Mad Social Scientist

I was reading up on spatial statistics for a friend today, and I finally understand why spatial correlation is not the "same" as time series correlation (or even the same as 2d time series): There is an arrow of time, but there is no arrow of space. That is, time always flows forwards. But spatial effects can mixed back and forth and every-which-way. In studying time series, the present can depend on the past, even every past period, but it can never depend on the future, nor can a chain of temporal causation lead back around to the present through other time periods. But in spatial statistics, there's no reason the spatial correlation can't lead anywhere, including back to "square one".

In the (extremely off) chance that someone is reading this who studies spatial statistics, does anyone know of a good article on the properties of spatial models with space-time lags. E.g., every step away in space is combine with a step back in time, so you never return to the same "place"?