Nehaveigur

When Mozart was my age, he was dead for five years already. He had gone from composing to decomposing.

He wasn’t the only composer to die young. Chopin died at 39, Gershwin at 38, Felix Mendelssohn at 38. This made me wonder if there’s a pattern. Are composers splendid torches that burn brightly and quickly?

To analyze this properly, you’d have to carefully compare the life span of each composer to the actuarial tables for the age and place they lived in, and my life, even though it’s longer than Mozart’s, is still too short for that. Instead, I asked three different AI agents to make a table for me. Here’s the prompt I used:

For each year between 1700 and 1900, pick one writer, one composer and one statesman (or politician) who was born that year and get their age at death. Return results in a table with the columns Year, Writer, Writer’s age at death, Composer, Composer’s age at death, Statesman, Statesman’s age at death. Limit the list to Europeans and North Americans. Start by looking for lists of writers, composers and statesmen that include birth and death dates.

By picking one composer for each year and matching them with a writer and a statesman, I was hoping to avoid a systematic error caused by the life expectancy changing over time. I also considered matching by country, but I thought that there wouldn’t be enough composers for this. I’d also have loved to include women, but there wouldn’t have been enough for statistical analysis either.

The three AIs I asked were:

• ChatGPT 4.0 Reason
• Gemini 2.0 Flash Thinking Experimental
• DeepSeek DeepThink R1

All three of them were free of charge. The only one that provided an adequate answer was Gemini. It initially tried to get away with providing a much smaller table, but when I insisted that I wanted all years between 1700 and 1900, it grudgingly made it.

AIs can be lazy. DeepSeek kept telling me that its server was too busy and ChatGPT kept asking me if I didn’t want to generate a slightly different table that it felt was easier to make until I gave up.

The list revealed that composers tended to have a shorter life (median: 66 years) than statesmen (median: 70 years), but about the same than writers (median: 65 years). The difference between composers and statesmen or between composers and writers was not statistically significant (p=0.10 and 0.41, respectively; one-tailed paired t-test).

I was surprised that composers’ median lives were only 4 years shorter than statesmen’s, especially since statesmen are not a random population sample. Unlike composers, who often do their best work young, men typically have to reach middle age before they make enough of a difference to be considered statesmen. The youngest statesman on my list was Nathan Hale, who died as a 21-year-old, executed by the British during the Revolutionary War. Reading his Wikipedia entry, I can only conclude that Gemini’s definition of a statesman is different from mine, because while he was admirably brave, I’d characterize him as a junior soldier rather than as a statesman.

In conclusion, AI can be fun but don’t fully trust it.