13

I would interpret $\sum_{i=1}^2 x_i + y$ as $x_1 + x_2 + y$, but I would interpret $\sum_{i=1}^2 x_i + y_i$ as $x_1 + y_1 + x_2 + y_2$. I realize this is a little inconsistent. Should the latter be written as $\sum_{i=1}^2 (x_i + y_i)$?

Or, in other words, does the sum operator have precedence over + and - ?

Nagel
  • 233
  • 1
  • 5

1 Answers1

10

It should be written with parentheses to avoid ambiguity, yes. If you think about the "sum" symbol as a function, it makes sense:

$$ \sum(\cdot) $$

This is a function which takes a list $\{x_1,x_2,\ldots\}$ of numbers (or other mathematical objects which you might want to add), and adds them in order. This list could be finite or infinite: the "sum" function figures out how long the list is, and adjusts its indexing accordingly (i.e. if there are 10 things in the list, your index will go from $1$ to $10$).

If you want to add a sequence which is itself the addition of two sequences, like your example of $\{x_1+y_1,x_2+y_2\}$, you'll need to drop the whole sequence into the function:

$$ \sum(\{x_1+y_1,x_2+y_2\})=\sum_{i=1}^2(x_i+y_i) $$

For finite sums of numbers, we always have the property that

$$ \sum(\{x_1,x_2,\ldots\}+\{y_1,y_2,\ldots\})=\sum(\{x_1,x_2,\ldots\})+\sum(\{y_1,y_2,\ldots\}) $$ however the following would have a different interpretation:

$$ \sum(\{x_1,x_2,\ldots\})+\{y_1,y_2,\ldots\} $$

Hence the parentheses!

Interesting, yet more advanced side note: "breaking up" a sum doesn't always work if the sequences are infinite!

icurays1
  • 17,161
  • 1
    Thanks! Does this mean that if I wrote $\sum_{i=1}^2 x_i + y_i$, you would read it as $x_1 + x_2 + y_i$? Or would you have understood that I meant $x_1 + y_1 + x_2 + y_2$? Or would you just have thought "Geez, what sloppy notation!" :) – Nagel Apr 16 '13 at 15:41
  • 2
    Most people would understand what is meant, but yes we would also think it was a bit sloppy =) I see $\int f+g$ all the time as well, similar confusion. – icurays1 Apr 16 '13 at 16:14
  • Right. But for integrals the d tends to take the ambiguity out of it. You would write $\int f + g , \mathrm{d}x$, and not $\int (f + g) \mathrm{d}x$, right? – Nagel Apr 16 '13 at 16:58
  • 1
    Correct, but sometimes people are lazy and don't put the 'dx'. – icurays1 Apr 16 '13 at 17:06
  • 1
    However, $\Sigma^2_{i=1} x_i + y$ would be $x_1 + x_2 + 2y$. If that is your intention, it would be preferable to write $\Sigma^2_{i=1} (x_i + y)$. Otherwise, one would write $y + \Sigma^2_{i=1} x_i$, which would be unambiguously taken to mean $x_1 + x_2 + y$ . – colormegone Apr 16 '13 at 17:19
  • 2
    Agreed. So $\sum_{i=1}^2 x_i + y$ would be considered somewhat ambiguous, or would most people take it to mean $y+\sum_{i=1}^2 x_i$ without hesitation? – Nagel Apr 16 '13 at 18:09
  • I would read it as $\Sigma^2_{i=1} (x_{i} + y)$. I suspect many people would, since the summation sign is often understood to be applied to "everything" to the right of it. – colormegone Apr 18 '13 at 01:51