Several years ago, we introduced the concept of fiat news on these pages.
It is a simple idea. In the same way that money created by fiat debases real money, news created by fiat debases real news.
Although it misinforms, fiat news should not be understood as misinformation, at least in the colloquial sense. News which contains false information or distorted interpretations of facts can be better thought of as counterfeit news. Like counterfeit money, enough counterfeit news can debase the real thing, too. Yet even considering how widespread counterfeit news has become, fiat news exists on such a massive scale that its power to debase is in a different category. Nor is fiat news synonymous with bias. We think bias represents a causal explanation for a very specific kind of recurring fiat news.
Most “media watchdogs” are in the business of identifying one of those two things: misinformation or bias. The problem with these efforts, beyond that they do not capture the full scope of actions which debase the information content of news, is that it is practically impossible to report on misinformation or bias in a manner that is itself not colored by the opinions of the author, or else designed to shape how the reader interprets facts and events. While they may in some cases offer a useful service in the face of blatant lies published through politically invested news outlets, too often they become yet another source of fiat news.
Why? Because fiat news is the presentation of opinion as fact. Fiat news is news which is designed not to provide information for the reader to process, but to provide interpretations of information for the reader to adopt. Fiat news is the primary vector for nudging, shaping common knowledge, or what everybody knows everybody knows. Fiat news is how governments, parties, corporations and other institutions in a free and always connected society meticulously shape it - then tell us that it was our idea.
By design, fiat news isn't always easy to spot. Outside of editorial pages, it is rare indeed that an expression of opinion as fact would include obvious phrases like “we believe” or “we think.” Instead, media outlets guide interpretations through more subtle means that may sometimes be as invisible to the author and editor as they are to the reader.
Several years ago, we also introduced what we called the narrative machine.
I have only begun to scratch the surface of this important piece, but as one who writes a paid-for newsletter I find there are a lot of red flags. Am I ascribing something inappropriately? Have I quoted somebody out of context? Do my facts fit? Will the reader be better informed after reading my stuff? Many places to look for shoes that fit. Thanks, Rusty.
Great starting point for a hugely important public discussion.
My point of reference is my analysis from 5-6 years ago of how Uber used these exact techniques to create massive Fiat Corporate Value (nearly $100 bn created out of thin air).
A couple of items you might consider adding to your list:
16: Ignore the Funding/Financial Interests of the Experts Interviewed (or the role of longstanding well-funded think tanks in political situations). Stories quote experts as if they are dispassionate analysts whose claims are backed by the kind of rigorous, peer-reviewed research you’d see in major academic settings and deliberately conceal are being paid large sums from the interests their claim supports. Tiny bits overlap with “Appeals to Authority” “Coverage Selection” and “Missionary Statements” but the overall problem here goes well beyond what you’ve included in those three items.
Only One Side of The story Gets Reported. Again, small bits included in other items, For years, MSM coverage of Uber exclusively reported that they were the greatest thing since sliced bread, and had succeeded because of cutting-edge technology that created huge productivity advantages. This also applies generally to MSM coverage of US overseas military actions in the last 25 years.
Never Any Attempt To Review the Bigger Picture. In isolation fiat news stories that simply repeat corporate or governmental claims are somewhat understandable given the pressures of the news cycle. But you never ever see a MSM media outlet step back after a number of years and examime whether the corporate or governmental claims uncritically published in the past actually turned out to be true. Was Uber actually the biggest thing since sliced bread? Did claims about weapons of mass destruction actually justify 20 years of war? Did those 20 years of war actually provide major benefits for the people of IRaq/Afghanistan?
Reject/Ignore Longstanding Measures of Bottom-Line Results. In 12 years Uber has lost $31 billion on its actual, ongoing taxi and delivery services and has yet to generate a single dollar of positive cash flow and no one can explain how it ever could produce sustainable profits. These are well-understood ways to measure corporate performance, but you won’t find a single MSM story that discusses them seriously. Even the Guardian’s recent Uber “expose” completely ignored competitive economics and financial results in order to focus on Stylistic/Cultural issues which provide a highly inaccurate picture of Uber’s performance. I’m sure everyone can quickly recall dozens of comparable issues in the coverage of military and political problems.
Maybe a couple of these could be subsumed in a modified version of your first 15 warning signs, but I’m guessing that list will end up expanding
Excellent. I’d appreciate explicit examples of each “face”. Maybe even in a separate doc? I think I get them all but real-world examples of each would help.
Which face is employed by that conclusion? I sense something dripping from it, but I don’t think it’s empathy.
Hi Roy. Always a good practice, but I’d say most newsletters fall squarely within the “obviously opinion” camp. While the rise in the aggregate volume of this kind of content is probably tipping the scales on the fiat news spectrum, I don’t know that it’s our primary concern. There should be a place for people to attempt to convince others in the various media; we’d argue that place is “where people know someone is seeking to convince them.” My guess is your newsletter falls squarely in that camp.
Hah! None whatsoever. I’m not a news outlet, and happy to assert some tendency toward self-aggrandization within certain professions. Couldn’t possibly be something that financial writers are guilty of, too.
Absolutely, Kevin! Definitely the plan over the next few pieces.
Thanks, Hubert! I think these are good ideas for checks for anyone who is thinking critically about a topic. Really good ideas, actually.
I also think the pitfalls of classifying them as fiat news are significant. In all three of these, you’re talking about framing through omission of one kind or another, which is absolutely a thing. I think you’re 100% on the right track and have zero disagreement with the specific examples. But framing through omission is a thing which presumes you can identify some objective baseline of “what someone should be covering” with respect to a particular topic. Not only am I not sure we can do that, I actively think that if I attempted to do it I would be programmatically introducing that unavoidably subjective assessment into the model. And I mean that both in the sense of “if I tried to do this personally in my own news consumption” and in the sense of “if I tried to incorporate this into a fixed fiat news model.”
And there are two answers to those two senses. To the former (i.e. our personal news consumption habits) I think it is very good to be aware of each of the ways mentioned by you that authors can frame through omission. I think it is also very good to be mindful that we are not doing question begging of our own. When we start doing the “but why aren’t you talking about” game, it is very often - or at least very often is for me, as I can’t speak for you - because we’ve already drawn some conclusions of our own and are just a bit miffed that they aren’t actively working to support our conclusions.
To the second sense (that is, modeling this more systematically), I presently think that the best way to track this is through the Coverage Selection face of fiat news. What we can model is the extent to which topics (e.g. actual financial results) are covered for one entity (e.g. your average S&P stock) at rates which they are not for another (e.g. Uber). We can model and show how that differs among outlets. We can show how it differs over time. We model topics and entities like this all the time, so this is a pretty vanilla part of what we’d potentially be looking at for Coverage Selection.
I think that’s ultimately going to get, say, 60% of what you’re talking about, which means I’m probably leaving 40% on the table to avoid false positives. Knowing that, how do you think we could improve on that to better capture the very real things you note without injecting too much subjectivity about baseline expectations of “proper” coverage ourselves?
Mapping out Content Context has to involve some mechanism that reads the headline(s) and compares them to the words used to promote the story on social media, specifically Twitter. Any revisions to the social media promotion–deleting of a tweet and replacing it with something more neutral, say–would provide another data point from which the magical algorithm (and to me it is indistinguishable from magic) that could indicate its fiat newsy-ness.
I’m picturing something simple like a point system where 0 is pure news and 1 is pure opinion, and each ‘face’ adds some weighted amount of points to the total score. I’m thinking this way because of course I have no bloody idea how the Hell NLP works or how this project is going to judge these things. That’s multiple levels above my comprehension. But in thinking broadly about how to judge context, which is by nature subjective, I would at least in concept try to assign a value to certain actions. So deleting a clickbait tweet and replacing it with something neutral would be worth 0.1 or whatever. (For comparison an article that is written from the first person and has a lot of ‘I believe’ or ‘it’s time for us to do X’ would score 1.0 right off the bat) You have to figure out how to assign a value to not just the words used but the actions around those words. Does the headline correspond with the actual language of the piece? Does the headline have someone’s name in it but that person is only mentioned once in a 700 word story? If so is that meaningful? If it is then why? If it isn’t then why not?
I will add, along a different line here, that Unsourced Attribution is probably the most dangerous of the 15 faces. That stories are believed ab initio even though nobody has put their name to them is frightening. We skipped past the ‘trust but verify’ model and went right into ‘because some guy who totally exists said so’. The media is obsessed with telling us that no, we wouldn’t know their girlfriend, she goes to another school. Maybe we don’t care right now because those types of stories are only hurting the Bad People
, but one day they’ll move on and something or someone you care about will be next.
I want to suggest a fiat news tactic I don’t think is covered by those here, forgive me if you think it is. I call it “speculation anchoring,” although it may have another name.
This is when an article suggest causality or connection between two things in its headline or in its opening paragraphs, but then goes on to say, wait, actually, there is no real evidence of this and the connection is purely speculative. Because a lot of readers won’t read carefully past the headline or the first paragraphs, the purported connection is established in the reader’s mind, due to the fact that an article ostensibly asserting the connection was published. What they will remember is, “I saw an article a while back saying X…” and will forget the details about evidence or lack of it. In this way the reader’s mind is “anchored” toward a false or weakly evidenced belief, even as the article may not explicitly argue for that belief. The reader thinks, “X is happening”, when in fact it is only someone speculating about X. I have sometimes even seen that the article will acknowledge evidence that directly refutes X, leaving it a mystery as to why the article was written - unless it were trying to establish a narrative rather than illuminate fact.
This technique is everywhere but I see it I think it is most rampant in articles about climate change - something weird is happening with the weather or in the natural environment. “Scientists think” this could be due to climate change. Read to see that there is no evidence that climate change has anything to do with it. Yet for some reason, you are still left with the impression, this is a climate change thing.
Here is a recent example about shark attacks on the rise recently on the East Coast, claiming in the subheadline that “Climate change may play a factor in sharks venturing closer to shore.” There is heading that reads “Global warming may play a factor.” The first sentence of the paragraph however, admits that “there is no data” to support this claim at all. Nevertheless, we still see several paragraphs of pure speculation on the topic, which includes implicit appeals to authority because the speculation is coming from “experts.” Even though there are more paragraphs establishing the numerous other possible causes, “Scientists have an explanation” according to the headline, but the article actually establishes that they DON’T have an explanation.
The article even goes on to say, there might not even be more sharks in the water, but just more people, conceding that even the apparent trend in shark behavior could be completely illusory. Moreover, anybody who has even semi-regular interaction with nature knows that wild animals sometimes deviate from the patterns we expect for no clear reason. But now there’s an article out there propagating the idea that “sharks are changing their behaviors because of climate change.”
The point system you mention makes me think of college rankings in U.S. News. What if there was some (respected) entity assigning point values to journalism in terms of the percentage of its content that was Fiat News? College admins would sacrifice their firstborn for higher rankings; I’d think a similar ranking system for Fiatness could be a valuable incentive for outlets to identify and reduce their Fiat content.
Hi David:
Please forgive me here as I have not gone back and updated my information this morning but there was a “Shark Attack Data Base”, available years ago, not remembering who put it out, but the analytical discussion I remember was based on the “time of day” theory ( low light = greater attacks) versus more opportunity ( more people in the water). There was the appearance of a trend towards a greater number of attacks taking place mid-day as the opportunity increased at those hours. Might take me a couple of days but will go looking.
The reason for my response was to comment that another form of what I call “moving the goalpost” on speculation ( resembles Moat and Baily) is when a string of speculative comments are offered under the guise of fact with a final, wildly speculative comment, slightly distanced from the main body of comments, tacked onto the end with a disclaimer such as “but that would be entirely speculation”.
I am sure most remember “GIGO”. Fiat news = junk food for the brain. Less filling, tastes great!
Postscript: International Shark Attack File , est. 1958, housed at the Florida Museum of Natural History
My fear is that media organizations would simply try to figure out how to keep doing exactly what they’re doing but without it being as easy to detect. In fact, I will go so far as to say that @rguinn and company should absolutely not disclose any of the proprietary methodology that will be used to identify fiat vs legitimate news.
This is definitely a thing, although it’s an open question whether is deserves its own category.
This practice may be even more pernicious than you suggest, as what we have discovered in streaming this data into our dataset is that headline fishing is huge. That is, even very “reputable” outlets regularly lead with exactly this kind of click-bait connection, then rotate after they’ve pushed it on social media (which often caches the original click-bait header for its preview snippets) to something else, then something else. There seems to be ombudsman indifference to anything that has to do with headline shenanigans, no mattet the source.
I DO think we intended and do capture this linguistically and behaviorally in two places. It will hit our question begging category and it will hit our rhetorical questions category. I think those are still correct personally, although this does give me several ideas to expand what we are catching. But curious, group: is this subset substantial enough to be its own thing?
I am curious what the group thinks about this. Even though we are explicitly modeling something other than bias, it is inevitable that we will be attacked for being inherently biased. They may even be correct. Just because I dont see how it would happen in context doesn’t mean it isn’t possible. So I get calls for transparency.
I also think D_Y is probably right. Both can be true.
What do you guys think?
I can see this happening, and I could see such positive effects. But to D_Y’s point, is this desirable? In the same way that restaurants hew to Michelin preferences and colleges hew to US News algorithms (i.e. arbitrarily weighted spreadsheets), does this lead to edge case optimization that doesnt change the fundamental thing?
Not rhetorical. Honestly don’t know.
You’re right that it could be considered a specific hybrid of rhetorical questioning, click-bait, and maybe question begging. Although, if I were to argue for a distinction I would say that speculation anchoring is a particular sleight of hand that relies on acknowledging an absence of evidence or contrary evidence to paradoxically create undue confidence in the original speculative assertion. It creates the illusion of “We looked into this question, and here is what we found.” If they just said, “This is because of X” then they would open themselves to criticism that this claim was unsupported. But if the author acknowledges that, saying, “I know this is unsupported, but think about it, it would make sense!” I think it tends to lower the reader’s defenses, because it makes seem like they investigated the question.
updated original post with source on shark attack data
Thanks for your response Rusty
My comment was focused on your larger objective “to inoculate the world against the weaponized narratives of Big Tech, Big Media and Big Politics”
I totally appreciate the huge effort you’ve put into building a system using natural language processing to evaluate massive quantities of text. My assumption that the battle to educate and inoculate the world would need to be fought on many fronts. This tool has obvious value in many situations including the graphs you’ve included tracking aggregate trends over time. But I don’t see how it can become the primary education and inoculation tool. Maybe that’s not what you meant but that’s what your reply implies.
I assume other fronts in this battle would include analysis of how Big Tech, Big Media and Big Politics (and Big Finance) actually develop and promulgate their narrative campaigns, how narrative promulgation has changed over time, and the incentives that get the media to create fiat news and uncritically endorse fiat news claims being pushed by politicians and corporations. As other commenters have noted, just assembling concrete examples would help get key fiat news concepts across to people who haven’t been thinking about them for years.
I had assumed that the “Fifteen Faces” taxonomy would be a key input to all of the components of the battle. Totally understand that additional faces I suggested (and in fact several of the Fifteen) may not lend themselves to an algorithm attempting to categorize the entire output of the media in a highly reliable manner. Counting the number of news stories featuring “bogeymen” is useful. Explaining how political/corporate interests manufacture bogeyman based narratives and then get the media to (implicitly) endorse those narratives is also important. I assume this taxonomy will continue to evolve, but doesn’t it need to serve all of the different components of the battle?
Up to you but I see negligible risk that publishing methodological details would allow medis types to figure out how to evade detection. As your original note suggested, journalism profs have been criticizing many of these things for years. When I’ve had direct discussions with earnest journalists from prestigious publications they are always shocked (shocked!) to learn their stories are promulgating manufactured narratives and groupthink among a category of journalist… Meta analysis isn’t going to change that kind of individual behavior.
Rusty:
If being attacked is unavoidable I would recommend you just own it. Haters gonna hate.
I have no doubt that you are being exhaustive in your methods and process. If a critique has merit and is helpful acknowledge it with gratitude. Let the other stuff slide. You have more important and better uses for your time .
I would ease into any transparency, again more on the time management side of working on developing your project instead of spending time defending it.
Building a Better Bullshit Detector is definitely going to get some attention. Not all positive.
One thing that brings a smile upon my face about all this is, you’re basically building as machine learning model something we did for fun later in university:
Identifying which logical fallacies an argument would contain. Fun times!
Of course, we did this by memorizing the Yuge poster from https://yourlogicalfallacyis.com/ we had in our break room, and then pointing it out manually
I’d naively say: be transparent. With the methods at the very least if not the exact key words (though ideally that too)
For many, the attempts to avoid getting flagged for manipulative language will involve them actually removing manipulative language.
For those that do avoid getting flagged without removing manipulation, it will likely be so convoluted or involve weird phrases to avoid matching your models that most readers will be tipped off by the out of place phrases.
And isn’t that the point? Not to eliminate manipulative news or identify all of it, but to make everyone (readers and journalists alike) more aware of it.
I doubt any level of transparency/secrecy will motivate propagandists to stop publishing propaganda.
Regardless of whether the modal is transparent, if it becomes influential there will be edge case optimization. A transparent model is easier for others to help update to account for that.
To reduce edge case optimization perhaps you could include a score of ‘linguistic normality’? Basically flagging articles that have unusually high concentrations of unusual phrases. Such twisted phrases may be an attempt to duck the model. But this would also be prone to false-positives, particularly for articles aimed at a technical niche (perhaps one could compare across a category)
Regardless, thank you for the fine work, and do let us know if we can help!
Sounds like the chicken and the egg argument maybe. A bit of this and a bit of that.
Though I don’t think being uneducated is the problem, it’s how education is prioritised that is.
It’s not like I ever got a crash course in doubting news information like we get taught how to evaluate data in science. News is just taught to be factual in and of it self. It takes a lot of self learning to pull yourself away from that idea.
I mean they did talk about propaganda in school, but its always in the past tense, like it doesn’t happen now.
It’s a bit of a can of worms I realise, politics being mixed into education will probably end like how religion mixed into education does. Promoting an agenda, rather than providing information.
I think the more significant thing that has changed now from back before, is the ability to leverage technology to reach a wider audience. And all the good and bad that entails. Not that people as individuals have gotten dumber, it’s just easier to craft a narrative for the herd.
And here is the head exploder, school is a form of propaganda. It teaches you what the state needs you to know to be a good member of society. Not necessarily what you need to know to be a better human being.
Another thing is that education is secondary to it’s purpose in freeing the adult workforce to increase productivity.
Even if you could empirically prove that having a stay-at-home parent could improve children’s education, I don’t think the state would be interested in making that happen because it would lock out 1/2 the population from potentially working.
The government is incentivised for Infinite growth at all costs with rewards of that growth to be siphoned to the top of the pyramid. And this system is very well optimised for that purpose.
Growth of the citizens’ wealth or growth of the governing class?
How wealthy are you if you can only share an narrow portion of your precious time with loved ones and your savings wither in the no yield salt flats of the broad consumer price inflation desert?
Academia, a governing class, pah. Let them teach cowering ears lent by command.
How painful, to wield that power and to be able only to exercise it upon children or to claw notionally upwards upon your colleagues thus proclaimed vagaries of thought or expression. The desperation to behold a grand vision before the hour glass leaks out.
Edit: Had a 3yr old future ET reader on my lap “helping” when writing the original post. I fixed a few things but will just leave the other things I feel the urge to change as I feel my point can be understood.
I think it’s inevitable that this will end up becoming similar to the white hat, black hat and gray hat hacker game. Where ET is a white hat and fiat news is the black hat. Over time -if this type of system reaches critical mass like firewalls/antivirus software did for protecting end users, the attack surface decreases significantly.
Please note this analogy is not perfect but it’s what came to mind when reading the post by D_Y.
In regards to D_Y’s second comment:
And to somewhat continue with the internet security analogy, this feels very much like the open source software debate.
Is proprietary/closed source software safer and more trustworthy than open source software? Microsft Windows/MacOS vs Linux/BSD: I think it depends. I personally default to open source options these days myself when the option is available and convienent. However, depending on what I need the software to do and how hard it will be to get running I don’t mind using proprietary solutions. Plus, even closed source software can be understood and manipulated without seeing the code itself.
Regardless of which I end up choosing I can honestly say: I don’t read the source code for nearly any of the open source projects I use. I depend on others to audit the code to confirm it does what the project says it does in both cases. So I personally don’t see a huge difference between the two options since I don’t really know what either code base is really doing behind the scenes.
However, I have found people over time that I somewhat trust who do read open source code. They write up reviews of their findings that I feel I can trust more than a closed source project which has corporate shareholders/investors who want ROI at any cost. That’s pretty much the deciding factor for me and why I try to use open source projects over closed source.
Open source has it’s own issues and isn’t always secure, whether it’s a good actor who contributes poorly written code or a bad actor intentionally writing code which is exploitable, it’s not a perfect solution when it comes to trust and security.
With that said, I do feel that if the true intent of the NEWS system is to make real change it’s best to educate the end user vs making them dependent on software without understanding what it does and trust it’s doing what is intended.
My suggestion: publish papers, publish the code and publish exactly what the ‘key words/phrases’ the system uses, why those words/phrases are important and how it predicts what is and isn’t fiat news.
Make is so the average person who is interested and dedicated to learn could identify what the NEWS system does and be able to do it manually if they wanted (outside of just reading ET of course). Then users could trust the system is doing what it was intended to do and use it simply for the convenience factor.
Educating the end user should be the most important mission factor in my opinion. And yes, using open source and publishing the methods the system uses will allow fiat news to possibly change up methods to avoid detection in some cases not to mention things that we can’t even think of at the moment, aka the Unknown Unknowns. NEWS will then have to update methods. Rinse and repeat.
Allow NEWS to become an open source project and have others to contribute, build on it and fork it to build new and different methods of detecting fiat news. This should be about changing the global perception of what news is today, what it could be and how ET will become the leader in changing news consumption at a global scale.
There is a recent article on the Unz Review about “cognitive infiltration (CI)” and Cass Sunstein (by Ron Unz). If I understand correctly, CI is about deliberately neutralizing fiat (?) news (some of which are ‘conspiracy theories’ which might actually be true) by infusing absurdities and fringe into the narratives. The article notes that the Internet was originally a Darpa project and wonders whether too-much-information overload wasn’t the original intent. Hopefully NEWS will be able to discern what is White/Black Hat as is mentioned above.
I think this article is quite in line with this topic and worth a read.
“In “The Ethics of Belief (1877)”, Clifford gives three arguments as to why we have a moral obligation to believe responsibly , that is, to believe only what we have sufficient evidence for, and what we have diligently investigated. His first argument starts with the simple observation that our beliefs influence our actions.” [my italics]
Fiat News Faces (narratives) used to get us to believe something so as to influence our actions.
“…in our capacity as communicators of belief, we have the moral responsibility not to pollute the well of collective knowledge.”
Seriously where do you guys find these amazing insights and perspectives? I’ve literally never thought of “the moral obligation of belief because belief drives/influences action.”
This at first glance literally blows my mind in terms of how true it is, because the people who have been the most wrong in history have held the strongest beliefs. Or maybe you could even break down “beliefs” down to “confidence”, confidence in being right. Which is a dangerous state to be in because - for lack of a better term - increases the volatility in relation to that knowledge. It allows for extreme actions to be taken because you feel more certain than you should be in some cases.
If nothing else, it does allow me to take a double take on my own personal beliefs. While I feel that I vet my beliefs quite regularly and stringently, making sure it’s based on “some type of evidence”, still, sometimes I’m wrong (shocker! I know). But I never thought of the moral obligation in relation to that.
Sounds stupid as I’m typing it out, but I do think I have felt the effect of being wrong and sharing that wrong view and the implications of that. I did regret having done that that deeply a few times. Not just the personal, “boohoo I’m wrong and my ego hurts”, but, “shit! I told someone actionable information that they may use to make life changing decisions.” Not that they have, or should, but that they could. Scary thought.
But where I think belief and evidence based beliefs diverge is in the state of “unknowing”. If you hold a belief but make sure you keep a part of yourself that still “doubts” it and maybe even discloses that doubt. I feel like that could be the best way to meet your moral obligation.
If nothing else, at least you tried to be right and told people if you weren’t sure before you had full confidence and shared it to the world.
I’m with you. After reading this article, I now think long and hard before I write or say out loud some belief and must have some evidence to back it up. Unfortunately I have to bite my tongue often now because my ear has become attuned to hearing people polluting the well.
Here’s one for you then…
I worked in finance and I hated that client psychology demanded that I project certainty about all things. “I don’t know” was taboo.
I enjoyed “The Undoing Project” by Michael Lewis which, in part, deals with how our minds work and our perception of how our minds work.
Lent it to one of my book sharing buddies with a much different world view of my own and they absolutely hated it.
I have believed for a couple of decades now that we live in a very uncertain world and seek to create certainty within our own personal narratives and the interplay between those two things is where most of the action is.
The more uncertain the times, the stronger the beliefs and stronger narratives emerge.
I agree whole heartedly. To me, human relationships are like plate tectonics. You might get mountains built, you might get rifts OR you might get earth quakes and volcanoes.