Martin Taylor 2018.05.10.12.59]
[From Erling Jorgensen (2018.05.10 1250 EDT)]
…
EJ: I would
make two suggestions, to try to keep the distinctions clean.
First, I wonder if we ought to resurrect the old CSGNet
language (from some 20+ years ago) of “controlling for” a
certain outcome. Maybe that can be the way we talk about those
environmental variables, that we are “controlling for them to
be in a certain state.” Then the term “control” by itself can
stay with the perceptions, where the internal matching to
references takes place.
I have always used "controlling for X" as a short form of
“Controlling a perception of X to be at a particular reference
value”, where the expansion of X describes both the nature of the
perception and the value of the reference. Example: “I am
controlling for the cup to be on the table” is short for “I am
controlling for my perception of the cup location to have the value
‘on the table’”. In effect, this implies that others will perceive
the cup to be on the table if my controlling works well, but it does
not imply that there exists such an environmental variable. The
variable whose perception I am controlling may well be in my
imagination, and not just because it is an illusion that I think
exists in the environment.
At least this is how I have always explained "controlling for" when
the need has seemed to arise.
Martin
Philip 5/10 17:16
What do you mean by
(1) this implies that others will perceive the cup to be on the table if my controlling works well, but it does not imply that there exists such an environmental variable.
(2) The variable whose perception I am controlling may well be in my imagination, and not just because it is an illusion that I think exists in the environment.
If you said “I want the cup to be on the table”, instead of “I am controlling for my perception of the cup location to have the value ‘on the table’”, what would be wrong?
···
On Thu, May 10, 2018 at 10:08 AM, Martin Taylor csgnet@lists.illinois.edu wrote:
Martin Taylor 2018.05.10.12.59]
[From Erling Jorgensen (2018.05.10 1250 EDT)]
…
EJ: I would
make two suggestions, to try to keep the distinctions clean.
First, I wonder if we ought to resurrect the old CSGNet
language (from some 20+ years ago) of “controlling for” a
certain outcome. Maybe that can be the way we talk about those
environmental variables, that we are “controlling for them to
be in a certain state.” Then the term “control” by itself can
stay with the perceptions, where the internal matching to
references takes place.
I have always used "controlling for X" as a short form of
“Controlling a perception of X to be at a particular reference
value”, where the expansion of X describes both the nature of the
perception and the value of the reference. Example: “I am
controlling for the cup to be on the table” is short for “I am
controlling for my perception of the cup location to have the value
‘on the table’”. In effect, this implies that others will perceive
the cup to be on the table if my controlling works well, but it does
not imply that there exists such an environmental variable. The
variable whose perception I am controlling may well be in my
imagination, and not just because it is an illusion that I think
exists in the environment.
At least this is how I have always explained "controlling for" when
the need has seemed to arise.
Martin