Seeking Density in the Gonzo Theater

Consider this thought experiment: what if you were only allowed 2000 words with which to understand the world?

With these 2000 words, you’d have to do everything. You’d be allowed to occasionally retire some words in favor of others, or invent new words, but you’d have to stick to the budget.

Everything would have to be expressible within the budget: everyday conversations and deep conversations, shallow thoughts and  profound ones, reflections and expectations, scientific propositions and vocational instruction manuals, poetry and stories, emotions and facts.

How would you use your budget? Would you choose more nouns or verbs? How many friends would you elevate to a name-remembered status? How many stars and bird species would you name? Would you have more concrete words or more reified ones in your selection? How many of the most commonly used words would you select? Counting mathematical symbols as words, how many of those would you select? Would you mimic others’ selections or make up your own mind?


When I read old texts, I am struck by the density of the writing. Words used to be expensive. You had to make one word do many things.

That last sentence contains a simple example. I originally had convey many meanings in place of do many things. For some readers, the substitution will make no difference. To others, it will make a great deal of difference.

We talk of dense texts as being layered.  They lend themselves to re-reading from many perspectives over a long period of time. Even as late as the nineteenth century, we find that the average professional writer wrote with a density that rivals the densest writing today.  With the exception of scientific writing — best understood as a social-industrial process for increasing the density of words — every other kind of writing today has become less layered. Most writing admits one reading, if that.

Dense writing is not particularly difficult. Merely time-consuming. As the word layering suggests, it is something of a mechanical craft, and you become better with practice. Even mediocre writers in the past, working with starter material no denser than today’s typical Top 10 blog post, could sometimes achieve sublime results by putting in the time.

If the mediocre can become good by pursuing density, the good can become great. Robert Louis Stevenson famously wrote gripping action sequences without using adverbs and adjectives. His prose has a sparse elegance to it, but is nevertheless dense with meaning and drama. I once tried the exercise of avoiding adverbs and adjectives. I discovered that it is not about elimination. The main challenge is to make your nouns and verbs do more work.


In teaching and learning writing today, we focus on the isolated virtue of brevity. We do not think about density. Traditions of exegesis — the dilution, usually oral, of dense texts to the point where they are consumable by many — are confined to dead rather than living texts.

We have forgotten how to teach density. In fact, we’ve even forgotten how to think  about it. We confuse density with overwrought, baroque textures, with a hard-to-handle literary style that can easily turn into tasteless excess in unskilled hands.

The 2000-word thought experiment, if you try it, will likely force you to consider density of meaning as a selection factor. Some words, like schadenfreude, are intrinsically dense. Others, like love, are dense because they are highly adaptable.  Depending on context, they can do many things.

Density is a more fundamental variable than the length of a text. It is intrinsic to writing, like the density of a fluid; what is known in fluid dynamics as an intensive property. The length of an arbitrarily delineated piece of text on the other hand, is an extensive property, like the volume of a specified container of fluid.

Choosing words precisely and crafting dense sentences is important. Choosing small containers is not.


Writing used to be a form of making. I sometimes wonder what it would be like to have to carve your thoughts onto stone tablets.  One of these days I am going to try carving the first draft of a post in stone.

Writing on paper is also an expensive luxury. There was a time when writers made their own paper and ink. You had to write with temperamental things like quills. The practice of calligraphy was not a writerly affectation. It was a necessary skill in the days of temperamental media.

The scribe was more of an archivist than a writer. The other ancestor of the writer, the bard-sage, was both composer and performer. The average person did not read, but relied on the bard or priest to expand upon and perform the written, archived word. Particularly good performances would lead to revisions of the written texts.

When fountain pens and cheap factory-made paper made their appearance, writers were able to waste paper, and as a consequence, written words. In the history of thought, the invention of the ability to waste words was probably as important as the invention of the ability — famously noted by Alan Kay — to waste bits in the history of programming.

With cheap paper was born that iconic image of the twentieth century writer — a writer sitting alone in a room, crumpling up a piece of writing in frustration, and tossing it into an overflowing waste-paper basket. Unlike the sage-bard, enacting old texts and beta-testing new ones through public oral performances, or the scribe, committing tested, quality-controlled and expensive texts to stone, the modern, pre-Internet writer was a resource-rich creature of profligate excess. 

The very idea of a “waste paper” basket would have been unthinkable at one time.


It is difficult today to get a sense of how expensive writing used to be. I once watched a traditional temple scribe demonstrate the process of making the palm-leaf manuscripts that were used in India until Islam brought paper-making to the subcontinent. That probably happened a few centuries after the Abbasids defeated the Tang empire at the Battle of Talas in 751 AD, and extracted the secret of paper-making from Chinese prisoners of war.

Palm leaves are easily the worst writing technology ever invented by a major culture. They make leather, papyrus, paper and silk look like space-age media by comparison. A good deal that seems strange about India as an idea suddenly makes sense once you get that the civilization was being enacted through this ridiculous medium (and equally ridiculous ones like tree bark) until about 1000 AD. Imagine a modern civilization that had to keep its grand narrative going using only tweets, and you get some sense of what was going on.

Here’s how you make palm-leaf manuscripts. First you cut little index-card sized rectangles out of palm fronds and dry them flat. Then you carefully use a needle to scratch out the text — typically a few lines per leaf.  Then you make an ink out of ground charcoal, carefully rub it into the scratches, and swab away the excess. Finally, you carefully pierce a hole through the middle (not the edge, since the thing is brittle) and thread a piece of string through a sheaf of loosely-related leaves.

Congratulations, you have a book.

Since the sheaf is more unstable than individual leaves, you have to plan for graceful degradation. Expect individual leaves to be lost or damaged. Expect accidental shuffling and page numbering turning to garbage. Expect new leaves to be inserted, like viruses. Don’t expect multi-leaf stories to remain stable. Expect narrative trunks to sprout branches added by later authors.

The palm leaf manuscript was brittle and easily damaged, available in one unhelpful size, with a lifespan of perhaps a few decades on average (carefully preserved ones lasted around 150 years I believe). After that you had to make a copy if you wanted to keep the ideas alive. If you were rich or powerful, you could get stuff carved onto stone or copper plates by slaves. If not, your best bet was to go with palm leaves and hope that people would descend on your home to make copies.


When you look at old writing technology, poetry suddenly makes sense.

It is  modular content that comes in fixed-length chunks, with redundancy and error-correcting codes built in. It is designed to be transmitted and copied across time and space through unreliable and noisy channels, one stone tablet, palm leaf or piece of handmade paper at a time. The technology was still unreliable enough that the oral tradition remained the primary channel. Writing began as a medium for backups. Scribes were the first data warehousing experts. They did more than merely transcribe the spoken word. They compressed, corrected and encrypted as well, and periodically updated texts to reflect the extant state of the oral tradition.

That is why verses are so eminently quotable outside the context of poems. Poems are extensive oral containers of arbitrary length, in some cases delineated after the fact. Verses are standardized containers designed to carry intense, dense, archival-quality words around.

Today we view traditional verse epics as single works. The Illiad has about 9000 verses. The Mahabharata has about 24,000. It makes far more sense to talk about both as data-warehoused records of extremely long — in both time and words — convergent conversations. They are closer to Google’s index than to books.

For the ancients, texts had to be little metered packets. But as paper technology got cheaper and more reliable, poetry, like many other obsolete technologies before and after, turned into an art form. Critical function turned into dispensable style. Meter and rhyme ceased to be useful as error-correcting coding mechanisms and turned into free dimensions for artistic expression.

Soon, individual verses could be composed under the assumption of stable, longer embedding contexts. Extensive works could be delineated  a priori, during the composition of the parts. And the parts could be safely de-containerized. Rhyming verse could be abandoned in favor of blank verse, and eventually meter became entirely unnecessary. And we ended up with the bound book of prose.

Technologically, it was something of a backward step, like reverting to circuit-switched networks after having invented packet switching, or moving back from digital to analog technology. But it served an important purpose: allowing the individual writer to emerge. The book could belong to an individual author in a way a verse from an oral tradition could not.


Poetry gets it right: length is irrelevant. You can standardize and normalize it away using appropriate containerization. It is density that matters. Evolving your packet size and vocabulary over time helps you increase density over time.

My posts range between 2000-5000 words, and I post about once a week here on ribbonfarm. But there are many bloggers who post two or three 300-word posts a day, five days a week. They also log 2000-5000 words.

So I am not particularly prolific. I merely have a different packet size compared to other bloggers, optimized for a peculiar purpose: evolving an idiosyncratic vocabulary. It seems to take several thousand words to characterize a neologism like gollumize or posturetalk But once that is done, I can reuse it as a compact and dense piece of refactored perception.

You could say that what I am really trying to do on this blog is compose a speculative dictionary of dense words and phrases. Perhaps one day this blog will collapse under its own gravity into a single super-dense post written entirely with 2000 hyperlinked neologisms, like a neutron star.

Poetry — functional ancient poetry, the cultural TCP/IP of the world before around 1000 AD — is necessarily a social process, involving, at the very least, a sage-bard, a scribe, an audience and a patron. The oral culture refines, distills, tests, reworks, debates and judges. Iterative performance is a necessary component. When oral exegesis of an unstable verse dies down, and memorization and repetition validate the quality of the finished verse, the scribe breaks out his chisel.

The prose book can stand apart from broader social processes in radically individual ways. It can travel from writer to readers largely unaltered, setting up a hub-spoke pattern of conversational circuits.


I’ve occasionally described my blogging as a sort of performance art. But something about that self-description has been bothering me. I have now concluded that if the description applies at all, it applies to a different kind of blogger, not me.

The Web obscures the crucial and necessary distinction between oral and written cultures.  Some bloggers perform and talk. Others are scribes. I think I am a scribe, not a performer.

Yet, there is no easy correspondence between pre-Gutenberg bard-sages and scribes and today’s bloggers. In the intervening centuries, we have seen the rise and fall of the individualist writer, working alone, filling waste-paper baskets.

History does not rewind. It synthesizes. The blogosphere, I am convinced, synthesizes the collectivist pre-Gutenberg culture of sage-bard and scribes with the individualist post-Gutenberg culture of paper-crumpling waste-paper-basket fillers.

In the process of synthesis, virtual circuits must ride once more on top of a revitalized packet-switched network. The oral/written distinction must be replaced by a more basic one that is medium-agnostic, like the Internet itself.


According to legend, the sage Vyasa needed a scribe to write down the Mahabharata as he composed it. Ganesha accepted the challenge, but demanded that the sage compose as fast as he could write. Wary of the trickster god, Vyasa in turn set his own condition: Ganesha would have to understand every verse before writing it down. And so, the legend continues, they began, with Vyasa throwing curveball verses at Ganesha whenever he needed a break.

The figure of Vyasa the composer is best understood as a literary device to represent a personified oral tradition (that perhaps included a single real Vyasa or family of Vyasas).

But the legend gets at something interesting about the role of a scribe in a dominantly oral culture. A second-class citizen like a minute-taker or official record-keeper, the scribe must nevertheless synthesize and interpret an ongoing cacophony in order to produce something coherent to write down. When the spoken word is cheap and the written word is expensive, the scribe must add value. The oral tradition may be the default, but the written one is the court of final appeal in case of conflict among two authoritative individuals.

There is a brilliant passage in Yes, Prime Minister, where the Cabinet Secretary Humphery Appleby  helps the Prime Minister, Jim Hacker, cook the minutes of a cabinet meeting after the fact, to escape from an informal oral commitment. Appleby’s exposition of the principle of accepting the minutes as the de facto official memory gets to the heart of the Vyasa-Ganesha legend:

Sir Humphrey: “It is characteristic of all committee discussions and decisions that every member has a vivid recollection of them and that every member’s recollection of them differs violently from every other member’s recollection. Consequently, we accept the convention that the official decisions are those and only those which have been officially recorded in the minutes by the Officials, from which it emerges with an elegant inevitability that any decision which has been officially reached will have been officially recorded in the minutes by the Officials and any decision which is not recorded in the minutes is not been officially reached even if one or more members believe they can recollect it, so in this particular case, if the decision had been officially reached it would have been officially recorded in the minutes by the Officials. And it isn’t so it wasn’t.”

The key point here is that the scribe must do more than merely transcribe. He must interpret and synthesize. I suspect the Vyasa-Ganesha legend was invented by the first scribe paid to write down the hitherto-oral Mahabharata, to legitimize his own interpretative authority in capturing something coherent from a many-voiced tradition, with each voice claiming the authority of a mythical Vyasa.


So if the modern blogosphere is neither the collectivist, negotiated recording of a Grand Narrative, arrived at via a conversation between scribes and sage-bards, nor the culture of purely individual expression that reigned between Gutenberg and Tim Berners-Lee, what is it?

For blogging to be performance art, the performer must live an interesting life and do interesting things. For a while I thought I qualified, but then I reflected and was forced to admit that my dull daily routine does not qualify as raw material for performance art.

How about this: instead of a half-coherent oral tradition or the relatively coordinated doings of the British Cabinet, the blogosphere is primarily an uncoordinated theater of large-scale individual gonzo blogging. As culture is increasingly enacted by this theater of decentered gonzo blogging instead of traditions that enjoy received authority, minute-taking scribe bloggers must increasingly interpret what they are seeing.

The first human scribe who wore the mask of Ganesha could reasonably assume that there was a coherent trunk narrative with discriminating judgments required only at the periphery.  He would only be responsible for smoothing out the rough edges of an evolving oral consensus. Equally Humphrey Appleby could hope for a coherent emergent intentionality in the deliberations of the cabinet.

But the scribe-blogger cannot assume that there is anything coherent to be discovered in the gonzo blogging theater. At best he can attempt to collect and compress and hope that it does not all cancel out.

There is another difference. When words are literally expensive, as words carved in stone are, anything written has de facto authority, underwritten by the wealth that paid for the scribe. Scribes were usually establishment figures associated with courts, temples or monasteries, deriving their interpretative authority from more fundamental kinds of authority based on violence or wealth.

With derived authority comes supervision. The compensation for lost derived authority is the withdrawal of supervision.  The scribe-blogger is an unsupervised and unauthorized chronicler in a world of contending gonzos. Any authority he or she achieves is a function of the density and coherence of the interpretative perspective it offers on the gonzo-blogging theater.


I wish I could teach dense blogging. I am not sure how I am gradually acquiring this skill, but I am convinced it is not a difficult one to pick up. It requires no particular talent beyond a generic talent for writing and thinking clearly. It is merely time-consuming and somewhat tedious.

Sometimes I strive for higher density consciously, and at other times, dense prose flows out naturally after a gonzo-blogger memeplex has simmered for a while in my head. I rarely let non-dense writing out the door. You need gonzo-blogging credibility to successfully do Top 10 list posts. I can manufacture branded ideas, but lack the raw material needed to sustain a personal brand.

Writing teachers with a doctrinaire belief in brevity urge students to focus. They encourage selection and elimination in the service of explicit intentions.  The result is highly legible writing. Every word serves a singular function. Every paragraph contains one idea. Every piece of prose follows one sequence of thoughts. There is a beginning, a middle and an end. Like a city laid out by a High-Modernist architect, the result is anemic. The text takes a single prototypical reader to a predictable conclusion. In theory. More often, it loses the reader immediately, since no real reader is anything like the prototypical one assumed by (say) the writer of a press release.

An insistence on focus turns writing into a vocational trade rather than a liberal art.

Both gonzo blogging and scribe blogging lead you away from the writing teacher.

Striving for density, attempting to compress more into the same number of words, inevitably leads you away from the legibility prized by writing teachers. Ambiguity, suggestion and allusion become paramount. Coded references become necessary, to avoid burdening all readers with selection and filtration problems. Like Humpty-Dumpty, you are sometimes forced to enslave words and chain them to meanings that they were not born with.


Dense writing creates illegible slums of meaning. To the vocational writer, it looks discursive, messy and randomly exploratory.

But what the vocational writer mistakes for a lack of clear intention is actually a multiplicity of intentions, both conscious and unconscious.

Francine Prose, in Reading Like a Writer, remarked that beginning novelists obsess about voice, the question of who is speaking. She goes on to remark that the more important question is who is listening?

The failure to ask who is listening is peculiar to pre-Internet book writers. You cannot possibly fail that way as a blogger.

The modern extensive-prose, word-wasteful book represents the apogee of a certain kind of individualism. An individualism that writes itself into existence through self-expression unmodulated by in-process feedback, something only entire cultures could afford to do in the age of stone-carved words. For this kind of writer, the reader was a distant abstraction, easily forgotten.

A muse was an optional aid to the process rather than a necessary piece of cognitive equipment. At most modern, pre-blogging book writers wrote for a single archetypal reader.

For the blogger, a multiplicity of readerly intentions is a given. At the very least, you must constantly balance the needs of the new reader against the needs of the long-time reader. Every frequent commenter or email/IM correspondent becomes an unavoidable muse. This post for instance, was triggered by a particularly demanding muse who accused me, over IM, of having gotten lazy over the last few posts and neglecting this blog in favor of my more commercial, less-dense writing.

She was right. Mea culpa. Having to pay the rent is not a valid excuse for failing to rise to the challenge of a tricky balancing act.

Density is the natural consequence of trying to say many things to many distinct people over long periods of time without repeating yourself too much or sparking flame wars. The long-time reader gets impatient with repetition and demands compaction of old ideas into a shorthand that can be built upon. The newcomer demands a courteous, non-cryptic welcome. Active commenters demand a certain kind of room for their own expansion, elaboration and meaning construction.

The exegesis of living texts is not the respectful affair that it is around dead ones. If you blog, there will be blood.


In the days of 64k memories, programmers wrote code with as much care as ancient scribes carved out verses on precious pieces of rock, one expensive chisel-pounding rep at a time.

In the remarkably short space of 50 years, programming has evolved from rock-carving parsimony to paper-wasting profligacy.

Still living machine-coding gray eminences bemoan the verbosity and empty abstractions of the young. My one experience of writing raw machine code (some stepper-motor code, keyed directly into a controller board,  for a mechatronics class) was enlightening, but immediately convinced me to run away as fast as I could.

But why shouldn’t you waste bits or paper when you can, in service of clarity and accessibility? Why layer meaning upon meaning until you get to near-impenetrable opacity?

I think it is because the process of compression is actually the process of validation and comprehension.  When you ask repeatedly, who is listening, every answer generates a new set of conflicts. The more you resolve those conflicts before hitting Publish, the denser the writing. If you judge the release density right, you will produce a very generative piece of text that catalyzes further exploration rather than ugly flame wars.

Sometimes, I judge correctly. Other times I release too early or too late. And of course, sometimes a quantity of gonzo-blogger theater compresses down to nothing and I have to throw away a draft.

And some days, I find myself staring at a set of dense thoughts that refuse to either cohere into a longer piece or dissolve into noise. So I packetize them into virtual palm-leaf index cards delimited by asterixes, and let them loose for other scribes to shuffle through and perhaps sinter into a denser mass in a better furnace.

It is something of a lazy technique, ultimately no better than list-blogging in the gonzo blogosphere. But if it was good enough for Wittgenstein, it’s good enough for me.


Get Ribbonfarm in your inbox

Get new post updates by email

New post updates are sent out once a week

About Venkatesh Rao

Venkat is the founder and editor-in-chief of ribbonfarm. Follow him on Twitter


  1. I really loved this piece. I am in particular interesting in your thoughts on writing (loved your Quora piece and would love to see you develop this stuff further. I think you’ve got a lot to add to the conversation. It’s thrilling stuff.

    • I considered cleaning up and posting that here, but am wary of slipping into instructional mode. If I could figure out a good model to somehow teach this stuff I would… I don’t want to call it either “writing” or “blogging” (there are people who do both much better than me). It’s more like a certain thinking style that requires writing, and is useful for certain purposes.

      • How about Rumi?

      • Pattern recognition and compression.

        I got very frustrated early on with some of the conventional wisdom you address here, particularly the notion that you need to identify a target audience and either provide something of obvious practical value or provide something delightfully entertaining.

        What you are doing, what I am trying to do, falls somewhere in between those extremes. It is neither wholly refined nor wholly unrefined. Much of it consists of the recognition and explication of *potentially* useful patterns, and the compression of those patterns over subsequent appraisals into increasingly useful form.

        The analogy to audio compression is actually pretty tight. Perform a piece of music once and it provides only a few moments of mental stimulation. Record it and it can spread to others. Digitize it and it can be further explored and remixed by others. Compress it and it can be accessed by more people on more devices in more disparate contexts.

        Writing, like recording, gets a set of ideas out of your head. Developing a language, like digitizing, modularizes the pattern so that others can remix it. Compression spreads ideas to more people in more contexts, allowing them to be accessed at multiple levels of depth, complexity, and fidelity depending on the application.

        This type of work is not nearly so bizarre and uncommon as the casual survey of blogging/marketing doctrine would suggest. It just took me a while to recognize the methods of similar writers and thinkers. For example, how frequently do you come across an article or a lecture that concludes with, “This is an important issue that we need to be having a conversation about!” That statement is a completely empty cop-out, but it gets used everyday in otherwise high quality work. Why?

        It is the accepted crutch that people use when they identify a pattern worth communicating but they don’t yet know what to do with it. My bet is that the people who resort to such conclusions are really pattern compression thinkers who feel compelled to conclude their remarks with some degree of certainty and authority to satisfy their more literal and linear readers.

        • Yes, I use the pattern recognition/compression explanation too, but it seems to be hard to understand for most people who have been ruined by bad writing teachers.

          • Maybe pitching wanna-be writer/bloggers is the wrong approach. As you note in your quora answer, there is a distinction between writers and thinkers. The audience drawn to ribbonfarm in general, as well as the arguments in this specific post, is going to be the thinker crowd rather than the writer. The more appropriate pitch might then be – writing as an amplifier of high level thinking. There is plenty of space to explore there that would incorporate technical writing issues without making such concerns the central focus (material explored in calculus of grit, learning curves, tempo, etc).

            There are plenty of naturally capable thinkers who are functionally crippled by their inability to compress (and therefore communicate/test) their ideas. One blogger, who I won’t name, I just removed from my RSS reader because his posts demanded far too much synchronization from the reader. The thinking was high quality but the writing too sparse to be worth the effort. He had to walk through his entire stream of consciousness to effectively deliver the punch line. In my own experience such lack of compression results in a dead end, crowding additional layers of insight out of consciousness.

          • Ho-Sheng Hsiao says


            This quora answer leads me to think that thinkers would want to learn expressiveness to be effective at thinking.

            A different narrative: thinkers tend to optimize towards correctness rather than minimal key insights.

          • Re: the writers/thinkers distinction… I think there’s no really way to understand this idea compression/density outside of attempts to communicate, whether to words, pictures, slides or any other medium. There is possibly something like “dense doing” but I can’t think how you can talk about the world of ideas without reference to a language of thought.

          • Ho-Sheng Hsiao says

            Venkat, I suppose if one frames physical confrontation as communication, then the language of doing is still subsumed in the language of thought.

        • Ho-Sheng Hsiao says

          And it isn’t just about compression. It’s about expressing the fractal nature of the big insights. Compression is an illusion.

          • Hosh, I’m not clear where you are seeing disagreement. Perhaps my phrasing, “writing as an *amplifier* of high level thinking” was imprecise. I didn’t mean amplify as a synonym for disseminate. I mean that the process of writing improves the quality of thinking…the process of externalizing thinking into formal language facilitates self-criticism.

            Also, you seem to be interpreting “compression” a bit differently than I am. I’m not suggesting that compression produces some ideal self-contained insight…just that compressing layers of minutia allows you to abstract and see more of the pattern.

          • Ho-Sheng Hsiao says


            Reading what I wrote now, looks like I need more practice in clear thought…

            On compression though, I am interpreting it differently. Like fractals, some of the important insights are self-similar. You see them in the details as well as the big themes. But that might just be me seeing nails everywhere when I’m holding a hammer.

          • That’s actually something of an interesting debate in the philosophy of mathematics right now… whether or not reality is “truly” dense in the amount of information/bits it contains or merely “fractally” dense (or to put it another way, whether reality can/cannot be reduced to a finite bit string much smaller than the apparent information content of the universe…).

            I incline to the weak view that at least the human corner of the universe is only fractally dense. Or to use a poetic oversimplification, there is possibly some very dense little book that could ideally be written by a super-human being, that could express the nature of reality…Erdos believed in something like that with respect to just math (he talked of theorems/proof as being in “God’s book” … expressed with maximal elegance and compactness).

        • That’s a pretty well-developed line of thought that I’ve encountered in multiple places, though I can’t think of any references off the top of my head. I go back and forth about it.

          Can you dance an idea for example? Can you choreograph or improvise a fight that expresses the “yin/yang” idea in purely action terms? Can you “read” an old real fight as an articulation of a particular dense idea in retrospect? Can you do any of this using pure action, skipping symbolic thought altogether?

          It’s an interesting theme. I haven’t yet made up my mind. It boils down to how deep the thought/action dichotomy goes. All dichotomies are false of course, but some are deeper/more fertile than others and you can do more with them before you start hitting paradoxes and contradictions.

          • Ho-Sheng Hsiao says


            For yin/yang in terms of action: I demonstrated a bit of the “hidden” “obvious” when you were here. You can still decode the body language and telegraphing in terms of verbal language, but that’s generally too slow. You tend to get good enough where you short-circuit that and simply know.

            It comes to what you mention in your book — people generally perceive with the eyes and manipulate with the hands as universal tactics. However, not everyone has visual dominance. There are sections of the population where someone first perceives through audio, tactile, and kinesthetic first before mapping it to visual. Likewise, manipulation has a strong default to tactile and kinesthetic, but … someone who primarily manipulates through sound might sing to a particular tone in harmony or disharmony, rhythm or syncopation.

            As far as reading a real fight as a particularly dense idea: yes. Most certainly. And your skill level determines the how much you see. That includes the different telegraphing and flows of emotion. Structure of the body in terms of locks, and balance. And over time as I get better, I’ve generally been able to perceive more. It also depends on what I have trained in. I can generally read standup and throws … but not much of what happens in the ground game for MMA/UFC fights that turn into wrestling. My friends who play on the ground can read ground games.

            One more example: the best way to learn pressure points is to have the teacher demonstrate it on you. Done enough times, you start naturally doing it to someone else. If I think a little harder about it, I can point to some of them in someone else because I get an echo of what it feels like in my body. If I try too hard, the skill disappears.

            From a different angle: Fingerspitzengefühl describes tactile/kinesthetic, not visual. Because that’s exactly how it feels like. There is usually so much narrative flowing by, you don’t have time to unpack them into words. Yet you are aware of the layers of complexity. The trick is not to dive into each, like with a depth-first or breadth-first search. That might be useful where correctness is vital (such as correct syntax for programming), but not useful where flow is more important (such as distilling your overall intent into syntax expression). You receive the whole thing at once, all layers and branches and abstractions.

            I pointed out that Quora answer for advanced mathematics because the process he describes more or less maps to these domains … though I might use tactile/kinesthetic rather than visual or audio representation. His intuition for quickly determining what he can answer mathematically is much the same when crossing hands with someone, or laying down stones on the Go board. Or a farmer working in the field. A mechanic touching my car. Writers distilling words out of ideas. Poets condensing nuance out of emotions.

          • Ho-Sheng Hsiao says

            Since math lends itself to visual symbols so well, it tends to attract people who have visual dominance. If all you have is visual cognition, then the notion of “dancing an idea” is just that: a notion. You can’t treat dance in any other way *but* visually. You work with the symbol of “dance” rather than experiencing the dance. For the professional dancers, though, they literally dance an idea. It’s a metaphor until it isn’t.

          • That’s fascinating. I am not tactile/kinesthetic at all, except in narrow domains like cooking and a little bit of drumming. I’ve been trying to change that a bit.

            I agree with you entirely that Fingerspitzengefühl is a tactile/kinesthetic concept. Actually, so is math in higher dimensions, where you lose visual intuition (it happens immediately in 4D) and you can only grope your way forward by using tactile/manipulative math. I don’t understand certain concepts like eigenvectors in visual terms at all (though they have visual connotations in 2/3D).

            For battlefield Fingerspitzengefühl, I’ve actually been attempting to build a math model, involving both maneuver and attrition warfare. The basic modeling approach is like fluid mechanics (building off the “flowing water” and path-of-least-resistance aspects of the battle). Though you can visualize the flows, you can make better sense of what is happening to the “flow” in tactile/kinesthetic ways. The “water” feels its way forward.

            This is speculation. I haven’t really gotten the model working yet.

          • Ho-Sheng Hsiao says

            Have you looked at the flow-field pathfinding that came out of the U. of Washington? It made it into Supreme Commander 2:

            If I remember correctly, the AI these days try to add threat assessments into the flow field. In the case for SC2, they layer different AI techniques.

          • Thanks for the link. Not familiar with that work. I am not interested in path-planning so much as maneuvering. Trying to get at the essence of ‘fight the enemy, not the terrain.’

            The last time I worked on this, I was trying to abstract maneuvering away from path planning in a 2-layer model with dumb obstacles in layer 1 and maneuvering adversaries in layer 2. This time, I am simplifying by mostly throwing layer 1 away.

          • Ho-Sheng Hsiao says

            It doesn’t necessarily have to be for path-finding around obstacles. You might be able to use it for threat and movement potential of the opposing force. Just like one of the plays you make on Go is to shape potential space and good shapes.

            On the other hand, modern Go AI uses Monte-Carlo method … maybe we’re just not there yet.

  2. In some ways this post reminded me of Leo Strauss and his ideas of esoteric writing; I don’t know if you read any of his stuff.

    Also the book “Scribal Culture and the Making of the Hebrew Bible” which talks about some of the stuff you brought up, but in the context of the Middle Eastern ancient cultures.

  3. Venkat,

    This was a thriller! You take-off from certain parts with intriguing alacrity and remind me of “Ghalib”! One of those pieces that I will come back to a few more times.


  4. This seems to have a certain desnity:

    “…scribe who wore the mask of Ganesha could reasonably assume that there was a coherent trunk narrative…”

    I’m glad you’ve remembered your readers here. It looks like I’m going to have to go back and read the whole post again. That’s usually a good sign, though I’m often left wondering whether it’s me who’s not bright enough to extract everything that I think’s in there, or you who fails to quite tie it all together.

    Many thanks again.

  5. With dense writing, typos become more dangerous. My first reaction to the small typo “nelogisms -> neologisms” was to think that it was some other of your neologism I had missed or forgotten.

  6. Alex Ragus says

    When words have to serve several meanings, they become ambiguous. In literature this leads to layers of meaning which are artistically valuable. But in argument, dialog, and science, this leads to misunderstanding between parties. It eventually produces cognitive dissonance in any person who thinks in that language.

    Another way to say this: with any language that allows artistic levels of layered meaning, you need to be especially watchful for thinking (and subsequent writing) that is un-intentionally ambiguous.

    Or to put it yet another way: be careful to distinguish between density and precision. If you’re a poet, then dense and unprecise writing is great. But if you’re a thinker, then your writing must be precise first, and dense second. Done unskillfully, this turns into legalese.

    related: paul graham’s criticism of modern academic philosophy that it is simply math, damaged by the imprecision of language:

    • I would argue that good poetry tends to be extremely precise. Precision in language follows precision in intention. Ambiguity can arise from both failed precision and from someone not being in the intended audience. A writer may not care whether a specific kind of reader understands or misunderstands something. There may even be a deliberate intention to confuse or deceive a specific audience.

      Though if the matter is consequential and the writer influential, there may be problems that arise from people who should not have been ignored in the first place.

      The relation between precision and density is a very strong one. Writing precisely is the only way to achieve density. Precisely chosen words fit together more densely. But again, if the reader is not in the intended audience, this precision and density may not be apparent. In fact, the writer may make an effort to hide the precision and density. The “Dancing Men” Sherlock Holmes story contains an interesting if rather obvious example of how that works (and fails).

      I think you are using “ambiguity” to refer to both multiple, layered intentions and sloppiness or carelessness. That’s bad in both poetry and science (and legalese for that matter). Sloppiness differs from ambiguity in not being dense, because you make up for lack of precision by adding a lot of fluffy qualifiers. Your point about legalese is perfect evidence for this. If you fail to accurately define the intentions and risk-management decisions behind a legal document, you cannot craft precise language. So you end up compensating by throwing in extra catch-all verbiage that is overly conservative. This stuff is typically not dense. Merely voluminous. Like swaddling or padding as an insurance policy.

      I think we may differ also on a fundamental assumption. I think both ambiguity and sloppiness are unavoidable in the process of science. As is cognitive dissonance and disagreements arising out of misunderstanding. You cannot use process to engineer conflict that arises from fundamental external uncertainties.

      I think this is also why I have never liked that particular Paul Graham essay. Some of his specific points about academic culture are spot-on, but he basically labors to miss the point of post-Wittgenstein analytical philosophy.

      • Alex Ragus says

        If dense communication is enabled by shared verbal associations and cognitive styles, then how can we continue to read Gilgamesh and get its intended, precise meaning? I don’t believe we can. But we continue to read it because it is sloppy and dense–we can project new layers of meaning onto the original text.

        Great literature is precise and dense for its intended audience, but it has staying power because it is so dense that there are a bunch of (sloppy) layers the author never consciously intended.

        Unlike science and Powertalk, literature and poetry are supposed to communicate to a fairly broad audience. If dense writing achieves that, it does so at least partially through (sloppy) ambiguity.

        My conjecture: dense, precise, wide-audience. Pick any two.

  7. atimoshenko says

    I’d agree with Alex Ragus above.

    I think there are at least two different types of density, and at least one of “apparent density”. With regards to the former, density can be achieved through compactness or through ambiguity. Compactness means not using two words where one word would do. It takes quite a bit of effort on behalf of the author, but saves effort for a focused reader. It’s a case of Einstein’s “everything should be made as simple as possible, but not any simpler”. Articles in The Economist are often a good example of this style. Ambiguity has one construct sentences and passages that make sense from multiple perspectives (this, I guess, is what you mean by “layering”). If well done, it requires additional effort both for the author and for the reader, but with the benefit of conveying the relationships and patterns between the multiple interpretations. It is common in mysticism (from fables to koans) and personally I am not a big fan – not only do I see no reason against conveying the same patterns and relationships in another form, but it also tends to be objectively impossible to distinguish between genuinely profound ambiguities and lunatic ravings (e.g. the ‘predictions’ of Nostradamus). Ambiguous faux-density, in other words, can be used by shameless authors to attempt to hide the fact that they do not know what they are talking about.

    Finally, there is an apparent density through re-framing or re-factoring. This is when an author comes to an insightful conclusion that cannot be elegantly expressed using existing concepts. So the author has to define new concepts through which a conclusion makes more sense. However, I would argue this is more akin to switching coordinate systems (or, for that matter, replacing a complex expression with a single variable for ease of manipulation) than a true addition of density – it feels more dense to people who have not previously experienced the new coordinate system or variable, but only because the new bases are not yet intuitive. Such re-framing can be incredibly useful (approaching motion through the frame of relativity) or it can be used to obfuscate (postmodernism generator) – inventing a set of bases that require ten words where one word would do in the default set.

    • I think you are missing something important in labeling refactoring/reframing as leading to “apparent density.”

      Refactoring is a new perspective, yes, but it isn’t just a different perspective. It’s not like moving along a 360 degree circle of positions around an object. Since we are talking about ideas here rather than literal ground realities, refactoring constructs what it sees. It is not like sensory observation.

      Just as code can become more compact for the same purpose through refactoring (i.e. “denser”), refactoring your perception of an external reality that must be perceived in reified terms leads to greater density when done well. The “view constructor” software inside your head for constructing that reified mental model becomes more compact.

      So there is an absolute sense in which refactoring leads to “better” views. You get more compact mental models of the same realities, which you should prefer by Ockham’s razor.

      But I agree with your critique of postmodernism. That is often dense simply because you are using an unfamiliar coordinate system. That’s why pomo accounts of things often seem to me to be the start of analysis that is then abruptly abandoned. The point of a new coordinate system to me is to make some analysis problem more tractable. Pomoists are content to merely do the transform and let it be.

      • To put this in psychological terms, “chunking” has been well demonstrated to correlate with high level performance in a wide variety of domains. Our working memory is limited and moving to alternative levels of abstraction allows us to explore cognitive leaps that are not discernible when working memory is tracking all the component parts.

        I was reading recently about study examining the memory capacity of elite chess players. As you might expect, when shown advanced formations that might occur in an elite level match, elite level players were able to track the position of every piece on the board in working memory and re-draw those boards from memory. However, when shown formations that would never occur in an elite level match they were no better than a control group at re-drawing the formations they were asked to remember.

        Their apparently flawless memories were all a function of compression. They are able to consider moves that average players can’t because in their minds they aren’t moving pieces against one another, they are moving formations against one another.

  8. > When fountain pens and cheap factory-made paper made their appearance, writers were able to waste paper, and as a consequence, written words. In the history of thought, the invention of the ability to waste words was probably as important as the invention of the ability — famously noted by Alan Kay — to waste bits in the history of programming. With cheap paper was born that iconic image of the twentieth century writer — a writer sitting alone in a room, crumpling up a piece of writing in frustration, and tossing it into an overflowing waste-paper basket.

    At least the 20th century writer bothered revising and discarding; one wonders about the 19th century writers like Dickens or Trollope ever did.

    > Striving for density, attempting to compress more into the same number of words, inevitably leads you away from the legibility prized by writing teachers. Ambiguity, suggestion and allusion become paramount. Coded references become necessary, to avoid burdening all readers with selection and filtration problems. Like Humpty-Dumpty, you are sometimes forced to enslave words and chain them to meanings that they were not born with.

    Have you read Melzer’s paper “On the Pedagogical Motive for Esoteric Writing” ?

  9. Thanks Andy and Gwern for the pointers to the Straussian literature. Am not familiar with it.

  10. If you’re not familiar with it, you might like CK Ogden’s “Basic English” concept:

    I’ve often thought that there’s room for a startup to build an app which teaches you that vocab plus basic grammar in your language of choice, using smart flash cards etc in an interactive/game format on mobile devices. (To be honest, I’ve often thought I should make my own such cards for Spanish after moving to the US, and that leads to the wish that someone would do the work for me.)

  11. Your thought experiment regarding limited words reminded me of something I read the other day in Stanley Fish’s “How to Write A Sentence: And How to Read One.” Fish was explaining Gertrude Stein’s refusal to use punctuation, to let the words alone suggest the structure and relationship of the ideas in the sentence. As I recall, Stein seemed to use fairly simple words, repeated with differing rhythms and antecedents, to express her ideas. You can acheive density by packing meaning into repetition and syntax, but man alive is that difficult to use as a blogger. The TL;DR-fairy is one of my muses, but she is cruel.

  12. I enjoyed this piece immensely, Venkat. Due to its density, I will be re-reading it in the near future.

  13. Do you think that tweeting regularly to express ideas improves ones writing density? At least I find that the 140 character constraint forces me to think more carefully about which words I choose and how one word can express as much as several individual (longer) words.

  14. Consider this thought experiment: what if you were only allowed 2000 words with which to understand the world?

    Interesting that you too left out the who is listening in your thought experiment. This choice will entirely determine the choice of your words. Is it a child, an alien civilization, the last man to survive your language, attitude and culture, is it a god, is it the average educated reader of your blog, the reader who is supposed to discover a great freelancing intellectual talent, is it possibly yourself, your double living in another time and age or branch of reality, is it you here and yet who desires the nomadic trails and some lightweight memorable thought baggage? Is it a matter of living well or of survival, is it an existential gift or an attempt of seduction?

    It seems to me that not only are those questions undecidable but I also wouldn’t want to decide them. Addressing a piece of the 2000 word text to each of them means that all the audiences are present in the text and this way I already preserved what was of the highest importance. This isn’t a complete solution for sure but the problem already got a little smaller.

  15. Ho-Sheng Hsiao says

    Since I’m still binging on Go: the better players do more with each stone. It is a way to accelerate tempo.

    Like all craft, the moves look simple. Simple app. Simple paintings. Simple tactics. Simple is hard.

    Strunk & White emphasized conciseness. This principle enfolds meanings for density. However, our generation lost this in translation. Now, we imitate Hemmingway like children dressing up in their parent’s clothings.

    Perhaps density is the secret of the Renaissance Man.

    • Yes, I actually started thinking about this when we were exchanging emails about how that same effect shows up even in tic-tac-toe.

      Hemmingway… Densely observed/imagined if not densely written I’d say.

  16. Happy to buy coffee for something as well written and original as this. It’s heartening to see Vyasa-Ganesh story revealed in this context. In Telugu they have a word for the rule book that kept poetry dense called ‘chandassu’ until modern poets like Sri-Sri threw away the shackles. The rule book may have come about because it was expensive to write but also to keep it inaccessible for the masses. It needed wealthy kings to patronize erudite scholars who would interpret past poetry and write new in it’s style.

  17. Do you know any software that can keep track of number of unique words used in a text?

    • That’s a trivial bit of software. For example, if you don’t mind punctuation and formatting being counted, here’s a shell script to count unique words on my website (a collection of files with the .page suffix):

      $ cat *.page | tr ‘[:space:]’ ‘\n’ | sort -u | wc
      102927 102934 1434454

  18. I came across Pico Iyer’s post arguing the benefits of long, winding sentences. And, while you may not have been talking explicitly about length of individual sentences, I thought it was an interesting article relevant to this one.,0,2137466.story (Note Brian Keaney’s comment as well!)

  19. no much to say