Decoupling vs Contextualizing Norms

[Chris_Leong]

John Nerst: “To a contextualizer, decouplers’ ability to fence off any threatening implications looks like a lack of empathy for those threatened, while to a decoupler, the contextualizer’s insistence that this isn’t possible looks like naked bias and an inability to think straight”[1].

Decoupling vs Contextualizing Norms: A cultural divide

A particularly thorny—yet very common—way for a discussion to break down is when participants strongly disagree about the correct scope of a discussion. If neither side is willing to compromise, progress often becomes impossible.

John Nerst identifies a difference in expectations that is particularly prone to causing such issues:

What these norms entail:[2]
Decoupling norms	People have a right to expect the truth of their claims to be considered on their own merits, with no obligation to pay heed to worries about the “broader context” or the “implications” of such speech. Insisting on raising these issues, despite a firm request to consider an idea in isolation, is likely a sign of careless reasoning or an attempt at deflection.
Contextualizing norms	Being a responsible actor necessarily involves considering certain contextual factors and implications when deciding what kinds of statements are acceptable. Not taking these factors into account most likely reflects limited awareness, a lack of care, or even deliberate evasion — especially if the speaker ignores an explicit request.

An example: Even-numbered birth year murderers

Suppose data showed that people born in even-numbered years committed murders at twice the rate of the general population. Can you state this directly and, if so, must you issue a disclaimer?

A decoupler would tend to see it as unreasonable to object to a direct statement of facts. Here’s an example of how someone with this viewpoint might think:

Surely, as a citizen in a free society, I should just be able to state the truth directly? After all, we’re adults. Additionally, we shouldn’t have to issue disclaimers all the time. This kind of compelled speech makes it hard to speak frankly. They amount to soft censorship in the short term and risk creating a slippery slope toward harsher censorship in the long term. Furthermore, it impedes the scientific and intellectual progress that has raised both living conditions and moral standards.

However, contextualizers tend to see the situation quite differently. Here’s one possible expression of this:^[3]

It would be deeply irresponsible to make statements that risk creating a stigma around even-numbered folk. Besides, is there any point in doing so? After all, you can’t just assume that people born in an even year are criminal by default! At the very least, you should issue a disclaimer to prevent bad actors from using your words to push bad faith narratives. It’s not as if that’s difficult! I’m not demanding that you say anything untrue, just that you exercise prudence with what you say regarding a few particularly charged issues.

Truth vs. Harmony

A word of caution: Beware dogmatism

For both norms, it’s easy to think of situations where insisting on them seems dogmatic. Scott Alexander’s excellent post, 📖 Weak men are superweapons , lays out how true statements can be weaponized to destroy a group’s credibility. If you have good reason to believe that someone is using this strategy against you, with the intent to cause serious harm, it would be shockingly naive to let them force you into strict adherence to decoupling norms.

On the other hand, it’s a very common strategy^[4] to frame every disliked action as part of someone’s agenda (neoliberal, cultural Marxist, far-right—take your pick).

Agendas are real, but wielding “universal counter-arguments” is one of the easiest ways to “mindkill” yourself, so I strongly encourage you to be wary here.

My position: Ultimately, it all comes down to wisdom

Contextualizers are correct that it would be rather naive to make certain true statements in a situation that is sufficiently highly charged. But what counts as sufficiently charged and what limitations are reasonable in such a case?

Unfortunately, there isn’t a simple answer here. It would be nice if there were, but I suspect that making the right choice ultimately requires wisdom.

Even if it is best for some conversations not to be maximally public, it still seems important for society’s epistemics to preserve at least some spaces for decoupling-style conversations.^[5] Such spaces create sites of resistance against cultural limitations arbitrarily imposed for political advantage, rather than genuinely serving the common good.

The original source (of this distinction)

🎁 Extras 📝 Recap — consolidate key learnings Executive summary[6]: Decoupling and contextualizing norms each capture something important: truth-seeking often requires evaluating claims apart from their implications, while responsible communication often requires attending to how claims will be heard and used. Neither norm can be applied mechanically. In charged contexts, judgment is needed to decide when broader implications are genuinely relevant and when they are being invoked to suppress inconvenient truths. Still, a healthy epistemic culture needs at least some protected spaces where claims can be examined in a strongly decoupled way. ❦ Key points: • Decoupling norms: Claims should, by default, be assessed on their truth and argumentative merits, rather than rejected because of their social implications or the motives attributed to the speaker. From this perspective, demands for disclaimers or contextual framing can look like deflection, bias, or soft censorship. • Contextualizing norms: Speech acts do not occur in a vacuum. Responsible speakers should sometimes consider how a claim may stigmatize people, empower bad actors, or interact with a charged political environment. From this perspective, ignoring such context can look naive, careless, or evasive. • Illustrative example: The hypothetical claim that people born in even-numbered years commit more murders brings the clash into focus. A decoupler sees a right to state a true fact directly; a contextualizer worries that the statement may create stigma or be weaponized unless carefully framed. • Beware dogmatism: Either norm can be misused. Rigid decoupling can leave one vulnerable to people weaponizing true statements for harmful ends; rigid contextualizing can turn into a universal objection, where any disliked claim is dismissed as serving some hidden agenda. • My stance: There is no simple rule for determining when a situation is “sufficiently charged,” or what constraints are justified. That judgment requires wisdom. • Why decoupling spaces are important: Even if some public conversations should be context-sensitive, society still needs spaces where people can reason in a more decoupled mode. Such spaces protect long-term epistemic health and resist arbitrary political constraints on inquiry. 📚 Recommended reading — delve deeper A Deep Dive into the Harris-Klein Controversy—John Nerst’s original (and excellent!) post ❦ Putanumonit—Explores the relationship between decoupling and mistake/​conflict theory ❦ Relevance Norms; Or, Gricean Implicature Queers the Decoupling/​Contextualizing Binary—Argues that the real distinction isn’t how much people contextualize, but what they consider to be “relevant context”. “The concept of “contextualizing norms” has the potential to legitimize derailing discussions for arbitrary political reasons by eliding the key question of which contextual concerns are genuinely relevant, thereby conflating legitimate and illegitimate bids for contextualization. Real discussions adhere to what we might call “relevance norms”: it is almost universally “eminently reasonable to expect certain contextual factors or implications to be addressed.” Disputes arise over which certain contextual factors those are, not whether context matters at all.” 👏 Acknowledgements — credit where credit is due Hat tip to prontab for sharing this article. He actually uses low decoupling/​high decoupling, but I prefer avoiding double negatives. Both John Nerst and prontab passed up the opportunity to post on this topic here, so I decided to pick up the baton.

1. ^

   This quote is slightly edited. It also serves as a TL;DR.

2. ^

   Honestly, this is more of a spectrum than a binary. However, it is easier to explain as a binary.

3. ^

   I’m sure many people will want to point out that this does not really represent the average view held by contextualizers. Sure, this only represents a more sympathetic contextualizer, but I think that’s perfectly fine as it makes sense to engage with the most defensible version of a viewpoint.

4. ^

   “Strategy” — I don’t mean to imply that it’s always, or even typically, consciously chosen.

5. ^

   Eliezer’s Local Validity as a Key to Sanity and Civilisation articulates the importance of such conversations well.

6. ^

   This recap was produced by hand-edited version of the SummaryBot output then using ChatGPT to iterate.

[Raemon]

Two years later, the concept of decoupled vs contextualizing has remained an important piece of my vocabulary.

I’m glad both for this distillation of Nerst’s work (removing some of the original political context that might make it more distracting to link to in the middle of an argument), and in particular for the jargon-optimization that followed (“contextualized” is much more intuitive than “low-decoupling.”)

This post has been object-level useful, for navigating particular disagreements. (I think in those cases I haven’t brought it up directly myself, but I’ve benefited from a sometimes-heated-discussion having access to the concepts).

I think it’s also been useful at a more meta-level, as one of the concepts in my toolkit that enable me to think higher level thoughts in the domain of group norms and frame disagreements. A recent facebook discussion was delving into a complicated set of differences in norms/​expectations, where decoupled/​contextualizing seemed to be one of the ingredients but not the entirety. Having the handy shorthand and common referent allowed it to only take up a single working-memory slot while still being able to think about the other complexities at play.

[Zack_M_Davis]

This post has been object-level useful, for navigating particular disagreements.

Can you give specific examples? I’ve basically only seen “contextualizing norms” used as a stonewalling tactic, but you’ve probably seen discussions I haven’t.

[Raemon]

The most recent example was this facebook thread. I’m hoping over the next week to find some other concrete examples to add to the list, although I think the most of the use cases here were in hard-to-find-after-the-fact-facebook-threads.

Note that much of the value add here is being able to succinctly talk about the problem, sometimes saying “hey, this is a high-decoupling conversation/​space, read this blogpost if you don’t know what that means”.

I don’t think I’ve run into people citing “contextualizing norms” as a reason not to talk about things, although I’ve definitely run into people operating under contextualizing norms in stonewally-ways without having a particular name for it. I’d expect that to change as the jargon becomes more common though, and if you have examples of that happening already that’d be good to know.

(Hmm – Okay I guess it’d make sense if you saw some of our past debates as something like me directly advocating for contextualizing, in a way that seemed harmful to you. I hadn’t been thinking there through the decoupled/​contextualized lens, not quite sure if the lens fits, but might make sense upon reflection)

It still seems like having the language here is a clear net benefit though.

[Zack_M_Davis]

as the jargon becomes more common

If the jargon becomes more common. (The Review Phase hasn’t even started yet!) I wrote a reply explaining in more detail why I don’t like this post.

[Raemon]

Cool! I found your new post pretty helpful. Will probably have more thoughts later.

[Trinley Goldenberg]

This is one of the Major splits I see in norms on LW (the other being combat vs. Nurture). Having a handy tag for this is quite useful for pointing at a thing without having to grasp to explain it.

[Kaj_Sotala]

My nomination seconds the things that were said in the first paragraphs of Raemon’s nomination.

[Trinley Goldenberg]

The core of this post seems to be this

• Decoupling norms: It is considered eminently reasonable to require your claims to be considered in isolation—free of any context or potential implications. An insistence on raising these issues despite a decoupling request are often seen as sloppy thinking or attempts to deflect.

• Contextualising norms: It is considered eminently reasonable to expect certain contextual factors or implications to be addressed. Not addressing these factors is often seen as sloppy or even an intentional evasion.

As Zack_M_Davis points out in his review, one of the issues with this definition is that there are infinite variations of “context” that can be added in any given situation “I’m wearing a hat while saying this”, and there are infinite implications that any given thing you say can have “this implies that the speaker has a mouth and could therefore say the thing.

However, I do think that there is an actual, useful, real distinction here that’s both important and doesn’t have another thing to describe. The thing I think this is pointing at is “How much you and others are willing to think about the consequences of what is said seperate from its’ truth value.”

This split is quite important in a community that cares strongly about truth, and strongly about the outcomes of the world, and being able to say “we’re in a decoupling space” or “this is a contextualzing conversation” or “I generally support decoupling as an epistemic norm” is quite an important shorthand to point at that thing.

I’d love to see the post cleaned up to make it clear that you’re talking about “contextualizing as understanding how your words will have an effect in the context that you’re in” and decoupling as “decoupling what you say from the effects it may create.”

[Taran]

-

[Trinley Goldenberg]

I think this is valid.

[Raemon]

This post seems to be making a few claims, which I think can be evaluated separately:

1) Decoupling norms exist
2) Contextualizing norms exist
3) Decoupling and contextualization norms are useful to think as opposites (either as a dichotomy or spectrum)

(i.e. there are enough people using those norms that it’s a useful way to carve up the discussion-landscape)

There’s a range of “strong” /​ “weak” versions of these claims – decoupling and/​or contextualization might be principled norms that some people explicitly endorse, or they might just be clusters of tendencies people have sometimes.

In the comments of his response post, Zack Davis noted:

It’s certainly possible that there’s a “general factor” of contextualizing—that people systematically and non-opportunistically vary in how inferentially distant a related claim has to be in order to not create an implicature that needs to be explicitly canceled if false. But I don’t think it’s obvious, and even if it’s true, I don’t think it’s pedagogically wise to use a politically-motivated appeal-to-consequences as the central case of contextualizing.

And, reading that, I think it may actually the opposite – there is general factor of “decoupling”, not contextualizing. By default people are using language for a bunch of reasons all jumbled together, and it’s a relatively small set of people who have the deliberately-decouple tendency, skill and/​or norm, of “checking individual statements to see if they make sense.”

Upon reflection, this is actually more in line with the original Nerst article, which used the terms “Low Decoupling” and “High Decoupling”, which less strongly conveys the idea of “contextualizer” being a coherent thing.

On the other hand, Nerst’s original post does make some claims about Klein being the sort of person (a journalist) who is “definitively a contextualizer, as opposed to just ‘not a decoupler’”, here:

While science and engineering disciplines (and analytic philosophy) are populated by people with a knack for decoupling who learn to take this norm for granted, other intellectual disciplines are not. Instead they’re largely composed of what’s opposite the scientist in the gallery of brainy archetypes: the literary or artistic intellectual.

This crowd doesn’t live in a world where decoupling is standard practice. On the contrary, coupling is what makes what they do work. Novelists, poets, artists and other storytellers like journalists, politicians and PR people rely on thick, rich and ambiguous meanings, associations, implications and allusions to evoke feelings, impressions and ideas in their audience. The words “artistic” and “literary” refers to using idea couplings well to subtly and indirectly push the audience’s meaning-buttons.

To a low-decoupler, high-decouplers’ ability to fence off any threatening implications looks like a lack of empathy for those threatened, while to a high-decoupler the low-decouplers insistence that this isn’t possible looks like naked bias and an inability to think straight. This is what Harris means when he says Klein is biased.

Although they’re interwoven, I think it might be worth distinguishing some subclaims here (not necessarily made by Nerst or Leong, but I think implied and worth thinking about)

• There exist a class of general storytelling contextualists

• There exist PR-people/​politicians/​activists who wield contextual practice as a tool or weapon.

• There exist “principled contextualizers” who try to evenly come to good judgments that depends on context.

My Epistemic State

Empirical Questions

There’s a set of fairly concrete “empirical” questions here, which are basically “if you do a bunch of factor analysis of discussions, would decoupling and/​or contextualization and/​or any of the specific contextual-subcategories listed above be major predictive power?”

The experiments you’d run for this might be expensive but not very confusing.

I would currently guess:

• “Decoupling factor” definitely exists and is meaningful

• Storytelling contextualists exist and are meaningful (though not necessarily especially useful to contrast with decouplers)

• PR-ists who wield context as tool/​weapon definitely exist (and decoupling is often relevant to their plans, so they have developed tools that allow them to modulate the degree to which decoupling fits into the conversational frame)

• I think I could name a few people at least attempting to be “fair, principled contextualists”, at least in some circumstances. I am less confident that this is a real thing because the alternative “secretly they’re just really effective or subtle PR-ists, either intentionally or not” is a pretty viable alternative

Conceptual Question

I have a remaining confusion, which is something like “what exactly is a contextualizer?”. I feel like I have a crisp definition of “decoupling”. I don’t have that for contextualizers. Are the three subcategories listed above really ‘relatives’ or are they just three different groups doing different things? Is it meaningful to put these on a spectrum with decouplers on the other side?

mr-hire suggests:

“How much you and others are willing to think about the consequences of what is said separate from its’ truth value.”

Which sounds like a plausibly good definition, that maybe applies to all three of the subcategories. But I feel like it’s not quite the natural definition for each individual subcategory. (Rather, it’s something a bit downstream of each category definition)

“Jumbled” vs “Contextual”

“High decoupling” and “low decoupling” are still pretty confusing terms, even if you get rid of any notion of “low decoupling” being a cogent thing. It occured to me, writing this review, that you might replace the word “contextual” with “jumbled”.

Contextual implies some degree of principled norms. Jumbled points more towards “the person is using language for a random mishmash of strategies all thrown together.” (Politicians might sometimes be best described as “jumbled”, and sometimes as “principled” [but, not necessarily good principles, i.e. ‘I will deliberately say whatever causes my party to win’]).

...

That’s what I got for now.

[Chris_Leong]

I really don’t like the term jumbled as some people would likely object much more to being labelled as jumbled than as a contextualiser. The rest of this comment makes some good points, but sometimes less is more. I do want to edit this article, but I think I’ll mostly engage with Zack’s points and reread the article.

[Raemon]

The OP comment was optimizing for “improving my understanding of the domain” more than direct advice of how to change the post.

(I’m not necessarily expecting the points and confusions there to resolve within the next month – it’s possible that you’ll reflect on it a bit and then figure out a slightly different orientation to the post, that distills the various concepts into a new form. Another possible outcome is that you leave the post as-is for now, and then in another year or two after mulling things over someone writers a new post doing a somewhat different thing, that becomes the new referent. Or, it might just turn out that my current epistemic state wasn’t that useful. Or other things)

Re: “Jumbled”

I think there’s sort of a two-step process that goes into naming things (ironically, or appropriate, which map directly onto the post) – first figuring out “okay what actually is this phenomenon, and what name most accurately describes it?” and then, separately, “okay, what sort of names are going to reliably going to make people angry and distract from the original topic if you apply it to people, and are there alternative names that cleave closely to the truth?”

(my process for generating names in that risk offending is something like a multi-step Babble and Prune, where I generate names aiming to satisfice on “a good explanation of the true phenomenon” and “not likely to be unnecessarily distracting”, until I have a name that satisfies both criteria)

I haven’t tried generating a maximally good name for Jumbled yet since I wasn’t sure this was even carving reality the right way.

But, like, it’s not an accident that ‘jumbled’ is more likely to offend people than ‘contextualized’. I do, in fact, think worse of people who have jumbled communication than deliberately contextualized communication. (compare “Virtue Signalling”, which is an important term but is basically an insult except among people who have some kind of principled understanding that “Yup, it turns out some of the things I do had unflattering motives and I’ve come to endorse that, or endorse my current [low] degree of prioritizing changing it.”)

I am a conversation consequentialist and think it’s best to find ways of politely pointing out unflattering things about people in ways that don’t make them defensive. But, it might be that the correct carving of reality includes some unflattering descriptions of people and maybe the best you can do is minimize distraction-damage.

[Zack_M_Davis]

Reply: “Relevance Norms; Or, Gricean Implicature Queers the Decoupling/​Contextualizing Binary” (further counterreplies in the comment section)

I argue that this post should not be included in the Best-of-2018 compilation.

[Wei Dai]

It occurs to me that “free speech”, “heterodoxy”, and “decoupling vs contextualising” are all related to intelligence vs virtue signaling. In particular, if you want to do or see more intelligence signaling, then you should support free speech and decoupling norms. If you want to do or see more virtue signaling, then you should support contextualising norms and restrictions on free speech. Heterodox ideas tend to be better (more useful) for intelligence signaling and orthodox ideas better for virtue signaling. (Hopefully this is obvious once pointed out, but I can explain more if not.)

[Kaj_Sotala]

AAAAAAAAA THIS THING IT HAS A NAME AT LAST!

(ahem.)

[Sniffnoy]

Decoupling, orthogonality, unbundling, separation of concerns, relevance, the belief that the genetic fallacy is in fact a fallacy, hugging the query.… :)

Not a new idea, but an important one, and worth writing explicitly about!

[Chris_Leong]

Any links to where this has already been discussed?

[Sniffnoy]

I mean, it’s kind of implicit in lots of the stuff on Less Wrong, isn’t it?

[quanticle]

Question: is there any reason to use the words “decoupling” rather than “coupling”? It seems to me that “low decoupling” is logically equivalent to “high coupling” and “high decoupling” is logically equivalent to low coupling. So in the spirit of simplification, would it not be better to state the distinction as being between “high coupling” people and “low coupling”?

[gjm]

To me, (1) “coupling” suggests specifically joining in pairs much more strongly than “decoupling” suggests specifically detaching pairs and (2) “coupling” suggests that the default state of the things is disconnection, whereas “decoupling” suggests that the default state is connection.

The usual scenario here is that (1) you have lots of things that all relate to one another, and that (2a) most people find it difficult to disentangle, or disapprove of disentangling, and that (2b) all really truly are connected to one another, so that considering them in isolation is a sometimes useful and effective cognitive trick rather than any sort of default.

For all those reasons I think “decoupling” is a better term than “coupling” here. (I also like the opposition decoupling/​contextualizing, as found in some of the earlier things Nernst links to, rather than more-decoupling/​less-decoupling. When faced with a pile of interrelated things, sometimes you want to decouple them and sometimes you want to pay special attention to the interrelations. It’s not as simple as there being some people who are good at decoupling and some who aren’t. Though of course most people are bad at decoupling and bad at contextualizing...)

[Chris_Leong]

Actually, I like Decoupling vs. Contextualising more too, especially as they become single words.

[Raemon]

Yeah I like those quite a bit more, actually (I actually thinking using them in the conversations on this post would make them easier to follow)

[Sniffnoy]

I definitely think Nerst has things the right way round, but I’m having trouble making explcit why. One reason though that I can make explicit is that, well, tangling everything together is the default. Decoupling—orthogonality, unbundling, separation of concerns, hugging the query—is rarer, takes work, and is worth pointing out.

[orthonormal]

I wish John Nerst could be convinced to crosspost here.

[Chris_Leong]

For what it’s worth, I tried.

[Raemon]

Curated for succinctly creating some useful handles for two concepts that have implicitly been coming up a lot. i think this has already been helpful to me when thinking about some confusing/​challenging conversations.

[Leon Lang]

I repeatedly refer people to this post, and they repeatedly tell me that it explains a great many of conversations in their real life in a way they previously found hard to pin down. It’s a great post.

[null]

I think this article is a considerable step forward, but it could benefit from some examples. I think I have a pretty good idea what this is about (and share the horror of being called out by a low-decoupler for being some kind of ism), but still.

[Chris_Leong]

Hmm, well the article has an example, but it is super long and I’m trying to avoid this becoming political. Any suggestions for examples?

[Daniel Kokotajlo]

The example you use is already CW-enough that high-decouplers may be suspicious or hostile of the point you are trying to make.

Then again, maybe anything elsewould be too far removed from our shared experience that it wouldn’t serve as a quick and powerful illustration of your point.

Here are some suggestions made with both of these points in mind:

--The original example Scott uses about a Jew in future Czarist Russia constantly hearing about how powerful Jews are and how evil Israel is.

--Flipping the script a bit, how about an example in which someone goes around saying “86% of rationalists are straight white men” (or something like that, I don’t know the actual number).

--Or: “Effective Altruists are usually people who are biased towards trying to solve their problems using math.”

Come to think of it, I think including one of those flip-script examples would be helpful in other ways as well.

[Ruby]

At the same time, I don’t want to fall for the Fallacy of the Undistributed Middle and assume that both perspectives are equally valid.

Minor possible quibble: based on the definition in the link given, I think Fallacy of the Undistributed Middle doesn’t refer to assuming a the deep wisdom position that two sides of a debate each have merit.

The fallacy of the undistributed middle (Lat. non distributio medii) is a formal fallacy that is committed when the middle term in a categorical syllogism is not distributed in either the minor premise or the major premise. It is thus a syllogistic fallacy.)

[Raemon]

Agreed with quanticle, but otherwise think this is a very helpful dichotomy to have a handle for.

[Will McPherson]

It would be deeply irresponsible to make statements that risk creating a stigma around even-numbered folk. Besides, is there any point in doing so? After all, you can’t just assume that people born in an even year are criminal by default! At the very least, you should issue a disclaimer to prevent bad actors from using your words to push bad faith narratives. It’s not as if that’s difficult! I’m not demanding that you say anything untrue, just that you exercise prudence with what you say regarding a few particularly charged issues.

This would be a more compelling (but not necessarily better) example if you imagine a totalitarian society that is already oppressing even-year people. Most people would avoid disseminating information that could be used to harm others, even if that information was true, and even if that harm was unjustifiable.

[Martin Sustrik]

Isn’t it just the distinction between how facts are used in scientific discourse (you state a fact, expect it to be confirmed or challenged) vs. how they are used in political discourse (carefully select other fact to augment it and suit your political narrative)? I guess Umberto Eco would have had something to say about that.

[Michael Pye]

I can see the models usefulness but I think you are impling they are equal. This seems wrong. Decoupling required more specific knowledge and concentration and is more analogous to Kahneman’s slow thinking. Obvious context can be thought of in a rich way but only after the individual ideas have been richly defined (often in previous thoughts). We cycle between the two approaches but I feel decoupling required the greater focus and is lacking when we discuss topics we are unfamiliar with (and a possible link with the Dunning Kruger effect). My understanding of cognitive load theory from education (limited working memory) also seems relavent. By limiting context we can more intensely analyse each aspect, repackage them efficiently and accurately before returning contextual information the whole problem. This seems to me the classic method of enlightenment thinking. Obviously understanding why other people might misunderstand d this is important however that is an argument around politics, rhetoric and persuasion not about clarity of thought.

Finally pure contextualisation is only about outcomes and decoupling only about process. Without understanding alternate approaches (which requires specialist knowledge) our assessment of the best method to achieve our outcome is likely flawed.

[zulupineapple]

Since this comes from the Harris-Klein debate, I should point to this recent post and especially the comments underneath it. To summarize, the “high decoupler” Harris is making errors. This is, of course, what happens when you ignore the context of the real world.

Now, there are perhaps better examples of disagreement between “high decouplers” and “low decouplers”, and perhaps those are still meaningful categories. But I’d be weary of conclusions made with high decoupling.

I propose an alternative view, where “low decoupling” is the objectively correct way to look at the world, and “high decoupling” is something you do because you’re lazy and unwilling to deal with all the couplings of the real world.

[Chris_Leong]

“”high decoupling” is something you do because you’re lazy and unwilling to deal with all the couplings of the real world”—I suspect you don’t quite understand what high decoupling is. Have you read Local Validity as a Key to Sanity and Civilisation? High decoupling conversations allow people to focus on checking the local validity of their arguments.

[zulupineapple]

High decoupling conversations allow people to focus on checking the local validity of their arguments.

We unfortunately live in a world where sometimes A implies C, but A & B does not imply C, for some values of A, B, C. So, if you’re talking about A and C, and I bring up B, but you ignore it because that’s “sloppy thinking”, then that’s your problem. There is nothing valid about it.

I suspect you don’t quite understand what high decoupling is.

High decoupling is what Harris is doing in that debate. What he is doing is wrong. Therefore high decoupling is wrong (or at least unreliable).

I get the feeling that maybe you don’t quite understand what low decoupling is? You didn’t say anything explicitly negative about it, but I get the feeling that you don’t really consider it a reasonable perspective. E.g. what is the word “empathy” doing in your post? It might be pointing to some straw man.

[Ben Pace]

High decoupling is what Harris is doing in that debate. What he is doing is wrong. Therefore high decoupling is wrong (or at least unreliable).

Upvoted for going all-in on a low-decoupling norm—I can’t tell whether that was intentionally funny or you’re genuinely living life by low-decoupling norms.

(Either way I think you’re passing the ITT for low-decoupling, so thanks.)

[zulupineapple]

If you think there is something funny about low decoupling, then you’re probably strawmanning it. Or maybe it was a straw man all along, and I’m erroneously using that term to refer to something real.

or you’re genuinely living life by low-decoupling norms.

I can’t say that I do. But I try to. Because high decoupling leads to being wrong.

[Chris_Leong]

“We unfortunately live in a world where sometimes A implies C, but A & B does not imply C, for some values of A, B, C. So, if you’re talking about A and C, and I bring up B, but you ignore it because that’s “sloppy thinking”, then that’s your problem. There is nothing valid about it.”—What kind of “implies” are you talking about? Surely not logical implications, but rather the connotations of words? If so, I think I know what I need to clarify.

I didn’t comment on what norms should be in wider society, just that low decoupling spaces are vital. I was going to write this in my previous comment, but I had to run out the door. John Nerst explains “empathy” much more in his post.

[zulupineapple]

I’m talking about the kind of “X implies Y” where observing X lead us to believe that Y is also likely true. For example, take A=”wet sidewalk” and C=”rain”. Then A implies C. But if B=”sprinkler”, then A&B no longer imply C. You may read this, also by Elizer and somewhat relevant.

John Nerst explains “empathy” much more in his post.

Yes, I’ve read that, and he is also strawmanning. Lack of empathy is not the problem with what Harris is saying. Did you read the comments I linked to? Or should I have quoted them here?

[Said Achmiz]

For example, take A=“wet sidewalk” and C=“rain”. Then A implies C.

But this is false. That was Eliezer’s whole point.

[zulupineapple]

When B is not known or known to be false, A implies C, and, when it is know to be true, A&B do not imply C. Surely we have no actual disagreement here, and I only somehow managed to be unclear that before I introduced B, it wasn’t known?

[Said Achmiz]

When B is not known … A implies C

No. This is wrong. This is what I am saying: when B is not known, A does not imply C. A can only imply C if B is known to be false.

Edit: In other words, A → (B ∨ C).

Edit 2: Spelling it out in more detail:

• A ⇒ (B ∨ C)

• (A ⇒ (B ∨ C)) ∧ A) ⇏ C

• (A ⇒ (B ∨ C)) ∧ (A ∧ ¬B)) ⇒ C

[Ben Pace]

Zulupineapple I feel like Said is trying to give a first lesson in propositional logic, a setting where all his statements are true. Were you trying to use the colloquial/​conversational meaning of the word ‘implies’?

[zulupineapple]

Yes, I explicitly said so earlier. And propositional logic makes no sense in this context. So I don’t understand where the confusion is coming from. But if you have advice on how I could have prevented that, I’d appreciate it. Is there a better word for “implies” maybe?

[zulupineapple]

Maybe you’re talking about the usual logic? I explained in the very comment you first responded to, that by “X implies Y” I mean that “observing X lead us to believe that Y”. This is a common usage, I assume, and I can’t think of a better word.

And, if you see a wet sidewalk and know nothing about any sprinklers, then “rain” is the correct inference to make (depending on your priors). Surely we actually agree on that?

[Said Achmiz]

Yes, I saw your definition. The standard sort of generalization of propositional logic to probabilistic beliefs does not rescue your claims.

And, if you see a wet sidewalk and know nothing about any sprinklers, then “rain” is the correct inference to make (depending on your priors). Surely we actually agree on that?

No. If you’re leaving propositional logic behind and moving into the realm of probabilistic beliefs, then the correct inference to make is to use the information you’ve got to update from your priors to a posterior probability distribution over the possible states of the world. This is all standard stuff and I’m sure you know it as well as I do.

The outcome of this update may well be “P(rain) = $large_number; P(other things, such as sprinklers, etc.) = $smaller number”. You would, of course, then behave as if you believed it rained (more or less). (I am glossing over details, such as the overlap in P(sprinkler, etc.) and P(rain), as well as the possibility of “hybrid” behaviors that make sense if you are uncertain between two similarly likely possibilities, etc.; these details do not change the calculus.)

Characterizing this as “A implies C, but (A ∧ B) does not imply C” is tendentious in the extreme (not to mention so gross a simplification that it can hardly be evaluated as coherent view).

Now, you might also be claiming something like “seeing a wet sidewalk does increase P(rain), but does not increase P(sprinkler)”. The characterization quoted in the above paragraph would be consistent with this claim. However, this claim is obviously wrong, so I assumed this wasn’t what you meant.

[zulupineapple]

So when I said “rain is the correct inference to make”, you somehow read that as “P(rain) = 1″? Because I see no other explanation why you felt the need to write entire paragraphs about what probabilities and priors are. I even explicitly mentioned priors in my comment, just to prevent a reply just like yours, but apparently that wasn’t enough.

Characterizing this as “A implies C, but (A ∧ B) does not imply C” is tendentious in the extreme (not to mention so gross a simplification that it can hardly be evaluated as coherent view).

Ok. How do you think I should have explained the situation? Preferably, in less than four paragraphs?

I personally find my explanation completely clear, especially since I expected most people to be familiar with the sidewalk/​rain/​sprinkler example, or something similar. But then I’m aware that my judgements about clarity don’t always match other people’s, so I’ll try to take your advice seriously.

[Said Achmiz]

How do you think I should have explained the situation? Preferably, in less than four paragraphs?

Assuming that “the situation” in question is this, from upthread—

We unfortunately live in a world where sometimes A implies C, but A & B does not imply C, for some values of A, B, C.

I would state the nearest-true-claim thus:

“Sometimes P(C|A) is very low but P(C|A,B) is much higher, enough to make it the dominant conclusion.”

Edit: Er, I got that backwards, obviously. Corrected version:

“Sometimes P(C|A) is very high, enough to make it the dominant conclusion, but P(C|A,B) is much lower [this is due to the low prior probability of B but the high conditional probability P(C|B)]”.

[zulupineapple]

Ok, that’s reasonable. At least I understand why you would find such explanation better.

One issue is that I worry about using the conditional probability notation. I suspect that sometimes people are unwilling to parse it. Also the “very low” and “much higher” are awkward to say. I’d much prefer something in colloquial terms.

Another issue, I worry that this is not less confusing. This is evidenced by you confusing yourself about it, twice (no, P(C|B), or P(rain|sprinkler) is not high, and it doesn’t even have to be that low). I think, ultimately, listing which probabilities are “high” and which are “low” is not helpful, there should be a more general way to express the idea.

[Michael Pye]

Zulupineapple do you release you have been engaged in a highly decoupled argument? Your point (made way back) that values contextual conclusions is valid but decoupling is needed to enhance those conclusions and as it is harder by an order of magnitude requires more practice and knowledge.

Personally I feel the terms abstract and concrete are more useful. Alternating between the two and refining the abstract ideas before applying them to concrete examples.