I See Stylists Everywhere

Cole: I see dead people.
Malcolm: In your dreams?
[Cole shakes his head no]
Malcolm: While you’re awake?
[Cole nods]
Malcolm: Dead people like, in graves? In coffins?
Cole: Walking around like regular people. They don’t see each other. They only see what they want to see. They don’t know they’re dead.
Malcolm: How often do you see them?
Cole: All the time. They’re everywhere.
The Sixth Sense.

In the movie The Sixth Sense, a boy Cole can see many ghosts, ghosts who are in denial about the fact that they are ghosts. I similarly see that most of our public intellectuals are in denial about the fact that they are mainly loved for their style, not their content. Let me explain.

A big reason that readers read (or listeners listen) is that reading transfers something from the mind of the writer to the mind of the reader, something that readers can then transfer to others if they themselves then write (or talk). I call that something the “content” of the writing.

All the other features of writing that can’t be as easily transferred, I call “style”. Readers may enjoy and admire the styles of things they read, but they can’t as easily add those styles to their own writing. Style is of course actually valuable, at least to some degree. Style can make content easier to enjoy, accept, and understand. The only question is how much effort to put into these different aspects of writing.

This difference between content and style is a continuum. For example, the main claim and argument of an essay is more content-like than are diversions and comments, which are less likely to be remembered or understood by readers. Subtext is even more style-like, as fewer readers can notice or repeat it. Subtle choice of vocabulary and sentence structure can be even more style-like.

(This way to define the distinction is unusual, but seems to me more concrete than most others you will find online.)

Content being more transferable than style has some big implications. Sometimes it is easy for writers who compete with you to read your writing and then put your content into their writing. In a way that few will complain about, because in these worlds there are not property rights in such content. In this case, you have less reason to invest in content. And this is in fact the situation for some kinds of intellectuals.

In contrast, this lack of property rights is less of a problem if you are a professional, paid by a client to write privately to them, on a topic of interest only to them, writings which they may then paraphrase to others.

In professions such as journalism, law, therapy, math, engineering, and business, people tend to adopt relatively standard and structured writing formats. For example, each essay is usually expected to clearly telegraph one main claim and argument. Many such professions also have their own specialized vocabularies and methods of inquiry, and their norms tend to prohibit many “fallacious” forms of argument. We tend to expect the best judges of professional writings are other professionals.

Such formats and habits tend to minimize opportunities for distinct personal writer styles and to maximize the ease with which readers can find and understand the distinctive relevant content of each case. That is, such professionals write as if they were interchangeable.

Such professions also often structure their relations to support property rights in writing content. Writers are either paid by specific clients to write on topics mainly of interest only to those clients, or professional norms of publication citation and priority help writers get credit for novel unique content.

Note that a memo to your boss is likely to be read only by that boss, who will paraphrase your content if it is passed up the management hierarchy. In which case your content will travel further than your style.

The writing style of many kinds of “public” intellectuals looks quite different from this. Such writers have much more distinctive personal styles, and allow themselves more digressions from a single main claim and argument. It is often unclear what exactly they are claiming, or which of their claims is central. And they often include examples, stories, and other discussions from which it seems hard to extract logical arguments. Such intellectuals also allow themselves a wider ranges of topics, methods of inquiry, and kinds of arguments. And for their writings we anchor more on the reactions of ordinary readers, rather than of similar professionals.

A straightforward interpretation of these differences is that these intellectuals emphasize style over content, relative to professional writing, because in their world it is much easier to own style, relative to content. Such intellectuals, however, usually resist such an interpretation of their own writing. If asked to explain why their writing looks so different, they point to the generality or value-orientation of their topics, or to their need to entice a wider range of more distractable readers.

Yes, they admit, they put a lot of effort into style. But this is only to make it easier for readers to enjoy, accept, and understand their content. They insist that the main reason that the world does, and should, attend to them, is their unique and valuable content. If one asks why then they don’t coordinate to create stronger property rights in content, they shrug, or suggest that the methods other professionals use won’t work for them.

I find the following observation especially telling. Many writers see themselves as specializing in content, and thus feel eager to team with writers who specialize more in style, to together make writings great in both content and style. But very few writers see themselves as style specialists, eager to team with content specialists. Almost all writers instead see themselves as having access to good enough content, thank you very much.

And thus we reach a Sixth-Sense-like situation. Just as in that movie, where a boy everywhere saw dead people, in denial about their being dead, I everywhere see stylists, in denial about their being stylists.

In that movie, sometimes one could convince a ghost that they were in fact a ghost. Could we similarly measure the degree of style emphasis of any given writer, and then use that to prove their style status to them?

One strategy would be to just pay other random writers to paraphrase essays, pay typical readers to read and rate those paraphrased versions, and then compare the popularity rank of original essays to that of paraphrased versions. (Using two different rank scales.) Writers whose actual essays rank higher than their paraphrased versions must be getting more of their popularity from style, which doesn’t get transferred in the paraphrasing process.

Of course it would be expensive to pay for these writings and readings. Maybe GPT will get good enough to do these things cheaper? Or maybe we could train such a system on a much smaller dataset of human reading and writing of paraphrased essays?

If the most popular (i.e., “top”) public intellectuals specialize the most in style, but still need good-enough content for their essays, that gives them a reason to look for content ideas by reading non-top intellectuals, who specialize more in content. Of course they might also need to read other top intellectuals, to track fashion trends in styles and topics.

This is my impression from attending social events with such top intellectuals. They don’t actually have more interesting ideas to discuss, compared to people a few ranks below them, and they spend most of their effort trying to track fashion trends and make connections. I noticed a similar thing when attending a group of investors interacting with a group of projects seeking funding; investors mostly tried to make connections and track trends in what the other investors were choosing, instead of evaluating projects themselves from first principles.

Some writers are especially interesting to top intellectuals, but are not top intellectuals themselves. These writers probably specialize more in good content. They will not get big book contracts, be invited as keynote speakers, or be assigned to top institutional positions. But they may still end up with outsized influence on what others come to think. Even if they don’t get much credit for that.

Note how content vs. style maps well onto experts vs. elites. Experts focus on content, while elites focus on style. Bosses are relatively elite, and they are often accused of presenting the content of their relative expert subordinates as if they had come up with it.

Added Jan5: Our norms generally tend to favor the production of things that would otherwise be underproduced. This predicts our norms favoring content over style in contexts with weak property rights on content. Which helps explain why we say that we favor it.

Many people think they find counter-examples to my definition of “content” vs. “style”, but they do not offer a concrete definition that accounts for all these examples. Seems to be just something folks “know it when they see it”, yet often don’t see the same thing.