ChatGPT can communicate better than a human. Sometimes.

So someone gave me the link to the public interface for the new AI. I decided to ask it a question that had always bugged me: What does E.M. Forster's phrase "only connect" actually mean? I posed this question once on my blog and got numerous replies, but not one gave a comprehensible answer to the question. It took ChatGPT to give me a usable answer.

"Only connect" is a phrase that comes from a novel called "Howards End" by E.M. Forster. The novel is about the relationships between different groups of people in England, and the idea behind the phrase is that people should try to connect with one another and understand each other, rather than remaining isolated and disconnected. In other words, it's about the importance of building relationships and forming connections with others.

OK, I went on, but why only connect? What else besides connect are you not supposed to do? I was even more astonished to get a straight answer to this one.

The phrase "only connect" is actually a bit of a paradox, because it suggests that the only thing that matters is connecting with others, but at the same time, it implies that there are other things that people should not do. In the context of the novel, the phrase is meant to convey the idea that relationships and connections with other people are the most important thing in life, and that everything else is secondary. So while the phrase doesn't specifically say what people should not do, it suggests that anything that doesn't involve building connections with others is less important.

All this must have come from somewhere, but unfortunately ChatGPT has no index to its source material, so I don't know what human was actually capable of explaining this. I got some equally useful answers to other questions of this kind, such as the meaning of the phrases "show me the money" (this was such a common catchphrase for a while that I actually watched the movie Jerry Maguire, from which it comes, for the sole purpose of finding out what the phrase meant, and came away unenlightened) and "off the wall" which I eventually figured out on my own, but which baffled me when I first heard it 45 years ago, and whose users couldn't explain it when I asked.

I decided to try Chat on some technical questions of musical terminology that I've noticed people getting wrong, and this was somewhat less successful. First I asked it if a piano quartet was a work for four pianos, a misapprehension I've encountered once or twice. It said accurately that "A piano quartet is a group of four musicians who play music together on the piano, violin, viola, and cello. It is not four pianos, but rather a combination of four different instruments," but it didn't say that the phrase "piano quartet" is actually shorthand for "quartet for piano and strings." I suggested it say this, and the Chat thanked me for the suggestion in such a way as to imply it already knew this but hadn't prioritized the information.

Then I asked it, "What instruments are in a string quintet?" because this is actually variable. It said correctly that "The most common combination of instruments in a string quintet is two violins, two violas, and a cello," but it didn't say anything about anything else. When I prompted it on that, it acknowledged there were other possibilities: "a string quintet could also include two violins, a viola, a cello, and a double bass," but all it said beyond that was that a combination with a harp or guitar isn't normally called a string quintet. It didn't offer two violins, one viola, and two cellos, which is the combination Schubert used in his String Quintet and is a little less unusual than the double-bass one.

Then I asked it the really tough one, "How many symphonies did Wolfgang Mozart write?" It answered in part, "It is difficult to say exactly how many symphonies he wrote, because some of his early works have been lost and there is some disagreement among scholars about which pieces should be considered symphonies." This is correct, though it didn't elaborate on the reasons for the disagreement (basically that Mozart tended to cobble together miscellaneous overtures and other pieces and use them as symphonies for concerts) or note a couple cases of questionable attributions, or a few past false attributions that have been crossed off the list.

Then it went on to say, "However, most sources agree that Mozart wrote between 41 and 51 symphonies," and this is entirely wrong. People often say 41 because that's the number on the "Jupiter," his last symphony, but what people who say that don't know is that the number was given not by Mozart but by an early 19th-century publisher who had an incomplete and inaccurate list. It's not only wrong, it's a shibboleth: anyone who claims to be a Mozart expert and says he wrote 41 symphonies has thereby revealed themselves to be no Mozart expert. Some omitted early works were later tacked on to the end of the list, and at one point it totaled 51, which must be where Chat got that number from, but that's also inaccurate. Neal Zaslaw's authoritative book Mozart's Symphonies lists 44 purpose-written symphonies, 21 of various cobbled-together kind, and 8 lost ones that we know about, making 73; and that's not counting 25 that have been incorrectly attributed to Mozart, 3 or 4 of which (one of them is uncertain) are among the list of 41. So watch it when it comes to Mozart.