CS 100 - Week 9 Lecture 2 - 10-19-12

MORE fallacies of insufficient evidence (Chapter 6 continued)

loaded question
----------------
*   a LOADED QUESTION contains an UNFAIR or QUESTIONABLE
    assumption;

    Have you stopped cheating on exams?
    Did you enjoy spoiling dinner for everyone else?
    Have you finally stopped flirting with so-and-so?
       (stated by your significant other)

*   the point is,
    there are REALLY multiple questions embedded
    in these,
    such that applying a SINGLE yes or no answer
        to ALL of them likely results in the answerer
	saying something they don't mean to;

    for the 1st example above, really TWO
    questions:
    1. Did you cheat on exams in the past?
    2. IF you did, have you stopped now?
    ...can't reasonably apply a single yes or no
       to BOTH of these;

*   (how to answer these?
    DON'T let the question form dictate your answer --
    DON'T preface the answer with yes or no --
    instead, indicate whether you agree or disagree
        with each implied characerization in succession;

QUESTIONABLE CAUSE
------------------
*   when an arguer claims, without sufficient evidence,
    that one thing is the cause of something else

*   three common varieties of this fallacy:
    *   the post hoc fallacy
    *   the mere correlation fallacy
    *   the oversimplified cause fallacy

*   post-hoc variety of questionable cause fallacy:

    committed when the arguer assumes, without
    adequate evidence, that because A occurred BEFORE
    B, A must have caused B.

    I drank tea, then my cold went away -- drinking
    the tea must have caused my cold to be cured!

    That person moved to town, then my cow died --
    that person hexed my cow to death!

*   the mere correlation variety of the questionable
    cause fallacy
   
    committed when the arguer assumes, without sufficient
    evidence, that because A and B REGULARLY occur 
    TOGETHER, then A must be the cause of B or vice versa.

    Whenever I party all night and eat eggs for breakfast,
    I flunk an exam I take that day.
    Eating eggs for breakfast must cause me to flunk 
    exams.

    ...the problem is the mistaken assumption that
    because two events occur regularly together --
    because they are regularly correlated (occur together) --
    that there MUST be a cause-and-effect relationship
    between them;

         ************************************
    BUT: CORRELATION DOES NOT IMPLY CAUSATION
         ************************************

*   the oversimplified cause variety of the questionable
    cause fallacy:
    
    (possibly the most common form of the questionable
    cause fallacy)

    committed when we assume, without adequate evidence,
    that A is the SOLE cause of B, when, in fact, there
    are SEVERAL causes of B

    Violent crime has declined steadily in recent years.
    Obviously, tougher imprisonment policies are working.

    ...this oversimplifies by ignoring other causes that
       have likely contributed to the decline --
       new policing strategies, change in population,
       etc.

    SAT scores have fallen sharply since the 1960's.
    Clearly, students are watching too much TV.

    ...again, an oversimplification, even if this might
       be one of several possible causes; more students
       take now than in the past, etc.

HASTY GENERALIZATION
--------------------
*   reminder: the term generalization means a statement
    that all or most things of a certain kind have a
    certain quality or characteristic

    all eneralds are green,
    most dogs are not dangerous,
    etc.

*   the fallacy of HASTY generalization occurs when we
    draw a GENERAL conclusion from a sample that is
    BIASED or TOO SMALL

    (biased, here, means a sample that is not
    REPRESENTATIVE of the target/claimed population as
    a whole)

    Do most Americans believe X? We asked >10,000 college
    students across the US, and less than 40% believed X.
    The conclusion is obvious: Most Americans do not
    believe X.

    ...but college students are not necessarily 
    representative of the entire American population;

    I've hired 3 people-of-X in the past 6 months.
    All 3 were [undesirable in some way]. So, all
    people-of-X must be [undesirable in some way].

    ...that's too small a sample;

SLIPPERY SLOPE
---------------
*   committed when we claim, without sufficient evidence,
    that a seemingly harmless action, if taken,
    will lead to a disastrous outcome

    *   We can't allow A, because A will lead to B,
        and B will lead to C, and we sure as heck
	don't want C!

WEAK ANALOGY
------------
*   occurs (as a fallacy) when an arguer compares two or
    more things that aren't really comparable in
    relevant respects (relevant to the conclusion, that is)

    lettuce is leafy and green and great on a burger!
    poison ivy is also leafy and green!
    ...so it also must be great on a burger!

    3 common patterns:
    *   2 things with several identified similarities
        (as above)
    *   several things but only 1 or 2 identied similarities
    *   simply to assert, WITHOUT further elaboration,
        that 2 cases are relevantly similar
...more on this on Tuesday!