Overbar notation for negation

There are many ways we can write down negatives (or more general “negations”). The most common way is to prefix terms with a “minus sign”, like 5unary_minus5 for the negative of 55, but I think minus signs are too cumbersome and confusing. On this page, I’ll explain why I instead prefer to draw a line/bar along the top of a term (which I pronounce as “negative”, rather than “minus” or “bar”).

In fact, any notation which affects the entire length of a term will have these advantages; for example we could write in a different colour. I just find overbars the most convenient (e.g. they don’t require swapping pens)!

Here are some examples of this notation:

Note that this bar notation is not original. Negatives are an “additive inverse”, and it is quite common to indicate other sorts of “inverse” with an overbar, for example:

Readability and aesthetics: long and short bars

Consider a long expression like this:

x(2+123.45+42y27¯+96)timesxplus2unary_minusplus123.45 times42 y2796

The bar clearly shows which parts are negative, without needing more parentheses; although we can add them if desired (in fact, grouping used to be written with such overlines; before parentheses became common!).

For comparison, using minus signs would give this:

x(2+((123.45+42y27))+96)timesxplus2unary_minusplus123.45 times42 y2796

Even if we don’t like such long bars, we could instead multiply by 1¯unary_minus1 and still require fewer parentheses than using a minus sign!

x(2+1¯(123.45+42y27)+96)timesxplus2timesunary_minus1plus123.45 times42 y2796

See the page on negatives for more discussion about 1¯unary_minus1.

Alignment and spacing

Adding bars doesn’t add any horizontal space, which makes it easier to align things. For example, here’s a small times-table written using bar notation. The contents of each cell is an independent MathML expression: no attempt has been made to align them, yet they manage to look pretty clear.

×× 2¯unary_minus2 1¯unary_minus1 00 11 22
2¯unary_minus2 44 22 00 2¯unary_minus2 4¯unary_minus4
1¯unary_minus1 22 11 00 1¯unary_minus1 2¯unary_minus2
00 00 00 00 00 00
11 2¯unary_minus2 1¯unary_minus1 00 11 22
22 4¯unary_minus4 2¯unary_minus2 00 22 44
Multiplication table with overbars for negation

Bars add a little vertical space to an expression, which causes a little wiggling; but only due to the thickness of the line (and some separation). In comparison, minus signs extend an expression horizontally by the entire length of the line (plus some separation). Here’s the same table using minus signs, and the wiggling between cells is much more pronounced!

×× 2unary_minus2 1unary_minus1 00 11 22
2unary_minus2 44 22 00 2unary_minus2 4unary_minus4
1unary_minus1 22 11 00 1unary_minus1 2unary_minus2
00 00 00 00 00 00
11 2unary_minus2 1unary_minus1 00 11 22
22 4unary_minus4 2unary_minus2 00 22 44
Multiplication table with minus signs for negation

This seems like a minor quibble, but grids of numbers/expressions are used all over the place: spreadsheets, long addition/multiplication, matrices, etc. When things line-up, we can skim them quickly and spot discrepencies; when they don’t, we must expend mental effort to keep track of items individually.

Negative digits

Thanks to the place-value system we use to write them, numbers themselves can be thought of as a grid of digits. Negative numbers are thus made of negative digits, which can make arithmetic simpler and more uniform. This is natural to express with overbar notation, but awkward and ambiguous using minus signs.

Known Conflicts

Bar notation is also used for things which have nothing to do with negatives, which makes for unfortunate clashes. Here are some I’m aware of:

Please suggest more clashes!

Repeating Decimals

Bars are sometimes used for repeating decimals, e.g. 1 . 23 ¯ to represent 1.23232323 . Since this varies between countries, it can be avoided in favour of a different notation.

p-adic Numbers

p-adic numbers have the same conflict as repeating decimals, except going to the left instead of right. If we pick one of the other common notations for repeating decimals, we should also use it for p-adic numbers, for consistency.