Clicky

- Supplement for "Where's today's Beethoven?"

Supplement for "Where's today's Beethoven?"

Data/calculations/charts

All of my data, calculations and charts are in this Google sheet.

Explanation of data sources

Science and art+literature+music, 1400-1950

Raw data is in the "Data 1 - acclaimed figs since 1400" sheet. Processed data, which I used for my charts, is in the "Data 2 - acclaimed figs since 1400" sheet.

I use the data set from Human Accomplishment, a book that aims to examine patterns in artistic and scientific achievement. I have a lot of issues with this book and largely disagree with its conclusions; the most relevant issue I have is that it presents its data on artistic and scientific figures as a good indicator of true excellence,1 whereas I am using them only as a measure of "critical acclaim" (see the main post), and think the latter is the right use of them.

But for purposes of measuring critical acclaim, I think the data set is useful:

  • It's generally hard to find data going back more than a couple hundred years, and this data set goes back well before the year zero (though it unfortunately ends in 1950).
  • It examines the amount of space that a number of reference works devoted to different artistic and scientific figures, interpreting this as indicating the perceived significance (which I often refer to as "acclaim") of these figures.
  • It provides data ranking the figures by "significance" (acclaim), along with information on where and when they did their work. It includes an "index" figure that attempts to convey how significant each figure is, rather than merely ranking them.
  • Generally, its top-ranked figure in a given category is unsurprising: Shakespeare for Western literature (there are separate rankings for Chinese literature, Arabic literature, etc.), Aristotle for Western philosophy, Confucius for Chinese philosophy, Michelangelo for Western art, Beethoven and Mozart (tied) for Western music, Darwin for biology, Galileo for astronomy and Isaac Newton for physics.

As discussed below, in this data, the highest per-capita "production of critically acclaimed ideas" by far comes from either ancient Greece, or Europe and the US after 1400. (As noted elsewhere, I am interpreting this as a fact about the "critical acclaim" being measured rather than a fact about objective excellence.) In the main piece, I have focused on trends within Europe and the US after 1400, and discussed ancient Greece separately.

Technological innovation, 1930-present

Not in spreadsheet; instead I simply screenshotted some charts for my post.

I used the very well-known and widely-cited paper Are Ideas Getting Harder to Find?, and also looked at this more recent post from New Things Under the Sun.

20th century film

Raw data is in the "Raw - film and albums" sheet. Processed data (for charts) is in the "Data 2 - film, TV, etc." sheet.

I used the TSPDT list, which was recommended by Luke Muehlhauser after he saw an initial draft citing the Sight & Sound list (the latter was first in Google results for "greatest films of all time", and also highlighted here). The TSPDT "top 100" produced a very similar chart to the "Sight & Sound" list (which was only 100); extending to the top 1000 changed the picture somewhat (in particular, making later periods look somewhat better), but not much. If you want to see the top 100 only, you can make a copy of my sheet and edit Column I so that it says "=countif('Raw - film and albums'!$E$3:$E$102,$A12)" instead of "=countif('Raw - film and albums'!$E$3:$E$1002,$A12)" (this just means changing ($E$1002 to $E$102).

Unlike nearly all of the lists below, the Sight & Sound list gave scores (not just rankings), and I used these to produce "Weighted" charts - not just for film but for all the below categories as well (the assumption here is that the difference between #1 and #2, #2 and #3, etc. is the same for all categories as it is for film). The "weighted" charts generally looked similar to the "unweighted" charts, and by the time I switched my underlying source from Sight & Sound to TSPDT, I didn't bother updating the "weighted" versions.

20th century music

Raw data is in the "Raw - film and albums" sheet. Processed data (for charts) is in the "Data 2 - film, TV, etc." sheet.

I used the Acclaimed Music list, which was recommended by Luke Muehlhauser after he saw an initial draft citing the Rolling Stone list (the latter dominated Google results for "greatest albums of all time"). Whether I used the Rolling Stone top 100, the Acclaimed Music top 100, or the Acclaimed Music top 1000 didn't seem to make much difference to how the chart looked; I ended up using the Acclaimed Music top 1000. (I never tried using all 3000; I thought that would be going too far, relative to my other lists, and giving too much of an edge to the modern world, since the sheer quantity of music is probably higher in more recent years.)

If you want to see the top 100 only, you can make a copy of my sheet and edit Column Q so that it says "=countif('Raw - film and albums'!$M$3:$M$102,$A12)" instead of "=countif('Raw - film and albums'!$M$3:$M$1002,$A12)" (this just means changing ($M$1002 to $M$102).

Again, I didn't bother updating the "weighted" version of the chart after switching sources, because I ended up not including "weighted" charts in my main post (they generally look similar enough to unweighted charts).

TV shows

Raw data is in the "Data 1 - film, TV, video games, rock, baseball" sheet. Processed data (for charts) is in the "Data 2 - film, TV, etc." sheet.

I wasn't able to find anything as comprehensive as the lists above, so I used the Rolling Stone list, which looked most appealing based on a Google search for "greatest TV shows of all time." As a reminder, I'm generally going here for professional critic-based scores rather than broad audience scores (which Metacritic rankings tend to use).

Video games

Data is in the "Data 2 - film, TV, etc." sheet.

I found a Wikipedia page that aggregates multiple "best video games of all time" list, and that was the best I was able to find for my purposes.

Books

Data in the "Data - great books" sheet

I used https://thegreatestbooks.org/, which is prominent on Google and seems to be a very systematic list based on other rankings.

"Effective population" calculations

As discussed in the main post, I wanted a way to capture the idea that the number of people with a realistic shot at being scientific and artistic innovators has risen over time, likely more than the population has risen. This proved challenging, as it's not at all straightforward to get data on this point.

I broke the "effective population" down into:

  • The population.
  • The "effective population multiplier," meant to track the percentage of the population with a reasonable level of education, health, etc. This is not a literal multiplier, and I don't have literal estimates of effective population; instead, this is an "index," meaning that growth in the figure from year to year is meaningful, but the units themselves are not meaningful.

In the "Effective population - early data points" sheet, I collect estimates of changes in literacy rates, urbanization, extreme poverty, and university degrees. I then (lower down in the sheet) line them up next to each other by time period and take my own "best guess" at annualized growth in the "effective population multiplier" over that period. (I only estimated annualized growth in these things, not the absolute number of people in the "effective population.") A rough rule of thumb I used was to err toward using the largest rate of change from urbanization, poverty, literacy, and university degrees, rather than the total product of all of these things (I'd guess they're correlated, so this could double-count) or the average (if the literacy rate grew slowly due to already being >90%, while urbanization grew quickly, I'd expect the "effective population multiplier" to grow quickly).

For data after 1930, instead of relying on these measures, I used the more detailed-seeming analysis on the "number of researchers" from Are Ideas Getting Harder to Find? (See "Effective population - researchers since 1930" sheet.) This data implied growth rates of 2.5-5% per year, which are very high compared to my estimates for much earlier years, but not terribly higher than my estimates for 1900-1920 based on growth in the number of university degrees. Still, I tried an alternate series where I simply capped "effective population multiplier" growth at 2% per year (see "Effective population multipliers" sheet) to see what my charts looked like with more modest, closer-to-long-run-history estimates of "effective population" growth. This didn't change the picture a ton, and I didn't love charts with plummeting "innovation productivity" lines mostly driven by high estimates of "effective population multipliers" that I wasn't confident in, so I ended up using this "alternate" series in the main piece (although charts for both are in the Google sheet).

The "Effective population multiplier" sheet has the master series I used for effective population multipliers, which were then multiplied by population. It includes some very clumsy interpolation in order to get a series that is never too jumpy, but roughly matches the estimates I made of what annual growth rates should look like in each time period.

Looking more widely for candidate "golden ages" (pre-1950)

A long time ago, I did some additional analysis of the pre-1950 data set discussed above to identify the leading candidates for a "golden age" of scientific/artistic accomplishment. I haven't cleaned up this analysis much, so it's a bit of a mess.

I looked at every combination of a country and a 50-year period (for example, "France from 200 AD to 250 AD") in the data set, and ranked them in terms of "total significant figures per capita." A "significant figure" is any figure that scored high enough on the critical acclaim metrics to make it onto the list of roughly 4000 figures listed in the data set. I also tried quite a few variants on this metric, but none of them seemed to change the picture a lot - details in footnote.2

I looked at the top 68 country-periods - every period with at least one "significant figure" per million people. Out of these periods, 100% of them were either from "ancient Greece" or from Europe after 1400. (As a reminder, I'm interpreting this data as data about critical acclaim, with all of the biases (including Western bias) that implies, rather than as data about "true excellence," which is how the author presents it.)

Here are the top 10 excluding Ancient Greece:

Country-period # sig figs Est. mlns population Sig figs per mln people Top 5 figures
Netherlands 1600-1649

23

4.24

5.42

Rembrandt van Rijn; Descartes, René; Rubens, Peter Paul; Dyck, Anthony van; Hals, Frans
Netherlands 1650-1699

18

4.90

3.67

Huygens, Christiaan; Swammerdam, Jan; Spinoza, Benedict; Vermeer, Jan; Leeuwenhoek, Antoni van
Norway 1850-1899

14

4.35

3.22

Ibsen, Henrik; Waage, Peter; Bjørnson, Bjørnstjerne ; Grieg, Edvard; Lie, Marius
Denmark 1800-1849

9

2.80

3.21

Ørsted, Hans; Andersen, Hans ; Oehlenschlaeger, Adam; Thorvaldsen, Bertel; Grundtvig, Nikolai
Netherlands 1500-1549

8

2.55

3.14

Erasmus; La Rue, Pierre de; Patenier, Joachim; Aertsen, Pieter; David, Gerard (Gheeraert)
Switzerland 1900-1949

22

8.11

2.71

Einstein, Albert; Pauli, Wolfgang; Piccard, Auguste; Kipfer, Paul; Karrer, Paul
Britain 1600-1649

44

16.77

2.62

Shakespeare, William; Harvey, William; Milton, John; Hobbes, Thomas; Donne, John
Belgium 1500-1549

8

3.21

2.49

Gossaert, Jan (Mabuse); Gombert, Nicolas; Lucas van Leyden; Massys, Quentin; Buus, Jacques
Britain 1750-1799

77

31.07

2.48

Watt, James; Herschel, William; Hutton, James; Cavendish, Henry; Priestley, Joseph
Netherlands 1400-1449

4

1.64

2.43

Dufay, Guillaume; Koster, Lauren; Binchois, Gilles; Malouel, Jean; [n/a]

And here's a grid showing every country other than ancient Greece that had at least one period in the top 50, for every period after 1400 (all top-50 periods for these countries were after 1400):

"Significant figures" (by critical acclaim) per million people

1400-1449 1450-1499 1500-1549 1550-1599 1600-1649 1650-1699 1700-1749 1750-1799 1800-1849 1850-1899 1900-1949
Austria

0.00

0.68

0.38

0.17

0.00

0.86

0.47

1.99

1.30

1.17

2.22

Belgium

1.80

1.80

2.49

1.14

1.63

0.76

0.00

0.00

0.20

0.79

0.57

Britain

0.21

0.09

0.62

1.10

2.62

1.57

1.55

2.48

1.89

1.32

1.02

Denmark

0.00

0.00

0.00

0.61

0.00

2.04

0.52

1.95

3.21

1.76

1.40

France

0.10

0.29

0.35

0.53

0.33

0.74

0.60

0.94

1.27

1.40

1.30

Germany

0.23

0.29

1.18

0.44

0.96

0.82

0.67

1.23

1.41

0.98

0.83

Italy

1.01

2.04

2.02

1.67

1.74

0.72

1.06

0.55

0.32

0.25

0.34

Netherlands

2.43

1.99

3.14

1.83

5.42

3.67

0.86

0.43

0.00

1.23

0.90

Norway

0.00

0.00

0.00

0.00

0.00

0.00

0.00

0.54

0.75

3.22

0.90

Sweden

0.00

0.00

0.00

0.00

0.33

0.27

1.25

1.40

1.57

1.06

0.54

Switzerland

0.63

0.55

0.96

0.86

0.82

0.37

2.38

1.43

0.56

1.72

2.71

USA n/a n/a

0.00

0.00

0.00

0.00

0.00

0.82

0.95

0.71

0.97

(I threw in the USA as well, even though it narrowly missed qualifying.)

Some things that jump out at me:

  • The Netherlands of the 17th century looks like the best candidate for an outlier within post-1400s Europe.
  • Putting that potential "golden place" aside, there are 29 other country-periods here (out of 130) with more than 1.5 significant figures per million people, which is within a factor of ~2 of the top-scoring (remaining) country-period.
  • So if we're looking for a culture with particularly impressive innovation, we're probably better off thinking of "Europe after 1400" as a bloc of country-periods in which such cultures were particularly likely, rather than focusing on a single country-period or a handful of them.
  • You could argue that Britain 1600-1649 should be considered an outlier comparable to the 17th-century Netherlands, as it makes the top 10 despite a much larger population compared to other country-periods. (In most periods, other countries with comparable or larger populations - France, Germany, US - were within a factor of 2-3 of Britain on significant figures per capita, but this wasn't the case for the 1600-1649 period.)
  • Major cities tended to account for a huge percentage of their country-period's significant figures. (I did some rough figures on these but didn't get them to the point of being presentable.)

And now for Ancient Greece's top 10 (I actually included 11 just because):

Time period # sig figs Sig figs per mln people Top 5 figures
-450 to -401

28

22.86

Hippocrates of Cos; Euripides; Democritus; Socrates; Aristophanes
-350 to -301

16

13.06

Aristotle; Theophrastus; Herophilus of Alexandria; Praxagoras of Cos; Heraclides Ponticus
-500 to -451

11

8.98

Empedocles; Aeschylus ; Sophocles; Pindar of Cynoscephalae; Parmenides of Elea
-550 to -501

9

7.35

Pythagoras of Samos; Anacreon of Teos; Aesop ; Simonides of Ceos; Theognis of Megara
-400 to -351

9

7.35

Plato; Eudoxus; Xenophon of Athens; Archytas of Tarentum; Hippias of Elis
-300 to -251

8

6.53

Aristarchus of Samos; Erasistratus; Ctesibius; Theocritus of Syracuse; Sostrastes of Cnidos
-250 to -201

6

4.90

Archimedes; Apollonius of Perga; Eratosthenes; Apollonius of Alexandria; Chrysippus of Soli
-50 to -1

6

4.90

Strabo of Amasia; Dionysius of Halicarnassus; Aenesidemos; Agesander; Athanodorus
-600 to -551

5

4.08

Thales; Sappho of Lesbos; Anaximander the Elder; Solon of Athens; Anaximander of Miletus
50 to 99

4

3.27

Hero of Alexandria; Plutarch (Boeotia); Epictetus; Dionysius Areopagita; [n/a]
-700 to -651

4

3.27

Homer; Hesiod; Archilochos of Paros; Tyrtaeus of Sparta; [n/a]

This assumes a population of 1.2 million people born (standing population of 500,000) for each 50-year period.3 You could argue that the "eligible population" was effectively much lower (compared to the post-1400 periods above) due to the lower percentage of the population that was educated, had civil rights, etc., which would make the figures here even more eye-popping.

Twitter Facebook Reddit More

Footnotes

  1. From near the end: "Excellence is not simply a matter of opinion, though judgment enters into its identification. Excellence has attributes that can be identified, evaluated, and compared across works. The judgments reached by those who are most expert in their fields, and who work from standards of excellence that they are willing to specify and subject to the inspection of logic, are highly consistent—so consistent that eminence in the various domains of accomplishment can be gradated with higher reliability than is achieved by almost any other measure in the social and behavioral sciences. When the rating of eminence is scrutinized against the reasons for that eminence, it also becomes apparent that those who rank highest are those who have achieved at the highest levels of their field ...

    I specify experts because they are the people who write the histories that carry the names forward into time. But my point applies to a broader audience. Indulge me in one more thought experiment, a familiar one: You will be stranded on a desert island, and you can take just 10 books and 10 music CDs. What do you choose? My prediction is that even people who don’t listen to classical music regularly will take Bach, Mozart, and Beethoven. Even people who haven’t picked up Shakespeare in years will take the collected works of Shakespeare. When we want something we can go back to again and again, we choose the same giants that the experts choose. My proposition about the literature, music, and visual arts of the last half century is that hardly any of it has enough substance to satisfy, over time."

    And from Chapter 5: "Let it be understood from the outset that I do not consider eminence and importance to be slightly glorified measures of fame, but more than that. They are reflections of excellence in human accomplishment. The Sistine Chapel keeps popping up because it is home to one of the greatest works of art ever to come from a human hand and mind.

    In whose opinion? Who is to say that some paintings are fine art and others are not? That some poems are greater than others? That some music is classical and other music is pop? That the achievements of some scientists are central and others are peripheral? In a world where judgmental has become an insult, who is to judge? We have a long and winding road to travel in this chapter. First, I will describe what I define as excellence in the sciences and arts respectively. Next comes a description of my reasons for concluding that standard historiometric methods do a pretty good job of identifying excellence in the terms I have set ..."

    He goes on to give a number of arguments for this position. I don't find them compelling, but won't elaborate further here. 

    • I tried 20-, 50- and 100-year time periods.
    • In addition to "significant figures per capita," I also looked at "total index per capita," where "index" is a measure provided in the data set that weighs more significant figures more highly - for example, the top-ranked figures have an index of 100, whereas the figures near the bottom have an index of 1. (There are also 136 figures - out of ~4000 - with index values below 1 or no index value.) An issue I had with this is that a single major figure in a low-population country-period can cause a high value for the entire country-period, so I ultimately decided that the raw significant-figures count was less noisy while not having dramatically different conclusions.
    • I also looked at 50-year periods (only) for "subregions" (mostly cities) rather than countries. I wasn't able to do per-capita analysis here due to not having population data for the cities; instead I looked at raw significant figure counts. In this case, since I had already done the rest of the analysis, I didn't bother to look at time periods before 1400. In this analysis, it was Paris (especially) and London that stood out, rather than the Netherlands.
    • The results of all of these alternate analyses are summarized here
  2. I haven't found any good data on this. Based on Wikipedia, I've assumed a standing population of 500,000. The population is generally multiplied by 2.5 to get the estimated people born over a given period, though this is the same for all regions so cancels out.