Although correlation and causality sound similar, they are in fact very different things, and not intrinsically linked. Variables can be correlated (positively or negatively), without one being responsible for the change in the other.
Correlations implies a relationship between the two, but this is different to causation. For example, ice cream sales and aggression may both increase during summer – but this does not mean that the more ice cream people eat the more aggressive they are likely to be. However, the increase (hopefully!) of heat could be responsible for causing both – and if a correlation were found between increased ice cream sales and the increased heat, this could be an indicator of causation, although not proof of. This is not 100%, but can be used as a clue towards where to investigate, but this is pretty much all that correlation can conclusively be used for.