Tuesday, September 21, 2010

More on contextually mediated behavior from an unusual source.

The book, Working with Families of Young Children with Special Needs (referenced in the previous posting), talks a great deal about Contextually Mediated Behavior. While there is nothing new to this concept, (it is what we have desired in plans, goals, and objectives, for years) this book adds some additional clarification to the issue. The article, as you will see, is from an unusual source. The focus is on Artificial Intelligence; however, it provides some further clarification and additional reasoning to argue the importance of providing the context for the behavior, if we are to expect the behavior to occur in a natural environment.


“The context in which an intelligent agent operates profoundly affects how it behaves. Not only is this intuitive, it has been shown to be the case by psychological and sociological studies.”

The above statement is profoundly logical and has been well researched and yet it is often missed in writing plans or objectives for either children or adults with developmental disabilities and/or mental health issues. If you are going to teach new behaviors, it simply must be, at least in the end, within the natural environment where you want the behavior to occur. While you may start in a contrived setting, you must move to the natural environment within the natural routines as quickly as is possible and safe. This natural environment must include the individuals who are typically in that environment with the systematic reduction of any individuals who are not natural to the environment. For example: if the behavior you are attempting to change typically occurs in the home; while the therapist may, if absolutely necessary, start in a clinical setting, the therapist must move into the home with the child (or adult client) and family or other persons naturally living in that environment. The interventions must be taught to the parents and/or other individuals naturally in that environment and the therapist must systematically and gradually discharge him or herself from that environment while empowering the parent and/or other individuals natural to the environment to assume the responsibilities of the intervention with the therapist returning on occasion as required on a consultation basis.

“The result is that these AI applications cannot capitalize on knowing what context they are in and how to behave in that context. To the extent they do so at all, they are forced to do situation assessment, without any clear notion of what the space of possible situations (contexts) might be. … Instead, it must waste effort constantly deciding if behavioral knowledge is appropriate for the situation (e.g., by checking antecedent clauses of potential rules, goals/preconditions of potential operators, etc.) An application cannot easily learn important information about how to behave in a context, since it doesn’t have any clear notion about what it means to be in the context.”

While an individual may have some or a great deal of understanding of what it means to be in their natural context, if you do not teach, and naturally reinforce, the new replacement behaviors, they may never learn the new skill in practical way as to make it useful and contextually mediated within the natural routines of their natural environment.

http://www.aaai.org/Papers/Workshops/1999/WS-99-14/WS99-14-017.pdf

No comments: