The case for and against analogies | Hacker News Reader

The case for and against analogies

dynomight.net

73 pointsggoo59 comments

My basic rule is that analogies are great for a sympathetic audience and bad for an antagonistic one. If you're trying to illuminate something to someone who wants to understand your point, an analogy is a great shortcut to building a tangible architecture in their mind. But if the other person has an opposing viewpoint and you're trying to dismantle that and replace it with yours, analogies usually end up just moving the goalposts in the argument.

This is brilliant.

I’d never thought about it such explicit terms, but of course using analogies with an unsympathetic interlocutor is basically building them a all-too-convenient strawman

That's a good rule. People on this site often make arguments using analogies, and those threads always devolve into arguments about the specifics of the analogy. I have an amazing screenshot I saved where the discussion turns into something about companies hiring clowns to yell at patrons. It's unintentionally comical. The participants are intelligent adults, talking about a domain that they all share some understanding of, yet the crux of their debate is the behavior of some obnoxious clowns.

Ultimately every debate boils down to the behaviour of some obnoxious clowns; it is analogous to, say, entropy.

> People on this site often make arguments using analogies, and those threads always devolve into arguments about the specifics of the analogy.

That is because analogies are only useful for illustration, not as a chain of reasoning to prove something.

Conversations almost universally become low quality once analogies are brought in. They should be used sparingly, and explicitly only for the purpose of illustrating an idea.

Software development is like building a house …

It's so common to see people arguing with an analogy, which should be obviously pointless. Proposing an alternate analogy is slightly better, but really you should just discuss the actual issue in front of you, it's much clearer.

I think analogy is useful in an argument for a few reasons:

1. It can reduce the problem space to the factors which the parties see as important.

2. It can uncover differences of opinion as to which factors are deemed important to the problem space. ie: "yeah, but your analogy discounts x!".

Of course, analogies can be used in bad faith as rhetorical tools to win an argument on grounds other than it's merits, and that is the real problem with them.

thats a good observation, yes, an antagonistic audience is likely to derail your point by focusing on the ways in which the things compared are dissimilar, even though analogies only compare the ways they are similar and requires the listener to figure out which ways those are, while ignoring the dissimilarities.

I often just mention that definition. I actually get really far by avoiding certain polarizing words and just saying the definition of the word. Flies right under the radar.

Yep, "false equivalence" is the response I always get from an antagonistic audience when I try to use an analogy. Of course an analogy is not an equivalence, but that gets conveniently ignored. These days I just make the most direct argument possible and avoid analogies altogether, or else we end up in an argument about what an analogy actually is. Analogies work a lot better during a disagreement if the person you're disagreeing with is operating in good faith, which unfortunately most humans aren't able to do when their views are being challenged.

(I started posting the below as a top-level comment, but in the end it arrived at the same point as yours, so leaving it as a reply…)

In Indian/Sanskrit literary theory (poetics), in the discussion of figures of speech (rhetoric, etc), similes are called upamā ("her face is like the moon", etc). The discussion of it in the literature is extensive and would fill several volumes (and I hardly know anything), but one thing recognized early is that in a simile/analogy, there needs to be a sādharaṇa-dharma, a shared property: the point is that there's something in common ("her face is beautiful, like the moon") while of course there is going to be a lot that is not (the intended meaning is not "like the moon, her face is pockmarked, full of craters", etc). In any given instance, this intended shared property may either be stated explicitly, in which case the simile is called "complete", or left implicit, in which case it's called "partial". Both can be highly effective.

In the context of this post, it seems that when making an argument, it may help for the analogy to be "complete": actually spell out what you intend by the analogy. Also, to echo the comment by @munificent, it is worth remembering that figures of speech are very effective as poetry, but poetry requires a sensitive and sympathetic reader (sahṛdaya). The same rhetoric can fail when used in argument with an unsympathetic audience.

The other thing to remember about analogies is that some people can be good at them, and some people can be bad at them.

They appear on IQ tests. Some are easier than others.

When making an analogy, you should try to work out if it’s an easy one or a hard one, to judge how much work you’re making your audience do.

Big +1. I think this is a really good distinction to keep in mind in general. Teaching and debating look very different.

E.g., it's going to come off as condescending if you try to "educate" someone on something where they already have a formed opinion and don't see you as an authority.

> analogies usually end up just moving the goalposts in the argument.

I appreciate the self referential nature of using an analogy to describe analogies.

Also "illuminate", "shortcut", "building a tangible architecture", and "dismantle and replace". :)

It doesn't need to be sympathetic vs antagonistic. I think the better dichotomy is scout mindset vs soldier mindset. Are they actively trying to change their beliefs or are they defending some notion? If the person you're talking to is in a scout mindset an analogy would be helpful, but to a soldier an analogy is like red meat. It's the perfect fodder for both sides to dig in and assail each other which is what they really want whether they realize it or not.

The case against analogies is like trying to teach something to someone without any kind of frame of reference. A well-crafted analogy is frequently one of the best tools a teacher (or lecturer on technical subjects) has to make an abstract concept relatable to the student/listener. To the article's point, however, a really good analogy is one where likeness of the things being related are readily apparent, or at least easy to explain. The example of Karen (or was it Beth?) conflating X::Y to A::B is a lack of apprehension on her part. This is more an issue of communication and lack of precise understanding; once that's cleared up Karen will get it rather quickly.

> The case against analogies is like trying to teach something to someone without any kind of frame of reference

If this was intentional... nice.

A bad analogy is like a bad analogy.

It was. :-)

In my opinion analogies should almost always be avoided. It is much more work to understand the analogy and map it back to the argument and if you are proposing the analogy you're asking the other debate partner to do all that work for you for free. Also you're asking him to be very gracious with your argument because of course the analogy is never a perfect fit. All while he's trying to win the same debate you are trying to win. So this is a recipe for anger and resentment. In 99.999% of cases, you should be able to explain something directly without the indirect approach via an analogy. Analogies may help in a teacher-subject system where both people know their role and the teacher is explaining very vast subjects very superficially.

I agree. I almost always feel that analogies make sense to the person saying them, but not to the person receiving. You often need to already know the topic at hand to understand why the analogy is relevant, but if you’re new to the subject it feels just confusing.

> So this is a recipe for anger and resentment.

I know this is a common idiom and makes complete sense in context, but I must point out that it's also an analogy. I just thought it was genuinely funny given the first sentence of your comment. Analogies are hard to escape! :)

My personal pet peeve is analogies that have no basis in reality: “It’s like sleeping on a cloud”. You’re trying to relate your experience to something that no one has ever experienced.

Consider the phenomena of the most formidable linguistic infestation of any language in the history of noises beyond clicks, grunts and twitching - the word: L I K E.

It's the omni analogy, used to equivocate everything conceivable, from punctuation, confusion, concrete or abstract meaning, nothing, something, to anything deemed unworthy of minimal effort or patience. Adverbs be damned and banished to hell!

We've entered a paradigm where everything is a similarity of itself and nothing is what it is. "How do you feel?" , one might inquire. Only to be informed that the feeling bears similarity (l i k e) to something else. "I feel l i k e shit", one may say. Certainly not shitty, but approximate to shit, with no defining parameters.

It is an age come where nothing can be what it is, but only a reference to another thing or its similar self. A linguistically programmed virtual reality, where no one need speak precisely if the smoke of ambiguity and mirrors of distortion can substitute.

Now we can have our thing that resembles a cake and do something similar to ingesting it too.

Horray.

>Take the analogy between worrying about artificial general intelligence (AGI) and worrying about traffic on Mars. This could be restated as “Worrying about the rise of AGI is silly because AGI isn’t going to happen for a long time, and any effort put into solving it is wasted.” Is that correct? Maybe, I don’t know! But what does the analogy add over a direct argument?

It makes the remoteness and silliness of the concern more apparent.

Note that this doesn't mean that the actual "rise of AGI" must actually be a silly notion or remote in time, for the analogy to be good.

The analogy is used to communicate what the person making it thinks is the case. If they want to communicate effectively that the "rise of AGI" is a silly concern for the far future, then it's a good analogy.

Whether in reality the "rise of AGI" is a serious and pressing concern, is orthogonal to how good the analogy is.

I don't know why the author seems to think that this analogy would be presented without further argument. It would be more like “Worrying about the rise of AGI is silly because AGI isn’t going to happen for a long time, and any effort put into solving it is wasted. Are you also worried about traffic on Mars?”

Maybe they are just someone that was really bad at the "X is to Y as A is to B" type SAT questions

It’s not good [use of] analogy because it doesn’t add anything. Everyone already understands the concept of problems being less urgent as their arrival is farther away in the future. They don’t need an analogy to understand that. Most often, you’d be responding to a strawman by acting like the core dispute involves your opponent not understanding that.

Yes, if you already accept that AGI dangers won’t be relevant until far into the future then that justifies a low priority.[1] But that’s exactly what’s under dispute, and the analogy doesn’t help resolve disagreement over how far it is!

[1] If the AGI alarmist accepts that it’s far off, the the dispute would be about something else, like the time-discounted danger. But in that case, the analogy still doesn’t work because the purported harm of AGI is much worse than that of traffic, and if you disagree about that, then, as before, the analogy doesn’t help resolve it.

>It’s not good [use of] analogy because it doesn’t add anything. Everyone already understands the concept of problems being less urgent as their arrival is farther away in the future. They don’t need an analogy to understand that.

You've missed the point though. This analogy is not there to make them understand that - it's there to help them feel it.

The analogy helps pin them to something the other can immediately feel as an absurd worry.

Abstract understanding is not the only thing an analogy is good for.

So…

- They already agree you should discount concerns in proportion to how far they are unto the future.

- You’re already aware of this agreement.

- You shift the focus of the debate in a way that makes you look right about something (discounting future worries) even though your opponent was equally right.

- All because, though they never wrong on that point, they didn’t “feel” the truth enough.

Yeah, that’s pretty much a textbook example of how to showboat at the expense of a productive dialogue. Don’t do it unless status is more important to you than understanding.

Reminds me of that scene in Thank You For Smoking where the dad shows he can “prove” chocolate is better than vanilla because “freedom good”. [1]

(Sorry for the analogy — even if you agreed that such behavior is toxic, maybe you didn’t “feel it enough”, and you think that’s a valid reason to make it.)

[1] https://youtube.com/watch?v=vjOrOMVFCbs

In my experience the good/bad-ness of analogies is often more about where you're trying to use them.

When used in education/teaching to simplify complicated concepts, analogies are invaluable.

When used in debates where you're trying to convince someone of something, analogies are often abused to subtly (or not so subtly) shift the thing that is being talked about, so as to make your argument sound more convincing.

I guess I'd agree with that, but worry that teaching can also function as indoctrination, that is to say when teaching you're also trying to convince someone of something, someone who probably doesn't have the mental resources to take your analogy apart. Get that analogy ground in there and it can be difficult to remove later.

Okay, but this sounds like it's only applicable to teaching about contentious topics, like history or politics, but it doesn't sound very applicable to teaching math or computer science.

There you're definitely not trying to indoctrinate or sell a story, you're very obviously just trying to explain a concept. This is the setting where analogies work perfectly well if used skillfully.

My case against analogies in tech is that they are abused at places where people want to avoid giving an actual explanation of how things really work. Some analogy + "read the code" is usually orders of magnitude less efficient than some wiki page with a couple of diagrams. If you want new team members to translate into dollars quickly, write things down.

"He is using an analogy like a drunk uses a lamp-post, for support instead of illumination."

There are limitations to analogies. And a big boundary line is experience. I can give you lots of analogies for some complicated math etc but if you want to move beyond the limitations of analogies there's nothing left but doing homework.

Not in the realm of science but still my fav attributed I believe attributable to fritz pearls: you can read it the menu, hold the menu, lick the menu, or eat the menu. But you didn't eat the meal

No model or analogy is a bijection on the real deal. I periodically have to remind paper pushers around the office that a complete Jira ticket doesn't mean anything. Organizing paperwork doesn't mean the problem is or can get solved

From a formal logic point of view, an analogy can be two types of argument. One is a valid form of argument, the other is not (i.e. it's fallacious).

The valid form is effectively an "argument to generalization", presented via an easier to understand "specialisation", but only keeping within the attributes of the general case.

E.g. if arguing X, and both X and Y are special cases of Z, an analogy is an argument on Z, presented through Y.

The fallacious variety is then any argument via a specialization Y, which makes use of attributes not present in the general case Z.

Given the above, the problem with using an analogy as an argument is as follows:

1. People tend to use the fallacious version, neglecting it does not apply to the general case. The response to such an analogy as a form of argument is simply to point out this fact. Unfortunately, people usually tend to focus on why the special attributes of Y are a bad argument in themselves, when in fact they should be ignored.

2. Even when people use the valid version, opponents will latch onto attributes, and make a motte-and-bailey attack. The correct defense to that is to reaffirm the validity of the "appeal to generalization" argument, and dismiss any attributes argued by the opponent which do not stem from the generalisation, thus invalidating the attack.

This of course is only useful in "discussions", where unlike a debate the aim is not to "win", but to refine arguments in search of better conclusions.

In a debate, one can therefore expect an analogy, even a logically valid one, to be a very weak debate strategy given the ease with which it can be attacked by a fallacy and the tediousness of the defense required to refute it, because this leaves the analogist in a position of constant, dry defense against a bombardment of vivid, emotionally charged motte-and-bailey arguments by their opponent, giving them ample opportunity to derail the debate to areas where they will have the higher ground.

So, I agree with another poster here. Analogies, particularly valid ones, are good for discussions ("sympathetic audiences"), but bad for debates ("antagonistic audiences").

> Maybe you could take your existing machines and distill them into their essential features so that you can talk about them without referring to the original contexts at all. But then you’ll see analogies between those abstractions and try to extract their features and suddenly you’ll wake up and find that you’ve invented category theory after which you won’t be able to communicate with normal people at all except to tell them how great category theory is and how important it is that they learn category theory. Not recommended.

My favorite line. I'm a huge fan of category theory.

Good at explaining concepts bad at convincing people who believe otherwise

I would say they can be good at conveying a general concept but not at conveying a lot of nuance and, as you say, if people disagree they'll respond by picking apart the analogy.

So a good analogy is like not finding the perfect partner life, but eventually settling with someone because they tick most of the boxes.

I randomly found this a while back, and I think it’s slightly simpler and more helpful: How Analogies Work

https://www.youtube.com/embed/n45D_zi3O5g?start=853&end=1450

“He was as tall as a 6’3” tree.” sounds like something an ML model would throw out if trained on analogies

“Analogies are like babies. Some of them are cute, some of them are ugly, but almost all of them suck.”

Analogies and idioms are heavily leaned on when you’ve substituted actual expertise for rhetoric.

That can be useful; when explaining something complex to someone lacking the same actual expertise, an analogy they can understand without it is potentially helpful.

I was going to correct this but I don’t want to disturb the awesome metajoke in its natural habitat.

An analogy is shorthand for: here is something relatively complicated. We can look at it with the same model that we have agreed upon for this other thing.

If the two things are similar enough in the ways that the model applies, then it's good. Otherwise, it's not.

analogies are like shortcuts: a good one gets everyone there quicker, everybody is happy, and you might even been thought clever. But a bad one gets you nowhere, or back where you started if you are lucky, tired and grouchy for the waste of time.

Analogies are great for suggesting ideas but useless for reaching firm conclusions.

Analogies are a good thing to convince people. When equations match, it is even good to introduce new concepts for people who already know other areas. Other than that, analogies are very rarely useful to explain anything.

> In a good analogy, both sides should have significant shared structure. The more that’s shared, and the deeper the structure, the better.

The alternative to analogies is to name the shared structures.

Analogies are not persuasive. If the audience isn't already on board with the argument at hand, then they're unlikely to be receptive to the analogy.

like them or not, analogies are one of the pillars of the American legal system. Being able to reason about them and create them (usually with motivated reasoning) is a highly valuable skill.

For those unaware, in a legal case where there is no precedant (something us technologists encounter often), you make the case that what you are doing is just like what X had done, and should be treated in kind.

I'm a fan of Douglas Hofstadter's (of Godel Escher Bach fame) view that "analogy is the core of cognition"†

On this view, I guess the question of whether to utilize analogy or not is reduced to the question of which analogy would be most useful, since all categorization (and thus, much of thought in general) is essentially analogical thinking.

† https://www.youtube.com/watch?v=n8m7lFQ3njk

See also his book on this subject (with Emmanuel Sander), Surfaces and Essences: Analogy as the Fuel and Fire of Thinking. I'm currently reading it and enjoying it. A lot of reviews complain about how padded with examples it is, but this is really the substance of the book and I appreciate it.

And he has a point. Analogies are common in fundamental physics. This particular lecture exemplifies that.

Lecture by Douglas Hofstadter: Albert Einstein on Light; Light on Albert Einstein https://www.youtube.com/watch?v=ePA1zq56J1I

It's a lot like the case for and against similes.

I have following 3 suggestions regarding analogies.

[1] Analogies in Fundamental physics

Lecture by Douglas Hofstadter: Albert Einstein on Light; Light on Albert Einstein : https://www.youtube.com/watch?v=ePA1zq56J1I

“Where does deep insight in physics come from? For those who view physics as a highly rational science grounded in strict mathematics, it is tempting to think that it comes from the purest and most precise of reasoning, following ironclad laws of thought that compel the clear mind completely rigidly. And yet the truth is quite otherwise.

One finds, when one looks closely at any major discovery in physics, that the greatest of physicists are the most daring and are constantly being guided by blurry, instinctive, nearly irrational mental forces. Albert Einstein ideally exemplifies this thesis.

In this talk, I will discuss the eternal mystery of light, which, over the course of millennia, was puzzled over, pondered on, and slowly worked out by a series of great minds, and finally, in the nineteenth century, was definitively settled with clarity and rock-solid certainty. And yet one day in the early spring of 1905, quite out of the blue, came an absurd-seeming new suggestion from an unknown Swiss patent clerk, third class, clashing violently with that rock-solid piece of collective wisdom. How did the brazen patent clerk come up with this crazy idea? How it was received by the physics world? What was its eventual fate? And what can we learn about the workings of the human mind from this twisty story filled to the brim with ironies?”

[2] Analogy as the Core of Cognition : https://www.youtube.com/watch?v=n8m7lFQ3njk

In this Presidential Lecture, cognitive scientist Douglas Hofstadter examines the role and contributions of analogy in cognition, using a variety of analogies to illustrate his points.

[3] And the following book,

Surfaces and Essences: Analogy as the Fuel and Fire of Thinking by Douglas R. Hofstadter, Emmanuel Sander

Analogy is the core of all thinking.

This is the simple but unorthodox premise that Pulitzer Prize–winning author Douglas Hofstadter and French psychologist Emmanuel Sander defend in their new work. Hofstadter has been grappling with the mysteries of human thought for over thirty years. Now, with his trademark wit and special talent for making complex ideas vivid, he has partnered with Sander to put forth a highly novel perspective on cognition.

Reply on news.ycombinator.com