On Thu, Jul 15, 2010 at 10:04 AM, Jörg Rhiemeier <[log in to unmask]> wrote:

> I hardly see any way things like phonology could influence a speaker's
> way of thinking, though.

It seems to me unlikely but not impossible; I suggested a possible
mechanism here:

On Wed, Jul 14, 2010 at 10:04 AM, Jim Henry <[log in to unmask]> wrote:
> E.g., maybe the
> phonaethetics of one conlang use the same phonemes and clusters
> thereof in words for bad things and concepts and in words for breeds
> of dog, while another conlang uses the same groups of phonemes in
> words for good concepts and things and in words for breeds of cat...

Conceivably that sort of phonaesethetic or phonosemantic association
could arise in a natlang naturally, either through a series of
coincidences, or more likely through a culture's biases at one stage
of history influencing the phonology of its language and then
continuing to influence a later generation's thinking unconsciously
even when the overt cultural markers for said bias have weakened or
disappeared over time. But in a natlang setting, that situation
(phonosemantic bias remaining strong when cultural bias has all but
disappeared) seems unlikely enough that I don't think you could
usefully test for this effect without using a pair of conlangs.

Ideally, for triple-blind purposes, the people who devise the conlangs
should be different from those who teach the conlangs to the
students/subjects and those who test them for bias in the conlang and
their native language; we don't want the teachers or experimenters to
know who's learned the experimental drug language and who's learned
the placebo language. But I'm not sure that's feasible or worth the
trouble; probably anybody smart enough to learn a conlang well enough
to teach it can figure out what phonosemantic biases the conlanger
built into it, at least if they know anything about the purposes of
the experiment.

--
Jim Henry
http://www.pobox.com/~jimhenry/