how wide would an LP be of every word ever spoken?

Clearly this can’t be calculated precisely, but I thought it would be an interesting thought/math experiment:

If we were able to put every word ever spoken by a human on one LP record, how wide would that LP be? And if we placed the earth at the center of the record, how far out would it go - what body in space would the edge be near?

I choose vinyl, since unlike a CD/DVD, I assume you can’t play nearly as much with compression, so the size should be easier to calculate if you know the total size.

To figure this out, we need to estimate how much the average person speaks in a day/year, how long the average person lives, and how many “average” people have lived in history.

Any ideas?

some starters:

[google - ‘how many humans in all history’] BBC estimates 107 billion people have ever lived , of whom about 7 billion alive now.

Average age - lets pick early middle age, because we can assume many of those people died younger and dragged the average down. - my guess 40 years old, of which 38 year are chatting at full steam

[google - ‘how much of your time do we spend talking’]- estimated 45% talking of 70-80% of waking time spent communicating, and assuming 16 hours awake. So [16 x 60 mins] x .75 x .45 = 324 mins talking per person per day. or 118,341 mins per person/year.

So our total would be - 107 billion people x 38 years x 118,341 = 4.8117E+17 minutes of blather.

Sometimes it feels like much more.


Probably the mo sensitive variable is average age. If it was 30 instead of 40, which could easily be plausible, that would knock it down about a quarter.

Ok, so if a 12 inch LP holds 22 minutes, what would the diameter of that LP be in kilometers? (It’s been a while since I had the math skills to calculate that…)

I’m looking at the same website you used, I think, and I think you got the percentage of time spent talking wrong. It says that we use 70-80% of our waking time communicating. Of that time spent communicating, it says we spend 30% of it speaking. So it’s .75 times .30, not .75 times .45 in your calculations. Even that strikes me as a vast overestimate. I don’t spend .75 times .30 time 16 hours (= 3.6 hours) each day talking, and I don’t think I’m a particularly quiet person:

Thanks Wendell.

Yes, I grabbed the wrong figure, so its 2/3 of whatever that number I provided was. That figure also factors in a proportion of old-man-self-muttering as well.

I’m not sure I’d buy the entire LP but if there was a 12" dance mix of me vaguely vocalising Ikea cabinet assembly instructions, I’d buy that!

If it’s just speech, you could almost triple the playtime since it requires only a single channel and the tracks could be much closer. On Techmoan, a 2 hour LP is reviewed:

Ok, so it seems we’ve done a good job of estimating the total amount of speech, and how much speech can fit on one normal LP. Any ideas how big the one that contained all of it would be?

That part’s easy enough. An ordinary LP holds 6e1 minutes. We’ve set the amount of speech as 3e17 minutes. So we need an LP with an area 5e15 times greater. That means the diameter must be greater by a factor of the square root of that, or about 7e7. An ordinary LP has a 1-foot diameter, so we’d be looking at a 70 million foot diameter. A mile is about 5000 feet, so that’s be about 13 thousand miles, not much bigger than the Earth’s diameter of 8,000 miles.

Though I think the speech figure is still too high. I interpret that stat as meaning that 30% of our communication time is talking (with the rest being gestures, writing, Instagram, etc.). And a lot of time we’re engaged in talk-communication, we’re listening to someone else, not talking ourselves.

Fascinating, thanks!

The logistics of constructing a giant turntable to play that record are daunting.

Better to just buy every record ever recorded.

At that size wouldn’t 33 1/3 RPM be approaching the speed of light?

It would have to be a constant linear veleocity not a constant angular velocity turntable. It it were constant angular the size relationship would be linear not square. Then it absolutely would be speed of light at the edge. The frequency response would be amazing.

The answer is about 18 feet diameter but there is a problem.

It would not be very big, let’s say the speaker speaks english only in their life. There are only about 170,000 words. If this person managed to speak all of them, let’s say at a average of 100 words/m that would equate to 1700 minutes, or 28 hours. Lets, and make the math easy, say that the person does not speak every single word, but let’s say 24/28 hours or 1 day’s worth of recording. At about 20 minutes per side we would need 72 records single sides or 36 using both sides. Now I’m guessing that the size of the ‘data’ area of a record is maybe 3 inches thick, and since RPM’s are constant, if we combine it into one record it would be 216 in of data area, or 18 feet in diameter, or 9 feet if double sided.

Checking to make sure the speed is OK at the outer surface:

Ignoring the center part taking up space, the RMP’s are 33.3 rpm’s the diameter is 18 ft, so the outside edge velocity would be 33.3x3.14x18= 1,882 ft/second which seems problematic for conventional vinyl, I would say it would tend to fly apart and another material would be needed. Even at half that for double sided it seems to be too fast. However it would qualify as a record before you attempt to play it, so 9 ft or 18 ft would be answers to satisfy the OP’s question.

We appear to have answered the question, so I’m going to throw this in. My immediate thought on reading the OP was “To Jupiter and Beyond the Infinite.”

kanicbird, I think you have misunderstood. The question was about every word ever spoken, not every unique word ever spoken. But even if the latter was the question, you have only counted English words. You would need to count the words in the other 6000 existing languages, plus the hundreds of thousands of extinct languages that have ever existed.

No it clearly says ‘spoken by a human’. ‘a’ being a singular human. Now I did make some simplifications that this person has only spoke English but considering I gave them speaking perhaps 80% of all english words I think it’s pretty reasonable overall as a down and dirty ballpark.

Re Multiple spoken times of the same word. That word was recorded so again fills te OP’s requirements.

Like it or not this is a valid answer to the OP’s Q. It may not be the only answer, but fundamentally it is correct.

Yes I did look for the easy way to do this and it did prove a interesting mathematical exercise, so that part also checks.

I do need to correct my math above, not that this changes my answer to the OP (just a notation of a potential issue), but the speed of 1,882 ft/second should have been 1,882 ft/min or 31 ft/sec or about 21 mph, which may not be a problem with vinyl. A needle would also have to be capable of this speed to read it. However all this part is beyond the requirements of the OP’s problem.

When Og the Caveman said “Ooga”, that was a word spoken by a human. When John the President said “Apollo”, that was also a word spoken by a human. So both would need to be included on the disc.

I think you are all forgetting something. As the size of the ‘LP’ gets larger, so does the circumference. Think the size of a 45 that has two songs, and a full LP that has 10. You could probably put an entire LP on one grove of a disk 100 yards wide.

No, I’m not gonna do the math.

No because the speed of rotation is the same.
Each revolution will be done in the exact same time (with perhaps relativistic effects if the edge gets near the speed of light)