Sheehan and Roles
Sheehan1 postulates that you stutter to the role, not to the person temporarily occupying the role. These roles are attached to positions, to statuses, and not to the person who is occupying the particular role at a point in time. In a speaking scenario, there are two roles: that of the speaker and that of the listener. Stuttering is therefore defined as a role specific self presentation disorder.
This is too simplistic in my opinion. What about speaking in front of an audience of individuals; do they make up one role en masse or are there multifarious roles corresponding to each member of the audience. In the later case, is the situation defined by the relationship between the speaker role and each listener role individually? What about the interaction between the individual listener roles?
My experience, which is personal and individual of course, is that my stuttering is not specifically defined by a fear of authority figures or specific situations. My disfluency can occur with anyone I am talking to, even close friends and family. I know I am not alone in having such broad stimuli to my disfluency. Wendell Johnson describes a similar scenario is his book, Because I Stutter (see links). Therefore while there may be some people to whom the role theory of stuttering may be applicable, it is certainly not universal, and should not be posited as a definitive theory.
In my own experience, I may be an example of an anti-role theory of stuttering, as I am often quite fluent when speaking to people in roles of high authority. This may be due somewhat to the adrenalin factor that may arise in such encounters. If I had to deal with these people on a more regular basis, the effect of the role may diminish and my speech could possibly go back to my average standard discourse fluency.
Sheehan gives an interesting example of role specificity taken from the research of Kinsey. Apparently people don't stutter when talking to a person they are sexually responsive to. Ipso facto, if a stutterer only spoke to people he or she were sexually responsive to they wouldn't ever stutter. In real life, this might be quite difficult to arrange! Again, from personal experience, this is not particularly true.
Footnotes: