I've always(-ish) thought that the proper way to typeset a vector which is labeled by a letter with a single character subscript, like v1, was to set the arrow over just the letter, like \vec{v}_1 (I realize the braces are superfluous here but I like to include them for clarity). But recently I have seen it argued that this is wrong, and that the proper way is to place the arrow over the combination of the letter and subscript, with \vec{v_1}. I have done this in the past but I consider the results of the former method to be better looking.

Is there a standard recommendation to do this one way or another in the TeX world? Or is it a matter of personal preference?

EDIT: to put the question another way, suppose you're proofreading a LaTeX document, and suppose that all the notation is clearly defined so that there is no ambiguity about what v-with-subscript-1-and-arrow-over-it means:

  • If the author consistently writes \vec{v_1}, would you globally change it to \vec{v}_1?
  • If the author consistently writes \vec{v}_1, would you globally change it to \vec{v_1}?
  • Or neither of the above (i.e. it doesn't matter so respect the author's original preference)?
  • 3
    It comes down to personal preference (or how you define notation at the beginning of the document) but I think \vec{v_1} is more clearly the vector v_1, \vec{v}_1 could be read as something like the 1st component of the vector v. It depends on the context what _i means of course.... Commented Feb 16, 2012 at 20:27
  • 8
    To keep the balance in the universe, I'll argue that \vec{v}_1 looks better, because the arrow is centered on the variable and does not appear shifted. Commented Feb 16, 2012 at 20:31
  • I think it should be consistent regardless of how many indexes there are and in the case in which you have many, the marks being over the symbol only would be the only way to go. But in this particular case, it should just come down to taste. Commented Feb 16, 2012 at 20:40
  • 3
    @DavidCarlisle: For vector components, you wouldn't write the vector arrow (vector arrows work like Perl 5 sigils in that respect).
    – celtschk
    Commented Feb 16, 2012 at 20:43
  • Do as you prefer, but definitely don't use both in the same document with different meanings. Commented Feb 17, 2012 at 8:20

2 Answers 2


I don't believe there's a standard recommendation to answer your question, but it's probably not just a matter of personal preference either. A main criterion for good writing -- in any field, not just in mathematics! -- is the avoidance of all (unnecessary) ambiguity. One approach to keeping ambiguity low is to make sure that all symbols and notational conventions are explained at the outset. For instance, if you wrote something like

Let $\{\vec{v}_1,\vec{v}_2,\dots,\vec{v}_n\}$ denote a set of
$n$ elements of some vector space $V$.

in your paper, it should be clear to all readers that each \vec{v}_i, i=1,\dots,n, is a vector and that the subscript i merely serves to distinguish among the $n$ vectors. No further clarity would be gained, in my view, if the arrow symbol were shifted to the right to make it straddle both the v glyph and the subscript.

In contrast, suppose that the vector space V happens to be R^n and \vec{v} is some n-tuple. Now, there might be some ambiguity as to whether \vec{v}_i denotes the i-th element of v (i.e., a scalar) or the i-th n-tuple out of some set of n-tuples. If you need to refer to both types of variables in your paper, you might achieve a slight improvement in clarity by shifting the vector arrow to the right whenever you want to emphasize that you're dealing with an n-tuple rather than with a scalar.

To be sure, my recommendation in the second case would be to find a different notational solution altogether, in order to avoid any possible ambiguity. For instance, I might write v_i to denote the scalar quantity, i.e., I'd leave off the arrow entirely. I think that's much more direct and doesn't rely on your readers being alert enough to figure out on their own the meaning of a right-shifted arrow.

  • Thanks for the response - and that is useful information, but in this case the meanings of the symbols have been clearly explained (they are vectors in R^3) and there's no ambiguity as to whether the symbols refer to components or anything else. It really is purely a typography question.
    – David Z
    Commented Feb 17, 2012 at 4:37

use the package esvect for really nice looking vectors:

Let $\{\vv{v_1},\vv{v_2}, \dots,\vv{v_n}\}$ denote a set of
$n$ elements of some vector space $V$.

Let $\{\vv{v}_{\!1},\vv{v}_{\!2},\dots,\vv{v}_{\!n}\}$ denote a set of
$n$ elements of some vector space $V$.

enter image description here

  • Thanks, but this doesn't actually address my question. Your examples don't make any comparative distinction between the two cases I'm trying to compare.
    – David Z
    Commented Feb 24, 2012 at 7:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .