TABLE 3.
Comparison of random guessing expected performance vs. Kanzi’s actual performance for pairs of sentences.
| Guessing expected # | Kanzi’s performance* | Guessing expected% | Kanzi% | |
| Both correct | 5 | 14 | 25% | 70% |
| One correct and the other incorrect | 10 | 5 | 50% | 25% |
| Both incorrect | 5 | 1 | 25% | 5% |
*As coded in the current analysis with respect to consistency with the grammar.