Methods of quantifying affect

From: Wesley R. Elsberry (welsberr@inia.cls.org)
Date: Thu Dec 21 2000 - 10:46:08 EST

  • Next message: Chris Cogan: "Re: Natural and Supernatural (was Chance and Selection)"

    Bertvan wrote:

    BV>TO (Wesley R. Elsberry)

     CC>No. They exist as states and processes in the real world (in
     CC>people's brains). They can even be *measured*.

     BV>Please tell me more about these scientific methods capable of
     BV>measuring hate and love in the human brain. :-)

    WRE>Research is being done to provide methods to quantify affect.
    WRE>One example that I know of is the research of W. Jackson Davis
    WRE>at the University of California, Santa Cruz, who has been
    WRE>working on quantifying affect in humans for some time now. By
    WRE>combining facial electromyography with standard subject slides
    WRE>used in psychology for self-report of affect, Davis has
    WRE>obtained good correlation between sub-threshold activation of
    WRE>facial musculature and self-report of affect. The pattern of
    WRE>activation indicates categories of affect (e.g., happy, sad,
    WRE>etc.) and the amplitude (or more precisely, the area within
    WRE>the envelope of the bursts) gives information on degree. In
    WRE>other words, Davis is able to both categorize and quantify
    WRE>affective states in humans via facial electromyography.

    BV>Hi Wesley. Possibly some scientists require electromyography
    BV>to determine whether a person looks happy, sad, etc., but most
    BV>of us have other (only partially understood) methods.

    I made no claim that any scientists "require electromyography
    to determine whether a person looks happy, sad, etc."
    The topic stated in my first sentence is what I was addressing:

     WRE>Research is being done to provide methods to quantify affect.

    This doesn't say or imply anything like BerthaJane's response.

    BV>However we read facial expressions, it is not entirely
    BV>experience, for even infants can do it. "A good correlation"
    BV>is pretty subjective.

    No, correlation is a standard method of statistical inference.

    BV>Scientific measurements such as gravity and the speed of
    BV>light, are called laws of nature because of their
    BV>consistency.

    Davis' research shows consistency as well. The precision is
    not yet as high as modern techniques for measuring g or c, but
    probably compares favorably with the precision of early
    techniques used for each of those.

    BV>All actors, and most of the rest of us, could fool
    BV>electromyography (by an act of free will).

    This is a statement made in ignorance of the technique in
    use by Davis, which examines *sub-threshold* activity.

    BV>I doubt such a "scientific measurement" would be admissible
    BV>as evidence in court.

    Well, I doubt that BerthaJane has any basis to render such a
    judgment on either legal or scientific grounds. But that's a
    secondary issue. Let me return to something more primary.
    What relevance does legal admissibility have to my topic? Let
    me post it again:

     Research is being done to provide methods to quantify affect.

    BV>"A good correlation" resembles the type of "scientific"
    BV>evidence offered for Darwinism, which is why some people
    BV>are skeptical of it.

    Correlation demonstrates the existence of a relationship.
    While someone may dismiss any correlation based solely upon
    how conveniently it fits his or her preconceptions, others are
    likely to consider the merits.

    BV>Furthermore, it was love and hate that Chris claimed could
    BV>be measured in the brain, which are much more complex than
    BV>"happy" or "sad".

    And Chris can provide details for those if he wishes.

    What I'm pointing out is that scientific inquiry is inclusive
    of affect, and there are successful approaches to putting this
    on a quantitative basis. Affect is related to the states of
    "love" and "hate", I think most would agree.

    Earlier, BerthaJane said:

    BV>Choice, free will, spontaneity, creativity, consciousness
    BV>and emotions were probably all necessary new ingredients.
    BV>They seem to distinguish life from non life. (You'll never
    BV>find any of them in a computer.)

    There are a variety of researchers who would also disagree on
    this statement. The volume, "Motivation, Emotion, and Goal
    Direction in Artificial Neural Networks" (Levine and Leven,
    editors) is available from Lawrence Erlbaum Associates, and
    will serve well as a starting point for further research.

    Wesley



    This archive was generated by hypermail 2b29 : Thu Dec 21 2000 - 09:55:14 EST