On Cultural Progress

Photo by Jeremy Yap on Unsplash

Ryan Murphy has written a provocative and meme-savvy piece for Works in Progress. In the article, Murphy argues that progress applies not just in scientific but also artistic endeavours. However, artistic progress is held back by a set of cultural assumptions, at which Murphy takes aim.

The article has prompted Reddit dissections, and verbal reflections by two of my favorite podcasters. I’ve written this essay because I agree with Murphy’s conclusion, disagree with his argument why, and propose a different argument in support of the fact that cultural progress is not only possible, but happening right now.

Steelmanning Murphy

Murphy opens with the attention grabbing rhetorical questions

If someone refused to use Google Maps and preferred to rely on a vellum chart from the 1400s, we would think they were very strange. If a bank communicated with its branches via telegraph, who would deposit money with them? If an army used guns from 1650 in a modern battle, they would meet certain death. (Murphy, 2021)

And yet, Murphy asserts, we do exactly this in the realm of arts and culture.

“It is commonplace to allege that not only is the art and media we produce no better than that of the near or far past, but that it is actually worse. These arguments are false.” (Murphy, 2021)

How do we know that cultural products are better now than in the past?

“Our revealed preferences mean something… if we can replicate a near approximation of a historical masterwork at a cost no higher than it was in the past, and we’re doing something else, that something else is probably better than the supposed historical masterwork.”(Murphy, 2021, emphasis mine)

The appeal to revealed preference may be stronger than Murphy lets on. For example, Murphy recommends that people watch Ford v Ferrari instead of Citizen Kane because Ford v Ferrari is more popular. I would add to this that for Ford v. Ferrari to be so popular, it must appeal to a wide variety of movie goers. Thrill seekers must enjoy it. People going on a date must enjoy it. Parents with their kids must enjoy it. The wider the revealed preference for a movie, the greater the range of preferences that it must have, successfully, appealed to. This tells us something about its quality.

Murphy deploys revealed preference to take exception with critics like Douthat, Hickel, Hirsch and Biskind. These critics argue the the pinnacle of cultural accomplishment was sometime well before the current day. Douthat for one has argued that the 1970s were “our civilization’s last great burst of creative energy”, to which I, and I’m sure Murphy would too, disagree.

To Murphy, lauding the past and poo-pooing the present is progress denialism. People today much prefer watching Ford v. Ferrari to Citizen Kane. People much prefer scrolling Instagram than going to the Louvre. People’s revealed preference shows us that the golden age is now, not 1970, not 1570, not any other period.

Murphy argues that this misplaced sense of a past golden age rests on nostalgia rather than cultural merit. He cites research into the nostalgia-age nexus, where our perception of the quality of cultural products coincides with their salience during our formative years.

To Murphy, it’s somewhat dangerous to laud antiquated works because of nostalgia and deference. If we cannot conceive of removing these revered cultural artifacts from our top ten lists, then we may be, at least tacitly, opposed to the idea of cultural progress.

“If one can’t imagine the removal of items from the Western cultural canon taking place, then one’s paradigm may be inconsistent with even the possibility of cultural progress.” (Murphy, 2021, original emphasis)

Murphy argues that in order to for culture to progress, we must be able to conceive of the possibility that masterpieces like Beatles music, Shakespearean drama, and Orson Welles’ filmography can and will be surpassed. To consider de-canonising any of the classics to be inconceivable is to be against progress — quoting Joel Mokyr

Progress is “an implied disrespect of previous generations… The same age that fostered a belief in progress shed its excessive respect for earlier thinkers, exuding a confidence that ‘we can do better.’ (Murphy, 2021)

Murphy adds a final, stronger claim to his thesis. Works such as Beatles music, Shakespearean drama and Citizen Kane have, in many way, already been surpassed. Contemporary music, stories and films employ more clearly written prose, more complex plots, and more sophisticated literary devices, as well as more photo-realistic images, higher-quality sound and more advanced special effects. The fact that today, as we speak, these classical works continue to be lauded despite their technological inferiority suggests that we as a culture are already incapable of conceiving of the idea of cultural progress.

As Murphy asserts, quote:

“Those studying Shakespeare today are metaphorically using the same scholarly methodology as medieval monks copying ancient texts line-by-line while thinking it’s teaching them science.” (Murphy, 2021)

An outline

I find a lot that’s interesting in this thesis. However, the argument is conceptually muddled and much is asserted that ought to have been demonstrated with reference to actual behaviour. My argument in short is:

1. Murphy doesn’t explain the purpose of art.
2. Revealed preference indicates popularity, not quality or progress.
3. Revealed preference is incoherent with what we think of as artistic quality.
4. Audiences’ revealed preferences lean toward the classics.
5. Praise for the classics doesn’t prevent cultural progress.
6. Artistic progress continues to be made to this day.
7. We underrate contemporary art because of biases.
8. Progress is positive sum.

1. What is the purpose of art?

Works in progress is a great publication. I adored Sue Marquez’ article in May 2021 calling for more work to eliminate lead pipes in the United States. Lead pipes are a huge problem, and if we invested more in locating and replacing lead pipes, then we would solve a serious problem that causes a lot of human suffering.

I also love Steven Pinker’s The Better Angels of our Nature. Violence has been a huge problem throughout human history, but if we doubled town on the trends of better governance, free markets, literacy, feminisation and rationality, then we could reduce a lot of violence that causes a lot of suffering.

In both cases we can measure progress by pointing to an outcome — developmental disorders, violent deaths. But what’s the better outcome that Murphy’s trying to progress? What is the outcome of art?

Murphy’s argument frames consumption as the outcome of art. Good art is what is our revealed preference, and what is popular is our revealed preference. I think this is fine as far as it’s necessary to debate cultural progress without getting bogged down in “what is art, anyway?”. However, it opens up the argument to other definitions of artistic progress.

Notwithstanding the ongoing debate within aesthetics, I propose an alternative measure of artistic progress that:

Cultural progress is the expanding capacity of our cultural repertoire to evoke our thoughts and feelings. Cultural stagnation or regression is the stagnant or shrinking capacity to evoke thoughts and feelings.

This definition puts thoughts and feelings at the center of artistic purpose, rather than popularity and revealed preference. Consuming a piece of art is necessary for art to evoke thoughts and feelings, but it is not sufficient. People often consume art that fails to evoke their thoughts or their feelings!

My definition doesn’t privilege either high culture or mass culture. It’s legitimate for a painting to evoke ‘refined’ thoughts and feelings. It’s also legitimate for a blockbuster film to evoke feelings of enjoyment, excitement and “gee-whizz”. Both are legitimate. What matters is how well the art evokes those thoughts and feelings.

Philosophers may argue whether some thoughts and feelings are better than others. Philosophers may argue whether the author’s intended thoughts and feelings are more important than those unintended. I’m sidestepping those debates for now. All thoughts and feelings are equal, what matters is the work’s capacity to evoke them.

This is a more satisfying metric of cultural progress, which we’ll return to later. I will argue that it’s more coherent than Murphy’s revealed preference.

2. Revealed preference indicates popularity, not quality or progress.

Murphy’s invocation of revealed preference as a measure of quality is an interesting one. If the classics are so good, and accessing them is so cheap, why don’t people access them all the time?

To address this, I ask you, the reader, what is your favourite movie? For me, the it’s 1996 movie Gattaca. I probably watch Gattaca once a year at most, but to be honest I think the last time was in 2014. You probably watch your favourite movie about as often as I watch mine.

You and I both say that we like our favourite film, and yet 99.9% of our spare time is spent not watching this movie. This raises the question: If we can recreate the experience of watching our favourite movie at a cost no higher than it was in the past, and we’re doing something else, we probably like that something else better than our supposedly favourite film.

If revealed preference is our measure of artistic success, then it’s paradoxical that we spend the vast majority of our spare time not watching our favourite film. It’s paradoxical that masterworks aren’t replayed night after night. But this isn’t a a paradox if evoking thoughts and feelings is the purpose of art.

• Many cultural works are excellent, but you only need to experience them once to appreciate them. Some works lose their edge through repeated viewings. It’s OK to want to move on and experience new things.
• Many works elicit difficult emotions like fear and sorrow. Just because you don’t enjoy experiencing those emotions doesn’t make the work bad. Many great works evoke very difficult emotions, but many people just want to come home and experience comfort and security.
• Many works can only evoke our thoughts and feelings after close engagement. Engaging with a work this way can be very rewarding, but it’s OK if you just don’t have the time or willpower to do so. It’s OK to come home and unwind with something easy.

Many great works are difficult, so it’s understandable that many people’s revealed preference is to not engage with them. That’s fine! But in Murphy’s article, engagement can be scorned.

“Those studying Shakespeare today are metaphorically using the same scholarly methodology as medieval monks copying ancient texts line-by-line while thinking it’s teaching them science.” (Murphy, 2021)

This trikes me as the artistic version of Chesterton’s Fence. I think you should only remove classics from the canon once you fully understand why they were put there. Murphy’s article does not even attempt to understand why Shakespeare is canonical.

It’s true that many people don’t read Shakespeare. Shakespeare is difficult for modern readers to understand. But difficulty is a weak reason to de-canonise a classic. I think Tyler Cowen is right about Shakespeare’s works

“Shakespeare is very likely the deepest thinker the human race has produced, so these are worth careful study!” (Cowen, 2017)

Accessible works are not intrinsically better, nor are inaccessible works worse. Murphy quotes from Adam Smith and Malcolm Gladwell to demonstrate how Gladwell’s 2005 Blink is more readable than Smith’s 1776 Wealth of Nations. But the reason that Gladwell is easier for us to read than Smith is not because the prose of one is objectively better, as Murphy claims, but the fact that English was a different language in 1776. The fact that a 21st century English speaker finds something easier to read does not make that work superior.

Between 2011–2014, Elena Ferrante published the Neapolitan Novels. They have sold over 10 million copies, and have been heralded as classics. But, they’re written in Italian. Murphy could have placed an excerpt from Ferrante’s novels next to Blink by Malcolm Gladwell and asked which was easier to read. To an English speaking audience, it’s clearly Gladwell. But this seems unfair, Ferrante’s prose is accessible to Italian speakers. If we took the time to learn Italian, then we could enjoy these novels too. It’s no different for Smith. The Wealth of Nations is written in 16th century Scots-English, a language different from our own. The fact that the reader must learn the language does not take away from the value of the Wealth of Nations or from Shakespeare. They are brilliant thinkers who reward close study.

Just because most people don’t engage with a difficult work doesn’t make the work bad. People may prioritise other things in their life than the time and effort of studying a difficult work. These are legitimate reasons to not engage with the classics. But if revealed preference is synonymous with artistic quality, they become a misplaced condemnation of the work itself.

3. Revealed preference is incoherent with what we think of as artistic quality.

Revealed preference doesn’t necessarily work even on its own terms. Reading Murphy’s piece I asked myself: over what time frame is revealed preference meant to indicate quality?

Murphy suggests that viewers will get more out of watching The Aviator than many supposed masterworks. This is an odd recommendation. The Aviator made $103 million domestically in 2004, reaching around 16.9 million tickets sold. By contrast, in 2004 there were a mere 4.4 million visitors to the Smithsonian in Washington D.C. The audience’ revealed preference suggests that The Aviator was better than all the works in the Smithsonian.

But if we look a few years later to 2007 it’s less clear. Theaters no longer screen The Aviator. Streaming services and DVD sales are hard to come by, but it’s hard to imagine the number of people who watch The Aviator each year would surpass half a million. In that same year, 7.1 million visitors attended the Smithsonian. The audience’ revealed preference suggests that the works of the Smithsonian Museum are better than The Aviator.

How can this be? The Aviator was a popular and therefore artistically superior film in 2004, but then a less popular, artistically worse film in 2007, despite the fact that The Aviator has not changed. The Smithsonian was artistically inferior in 2004, but superior in 2007 despite having not changed. This rapid change in quality, despite neither work changing at all, is incoherent.

—

The effect is even more extreme in laboratory settings. Researchers at Columbia University created an online music marketplace for over 14,000 teenage participants. The marketplace consisted only of obscure indie songs none of the participants had heard before. Participants were asked to listen to songs, rate them, and could reveal their preference for certain songs by downloading them to keep.

The popularity of each song within a given cohort was random. Downloads created a cascade effect where high downloads induced participants to try the song, leading to more downloads. For one group of participants, one song reached number one out of 48 songs, while in another group the same song ranked 40th. As the paper noted, “The best songs rarely did poorly, and the worst rarely did well, but any other result was possible.”

How can this be? Each cohort’s revealed preference for each song varied wildly. This would imply that each song’s quality varied wildly, despite the song remaining identical across cohorts. This wild variation in revealed preference despite the song not changing at all seems incoherent.

—

On 10 December 2020, CD Projekt Red released Cyberpunk 2077, a highly anticipated video game. Within twelve hours of release it had over one million players on Steam, but initial reviews were scathing. There game was full of bugs, clunky combat, and was unplayable on last-gen consoles (CD Projekt Red offered these players a refund).

Cyberpunk 2077 has gotten better since launch. CD Projekt Red have released 11 patches, with a 12th in development. And yet, in June 2021 the number of concurrent Steam users peaked at 15,703 players, it’s lowest month on record. These patches are meant to have made Cyberpunk 2077 better, but each new patch has seen the player base shrink. Players’ revealed preference is that buggy, unplayable Cyberpunk 2077 was better than patched and working Cyberpunk 2077. This seems incoherent with what we think of as artistic quality.

—

These counter-examples underscore that revealed preference is an incoherent measure of a work’s quality. To embrace this approach is to reach strange conclusions where that quality of a work varies considerably even when the work remains unchanged, that quality is random, and that unambiguous improvements to a work make it worse.

4. Audiences’ revealed preferences is for classics.

Even if revealed preference were a coherent measure of artistic quality, it would not necessarily support Murphy’s conclusion that new art is better art.

Last year my housemate bought a Disney + subscription. First we watched The Little Mermaid (1989), Sleeping Beauty (1959), then Snow White & the Seven Dwarfs (1939). At various points I found them watching Pinocchio (1940) and Cinderella (1965). Despite having a Netflix account, and its award winning original films such as Roma (2018) and The Irishman (2019), they preferred watching old movies on Disney +.

Which is not to say that Disney + has no new content, but the new content embodies the reverence for old classics that Murphy lambasts. Beauty & The Beast (2017) at its most ambitious added a blink-and-you’ll-miss-it nod to same-sex attraction, and fixed a ‘plot hole’ that no one knew about the Beasts’ castle that was only a day’s ride from the village. The Lion King (2019) updated the graphics of The Lion King (1994) and added one line to show that Nala could be a boss too, but otherwise no change.

The latest Star Wars films fare little better. The Force Awakens remade A New Hope. Rogue One used CGI to recreate Carrie Fisher and Peter Cushing exactly as they appeared in 1977. Disney tried to move the franchise forward in The Last Jedi, but fans revolted against these changes. Fans would rather that Star Wars repeat 1977 again and not progress at all.

Murphy appeals to revealed preference to argue that we are progressing as a culture, but our revealed preference is very often for classics and reverent recreations of the classics.

5. Praise for the classics doesn’t prevent cultural progress.

Murphy raises another interesting point in the essay that our view of the classics conflates quality with what was ground breaking technology at the time. At one point Murphy argues that Citizen Kane is overrated because people often praise Citizen Kane for its ground breaking technology.

In fairness, many people do say this and it’s very annoying. Ask Reddit and responders cite Kane’s innovation. Crash Course says Kane was groundbreaking. Roger Ebert too

“Citizen Kane” is more than a great movie; it is a gathering of all the lessons of the emerging era of sound, just as “Birth of a Nation” assembled everything learned at the summit of the silent era.” (Ebert, 1998)

Murphy leans on the over-praise of Kane to show how critics conflate past accomplishment with contemporary enjoyment. But this doesn’t prove anything about cultural progress more broadly. Groundbreaking movies can and are declassified as greats, Kane is the outlier in how long it has survived.

In 1998 the American Film Institute (AFI) released 100 Years, 100 Movies, a poll of the 100 best American films according to 1,500 of its members. In 2007 they released an updated list in which twenty three movies were removed, including:

• The Jazz Singer was the first feature film with synchronised sound. For this accomplishment, The Jazz Singer was added to the 1998 list, then removed in 2007.
• Birth of a Nation was once praised as “the summit of the silent era”. It was added to the 1998 list, then removed in 2007.
• Close Encounters of the Third Kind pioneered motion control photography, was added in 1998, and removed in 2007.
• Stagecoach pioneered cinematic stunt work in 1939, was listed in 1998, and removed in 2007. That said, The Lord of the Rings, which was added in 2007, does have pretty good stunt work.

Nor is willingness to de-list cultural classics confined to film. In 2005, Universal Publishing released 1001 Albums You Must Hear Before You Die. There have been three more editions, the latest published in 2018. Seventy seven albums have been de-listed over the thirteen years, all of them in order to make room for new albums that came out during that period.

Had Murphy’s article explored further, it would have been able to demonstrate that technological innovation does not guarantee even temporary status as a great film.

• Tron was praised on release for its groundbreaking CGI but has never broken into consideration as a great film qua film.
• The first film to use all digital colour grading was O Brother Where Art Thou, which by consensus is in the bottom half of all Coen Brothers’ films.
• Vertigo, The Godfather, Casablanca and City Lights are argued to be great films. They made, at best, minor advances in the technology of movies.
• City Lights is considered one of Charlie Chaplin’s great films, yet it was a silent film that poked fun at ‘talkie’ pictures after the technology had been mainstream for over three years.

Technological innovation is neither necessary nor sufficient for a movie to be perceived as great. Murphy’s criticism that technological praiseworthiness clouds our assessment of film quality is not born out in practice.

6. Artistic progress continues to be made to this day.

But here’s the thing, I think Murphy is right that culture is getting better. Murphy is wrong about why, but the fact is that we are living through the best time in all history to enjoy culture.

Earlier I proposed that the purpose of culture is to evoke thoughts and feelings. Cultural progress is the expanding capacity of our cultural repertoire to evoke our thoughts and feelings. And there has never been a more exciting time for the arts to evoke our thoughts and feelings.

• If you enjoy horror, you can watch It Follows and The Babadook, The Thing and The Shining. Old or new, each of these films evoke fear and dread.
• If you enjoy suspense, you can read And Then There Were None, Misery and Red Dragon, you can watch Vertigo and Wages of Fear. Old or new, each of these works continues to evoke tension.
• If you enjoy animation and gorgeous visuals, you can watch Into the Spiderverse, Spirited Away or Snow White. Old or new, each is beautiful.
• If you enjoy spooky vampires you can watch Nosferatu if you enjoy spooky Swedish vampires you can watch Let the Right One In, if you enjoy Mormon vampires you can watch Twilight. Where the range of vampire films was once narrow, it is now wide and caters to many tastes.
• If you enjoy American epics, you can watch Citizen Kane and The Godfather, you can read Last of the Mohicans and Freedom. It’s a sign of cultural progress that authors continue to evoke new, complex thoughts and feelings towards the world’s last super-power.
• If you enjoy action, you can watch Mad Max Fury Road, Terminator 2 and Die Hard. They’re all awesome.

Thirty years ago, if you enjoyed a certain genre of art, there may have been one great work to enjoy. Today, there are dozens. That is progress.

7. We underrate contemporary art because of biases

Despite the fact that masterpieces continue to be released, many doubt that today’s cultural output is as good as that of the past. These doubters are wrong. They are wrong because they’ve fallen prey to a suite of psychological biases.

We are prone to recency bias, where we put too much emphasis on recent memories. A lot of cultural products in any era just aren’t that good, but we consume them because they’re novel and because other people are consuming them. This recency bias means that contemporary mediocrities are fresh in our memories while we forget about past mediocrities altogether.

Recency bias is compounded by availability heuristic, where we judge the frequency of something by how readily it comes to mind. We overestimate the frequency of exciting but rare events like shark attacks because they are vivid and memorable. Great pieces of culture are memorable. Mediocre ones forgettable. So when we’re asked to name a song, film or painting from the past, we tend to remember only the good ones. The availability heuristic then leads us to think that because good culture from the past is easier to remember, there must have been more of it.

The past had just as many dud songs, films and other works as we have now. The only difference is that we forget about past duds, we only remember past masters, and we are all too aware of recent mediocrities. This creates a false impression that the past produced more classics than today.

The truth is that we are producing as many great cultural works today as we ever had. Our culture’s capacity to evoke thoughts and feelings continues to grow. Each new masterpiece is a contribution to the cultural repertoire.

8. Progress is positive sum

My final objection relates to the nature of progress as Murphy describes it. I may be reading more into Murphy’s words than I think is fair, so consider my epistemic status of this section far less certain.

Murphy opens the piece with a set of rhetorical questions

“If someone refused to use Google Maps and preferred to rely on a vellum chart from the 1400s, we would think they were very strange. If a bank communicated with its branches via telegraph, who would deposit money with them? If an army used guns from 1650 in a modern battle, they would meet certain death”

This is an impoverished way to think of progress. Progress is not about the tools we use, like vellum charts, telegraphs and muskets. Progress is about better outcomes — like reducing inter-state violence, or reducing developmental disorders associated with lead-contaminated drinking water.

Consider how Murphy’s rhetorical question would come across if posed differently.

• If a construction worker built your house using a hammer invented in the 1500s, you’d think they were very strange — but this is exactly what a claw hammer is and construction workers use them all the time because they achieve the outcome of building sturdy buildings.
• If your doctor recommended you prevent the spread of fomites using a technique developed during the Islamic Golden Age, they would become a super-spreader — but the invention of soap goes back to medieval Syria (if not earlier!) and we still use soap because it achieves the outcome of killing viruses.
• If someone refused to use Google Maps… — which is what hikers do on multi-day hikes. You won’t have mobile reception and even if you did Google Maps provides an inferior description of terrain contours, hazards and search-and-rescue coordinates. Google Maps is awesome for achieving the outcome of driving across town. But if you go hiking, you must take a topographic map or else the outcome may involve getting lost and dying.

My concern with Murphy’s focus on means rather than outcomes is that it paints a picture of progress that is zero-sum. For map navigation to progress, vellum charts must be removed. For war fighting to progress, old machinery must be abandoned. For culture to progress we must remove old classics from the cultural canon.

Alyssa Rosenberg has expressed why this is a tragic view of progress better than I could. She was writing in 2017 when La La Land was accidentally awarded the Oscar meant for Moonlight:

“I hate that the contest between “Moonlight” and “La La Land” … emphasized the sense that they were locked in a zero-sum contest, and that for “Moonlight” to win, “La La Land” not merely had to lose, but to be defeated.

The truth is that both “Moonlight” and “La La Land” are highly moving, original works… <by> directors of tremendous vision who are going to compete for Academy Awards again, maybe even against each other.

Let’s hope the next time they do, they aren’t pitted against each other in a way that mirrors an ugly idea about… progress that is ascendant in our politics: that the only way for members of one community to advance is at the expense of another.” (Rosenberg, 2017, emphasis original)

Replace La La Land with Citizen Kane and Ford v Ferrari with Moonlight and I feel like Rosenberg is describing Murphy’s article. This zero-sum approach is not supportive of cultural progress. What cultural progress needs is for us to think about culture as a positive sum endeavour. I think Tyler Cowen is right about this too, so I’ll give him the last word:

“Seeing Roger McGuinn perform has not detracted from my memories of seeing Paul McCartney or for that matter seeing Vladimir Horowitz… In many cases good experiences can even complement each other, rather than detract. (Cowen, 2005)

Great piece of culture do not complete — they complement each other. Both old and new works are a legitimate way to evoke our thoughts and feelings. We are making cultural progress right now because new works are expanding our cultural repertoire’s capacity to evoke our thoughts and feelings. Praising the great works that have come before us doesn’t detract from that progress.