- Email: [email protected]
The role of predictive accuracy in serial autoshaping with pigeons
Behavioural Processes, 10 (1985) 195-200 0 1985 Elsevier Science Publishers B.V.
195
SELECTED ABSTRACTS OF PAPERS CONFERENCE OF THE EXPERIMENTAL
PRESENTED AT THE ANNUAL ANALYSIS OF BEHAVIOUR GROUP
University of Sussex, U.K. 9-11 April 1984
THE ROLE OF PREDICTIVE
ACCURACY
IN SERIAL AUTOSHAPING
WITH PIGEONS
L. COLLINS Department
of Psychology,
University
reported
In a series of experiments, responding
by pigeons
serial compound schedule),
a low response
trials. schedule
followed
the first element
(A) of a
by B and food (AB+A'
by B and food (AB+ schedule).
some possible
explanations
our previous
for
findings
and
during A, in the AB+A' schedule,
was
of A with food. This was demonstrated by +/O schedule, in which A was
rate during A, in an AB by B which
itself was followed
by food on 50% of the
The second experiment revealed that, after training in the AB+A' +/o rather than the AB schedule, A is a more effective reinforcer for
higher-order
conditioning.
that A, in the former strength
pairing
CFl 1XL
et al (1983), autoshaped
followed
replicated
that the high rate of responding
consistently
during
followed
examines
The first experiment
not due to the intermittent finding
by Collins
was found to be higher
series of experiments
these results.
P.O. Box 78, Cardiff
(AB) when it was intermittently
than when it was consistently
The present
revealed
College,
This result was taken to confirm
schedule,
than in the latter.
procedure,
in which subjects
corresponded sequences
supports
a higher
level of associative
The final experiment received
the prediction
was a within-subjects
four sequences
of events that
to the two schedules of the previous experiments. + , AB', (which together represented the AB+"
were AB
previously
studied),
As would be expected were higher simultaneous C and A. responding
CD+ and Co (which together
mirrored
on the basis of the preceding
during C than A. choice procedure
the AB+A' sequence).
results,
response
rates
In the second stage of this experiment, was employed,
It was found that the stimulus during conditioning (A).
These event schedule
alternative
stimulus
alternative
index of associative
between
which sustained
(C), was reliably
This choice measure strength.
a
the two first elements,
chosen
the higher
rate of
in preference
was considered
as an
to the
196
The results of these three experiments are discussed in terms of the
Pearce-Hall (1980) model of conditioning, the first CS is seen as an important
1.
Collins,
Pearce,
the predictive
accuracy
of its associative
L., Young, D.B., Davies, K. and Pearce,
exp. Psychol.,
2.
whereby
determinant
J.M.
(1983)
of
strength
Q. Jl.
358, 275-290.
J.M. and Hall, G.
(1980)
Psychol.
Rev., 87, 532-552
Conversation Analysis and the Empirical Study of Verbal Behaviour.
U.T. PLACE Department of Philosophy, University of Leeds, Leeds LS2 9JT.
From the standpoint of the experimental analysis of behaviour, B. F. Skinner's Verbal Behavior (1957) has been criticised for its failure to generate a significant ongoing programme of empirical research. The critique of Skinner's analysis (Place, 1981) suggests that this is due (a) to his failure to draw an effective distinction between words and sentences and (b) to the use of a taxonomy based on the antecedents rather then the consequences of verbal behaviour. The mand, however, is one concept introduced in Verbal Behavior which not only has to be interpreted as the utterance of a sentence, but is also defined in terms of its characteristic effect on the behaviour of the listener. In Chapter 3 of Verbal Behavior 'a mand' is defined as a verbal operant whose emission by the speaker is reinforced by the listener's response in emitting the behaviour specified in the first speaker's mand. But since the listener's response cannot be relied upon, unless that too is reinforced, it follows that the mend is effectively defined in terms of its position in a three part verbal transaction consisting of (1) the speaker's mend, (2) the listener's response and (3) the first speaker's reinforcement of the listener's response. On pages 38-39 of Verbal Behavior three examples of such three part verbal transactions are presented in diagrammatic form. Such three part verbal transactions are examples of what would be described in the terminology of Conversation Analysis as 'a three turn sequence' (Levinson 1983) consisting of an 'adjacency pair' (Schegloff
195
SELECTED ABSTRACTS OF PAPERS CONFERENCE OF THE EXPERIMENTAL
PRESENTED AT THE ANNUAL ANALYSIS OF BEHAVIOUR GROUP
University of Sussex, U.K. 9-11 April 1984
THE ROLE OF PREDICTIVE
ACCURACY
IN SERIAL AUTOSHAPING
WITH PIGEONS
L. COLLINS Department
of Psychology,
University
reported
In a series of experiments, responding
by pigeons
serial compound schedule),
a low response
trials. schedule
followed
the first element
(A) of a
by B and food (AB+A'
by B and food (AB+ schedule).
some possible
explanations
our previous
for
findings
and
during A, in the AB+A' schedule,
was
of A with food. This was demonstrated by +/O schedule, in which A was
rate during A, in an AB by B which
itself was followed
by food on 50% of the
The second experiment revealed that, after training in the AB+A' +/o rather than the AB schedule, A is a more effective reinforcer for
higher-order
conditioning.
that A, in the former strength
pairing
CFl 1XL
et al (1983), autoshaped
followed
replicated
that the high rate of responding
consistently
during
followed
examines
The first experiment
not due to the intermittent finding
by Collins
was found to be higher
series of experiments
these results.
P.O. Box 78, Cardiff
(AB) when it was intermittently
than when it was consistently
The present
revealed
College,
This result was taken to confirm
schedule,
than in the latter.
procedure,
in which subjects
corresponded sequences
supports
a higher
level of associative
The final experiment received
the prediction
was a within-subjects
four sequences
of events that
to the two schedules of the previous experiments. + , AB', (which together represented the AB+"
were AB
previously
studied),
As would be expected were higher simultaneous C and A. responding
CD+ and Co (which together
mirrored
on the basis of the preceding
during C than A. choice procedure
the AB+A' sequence).
results,
response
rates
In the second stage of this experiment, was employed,
It was found that the stimulus during conditioning (A).
These event schedule
alternative
stimulus
alternative
index of associative
between
which sustained
(C), was reliably
This choice measure strength.
a
the two first elements,
chosen
the higher
rate of
in preference
was considered
as an
to the
196
The results of these three experiments are discussed in terms of the
Pearce-Hall (1980) model of conditioning, the first CS is seen as an important
1.
Collins,
Pearce,
the predictive
accuracy
of its associative
L., Young, D.B., Davies, K. and Pearce,
exp. Psychol.,
2.
whereby
determinant
J.M.
(1983)
of
strength
Q. Jl.
358, 275-290.
J.M. and Hall, G.
(1980)
Psychol.
Rev., 87, 532-552
Conversation Analysis and the Empirical Study of Verbal Behaviour.
U.T. PLACE Department of Philosophy, University of Leeds, Leeds LS2 9JT.
From the standpoint of the experimental analysis of behaviour, B. F. Skinner's Verbal Behavior (1957) has been criticised for its failure to generate a significant ongoing programme of empirical research. The critique of Skinner's analysis (Place, 1981) suggests that this is due (a) to his failure to draw an effective distinction between words and sentences and (b) to the use of a taxonomy based on the antecedents rather then the consequences of verbal behaviour. The mand, however, is one concept introduced in Verbal Behavior which not only has to be interpreted as the utterance of a sentence, but is also defined in terms of its characteristic effect on the behaviour of the listener. In Chapter 3 of Verbal Behavior 'a mand' is defined as a verbal operant whose emission by the speaker is reinforced by the listener's response in emitting the behaviour specified in the first speaker's mand. But since the listener's response cannot be relied upon, unless that too is reinforced, it follows that the mend is effectively defined in terms of its position in a three part verbal transaction consisting of (1) the speaker's mend, (2) the listener's response and (3) the first speaker's reinforcement of the listener's response. On pages 38-39 of Verbal Behavior three examples of such three part verbal transactions are presented in diagrammatic form. Such three part verbal transactions are examples of what would be described in the terminology of Conversation Analysis as 'a three turn sequence' (Levinson 1983) consisting of an 'adjacency pair' (Schegloff