4 ms·

Even among words that do theoretically have an ordering preference, some people have different habits. I knew someone who habitually said "three or two", rather

Even among words that do theoretically have an ordering preference, some people have different habits. I knew someone who habitually said "three or two", rather than "two or three". Leaving aside the effect of prosody mentioned in another comment, I think the rest of it is habit, together with the brain's tendency to group things. If you're used to hearing "plus or minus", your brain may be grouping it together into a phrase whose meaning you understand directly without decomposition, so if you hear "minus or plus" there's a moment of having to map the components to the composite meaning, together with wondering if the speaker intends the difference to have a meaning. (In mathematics, there's a reason to have both ± and ∓, since you can use them both in the same equation when you need "minus when the other one is plus, and plus when the other one is minus". For instance, a±b∓c means "a+b-c or a-b+c".) I don't know if this has a name, but if it does I'd love to know it.

LordDragonfang 2y ago
> For instance, a±b∓c means "a+b-c or a-b+c". I feel like this would be less confusing if rendered as "±(a+b-c)". It's the same amount of cognitive effort for figuring out the second term/factor, but you get the first one for free.
- z3dd 2y ago
  that makes a and -a
- billyjmc 2y ago
  What you’ve suggested is not equivalent. The a is always taken without negation, and that’s not the case in your rendering.
  - LordDragonfang 2y ago
    Ah. "a±(b-c)" then. The fact I made that mistake in the first place just makes me more convinced the original form is unnecessarily complex to parse.
    - stabbles 2y ago
      You're too fixated on the example. It's obviously used in cases where you cannot group like that: (-b ± sqrt(b² - 4ac)) / 2a = 2c / (-b ∓ sqrt(b² - 4ac))
huhtenberg 2y ago
"three or two" means "likely three, maybe two" "two or three" means "likely two, maybe three" Here, the ordering is significant and has nothing to do with a habit.
- JoshTriplett 2y ago
  That's not a common connotation; I've never once seen anyone use those phrases to mean that. "two or three" does not typically have any connotation of either being more likely.
  - huhtenberg 2y ago
    Might be something specific to where you are? What I described is how this construct is used in every place I lived.
    - arrowsmith 2y ago
      Have you ever lived in the UK? Because I've never thought of "two or three" as meaning that two is more likely than three. Are you sure that other people actually mean this when they say it, or are you just adding the extra connotation in your head? Also I'm not sure I've ever heard anyone say "three or two". That sounds weird, like saying "fork and knife".
      - huhtenberg 2y ago
        Not in the UK, no. But I did live in Canada, the US and across Europe. Not that it's commonly used and not with 3 and 2 specifically, but when used and the pair is inverted, it's meant to convey that the former is more likely. That's just like ... common sense, really.
      - alarge 2y ago
        Lest anyone draw the conclusion that this is a US thing, it certainly is not. I've lived in many places throughout the US and have never heard anyone say "three or two". Nor does "two or three" carry the connotation that two is likelier. A closer fit to that would be "a couple", which formally means two, but sometimes means "a small number".

LordDragonfang 2y ago

> For instance, a±b∓c means "a+b-c or a-b+c". I feel like this would be less confusing if rendered as "±(a+b-c)". It's the same amount of cognitive effort for figuring out the second term/factor, but you get the first one for free.

z3dd 2y ago
that makes a and -a
billyjmc 2y ago
What you’ve suggested is not equivalent. The a is always taken without negation, and that’s not the case in your rendering.
- LordDragonfang 2y ago
  Ah. "a±(b-c)" then. The fact I made that mistake in the first place just makes me more convinced the original form is unnecessarily complex to parse.
  - stabbles 2y ago
    You're too fixated on the example. It's obviously used in cases where you cannot group like that: (-b ± sqrt(b² - 4ac)) / 2a = 2c / (-b ∓ sqrt(b² - 4ac))

huhtenberg 2y ago

"three or two" means "likely three, maybe two" "two or three" means "likely two, maybe three" Here, the ordering is significant and has nothing to do with a habit.

JoshTriplett 2y ago
That's not a common connotation; I've never once seen anyone use those phrases to mean that. "two or three" does not typically have any connotation of either being more likely.
- huhtenberg 2y ago
  Might be something specific to where you are? What I described is how this construct is used in every place I lived.
  - arrowsmith 2y ago
    Have you ever lived in the UK? Because I've never thought of "two or three" as meaning that two is more likely than three. Are you sure that other people actually mean this when they say it, or are you just adding the extra connotation in your head? Also I'm not sure I've ever heard anyone say "three or two". That sounds weird, like saying "fork and knife".
    - huhtenberg 2y ago
      Not in the UK, no. But I did live in Canada, the US and across Europe. Not that it's commonly used and not with 3 and 2 specifically, but when used and the pair is inverted, it's meant to convey that the former is more likely. That's just like ... common sense, really.
    - alarge 2y ago
      Lest anyone draw the conclusion that this is a US thing, it certainly is not. I've lived in many places throughout the US and have never heard anyone say "three or two". Nor does "two or three" carry the connotation that two is likelier. A closer fit to that would be "a couple", which formally means two, but sometimes means "a small number".