From ancient Greece 32 tragedies are available to us. These plays, from the 5th century BCE, mostly don’t have a good ending. Hence our use of the word tragedy.

The 32 remaining are from just three authors, Euripides (18), Aeschylus (7) and Sophocles (7).
Hundreds of tragedies have been written, we know of some 300 more through fragments and titles from just those three authors. There were other authors, and we know the number of festivals and stagings etc. that we know took place implies there have been hundreds more than that. Just at the Athens festivals alone an estimated 648 different tragedies have been performed over a 70 year period in the 5th century BCE. Just 1 or 2 percent of tragedies written and staged are left to us. A tragedy in its own right.

Why these 32 works? These three authors mostly because a century onwards, the Greeks themselves held them in high regard. But beyond that, why these out of the 300 or so the 3 authors wrote? Mostly because of the Romans. 24 of 32 remaining works were selected by Roman grammatic scholars 600 years later, becoming canon in the 2nd and 3rd centuries CE. All 7 from both Aeschylus and Sophocles, and 10 by Euripides. Likely because of their link to the Homeric tales, which was central in Roman educational efforts. The number 24 isn’t a coincidence either, aiding memory techniques and echoing the 24 chapters in the Iliad and the Odyssee. Of these 24 we have multiple manuscripts and comments and marginalia, because they were taught and used.

This filter, 600 years after the tragedies were written, and 1800 years ago to us means that neither the ancient Greeks nor we ourselves have had any influence on the logic of this selection. The 24 handed down to us through the Roman filter, weren’t selected because the Greeks thought them most representative, nor because we think they are the most outstanding work. And possibly not because of esthetic reasons at all, but because of Roman educational preferences in teaching Greek grammar.

The other 8, all by Euripides, have come to us via a different way. Their titles are in alphabetic order, ε to ι, a piece of what once would have been a full overview of Euripides work, being copied together by some scribes. No selection criteria, just coincidence that one part of these ‘collected works’ survived. No annotations or comments either, just the works.

We know tragedies end badly, right? The ancient Greek word, originally meaning something like ‘goat song’, has come to mean that very thing to us. But it’s not that the ancient Greek playwrights were all depressed or nihilists. It seems it’s just that the Romans selected 24 works that mostly ended badly. Contemporary sources from Ancient Greece tell us tragedies did also end happily, and that such plays were very popular. Yet, the Roman selection 600 years later definitely picked mostly ones ending in, well, tragedy.

Comparing the two groups of plays by Euripides gives an enormous contrast though: of the 10 Roman selected works by Euripides only 2 end happily (20%), of the 8 ‘alphabetical’ ones by the same author, 7 end happily (88%). A reverse image!

The Roman selection is what made tragedies tragedies for us. A group of people 600 years removed from the source of these plays and 1800 years removed from us. For their own, unknowable to us, reasons. It’s a survival bias, not an inherent trait of the plays concerned.

And we’re stuck with that choice.

Tragic.

Also highly fascinating, I find.

Sources: Libraries of the Mind by William Marx (2025), page 3 and 112-116, Greek Tragedy on Wikipedia.