# Sign convention

In physics, a sign convention is a choice of the physical significance of signs (plus or minus) for a set of quantities, in a case where the choice of sign is arbitrary. "Arbitrary" here means that the same physical system can be correctly described using different choices for the signs, as long as one set of definitions is used consistently. The choices made may differ between authors. Disagreement about sign conventions is a frequent source of confusion, frustration, misunderstandings, and even outright errors in scientific work. In general, a sign convention is a special case of a choice of coordinate system for the case of one dimension.

Sometimes, the term "sign convention" is used more broadly to include factors of i and 2π, rather than just choices of sign.

## Relativity

### Metric signature

In relativity, the metric signature could either be + − − − or − + + +. A similar dual convention is used in higher-dimensional relativistic theories. The choice of signature is given a variety of names:

+ − − −:

− + + +:

Regarding the choice of − + + + versus + − − −, a survey of some classic textbooks reveals that Misner, Thorne and Wheeler (MTW) chose − + + + while Weinberg chose + − − −[citation needed] (with the understanding that the first sign corresponds to "time"). Subsequent authors writing in particle physics have generally followed Weinberg, while authors of papers in classical gravitation and string theory have generally followed MTW (as do most Wikipedia articles related to relativistic physics). Nevertheless, the Weinberg form is consistent with Hyperbolic quaternions, a forerunner of Minkowski space.

The signature + − − − would correspond to the following metric tensor:

$\begin{pmatrix} 1 & 0 & 0 & 0 \\ 0 & -1 & 0 & 0 \\ 0 & 0 & -1 & 0 \\ 0 & 0 & 0 & -1 \end{pmatrix}$

whereas the signature − + + + would correspond to this one:

$\begin{pmatrix} -1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 1 \end{pmatrix}$

### Curvature

The Ricci tensor is defined as the contraction of the Riemann tensor. Some authors use the contraction $R_{ab} \, = R^c{}_{acb}$, whereas others use the alternative $R_{ab} \, = R^c{}_{abc}$. Due to the symmetries of the Riemann tensor, these two definitions differ by a minus sign.

In fact the second definition of the Ricci tensor is $R_{ab} \, = {R_{acb}}^c$. The sign of the Ricci tensor does not change, because the two sign conventions concern the sign of the Riemann tensor. The second definition just compensates the sign and it works together with the second definition of the Riemann tensor (see e.g. Barrett O'Neill's Semi-riemannian geometry).

## Other sign conventions

It is often considered good form to state explicitly which sign convention is to be used at the beginning of each book or article.