>His coding of languages into weak/strong FTR is based on the
>predominant form of future marking used in the language. So even
>though a language might have overt future tense, speakers of that
>language might show a strong preference for referring to the future
>with phrases like "I am eating tomorrow" rather than "I will eat

But in the video he reasoned with information that *has* to be expressed in
a particular language vs. information which *may* be expressed. In Chinese
you cannot say 'uncle' without expressing whether it's an uncle on the
mother's or father's side and so on. But 'I am eating tomorrow' and 'I will
eat tomorrow' convey exactly the same amount of information because of the
temporal adverb.

From the statistical point of view, you'd need to start off with a random
sample of language (preferably unrelated), but the diagram given at 9:04 in
the video is clearly based on the available savings data which happens to be
for European countries mostly.

>Now wait a minute.  I'm also skeptical of this study, and I'd like to see
>more work done, but if we rejected every hypothesis that required us to
>control for other variables (the whole "all else being equal" bit), then a
>whole lot of sociological and economic research simply could not be done.

Mhm, maybe they shouldn't have been done then...