of principles, but it’s not okay for them to object to my feels on principles or my principles based upon feels”

        Which isn’t a healthy public debate. It’s just wanking.Report

      • Glyph in reply to Vikram Bath says:

        @brandon-berg – Well, sure, if everyone already understands what is being discussed then analogy is probably not needed. But if that’s the case, it probably wouldn’t be discussed at all.

        All I meant was, if I use an analogy and it’s a bad one (or, it breaks down past a certain point) and you point out that it’s bad or breaks down, and I admit yeah, it’s not perfect, we each still have moved closer to understanding what the other person is getting at (what their gestalt view is) and what the actual contours of the question (and points of contention) are.

        So long as we are both being honest about the tool and what we are trying to do with it (and what it even can do), it’s a perfectly cromulent tool. Vikram’s contention is that it wastes more time than it saves, and it can, but it doesn’t have to (and I would suspect on net that it probably saves time overall, by bringing useful informational contours to the surface).Report

      • Brandon Berg in reply to Vikram Bath says:

        @glyph What I meant was that it seems to me that you’d have to understand the thing being analogized to follow the rules you outlined.

        Maybe you’re right. It’s just that I see analogies being abused a lot. The worst is when someone responds to an analogy by extending it in some way that doesn’t make sense, and then the first person responds by trying to explain the right (or “right”) way to extend the analogy, and then they’re arguing about the analogy in ways that are of dubious relevance to the underlying topic. Extended analogies are the Nickelback of debate.

        But maybe the issue is that most people just suck at using analogies and don’t understand their limits, and that someone with good analogy skills isn’t vulnerable to this sort of thing. We should find these people and clone them.Report

      • Glyph in reply to Vikram Bath says:

        @brandon-berg What I meant was that it seems to me that you’d have to understand the thing being analogized to follow the rules you outlined.

        I guess my view is the way we even get (close) to a consensus understanding of the thing being analogized, is via an analogy (which one interlocutor moots, and the other accepts/rejects/modifies) to something the interlocutors both already (largely) have consensus on.

        Analogy is such a root bit of how we understand the world and each other that any attempt to weed it out is probably doomed.

        It’s just that I see analogies being abused a lot…(snip)

        Yeah, I am probably one of the worst offenders in this arena. I apologize, that’s got to be frustrating. But it’s just how my brain works.

        FWIW, I don’t mind people calling mine out when they are bad or incomplete, or proposing an alternate analogy; I appreciate it. It’s helpful to me.

        My dad was a CPA. He made his living making sure the numbers on the page went where they were supposed to go, and did what they were supposed to do.

        Yet he couldn’t help me with my school math homework, because he knew no way to explain the problem, other than to re-state only that which appeared on the page….which doesn’t help the person who doesn’t already understand what’s on the page.

        Symbols, of any kind, are always and only symbols; if you want them to be understood by someone who doesn’t already recognize them, sometimes stories are necessary (and in stories some truths will be lost or obscured, but others can be brought to light.)

        You know, it’d be an interesting experiment – next time Vikram does a post on something controversial, he should enforce a strict “no analogies” rule in the combox. See if people seem to understand each other better or not. That’ll tell us whether they are a net gain or a net drag.

        I know which way I’d bet.Report

      • next time Vikram does a post on something controversial, he should enforce a strict “no analogies” rule in the combox. See if people seem to understand each other better or not. That’ll tell us whether they are a net gain or a net drag. I know which way I’d bet.

        I’d be willing to try that. At least we wouldn’t be debating whether drinking to excess if you are a young woman is more like leaving your car door unlocked or more like leaving the keys in the ignition with the door unlocked, etc.

        Of course, it’d be nice if we had some sort of control group.

        Not to put you on the spot, and it’s totally understandable if you don’t have any ready examples, but can you point to a place online where two people have disagreed, one person gave an analogy, and the other person then got it and it clarified things for them and they then agreed?Report

      • Mike Schilling in reply to Vikram Bath says:

        Schilling seems to me to be the person *most* likely to come up with seemingly-unrelated-but-actually-valid connections to whatever the subject is.

        That’s because everything is like baseball.Report

    • Jim Heffman in reply to Patrick says:

      “Why? This is the thing in the fungible-money v. insurance conversation that I haven’t been able to get anybody to articulate. Other than “it just is”.”

      If it doesn’t matter whether the money comes from the employer or the employee, then why is there a legal mandate that it come from the employer?Report

      • Kim in reply to Jim Heffman says:

        There isn’t. The company can elect to not insure its workers, in which case that Costs Them Money (in that they get a tax break for insuring workers). HL wanted the best of both worlds — to get tax break,and not end baby Jesus’s life.Report

    • James Hanley in reply to Patrick says:

      This is the thing in the fungible-money v. insurance conversation that I haven’t been able to get anybody to articulate. Other than “it just is”.

      That’s quite a statement, @patrick. I’d have no trouble believing you if you said nobody had made an argument that persuaded you, but to say nobody said anything more than “it just is” sure isn’t how I remember those discussions.Report

      • Patrick in reply to James Hanley says:

        This post of yours is as far back as I was willing to dig at the moment.

        I don’t see anybody offering much of anything assigning culpability to the employer. Indeed, almost all the contributions assigning culpability are assigning it to the employee… some arguing about the particulars, but nobody switching it around.

        Granted, that’s only one of a lot of different posts on HL (and that one is not directly about Hobby Lobby), but I don’t remember anybody laying out an affirmative case for *why* the employer bears culpability other than them asserting that it does.

        Which, okay, granted, you can say, “I feel culpable, but that’s not much more articulation than “it just is”.Report

    • zic in reply to Patrick says:

      I have repeatedly tried to come up with medical analogies that draw the distinction between women’s health care (not just contraception) and some necessity of men (that isn’t an assured necessity); and there is no analogy; not even one that provides enlightenment.

      On this exchange:

      Going back to birth control, buying insurance for an employee that covers some particular thing is different than paying an employee fungible dollars she uses to buy that same thing.

      Why? This is the thing in the fungible-money v. insurance conversation that I haven’t been able to get anybody to articulate. Other than “it just is”.

      the thing isn’t how the money is paid, compensation as wages or as benefits; it’s keeping a woman’s health-care transactions within that transaction system we have in place. I could care less who pays for the coverage; I just want her to treat it like all her other preventive medical needs; that is a completely reasonable requirement. If you make it different due to your own moral judgements, you are imposing those judgements on her; segregating her care in a discriminatory fashion. To offer analogies (just to tease Vikram,) it’s the same underlying civil-rights violation as ride on the back of the bus or civil unions but not marriage, or neighborhood red-lining; treating one class of people as different.

      I know there are those who disagree, too. But if something about that coverage bugs an employer, the burden of that disturbance should not fall on the employee.Report

      • veronica d in reply to zic says:

        @zic — Right. This. No analogy can ever be given to white, cis men because they do not live under oppression. To them, oppression is an alien experience, and to understand, they would have to be willing to set aside their preconceptions, their self-interests, their self-invested sense of privilege, and to actually listen. But then, they don’t need to listen, since naked power means they already win, and the barest effort will hide naked power behind the operations of the status quo, which all feels very nice to the self-satisfied white, cis man. And on it goes.Report

      • zic in reply to zic says:

        @veronica-d

        To them, oppression is an alien experience, and to understand, they would have to be willing to set aside their preconceptions, their self-interests, their self-invested sense of privilege, and to actually listen.

        I actually think many (if not most) men want to do this, and try to do this much of the time. I know men who self-check, who seek better understanding. Analogy is an important part of that process, since they can not be not-man, unless they go through transition. That we’ve come as far as we have is testament to men’s willingness to empathize with women, to allow women access to the full privilege of adult human.

        But misogyny is baked so deeply into our culture that it’s often hard to recognize and root out, and different people have different levels of comfort with change. Rooting out misogyny is change. I can understand the discomfort.

        But the misogyny itself is still wrong. I’m discomforted presuming it the default position for most men. But I also think most men (and probably most women) don’t understand ‘equal’ to some degree, simply because unequal has been the tradition for most of history.Report

      • veronica d in reply to zic says:

        @zic — I dunno. I run into too many men who are like brick walls. They can’t see through any lens but their own.

        For example, recent events ’round these parts.Report

      • Jaybird in reply to zic says:

        You know, there are sub-groups within the white cishet men. Perhaps analogies like “you know how the jocks treat the burners? The nerds? The av-club? It’s like that” might work.

        Well, not for the jocks, of course. They’re stupid to the point where they’re unable to understand anything that isn’t related to their balls.

        But the former burners, the former nerds, and the former av-club kids might be made to understand exactly how poorly they’re wielding the naked power in their hands.Report

      • veronica d in reply to zic says:

        @jaybird — First, I’m bored with #notallmen. I know that already. Second, the former nerds? Like these guys: http://wehuntedthemammoth.com/2014/09/08/zoe-quinns-screenshots-of-4chans-dirty-tricks-were-just-the-appetizer-heres-the-first-course-of-the-dinner-directly-from-the-irc-log/

        You know, those big old drive a woman to suicide unless we can rape her first nerds?

        Nice fella. They seem primed to understand.

        Well, I guess #notallnerds. Obviously.

        You know, I’m really not in the mood for this right now. I’ll say one thing: if you think you get this stuff, maybe you kinda halfway do. It is possible to listen and get it right. But check in a lot. Ask, “Hey, how am I doing?” And don’t think the feedback you get is something to argue about.

        And if the woman just fucking gives up and goes away, I guess you win. Or something.Report

      • veronica d in reply to zic says:

        I think those will be my last words on this forum.Report

      • zic in reply to zic says:

        @veronica-d

        I hope not.Report

      • Stillwater in reply to zic says:

        I think those will be my last words on this forum.

        Damn. That sucks.Report

  2. Mad Rocket Scientist says:

    Let’s be honest, people generally suck at forming analogies. It’s like if you were to…Report

  3. kirk says:

    For what it’s worth, I’m like the guy in the XKCD cartoon, complaining that the rubber-sheet analogy makes no sense. Every time some science show uses it to illustrate curved space I groan.Report

    • Vikram Bath in reply to kirk says:

      For me, the issue is which way is it being pulled down. I get that mass creates a “gravity well”, but the pictures of it assume that there is a down direction, and that direction is only that way because that’s the way they drew it.Report

    • Boegiboe in reply to kirk says:

      The rubber sheet analogy is a very good one, because it demonstrates deforming a two-dimensional sheet in the third dimension. This can allow a student to get their head around the idea of spacetime being deformed if they hadn’t managed that before.

      Once the student has absorbed the idea that spacetime is deformed by matter and energy, they can later discard the rubber sheet in favor of some analogy or equation that works better.Report

      • Jim Heffman in reply to Boegiboe says:

        In fact, the rubber sheet analogy is entirely accurate when you consider that it’s a two-dimensional space (the rubber sheet) contained in a three-dimensional universe (ours). There is no way for people in the two-dimensional universe to perceive three-dimensional deformation, so it looks to them as though mass has some mysterious attractive power.Report

      • morat20 in reply to Boegiboe says:

        Visualizing past 3 dimensions is rather difficult for most people. You need lots and lots of math that you understand cold, and frankly aside from a few geniuses I bet most of the people doing real work in multi-dimensional space either do it entirely mathematically (and thus don’t try at ALL to visualize it, other than whatever mathematical representations are in use) or use some sort of internal analogy or mental shortcut.

        There’s a reason that classic mechanics problems are often done using word problems — a tank in position A moving at such and such a vector way fires a projectile at a tank B (located here and moving at this vector). Given a muzzle velocity of V, what angle should the turret be elevated at?

        Because people can model that in their heads, although the solution is generic to any moving body problems. (Well, those not involving ridiculously high gravities or ridiculously high speeds).Report

      • Kim in reply to Boegiboe says:

        morat20,
        and reducing dimensions (or solving in one dimension at a time) is a standard math/physics trick.Report

      • Brandon Berg in reply to Boegiboe says:

        But the issue is that it’s not clear why spacetime deformation should lead to attraction. The rubber-sheet analogy assumes an unexplained downward pull.Report

      • Mike Schilling in reply to Boegiboe says:

        they can later discard the rubber sheet

        Once they’ve been physics-trained.Report

      • Mike Schilling in reply to Boegiboe says:

        The rubber-sheet analogy assumes an unexplained downward pull.

        It doesn’t really. The analogy would also work if the sheet were deformed upwards. Think of the rubber sheet as the entire world, so that the marble rolling on it has no choice other than to stay put on it. The marble can’t either rise above it or fall below it, because above it and below it don’t exist.

        A marble rolling along a flat sheet will move in a straight line, unless force is applied to it or it encounters an obstacle. When it reaches a gravity well (or gravity peak: as I said above, either one works), it also continues to travel in a straight line suitably defined; that is, it takes the shortest distance between any two points it passes through. In three dimensions. If you look at the path in two dimension, it appears to be a curve, which isn’t the shortest path. But the alternative “straight line” which seems shorter in two dimensions is actually longer in three.

        To use yet another analogy, think about airline routes. The great circle routes, which really are the shortest distances between two cities, look longer on a standard Mercator map, but that’s because the Mercator projection distorts distances when translating the earth’s three-dimensional surface into a two-dimensional representation. The rubber sheet analogy says that our idea of space as three-dimensional distorts the more general idea of distance in four-dimensional space affected by gravity fields.Report

      • Jim Heffman in reply to Boegiboe says:

        “But the issue is that it’s not clear why spacetime deformation should lead to attraction. The rubber-sheet analogy assumes an unexplained downward pull.”

        The intent of the rubber-sheet analogy is not to explain how gravity happens, it’s meant to explain what gravity does.Report

      • kirk in reply to Boegiboe says:

        Absent something pulling down on it, placing a ball on the sheet would cause absolutely nothing to happen. The sheet would still be flat, and nothing would go towards the object.

        So, the analogy is bogus.

        As for the reality behind the analogy (curved space? gravitons?), I doubt that even 1/10,000 people understand any of it. Lord knows I don’t. Frankly, I’m skeptical that anyone here does.Report

      • Kim in reply to Boegiboe says:

        Kirk,
        you’ve got at least two people currently working in physics-based work, and I’m also trained as a physicist. I understand the analogy (mostly through understanding black holes, and gravitational lensing).Report

      • Jim Heffman in reply to Boegiboe says:

        kirk,
        you are the guy the xkcd comic is about.

        the point of the rubber sheet analogy is not “here’s what makes gravity happen”, the point of the rubber sheet analogy is “here’s an illustration of how the apparent attractive force inherent to mass is actually the result of a higher-dimensional distortion of three-dimensional spacetime”.Report

      • Mike Schilling in reply to Boegiboe says:

        Frankly, I’m skeptical that anyone here does.

        It’s not rocket science. Which works out to be a point in your favor, because two of the commenters on this thread are honest-to-God professional rocket scientists.Report

      • Mad Rocket Scientist in reply to Boegiboe says:

        Thank you Jim.Report

      • morat20 in reply to Boegiboe says:

        Frankly, I’m skeptical that anyone here does.

        Einstein’s stuff is a century old. It’s well trodden ground. Anyone with an undergraduate’s degree in physics understands the stuff the rubber sheet analogy is all about.

        It might have been cutting edge 100 years ago, but Newton was cutting edge in his time (he invented entirely new forms of math to do his work!). I learned Newton’s bleeding edge material in high school, when I did projectile motion and Cal I.Report

  4. Wardsmith says:

    Every time I read the word analogy I stop at the ‘g’ and read analog. As opposed to digital that is. Analog measurements are always estimates while the digital meter gives you a very precise number. Of course the digital meter is /always/ wrong, because we live in an analog world. 🙂

    Using analogs to help describe something means we are purposefully obfuscating the underlying precision of the original. Cartoonish images often work better than HD photographs at communicating the fundamental point, because with all that precision comes distraction.Report

    • Kim in reply to Wardsmith says:

      the digital reading is always wrong because you change the system by observing it.Report

    • James Hanley in reply to Wardsmith says:

      Well said, @wardsmith.Report

    • Chris in reply to Wardsmith says:

      Good analogy, dude.Report

    • zic in reply to Wardsmith says:

      Nice, @wardsmithReport

    • Murali in reply to Wardsmith says:

      @wardsmith
      I thought it is the other way around. Digital is necessarily imprecise because the waveform is either rounded up or down in order to get either 0 or 1. The original signal is always in analogue. Think about it this way: early analogue methods of recording/capture were imprecise not because analogue is necessarily imprecise, but because our recording methods were crude. Our human brain has a difficult time dealing directly with very highly specific quantities. So, when we look at an analogue ammeter, we have difficulty locating the precise position of the needle. Digital ammeters appear to be better only because we have better sampling methods. This allows us to get an approximation that can be presented as a series of numbers which have a greater degree of accuracy than we could detect with our eyes and the needle. But this should not fool us into thinking that analogue devices are the ones doing the approximation. It is us who do the crude approximation from the very accurate analogue signal. The digital signal is by its very nature inaccurate because it is an approximation of the true analogue signal.Report

      • Chris in reply to Murali says:

        Precisely. Arguing that digital has more fidelity than analog implies an understanding of neither.

        Analogies are such a central part of the way we think because they help us to see structural relationships. They don’t lose fidelity, they add depth by highlighting what might not have obvious.Report

      • Chris in reply to Murali says:

        I imagine what Ward is thinking of, in his analogy between analogy vs. something (literal representation?) and the analog-digital distinction is rather simple analogs, like a thermometer that represents temperature by the volume of mercury vs. a digital thermometer that can give you relatively precise readings without doing much work. But of course, the mercury thermometer is more accurate if you look at it close enough, because it isn’t limited to a pre-set level of precision, a specific number of digits after the decimal point. If you could measure it accurately, it would give you very fine measurements. Any increase in precision of a digital thermometer can only approximate the precision of the analog thermometer if you measure it well enough. A digital thermometer is just a lazy approximation of an analog thermometer.Report

      • dhex in reply to Murali says:

        “They don’t lose fidelity, they add depth by highlighting what might not have obvious.”

        i wouldn’t say fidelity is the core reason behind analogies; persuasion is.Report

      • Chris in reply to Murali says:

        I was discussing Ward’s analogy, for which fidelity was at issue.

        Now, this context the purpose of analogies is to persuade, but that’s merely one use of analogy. At the highest level of abstraction, the purpose of analogy is to compare one domain to another in order to transfer knowledge about relations from one to the other (or in some cases, from both to each other), in such a way that it facilitates inferences about the domain to which the knowledge is being transferred. They can be used to understand, to teach, to persuade, to clarify, to create, and all the antitheses of those things if the user of analogy is dishonest or mistaken or careless.

        There is a huge literature, the very tip of the tip of the tip of which I’ve linked in this thread, about how analogies are used (including how they are used to change people’s minds in ways that facilitate knowledge and creativity). Like any cognitive and rhetorical device, they can distort and deceive, but they are no more prone to that than any other, and their usefulness, and their naturalness, makes them incredibly important and powerful tools for communication and understanding.Report

      • Wardsmith in reply to Murali says:

        @murali
        What Chris said below. I thought I was being cute by saying the digital was more precise but /always/ wrong. Apparently cute doesn’t cut it for your precision meter 😉

        I would explain more but typing on an ipad is frustrating at best. Maybe when I get to a real keyboard I can flesh this out.Report

      • Murali in reply to Murali says:

        @wardsmith

        *Sigh*

        obviously[1], if someone is going to be cute by making an analogy for analogies, it is appropriate to post a response questioning the propriety of the analogy.

        [1]so, not so obvious in hindsight?Report

    • Mike Schilling in reply to Wardsmith says:

      That is excellent.Report

  5. nevermoor says:

    Unsurprisingly, given my profession, I find this argument completely unpersuasive.

    When people who disagree are trying to convince a neutral third party, it’s hard to see what would be more effective than saying “we all know that situation A has outcome X, and our situation is like A, so outcome X is correct.” Obviously there will be a retort, usually explaining why our situation is more like B than A, but at least then you are examining the details of the situation at hand within a framework that leads to one outcome or another.

    I think it’s telling that the argument the OP describes as “brilliant” is purely conclusory. I agree it is a description of the central issue that includes Megan’s opinion, but it has exactly zero persuasive power. If you disagree with her, you simply say “there is no way that Caterpillar could not have reasonably known what the IDF was going to do with that bulldozer, and hence, no grounds for opposing a suit.” And we are nowhere. You get somewhere by saying “C corp sold a similarly common product and was found liable for reasonably anticipatible misuse” or “D corp sold a similarly common product and was not found liable” and then examining the connections between those precedents and your case.

    This, of course, does not mean all analogies are good. “X is like naziism” is almost always a bad analogy that does have the effects described in the OP. I’m just not sure how you can make a convincing argument without analogizing your facts to resolved precedent/analogy.Report

    • Wardsmith in reply to nevermoor says:

      What is your profession if I might ask?Report

      • Patrick in reply to Wardsmith says:

        Professional Analogist?Report

      • Glyph in reply to Wardsmith says:

        Hopefully he’s not a professional twice over— an analogist and a therapist. Those business cards can almost get you arrested.Report

      • Kim in reply to Wardsmith says:

        glyph,
        and people say rape jokes are never funny.
        (disclaimer: making fun of people who have been raped is probably going to get you punched. Do not recommend).Report

      • Glyph in reply to Wardsmith says:

        You know, I’m not even much of an Arrested Development fan, I stopped after the first season or two. But I am a David Cross fan.

        I don’t even know that that was a “rape joke” so much as a “not thinking things through when you are making your business cards” joke – does the appearance of a word de facto make the joke “about” that word (or more precisely, about the act which the word names)?

        I mean, it does in some sense, since the word has to be *something* that hits a nerve, but in another sense the joke isn’t really about that word or the named act at all, but about a hapless character who would inadvertently do something so clueless (see also, Curb‘s “Beloved Aunt” episode).Report

      • nevermoor in reply to Wardsmith says:

        Attorney.Report

    • Burt Likko in reply to nevermoor says:

      I suspect that the OP’s use of the word “Brilliant” to describe this argument is intended to express relief that McArdle has dispensed with questionably apposite analogies and instead directly addressed the actual subject of the discussion, rather than to characterize the strength of the argument itself.

      The argument thus described assumes the truth of its factual precedent: had McArdle’s argument been phrased “… if it is true thatthere is no way that Caterpillar could have reasonably known what the IDF was going to do with that bulldozer, and hence, there would be no grounds for a suit” and the author would have found it equally “brilliant” because it’s no longer talking about Zyklon B or Fidel Castro, but instead is squarely addressing the subject matter itself: whether Caterpillar is morally blameworthy for doing business with the IDF.Report

      • nevermoor in reply to Burt Likko says:

        Maybe. But how do we know that? Because Megan says so? Because other companies that did not know were not subject to liability?

        Stripped of both support and opinion, it becomes an OK sentence for an introductory paragraph, assuming you can support it as a frame for the argument. Which I have trouble seeing someone do without analogy.Report

      • Murali in reply to Burt Likko says:

        You appeal to an even deeper principle. Vickram’s point is that analogies seem to do a lot of convincing without doing any justifying. All the apparent work that is done with analogical arguments is done by considerations of whether two cases are relevantly similar. That is to say, the analogy works only if the details according to which the cases are distinguishable from one another are irrelevant to the argument being made. However, the relevance/irrelevance of these details is not argued for. These assumptions are often illegitimately smuggled in. Any adequate explanation for why some details are relevant and others aren’t would solve this problem, but also make the analogy unnecessary.Report

      • nevermoor in reply to Burt Likko says:

        This seems backwards to me. The point of an analogy is to look to something that everyone agrees on and connect it to (at least part of) the thing you’re discussing.

        If done at all well, you’re at least narrowing the discussion and at best resolving it. If done poorly, you’re doing it poorly. No method is immune to poor application. So I agree that if you just say “A is like B” you haven’t convinced anyone. But no single sentence could, certainly not the Megan sentence quoted in the OP.Report

  6. I actually think that argument by analogy tends to be a more productive form of argument than other forms. It’s not necessarily all that likely to persuade any given interlocutor on any given issue, but that’s true of all forms of argument. It’s generally quite rare to successfully change someone’s mind using any form of argument. Instead, successful persuasion often times can be a matter of trial and argument, because we never know which form of argument is most capable of persuading a given person or clicking with a given person – if any – and what works to persuade one person may not be persuasive to another person (and vice versa).

    To be sure, there are some arguments by analogy that are horrible, but that’s true of any form of argument, and I’m not at all convinced that horrible arguments by analogy are any more likely than horrible uses of other forms of argument.

    But a well-constructed argument by analogy can, I think, be more likely to be effective than well-constructed arguments in other forms. A well-constructed argument by analogy makes your position more relatable to your interlocutor and, ideally, helps them see where you’re coming from or at least why you think your position is important.

    While there are always differences, those differences, too, can be a source of productivity. Indeed, they just about always should be a way of making a discussion productive as long as the participants in the discussion are willing to let the discussion go on beyond an initial assertion of “these things are different.” A response of “these things are different because” should immediately be followed with “why do you think those differences are important and relevant?” An honest response to that question immediately brings the real issues better into focus, allows the person who originally asserted the analogy to have a better understanding of where the disagreement really lies (and in some circumstances may even get the original analogizer to see where his own position is flawed or even wrong). Simply put, a well-constructed argument by analogy should at least narrow the issues even if it doesn’t succeed in persuasion.

    Even if the response to the argument by analogy is simply “these things are different,” and the interlocutor refuses to explain why they’re different and why those differences are relevant or important, you at least quickly learn that your interlocutor is simply uninterested in discussing the issue so you can stop wasting your time.Report

    • Stillwater in reply to Mark Thompson says:

      Thanks for this comment Mark. That’s how it seems to me, too.Report

    • Murali in reply to Mark Thompson says:

      The thing about deductive and inductive arguments is that there is a way to formally specify how the conclusion follows (either necessarily or probabilistically) from the premises. Analogical arguments are less about reasoning than persuasion. As a result, it may have its place in the lawyer’s toolkit, but it really shouldn’t be there in the scientist’s or philosopher’s tool kit.Report

      • Road Scholar in reply to Murali says:

        @murali, As a result, it may have its place in the lawyer’s toolkit, but it really shouldn’t be there in the scientist’s or philosopher’s tool kit.

        It can certainly serve as a means to help you get a grip on a problem. I was just reading an article where a scientist was working on a problem involving fluid vortices in the ocean. He noted that they had a superficial resemblance to black holes in space. Investigating further he found that the same equations governed the flow of water around the perimeter of a vortex and the path of photons around the event horizon of a black hole. This gave him the crucial insight that led to a method for detecting and mapping vortices from satellite data.

        The purpose of analogies isn’t to serve as a link in a chain of logic, but instead to jumpstart the exploration of a new problem domain given known results in a similar domain.Report

  7. Jim Heffman says:

    The purpose of analogy, when used in argument (rather than the xkcd comic’s example of education) is to recontextualize the argument and to illuminate the flaws in reasoning.

    I might say “global warming is a complicated subject, and I believe that you’re drawing the wrong conclusion from the many complex facts. Let me present a similar set of conditions using a situation where the relevant physical processes are easier to understand”.

    Or I might say “you’re an idiot and your conclusions do not follow from your evidence. Let me present a similar set of conditions using a situation that is different in particulars but similar in most ways, and tell me whether your conclusion is still the same. If not, then you must consider what differences exist that cause your conclusion to be valid in one place but not in another.”Report

    • Jim Heffman in reply to Jim Heffman says:

      “Taking nude selfies with your phone is not the same as doing various things with your house keys, and no amount of refining or modifying the comparison will suddenly make the analogy insightful for those who disagree with you.”

      The intent here is to reply to people who claim that calls for personal responsibility are “victim-blaming” by showing a similar situation (a person who engages in risky, not-secure behavior) and asking whether the same reasoning applies (pointing out the risky behavior is accusatory toward the victim and exculpatory toward the criminal, and the one who points it out thinks the victim deserved what they got).Report

  8. zic says:

    Perhaps there are two types of analogies: the unusual event, “he drove that car like he was blasting off in a rocket,” emotionally indicates what was going on, but it obscures because very few people have experienced lift off. Where it informs, it does so because most of us have probably imagined lift off, and riding in a rapidly-accelerating car is one of the most common experiences of g-force.

    Upthread, @wardsmith spoke of the word “analog” at analogy’s heart; and the importance is painting emotional significance — not necessarily the same, but evoking similar things in some limited, analog way. This is a very important human skill — pattern recognition, where one pattern produces a certain outcome, so we intuit that another, similar pattern might produce a similar outcome. A huge learning-curve advantage; though the patterns won’t always hold, if enough of them do, you don’t have to learn the whole system over for each; you can predict potential results.

    But it is a gross approximation, too; not a scientific result. A guestimate.Report

    • Jim Heffman in reply to zic says:

      “he drove that car like he was blasting off in a rocket”

      …is simile, not analogy. Analogy requires a “therefore” because the intent is to demonstrate the generality of a particular chain of reasoning. (Or, through showing how a particular chain of reasoning leads to ridiculous results, to demonstrate the invalidity of that reasoning.)Report

      • zic in reply to Jim Heffman says:

        Hmmm. That simply seems like an attempt to discredit what I’ve said without actually addressing what I suggested — that analogical thought is crucial to intuitive thinking. Sorry, @jim-heffman, since you point out that my example is a simile, I’d be remiss not to point out that a simile is one type of analogical language. There are others — metaphor, exemplification, comparison, parable. All help us understand one thing by drawing off the knowledge we’ve already gained from another.Report

      • Jim Heffman in reply to Jim Heffman says:

        Simile is required for analogy, but analogy is more than simile.Report

  9. Tod Kelly says:

    I think you’re attempting to shove analogies into boxes they weren’t built for and declaring them faulty for it. Analogies are not really a way to score debating points so much as they are a way of communicate perspective.

    Let’s say Commenter X and I argue over the merits of legalized SSM. Regardless of my arguments Commenter X has a working assumption with why I support it, and that assumption in turn feeds into his own opposition — maybe he thinks I’m afraid of the PC police, maybe he thinks it’s because I hate religious people, maybe he thinks I’m a Libertine. During our discussion, I will probably explain to him why, to me, the reasons I feel he is on the wrong side of SSM are the precise reasons why I feel this country was on the wrong side of interracial marriages a generation or two ago.

    I do not use the interracial marriage analogy because I think that skin pigmentation is the same thing as sexual orientation; nor do I use it because think Harvey Fierstein and Samuel L. Jackson are interchangeable on any level. And I certainly do not expect Commenter X to accept my analogy as something akin to the Lord’s Gospel or a mathematical proof. I use the analogy because it communicates a very deep truth from my perspective. Commenter X is free to accept or reject that perspective, of course. However, the hope is that through the use of analogy I can get him to stand and look at the issue from a slightly different place than he was standing prior — and that this might at best get him to see things my way, or at worst understand why I see them that way.

    This seems to me to be the very opposite of something to be avoided when people who don’t see eye to eye take the time to discuss their disagreements.Report

  10. Jaybird says:

    I’m trying to think of examples where analogies end up doing more harm than good.

    The only example that I could come up with was “Religion” and institutionalized attempts to immanetize the analogy.

    Surely my imagination is limited.Report

  11. ScarletNumbers says:

    I think that crafting intelligent, germane analogies is a sign of verbal intelligence.

    I think that not liking analogies is Aspbergery.Report

  12. Stillwater says:

    and no amount of refining or modifying the comparison will suddenly make the analogy insightful for those who disagree with you.

    Exactly. For some people being presented with an analogy feels like being bashed in the head with the very same opaque surface limiting their vision.Report

  13. Kolohe says:

    OK, Mr. Bath, you win

    Homework
    Directions: Now that we have read about two men of power who abused their power in various ways, we will compare and contrast them and their actions. Please refer to your texts, “Fighting Hitler – A Holocaust Story” and “Bush: Iraq War Justified Despite No WMD” to compare and contrast former President George W Bush and Hitler. We will use this in class tomorrow for an activity!

    Report