First Person
Approaches
Plan for today
Assess new neurophenomenology
movement in consciousness science
Learning from historical critiques
Signal Detection Theory
How to measure
consciousness?
From first class –
You ask people, right?
Subjective measures seem obvious
Behaviours are just behaviours, but
Phenomenological/
Introspective approaches
Reflect on the contents of their experience
Training makes you better able to observe
and describe the contents and structure of experience
Maybe analogous to other scientific training?
BUT problems arise when subjective reports
are seen as the only proper way to study consciousness
Ramsoy and Overgaard
History of unconscious perception research
is way more complicated than they suggest….
Sidis etc 1900 using subjective measures,
problems with subjective measures and introspectionism early 20th Century, Signal
Detection Theory 1950s, resurgence of
subjective measures 1980s, Signal Detection Theory again, and now
Subliminal perception
1. Use some way of ensuring that subjects are not conscious of stimuli
2. Then measure whether stimuli still influence behaviours
Sometimes use reports for 1, forced choice tasks or priming for 2
Sometimes use forced choice tasks for 1 and 2
Sometimes use anaesthesia for 1
Subliminal perception
Also found in: Prosopagnosia (inability to recognise faces) – through heightened GSR for familiar faces
Subliminal perception
All find some degree of processing of a
stimulus in the absence of consciousness of the stimulus
=> Unconscious perception exists!
But are we really sure that this is
unconscious perception, not just low level conscious perception?
Perceptual Awareness Scale
(PAS)
Consciousness is graded, but many
response scales are not
Often dichotomous (conscious or not)
Response scales also measure the wrong
things
Difference between performing a task, or
being confident about a response, and
consciously perceiving something more or less clearly
PAS
Response scale developed using
feedback from subjects
Supposed to be intuitive and easy to use
Supposed to better reflect grades of
consciousness than any other scales
Supposed to uncover cases of low level
PAS
Subjects shown a stimulus at some
variable stimulus duration
Subjects first told to report what they
think is presented in terms of shape, colour and position (even if they think nothing is there)
Then subjects asked to report their
Problems with the PAS (1)
Signal Detection Theory has been used fordecades to uncover lower grades of perception
SDT uses graded response scales
More exhaustive than the PAS
Used to routinely identify cases of low level conscious perception
PAS is not new, and objective measures are now
PAS Problems (3)
Using the PAS requires training
Introspective training has a bad history in psychology (as we saw earlier)
Responses are always biased, training
gives you different ones
Dodge 1912
“No psychological scheme has been too
absurd to be supported by introspection. It shows fashions like hysteria and the
delusions of the insane. Even the
fundamental categories of consciousness change with the years, while new and
previously totally unsuspected facts may be readily introspected as soon as there is
PAS Problems (4)
The distinction between introspective
measures and ‘behavioural’ measures supports a kind of dualism
Introspection is seen as the only way to
understand consciousness
But then how can we ever test its
reliability?
Dodge
“On these considerations the
methodological dogma that all mental reality is subjectively observable and conversely that the subjectively
Dodge on integration
“In view of all the evidence I believe that
introspection is only one of the indicators of mental reality. It is a real and important
indicator of peculiar value in special fields but it is only one of many. Equally real I
believe is every pathological or neurological fact, every result of practice, training or
Your questions (1)
Could functional imaging techniques provide some
external validity by correlating neural events with introspective reports?
Aren't many of the criticisms applicable to a variety of
other disciplines? The reliability of a method is always a problem that has to be addressed.
Standardization of scientific instruments requires
quantification and the reference to a common measure. Is it possible to do so with mental reality?
What will happen in case of a switch from pure
Your questions (2)
If I train to be a introspective psychologist, I change the way I experience things, therefore the results of my introspections will not be applicable to untrained people. This seems not to happen for natural sciences where subject and object of study are two separated things and my knowledge about the object of my study change my theory but not the object itself. Is it a problem that is possible to overcome?
How can the concepts of 'conscious' and 'unconscious' be defined without any reference to subjective experience? What does the possibility of gnosanopsia (awareness without discrimination) and agnosopisa (discrimination without awareness) tell us about
objective measures of 'conscious' and 'unconscious'? Is a graded categorization of conscious and unconscious
Integration
Integrating and comparing subjective
and objective measures seems very sensible
BUT what do you say when someone
insists that they see ‘all the letters’ but can only identify 4 of them
OR when someone says they cannot see
Integration
Is there a principled way of dealing with inconsistencies between reports and
behaviours?
Which do you trust as the more reliable measure, when and why?
Can people just be plain wrong in what they report?