Re: [tied] areal and genetic groupings

Piotr wrote:
<<[Areal grouping] refers to a cluster of languages that are or used to be
close to one another geographically and which show common _diffused_ features
(as opposed to shared innovations inherited from a common ancestor.>>

Does this mean that languages not "geographically close" but nevertheless in
contact cannot be grouped together as areal?
For example, if Yankee traders were in contact with Spanish speakers on a
particular West Indies island, would it be incorrect to classify them as
areally grouped even if New England and the West Indies Spanish showed common
diffused features? Or is the mere occasional presence of Yankee traders in
the West Indies considered "geographical closeness"?

I'm always surprised by this geographical thing. Historically, the majority
of speakers of two languages don't need to be concentrated adjacent or
"close" to one another to have an influence on one another. E.g., the
British and India. In prehistory, evidence of long range contact might
disappear, but prolonged mutual influence may have occurred over well-used
trade routes.

I realize that linguists often use "contact" with a different meaning. But
wouldn't contact rather than geographical closeness be the real criterion?

Steve Long