Coopers Law

"All machines are amplifiers."

This is to say, all machines amplify some human ability. The wheel, the lever, the pulley, the AddingMachine?, and so on.

A curious fact about Cooper's Law is that nobody knows who Cooper was. If your Google skills outclass mine, maybe you can find out.

Counter-examples: speedbumps, blindfolds, earmuffs etc.

Speedbumps amplify human ability to live safely/quietly. Blindfolds/earmuffs amplify human ability to ignore unpleasant stimuli.

VotingMachine?s present an interesting conundrum. Plainly it is possible to vote without them. Plainly it is possible to count votes without them. India, the world's most populous democracy, has no such things as voting machines.

So what is it that voting machines amplify?

I would suggest that the purpose of voting machines is to amplify the efficiency and correctness of the voting process.
Perhaps, but what is it they actually amplify?

Also only because you can amplify something doesn't mean that you have to.

so a voting machine is an amplifier with a gain of 1 (or possibly less than 1)?

If the machine is properly characterized, what it amplifies becomes evident. Voting machines embody memory. They make it possible to remember what millions of people said at a particular time. This may not be the most refined distillation of what they do, but it gives a human ability context.
I'm sorry, but I don't understand this. Human verified accounting of inked paper slips is demonstrably a far less fallible form of memory than voting machines. It seems more likely that what voting machines amplify is selective memory.
There seems to be an assumption that "voting machines are inherently flawed." Voting, in its simplest form, is the simple recording and tallying of choices. A simple voting machine can do that without any degradation, and can render tallies faster than a bunch or people.
- Actually, no, it can't. The reason for this is that humans are inherently greedy and dishonest. They will always try to rort the system. So voting inherently requires human review of each vote, and of each reviewer of each vote, and so on. The the only way to reliably accomplish that review is socially, in a counting room.
- As to speed, none of the extremely large democracies that don't use voting machines have any problem obtaining election results within a few hours or at worst days of the election. Speed is not the problem. Neither is cost, since the counters are volunteers and will always scale with the number of voters. The only thing voting machines can possibly amplify is bias.
Voting on a grand scale is simply a larger problem of logistics, and machines or systems of machines can be scaled to match. It's not even a hard problem, just big. When you make security and secrecy part of the problem, and when you have to solve cheating and dishonesty along with overcoming the learning curve for users, the problem becomes more complex. Part of the problem is clearly not the machine but the fox guarding the hen house: "Quid custodiet ipsos custodes?" So, to establish our foundation assumption ("voting machines are a fundamentally flawed concept" or whatever), is the fallibility cited above simply one of implementation, or is the assertion one of "voting machines can't do the job better than a bunch of people?"
- It's a social duty. Ergo voting machines can't do it better than a bunch of people. These machines introduce nothing but vulnerability into a social system that otherwise works well.

I think the above assertion about voting machines embodying memory is off-base. As I see it, a voting machine amplifies the ability of a human to count large numbers of votes - ostensibly, without error. Remembering what was voted should be factored out to elsewhere - in the case of "old-fashioned" voting, paper totalling sheets.

They inherently amplify your "ostensibly".

Well, hmmm. The computers used for the census do count, but what makes the census computer useful is its ability to group and regroup the data, creating various tallies. Voting machines, by themselves (in the applications I've seen), don't do the counting or tallying but rather simply store the choices made by the voters. The (stored) votes are then transferred to a central box where they're tallied. The voting machine stations combined with the central tally box could be considered a complete system, but we still have the storage/remembering thing as separate from the tally. Like the census thing, there are more things than just counting that the tally box might do (demographic analysis and so on).

Let the demographics happen after the count. That's what most democracies do. Election night is the same experience for their citizens as for Americans except for one thing - they don't doubt the accuracy of the count.

So, considered as a system, voting machines do count, but surely the storage feature is important. In its purest form, voting is just a count, but in my experience the ballots are not discarded for some time. The voting machine (system) seems to fulfill the counting function as well as the archiving function, with the added benefit that the election can be sliced and diced to glean deeper information. Am I still wandering in the woods?

Yes. Once tallies have been made by a social voting/counting process, what's produced is information. You can stuff that into any excel spreadsheet and do whatever you like with it. Spreadsheets amplify your ability to analyze information. A voting machine is not an excel spreadsheet.

So, a voting machine implements (amplifies) - or tries to - trusted counting.