I suspect it might have to do with (the representation of the thing) and (the thing) tending to blend together in people’s minds. Once you’ve learned to read fluently, seeing a string of writing will make you think of the meaning of the words rather than the underlying letters. And especially someone who is only familiar with one writing system is likely to see things not as a property of the writing system, but as a property of the words themselves. So instead of thinking “this writing system makes this word hard to spell”, they’ll just think “this word is hard to spell”.
In a similar way, I would expect the average person only familiar with Roman numerals to think not “our number system makes it hard to write down numbers efficiently”, but just “it’s hard to write down numbers efficiently”. In order to realize that the difficulty is a property of number system, you first need the idea that it’s possible for a number system to represent numbers more efficiently than you are currently doing, which is exactly the idea that you are missing if nobody has invented a better number system yet.
That explanation does still leave it a bit confusing why the abacus didn’t work as an example of an alternative number system. The one thing that comes to mind is that the abacus is a device for doing calculations by physically manipulating the beads, while Roman numerals are something that you write down. There are a lot of mathematical equivalencies that seem obvious to us but needed to be explicitly learned—it’s not immediately obvious to all children that 2 times 4 and 4 times 2 are the same thing, for instance. Likewise, if a culture doesn’t have the abstract concept of “a representational system” yet, it may not be very obvious to them that an abacus and a system for writing down numbers have anything to do with each other. “They’re different things for different purposes” may be the default thought.
I suspect it might have to do with (the representation of the thing) and (the thing) tending to blend together in people’s minds. Once you’ve learned to read fluently, seeing a string of writing will make you think of the meaning of the words rather than the underlying letters. And especially someone who is only familiar with one writing system is likely to see things not as a property of the writing system, but as a property of the words themselves. So instead of thinking “this writing system makes this word hard to spell”, they’ll just think “this word is hard to spell”.
In a similar way, I would expect the average person only familiar with Roman numerals to think not “our number system makes it hard to write down numbers efficiently”, but just “it’s hard to write down numbers efficiently”. In order to realize that the difficulty is a property of number system, you first need the idea that it’s possible for a number system to represent numbers more efficiently than you are currently doing, which is exactly the idea that you are missing if nobody has invented a better number system yet.
That explanation does still leave it a bit confusing why the abacus didn’t work as an example of an alternative number system. The one thing that comes to mind is that the abacus is a device for doing calculations by physically manipulating the beads, while Roman numerals are something that you write down. There are a lot of mathematical equivalencies that seem obvious to us but needed to be explicitly learned—it’s not immediately obvious to all children that 2 times 4 and 4 times 2 are the same thing, for instance. Likewise, if a culture doesn’t have the abstract concept of “a representational system” yet, it may not be very obvious to them that an abacus and a system for writing down numbers have anything to do with each other. “They’re different things for different purposes” may be the default thought.