Basically: Don't crystallize too early, have a primordial soup of notes that you coagulate/congeal bit by bit. Take little iterative steps on local slices, don't try to construct the final product from the get-go.
This method came quite naturally to me during my writing-based personal projects since I have no deadline or anything and am literally just collecting thousands of little A6/A7 notes that I capture as they pop into my head. I can take all the time I want to stew on them and have a structure bubbling up all on its own.
I suppose some of the beauty of this is that the "notes" are easy. Whereas writing is hard. So the more you can leverage these fragments, the better.
I wonder how the non-fiction element plays into it. I guess if you have a lot of fragments the hard part is organizing and then glueing them together. But in non fiction a lot of this is done or you. You are given the order and a lot of the glue. The glue you are missing is probably relatively obvious ("why did they do this?") and you can get that information and include it, or you cant and there is nothing to be done about it ("the motive is unknown" etc).
Whereas in fiction these are all unknown. You have to decide. And you have to make it compelling, and believable, and maybe astonishing and otherworldly at the same time.
In some limited experience in fictional writing, this is the hard part. I have all these fragments (this happened, this happens, there is this dynamic with 2 characters, some broad themes I want to hit on, etc). Makes me think about if non-fiction stories could be used as a sort of "seed" for the glue. I feel like the content is pretty highly coupled to that sort of pattern though.
> But in non fiction a lot of this is done or you. You are given the order and a lot of the glue. The glue you are missing is probably relatively obvious ("why did they do this?") and you can get that information and include it, or you cant and there is nothing to be done about it ("the motive is unknown" etc).
I published a long time ago and ended up doing some deep reporting on a couple of issues. And honestly, it’s a lot closer to writing fiction than I would have thought going in. In ways, it’s actually a lot harder.
Starting out, I believed roughly what your quote said - that I could get into a deep piece and all the glue would be there. In practice, it’s a lot messier because you try to tell the truth related by witnesses with unpredictable self interests. So as you go through the process, you find multiple truths that can each be corroborated with other sources.
It’s a lot more like building with Lego. A piece could be the nose of a horse or the gutter of a house. It’s where judgment, ethics and good note taking practices come in. Witness problems are a big reason why journalists develop their own note taking style to code interviews - an outlining style gives the freedom to consider each fact as a piece of a larger build.
Edit - If it makes you feel better about writing fiction, I wouldn’t worry so much about plausibility. I wasn’t a great journalist but have heard (and verified) stories that the most creative minds would call impossible. Good characters can make the implausible into magic.
As another Emacs user it'll remain immensely funny to me how Emacs can be shoe-horned into almost any conversation. Probably due to the fact that even when it's not the best tool for the job, by any stretch, it can be kludged into something useful enough.
If there were another writer of non-fiction as deeply researched I'd compare McPhee to, it would be Robert Caro. I already knew from Caro's memoir Working that Caro did not use a tape recorder in his interviews with subjects, and from this article about McPhee's method, I learned that McPhee does not either. I'm a bit surprised: I'd have thought for such deep research one would want a recording to refer back to, but both seem to feel that the drawbacks of influencing their subjects outweigh their benefit.
McPhee was recommended as someone whose writing "makes boring things interesting". I did enjoy The Curve of Binding Energy (nuclear science) and to some extent Coming out of the Country (Alaska). Both of those featured interesting vignettes and colorful characters which propelled along the narrative.
However, I then turned to his magnum opus on geology, Annals of the Former World. That was a long slog which, although I enjoyed moments of it, now I wonder if my time wouldn't have been better spent reading something more interesting.
I own every book McPhee published and have read each one at least twice. He is, without question, the finest writer of non-fiction I know. Annals, as you may know, was originally published as 4 separate volumes, each covering a particular US region. Assembling California is my absolute favorite McPhee work. I have a layman’s interest in geology and plate tectonics that I developed specifically because of this book.
I'm currently reading Assembling California (California resident, so I've seen all the things he discusses in the book, and wondered, like why is Half Dome so big and grey??). Like you I am rapidly developing a layman's interest in plate tectonics.
In every chapter there is a passage like this:
If you could pull up an acre of abyssal plain anywhere in the world -- lift into view a complete column of the ocean floor, from the accumulated sediments at the top to mantle rock at the base -- you would find the sheeted dikes about halfway down. In contrast to the rock columns you find all over the continents -- giddy with time gaps among lithologies of miscellaneous origin and age -- this totem assemblage from the oceans tells a generally consistent story. At its low end is peridotite, the rock of the mantle, tectonically altered in several ways on departure from the spreading center. Above the mantle rock lie the cooled remains of the great magma chamber that released flowing red rock into the spreading center. The chamber, in cooling, tends to form strata, as developing crystals settle within it like snow -- olivine, plagioclase, pyroxene snow -- but above these cumulate bands it becomes essentially a massive gabbro shading upward into plagiogranite as the magmatic juices chemically differentiate themselves in ways that relate to temperature. Just above the granites are the sheeted dikes of diabase, which keep filling the rift between the diverging plates. Above the sheeted dikes, where the fluid rock actually entered the sea, the suddenly chilled extrusions are piled high, like logs outside a sawmill. Because these extrusions have convex ends that bulge smoothly and resemble pillows, they are known in geology as pillow lavas. Above the pillows are the various sediments that have drifted downward through the deep sea -- umbers, ochres, cherts, chalk. Unlike the rest of the crust-and-mantle package, the sediments may hint at the surrounding world.
...
This guy basically writes in a way that transforms the book I am holding in front of my face, physically into a rock. Like some kind of magic trick. He clearly had so much fun writing this, it's amazing and very fun... if rocks tickle you even a little bit this book is worth reading.
Folks, in case it wasn't clear, I was complaining about people downvoting a critical opinion, I was not myself downvoting a critical opinion. I'm glad to see the trend on the original comment has reversed.
This is such a great piece and I admire Somers Atlantic writing. I just want to say, this process feels similar to the process Michael Lewis used in his later books, particular with "Going Infinite", where he embedded deeply and for a long time with FTX, emerging to write just as the story shifted from FTX legitimizing the world of cryptocurrency to becoming its chief villain.
So I do wonder, when you dig in that deeply before coming close to getting any work product, how do you ensure that you'll be able to cut your losses, as opposed to what Lewis did, trying (and failing) to salvage a narrative that had been demolished by events and other reporting?
> McPhee usually had one person at the center of each piece, so he would aim to spend a lot of time with that person … stay at their cottage for a season
Even back when every household received a morning paper I cannot fathom how a single article could command such a high pay.
For a magazine like the New Yorker, there was money. You might be interested in Bryan Burrough's experience writing for Vanity Fair in the 90s and 2000s.
> For twenty-five years, I was contracted to produce three articles a year, long ones, typically ten thousand words. For this, my peak salary was $498,141. That’s not a misprint—$498,141, or more than $166,000 per story. Then, as now, $166,000 was a good advance for an entire book. Yes, I realized it was obscene. I took it with a grin.
> Even back when every household received a morning paper I cannot fathom how a single article could command such a high pay.
He wrote for the New Yorker, which is a magazine rather than a newspaper. The number of long-form literary nonfiction pieces that the New Yorker runs every year is drastically fewer than the number of news articles produced to fill a daily newspaper in just a couple weeks.
Roger Zelazny, the science-fiction author, relates a story of how he wrote one story that fulfilled the briefs of three anthologies/magazines, sold it three times, and it was enough to pay for a cruise.
Less cynically, John McPhee would have been fine if what he made from the New Yorker merely covered his expenses, since he would also publish his work in books.
> I should say that in those nightly encounters I rarely go straight from notes to full sentences. I bridge the gap with a kind of pseudowriting. In effect it’s an outline that takes the form of paragraphs, mixed in with the raw notes. I write good full sentences, broken half-sentences, non-sentences that also include todos in them.
Back when i wrote a lot, i often did something like this. I used a program called Acta [1], which was described as an outliner. Basically a text editor where paragraphs form a tree, and nodes can be collapsed. I'd start with the high-level structure and recursively flesh it out, and just keep doing that until i had nodes at the level of individual sentences. Then i'd just write a sentence for each node, and concatenate them. Often i'd take a shortcut and just write a whole paragraph. But a lot of the time, i wouldn't - a difficult passage becomes easy when you're incrementally breaking it down, reordering it, revising it, etc.
> So much of writing is managing your own emotions. The virtue of “pseudowriting” is that it helps you preserve hope for as long as possible—hope that what you will eventually put in place of those square brackets will be good. Hope keeps you coming back: it is more pleasant and low-stakes to pseudowrite than to fix actual language into the draft; and it is less daunting to return to a document where it feels like all that’s left is for you to fill in some blanks. Do that enough times and you will, in fact, end up with something you can read top to bottom.
This describes how I write a new chunk of research code, often. I'll type along until I get to something like "oh, I'll need to calculate the foo of the widget here," and I'll just put a non-existent function call calculateFoo(widget) there until later, when I'll come back and fill it in. I feel like it keeps it manageable; I'm choosing the level of abstraction that I'm drafting code at, and I come in and fill in the details later. I hadn't connected this idea to the journal articles that I am working on; I typically feel somewhat guilty when I add a FIXME in my LaTex document, but with this framing I see now that that is probably the better way to do it than aiming for a finished paragraph from the get-go. The square brackets and placeholders also seem much nicer that the FIXME I was using. Glad to have seen this at a timely moment for me!
My immediate reaction is to notice that this method is actually closely mirrors Barbara Minto’s Pyramid Principle, a go-to tool for McKinsey consultants for the last 40 years or so. One way to build so-called Pyramids is to go bottom up: gather raw facts, aggregate them into themes, themes align into arguments, and only then does the polished narrative (or the key idea) emerge. As far as I can tell, McPhee recommends a very similar approach.
It's the "too much information" style of writing. The subject has to be worth it. I've read his "Basin and Range" (geology), "The Deltoid Pumpkin Seed" (aircraft design), and "The Curve of Binding Energy" (nuclear weapons). Only the last is worth reading. He went off on a years-long geology tangent, interesting only if you're really into geology.
William Langewiesche had a similar style. He died recently, and leaves behind some good writing on aviation. He was the son of Wolfgang Langewiesche, known for writing "Stick and Rudder", a very well written book for pilots on how airplanes really work.
John McPhee is a treasure. If you haven't read any of his work, I would. And if you don't want to dive into a full book, he has a number of collections of short stories.
What I didn’t realize until a few months ago about writing, James summarized as
> while “writing a first draft” is intimidating, “reading through all your notes” or moving note cards around on a table, contemplating structure, is not. In fact these tasks are kind of delightful.
This seems similar to the method used by Robert Greene (and Ryan Holiday who learned it from him, and many others.) I just started using it (as in: I bought notecards and a box to store them.)
> William L. Howarth, who edited the collection, described McPhee’s method for producing what the New Yorker called “fact pieces,” or deeply reported nonfiction.
He is an essayist and he wrote essays. Avoiding the word “essay” in favor of others is fine but it’s striking in a conversation about John McPhee, of all people.
This is about journalism, in an age when journalism is being bulldozed by cheap content, AI hallucinations and other clickbait tactics. Journalism still has the idea of finding facts, things that correspond to something in reality.
It’s been shown that people in fact prefer to have their beliefs confirmed, rather than challenged by investigative reporting, hence the decline in real journalism, because it is both more difficult and in the end less popular across a large audience.
I disagree with the cynical take in your second sentence.
I think people prefer to be right. Any animal that wants to walk the Earth successfully has an intrinsic need to understand the world as it is, not as they wish it to be. Wanting the fruit tree to be right outside the cave doesn't make it so.
But that desire to have correct beliefs is complex in humans where we also have beliefs about ourselves and others have beliefs about us. We don't want to just be right, but believe ourselves good at being right, and be known to others as someone who is usually right.
That puts some hysteresis in our belief system where we have a tendency to disregard evidence that goes against out beliefs for a while until it's sufficiently compelling.
But, also, no one wants to be a gullible rube or an out-of-touch idiot. We will update our beliefs when in a social and information environment that permits us to do so. The problem today is largely that we have so much choice of our social and information environment, then it's much easier to select an echo chamber that lets us think we're already right then be confronted with the reality that we're wrong.
The decline in investigative journalism can, I think, be explained largely in economic terms. The news has always been driven by ads. Even when people paid for papers, the subscription cost didn't fully cover the newspapers' expenses. The classified ads filled that gap.
Then the Internet arose and transferred advertising out of classifieds and into Craigslist, eBay, and others. And it transferred human attention out of newspapers and into social media. There simply weren't enough resources left to fund expensive investigative journalism.
You are right, I am being casual in asserting what “people prefer”.
You assert that people prefer to be right, I agree but in practice “being right” requires having to accept changes in your beliefs, and that’s painful.
Also “being right” requires some effort to check what the facts are, and that is arduous when people are busy with life.
As a result I think people prefer to feel like they’re right rather than dig deeper, and so they prefer information sources that tend to confirm their beliefs.
Behavior makes a lot more sense when you realize that humans are trying to have as accurate a view of the world as possible while also conserving cognitive resources. Running a brain is actually astonishingly expensive in terms of calories, so we've evolved to be judicious in how much thought we put into things.
Changing an existing belief is hard because we already spent effort acquiring the belief in the first place. Throwing that out should be expensive because otherwise we risk thrashing where we are constantly vacillating between competing beliefs. It makes more sense in terms of efficiency and being able to take action if there is some hysteresis and beliefs are sticky.
Note that while people don't change beliefs easily, they do acquire them pretty easily. If I tell you something that doesn't directly conflict with an existing belief, it's easy to absorb.
I enjoyed your thoughts on people preferring to be right but having competing personal interests.
I have started and stopped writing several chains of thought related to it. But it makes me think about what categories of things we may prefer to be right about. And how they do or do not fit into the model you described. For example, empirical (the elephant weights more than a mouse), pragmatic ("hot water is good for your health" in Chinese culture), moral (murder is bad), etc. "Truth", "being right" etc. are just loaded terms I guess. The claims about being able to function effectively seem largely true for the more empirical stuff but the sort of information silos discussed online draw more idealogical lines.
Also made me think how a nihilist or social constructivist might behave or model things differently. it just seems like if you have a more relativist interpretation of truth, being right and believing in something is just the same thing. So it would be literally impossible for those things to be in conflict (either naturally or from the perspective of the person).
The interesting part is not so much that it’s KEDIT (not that surprising, when these things get mentioned it’s kedit, one of the other xedit editors, xywrite, or wordstar) but that he had someone write note management macros for it with what sounds like a really idiosyncratic workflow.
This approach sounds very similar to the construction of grounded theory in ethnography/anthropology -- something I've always wanted to practise but never had the patience for!
The opening to John McPhee's essay "Los Angeles Against the Mountains" is part of his book The Control of Nature. This first section of that essay is one of my favorite pieces of writing I've ever read.
In Los Angeles versus the San Gabriel Mountains, it is not always clear which side is losing. For example, the Genofiles, Bob and Jackie, can claim to have lost and won. They live on an acre of ground so high that they look across their pool and past the trunks of big pines at an aerial view over Glendale and across Los Angeles to the Pacific bays. The setting, in cool dry air, is serene and Mediterranean. It has not been everlastingly serene.
On a February night some years ago, the Genofiles were awakened by a crash of thunder--lightning striking the mountain front. Ordinarily, in their quiet neighborhood, only the creek beside them was likely to make much sound, dropping steeply out of Shields Canyon on its way to the Los Angeles River. The creek, like every component of all the river systems across the city from the mountains to ocean, had not been left to nature. Its banks were concrete. Its bed was concrete. When boulders were running there, they sounded like a rolling freight. On a night like this, the boulders should have been running. The creek should have been a torrent. Its unnatural sound was unnaturally absent. There was, and had been, a lot of rain.
The Genofiles had two teen-age children, whose rooms were on the uphill side of the one-story house. The window in Scott's room looked straight up Pine Cone Road, a cul-de-sac, which, with hundreds like it, defined the northern limit of the city, the confrontation of the urban and the wild. Los Angeles is overmatched on one side by the Pacific Ocean and by the other by very high mountains. With respect to these principal boundaries, Los Angeles is done sprawling. The San Gabriels, in their state of tectonic young, are rising as rapidly as any range on earth. Their loose inimical slopes flout the tolerance of the angle of repose. Rising straight up out of the megalopolis, they stand ten thousand feet above the nearby sea, and they are not kidding with this city. Shedding, spalling, self-destructing, they are disintegrating at a rate that is among the fastest in the world. The phalanxed communities of Los Angeles have pushed themselves hard against these mountains, an aggression that requires a deep defense budget to content with the results. Kimberlee Genofile called to her mother, who joined her in Scott’s room as they looked up the street From its high turnaround, Pine Cone Road plunges downhill like a ski run, bending left and then right and then left and then right in steep christiania turns for half a mile above a three-hundred-foot straightaway that aims directly at the Genofiles’ house. Not far below the turnaround, Shields Creek passes under the street, and there a kink in its concrete profile had been plugged by a six-foot boulder. Hence the silence of the creek. The water was now spreading across the street. It descended in heavy sheets. As the young Genofiles and their mother glimpsed it all in the all but total darkness, the scene was suddenly illuminated by a blue electrical flash. In the blue light they saw a massive blackness, moving. It was not a landslide, not a mudslide, not a rock avalanche; nor by any means was it the front of a conventional flood. In Jackie’s words, “It was just one big black thing coming at us, rolling, rolling with a lot of water in front of it, pushing the water, this big black thing. It was just one big black hill coming towards us.”
In geology, it would be known as a debris flow. Debris flows amass in stream valleys and more or less resemble fresh concrete. They consist of water mixed with a good deal of solid material, most of which is above sand size. Some of it is Chevrolet size. Boulders bigger than cars ride long distances in debris flows. Boulders grouped like fish eggs pour downhill in debris flows. The dark material coming towards the Genofiles was not only full of boulders, it was so full automobiles it was like bread dough mixed with raisins. On its way down Pine Cone Road, it plucked up cars from driveways and the street. When it crashed into the Genofiles’ house, the shattering of safety glass made terrific explosive sounds. A door burst open. Mud and boulders poured into the hall. We're going to go, Jackie thought. Oh my God, what a hell of a way for the four of us to die together.
The parents' bedroom was on the far side of the house. Bob Genofile was in there kicking through white satin draperies at the paneled glass, smashing it to provide an outlet for the water, when the three others ran in to join him. The walls of the house neither moved nor shook. As a general contractor, Bob had built dams, department stores, hospitals, six schools, seven churches, and this house. It was made of concrete block with steel reinforcement, sixteen inches on center. His wife had said it was stronger than any dam in California. His crew had called it "the fort." It those days, twenty years before, the Genofiles' acre was close by the edge of the mountain brush, but a developer had come along since then and knocked down thousands of trees and put Pine Cone Road up the slope. Now Bob Genofile was thinking, I hope the roof holds. I hope the roof is strong enough to hold. Debris was flowing over it. He told Scott to shut the bedroom door. No sooner was the door closed than it was battered down and fell into the room. Mud, rock, water poured in. It pushed everybody against the far wall. "Jump on the bed," Bob said. The bed began to rise. Kneeling on it--on a gold velvet spread--they could soon press their palms against the ceiling. The bed also moved towards the glass wall. The two teen-agers got off, to try to control the motion, and were pinned between the bed's brass railing and the wall. Boulders went up against the railing, pressed it into their legs, and held them fast. Bob dived into the muck trying to move the boulders, but he failed. The debris flow, entering through the windows as well as the doors, continued to rise. Escape was still possible for the parents but not for the children. The parents looked at each other and did not stir. Each reached for and held one of their children. Their mother felt suddenly resigned, sure that her son and daughter would die and she and her husband would quickly follow. The house became buried to the eaves. Boulders sat on the roof. Thirteen automobiles were packed around the building, including five in the pool. A din of rocks kept banging against them. The stuck horn of a buried car was blaring. The family in the darkness in their fixed tableau watched one another by the light of a directional signal, endlessly blinking. The house had filled up in six minutes, and the mud stopped rising near the children's chins.
> This was most of what I knew about nonfiction writing when I managed to land an assignment, on spec, to profile Douglas Hofstadter for a piece in the Atlantic’s print magazine.
I suspect a large contingent here will really hate this suggestion but here it goes:
The McPhee method sounds like a great framework for making writing prompts. That is, prompts for LLMs to write things.
>in stage one he accumulates notes; in stage two he selects them; in stage three he structures them; and in stage four he writes. By the time he is crafting sentences the structure of the piece as a whole, and of each section, even paragraph, and the logic connecting them all, is already determined, thanks to the mechanical work done in the first three stages. McPhee is on rails the whole time he writes his first draft. From there it’s all downhill and the standard thing that everybody does: revision, revision again, then refinement—a sculptor with ax, then knife, then scalpel.
I know hackernews kinda hates LLMs but I don think this idea has to be so offensive. Much of the work and value from the author is in collecting these fragments and structuring them. Purely from a communication standpoint, I have no issues whatsoever with an LLM stitching them together and choosing the vocab and grammar.
OP describes this ^ long arduous process and then notes:
> Your writing can only be as good as your taste.
That is, using an LLM to help with "collecting these fragments and structuring them" might be okay--should the writer still be able to deeply immerse--but the "stitching them together and choosing the vocab and grammar" of a faux writer LLM is likely to leave a bad taste in the mind of some readers (e.g. those HN types who "kinda hates LLMs").
That's interesting because I thought the exact opposite. How could the LLM generate the notes? Those are from the person's direct observations, investigation, etc. right? So the idea would be to farm that part out? As off-putting as LLM style writing is, letting the LLM make the content seems reckless and error prone.
I do agree that the LLM style vs. the authors style is off-putting. I have used LLMs to help me write things and I do not like how its not "my" voice. I see no reason it couldnt use my voice given some of my writing samples. In any case, I've found it very easy to revise it in my own words. This is part of the "revise" stage mentioned in the process. In addition, this step is barely necessary for some things, such as technical writing.
Basically: Don't crystallize too early, have a primordial soup of notes that you coagulate/congeal bit by bit. Take little iterative steps on local slices, don't try to construct the final product from the get-go.
This method came quite naturally to me during my writing-based personal projects since I have no deadline or anything and am literally just collecting thousands of little A6/A7 notes that I capture as they pop into my head. I can take all the time I want to stew on them and have a structure bubbling up all on its own.
I suppose some of the beauty of this is that the "notes" are easy. Whereas writing is hard. So the more you can leverage these fragments, the better.
I wonder how the non-fiction element plays into it. I guess if you have a lot of fragments the hard part is organizing and then glueing them together. But in non fiction a lot of this is done or you. You are given the order and a lot of the glue. The glue you are missing is probably relatively obvious ("why did they do this?") and you can get that information and include it, or you cant and there is nothing to be done about it ("the motive is unknown" etc).
Whereas in fiction these are all unknown. You have to decide. And you have to make it compelling, and believable, and maybe astonishing and otherworldly at the same time.
In some limited experience in fictional writing, this is the hard part. I have all these fragments (this happened, this happens, there is this dynamic with 2 characters, some broad themes I want to hit on, etc). Makes me think about if non-fiction stories could be used as a sort of "seed" for the glue. I feel like the content is pretty highly coupled to that sort of pattern though.
> But in non fiction a lot of this is done or you. You are given the order and a lot of the glue. The glue you are missing is probably relatively obvious ("why did they do this?") and you can get that information and include it, or you cant and there is nothing to be done about it ("the motive is unknown" etc).
I published a long time ago and ended up doing some deep reporting on a couple of issues. And honestly, it’s a lot closer to writing fiction than I would have thought going in. In ways, it’s actually a lot harder.
Starting out, I believed roughly what your quote said - that I could get into a deep piece and all the glue would be there. In practice, it’s a lot messier because you try to tell the truth related by witnesses with unpredictable self interests. So as you go through the process, you find multiple truths that can each be corroborated with other sources.
It’s a lot more like building with Lego. A piece could be the nose of a horse or the gutter of a house. It’s where judgment, ethics and good note taking practices come in. Witness problems are a big reason why journalists develop their own note taking style to code interviews - an outlining style gives the freedom to consider each fact as a piece of a larger build.
Edit - If it makes you feel better about writing fiction, I wouldn’t worry so much about plausibility. I wasn’t a great journalist but have heard (and verified) stories that the most creative minds would call impossible. Good characters can make the implausible into magic.
...and use Emacs. Very important.
As another Emacs user it'll remain immensely funny to me how Emacs can be shoe-horned into almost any conversation. Probably due to the fact that even when it's not the best tool for the job, by any stretch, it can be kludged into something useful enough.
And yeah, regarding writing (fiction and non-fiction) there's Emacs Writing Studio which people praise and I haven't personally used: https://github.com/pprevos/emacs-writing-studio
There's also some packages like Olivetti or Writeroom-mode for just a "Zen mode" writing experience.
If there were another writer of non-fiction as deeply researched I'd compare McPhee to, it would be Robert Caro. I already knew from Caro's memoir Working that Caro did not use a tape recorder in his interviews with subjects, and from this article about McPhee's method, I learned that McPhee does not either. I'm a bit surprised: I'd have thought for such deep research one would want a recording to refer back to, but both seem to feel that the drawbacks of influencing their subjects outweigh their benefit.
McPhee was recommended as someone whose writing "makes boring things interesting". I did enjoy The Curve of Binding Energy (nuclear science) and to some extent Coming out of the Country (Alaska). Both of those featured interesting vignettes and colorful characters which propelled along the narrative.
However, I then turned to his magnum opus on geology, Annals of the Former World. That was a long slog which, although I enjoyed moments of it, now I wonder if my time wouldn't have been better spent reading something more interesting.
I own every book McPhee published and have read each one at least twice. He is, without question, the finest writer of non-fiction I know. Annals, as you may know, was originally published as 4 separate volumes, each covering a particular US region. Assembling California is my absolute favorite McPhee work. I have a layman’s interest in geology and plate tectonics that I developed specifically because of this book.
I'm currently reading Assembling California (California resident, so I've seen all the things he discusses in the book, and wondered, like why is Half Dome so big and grey??). Like you I am rapidly developing a layman's interest in plate tectonics.
In every chapter there is a passage like this:
If you could pull up an acre of abyssal plain anywhere in the world -- lift into view a complete column of the ocean floor, from the accumulated sediments at the top to mantle rock at the base -- you would find the sheeted dikes about halfway down. In contrast to the rock columns you find all over the continents -- giddy with time gaps among lithologies of miscellaneous origin and age -- this totem assemblage from the oceans tells a generally consistent story. At its low end is peridotite, the rock of the mantle, tectonically altered in several ways on departure from the spreading center. Above the mantle rock lie the cooled remains of the great magma chamber that released flowing red rock into the spreading center. The chamber, in cooling, tends to form strata, as developing crystals settle within it like snow -- olivine, plagioclase, pyroxene snow -- but above these cumulate bands it becomes essentially a massive gabbro shading upward into plagiogranite as the magmatic juices chemically differentiate themselves in ways that relate to temperature. Just above the granites are the sheeted dikes of diabase, which keep filling the rift between the diverging plates. Above the sheeted dikes, where the fluid rock actually entered the sea, the suddenly chilled extrusions are piled high, like logs outside a sawmill. Because these extrusions have convex ends that bulge smoothly and resemble pillows, they are known in geology as pillow lavas. Above the pillows are the various sediments that have drifted downward through the deep sea -- umbers, ochres, cherts, chalk. Unlike the rest of the crust-and-mantle package, the sediments may hint at the surrounding world.
...
This guy basically writes in a way that transforms the book I am holding in front of my face, physically into a rock. Like some kind of magic trick. He clearly had so much fun writing this, it's amazing and very fun... if rocks tickle you even a little bit this book is worth reading.
It's giving me little crunchy Dwarf Fortress dopamine tingles.
For me, it’s "The Pine Barrens". It’s boring, but he did make it interesting. For a while, at least.
Sharing a critical opinion? That's a downvote for you! (Sheesh)
Folks, in case it wasn't clear, I was complaining about people downvoting a critical opinion, I was not myself downvoting a critical opinion. I'm glad to see the trend on the original comment has reversed.
This is such a great piece and I admire Somers Atlantic writing. I just want to say, this process feels similar to the process Michael Lewis used in his later books, particular with "Going Infinite", where he embedded deeply and for a long time with FTX, emerging to write just as the story shifted from FTX legitimizing the world of cryptocurrency to becoming its chief villain.
So I do wonder, when you dig in that deeply before coming close to getting any work product, how do you ensure that you'll be able to cut your losses, as opposed to what Lewis did, trying (and failing) to salvage a narrative that had been demolished by events and other reporting?
Another lesson you could take away there: It's possible to do this sort of total deep immersion for months, and also completely miss the real story.
> McPhee usually had one person at the center of each piece, so he would aim to spend a lot of time with that person … stay at their cottage for a season
Even back when every household received a morning paper I cannot fathom how a single article could command such a high pay.
For a magazine like the New Yorker, there was money. You might be interested in Bryan Burrough's experience writing for Vanity Fair in the 90s and 2000s.
> For twenty-five years, I was contracted to produce three articles a year, long ones, typically ten thousand words. For this, my peak salary was $498,141. That’s not a misprint—$498,141, or more than $166,000 per story. Then, as now, $166,000 was a good advance for an entire book. Yes, I realized it was obscene. I took it with a grin.
https://yalereview.org/article/burrough-vanity-fair-graydon-...
> Even back when every household received a morning paper I cannot fathom how a single article could command such a high pay.
He wrote for the New Yorker, which is a magazine rather than a newspaper. The number of long-form literary nonfiction pieces that the New Yorker runs every year is drastically fewer than the number of news articles produced to fill a daily newspaper in just a couple weeks.
How do you think Googlers command such high pay?
Roger Zelazny, the science-fiction author, relates a story of how he wrote one story that fulfilled the briefs of three anthologies/magazines, sold it three times, and it was enough to pay for a cruise.
Things were different back in the day.
Less cynically, John McPhee would have been fine if what he made from the New Yorker merely covered his expenses, since he would also publish his work in books.
> I should say that in those nightly encounters I rarely go straight from notes to full sentences. I bridge the gap with a kind of pseudowriting. In effect it’s an outline that takes the form of paragraphs, mixed in with the raw notes. I write good full sentences, broken half-sentences, non-sentences that also include todos in them.
Back when i wrote a lot, i often did something like this. I used a program called Acta [1], which was described as an outliner. Basically a text editor where paragraphs form a tree, and nodes can be collapsed. I'd start with the high-level structure and recursively flesh it out, and just keep doing that until i had nodes at the level of individual sentences. Then i'd just write a sentence for each node, and concatenate them. Often i'd take a shortcut and just write a whole paragraph. But a lot of the time, i wouldn't - a difficult passage becomes easy when you're incrementally breaking it down, reordering it, revising it, etc.
[1] https://www.a-sharp.com/acta/
Sounds a bit like Gingko, which I used to really love:
https://gingkowriter.com/
Same, I really loved that they finally innovated on text editing interface. Couldn't get used to be honest.
As usual, there's an Obsidian plugin nowadays trying to mimic (some of?) it: https://github.com/ycnmhd/obsidian-lineage
> So much of writing is managing your own emotions. The virtue of “pseudowriting” is that it helps you preserve hope for as long as possible—hope that what you will eventually put in place of those square brackets will be good. Hope keeps you coming back: it is more pleasant and low-stakes to pseudowrite than to fix actual language into the draft; and it is less daunting to return to a document where it feels like all that’s left is for you to fill in some blanks. Do that enough times and you will, in fact, end up with something you can read top to bottom.
This describes how I write a new chunk of research code, often. I'll type along until I get to something like "oh, I'll need to calculate the foo of the widget here," and I'll just put a non-existent function call calculateFoo(widget) there until later, when I'll come back and fill it in. I feel like it keeps it manageable; I'm choosing the level of abstraction that I'm drafting code at, and I come in and fill in the details later. I hadn't connected this idea to the journal articles that I am working on; I typically feel somewhat guilty when I add a FIXME in my LaTex document, but with this framing I see now that that is probably the better way to do it than aiming for a finished paragraph from the get-go. The square brackets and placeholders also seem much nicer that the FIXME I was using. Glad to have seen this at a timely moment for me!
My immediate reaction is to notice that this method is actually closely mirrors Barbara Minto’s Pyramid Principle, a go-to tool for McKinsey consultants for the last 40 years or so. One way to build so-called Pyramids is to go bottom up: gather raw facts, aggregate them into themes, themes align into arguments, and only then does the polished narrative (or the key idea) emerge. As far as I can tell, McPhee recommends a very similar approach.
"McPhee has been my model. He's the most elegant of all the journalists writing today, I think."
— Tracy Kidder, author of The Soul of a New Machine and many other great writings.
It's the "too much information" style of writing. The subject has to be worth it. I've read his "Basin and Range" (geology), "The Deltoid Pumpkin Seed" (aircraft design), and "The Curve of Binding Energy" (nuclear weapons). Only the last is worth reading. He went off on a years-long geology tangent, interesting only if you're really into geology.
William Langewiesche had a similar style. He died recently, and leaves behind some good writing on aviation. He was the son of Wolfgang Langewiesche, known for writing "Stick and Rudder", a very well written book for pilots on how airplanes really work.
John McPhee is a treasure. If you haven't read any of his work, I would. And if you don't want to dive into a full book, he has a number of collections of short stories.
I absolutely loved this:
> if you tell someone you’re a journalist they’re going to believe you. Your job is to honor their trust.
A lot of life wisdom wedged into those two sentences, I think, that goes beyond journalism.
What I didn’t realize until a few months ago about writing, James summarized as
> while “writing a first draft” is intimidating, “reading through all your notes” or moving note cards around on a table, contemplating structure, is not. In fact these tasks are kind of delightful.
Thanks for sharing.
This seems similar to the method used by Robert Greene (and Ryan Holiday who learned it from him, and many others.) I just started using it (as in: I bought notecards and a box to store them.)
This would be a good overview of the method: https://billyoppenheimer.com/notecard-system/
And Umberto Eco wrote a book on this model before Greene. "How to Write a Thesis"
Thank you, now I have to get this and put it in my already too high pile of books to read !
> William L. Howarth, who edited the collection, described McPhee’s method for producing what the New Yorker called “fact pieces,” or deeply reported nonfiction.
He is an essayist and he wrote essays. Avoiding the word “essay” in favor of others is fine but it’s striking in a conversation about John McPhee, of all people.
“Fact pieces.” Ick.
This is about journalism, in an age when journalism is being bulldozed by cheap content, AI hallucinations and other clickbait tactics. Journalism still has the idea of finding facts, things that correspond to something in reality.
It’s been shown that people in fact prefer to have their beliefs confirmed, rather than challenged by investigative reporting, hence the decline in real journalism, because it is both more difficult and in the end less popular across a large audience.
I disagree with the cynical take in your second sentence.
I think people prefer to be right. Any animal that wants to walk the Earth successfully has an intrinsic need to understand the world as it is, not as they wish it to be. Wanting the fruit tree to be right outside the cave doesn't make it so.
But that desire to have correct beliefs is complex in humans where we also have beliefs about ourselves and others have beliefs about us. We don't want to just be right, but believe ourselves good at being right, and be known to others as someone who is usually right.
That puts some hysteresis in our belief system where we have a tendency to disregard evidence that goes against out beliefs for a while until it's sufficiently compelling.
But, also, no one wants to be a gullible rube or an out-of-touch idiot. We will update our beliefs when in a social and information environment that permits us to do so. The problem today is largely that we have so much choice of our social and information environment, then it's much easier to select an echo chamber that lets us think we're already right then be confronted with the reality that we're wrong.
The decline in investigative journalism can, I think, be explained largely in economic terms. The news has always been driven by ads. Even when people paid for papers, the subscription cost didn't fully cover the newspapers' expenses. The classified ads filled that gap.
Then the Internet arose and transferred advertising out of classifieds and into Craigslist, eBay, and others. And it transferred human attention out of newspapers and into social media. There simply weren't enough resources left to fund expensive investigative journalism.
You are right, I am being casual in asserting what “people prefer”.
You assert that people prefer to be right, I agree but in practice “being right” requires having to accept changes in your beliefs, and that’s painful.
Also “being right” requires some effort to check what the facts are, and that is arduous when people are busy with life.
As a result I think people prefer to feel like they’re right rather than dig deeper, and so they prefer information sources that tend to confirm their beliefs.
Behavior makes a lot more sense when you realize that humans are trying to have as accurate a view of the world as possible while also conserving cognitive resources. Running a brain is actually astonishingly expensive in terms of calories, so we've evolved to be judicious in how much thought we put into things.
Changing an existing belief is hard because we already spent effort acquiring the belief in the first place. Throwing that out should be expensive because otherwise we risk thrashing where we are constantly vacillating between competing beliefs. It makes more sense in terms of efficiency and being able to take action if there is some hysteresis and beliefs are sticky.
Note that while people don't change beliefs easily, they do acquire them pretty easily. If I tell you something that doesn't directly conflict with an existing belief, it's easy to absorb.
I enjoyed your thoughts on people preferring to be right but having competing personal interests.
I have started and stopped writing several chains of thought related to it. But it makes me think about what categories of things we may prefer to be right about. And how they do or do not fit into the model you described. For example, empirical (the elephant weights more than a mouse), pragmatic ("hot water is good for your health" in Chinese culture), moral (murder is bad), etc. "Truth", "being right" etc. are just loaded terms I guess. The claims about being able to function effectively seem largely true for the more empirical stuff but the sort of information silos discussed online draw more idealogical lines.
Also made me think how a nihilist or social constructivist might behave or model things differently. it just seems like if you have a more relativist interpretation of truth, being right and believing in something is just the same thing. So it would be literally impossible for those things to be in conflict (either naturally or from the perspective of the person).
Don't miss the discussion of McPhee's text editor. I would love to get more details about that. Catnip for this crowd.
The interesting part is not so much that it’s KEDIT (not that surprising, when these things get mentioned it’s kedit, one of the other xedit editors, xywrite, or wordstar) but that he had someone write note management macros for it with what sounds like a really idiosyncratic workflow.
I suppose Scrivener would be the modern equivalent.
https://www.literatureandlatte.com/scrivener/overview
This approach sounds very similar to the construction of grounded theory in ethnography/anthropology -- something I've always wanted to practise but never had the patience for!
The opening to John McPhee's essay "Los Angeles Against the Mountains" is part of his book The Control of Nature. This first section of that essay is one of my favorite pieces of writing I've ever read.
In Los Angeles versus the San Gabriel Mountains, it is not always clear which side is losing. For example, the Genofiles, Bob and Jackie, can claim to have lost and won. They live on an acre of ground so high that they look across their pool and past the trunks of big pines at an aerial view over Glendale and across Los Angeles to the Pacific bays. The setting, in cool dry air, is serene and Mediterranean. It has not been everlastingly serene.
On a February night some years ago, the Genofiles were awakened by a crash of thunder--lightning striking the mountain front. Ordinarily, in their quiet neighborhood, only the creek beside them was likely to make much sound, dropping steeply out of Shields Canyon on its way to the Los Angeles River. The creek, like every component of all the river systems across the city from the mountains to ocean, had not been left to nature. Its banks were concrete. Its bed was concrete. When boulders were running there, they sounded like a rolling freight. On a night like this, the boulders should have been running. The creek should have been a torrent. Its unnatural sound was unnaturally absent. There was, and had been, a lot of rain.
The Genofiles had two teen-age children, whose rooms were on the uphill side of the one-story house. The window in Scott's room looked straight up Pine Cone Road, a cul-de-sac, which, with hundreds like it, defined the northern limit of the city, the confrontation of the urban and the wild. Los Angeles is overmatched on one side by the Pacific Ocean and by the other by very high mountains. With respect to these principal boundaries, Los Angeles is done sprawling. The San Gabriels, in their state of tectonic young, are rising as rapidly as any range on earth. Their loose inimical slopes flout the tolerance of the angle of repose. Rising straight up out of the megalopolis, they stand ten thousand feet above the nearby sea, and they are not kidding with this city. Shedding, spalling, self-destructing, they are disintegrating at a rate that is among the fastest in the world. The phalanxed communities of Los Angeles have pushed themselves hard against these mountains, an aggression that requires a deep defense budget to content with the results. Kimberlee Genofile called to her mother, who joined her in Scott’s room as they looked up the street From its high turnaround, Pine Cone Road plunges downhill like a ski run, bending left and then right and then left and then right in steep christiania turns for half a mile above a three-hundred-foot straightaway that aims directly at the Genofiles’ house. Not far below the turnaround, Shields Creek passes under the street, and there a kink in its concrete profile had been plugged by a six-foot boulder. Hence the silence of the creek. The water was now spreading across the street. It descended in heavy sheets. As the young Genofiles and their mother glimpsed it all in the all but total darkness, the scene was suddenly illuminated by a blue electrical flash. In the blue light they saw a massive blackness, moving. It was not a landslide, not a mudslide, not a rock avalanche; nor by any means was it the front of a conventional flood. In Jackie’s words, “It was just one big black thing coming at us, rolling, rolling with a lot of water in front of it, pushing the water, this big black thing. It was just one big black hill coming towards us.”
In geology, it would be known as a debris flow. Debris flows amass in stream valleys and more or less resemble fresh concrete. They consist of water mixed with a good deal of solid material, most of which is above sand size. Some of it is Chevrolet size. Boulders bigger than cars ride long distances in debris flows. Boulders grouped like fish eggs pour downhill in debris flows. The dark material coming towards the Genofiles was not only full of boulders, it was so full automobiles it was like bread dough mixed with raisins. On its way down Pine Cone Road, it plucked up cars from driveways and the street. When it crashed into the Genofiles’ house, the shattering of safety glass made terrific explosive sounds. A door burst open. Mud and boulders poured into the hall. We're going to go, Jackie thought. Oh my God, what a hell of a way for the four of us to die together.
The parents' bedroom was on the far side of the house. Bob Genofile was in there kicking through white satin draperies at the paneled glass, smashing it to provide an outlet for the water, when the three others ran in to join him. The walls of the house neither moved nor shook. As a general contractor, Bob had built dams, department stores, hospitals, six schools, seven churches, and this house. It was made of concrete block with steel reinforcement, sixteen inches on center. His wife had said it was stronger than any dam in California. His crew had called it "the fort." It those days, twenty years before, the Genofiles' acre was close by the edge of the mountain brush, but a developer had come along since then and knocked down thousands of trees and put Pine Cone Road up the slope. Now Bob Genofile was thinking, I hope the roof holds. I hope the roof is strong enough to hold. Debris was flowing over it. He told Scott to shut the bedroom door. No sooner was the door closed than it was battered down and fell into the room. Mud, rock, water poured in. It pushed everybody against the far wall. "Jump on the bed," Bob said. The bed began to rise. Kneeling on it--on a gold velvet spread--they could soon press their palms against the ceiling. The bed also moved towards the glass wall. The two teen-agers got off, to try to control the motion, and were pinned between the bed's brass railing and the wall. Boulders went up against the railing, pressed it into their legs, and held them fast. Bob dived into the muck trying to move the boulders, but he failed. The debris flow, entering through the windows as well as the doors, continued to rise. Escape was still possible for the parents but not for the children. The parents looked at each other and did not stir. Each reached for and held one of their children. Their mother felt suddenly resigned, sure that her son and daughter would die and she and her husband would quickly follow. The house became buried to the eaves. Boulders sat on the roof. Thirteen automobiles were packed around the building, including five in the pool. A din of rocks kept banging against them. The stuck horn of a buried car was blaring. The family in the darkness in their fixed tableau watched one another by the light of a directional signal, endlessly blinking. The house had filled up in six minutes, and the mud stopped rising near the children's chins.
> This was most of what I knew about nonfiction writing when I managed to land an assignment, on spec, to profile Douglas Hofstadter for a piece in the Atlantic’s print magazine.
That sounds like an impressively rapid climb.
This is a beautiful website. The article and homepage.
It looks directly inspired by the design sensibility and aesthetics of Edward Tufte.
I suspect a large contingent here will really hate this suggestion but here it goes:
The McPhee method sounds like a great framework for making writing prompts. That is, prompts for LLMs to write things.
>in stage one he accumulates notes; in stage two he selects them; in stage three he structures them; and in stage four he writes. By the time he is crafting sentences the structure of the piece as a whole, and of each section, even paragraph, and the logic connecting them all, is already determined, thanks to the mechanical work done in the first three stages. McPhee is on rails the whole time he writes his first draft. From there it’s all downhill and the standard thing that everybody does: revision, revision again, then refinement—a sculptor with ax, then knife, then scalpel.
I know hackernews kinda hates LLMs but I don think this idea has to be so offensive. Much of the work and value from the author is in collecting these fragments and structuring them. Purely from a communication standpoint, I have no issues whatsoever with an LLM stitching them together and choosing the vocab and grammar.
OP describes this ^ long arduous process and then notes:
> Your writing can only be as good as your taste.
That is, using an LLM to help with "collecting these fragments and structuring them" might be okay--should the writer still be able to deeply immerse--but the "stitching them together and choosing the vocab and grammar" of a faux writer LLM is likely to leave a bad taste in the mind of some readers (e.g. those HN types who "kinda hates LLMs").
That's interesting because I thought the exact opposite. How could the LLM generate the notes? Those are from the person's direct observations, investigation, etc. right? So the idea would be to farm that part out? As off-putting as LLM style writing is, letting the LLM make the content seems reckless and error prone.
I do agree that the LLM style vs. the authors style is off-putting. I have used LLMs to help me write things and I do not like how its not "my" voice. I see no reason it couldnt use my voice given some of my writing samples. In any case, I've found it very easy to revise it in my own words. This is part of the "revise" stage mentioned in the process. In addition, this step is barely necessary for some things, such as technical writing.
[dead]
This is a well-written, interesting article. The only issue I have is with this:
> He said, if you tell someone you’re a journalist they’re going to believe you. Your job is to honor their trust.
Sadly, that is no longer true. Today's "journalists" far too often see their job as proselytizing rather than reporting.