Common use of Problem Areas Clause in Contracts

Problem Areas. Phase 1 2 3 4 Total 1. Nodes with C 9.20 2.67 8.38 5.96 6.73 2. Position of C 0.80 0.67 1.80 3.25 1.92 3. Adjuncts with C 4.80 2.00 3.59 4.88 4.17 4. Unit expressions 4.00 6.67 4.79 3.52 4.38 5. Attributes 36.40 48.67 27.54 34.15 35.90 6. Settings 8.40 10.67 8.98 11.38 10.04 7. Errors 3.60 7.33 8.98 2.98 4.91 8. Other cases 32.80 21.33 35.93 33.88 31.94 9. C/T v. F 72.40 82.00 62.87 69.65 71.15 10. C v. T 22.00 11.33 33.53 28.18 24.79 Table 3 gives a list of particular theoretical areas that we have especially focused on. We have taken into account only those cases where the annotators disagreed between two values. We consider the cases where each annotator used a different value for contextual boundness to be too unclear, moreover they occur very scarcely in the data (they constitute in each phase less than one percent of all cases). The percentages in the table say how frequent individual phenomena are relative to all disagreements. Table 3: Problem areas and proportions of disagreements Individual disagreements in annotation are not considered as errors, but as potential values of contextual boundness – we never can accurately determine the contextual boundness of a node, because we cannot dispose of ample enough (complete) context for a univocal decision. Therefore, the values of contextual boundness of individual nodes are only to be taken as more or less probable. The distribution of diverging annotations of a node serve to evaluate these tendencies.