Attribute listing, Crawford's work (cited in [#!creat:finke!#]), begins with a reasoner considering an object which they think can be the springboard to a new, creative artifact. Each of the object's primary attributes are examined, with the reasoner thinking about ways in which each could be manipulated. Finally, the reasoner filters out the manipulations which lead nowhere; those that remain are considered to be creative.
Morphological synthesis, Allen's technique (cited in [#!creat:finke!#]), begins in the same fashion as attribute listing. The reasoner considers both attributes and commonly accepted values for each. Then, new combinations of attributes and values are examined. The reasoner looks for unknown combinations which may prove to be creative. While this is similar to attribute listing, in morphological synthesis, the reasoner does not concentrate on changing attributes one-at-a-time; instead, interesting combinations are concentrated on. This leads to faster generation of potentially creative artifacts.
Consider a piece of luggage, an everyday mundane sort of object. To perform attribute listing on it, a reasoner would list all the primary attributes of the object and their prototypical values; in essence, the reasoner would bring to mind an ``average'' piece of luggage. It is made of leather, about two foot by two foot in size, has a soft body construction, one large compartment for storage, and is carried by means of a handle. After these major attributes are listed, the reasoner starts to consider altering them, one at a time. Some possible changes are: cloth may be substituted for leather, the dimensions may become larger or smaller, the the number of compartments may change, or wheels may be used in addition to the handle. The reasoner may have to ``clean up'' the new product in some situations; for example, if the number of compartments become two, the reasoner must decide how the two compartments will be integrated. Morphological synthesis would go about the process in much the same manner, with interesting combinations of attributes considered rather than single ones. The idea is the same as in attribute listing, the focus is just a little larger. So, the reasoner may decide to consider a large piece of luggage with lots of compartments versus a small piece of luggage with only one. As more attributes are added to the set under consideration, the number of possibilities grow.
The drawback to these approaches is that a reasoner will never develop an R-novel object, just relatively close variations of the original concept. As discussed earlier, this is due to the limitation of only considering primary attributes. As a result, E-novel artifacts are the only ones which can be created. The main advantage of morphological synthesis is the capability of attribute combinations; this decreases the time needed to produce potentially creative artifacts. Notice, for instance, that neither of the two techniques allow the introduction of new attributes. Yet, one could imagine a number of additions to a piece of luggage which would be useful. Many travelers, for example, forget to bring an alarm clock with them. One could conceivably add a small alarm/radio to a piece of luggage; since it is unlikely that a traveler would forget their luggage, an alarm clock would always be ``remembered.'' Or, a mirror could be added to the inside of one of the compartments. Another possible addition is the idea of dedicated compartments of the suitcase being used for specific items.
As mentioned earlier, one of the core tasks which my theory claims is part of the creative understanding process is the ability to dynamically create new concepts as needed. The method which is accomplishes this in my work is known as function-driven morphological synthesis or FMS. As the name implies, this is an extension of Allen's morphological synthesis. In the ISAAC model, there is no longer a restriction that only primary attributes be manipulated. Instead, any attribute may be altered. This does eliminate the benefit of the restriction; namely, the reasoner is no longer ensured of developing only ``reasonable'' concepts. That is the purpose of the ``function-driven'' portion of the technique. As Section 6.2.3 will describe, by requiring that resulting concepts share the same function as the original one being consider and by restricting the modifications through the specific knowledge organization which is used (the topic of Chapter 4), FMS is constrained to produce only useful concepts rather than bizarre ones. As a result, my extension to Allen's work overcomes the limitations while maintaining the usefulness of it.
The FMS technique permits the modifications to the suitcase which was discussed earlier, both the ones which morphological synthesis and attribute listing could produce and the ones which I described as being impossible for those techniques. Since the function of the suitcase is to act as a mobile closet/bathroom counter for a traveler, the extensions which increase its functionality will be considered. In fact, even more revolutionary modifications can be considered. It is possible to imagine, for example, endowing a suitcase with volition and a method of independent motion, thereby allowing the problem of lost luggage to be eliminated. But, this modification would never be considered by the earlier techniques on which FMS is based. Of course, one may debate the reasonableness of this modification, given that such a device could not easily exist in the current world; however, since ISAAC is a reading system, the ability to understand such a concept could prove useful; at least one fantasy novel incorporates the idea of an independently mobile suitcase for the precise reason I described ([#!story:colour!#]).