'It is generally not convenient to shape skillful behavior with a negative reinforcer because it elicits many respondents which may conflict, be incompatible with the behavior to be shaped' (Holland & Skinner, The Analysis of Behavior, McGraw-Hill 1961, p. 219 § 32-37).
'[T]he greater the severity of the punishment (after some minimum intensity is surpassed), the lower the response rate' (Ibid., p. 269, § 40-39).
'[P]unishment does not actually eliminate behavior from a repertoire, and its temporary achievement is obtained at tremendous cost in reducing the over-all efficiency and happiness of the group' (Skinner, Science and Human Behavior, Macmillan 1953, p. 190.
Piper: Skinner thought punishment was effective for certain purposes.
Let's see the reference.
Let's start with the 'Analysis of Behaviour' by Skinner and Holland. Now in this work Skinner et al say that punishment cannot be defined in terms of eliciting a particular effect. The effect of a punishment they say can only be determined via experiement. However they do offer as an example of punishment the following:
"A child reaches out and touches a flame. The child's behaviour was punished by a natural contingency" (Set 37).
Are you going to deny the efficacy of that? Seems pretty obvious to me that punishment was effective in such a situation in stopping a child touching fire (and consequently suffering 'pain').
Seems to me you are taking 'effective' to mean 'effective shaping behaviour', whereas i take effective to mean 'effective in producing a particular (or 'desired') response'.
But lets see what else B.F. has in mind.
In 'About Behaviourism', Skinner describes the effect of punishment as follows:
"A punished person remains 'inclined' to behave in a punishable way, but he avoids punishment by doing something else instead, possibly nothing more than stubbornly doing nothing" (at 62).
Skinner then enumerates possible feelings that punishment can produce in the victim: shame, sin, guilt (He attaches these to particular institutions). Such feelings Skinner says arise when 'he behaves or tends to behave in punishable ways'. Again punishment is effective to produce a particular response.
So punishment IS effective in eliciting particular responses. (presumably particular responses to a punishment can be determined via experiment and then repeated a nauseum).
Can punishment be used to shape behaviour? (effective in your sense). Well, i think Skinner would have to say 'yes, in some circumstances' to this.
In 'A New Preface to Beyond Freedom and Dignity' (in 'Recent Issues in the Analysis of Behaviour), Skinner specifically advocates punishment as a means of stopping people from smoking. He says:
"punish smoking with criticism, complaints, restrictions on where one may smoke and heavy taxes on cigarettes" (at 118)
But here skinner is advocating punishment in addition to reinforcement of not smoking and commendations for those who do stop. So it would seem that he thinks punishment in combination with positive reinforcers CAN be used to shape behaviour.
But i think we can take this further. As i said earlier he thought punishment was effective to produce feelings of 'sin'. Now, it seems to me that a person who has been punished by a religous organisation can feel 'sin' many years (in not indefinitely) after the original punishment(s) (Ever heard of the Catholic Church Barry?). Without any further contact with the organisation, that is without further instances of punishment.
So again if somebody is looking to piss somebody off or make them act in certain ways then punishment CAN according to Skinner DO these things, albeit he thinks that it is not a particulary efficient way of shaping behaviour (i.e. controlling people).
BTW Barry, I think reinforcement is a nonsense deigned to attempt to make psychology deterministic and thus scientific. Why does a pigeon raise its head in a Skinner box to get food? Is it because such behaviour has been reinforced? Or is it simply because the pigeon EXPECTS IT? The latter i suspect.
So you can stop with the 'reinforcement schedule' crap.