Co-ordinated activation of both a particular category and the instances relation will select the locations that belong to that category. Commands can select locations belonging to multiple categories in ways that correspond to "and" and "or" connectives of Boolean algebra.
The methods are practical. There is a
"tricky step" in the adjacent
diagram. The author finds it convincing;
but, even if you don't, you can still use
instance relations (including the trick)
to organize information in the examples:
1. representation of syntax
2. library research system
The simplified notation used for instances can be generalized to notate other relations. This generalized notation leads to the introduction of generalized relations discussed in connection with device_2.
The legal system's central element is the case, a decision of a court of appeal; cases are published in series of volumes organized according to the court jurisdiction and the date of issue of the decision.
In general, each decision is a bloc of text that decomposes into blocs of text:
Legal publishers assign each holding to one or more classes; and the classes are organized in a digest with a hierarchical structure of topics.
Words and phrases, keywords, appear in the blocs of text. Some keywords, like "California Civil Code §1649" refer to statutes, rules of law enacted by the legislature. Statutes are organized into codes, compendia of statutes with their own hierarchical structures of topics.
Although a thesaurus is not commonly
used in legal research, one could easily be constructed
and added to the system.
It is possible to conduct research with this system through mass action techniques more powerful than the item-by-item methods conventionally used. The system could be installed on a conventional mainframe-sized computer, although beyond the reach of personal computers presently available.
For example, a researcher could begin with a set of keywords (perhaps expanded with the thesaurus) and obtain all holdings that include 3 or more of the keywords. From the set of holdings so obtained, a new set could be constructed based on citation clusters. Digest classes obtained from the clusters could then be used to assemble a set of "leading cases" obtained through additional steps on the citation relations. Leading cases could be ranked according to the number of times each is cited in other cases derived from the digest classes. Possibly significant statutes could also be identified and ranked according to frequency of appearance.