It's worth remembering that APL and SML, amongst others, distinguish between the sign used for a negative literal (¯1 in APL, ~1 in SML) and the sign used for subtraction (the hyphen/minus in both of them). It doesn't seem to be a hard thing to get your head around in practice. From having worked on a Prolog system, I can tell you that the fact that -1 is a single token except when it isn't, yet -X is always two, caused headaches for implementors and confusion for users. In Smalltalk, -1 is a number, but x-1 is three tokens, not two. (You have to keep track of what the previous token was to tell what to do.) If I were making suggestions for Haskell' (other than please, pretty please with knobs on, let me keep n+k), one of them would be to introduce the character U+00AF (chosen because it's 8859-1, -4, -8, -9, and -15 at least) as a unary minus sign, allowing it to be used for exponent signs as well, so that ¯x - ¯1.0e¯10 is allowed. Then Haskell'' could remove the unary - . In the mean time, the unary - / binary - issue is something you run into hard ONCE, and then avoid easily enough, not unlike forgetting the back-ticks in x `div` y.