[Date Prev][Date Next][Thread Prev][Thread Next]
[Date Index] [Thread Index] [Author Index]
Return to main CEDA-L Archive Page

bridge one

Ken wrote an excellent response.  as usual, i believe there's more to
the tale and that the story ends differently.  but before that, three
items:

a)  i hope (and expect) the Hoe/Broda-Bahm fiat paper is excellent: good
thinking on that subject could be worth much.

b)  take Ken's warning, but there is no imminent danger.  i'll keep
plan-plan out of the discussion.  that discussion becomes relevant if
and only if all of the plan-focus bridges get burned.

the outstanding objection (Ken's Objection, i like to call it) to the
plan-plan way of doing things is that it abandons rational evaluation of
a single policy.  of course, if there is no such thing as rational
evaluation of a single policy (i.e., plan focus means we are fooling
ourselves into thinking that we can rationally evaluate action - whether
"X should do Y")... 

of course, there are many bridges left to burn before we get to THAT
place.  most significantly, "same-agent competitive counterplans" seems
to me to be built of tougher stuff.

c)  i will grant one of Ken's arguments from 2 years ago which he makes
again.  there was and is wisdom in his view that "All evaluation
requires a comparative base - to call something good/bad/ugly is to
compare it.  Evaluation of a prospective policy change also requires a
comparison."  

the broader project can be understood fairly simply as:

is there a way to specify those and only those "comparative bases"
against which to evaluate a posited action (by the posited actor) such
that our evaluations will result in a rational decision about whether or
not the posited action ought to be taken (by the posited actor)?  

and if such can be done, then plan focus will have in its defense the
force of Ken's Objection and we will have, incidentally, figured out how
to do policy evaluation well.  and if it is shown that such cannot be
done, then welcome to a postmodern debate world.

now, the first bridge is the one Ken has, over the past 2 years, moved
to defend.  it is also the bridge Dallas Perkins may wish to defend.  it
is, alas, a structure built of kindling.

A) the definition of "should".

well, i offer a simple counterdefinition from the Hypertext Webster's
(http://gs213.sp.cs.cmu.edu/prog/webster?should):

"used in auxiliary function to express obligation, propriety, or
expediency. {'tis commanded I should do so - Shakespeare}{this is as it
should be - H.L. Savage}{you should brush your teeth after each meal}."

now Ken offers only one definition of "should" and it is Dallas'
definition from an article looking to make the "aff plan focus, no fiat,
no counterplans" argument.  while i would defer to the esteemed one in
some matters, he ain't no lexicographer.  

this issue is important only because the sole defense of the commonsense
framework proffered by Ken is that "should MEANS ceteris paribus".  if
should does not mean THAT, if should means what we and Websters think
that it means, then the bridge continues to burn brightly.

in short, it seems to me that we believe what Webster says "should"
means, that "should" means "oughta".  
as in "you should brush your teeth after each meal" 
MEANS "you oughta brush your teeth after each meal."

1) not closing open questions

if it is the case that "should" just means "oughta", then a subsequent
inquiry regarding the appropriate means to evaluate "should" claims is
necessary.  in short, we are left to ask, "how should we decide whether
or not 'should'?"

on the other hand, if "should" MEANS "incremental should" or
"counterfactual should" then the question of how we are to verify
"should" claims is decided definitionally as "use the ceteris paribus
assumption" or "use a counterfactual analysis."

the problem with these "closed" definitions of "should" is that they
make arbitrary commitments about how we ought to verify "should" claims.

one could, for example, understand "should" AS meaning "Ten Commandments
should" or "affirmative always wins should" wherein we verify "should"
claims by seeing whether or not the Bible states the "should" claim or
by just agreeing that "should" claims are true.  perhaps "plan-plan
should"?

it seems to me that "should" just means "oughta" and there remains an
open and subsequent question of how to verify "should" claims.  we are
"just" making a claim as to the correct action to be taken, not also
making a claim as to the method of deciding whether or not that is the
correct action to take.  so when we say "you should brush your teeth
after each meal." we JUST mean, "the correct action for you to take is
to brush your teeth after each meal." and then we try to figure out how
to decide whether or not that claim is true.

one last argument in this section is the most basic.  tying the
definition of "should" to any particular means of verification of
"should" claims runs the risk that our means of verification is foolish. 
so suppose that the way folks make decisions is to consult 1-900-number
astrologers and the definition of "should" is closed to mean "psychic
should" and you ask yourself "should i brush my teeth after each
meal?"...

2) the OTHER should/would distinction

the trouble with the definition of "should" claims as counterfactuals
(and thus as ceteris paribus claims) move is that Lewis and subsequent
authors don't argue that "should" claims ought to be analyzed as
counterfactual claims.  they argue that WOULD is the paradigm modal
auxiliary indicating the subjunctive conditional subject to a
counterfactual analysis.  the evidence and analysis from a genius
philosopher like Lewis which argues that "should" claims ARE
counterfactual claims is worth Hoe's weight in platinum.  i looked very
hard for that card and that analysis when introducing the whole
counterfactual thing years ago.  didn't find it.  Ken hasn't either. 
just a couple of cogent sentences and i'd probably join the fire brigade
myself.

that's why i say that:

	"The analysis of "If you drink the coffee, your alertness 
	will increase." requires a different framework than "You 
	should drink the coffee."

the difference is that the CAUSAL claim uses the present tense of WOULD
while the policy claim uses SHOULD.  Lewis and others say WOULD
indicates a counterfactual claim, that is, a causal claim for which the
appropriate method of verification is counterfactual analysis.  they
don't say that about "should":  Dallas "says" that about "should",
but...

that is also why i argue that counterfactual analysis treats all actions
except the action posited as CAUSALLY entailed or implied, rather than
treating at least some actions as also subject to CHOICE.

the ceteris paribus/counterfactual analysis treats the "i should drink
the coffee" claim this way:  

"if i drink the coffee, then i WILL... compared to if i do not drink the
coffee then i WILL..." 

that is, with all actions as causally determined except the posited
action.  the opportunity cost analysis, in comparison, leaves open to
CHOICE at least some other actions, in particular the actions of the
relevant actor.

B) the commonsense framework is common sense

yes, i begin there by acknowledging this claim.  our pre-theoretical
intuitions are to evaluate claims about whether to take a posited action
by using a ceteris paribus or counterfactual analysis.  i suspect that
this thought pattern may even be wired in.  the problem with this is
that we make bad decisions about what actions to take by using a
logically invalid framework.

the "you should put your money in a savings account" and "i should drink
the coffee" examples show that.  there is no answer to them attempted.

C) natural language

i grant without reservation the Dunning and Parpal (1989) claim Ken
offers:

	"assessing the consequences of actions and events often requires
	comparing a mental simulation of the world in which an action is
	present to one in which the action is absent."

in fact i take it as uncontroversial that ANY attempt to show 
X should do Y
will necessarily involve a comparison between a world or worlds in which
X does Y 
and a world or worlds in which
X does not do Y.

the 64 billion dollar question remains unanswered (and i now suspect may
be UNANSWERABLE):  which world or worlds in which "X does Y" compared to
which world or worlds in which "X does not do Y"?

my argument is that the answer: 

"the world as it is modified only by "X does Y" compared to the world as
it is" 

is clearly incorrect.  that "mental simulation" results in terrible
decisions.  that's why it makes no sense to make ceteris paribus the
"default comparison" much less the comparison by definition or the
comparison by a "plan focus, no fiat, no counterplans" paradigm.

D) Ken's examples of why competition is problematic

thank you, but i'm not at those bridges yet...

in any case, the examples all seem to me problematic merely in virtue of
the fact that they assume a fiating of solvency rather than a fiating of
action.  so, i DO think that if Ken is evaluating "I should accept
tenure at TSU." and the status quo is "I will remain an untenured
professor at TSU." that "I should run for President." ought to be
considered.

the real cost of accepting tenure isn't remaining untenured, all else
being equal, the real cost of accepting tenure is the value of the best
ACTION which would be available to Ken otherwise but would have to be
foregone by Ken if he accepted tenure.  nothing murky there: impacts of
becoming Prez times chance that the action of running for Prez makes him
Prez compared to tenure advantages. same answer for "devote my time
attempting to cure cancer."

and i will leave my response here:  that's why opportunity cost points
to counterplans which are "same agent, competitive".  the actor is Ken
and he should have choice about his actions:  he ought not to be treated
as a puppet who is stuck being an untenured professor at TSU ala the
status quo.  at least Ken should instead be treated as having choice
over what actions he will take if he were not to accept tenure.  like i
said, that one is made of tougher stuff.

and when and if we get to the bridge that gives fiat to Ms. Broda-Bahm,
the torch has now been lit.

thanks for reading,
michael korcok
Archive created by Jonathan Stanton (jonathan@cs.jhu.edu)
Return to main CEDA-L Archive Page